S e v e n t h I n t e r n at i o n a l C o n f e r e n c e o n

AEROSPACE SCIENCE & ENGINEERING

Emerging Technologies in Aerospace Engineering & Geographic Information Science
N at i o n a l C e n t e r o f G I S & S pa c e A p p l i c at i o n s
Institute of Space Technology
Islamabad, Pakistan

ICASE Proceedings

Institute of
Space Technology
Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

PROCEEDINGS
Seventh International Conference on Aerospace
Science and Engineering
December 14 – 16, 2021
Institute of Space Technology
Islamabad, Pakistan

Organized by
National Center of GIS and Space Applications (NCGSA)
Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Editorial Board

Dr. Najam Abbas Naqvi

Conference Secretary
Institute of Space Technology,
Islamabad, Pakistan

Mr. Usama Ahmed

Technical Program Manager
Institute of Space Technology,
Islamabad, Pakistan

Mr. Faaiz Ahmed Jeelani

Business Development Manager
National Center of GIS & Space Applications
Islamabad, Pakistan

Ms. Aqsa Zouq

Manager Research Content
National Center of GIS & Space Applications
Islamabad, Pakistan

Ms. Mahnoor Ajmal

Manager Proceedings
Institute of Space Technology,
Islamabad, Pakistan

International Standard Book Number

E-Copy: 978-969-9448-05-8
Hard Copy: 978-969-9448-06-5

Copyright © 2022
This document is subject to the copy right. All rights are reserved with Institute of Space Technology
(IST), Islamabad, Pakistan. IST can use, publish or reproduce this document in any form under the
copyright law and no part of this document can be reproduced by anyone without the prior written
permission of IST.
Institute of Space Technology (IST), Islamabad, Pakistan.

Seventh International Conference on Aerospace Science & Engineering II

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Preface
Institute of Space Technology (IST), one of the premier institutes in Pakistan, is a federally chartered,
degree awarding Institute offering graduate and baccalaureate programs in Aerospace, Avionics,
Electrical, Mechanical, Materials Science & Engineering, Space Science, Physics, Mathematics, Remote
Sensing & GISc, Astronomy & Astrophysics, Global Navigation Satellite Systems (GNSS) and Applied
Mathematics & Statistics.
National Center of GIS and Space Applications (NCGSA), a project of the Higher Education Commission
(HEC) of Pakistan, led by the Institute of Space Technology is a consortium of seven research laboratories
in seven Higher Education Institutions of Pakistan. NCGSA encompasses endeavors that strive to achieve
and sustain human, technological, research, and entrepreneurial capacities in the areas of space science,
technology, and its applications.
Institute of Space Technology, being the pioneer institute in Aerospace Science & Engineering envisioned
an international forum wherein scientists, researchers, engineers, academicians, private and public industry
professionals, entrepreneurs, and students from all across the globe could interact and share the latest
research trends related with Space Science, Technology and its Applications. Consequently, the first
International Conference on Aerospace Science and Engineering (ICASE) was organized in 2009.
Since its inception ICASE is a regular biennial event of IST which provides a platform to share research
experiences, foster academia-industry linkages, and evaluate the latest findings in the emerging disciplines
of Space Science and Technology.
ICASE 2021 features a diverse blend of ten thematic areas including Aeronautics and Astronautics,
Satellite Technology, Information & Communication Technologies, Positioning, Navigation & Timing,
Geographic Information Science, Remote Sensing & Photogrammetry, Environment & Climate Science,
Applied Physics & Mathematics, Astronomy, Astrophysics, & Astrobiology, Space Law, Management &
Outreach.
A wide variety of sessions including Technical Sessions, Plenary Sessions, Leaders’ Session, Panel
Discussions, Workshops & Trainings, Meet the Scientists Forum, Forum 360: STEM Education, STEP
Stakeholders' Forum, Scientific Networking Session, Science at Glance Session, Emerging Technologies
Summit, Product & Poster Exhibition, Symposiums, GSA Summit and Star Gazing shall engage students
and professionals from all domains of ‘Emerging Technologies in Aerospace Science and Geographic
Information Science’ in the mesmerizing three day conference.
ICASE 2021 will host more than 45 International and National keynote speakers from up to 18 different
countries namely France, Spain, Netherlands, Belgium, Italy, Turkey, Croatia, New Zealand, China, Japan,
Indonesia, Philippines, Saudi Arabia, Egypt, Jordan, Morocco, Latvia and Niger
Owing to COVID 19 and subsequent travel restrictions around the world, ICASE 2021 is planned
intricately to be executed seamlessly in hybrid mode. To ensure that the platforms remain inclusive at all
levels and maximum engagement from around the world is ensured, a number of talks by invited keynote
speakers and guests shall be delivered virtually.

Conference Secretary
January 2022

Seventh International Conference on Aerospace Science & Engineering III

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Table of Contents

1 Conference Organization VII

2 List of Papers Published in IEEE Xplore XX

3 List of Posters XXIV

4 Research Papers 1

Application of Lotus Root Bionic Structure for Increment in Impact Load Bearing
4.1 Capability of a Satellite’s Return Capsule 2
Obaid Ur Rehman, Shahzad Hameed, Zuhaib Hassan
Recent Advancements in Thermal Barrier Coatings (TBC) for High-Temperature Gas
4.2 Turbines 6
Zahid Mehmood, Prof. Dr. Ali Sarosh, Osama Ali Ahmed Awan
Design & Development of Static-Thrust Test-Bench for Aviation Based Piston Engines
4.3 15
Syed Muhmmad Basit Ali, Usama Saleem, Irtiza Ali
Nose Landing Gear Conceptual Design and Analysis for M.A.L.E U.C.A.V
4.4 Asim Mehmood, Gohar Majeed, Kashif Mehmood, Usama Rehman, Muhammad Adeel 19
Khan
Design and Analysis of an Effective Ventilation System to Reduce COVID-19 Spread in A
4.5 Confined Room 26
Salem Mohammed Alohani, Muzaffar Habib, Muhammad Anas
Development of Low Cost Wind Tunnel for Measuring Lift & Drag
4.6 42
Hira Aftab, Beenish Batul
Resolution Improvement in GNSS Based Synthetic Aperture Radar Using GPS L1 Signals
4.7 from Different Look Angles 48
Muhammad Shoaib Ul Hassan, Hassan Sadiq, Najam Abbas Naqvi, Mohammad Usman
Evaluation of Flight Mechanics Performance of a Fighter Aircraft Using
4.8 Computational Fluid Dynamics 52
Asra Imtiaz, Hussna Zainab, Dr. Hossein Raza Hamdani
CFD Analysis of Multi-Element Natural Laminar Flow Air Foil NLF (1)-0115 for A
4.9 Fixed-Wing UAV 56
Absar Ahmed Khan
An Overview of Passive Flow Control Techniques in Supersonic Flows
4.10 66
Zukhruf Liaqat, Aamer Shahzad
Possible Seismo-Ionospheric Anomalies Associated with the Earthquakes Occurred in
2018 Using GNSS TEC: A Statistical Analysis
4.11 71
Muhammad Arqim Adil, Munawar Shah, Asim Khan, Ejaz Nazir, Mohamed Freeshah,
Hamza Bin Babar
Analysis of The Atmospheric Anomalies Associated with Mw>6.0 Continental
4.12 Earthquake in The Afghanistan and Tajikistan 76
Daboor Sultana, Dr. Munawar Shah

Seventh International Conference on Aerospace Science & Engineering IV

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Analysis and Comparison of Maximum Power Point Tracking Algorithms for Electrical
4.13 Power System of the CubeSat’s 79
Usman Azeem, Dr. Rehan Mahmood
Formation Flying Strategy for Attitude Determination and Control of CubeSat
4.14 86
Muhammad Abubakar Arshad, Dr. Rehan Mahmood, Dr. Mohammad Tahir
Overview of ICUBE-N: National CubeSat of Pakistan
4.15 Usman Imran, Muhammad Asfandyaar Arshad, Arooj Shafique, Rehan Mahmood, Rizwan 91
Mughal, Hayat Muhammad Khan
Design and Development of a Portable Dish Antenna for SATCOM Man-Pack Terminal
4.16 96
Shujaat Hussain, Awais Munawar Qureshi, Rizwana Kausar, Mohsin Javaid
Design and Simulation of Cost-Efficient Reaction Wheel for 1U CubeSat
4.17 104
Pawan Kumar, Oscar Lionel
Comparative Study of Image to Image Translation Models for Abnormal Weather Effects
4.18 Removal 112
Maria Siddiqua, Naeem Akhter, Javaid Khurshid
Pick and Place Hand-Gestured-Robot for Physically Challenged People
4.19 117
Muhammad Ihtisham Babar, Muhammad Ibtisam Babar, Azeema Azhar
A Review of Vulnerabilities and Cyber-Security Threats Associated with Overseas
4.20 Telecom Companies’ Operations 124
M Naveed Tahir, Areeba Azhar, Adnan Maqsood
Sentiment Analysis of Social Media Text For Selective World Events And Geographical
4.21 Segregation 132
Muhmmad Ahsan khan
Autonomous Drone Navigation using Deep Convolutional Neural Network
4.22 Muhammad Arif Arshad, Saddam Hussain Khan, Muhammad Waleed Khan, Suleman 140
Qamar, Asifullah Khan
Parameter Identification and Optimization of Permanent Magnet DC Motor using
4.23 Nonlinear Least Square Error Method 146
Raja Muhammad Sohail Khan, Hamza Malik, Hammad Munawar
Threats to the Navigation of Drone Based Delivery Services
4.24 151
Imtiaz Nabi, Salma Zainab Farooq, Sunnyaha Saeed
Investigation of Pre-Seismic Ionospheric and Atmospheric Anomalies before Mw 6.1 El
Salvador Earthquake on October 28, 2018
4.25 156
Abdur Rafeh Abbasi, Abdul Hanan Hamid, Muhammad Arqim Adil, Daboor Sultana,
Saqib Mehdi, Erman Senturk
Uplifting Agriculture through Information and Communication Technology and
Geospatial Applications: A Review of Potential Tools and Techniques
4.26 163
Muhammad Adnan Shahid, Saif Ullah Khan, Subhe Seeme, Muhammad Danish Majeed,
Muhammad Saqlain
Atmospheric Monitoring Using Artificial Intelligence Models & Applications
4.27 168
Muhammad Mubeen, Dr. Shahid Mahmood, Moosa Bin Saif
Unsupervised Classification of Hyperspectral Images using PCA and K-Means
4.28 173
Ayesha Malik, Mamoona Waheed
Implementation Of The Maximum Cross Correlation (MCC) Method In Python For
4.29 Ocean Current Estimation Over The Arabian Sea 177
Sheeza Khalid, Waqas A. Qazi

Seventh International Conference on Aerospace Science & Engineering V

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Landslide Hazard Assessment And Monitoring In Azad Kashmir Using Sentinel-1 InSAR
4.30 Data 181
Labeeqa Zaheer, Waqas A.Qazi
Modeling and Simulation of BSAR (Bi Static Synthetic Aperture Radar) System Using
4.31 Reflected BeiDou B1I Signals 187
Hassan Sadiq, Muhammad Shoaib ul Hassan, Muhammad Usman
Flood Mapping and Crop Damage Assessment using Sentinel-1 SAR GRD and Optical
4.32 Data through Google Earth Engine 191
Tariq Abbas, Waqas A. Qazi
Prototype of an Autonomous Farming Robot: A Smart Farming System
4.33 Moaz Rao, Muhammad Shaarif Younas, Nauman Latif, Ahmad Hammad, Sadia, Ilyas 197
Mahmood Qureshi
Design and CFD Analysis of Outer Aerodynamics of 10kw Horizontal-Axis Wind
Turbine
4.34 202
Noman Ahsan, Saqlain Abbas, Muhammad Ahmad, Ijaz Ali, Zulkarnain Abbas,
Muhammad Umair Shahid
Autonomous Unmanned Aerial Vehicle for Agricultural, Security, Emergency and Other
Uses
4.35 206
Dr. Amir Rashid Chaudhry, Javid Rashid Ch, Dr. Imran Tahir, Tayyaba Shafaq, Imaan
Shahid Raja, Salman Amir, Zohair Anjum, Nabeel Uddin Sheikh
Simulations for Boundary Layer Flows and Heat Transfer of Cross Fluid over Moving
4.36 Flat Plate 213
Mazhar Shehzad, Ammar Mushtaq
Design, Analysis and Fabrication of Composite Curing Compressive Mold for Aerospace
Applications
4.37 220
Muhammad Ayaz Ahmad, Sabih Ahmad Khan, Syed Irtiza Ali Shah, Hamza Akbar Khan,
Taimur Ali Shams
Libration Points of Space
4.38 228
Ms. Hadia Batool, Ms. Sumaira Abbasi
Deinococcus Radiodurans as A Chassis for Synthetic Biology Experiments in Space
4.39 Biology 232
Muneef U. Rehman, Khudeja Salim, Faisal F. Khan
Event Rate of Extreme Mass Ratio Inspiral Systems in M32 Galaxy
4.40 241
Muhammad Waqas, Asad Ali, Rafia Sarwar
Strategizing Pakistan’s Space Policy for Defensive Discourse: A Case Study of Pulwama-
4.41 Balakot Crisis 248
Noor Ul Huda Atif
Multip-GNSS Navigation Performance Assessment and Improvement Quantification
4.42 Arif Hussain, Syed Hadi Hussain Shah, Hina Magsi, Arslan Ahmed, Faheem Akhtar 252
Chachar
GNSS TEC based Ionosphere Anomalies Related to the 2019 Mw 6.2 Thailan Earthquake
4.43 259
Shireen Mushtaq, Rimsha Ashraf, Dr. Munawar Shah, Punyawi Jamjareegulgran
Investigation of Seismo Ionospheric Anomalies Associated with the Mw 6.7 Mongolia
4.44 Earthquake 265
Shireen Mushtaq, Rimsha Ashraf

5 Author Index 273

Seventh International Conference on Aerospace Science & Engineering VI

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Conference Executive Committee

Name Designation Conference Role

Maj. Gen. Rehan Abdul Baqi (H.I)

Vice Chancellor, IST Conference Chair
(Retd.)

Dr. Najam Abbas Naqvi Chairman NCGSA Conference Secretary

Technical Committee
Dr. Suhail Akhtar Dean, IST
Chair

Mr. Mirza Kashif Begg CFO, IST Treasurer

Wg Cdr Ishaat Saboor Khan (Retd.) Registrar, IST Coordination Chair

Engr. Amer Azam Qazi Director Administration, IST Administration Chair

Brig. Zia Ahmed (Retd.) Director Human Resource, IST Human Resource Chair

Conference Facilities
Mr. Rehan Rasheed Director Civil Works
Chair
Conference Audit
Col Amir Najam Riaz (Retd.) Addl. Director Audit, IST
Chair

Mr. Zulfiqar Ahmed Head of Department, IT IT Facilities Chair

Technical Sponsor
Dr. Muhammad Uzair Khan Chairman IEEE Islamabad Section
Chair

Seventh International Conference on Aerospace Science & Engineering VII

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Technical Committee

Name Designation & Affiliation Conference Track

Dean Technical Committee

Dr. Suhail Akhtar
Institute of Space Technology Chair
Head of Department
Dr. Syed Hossein Raza Hamdani
Aeronautics and Astronautics, IST
Professor
Aeronautics &
Dr. Iqbal Rasool Memon Dept. of Aeronautics and
Astronautics
Astronautics, IST
Head of Department
Dr. Abdul Waheed
Avionics Engineering, IST
Head of Department
Dr. Asif Israr
Mechanical Engineering, IST Mechanical Engineering
Professor for Aerospace
Dr. Muhammad Zubair Dept. of Mechanical Engineering, Applications (Aero-Astro)
IST
Head of Department
Dr. Ibrahim Qazi Material Science & Engineering,
IST
Professor Aerospace Materials
Dr. Wilayat Hussain Dept. of Material Science & Design and Engineering
Engineering, IST (Aero-Astro)
Assistant Professor
Dr. Abdul Basit Saim Dept. of Material Science &
Engineering, IST
Professor
Dr. Qamar-ul-Islam Dept. of Electrical Engineering,
IST
Principal Investigator/ Lab Satellite Technology
Director
Dr. Rehan Mahmood
Small Satellite Technology &
Research Lab, NCGSA
Head of Department
Dr. Khurram Khursheed Information &
Electrical Engineering, IST
Communication
Head of Department Technologies
Dr. Altaf Hussain
Computer Science, KICSIT
Professor
Dr. Jamshed Riaz Dept. of Aeronautics and
Astronautics, IST Positioning, Navigation &
Timing
Dr. S. Zaineb Farooq Assistant Professor

Seventh International Conference on Aerospace Science & Engineering VIII

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Dept. of Electrical Engineering,

IST
Assistant Professor
Dr. Farooq Hayder Dept. of Aeronautics and
Astronautics, IST
Dean
Faculty of Information and
Dr. Bakhtiar K. Kasi
Communication Technology,
Geographic Information
BUITEMS
Science
Associate Professor
Dr. Muhammad Shafique National Center of Excellence in
Geology, UoP
Head of Department
Dr. Zahir Ali National Center for Remote
Sensing & Geo Informatics
Chief Technology Officer Remote Sensing &
Dr. Waqas Ahmed Qazi Photogrammetry
Offshore Monitoring Ltd, Cyprus
Associate Professor
Dr. Saleem Ullah
Dept. of Space Science, IST
Principal Investigator
Dr. Zia ul Haq Remote Sensing, GIS and Climatic
Research Lab, NCGSA Environment & Climate
Science
Assistant Professor
Dr. Shahid Iqbal
Dept. of Space Science, IST
Head of Department
Dr. Salman Ahmad Applied Mathematics and
Statistics, IST Applied Physics &
Mathematics
Associate Professor
Dr. Fazeel Mehmood Khan
Dept. of Space Science, IST
Principal Investigator
Dr. Saeeda Sajjad Space & Astrophysics Research
Lab, NCGSA Astronomy, Astrophysics
& Astrobiology
Assistant Professor
Dr. Imran Khan
Dept. of Space Science, IST
Asia Pacific Space Cooperation
Ms. Ayesha Jagirani
Organization, China
Head of Department Space Law, Management
Dr. Ausima Sultan Malik
Humanities Department, IST & Outreach
Assistant Professor
Dr. Ayesha Bukhari
Dept. of Humanities, IST

Seventh International Conference on Aerospace Science & Engineering IX

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

International Scientific Committee

Name Designation Affiliation

King Abdulaziz University, Saudi

Dr. Abdulrahman H. Bajodah Professor
Arabia
The National Authority of Remote
Dr. Tamer Mekky Ahmed Research Associate
Sensing and Space Sciences
Habib Professor
(NARSS), Egypt
Istanbul Technical University (ITU),
Dr. Alim Rustem Aslan Professor
Turkey
Sorbonne Université, Ecole
polytechnique,Institut Polytechnique
Dr. Christine, Marie, Eugenie de Paris, Université Paris Saclay,
Senior Scientist
Amory Observatoire de Paris, CNRS,
Laboratoire de Physique des Plasmas
(LPP), France
Nanjing University of Information
Associate Professor / Science and Technology (NUIST),
Dr. Andres Calabia Aibar
Visiting Professor China Universidad Politécnica de
Madrid (UPM), Spain
Department of Geophysics Services
Advisor of Geophysics & Research, M/s Royal Institute of
Dr. Jean L. Rasson
research Meteorology, B5670, Dourbes,
Belgium

Dr. Mark Van Der Meijde Professor University of Twente, Netherlands

Geophysics, Sekolah Tinggi

Researcher and
Dr. Relly Margiono Meteorologi Klimatologi dan
Lecturer
Geofisika (STMKG), Indonesia
Head of GIS & RS
Dr. Ali Nouh Ali Ma'bdeh Al - Bayt University, Jordan
Section

Dr. Fadia W. Al-Azawi Assistant Professor Al-Karkh University of Science, Iraq

University Abdou Moumouni

Dr. Mounkaila Mohamed Lecturer
Niamey, Niger

Dr. Andris Slavinskis Associate Professor UT Tartu Observatory, Estonia

Space Generation Advisory Council

Mr. Hamza Hameed Team Lead - Legal
(SGAC), Italy
Islamic World Educational, Scientific
Advisor Science &
Dr. Muhammad Sharif and Cultural Organization (ICESCO),
Technology
Morocco

Seventh International Conference on Aerospace Science & Engineering X

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Northwestern Polytechnical
Dr. Chen Jie Associate Professor
University (NPU), China
Technical University of Madrid,
Dr. Gustavo Alonso Rodrigo Professor
Spain
Northwestern Polytechnical
Dr. LIU Cun-liang Professor
University, China
Shanghai Jiao Tong University,
Dr. Shufan Wu Professor
China
Northwestern Polytechnical
Dr. Zhigang Gao Associate Professor
University, China
Northwestern Polytechnical
Dr. Huiqing Fan Professor
University (NPU), China
University of Rijeka & Krapina
Dr. Renato Filjar Professor University of Applied Sciences,
Croatia

Dr. Dinesh Manandhar Associate Professor The University of Tokyo, Japan

Associate Research
Dr. Xiaoliang Wang Shanghai Jiaotong University, China
Fellow
Philippines Space Agency,
Dr. Eng. Ariel C. Blanco Director
Philippines
Remote Sensing
Dr. Salman Ashraf GNS Science, New Zealand
Scientist

Seventh International Conference on Aerospace Science & Engineering XI

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

National Scientific Committee

Name Designation Affiliation

National University of Sciences and

Dr. Adnan Maqsood Associate Professor
Technology, Islamabad, Pakistan
Institute of Avionics and Aeronautics
Dr. Ali Sarosh Associate Professor
(IAA), Air University, Islamabad

Dr. Ali Javed Hashmi Assistant Professor Air University, Islamabad, Pakistan

Dr. Muhammad Kamran University of Central Punjab Lahore,

Associate Professor
Saleem Pakistan
Research Institute for Microwave and
Millimeter-Wave Studies (RIMMS)
Dr. Nosherwan Shoaib Associate Professor National University of Sciences and
Technology (NUST), Islamabad,
Pakistan
School of Electrical Engineering and
Dr. Syed Ali Hassan Associate Professor
Computer Science (SEECS)
FAST - National University of
Dr. M. Uzair Khan Associate Professor
Computer and Emerging Sciences
embedINN Pvt. Ltd., Islamabad,
Dr. Abdul Malik Khan Director
TecSoB LLP, Islamabad
Sukkur IBA University, Sukkur,
Dr. Arslan Ahmed Assistant Professor
Pakistan
Dr. Muhammad Ayyaz Pakistan Space & Upper Atmosphere
General Manager
Ameen Research Commission (SUPARCO)
Punjab Information Technology Board,
Ms. Maria Zubair GIS Specialist
Lahore, Pakistan
Department of Computer Sciences,
Dr. Dostdar Hussain Associate Professor
Karakoram International University
Department of Geological Engineering,
Prof. Dr. M. Farooq Ahmed Professor/Chairman
UET, Lahore
Theoretical Physics Division,
Dr. Nadia Imtiaz Principal Scientist
PINSTECH (PAEC), Pakistan
Head of Environmental
Dr. Salar Ali University of Baltistan, Skardu
Science Department

Dr. Shafqat Ali Assistant Professor GIK Institute, Topi, Swabi, Pakistan

Seventh International Conference on Aerospace Science & Engineering XII

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Department of Computer and

Prof. Dr. Muhammad Ali Information Systems Engineering,
Professor
Ismail NED University of Engineering and
Technology
University of Karachi, Karachi,
Dr. Muhammad Jawed Iqbal Professor
Pakistan
Department of Strategic Studies
Dr. Rubina Waseem Assistant Professor National Defence University,
Islamabad, Pakistan

Seventh International Conference on Aerospace Science & Engineering XIII

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Technical Review Committee

Sr.
Reviewer Name Designation Affiliation / Institution
No.
Institute of Space Technology,
1 Dr. Abdul Faheem Khan Associate Professor
Pakistan
Institute of Space Technology,
2 Dr. Abdul Kabir Assistant Professor
Pakistan
GNSS & Space Education Research
Head of Department and
3 Dr. Abdul Waheed Lab, Institute of Space Technology,
Principal Investigator
Pakistan
Dr. Abrar-Ul-Haq Institute of Space Technology,
4 Associate Professor
Baluch Pakistan
Tenured Associate National University of Sciences and
5 Dr. Adnan Maqsood
Professor Technology, Pakistan
Agricultural Remote Sensing Lab,
Head of Department and
6 Dr. Adnan Shahid University of Agriculture
Principal Investigator
Faisalabad, Pakistan

7 Dr. Ali Saroosh Associate Professor Air University, Islamabad, Pakistan

Institute of Space Technology,

8 Dr. Amir Gilani Associate Professor
Pakistan
Institute of Space Technology,
9 Dr. Anjum Tauqir Professor
Pakistan
Institute of Space Technology,
10 Dr. Asad Ali Associate Professor
Pakistan
Spatial Decision Support System
Lab, Balochistan University of
Dean and Principal
11 Dr. Bakhtiar K. Kasi Information Technology,
Investigator
Engineering and Management
Sciences, Baluchistan, Pakistan
Institute of Space Technology,
12 Dr. Ch. Bilal Ahmad Assistant Professor
Pakistan
Institute of Space Technology,
13 Dr. Danish Ali Sunny Assistant Professor
Pakistan
Institute of Space Technology,
14 Dr. Farrukh Aziz Bhatti Assistant Professor
Pakistan
Dr. Fazeel Mahmood Institute of Space Technology,
15 Associate Professor
Khan Pakistan
Dr. Hayat Muhammad Institute of Space Technology,
16 Assistant Professor
Khan Pakistan

Seventh International Conference on Aerospace Science & Engineering XIV

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Institute of Space Technology,

17 Dr. Ihtzaz Qamar Professor
Pakistan
Institute of Space Technology,
18 Dr. Imran Ali Khan Assistant Professor
Pakistan
Dr. Iqbal Rasool Institute of Space Technology,
19 Professor
Memon Pakistan
Institute of Space Technology,
20 Dr. Irfan Majid Assistant Professor
Pakistan
Professor (Adjunct Institute of Space Technology,
21 Dr. Jamshed Riaz
Faculty) Pakistan
Remote Sensing GIS & Climate
Assistant Professor / Co-
22 Dr. Khalid Mehmood Research Lab, University of Punjab,
Principal Investigator
Lahore, Pakistan
Small Satellite Technology &
Head of Department and
23 Dr. Khurram Khurshid Research Lab, Institute of Space
Co-Principal Investigator
Technology, Pakistan
Institute of Space Technology,
24 Dr. Komal Nain Sukhia Assistant Professor
Pakistan
Institute of Space Technology,
25 Dr. Madiha Tahir Assistant Professor
Pakistan
Institute of Space Technology,
26 Dr. Maryiam Javed Associate Professor
Pakistan
Institute of Space Technology,
27 Dr. Moazam Maqsood Assistant Professor
Pakistan
Dr. Muhammad Abdul Institute of Space Technology,
28 Assistant Professor
Basit Pakistan
Dr. Muhammad Farooq Institute of Space Technology,
29 Assistant Professor
Haydar Pakistan
Institute of Space Technology,
30 Dr. Muhammad Naeem Professor
Pakistan
Institute of Space Technology,
31 Dr. Muhammad Nawaz Associate Professor
Pakistan
GIS & Space Applications in
Dr. Muhammad Associate Professor /
32 Geosciences Lab, University of
Shafique Principal Investigator
Peshawar, Pakistan
Institute of Space Technology,
33 Dr. Muhammad Shakir Assistant Professor
Pakistan
Institute of Space Technology,
34 Dr. Muhammad Umer Assistant Professor
Pakistan
Institute of Space Technology,
35 Dr. Muhammad Usman Assistant Professor
Pakistan

Seventh International Conference on Aerospace Science & Engineering XV

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Senior Assistant Professor

36 Dr. Muhsan Ehsan Bahria University, Pakistan
& Program Coordinator
Institute of Space Technology,
37 Dr. Munawar Shah Assistant Professor
Pakistan
Pakistan Institute of Nuclear
38 Dr. Nadia Imtiaz Principal Scientist
Science & Technology, Pakistan
National Center of GIS & Space
Head of Department /
39 Dr. Najam Abbas Naqvi Applications, Institute of Space
Chairman
Technology, Pakistan
Institute of Space Technology,
40 Dr. Naseem Ahmad Assistant Professor
Pakistan
Assistant Professor / Head Prime Institute of Health Sciences,
41 Dr. Nazia Khatoon
of Department Islamabad, Pakistan
National University of Sciences and
42 Dr. Nosherwan Shoaib Associate Professor
Technology, Pakistan
Institute of Space Technology,
43 Dr. Qamar ul Islam Professor
Pakistan
Spatial Decision Support System
Director ORIC & Co- Lab, Sardar Bahadur Khan
44 Dr. Rabia Zafar
Principal Investigator Women's University, Quetta,
Pakistan
Institute of Space Technology,
45 Dr. Raees Fida Swati Assistant Professor
Pakistan
Institute of Space Technology,
46 Dr. Rahila Naz Associate Professor
Pakistan
Small Satellite Technology &
Assistant Professor &
47 Dr. Rehan Mahmood Research Lab, Institute of Space
Principal Investigator
Technology, Pakistan
Institute of Space Technology,
48 Dr. Rizwan Mughal Associate Professor
Pakistan
Institute of Space Technology,
49 Dr. Sadia Saeed Assistant Professor
Pakistan
Institute of Space Technology,
50 Dr. Sadia Shakil Assistant Professor
Pakistan
Assistant Professor / Institute of Space Technology,
51 Dr. Saeeda Sajjad
Principal Investigator Pakistan
Institute of Space Technology,
52 Dr. Sajid Ghaffar Assistant Professor
Pakistan
Institute of Space Technology,
53 Dr. Saleem Ullah Associate Professor
Pakistan

Seventh International Conference on Aerospace Science & Engineering XVI

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Dr. Salma Zainab Institute of Space Technology,

54 Assistant Professor
Farooq Pakistan
Institute of Space Technology,
55 Dr. Salman Ahmad Head of Department
Pakistan
National University of Sciences and
56 Dr. Salman Atif Assistant Professor
Technology, Pakistan
GIS & Space Applications in
Head of Department / Co- Geosciences Lab, Shaheed BB
57 Dr. Sami Ullah Dawar
Principal Investigator University Sheringal, Dir Upper,
Pakistan
Institute of Space Technology,
58 Dr. Suhail Akhtar Dean
Pakistan
Institute of Space Technology,
59 Dr. Syed Ali Irteza Assistant Professor
Pakistan
Dr. Syed Hossein Raza Institute of Space Technology,
60 Head of Department
Hamdani Pakistan
Dr. Syed Wilayat Institute of Space Technology,
61 Professor
Husain Pakistan
Institute of Space Technology,
62 Dr. Talha Irfan Khan Assistant Professor
Pakistan
Institute of Space Technology,
63 Dr. Umair Ali Assistant Professor
Pakistan
Institute of Space Technology,
64 Dr. Umer Sohail Assistant Professor
Pakistan
Institute of Space Technology,
65 Dr. Waqas Ahmed Qazi Associate Professor
Pakistan
Institute of Space Technology,
66 Dr. Waqas Ashraf Associate Professor
Pakistan
Remote Sensing GIS & Climate
Associate Professor /
67 Dr. Zia ul Haq Research Lab, University of Punjab,
Principal Investigator
Lahore, Pakistan
Sukkur Institute of Business
68 Dr. Arslan Assistant Professor
Administration, Pakistan
National Engineering and Scientific
69 Mr. Danial Amin Assistant Manager
Commission, Pakistan
Mr. Izhar Hussain Institute of Space Technology,
70 Assistant Professor
Kazmi Pakistan
Mr. Muhammad Institute of Space Technology,
71 Lecturer
Nadeem Khan Pakistan
Institute of Space Technology,
72 Mr. Saad Rifhat Qureshi Lecturer
Pakistan

Seventh International Conference on Aerospace Science & Engineering XVII

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Institute of Space Technology,

73 Ms. Saima Siddiqui Assistant Professor
Pakistan
National Engineering and Scientific
74 Mr. Shakaib Ahmed General Manager
Commission, Pakistan
Institute of Space Technology,
75 Ms. Aneeqa Abrar Lecturer
Pakistan
National University of Sciences and
76 Ms. Maria Mehmood PhD Scholar
Technology, Pakistan
National Center of GIS & Space
Mr. Muhammad Aqib Research Associate / Team
77 Applications, Institute of Space
Khan Lead
Technology, Pakistan

Seventh International Conference on Aerospace Science & Engineering XVIII

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Conference Management Team

Sr. No. Name Conference Role

1. Dr. Najam Abbas Naqvi Secretary ICASE 2021

2. Dr. Abdul Waheed Manager Technical Review Committee

3. Mr. Usama Ahmed Technical Program Manager

4. Mr. Faaiz Ahmed Jeelani Public Relations Manager

5. Mr. Nouman Sheikh Procurement Manager

6. Mr. Aqib Khan Technical Sessions Manager

7. Ms. Aqsa Zouq Paper Review & Research Content Manager

8. Mr. Faran Ali Operations Manager

9. Ms. Mahnoor Ajmal Workshops Manager

10. Ms. Faiza Badar Graphics Manager

11. Mr. Muhammad Afnan Web Manager

12. Mr. Sohaib Ahmed Protocol Manager

13. Mr. Oscar Lionel Hospitality Manager

14. Ms. Ayesha Malik Registrations Manager

15. Mr. Amir Faizan Malik Content Writing Manager

16. Mr. Danial Raza Kazmi Forums Manager

Seventh International Conference on Aerospace Science & Engineering XIX

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

List of Papers Published in IEEE Xplore

International Standard Book Number
Xplore Compliant: 978-1-6654-8319-3
Print: 978-1-6654-8318-6

Sr.
Author Name & Affiliation Research Title
No.
Dr. Sami Ullah
Time Series Analysis of Bindi Dheraja Riverine Forest of
1 Shaheed Benazir Bhutto Sukkar, Sindh, Pakistan using Remote Sensing
University, Pakistan
Khadeejah Motiwala
Gamma-ray Burst High-latitude Emission: Simulating the
2 Institute of Space Technology, Propagation Effect
Pakistan
Yusha Anis
Developing Heat Index for Heat Wave Hazard Mapping:
3 Pakistan Space & Upper
Atmosphere Research A Case Study of Sindh Province, Pakistan
Commission, Pakistan
Ali Hanif
Deep Learning Based Radar Target Classification Using
4 National University of Sciences Micro-Doppler Features
and Technology
Syed Roshaan Ali Shah

5 Pakistan Space & Upper Deep Learning on High Spatial and Temporal Cadence
Atmosphere Research Satellite Imagery for Field Boundary Delineation
Commission, Pakistan
Muhammad Asim Mustafa
Using Optical Remote Sensing and Radar Altimeter Data
6
Institute of Space Technology, for Lake Volume Estimation of Manchar Lake, Pakistan
Pakistan
Muhammad Ihtisham Babar
Robust Non-Linear Dynamic Inversion Control System To
7 Pakistan Aeronautical Complex, Improve Stability Of Large Aircraft In Crosswind
Pakistan
Muhammad Ihtisham Babar PanAir Study of Variation in Canard Location and
8 Pakistan Aeronautical Complex, Dihedral on Lift Characteristics of a Close-Coupled Wing-
Canard Configuration
Pakistan
9 Afaq Ahmed

Seventh International Conference on Aerospace Science & Engineering XX

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Bahria University Islamabad, Blockchain Based RFID Prototype For Medicine

Pakistan Traceability along the Healthcare Supply Chain
Asim Mehmood
Main Landing Gear Conceptual Design and Analysis for
10 Institute of Space Technology, M.A.L.E U.C.A.V
Pakistan
Dr. Awais Munawar Qureshi
Design and Development of a Vivaldi Antenna Array for
11 National University of Sciences Airborne X-Band Applications
and Technology, Pakistan
Hadiqa Khan
Probabilistic Assessment of Storm Surge Hazard along
12 Pakistan Space & Upper
Atmosphere Research Coastal Regions of Sindh Province, Pakistan
Commission, Pakistan
Hina Magsi
Investigation of Ionospheric Scintillation using BeiDou
13 Sukkur Institute of Business during Geomagnetic and Non-geomagnetic Storm
Administration University, Conditions Over Pakistan
Pakistan
Obeid Muhammad Usmani Mathematical Modelling and Analysis of Submarine
14
Propeller Shaft
Air University, Pakistan

Zarrar Haider Control System Mathematical Model Design of

15 National University of Sciences Ornithopter Flapping Wing Unmanned Air Vehicle for
Agile Performance
and Technology, Pakistan
Muzamil Ali
Design and Development of a High Precision Thrust
16 Pakistan Institute of Engineering Measurement System
and Applied Sciences, Pakistan
Danial Ahmed Mir
17 Modern and Classical Pitch Control for Yak54
Institute of Space Technology,
Pakistan
Nadia Jabeen
Detection of Thermal Fronts in the Arabian sea through
18 Institute of Space Technology, SAR (Synthetic Aperture Radar) imagery
Pakistan
Muhammad Shahzad Anjum
19 The Spectral Parameters of Gamma-Ray Blazars
Xiamen University, Malaysia

20 Zohaib Altaf

Seventh International Conference on Aerospace Science & Engineering XXI

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Aerodynamic Shape Optimization of Doubly Offset

Air University, Pakistan
Serpentine Diffuser using Response Surface Methodology
Sadia Azhar
21 Modeling and Analysis of a Vibratory Bowl Feeder
Air University, Pakistan

Sadia Maryam

22 National University of Sciences Simulations of Thin Gray Fluid Over a Moving Flat Plate
and Technology, Islamabad,
Pakistan
Fahad Sagheer
Investigations of mechanical properties of Sn-3.5Ag-0.5La
23
University of Engineering and lead-free solder alloy for green electronics
Technology, Peshawar, Pakistan
Muzaffar Habib
Autonomous control of a Quadcopter using Machine
24
National University of Sciences Learning algorithm
and Technology, Pakistan
Obaid-ur-Rehman
UAV Assessment of Crop Evapo-transpiration Dynamics
25 Pakistan Space & Upper in Winter Wheat and Barley under Varying Pressures of
Atmosphere Research Fungal Diseases
Commission, Pakistan
Syed Saddam ul Hassan Generalised Modelling of Sound Signatures for
26 Characterization of Multi-copter Unmanned Air Vehicles
Air University, Pakistan based on Aero-acoustics Measurements and CFD Analysis

Jabir Shabbir Malik Analysis and Estimation of Zenith Wet Delay and Zenith
Beijing Institute of Technology, Tropospheric Total Delay at Earth’s different
27
Geographical Areas
China
Dr. Muhammad Shahid Iqbal
Climate Change an Emerging Risk: A Case Study of Sutlej
28 Institute of Space Technology, River Basin, Pakistan
Pakistan
Sheher Bano Design and implementation of dual-axis solar tracking
29
system for maximizing the efficiency of solar cells
Air University, Pakistan

Muhammad Taimur Khan

Storm time ionospheric variations from GNSS TEC and
30 Institute of Space Technology, Swarm satellites
Pakistan
31 Amna Hafeez

Seventh International Conference on Aerospace Science & Engineering XXII

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Machine learning based thermal anomaly detection

Institute of Space Technology,
associated with three earthquakes in Pakistan using
Pakistan
MODIS LST
Muhammad Aitessam Ahmed
Design and Development of Audio Processing and Speech
32
Institute of Space Technology, Recognition Algorithm
Pakistan
Hoor Bano
Attitude Estimation & Control of a CubeSat Using Linear
33
Institute of Space Technology, Quadratic Gaussian Approach
Pakistan
Syed Riaz un Nabi Jafri
Campus Terrain Surveying and Mapping using Low Range
34 NED University of Engineering 2D Laser Scanners
and Technology, Pakistan
Khizar Ahmad Baqai Review on “ADVANCE SEAPLANE CONCEPTUAL
35 DESIGN ADAPTING TRIMARAN BOAT HULL
Superior University, Pakistan CONCEPT

Muhammad Asfandyar Arshad

Small Satellite Technology
36 Research Lab, National Center Dynamic Analysis of 3U deployer (iPOD) Structure
of GIS & Space Applications,
Pakistan
Maria Mehmood
Performance Analysis of global ionosphere/TEC models
37 National University of Sciences over Islamabad
and Technology, Pakistan
Syeda Saleha Fatim Ali
Geospatial Application of Groundwater Quality Index for
38
Lahore University of Lahore Metropolitan, Pakistan: A Retrospective Approach
Management Sciences, Pakistan

Seventh International Conference on Aerospace Science & Engineering XXIII

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

List of Posters
Sr.
Author Name & Affiliation Research Title
No.
Jahanzeb Tariq Khan
Impact of Sampling Delays and Communication
1
National University of Sciences and Breakdown on the Stability of Spacecraft Formation
Technology, Pakistan
Asim Khan
GPS TEC anomalies associated with Mw>7
2
Institute of Space Technology, earthquakes in Japan
Pakistan
Sana Asif
3 Design of low cost AIS receiver for ICUBE-N
Institute of Space Technology,
Pakistan
Dr. Azhar Imran Mudassir
Rise of Educational Software and Impact of COVID-
4
19 Outbreak in Pakistan
Air University, Islamabad, Pakistan

Masood Bari
Find The Shortest Route With The Help Of Network
5
Analysis
University Of Karachi, Pakistan

Muhammad Abdul Moiz Computational Study of Fe-Doped ZnO as a

6 Photocatalyst to Reduce the Environmental Pollution
Institute of Space Technology,
from Textile Industry
Pakistan
Haniya Usmani
7 The Perks Of Space Exploration
Generation's School, Pakistan

Quratulain Danish
Network analysis of DEGs with putative roles in
8
CECOS University of IT & neuronal development under microgravity conditions
Emerging Sciences, Pakistan

Seventh International Conference on Aerospace Science & Engineering XXIV

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Research Papers

Seventh International Conference on Aerospace Science & Engineering 1

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Application of Lotus Root Bionic Structure for

Increment in Impact Load Bearing Capability Of a
Satellite’s Return Capsule
Obaid Ur Rehman Shahzad Hameed Zuhaib Hassan
School of Astronautics,Beihang School of Astronautics, Beihang School of Astronautics,Beihang
University University University
Beijing, China Beijing, China Beijing, China
[email protected] [email protected] [email protected]

Abstract— Nature has been astonishingly inspired man to has to bear the ground or water impact load but also the high
look around and learn in various fields since ages. Lotus root g loads. Reentry speed typically is higher than the First
design, being one of the famous inspiring techniques in recent cosmic velocity of earth which is also termed as orbital
era has been studied and applied to the structure of a Satellite’s velocity. Space return capsules generally are blunt-bodies
Reentry capsule. The paper presents a study related to
without wings that re-enters the Earth's atmosphere. The
improvement in the impact stress bearing capability in structure
that strikes and lands on earth with a relatively high velocity primary difference between a Satellite and Reentry Capsules
despite the use of parachute. For achieving better structural is the capability to survive reentry and return a payload to the
designs, bionics have been playing a significant role in almost Earth's surface from space. Return vehicles can broadly be
every field. This paper presents a stress reduction of 10.18 % by classified under two different subheads that are manned and
the application of using a lotus root concept in the Return unmanned capsules. If viewed in a general perspective they
Capsule. Moreover the distribution of stresses in the cross- both have several similarities and commonalities however
sectional area of Reentry vehicle was analyzed and presented they also differ in many sub-systems and technicalities. A
crewed space capsule necessarily has to attain the ability to
Keywords--Bionic structure, Re-entry capsule, FEM Analysis
sustain life in the vacuum, often with a harsh and extremely
I. INTRODUCTION demanding thermal and radiation environment. Many
subsystems are required to accomplish this task including
Exploration of space has helped man to widen its horizon.
reaction control system, LSE (life saving and support
Human exploration has now surpassed the earth’s close space
equipment), weightlessness restraint seats for astronauts,
and progressively motivated researchers to look forward into
Docking Mechanisms, Communication antennas, Launch and
the Deep Space. Exploration methods have slowly evolved
abort systems Parachute and the list goes on. However the
from initial remote sensing explorations to landing
unmanned capsule comparatively has a very less complex
expeditions. One of the major purposes is to approach the
system as the life saving and much other equipment can be
extra-terrestrial revolving bodies, collect sample and return
easily omitted from the unmanned one. Current examples of
them back for further deep advance researches. NASA and
crewed capsules include Soyuz, Shenzhou, Orion, CST-100,
USSR’s space programs have been working on their space
and Dragon 2. The unmanned capsules are Hayabusa Muses-
explorations since back in 1950’s and 1960’s however many
C, Stardust, Genesis and many others.
other countries have joined upon in 90’s and second
millennium like the Chinese CNSA and Japanese JAXA.

Light Reentry Vehicles have now become an

indispensable part of deep space and celestial world
exploration due to their imperative capabilities and
characteristics. As a general process it can be summarized
that a satellite is deployed in space to perform a landing on
another planet or a designated comet and after the successful
collection of sample it returns back to earth which itself is a
complex, but vital procedure which can cause catastrophic
failures if not planned and executed appropriately. An Fig. 1. Nominal Stardust capsule entry sequence
extensive in depth study and a broad range of knowledge of
reentry technology deems necessary for carrying sample from EDL process can be distributed into 4 major steps as
the interplanetary space [1]. After the releasing of Space 1. Separation and Spinning of capsule
capsule till the time it lands back to earth, is just a few hours 2. Atmospheric Reentry
job generally termed as EDL (Entry Descent and Landing) 3. Descend
procedure [2]. Figure 1 briefly explains the procedure of EDL 4. Landing
procedure of Stardust Capsule. Despite bearing the high Un-manned Space capsules have typically been about 1
Reentry heat, during the EDL process the structure not only meters or lesser in diameter and the manned can be up to 5

Seventh International Conference on Aerospace Science & Engineering 2

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

meter or more. The unmanned Reentry capsule design has to verified that presence of oval holes in lotus roots causes the
be structurally strong as well as volumetrically efficient, so it huge amount of compressive stresses available in cross
is typically feasible to create small capsules having sectional area occurred at the ends of major axis of the actual
performance which is comparable to lifting body in all but holes with in the lotus root. This was a very important
lift-to-drag ratio for less cost. Space capsules are built in conclusion because once the oval holes were re-simulated as
manner so they are well suited to reentries possessing high circular ones the location of maximum compressive stress
energy. changed to the location where short axis of oval holes was
present.
The entry velocities of space capsules have considerably
high values. To reduce this high velocity, the schematic shape Architectural bionics improves the appearance of modern
of capsules are designed such to generate drag which in turn structures. It works on the basis of pristine specimens of flora
slows the capsule. Table 1 shows the comparison of three and fauna, which are used to create innovative patterns. A
famous return capsules Hayabusa, Genesis and Stardust [3]. few major trends in the structural evolution due to field
of bionics are: bio based materials, biomechanics, semi-
TABLE I. DATA COMPARISION OF STARDUST,GENESIS AND bionic architectures, bionic urbanism, green-architecture,
HAYABUSA REENTRY CAPSULES
bionic architectural ecology, architectural transplantism, and
Description Stardust Genesis Hayabusa various others. They permeate and enhance one another [6].
Reentry initial
12.5 10.7 12.04
Velocity (km/s)
Reentry capsule mass II. METHODOLOGY
46 200 18
(kg)
Reentry angle A. Generic return capsule structure (Case-1)
-8.2 -10 -12.35
(degree)
Spin velocity (rpm) 15 15 2 The main structure of a capsule contains a sample
Parachute opening container, parachute bay, electronic bay and a heat ablative
1.4 1.8 0.8
Mach No shield on the outer body. The aft end of a capsule enters first,
Diameter big bottom interacting and impacting the atmosphere thus withstands the
0.811 1.52 0.4
(m)
Radius of curvature g-forces induced. The rounded shape typically used for aft
of spherical segment 0.22 0.44 0.2 end also known as the blunt body of a capsule forms a shock
with large base (m) wave that keeps most of the heat away from the body, though
Peak Reentry Heat 11.9 to a thermal protective system is still necessary thus ablative
7.5 11
flow (MW/m2) 12.6
Outer layer
shields are used for it. Space capsule must reenter at a precise
Front end thermal Carbon- angle of attack to prevent a skip off the surface of the
C/C and
protection material PICA Phenolic
type
inner is
material
atmosphere of destructively high accelerations. While
C/fiber Descend phase, as shown in step 3 of Figure 1, the ablative
Thickness of Thermal
shield separates from the structure after its desired results
Protective Layer 58.4 60 30
(mm) within Reentry phase (step 2 of figure 1), thus the remaining
Parachute structure impacts on the ground. Remaining structure
Type of decelerating Tapered
parachute
with
umbrella
Nothing contains the parachute attached to a sample container having
umbrella a structural layer that strikes the ground.
Main parachute Tapered Cross
Parachute
Umbrella type umbrella umbrella

Despite using the blunt shape and making use of

parachutes the velocity of capsules remain still on the higher
side at about 8m/s at the touch down [3]. The structure bears
a shear impact load when it collides with the ground or
splashes in water. To cater for this load, this paper presents a
new innovative scheme by using the nature inspired structure
of a lotus root.
Fig. 2. Re-entry capsule model
Since life is estimated to have appeared on earth,
nature has gone through extreme evolution over a huge span Taking leads from the three aforementioned Return
of 3.8 Gyr [4]. Nature has forced evolution of objects with Capsules namely Stardust, Genesis and Hayabusa MUSES-
high performance using commonly found materials C, a generic structure of a capsule was modeled in ANSYS
functioning in both the macro and Nano-scale domains. [7]. The model was of the dimension 200mm by 200 mm and
Developing better understanding of the functions provided by weighs 6.382 kgs. Figure 2 shows a model in ANSYS
matter and processes evolved in nature can help mankind to software which wherein an entry capsule is visible before the
imitate and produce optimized structures, devices and impact onto the ground. The surface beneath the capsule has
processes. ‘Bio-mimetic’ is referred to a biologically inspired been modeled as sand for realistic approach.
design derived from nature. During research, the author could
not find any study in which the lotus roots bio-metric has been Since by the time the capsule reaches the earth’s
used in the field of return capsule's structure. surface the ablative material has already been utilized and
Yiyuan Zu [5] has discussed Applications and Engineering detached from the capsule, hence the material used for
Analysis of Lotus Roots and their advantages. The study simulation was considered Aluminum AL 2024-T4 and the
ground was modeled as SAND. The material selected for the

Seventh International Conference on Aerospace Science & Engineering 3

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

capsule i.e. AL 2024-T4 has density of 2785 kg/m3, its shear into ANSYS on the model of Reentry capsule as shown below
modulus is 2.86x1010 Pa and the Poisson ratio is 0.33. in Figure 5.

Fig. 3. Reentry Model Y-Z plane cut section (Case-1) Fig. 5. Re-entry capsule with Lotus root base(Case-2)

Figure 3 depicts an inside view of the reentry model The reduction of this mass in a bionic pattern causes
by cutting a section in Y-Z plane. This section shows an the weight reduction in capsule. The weight is reduced by
internal closed chamber which contains the extra-terrestrial 2.522% and the final capsule weighs 6.221 kgs.
materials hence the basic aim of our research is to make sure
that this inside chamber remains safe even after the impact. D. Von Mises Stress
Conservation of momentum and energy is an
B. Structure of a lotus root inevitable part of the physics of impact. The force which tends
Lightening holes have been purposefully used in to decelerate a mass of a moving object when it strikes
aerospace industry since ages to lighten the structure yet another structure, the force satisfies conservation of
maintain its structural integrity. Lotus root can be seen to have momentum. The impacting body’s kinetic energy will be
the same principle applied in it. Understanding the internal partially converted to strain energy inside the target and will
oval holes containing structure of a lotus root helps us to partly be dissipated through friction and local plastic
design a better structure. Figure 4(a) demonstrates a fresh deformation however the strain energy is radiated away as
lotus root sliced carefully to visualize the holes. stress waves. To verify that the model sustains the impact load
of structure and ground the Von-Mises stress criterion will be
used. This criterion is also termed as maximum distortion
energy criterion is associated with the change in shape of the
material and is normally applied to ductile materials like
metals.

(𝝈𝟏𝟏 −𝝈𝟐𝟐 )𝟐 −(𝝈𝟐𝟐 −𝝈𝟑𝟑 )𝟐 −(𝝈𝟑𝟑 −𝝈𝟏𝟏 )𝟐 +𝟔(𝝈𝟏𝟐 𝟐 +𝝈𝟐𝟑 𝟐 +𝝈𝟑𝟏 𝟐 )

𝝈𝒗 𝟐 = √
𝟐


Where σv is the Von-Mises stress as σ11 , σ22 , σ33 , σ12 , σ23

and σ31 are the six components of stress tensor. Yielding
occurs when the equivalent stress reaches the yield strength
Fig. 4. (a) Lotus root internal structure (b) Lotus root points translated in of the material.
Ansys software
III. RESULTS AND DISCUSSIONS
Structure of lotus root is quite complex as evidently
A. Impact Analysis setting
visible in the figure above, the sliced lotus root had 9 larger
oval holes that were translated in ANSYS software as shown
in Figure 4(b). To guarantee an acceptable degree of A short time duration impulse event was generated.
consistency between the model and actual structure and The analysis was carried for 0.01 second where in the capsule
effectively improve the efficiency of FEM analysis, the hit the ground. Finite element model cut in plane to view the
model of lotus root was simplified. cross-sectional area of the capsule and the meshes are shown
in Figure 6. An automatic generated mesh was used by means
The smaller holes were deliberately disregarded as of ANSYS Workbench meshing tool. Areas where high
the manufacturing process of smaller ones would be difficult. precision was a requirement, the meshing tool automatically
These holes were then translated on the base of space capsule adjusted the mesh density so that a high quality based mesh
already modeled in Section 2.1. The lotus root structure has is guaranteed [8]. The structure of Case 1 and Case 2 had
already been studied in detail by Yiyuan Zhu [5] and its 2D different numbers of mesh and nodes due to their different
structural analysis also proved that shape of holes in the lotus requirements and are shown in figure 6(a) and figure 6(b)
root causes significant changes in the stress states thus respectively. The minimum size of auto generated mesh is
inspiring creation of new structures and materials. relatively smaller vis a vis the overall size of the capsule thus
the requirement of calculation accuracy is met. ANSYS
C. Lotus root bionic structure application in Return Explicit is used to solve the impact of Reentry capsule.
Capsule structure (Case-2)
As a new scheme the lotus root sketch was
duplicated by the extraction of points and the scaled down

Seventh International Conference on Aerospace Science & Engineering 4

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

the stress concentration in container containing extra-

terrestrial material.

Fig. 6. (a)Mesh Case-1 (b) Mesh Case-2

For the boundary conditions the gravity was applied

in –Z direction i.e. directing towards the ground. A velocity
of 8000mm/s (8m/s) was imparted to the Return capsule. The
sand block was fixed from remaining 5 surfaces so as to Fig. 9. Lotus root section on base of Return capsule
simulate an area in which the capsule lands.
IV. CONCLUSIONS
B. Von-Mises Stress comparison By analyzing the FEM results of the model, it can be
The solution for equivalent Von-Mises stress as per concluded that the structural design of the capsule meets the
ANSYS solver was analyzed and the values were compared essential requirements of strength, no structural damage will
to verify that model of reentry vehicle will be able to sustain occur because the high stress zones are focused on the lotus
the impact loads. The von-Mises maximum value was found root edges. Because of the oval holes of lotus root, the largest
to be 41.067 MPa for case-1 and 122.53 MPa for case 2. Both stress in the cross-sectional area of structure occurred near the
values were lesser than the yielding stress value of Aluminum lotus root hole as evident in Figure 9. Major conclusions are
AL 2024-T4 which is 255 MPa thus conforming that the as follows:
structure remains safe in both cases.
a) In both cases, the maximum stress value remained
lesser than the yielding stress value.
b) A 10.18 % stress reduction was observed in Case-2
due to the bionic lotus root structure.
V. REFERENCES
[1] Yoshifumi Inatani, “Reentry Capsule for Sample Return from
Asteroids in the Planetary Exploration Missions” J. Phys.: Conf. Ser.
(a) Nephogram:Case-1(generic structure)
1005 012049(2018)
[2] Prasun N. Desai, Dan T. Lyons, “Entry, Descent, and Landing
Operations Analysis for the Stardust Re-Entry Capsule”, NASA
Langley Research Center, Hampton, VA, 23681-2199, Jet Propulsion
Laboratory, Pasadena, CA 91109-8099
[3] GENG Yunfei, GAOJie, LUYadong, LiQi “Demostration of Light
Reentry Vehicle Development” Journal Spacecraft Engineering, Vol 23
No 6, Page 99 (2014)
[4] B. Bhushan, “Bio-mimetics: Lessons from nature an overview,”
(b) Nephogram:Case-2 (lotus root bionic structure) Philosophical Transactions of the Royal Society A: Mathematical,
Physical and Engineering Sciences, vol. 367, no. 1893, pp. 1445–1486,
Fig. 7. (a) & (b) Von-Mises stress Nephograms (2009).
[5] Yiyun Zhu, Chang Jiang Wang and Diane Mynors, “Applications and
Engineering Analysis of Lotus Roots under External Water Pressure”,
The basic aim of this paper is to reduce the stress on Hindawi Publishing Corporation, Applied Bionics and Biomechanics,
the sample container by application of Lotus root bionic Article ID 2386982 (2016)
structure application. Comparing both capsule containers in [6] Irina Mayatskaya, Victor Eremin, “Bionics and the choice of rational
case 1 and 2 a maximum value of 41.067 MPa and 36.886 structural form” Volume 110, 2019, International Science Conference
MPa were calculated respectively which makes a significant SPbWOSCE-2018, E3S Web of Conferences, Article No 01042 (2019)
stress reduction of 10.18%. [7] ANSYS Software, ANSYS Academic Research Release 18.0. ANSYS
Inc., USA.
[8] Mou WJ, Yang JZ, Chen JY, et al. Finite element analysis for power
car skeleton based on ANSYS Workbench[J]. Science Technology and
Engineering, 22:5592-5594 (2010)

Container:Case-1 (b) Container:Case-2

Fig. 8. Stress Nephograms of Re-entry vehicle’s container

The basic reason of this stress management can be

explained by closely examining the lotus root section in
Figure 8 as elaborated in Figure 9. This special region
distributes the stress vectors in such a way that the stress
values are higher towards the lotus root holes thus reducing

Seventh International Conference on Aerospace Science & Engineering 5

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Recent Advancements in Thermal Barrier Coatings

(TBC) for High-Temperature Gas Turbines
Zahid Mehmood Prof. Dr. Ali Sarosh Osama Ali Ahmed Awan
Department of Aerospace Engineering Department of Mechanical and Department of Mechanical and
CAE, NUST Aerospace engineering Aerospace engineering
H-9 Islamabad, Pakistan Air University Air University
[email protected] E-9 Islamabad, Pakistan E-9 Islamabad, Pakistan
[email protected] [email protected]

Abstract— Gas turbines can produce a tremendous metal on the TBCs (Thermal Barrier Coatings) and sweeps
amount of energy and have wide industrial applications down to substrate material [2, 3]. HTHC is due to Na2SO4
due to their small size and lower weight. Their efficiency (sodium sulfate), NaCl (sodium chloride), and V2O4
depends on two factors i.e., turbine inlet temperature
(Vanadium Oxide). The impurities in fuel can lower the
and thermodynamic efficiencies. By increasing the
turbine inlet temperature (TIT), turbine blades face melting point of Na2SO4, moreover, during combustion
high-temperature hot corrosion (850 ℃ to 950 ℃) that NaCl can form Na2SO4 by reacting with SO2 (sulfur dioxide)
limits the life of the turbine blade. During hot corrosion, and oxygen as shown by the following reaction [4]. Presence
molten sodium salts and vanadium oxide condense over of Vanadium accelerates the solubility of turbine blade
the surface of turbine blades, consume protective oxide protective oxide in Na2SO4 [5]. It starts by breaking the
layers, and diffuse into the microstructure of superalloy. oxide layer, followed by depletion of chromium content of
As a result, fatigue, and creep phenomena get
substrate material, substrate oxidation loss of material
accelerated leading to fracture. A survey of past few
years shows that the use of thermal barrier coatings integrity, which finally leads to blade failure along with
(TBC) has proven to be an effective strategy to avoid material strength.
premature failure. In this paper, an extended survey has
been presented about the improvement of hot corrosion 2NaCl (l) + SO2 (g) + O2 (g) → Na2SO4 (l) + Cl2 (g)
resistance by TBCs by reinforcement with
Nanoparticles, rare earth elements, gradient aluminum
layers, and metallic bond coatings. The performance B) Type-2: LTHC (low temperature hot corrosion)
parameters of different reinforcements have been
evaluated by XRD techniques, SEM spectroscopy, and In the subsequent type, hot corrosion is studied for the
corrosion kinetics curves. temperature limit from 650 to 800 ℃ [6]. A high partial
pressure SO3 (sulfur trioxide) formed as a byproduct of
Keywords— High-temperature hot corrosion, low- combustion leads to LTHC Reaction.
temperature hot corrosion, fatigue life, creep rate, thermal
barrier coatings, nanostructured coatings, rare earth elements,
gradient coatings Research has shown that HTHC effects the life of the
turbine blade. This paper presents a survey of the
I. INTRODUCTION performance of Nano-structured TBCs, rare earth elements,
A gas turbine is used to produce a tremendous amount of and multilayer TBCs against HTHC. It has been found in
energy and has many useful applications due to their smaller recent research that the additions of Nano-structured TBCs,
size and weight. Its efficiency depends on two factors: i.e., rare earth elements, and multilayer TBCs provide better
higher turbine inlet temperature and efficiency of performance as compared to the conventional coatings. This
compressor and turbine. The increase in inlet temperature is study has employed XRD, SEM, and corrosion kinetic
driven by improvement in turbine blade metallurgy, curves to evaluate several types of coatings.
coatings, and cooling techniques. Turbine efficiency and
blade life are enhanced by hot corrosion, erosion, thermo-
mechanical creep, and fatigue properties of turbine blades.
Hot corrosion is accelerated due to molten deposits which
harm the blade surface [1]. Mainly, hot corrosion has two
types which are discussed in this research work:

A) Type-1: HTHC (high-temperature hot corrosion)

The first type is where the hot corrosion is observed from

850 to 950 . The chemical reaction starts due to molten

Seventh International Conference on Aerospace Science & Engineering 6

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

turbine blades against thermal shock, hot corrosion, and

erosion degradation [16-18]. A TBC consists of a low
thermal conductivity topcoat, which protects the substrate
against hot corrosion, and the metallic bond coat, which
provides adhesion between the topcoat and substrate. The
common methods for TBCs applications are EB-PVD
(electron beam-physical vapor deposition), HVOFC (high-
velocity oxygen fuel coating), and APS (atmospheric plasma
spraying) [7, 19, 20]. APS is a widely used deposition
because of cost-effectiveness and higher efficiency [21, 22].

Yttria-Stabilized Zirconia (YSZ) is the most widely used

TBC [23, 24]. However, its performance degrades at the
temperature above 1200℃ in the presence of corrosive
environment and fuel impurities. These conditions
deteriorate TBCs by thermal cycling and spallation by hot
corrosion [25-28]. Before the occurrence of hot corrosion
Figure 1 Basquin curve representing the number of cycles to failure reaction, ZrO2 (Zirconia) is stabilized in the tetragonal phase
for superalloy DZ125 in air and the presence of hot corrosive
and stabilized by Y2O3 (Yttria) [29, 30]. During hot
environment [7].
corrosion, the acidic salts deplete the Y2O3 which lead to
phase transformation of tetragonal to monoclinic phase for
II. INFLUENCE OF HOT CORROSION ON TURBINE BLADE LIFE Zirconia. Additional stresses and cracks are added in the
With exceptional strength properties at elevated coatings due to this phase transformation. Corrosion
temperature, Ni (Nickel) based super-alloys are usually used products induce stresses in TBCs, by reducing their thermal
as a substrate material for turbine blades. chemical contraction upon cooling [31]. Wide-scale research has been
composition of refractory elements effects the mechanical done to improve the effectiveness of TBCs. These research
properties of the substrate. Cr (chromium) and Ti (Titanium) works have focused on Nanostructured TBCs and rare earth
are necessary for resistance against hot corrosion [8]. Cr metals and compounds.
restricts the Na2SO4 molten deposits by forming Cr2O3
(chromium oxide). Other materials like Nb (Niobium) and A) Nanostructured TBCs
Mo (Molybdenum) serves up to 1200 °C and 1300°C, in Nanostructured YSZ coatings have shown an increase in
uncooled systems, respectively. However, hot corrosion thermal fatigue, bonding strength, and decreased the thermal
challenges the turbine blades and disks as they reduce their conductivity of TBCs in comparison to conventional
fatigue life [9-11]. The fatigue testing of Ni superalloy coatings [32-34]. Hot corrosion experiments of
DZ125 also indicates lower fatigue strength in the presence Nanostructured YSZ TBCs have revealed better resistance
of a corrosive environment [12]. The Basquin curve in the presence of a corrosive environment.
presented in Fig. 1 presents a lower fatigue life of DZ125 Accelerated corrosion is performed on conventional and
superalloy in a hot corrosive environment. nanostructured coating specimens in the presence of
55 wt. % V2O5 and 45 wt. % Na2SO4. The specimens are
Furthermore, hot corrosion experiments done on Ni heated at 1000°C for 30 hours to capture the hot corrosion
superalloy 617 has shown that hot corrosion accelerates the effects. The process can be explained by following reactions
creep rate in corrosive environments [13]. During hot in the existence of molten salts [35].
corrosion, M23C6 carbides are precipitated at grain NaVO3 formed during the experiments has a low melting
boundaries, which depletes Cr and cause inter-granular point and it easily penetrates the coatings. Where it reacts
corrosion. These carbides enhance the fatigue and creep rate with Y2O3 to transform ZrO2 from tetragonal to the
of turbine blades due to which, grain boundaries are monoclinic form of YVO4.
weakened by the cyclic loadings and lead to fracture.
Therefore, to enhance the turbine blade life, resistance ZrO2.Y2O3 (s) + 2NaVO3 (I) → m-ZrO2 (l) + 2YVO4 (s) +
against corrosion is essential. Thermal barrier coatings Na2O (I)
(TBCs) are utilized as resistance against hot corrosion [14].
In addition to the above reaction, V2O5 also reacts readily
III. HOT CORROSION RESISTANT TBCS
with Y2O3 at a temperature above 800ºC to transform
TBCs protect the turbine blade against hot corrosion and tetragonal Zirconia to monoclinic form.
thermal stresses, thus enhancing the efficiency and the
operating temperature of the gas turbine [15]. They protect ZrO2.Y2O3 (s) + V2O5 (I) → m-ZrO2 (l) + 2YVO4 (s)

Seventh International Conference on Aerospace Science & Engineering 7

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

porosity and lower γ-phase in nano Al2O3 resist chemical

The stresses and cracking developed by hot corrosion are reaction, which gives it an advantage over micro Al2O3.
more pronounced in conventional coatings while After the hot corrosion test, thermal stresses and cracking
nanostructured coatings resist transformation as shown in occur in CSZ with the transformation of ZrO2 and formation
Fig. 2. High surface roughness and porous zones are formed of irregular shape products of YVO4, CeVO4, and CeO3. The
in nanostructured coatings, which prolong reaction with extent of degradation in CSZ/ micro Al2O3 is lower than
molten salts and resist ZrO2 transformation. In conventional CSZ due to the lesser transformation of zirconia. However,
coatings, cracking starts earlier after 12 hours of disclosure no damage and hot corrosion products are found in layered
to hot corrosion and delamination occurs after 24 hours. composite CSZ/nano Al2O3 coatings.
However, nanostructured coatings demonstrated resistance It is found in research that corrosion resistance is improved
to hot corrosion as the growth of crack is delayed to due to less transformation of ZrO2 in gadolinia
24 hours. (Gd2O3), ytterbia (Yb2O3) and yttria co-stabilized ZrO2
(GdYb-YSZ). Yb2O3 limits the intensity of GdYb-YSZ to
react with molten salts [50]. It acts as a stabilizer for
tetragonal ZrO2 while Y2O3 and Gd2O3 react with molten
salts. After the hot corrosion test, the monoclinic phase
content of ZrO2 in GdYb-YSZ coatings is also less than
YSZ coatings. Similarly, Gd2Zr2O5 + YSZ
and Gd2Zr2O5 coatings offer higher corrosion resistance than
YSZ coatings, because of less depletion of Y2O3. Moreover,
Figure 2 Hot carrion test on Nano-structured and conventional development in thermal stresses and cracking in Gd2Zr2O5 +
coatings [36].
YSZ and Gd2Zr2O5 coatings takes a longer time in the
corrosive environment than YSZ coatings [51].
In recent study, a layer gradient nanostructured Sc2O3_Y2O3 The hot corrosion test of titania stabilized zirconia (TiSZ)
co-stabilized Zr2O3 (ScYSZ) was compared with has revealed the superior performance of TiSZ in
conventional coatings to improve performance of thermal comparison with YSZ and CSZ [52]. XRD Peaks of coatings
barrier coatings. The thermal cycling life of conventional obtained after the hot corrosion test indicate no monoclinic
coatings is less than that of ScYSZ. Also the average phase of ZrO2 in the matter of TiSZ as shown in Fig. 3.
insulation temperature of ScYSZ is higher than that of
conventional coatings [37].

B) Effect of Rare Earth Elements and Compounds

The addition of rare earth elements or compounds stabilizes

YSZ and offers better corrosion resistance than conventional
YSZ coatings [38-42]. In experiments, it was found that
Scandia and Yttria co-stabilized Zirconia (ScYSZ) offer
improved corrosion resistance in comparison with
Nanostructured YSZ and Scandia [43, 44]. ScYSZ has
different hot corrosion behavior from YSZ coatings [45]. In
the matter of YSZ coatings, spallation occurs in topcoat
while in ScYSZ coatings spallation occurs in bond coat.
ScYSZ delays the phase transformation of zirconia [43].
Figure 3 XRD patterns after exposure to hot corrosive
Also, in ScYSZ, spallation occurs in the bond coat while in
environments (a) convention YSZ coatings (b) CSZ coating and (c)
YSZ coatings it occurs in the topcoat. Research has shown TiSZ coatings [52].
that Nanostructured Ceria and Yttria co-stabilized Zirconia
(CYSZ) exhibits better corrosion resistance than Tantalum (Ta) doping on YSZ lowers thermal conductivity
conventional coatings[46]. CeO2 resists the reaction with the and stabilizes YSZ up to 1500 °C due to strong
molten salts and obstructs diffusion [47, 48]. interaction between Y3+ and Ta5+ ions. The hot corrosion
resistance of Y2 O3 and tantalum oxide (Ta2O5) co-doped
The hot corrosion resistance is enhanced for Ceria stabilized zirconia coating (TaYSZ) is found superior to YSZ coatings.
zirconia (CSZ) coatings due to the overlay of alumina The reaction of a molten mixture of Na2SO4 and V2O5
(Al2O3), as it reduces the penetration of molten salts and with TaYSZ coating is weak and reaction products
decreases the depletion of Zirconia stabilizers [49]. Higher are Na2TO3, TaVO5, and Ta9VO25.

Seventh International Conference on Aerospace Science & Engineering 8

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

In contrast to YSZ coatings, the TaYSZ coating resist phase better performance of coatings against hot corrosion
transformation and the corrosion products are small. performance. It is given by the following relation
Therefore, thermal cracking and stresses in TaYSZ coatings
are not induced due to their higher corrosion resistance. The Kp=((∆W/A)2)/t
Scanning Electron Microscope (SEM) images of both
coatings, developed following the hot corrosion test, are Where, ‘t’ is the hot corrosion time. and ∆W/A is the weight
shown in Fig. 4. gain per unit surface area.

IV. HOT CORROSION RESISTANT METALLIC BOND COATINGS

NiCoCrAlY and NiCrAlYSi coatings are renowned as hot
corrosion and oxidation resistant coatings and are also used
in bond coats of Thermal Barrier Coatings (TBC) [55].
Al2O3 is formed due to the depletion of bond coatings. Al2O3
layer is a protective layer, but it is not continuous [56].
Thus, it causes spallation.
Al2O3 reacts with Na2O to form AlO2-, reprecipitating the
Figure 4 SEM images of (a) conventional YSZ coating and (b) Al2O3 and releasing the O2-. O2- accelerates the formation of
TaYSZ coating developed after hot corrosion tests are presented Al2O3 thus consuming more Al from the coating. In the
[53]. presence of NaCl, Al2O3 is additionally dissolved by the
oxychlorination process. To avoid deterioration in
C) Hot Corrosion Comparison of Base Metal, Bond mechanical properties, Al content is kept lower than 5 wt.
Coating, and Multiple Layered TBCs %. The application of nanostructured reinforced coatings,
doping in rare earth metals and composite/gradient coating
In one study, the hot corrosion performance of a five-layered
is also investigated.
coated specimen with LZ as top layer
(LZ+YSZ/YSZ/YSZ+BC), duplex YSZ (YSZ/BC), base
Al2O3 + 1/2O2 + NaCl → 2NaAlO2 + Cl2
metal Inconel 738(BM), NiCrAlY bond coat (BC), , and
duplex Lanthanum Zirconate (LZ/BC) is compared [54]. A) Nanostructured Reinforced Coatings
The hot corrosion resistance of coated specimens is in the The addition of nano Al2O3, nano SiO2, and nano CeO2
following order. exhibits better corrosion resistance in comparison to normal
coatings. The mass change of these nano-particle coatings
Five Layered coating> YSZ> LZ> BC> BM gauges the corrosion performance of such coatings. Mass
change is defined as the difference of mass gained from
The exposure of BM to a molten mixture of Na2SO4 and scale formation to mass loss by spallation of oxide scales.
V2O5 forms Na2VO3 which acts as a catalyst in the oxidation Mathematically it is written as
of metallic ions and forms Al2O3, Cr2O3, .
The spinels NiAl2O4 and NiCr2O4 formed lead to mass gain, Mc = (m1-m0 *A0/A)/ A1
severe strains, and spallation of oxide layers. The BC offers
superior corrosion resistance than the BM due to the Here, Mc is the mass change of coating; m1, mass change of
development of Al2O3 and Cr2O3 layers, which prohibits the coated sample; A1, superficial area of the coated part of the
diffusion of oxidizing agents. The spallation of these oxide sample; m0, mass change of uncoated sample; 𝐴, superficial
layers occurs due to thermal stresses and chemical reaction area of the uncoated sample; A0, superficial area of the
with molten salts. uncoated part of the test sample. In the presence of salt (75
Due to hot corrosion in BC, the Al and Ni oxide layers are w𝑡 % Na2𝑆O4/ 25 𝑤𝑡 %K2𝑆O4) at 1050 °C, the mass change
dissociated because of the fluxing mechanism. (corrosion kinetics) curves of coatings with and without
YSZ shows improved hot corrosion resistance than the BC nanoparticles are exhibited in Fig. 5.
and worse than LZ. However, LZ has low hot corrosion Fig. 6. depicts that nanoparticle coatings have better
resistance at a higher number of cycles than YSZ, due to corrosion resistance with enhanced thermal fatigue. Whereas
propagation of cracks and low fracture toughness. The five- CeO2 has the superior hot corrosion resistance. The addition
layered coating exhibits better hot corrosion resistance than of nanoparticles serves as the nucleus for heterogeneous
YSZ and LZ coatings. The top layer reacts with the molten nucleation and results in microstructure refinement.
salts and protects the YSZ coatings, thus it has better Nanoparticles also enhance corrosion resistance and
corrosion resistance than the YSZ and LZ coatings. improve thermal fatigue. Coatings without nanoparticles are
The parabolic corrosion rate constant ‘Kp’ measures the hot affected by spallation and deformation of a surface while no
corrosion performance of coatings. Lower ‘Kp’ indicates

Seventh International Conference on Aerospace Science & Engineering 9

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

spallation is observed for coatings reinforced by corrosion resistance of conventional coatings was compared
nanoparticles as shown in Fig. 5. with nano-GZ coatings. It was found that the hot corrosion
resistance of conventional coatings is less that of GZ
coatings. As nano GZ coatings acts as a barrier to pores and
micro-cracks which play a major role in infiltration of
molten salts to the TBCs [58].

B) Doping with Rare Earth Elements

Co-Al-Y-Ce coated samples have better corrosion resistance
and lesser mass gain as compared to Co-Al and Co-Al-Y
coated specimens [59]. Co accelerates to the outer layer, and
also a synergetic effect is produced to resist hot corrosion
due to the doping of Y and Ce [60]. Co enhances the
formation of Al2O3 in NiAl coatings and lowers the diffusion
rate of sulfur [61].
Experiments have also revealed that Co-Al-Si coatings have
better corrosion resistance and lower mass gain than the Co-
Figure 5 The corrosion kinetics (mass change) curves of
NiCoCrAlY coatings [57]. Al coatings [59, 62]. The reason being is the formation of ꞵ-
NiAl which has higher Al content and it facilitates the
formation of Al2O3 [63].
A

Figure 8 XRD pattern (a) single coating on γ-TiAl (b) double

coating on γ-TiAl [67].

[Si]+ γ́ [Ni3Al] → δ[Ni2Si] + ꞵ[NiAl]

D
Moreover, SiO2 is formed as a byproduct and is resistant to
molten salts. Si restricts inter diffusion of an element
between the substrate and the coating. Also, it improves the
oxidation resistance of NiAlHf coatings [64].
Pt modified coatings also exhibit better corrosion resistance
and lower mass gain due to the transformation of 𝜃- Al2O3
Figure 6 SEM images of (A) NiCoCrAlY coating without to α- Al2O3 at initial stages of hot corrosion and formation of
nanoparticles, (B) with nano Al2O3, (C) nano SiC and (D) nano Cr2O3 which hinders oxygen, sulfur, and chlorine [65].
CeO2 are displayed [57].
C) Composite / Gradient NiCoCrAlYSi Coatings
In recent study, APS technique was applied to deposited The composite coatings, inner NiCoCrAlYSi layer, and
nano-Gd2Zr2O7 (nano-GZ) powder after synthesizing by co- outer AlSiY layer exhibit better corrosion
precipitation method on the nickel-based substrate. The hot resistance[66]. The gradient NiCoCrAlYSi coatings are

Seventh International Conference on Aerospace Science & Engineering 10

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

prepared by depositing NiCoCrAlYSi target first for forty VII. RECOMMENDATIONS

microns and then Al target for ten microns. Normal
NiCoCrAlYSi coatings have a thickness of forty microns. A standard hot corrosion testing procedure can be derived as
During hot corrosion, Al is consumed rapidly from the the parameters vary between research works. Fatigue and
normal coating as it offers less resistance to Al diffusion. In creep testing of modified coatings may also incorporate in
composite and gradient coatings, more ꞵ-NiAl phase is hot corrosive environments, as they are the most affected.
formed which is corrosion resistant. It also serves as an Al
reservoir in reaction with corrosive salts thus increasing VIII. REFERENCES
resistance of composite and gradient coatings. Experiments
have shown that gradient coatings have lower mass gain and [1] L. Zheng, Z. Maicang, and D. Jianxin, "Hot
higher corrosion resistance. In a recent study, an alloy with corrosion behavior of powder metallurgy Rene95
chemical composition of Ti-48Al-2Nb-2Cr known as γ-TiAl nickel-based superalloy in molten NaCl–Na2SO4
is employed as substrate. One sample contain one layer and salts," Materials & Design, vol. 32, no. 4, pp. 1981-
1989, 2011.
second contains two layers. The experiment was conducted
[2] M. Siba, W. Wanmahmood, M. Z. Nuawi, R.
to investigate the microstructure and isothermal oxidation Rasani, and M. Nassir, "Flow-induced vibration in
resistance at 1050℃ [67]. It was found that due to presence pipes: Challenges and solutions—A review,"
of oxide layers of ZrO2 and Y2O3, better high isothermal Journal of Engineering Science and Technology,
oxidation resistant TBCs are formed. Fig. 8. shows XRD vol. 11, no. 3, pp. 362-382, 2016.
experimental results of the single and double layered [3] S. Li, X. Yang, and H. Qi, "High-temperature hot-
corrosion effects on the creep–fatigue behavior of a
coatings on γ-TiAl substrate [67].
directionally solidified nickel-based superalloy:
Mechanism and lifetime prediction," International
Journal of Damage Mechanics, vol. 29, no. 5, pp.
V. CONCLUSION 798-809, 2020.
Hot corrosion occurs from 650 °C to 950 °C and in the [4] J. Sahu, B. R. Kumar, S. Das, N. Paulose, and S.
Mannan, "Isothermal high temperature low cycle
existence of molten sodium salts and vanadium oxide.
fatigue behavior of Nimonic-263: Influence of type
Resistance against hot corrosion is essential for avoiding I and type II hot corrosion," Materials Science and
premature failures in turbine blades. Nanostructured, rare- Engineering: A, vol. 622, pp. 131-138, 2015.
earth element co-doped, and multiple layered TBCs have [5] F. Pettit, "Hot corrosion of metals and alloys,"
improved hot corrosion resistance. These coatings avoid the Oxidation of Metals, vol. 76, no. 1-2, pp. 1-21,
phase transformation of zirconia to monoclinic phase from 2011.
[6] V. Mannava, A. S. Rao, N. Paulose, M. Kamaraj,
tetragonal phase and the formation of large crystalline
and R. S. Kottada, "Hot corrosion studies on Ni-
corrosion products. However, conventional YSZ fails in base superalloy at 650 C under marine-like
resisting these two phenomena. Hot corrosion also degrades environment conditions using three salt mixture
metallic bond coatings by spallation of protective alumina (Na2SO4+ NaCl+ NaVO3)," Corrosion Science,
layer and depletion of aluminum. The addition of Nano vol. 105, pp. 109-119, 2016.
reinforced particles, rare earth elements, and gradient [7] R. Steinbrech, V. Postolenko, J. Mönch, J.
coatings improve the hot corrosion resistance of metallic Malzbender, and L. Singheiser, "Testing method to
assess lifetime of EB-PVD thermal barrier coatings
bond coats. These additions increase aluminum content,
on tubular specimens in static and cyclic oxidation
stabilize the protective layer, and prohibit penetration of tests," Ceramics International, vol. 37, no. 1, pp.
elements between coating and substrate. XRD, SEM, and 363-368, 2011.
corrosion kinetic curves are used to evaluate diverse types of [8] J. Chang, D. Wang, T. Liu, G. Zhang, L. Lou, and
coatings. J. Zhang, "Role of tantalum in the hot corrosion of
a Ni-base single crystal superalloy," Corrosion
VI. FUTURE WORK Science, vol. 98, pp. 585-591, 2015.
[9] S. Li, X. Yang, H. Qi, J. Song, and D. Shi, "Low-
In the view of the above discussion, it is possible that the hot
temperature hot corrosion effects on the low-cycle
corrosion resistance of turbine blades can be enhanced by fatigue lifetime and cracking behaviors of a powder
the addition of nanoparticles, rare-earth metals, and gradient metallurgy Ni-based superalloy," International
coatings. However, there is a need to analyze their effects on Journal of Fatigue, vol. 116, pp. 334-343, 2018.
the properties of coatings, the cost-effectiveness of [10] P. Jena, R. Singh, L. Mahanta, S. Paswan, and J.
modifications, and the comparison of coating techniques like Sahu, "Low cycle fatigue behaviour of nickel base
EB-PVD, HVOFC, APS, etc. superalloy IN 740H at 760° C: Influence of fireside
corrosion atmosphere," International Journal of
Fatigue, vol. 116, pp. 623-633, 2018.
[11] S. Kargarnejad and F. Djavanroodi, "Failure
assessment of Nimonic 80A gas turbine blade,"

Seventh International Conference on Aerospace Science & Engineering 11

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Engineering Failure Analysis, vol. 26, pp. 211-219, [26] I. N. Qureshi, M. Shahid, and A. N. Khan, "Hot
2012. corrosion of yttria-stabilized zirconia coating, in a
[12] X. Yang, S. Li, and H. Qi, "Effect of high- mixture of sodium sulfate and vanadium oxide at
temperature hot corrosion on the low cycle fatigue 950 C," Journal of Thermal Spray Technology, vol.
behavior of a directionally solidified nickel-base 25, no. 3, pp. 567-579, 2016.
superalloy," International journal of fatigue, vol. [27] L. Gao, H. Guo, L. Wei, C. Li, S. Gong, and H. Xu,
70, pp. 106-113, 2015. "Microstructure and mechanical properties of yttria
[13] A. Homaeian and M. Alizadeh, "Interaction of hot stabilized zirconia coatings prepared by plasma
corrosion and creep in Alloy 617," Engineering spray physical vapor deposition," Ceramics
Failure Analysis, vol. 66, pp. 373-384, 2016. International, vol. 41, no. 7, pp. 8305-8311, 2015.
[14] R. Synnott, Y. Panchenko, and A. Paczoski, [28] Y. Yin et al., "Hot corrosion behavior of the La2
"Turbine blade with hot-corrosion-resistant (Zr0. 7Ce0. 3) 2O7 ceramic in molten V2O5 and a
coating," ed: Google Patents, 2020. Na2SO4+ V2O5 salt mixture," Journal of Alloys
[15] R. Ahmadi-Pidani, R. Shoja-Razavi, R. and Compounds, vol. 689, pp. 123-129, 2016.
Mozafarinia, and H. Jamali, "Laser surface [29] Y. Hui et al., "High-temperature corrosion behavior
modification of plasma sprayed CYSZ thermal of zirconia ceramic in molten Na2SO4+ NaVO3
barrier coatings," Ceramics International, vol. 39, salt mixture," Ceramics International, vol. 42, no.
no. 3, pp. 2473-2480, 2013. 1, pp. 341-350, 2016.
[16] N. P. Padture, "Advanced structural ceramics in [30] H. Huang, C. Liu, L. Ni, and C. Zhou, "Evaluation
aerospace propulsion," Nature materials, vol. 15, of microstructural evolution of thermal barrier
no. 8, p. 804, 2016. coatings exposed to Na2SO4 using impedance
[17] C. Ramachandran, V. Balasubramanian, and P. spectroscopy," Corrosion Science, vol. 53, no. 4,
Ananthapadmanabhan, "Thermal cycling behaviour pp. 1369-1374, 2011.
of plasma sprayed lanthanum zirconate based [31] M. Saremi and M. Habibi, "An Investigation on
coatings under concurrent infiltration by a molten Hot Corrosion Resistance of Plasma Sprayed YSZ
glass concoction," Ceramics International, vol. 39, ‐Ceria TBC in Na2SO4+ V2O5 at 1050° C,"
no. 2, pp. 1413-1431, 2013. Supplemental Proceedings: Materials Processing
[18] R. Ahmadi-Pidani, R. Shoja-Razavi, R. and Energy Materials, vol. 1, pp. 429-437, 2011.
Mozafarinia, and H. Jamali, "Improving the [32] A. Keyvani, M. Saremi, M. H. Sohi, Z. Valefi, M.
thermal shock resistance of plasma sprayed CYSZ Yeganeh, and A. Kobayashi, "Microstructural
thermal barrier coatings by laser surface stability of nanostructured YSZ–alumina composite
modification," Optics and Lasers in Engineering, TBC compared to conventional YSZ coatings by
vol. 50, no. 5, pp. 780-786, 2012. means of oxidation and hot corrosion tests,"
[19] L. Gao, L. Wei, H. Guo, S. Gong, and H. Xu, Journal of alloys and compounds, vol. 600, pp.
"Deposition mechanisms of yttria-stabilized 151-158, 2014.
zirconia coatings during plasma spray physical [33] H. Jamali, R. Mozafarinia, R. Shoja Razavi, R.
vapor deposition," Ceramics International, vol. 42, Ahmadi-Pidani, and M. Reza Loghman-Estarki,
no. 4, pp. 5530-5536, 2016. "Fabrication and evaluation of plasma-sprayed
[20] Y. Li, Y. Xie, X. Liu, and X. Zheng, "Effect of nanostructured and conventional YSZ thermal
physical vapor deposited Al2O3 film on TGO barrier coatings," Current nanoscience, vol. 8, no.
growth in YSZ/CoNiCrAlY coatings," Ceramics 3, pp. 402-409, 2012.
international, vol. 38, no. 6, pp. 5113-5121, 2012. [34] H. Jamali, R. Mozafarinia, R. S. Razavi, and R.
[21] P. R. AHMADI, R. Razavi, R. Mozafarinia, and H. Ahmadi-Pidani, "Comparison of thermal shock
Jamali, "characterization of ceria and yttria resistances of plasma-sprayed nanostructured and
stabilized zirconia thermal barrier coatings on in conventional yttria stabilized zirconia thermal
738 superalloy," 2012. barrier coatings," Ceramics International, vol. 38,
[22] X. Chen et al., "Thermal cycling failure of new no. 8, pp. 6705-6712, 2012.
LaMgAl11O19/YSZ double ceramic top coat [35] S. Yugeswaran, A. Kobayashi, and P.
thermal barrier coating systems," Surface and Ananthapadmanabhan, "Initial phase hot corrosion
Coatings Technology, vol. 205, no. 10, pp. 3293- mechanism of gas tunnel type plasma sprayed
3300, 2011. thermal barrier coatings," Materials Science and
[23] S. Bose, High temperature coatings. Butterworth- Engineering: B, vol. 177, no. 7, pp. 536-542, 2012.
Heinemann, 2017. [36] H. Jamali, R. Mozafarinia, R. Shoja-Razavi, and R.
[24] C. Roy, M. Noor-A-Alam, A. Choudhuri, and C. Ahmadi-Pidani, "Comparison of hot corrosion
Ramana, "Synthesis and microstructure of Gd2O3- behaviors of plasma-sprayed nanostructured and
doped HfO2 ceramics," Ceramics International, conventional YSZ thermal barrier coatings
vol. 38, no. 3, pp. 1801-1806, 2012. exposure to molten vanadium pentoxide and
[25] L. Guo, M. Li, and F. Ye, "Comparison of hot sodium sulfate," Journal of the European Ceramic
corrosion resistance of Sm2Zr2O7 and (Sm0. 5Sc0. Society, vol. 34, no. 2, pp. 485-492, 2014.
5) 2Zr2O7 ceramics in Na2SO4+ V2O5 molten [37] W. Fan et al., "Microstructural design and thermal
salt," Ceramics International, vol. 42, no. 12, pp. cycling performance of a novel layer-gradient
13849-13854, 2016. nanostructured Sc2O3–Y2O3 co-stabilized ZrO2

Seventh International Conference on Aerospace Science & Engineering 12

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

thermal barrier coating," Journal of Alloys and sprayed thermal barrier coatings," Ceramics
Compounds, p. 154525, 2020. International, vol. 40, no. 3, pp. 4579-4590, 2014.
[38] L. Guo et al., "Calcium-magnesium-alumina- [50] L. Guo, C. Zhang, M. Li, W. Sun, Z. Zhang, and F.
silicate (CMAS) resistance property of Ye, "Hot corrosion evaluation of Gd2O3-Yb2O3
BaLn2Ti3O10 (Ln= La, Nd) for thermal barrier co-doped Y2O3 stabilized ZrO2 thermal barrier
coating applications," Ceramics International, vol. oxides exposed to Na2SO4+ V2O5 molten salt,"
43, no. 13, pp. 10521-10527, 2017. Ceramics International, vol. 43, no. 2, pp. 2780-
[39] F. Wang, L. Guo, C. Wang, and F. Ye, "Calcium- 2785, 2017.
magnesium-alumina-silicate (CMAS) resistance [51] M. Habibi, L. Wang, and S. Guo, "Evolution of hot
characteristics of LnPO4 (Ln= Nd, Sm, Gd) corrosion resistance of YSZ, Gd2Zr2O7, and
thermal barrier oxides," Journal of the European Gd2Zr2O7+ YSZ composite thermal barrier
Ceramic Society, vol. 37, no. 1, pp. 289-296, 2017. coatings in Na2SO4+ V2O5 at 1050° C," Journal
[40] X. Xiaoyun, G. Hongbo, G. Shengkai, and X. of the European Ceramic Society, vol. 32, no. 8, pp.
Huibin, "Hot corrosion behavior of double- 1635-1642, 2012.
ceramic-layer LaTi2Al9O19/YSZ thermal barrier [52] M. Habibi and S. Guo, "The hot corrosion behavior
coatings," Chinese Journal of Aeronautics, vol. 25, of plasma sprayed zirconia coatings stabilized with
no. 1, pp. 137-142, 2012. yttria, ceria, and titania in sodium sulfate and
[41] A. Bhattachaya, V. Shklover, K. Kunze, and W. vanadium oxide," Materials and Corrosion, vol.
Steurer, "Effect of 7YSZ on the long-term stability 66, no. 3, pp. 270-277, 2015.
of YTaO4 doped ZrO2 system," Journal of the [53] M. Habibi, L. Wang, J. Liang, and S. Guo, "An
European Ceramic Society, vol. 31, no. 15, pp. investigation on hot corrosion behavior of YSZ-
2897-2901, 2011. Ta2O5 in Na2SO4+ V2O5 salt at 1100° C,"
[42] S. Li, Z.-G. Liu, and J.-H. Ouyang, "Study on hot Corrosion science, vol. 75, pp. 409-414, 2013.
corrosion reactions between SmYbZr2O7 ceramic [54] C. Ramachandran, V. Balasubramanian, and P.
and vanadium pentoxide at temperatures of 600– Ananthapadmanabhan, "On the cyclic hot corrosion
1000° C in air," Materials Chemistry and Physics, behaviour of atmospheric plasma sprayed
vol. 130, no. 3, pp. 1134-1138, 2011. Lanthanum Zirconate based coatings in contact
[43] M. Loghman-Estarki, R. S. Razavi, H. Edris, S. with a mixture of sodium sulphate and vanadate
Bakhshi, M. Nejati, and H. Jamali, "Comparison of salts: A comparison with the traditional Ysz duplex
hot corrosion behavior of nanostructured ScYSZ and NiCrAlY coated samples," Vacuum, vol. 97,
and YSZ thermal barrier coatings," Ceramics pp. 81-95, 2013.
international, vol. 42, no. 6, pp. 7432-7439, 2016. [55] L. Shi et al., "Influences of MCrAlY coatings on
[44] H.-f. Liu, X. Xiong, X.-b. Li, and Y.-l. Wang, "Hot oxidation resistance of single crystal superalloy
corrosion behavior of Sc2O3–Y2O3–ZrO2 thermal DD98M and their inter-diffusion behaviors,"
barrier coatings in presence of Na2SO4+ V2O5 Journal of Alloys and Compounds, vol. 649, pp.
molten salt," Corrosion science, vol. 85, pp. 87-93, 515-530, 2015.
2014. [56] K. Zhang, M. Liu, S. Liu, C. Sun, and F. Wang,
[45] M. R. Loghman-Estarki, M. Nejati, H. Edris, R. S. "Hot corrosion behaviour of a cobalt-base super-
Razavi, H. Jamali, and A. H. Pakseresht, alloy K40S with and without NiCrAlYSi coating,"
"Evaluation of hot corrosion behavior of plasma Corrosion Science, vol. 53, no. 5, pp. 1990-1998,
sprayed scandia and yttria co-stabilized 2011.
nanostructured thermal barrier coatings in the [57] H. Wang, D. Zuo, G. Chen, G. Sun, X. Li, and X.
presence of molten sulfate and vanadate salt," Cheng, "Hot corrosion behaviour of low Al
Journal of the European Ceramic Society, vol. 35, NiCoCrAlY cladded coatings reinforced by nano-
no. 2, pp. 693-702, 2015. particles on a Ni-base super alloy," Corrosion
[46] M. Hajizadeh-Oghaz, R. S. Razavi, A. Ghasemi, Science, vol. 52, no. 10, pp. 3561-3567, 2010.
and Z. Valefi, "Na2SO4 and V2O5 molten salts [58] M. Bahamirian, S. Hadavi, M. Farvizi, M.
corrosion resistance of plasma-sprayed Rahimipour, and A. Keyvani, "Enhancement of hot
nanostructured ceria and yttria co-stabilized corrosion resistance of thermal barrier coatings by
zirconia thermal barrier coatings," Ceramics using nanostructured Gd2Zr2O7 coating," Surface
International, vol. 42, no. 4, pp. 5433-5446, 2016. and Coatings Technology, vol. 360, pp. 1-12, 2019.
[47] J. Kärger, D. M. Ruthven, and D. N. Theodorou, [59] Z. Liu, X. Zhao, and C. Zhou, "Improved hot
Diffusion in nanoporous materials. John Wiley & corrosion resistance of Y–Ce–Co-modified
Sons, 2012. aluminide coating on nickel base superalloys by
[48] M. Saremi, A. Keyvani, and M. Heydarzadeh Sohi, pack cementation process," Corrosion Science, vol.
"Hot corrosion resistance and mechanical behavior 92, pp. 148-154, 2015.
of atmospheric plasma sprayed conventional and [60] H. Guo, D. Li, L. Zheng, S. Gong, and H. Xu,
nanostructured zirconia coatings," in International "Effect of co-doping of two reactive elements on
Journal of Modern Physics: Conference Series, alumina scale growth of β-NiAl at 1200° C,"
2012, vol. 5: World Scientific, pp. 720-727. Corrosion science, vol. 88, pp. 197-208, 2014.
[49] M. Nejati, M. Rahimipour, and I. Mobasherpour, [61] Q. Fan, S. Jiang, H. Yu, J. Gong, and C. Sun,
"Evaluation of hot corrosion behavior of CSZ, "Microstructure and hot corrosion behaviors of two
CSZ/micro Al2O3 and CSZ/nano Al2O3 plasma Co modified aluminide coatings on a Ni-based

Seventh International Conference on Aerospace Science & Engineering 13

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

superalloy at 700 C," Applied surface science, vol.

311, pp. 214-223, 2014.
[62] H. He, Z. Liu, W. Wang, and C. Zhou,
"Microstructure and hot corrosion behavior of Co–
Si modified aluminide coating on nickel based
superalloys," Corrosion Science, vol. 100, pp. 466-
473, 2015.
[63] C. Fu, W. Kong, and G. Cao, "Microstructure and
oxidation behavior of Al+ Si co-deposited coatings
on nickel-based superalloys," Surface and Coatings
Technology, vol. 258, pp. 347-352, 2014.
[64] P. Dai, Q. Wu, Y. Ma, S. Li, and S. Gong, "The
effect of silicon on the oxidation behavior of
NiAlHf coating system," Applied surface science,
vol. 271, pp. 311-316, 2013.
[65] R. Liu, S. Jiang, H. Yu, J. Gong, and C. Sun,
"Preparation and hot corrosion behaviour of Pt
modified AlSiY coating on a Ni-based superalloy,"
Corrosion Science, vol. 104, pp. 162-172, 2016.
[66] J. Ma, S. M. Jiang, J. Gong, and C. Sun, "Hot
corrosion properties of composite coatings in the
presence of NaCl at 700 and 900° C," Corrosion
science, vol. 70, pp. 29-36, 2013.
[67] S. Nouri, S. Sahmani, M. Asayesh, and M. M.
Aghdam, "Microstructural characterization of YSZ-
CoNiCrAlY two-layered thermal barrier coating
formed on γ-TiAl intermetallic alloy via APS
process," Intermetallics, vol. 118, p. 106704, 2020.

Seventh International Conference on Aerospace Science & Engineering 14

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Design & Development of Static-Thrust Test-Bench

For Aviation Based Piston Engines
Syed Muhmmad Basit Ali Usama Saleem Irtiza Ali
Institute of Aviation Studies University of Management & University of Management &
University of Management & Technology Technology
Technology Lahore, Pakistan Lahore, Pakistan
Lahore, Pakistan [email protected] [email protected]
[email protected]

Abstract— Internal combustion engines have been pioneers an electronic load cell to determine the amount of thrust, and
in the aviation industry, use of piston engines for aircraft a data acquisition system to obtain data from the test rig [5].
propulsion, from propeller-driven bi-planes to turbo-prop,
commercial and cargo airliners. To provide an adequate amount II. LITERATURE
of thrust piston engine rotates the propeller at a specific rpm,
allowing enough mass airflow. Thrust is the only forward-acting A. EASA Certification Requirements
force of an aircraft that helps heavier than air bodies to fly. Every regulatory authority describes the generally accepted
Depending on the mathematical model and variables included in means of compliance and guidance material over both
that with correct measurement. Test-benches have been a designing and maintenance of aircraft. Certain equipment
bench-mark in the aerospace industry to analyze the results
must be certified by the regulatory authority to get into
before a flight, having paramount significance in reliability and
safety engineering. Depending on the mathematical model and service. Having in mind these limits or factors which has to
variables included in that with correct measurement. be considered to design a mount that is safe and operable.
Calculation of thrust from a piston engine also depends on Engine mount should be fail-safe and have adequate strength
environmental changes, the diameter of the propeller, and the to bear all those arose stresses during engine run. EASA
density of air. The project would be centered on piston engines publishes these requirements in its Certificate Specification –
used in the aviation industry for light aircraft and UAVs. Static Engines, Annex VII to ED Decision 2020/006/R [6].
thrust tests are performed on the ground and safety concerns Some of the factors are as follows:
hold paramount importance. The execution of this study
 Fire Proofing
involves research, design, manufacturing, and results based on
reverse engineering initiating from virtual design, analytical  Vibration Isolation
analysis, and simulations. The final evaluation of results  Corrosion Prevention
gathered from various methods such as co-relation between  Strength
conventional mass-spring and digital loadcell. On average we
 Fail-Safe
received 19kg of thrust (25+ engine run-ups – around 40 hours
of engine run) only 10% deviation from analytically calculated B. Existing Designs Study Leading to Design Parameters
thrust – providing 90%+ accuracy.
After studying more than 25 patents and more than 60
Keywords—Aeronautics, Evaluation, Static-Thrust, Test- research papers, the design parameters concluded involves
Bench, Piston Engine. these guidelines:
I. INTRODUCTION ● The test-bench consists of three Basic Units:
Thrust, a force that acts on the reaction principle by 1. Fixed Holding Structure
Newton’s 3rd law. When an object accelerates mass in a 2. Movable Frame (Engine) or Aircraft
direction, there is a repulsive force in the opposite direction
[1]. Measurement of thrust can be achieved by two methods, 3. Load Cell & Display Unit.
a direct way where the actual force produced by the propulsor ● The design should be robust to bear all kinds of loads
to an instrument designed to respond to that particular (weight, thrust, and vibrational loads).
stimulus or an inferential way where two or more
characteristic of the system is measured than through a ● Aim to develop a mobile test-bench – having fewer
mathematical model thrust is calculated [2]. limitations to airflow and structural loads.
● Design should permit the forward/backward motion
Ground tests are usually done on a test rig to evaluate the and restrict the sideways motion.
desired characteristics and structural integrity of the piston
● The system developed should hold at least 90%
engine before installation. Static tests assist to ascertain the
accuracy.
engine’s actual parameters are the same as the design
parameters specified in the datasheet [3]. As airborne ● The material selected should resist fatigue cracks and
aircraft's thrust measurement is based on several individual corrosion.
measurements. Only one possible method of measuring the
actual thrust of a piston engine excluding the altitude and III. METHODOLOGY
airspeed influences is a static test on the ground with the help Initiated from literature research considering the practices
of a load cell [4]. The test rig is a metallic structure, having which are performed in the measurement of the static thrust of
the capability to withstand loads imposed by a piston engine, piston engines. With consulting some of the major key

Seventh International Conference on Aerospace Science & Engineering 15

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

industrial available prototypes, analyzing the outcomes and kinds of propellers can be used (pusher/attractor). As
shortcomings of the respective model. A test-stand was in test and evaluation, we have installed both
virtually designed in the view of those parameters specified in configurations with a three-blade propeller as well,
literature and market study, with the help of Solidworks. The and the strength and safety indicated were remarkable.
initial material consideration was stainless steel but due to
some capabilities issues and cost analysis, mild steel was  The engine provided in the prior design is STHIL MS-
chosen as per the analysis. After design approval, the material 070 engine. Originally, manufactured for a chainsaw,
considered was analyzed for structural loadings. The facilities modified for MCC paraglider with a reduction drive
such as machining, welding, and CNCs were used to to provide 1900 rpm reduced from around 7000 rpm.
manufacture an assembly. In the process of designing, four MS-070 is a two-stroke engine, single-cylinder piston
designs were proposed from which one is selected for the engine. The throttle is coupled with a wire to the top-
assembly, due to its robust nature and exceptional capabilities left of the forward stand. Engine specifications
to resist buckling and other structural deformities. provided by the manufacturer are as follows:

A. Software Design TABLE 1: ENGINE SPECIFICATIONS

The preliminary conceptual and detailed designs were STHIL MS-070 Engine Specifications
established using reverse engineering principles – dictating for
Power 4.8 kW
a 3d software design as the initial step. All designs were made
component by component then assembled through solidworks Crankshaft RPMs 6800
3d. The final assembly is as follows: Weight 10.2 kg
Bore | Stroke 58 mm | 40mm
Ignition System Flywheel magneto
ignition system
Displacement 105.7 cm
Fuel Capacity 1.2 liters

 Vibration isolators are of industrial grade with two

threaded ends, attached to separate plates and the
whole assembly is reinforced by an elastomer material.
C. Analytical Analysis
The analytical analysis of the test-bench involves stress
calculations, deflection calculations, load calculations,
critical buckling loads, thrust loads and various yield limits –
The beams selected for the test bench are constituents of mild
steel AISI 1212, also referred ass low-carbon steel. Material’s
1 Final assembly of test-bench (Virtual). mechanical and physical properties as per analytics are listed
below [7]:
B. System Description
 Engine Attachment - The engine has three attachment TABLE 2: MATERIAL LIMITATIONS & ANALYTICAL RESULTS
points with the forward stand, having 4 fasteners Property Amount (Metric)
securing the engine with three vibration isolators. As
a safety redundancy of attachment, there are two Hardness 126
titanium plated wires wound twice between the engine Tensile Strength, Ultimate 440 MPa
and the attachment stand. There are three vibration Tensile Strength, Yield 370 MPa
isolators two of the same size installed upstream at the Maximum Elongation at Break 15%
upper attachments, bolted on each side with spring Young’s Modulus 205 GPa
washers to have further vibrational damping
Bulk Modulus 140 GPa
characteristics. On the third attachment at the lower
side, there is an iron plate having two bold Poisons Ratio 0.290
attachments with the engine and a significantly larger Shear Modulus 80 GPa
vibration isolator connecting that mounting plate to Density 7.87 g/cc
the test stand. Due to uneven attachment points, the
most reliable attachments points were considered
neglecting the symmetry of attachments.
 Load Cell & Digital Scale - To avoid a significant 𝑊𝑥
 (3𝑙 2 − 4𝑥 2 )
amount of losses, the sensor placements are of 48𝐸𝐼
paramount importance, in the following test bench we
𝑊𝑥
have established a way to install a load cell in line with S= (2)
2𝑍
the crankshaft. To get the thrust vector (force) in line
with the load cell. The load cell is directly coupled 𝑛𝜋2 𝐸𝐼
with a digital scale to provide the real-time readings  Fcr  
𝐿2
of thrust since the project utilizes an s-type load cell
measuring both compression and tension forces – both

Seventh International Conference on Aerospace Science & Engineering 16

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

(1) represents the equation used to calculate stress among

the beams, while (2) represents the mathematical model for
calculating maximum deflection in the beam under respective
load, and (3) represents the Euler’s mathematical model to
calculate critical buckling load of the beams. All results were
analyzed and came under the limits of material and
dimensional characteristics [8].
 E = Modulus of Elasticity
 I = Moment of Intertia
Fig. 3 Static Structural Simulation - Total Deformation
 W = Load
E. Thrust Calculation
Engine RPM Vs Thrust The mathematical model used to analyse the thrust values
of the engine is as follows:
25 Area of Propeller: 𝐴 = 𝜋𝐷2 /4 = 1809.55 in2 => 1.17 m2
20 Engine Power: P = 4.8 kW => 6.43691 hp => 489.66 kg-m/s
Thrust (kg)

Air Density (ambient): ρ = 1.225 kg/m3

15
10 Static Thrust = T =
5 𝜋(0.0254 . 𝑑)2 1𝑚𝑖𝑛 2 𝑑 1.5
1.225 [(𝑅𝑃𝑀 . 0.0254 . 𝑝𝑖𝑡𝑐ℎ . ) ]( )
4 60𝑠𝑒𝑐 3.29546 .𝑝𝑖𝑡𝑐ℎ
0
T = 21.433 kg
Legend:
RPM
0.0254 & 3.29546 = Conversion Factor
RPM = Propeller Speed
Fig. 1 Engine Vs RPM – Analytical Results Pitch = Distance covered in one revolution.
F. Belt Reduction Drive
𝐷1 8
Engine RPM Vs. Time Ratio of Belt Diameters: =
𝐷2
=
2.25
= 3.56
𝐷𝑟𝑖𝑣𝑒 𝑅𝑃𝑀 6800
2500 Speed of Large Pulley (Propeller) = = =
𝑅𝑎𝑡𝑖𝑜 3.56
2000 1913.11 𝑅𝑃𝑀
The real-time measurement of propeller rpm has been done
1500
RPM

which came across 1910 revolutions per minute, neglecting

1000 the instrumentation errors.
500
0
1 2 3 4 5 6 7 8 9 10 11 12
Time

Fig. 2 RPM Vs Time – Real-Time values from RPM Sensors

D. Finite Element Model Analysis

Finite element model was built and static structural
simulations resulted similar results as per analytical:

Fig. 5 RPM Reduction Gauge Illustration

IV. WORK TECHNIQUES & APPARATUS

The aspect having paramount importance to specify the
work techniques is a suitable selection of material, a
comprehensive study was done to select mild steel beams as
structural members of test-bench due to its resistance for
deformation, certain elasticity, and vibration resistance
capabilities. After a suitable material is selected the work
techniques to join them have paramount importance to avoid
Fig. 2 Static Structural Simulation - Equivalent Stress
any kind of convenience of breakdown [9]. In order to achieve
a structure that is capable of withstanding all the involved
loads like weight, thrust, and structural friction strongest
joints were made. Following are the work techniques
involved:

Seventh International Conference on Aerospace Science & Engineering 17

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

 Grinding From the research to designing, to measure static-thrust

 Machining (CNC/Milling/Lathe) of aviation-based piston engines, there are not many
complexities. A basic consideration of sustainable design and
 Welding (SMAW) testing methods should be decided. A material to withhold the
A. Load Cell weight and thrust loads of the engine. To achieve all these
benchmarks every step involved in this project should be
To measure the static thrust of an engine on a test bench taken with extreme precision. From the data analysis, it is
many sensors play a vital role. The sensors or modules that concluded that the project can be performed with
need to be configured in the thrust bench mechanism need to extraordinary results.
detect and read any slight changes in the environment. These
sensors do not operate alone and require a mean of an TABLE 2: RESULTS
electronic hub or board through which its physical or
electrical signals can be interpreted and then displayed System Results
Sr. No.
A 500 kg load cell that uses an HX711 amplifier to Configuration Digital Conventional
operate. It is an S-type load cell that can bear up to 500kg of 1 With Wheels 14 15
force. The load cell provides electrical signals to the 2 Without Wheels 19 20
amplifier, the amplifier measures the electrical resistance that
a.
changes in the load cell due to the force being applied. Final testing and evaluation results of the test-bench

Through this, we can measure the exact change in force being ACKNOWLEDGMENT
applied by the engine of the thrust bench over time.
First, all the gratitude to the most gracious and almighty
B. RPM Sensors Allah and, our sincere gratitude to Sir. Nadir Munir for an
RPM sensors are internal electronic devices that play a upbeat inspiration, healthy criticism which helped us along
major part in an engine’s functionality. The sensors measure the way, and guidance throughout this project. And sincere
the number of cycles an engine crankshaft achieves per gratitude towards Prof. Dr. Ahmed Aizaz, Associate Vice
minute. This provides information about, where the cylinder Principal, IAS for his valuable experience and guidance
heads are located and can be fired accordingly. Two rpm throughout this journey.
sensors would be integrated one for engine rpm and the other Utmost respect & gratitude towards Sir. Ahmed Rehan,
for propeller rpm. TechnoCrafts for sponsoring this project, exceptional
The RPM sensor in our thrust bench mechanism is going practical trainings, and for their technically advanced facility.
to be an electronic spark sensor configured with a display.
For the project, we are going to use a Digital Tachometer
using an optical sensor with a separate power source for REFERENCES
measuring the RPM of the engine. Simply we have interfaced [1] National Academy of Sciences, "Commercial Aircraft Propulsion &
Engineering Research," National Academy of Engineering, vol. 6.
rpm sensor module with a battery and an integrated digital 2016. National Academy of Sciences.
module for display. The optical sensor module consists of an [2] J O. Babayomi, O. C. O, Low-cost ground test equipment for piston
optical receiver and transmitter signal (sunlight or separate) engines. 2013 IEEE International Conference on Emerging &
in a single pair that can work as a digital tachometer Sustainable Technologies for Power & ICT... Imo State, Nigeria.
[3] Mingtai, C., Static Thrust Measurement for Propeller-driven Light
V. CONCLUSTION & RESULTS Aircraft. Shanghai, Shanghai Institute of Aerospace Research – China,
2012.
All the tests performed at different times of the day and at [4] Ribeiro, A., Design and methodology of Thrust testing - Master
different environmental conditions, the team was able to Dissertation, 2013 Brazil: Sao Paulo University.
synthesize a result of 19 kg of thrust on a digital scale. The [5] Miguel, A., Development of a Test Bench for UAV Propulsion Systems
results were co-related with another loadcell and a manual Characterisation. 1st ed. Lisboa: Tecnico Lisboa, pp.22-47, 2017.
mass-spring system, we received the highest thrust of 20 kg [6] EASA, Certification Specifications and Acceptable Means of
(averaging all the results taken). Not a significant loss from Compliance (CS-E) - Annex VII to ED Decision 2020/006/R. 6th ed.
2020, Cologne: EASA.
the analytical calculation, which in terms of proving the
[7] Lane, L., UAV Engines Ltd., 2019, United Kingdom:
capability and accuracy of the test-bench designed. The www.uavenginesltd.co.uk. [Online]
assembly was modified from wheels to linear couplings to [8] ÇOBAN, S., Unmanned Aerial Vehicles (UAVs) According to Engine
avoid frictional losses – which are represented in the table Type. Journal of Aviation, III(2), 2018, pp. e-ISSN: 2587-1676.
below. [9] H. Wittenberg, E. T., 2002. Flight Physics: Essentials of Aeronautical
Disciplines and Technology, with Historical Notes, Springer: ISBN
978-1-4020-8664-9.

Seventh International Conference on Aerospace Science & Engineering 18

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Nose Landing Gear Conceptual Design and Analysis for M.A.L.E

U.C.A.V
Asim Mehmood Gohar Majeed Kashif Mehmood Usama Rehman Muhammad Adeel Khan
Institute of Space Institute of Space National University of Institute of Space Politecnico di Torino
Technology Technology Science and Technology Technology Turin, Italy.
Islamabad,Pakistan. Islamabad,Pakistan. Islamabad,Pakistan. Islamabad Pakistan [email protected]
[email protected] [email protected] [email protected] usamarehman987@gmail.
com

Abstract— Among the major components for any Unmanned and at the most critical phase of landing. Landing Gear Strut
Combat Aerial Vehicle (UCAV) landing gear is of significant acts as a mechanical shock absorbing structure and a channel
importance, as it performs the function of supporting weight, to transfer loads to the airframe in the fuselage which is a
shock-absorbing structure, taxiing, and providing stable ground stronger part as compared. It absorbs the impact energy and
support during the nonoperational period, Takeoff and at the transmits it to stronger structural members of the airframe to
most critical phase of Landing. This study presents the dissipate this energy in order to avoid the failure of
Conceptual Design, Preliminary Structure Design, Sizing and components. [1] A highly preferred landing gear strut
Stress Analysis, Finite Element Analysis, Drag Strut design, and comprises of high strength, low weight, low level of
Design of Oleo Pneumatic Shock Absorber which is
complexity, and economic feasibility. Thus a landing gear
incorporated into Nose Landing Gear (NLG) strut for Medium
consisting of these qualities will result in increased efficiency
Altitude Long Endurance Unmanned Combat Aerial Vehicle
(MALE UCAV). The MALE UCAV design was carried out by of UAV [2].
different groups and this study is a part of that program. The TABLE I COMPARISON OF DESIGN PARAMETERS
conceptual design was carried out by fulfilling the preset
constraints of Nose Landing Gear weight, its height, and
Design
location. Tri-cycle layout configuration was selected and Parameter MQ-9 Reaper Ch-5
UCAV
geometric parameters such as Wheel track and Wheel base and
other parameters were worked out. A detailed study of all the Length 36.1 ft. 36.75 ft. 35 ft.
possible loading conditions was conducted and the most critical
Wingspan 65.7 ft. 68.8 ft. 62 ft.
and maximum loads were selected for sizing and stress analysis.
Two candidate materials were selected and based on the Normal Height 12.6 ft 12.47 ft 12.6 ft
stress and Shear stress criteria, simultaneously sizing and stress
MTOW 10,494 lb. 6724 lb. 6433 lb.
analysis was carried out leading to the desirable design of the
Nose Landing Gear strut. The finalized design was modeled and Range 1200 mi 6,214 mi 2486 mi
FEA was carried out and results were obtained. Endurance 14 hrs. 35 hrs. 20 hrs.

Keywords—landing gear, Nose Landing Gear, landing gear Service Ceiling 50,000 ft. 30,000 ft. 25,000 ft.
conceptual design, strut sizing, sizing and stress analysis, landing
loads, finite element analysis
II. CONCEPTUAL DESIGN
I. INTRODUCTION
Landing gear configurations under our consideration are a)
Unmanned Aerial Vehicle (UAV) serve dozen of purposes Tail Wheel Type b) Tandem Type c) Tricycle Type. Tail
ranging from scientific, commercial to military applications. Wheel type configuration has the Main or Front landing gear
The majority of functions include transport, research and positioned a little ahead of the Center of Gravity (C.G) of
development, reconnaissance, and some form of remote UAV which creates the need for another gear for empennage.
sensing. This study deals with the preliminary design and In Tandem type configuration the main landing gear and tail
static analysis of the main landing gear system for a MALE gear are arranged in a manner that both are aligned in a straight
UCAV (Medium Altitude Long Endurance Unmanned line along the longitudinal axis of the aircraft [2]. Based on the
Combat Aerial Vehicle) adhering to CS-23 Certificate requirements and its use in similar types of UCAV such as
Specification. The designed UCAV for which landing gear is CH-5 and time tested MQ-9 Reaper, our chosen configuration
designed in this study belongs to fixed-wing Combat UAVs is the Tricycle- type of configuration, one wheel is placed at
mainly used for defense applications. The designed UAV is the nose of UAV and the other two main wheels are connected
closely related in terms of design parameters to the Chinese either to the wings or the fuselage of the UAV. This
CH-5 and American MQ-9 Reaper, and owing to this configuration prevents the UAV from nosing over as well as
resemblance the design of landing gear will be closely in avoids the ground looping due to evenly distribution of weight
proximity of landing gear of these reference UCAVs in and the well-balanced center of gravity. [2]
TABLE I.
Landing gear design is carried out at the end of the design
UAVs are remotely controlled aircrafts which increase the cycle for an aircraft the reason being the number of inputs
risk of crash on landing several times, thus a reliable and required to carry out its design procedure. Moreover, aircraft
strong landing gear system is of paramount importance. Like center of gravity both the forward and aft positions must be
any gear system of a vehicle, landing gear provides a means available [3].The conceptual design is carried out based on
for the UAV to perform its basic functions of Takeoff, predefined constraints which are tabulated in TABLE II.
Landing, and Taxiing of UAV. Stable ground support is
provided by Landing gear during the non-operational period

Seventh International Conference on Aerospace Science & Engineering 19

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

TABLE II INITIAL DESIGN CONSTRAINTS each condition fulfilling the requirements of each condition.
Selection of relevant landing conditions and calculation of
Parameter Value loads in each condition is in itself a detailed discussion and out
UCAV (𝐅𝐰𝐝) 𝑿𝑪.𝑮 17.92 ft. of scope for the present discussion. Landing loads are
UCAV (𝐀𝐟𝐭 )𝑿𝑪.𝑮 18.11 ft.
tabulated in TABLE IV.
Nose Landing gear length 5 ft. TABLE IV LANDING LOADS

Nose Landing gear total weight 62.52 lb. Main Gear Reactions
Maximum Landing weight 𝒎𝑳 5870 lb. (N)
No Load Condition
X Y Z
(𝐍. 𝐋. 𝐆) 𝑿𝑪.𝑮 8.2 ft. Level landing
1 2576.70 0 - 6054.28
(𝐌. 𝐋. 𝐆) 𝑿𝑪.𝑮  18.5 ft. (three point contact)
2 Braked roll 0 0 -14291.64
𝑳𝒂𝒕𝒆𝒓𝒂𝒍 𝑻𝒊𝒑 𝒐𝒗𝒆𝒓 𝒂𝒏𝒈𝒍𝒆 𝜳 64.3 deg.
Supplementary Nose
3 5448.85 0 -6811.06
Wheel (aft)
Supplementary Nose
4 2724.42 0 -6811.06
Wheel (forward)
Supplementary Nose
5 0 -4767.74 -6811.06
Wheel (side)

IV. NOSE LANDING GEAR

The total weight of the aircraft is born by the landing gear
system and approximately 5-10% of this total weight is taken
by Nose Landing Gear. It is one of the most significant
components, enables the steering and taxing of the aircraft.
The nose landing gear comprises of main strut with
Fig. 1 Parameters for Landing gear configuration incorporated oleo-pneumatic shock absorber and a drag strut
as well. The main strut shock absorber absorbs the impact
Conceptual design parameters necessary to accurately energy and dissipates it effectively whereas a drag strut takes
configure the landing configuration were worked out on a part of the loading along with providing strength and fixes
following the literature for conceptual design provided in [1] the position by restricting the sideways movement of the main
[4]. F is the wheel base, L is the distance amid Forward C.G strut. The nose landing gear strut is designed in accordance
and NLG, and N is the distance amid Aft C.G i.e. (C.G with the given set of constraints and also meeting the stability
corresponding to empty weight) and NLG. M is the distance requirements in accordance with CS-23 certificate
between Aft C.G and MLG [4]. TABLE III contains the specification. Some constraints like lateral stability,
worked out design parameters Horizontal Stability, and Ground clearance have been satisfied
TABLE III DESIGN PARAMETERS
in working of conceptual design parameters. The remaining
constraints such as weight and length will be satisfied during
Parameter Selected Value
sizing of N.L.G. The designing of the N.L.G focuses on the
dimensional sizing of the main strut, the drag struts, and
Wheel Base (F) 10.8 ft.
piston-cylinder assembly of the oleo-pneumatic shock
Wheel Track 10.39 ft. absorber that absorbs the major portion of energy due to
Parameter L 9.72 ft. impact during the landing of aircraft. The design of single
Parameter N 9.91 ft. acting oleo-pneumatic shock absorber follows the detailed
procedure as described by Norman S. Currey [4] . The
Turnover Angle (∅) 25.7 deg
procedure to achieve the strut design needs not just designing
Tip back Angle (𝜶𝒄 ) 15 deg the strut to bear load but also to meet the weight (62.52 lb.)
Height (𝑯𝒇 ) 4.47 ft. and length (5 ft.) constraints. The first step or sizing of the strut
and cylinder for a shock absorber is carried out following
methods of beam designing. The strut is designed to bear the
III. LANDING LOADS respective normal and shear stresses whereas the cylinder also
bears hoop stress. The working stresses are limited using a
Subsection C - Structure of the CS-23 Certification
F.O.S (factor of Safety) = 2.3..
Specifications postulates the use of Limit ground loads that
can be used in the designing cycle of the landing gear. The A. Landing Gear Schematic Layout and Axis System
limit ground loads indicated in the subsection are recognized The Nose Landing gear layout depicts the oleo-pneumatic
as inertia forces and loads acting externally and which are born shock absorber and a Drag strut at an angle of 33.44 attached
by the aircraft structure. Every external reaction and ground to the main strut just above the starting point of the Cylinder.
load condition should be in balance with the angular and linear Positive X-axis starting from the nose towards the tail of
forces of inertia in a balanced or conventional way. The UCAV. Positive Y-axis is taken along the right-wing chosen
requirements of ground load postulated in subsection C should by keeping the same orientation. To complete the triad
be fulfilled at the max designing weight, excluding that following the right-hand rule positive Z-axis points vertically
sections CS 23.479, 23.481, and 23.483 may possibly be
fulfilled with at the designed landing weight permitted under
section CS 23.473 subsection (b). [5]
Taking into account the most relevant landing conditions
for the main landing gear, landing loads were calculated for

Seventh International Conference on Aerospace Science & Engineering 20

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

downward as depicted in Fig 2. All dimensions are taken in For the present study, a single-acting shock absorber is
feet. designed according to the procedure described by Norman S.
Currey [4].
V1 is extended volume
V3 is compressed volume
P1 is extended pressure 60<P1<300 psi
P2 is Static pressure
P3 is compressed pressure
𝑆𝑡𝑟𝑜𝑘𝑒 (𝑆) = 0.32𝑚 = 12.6𝑖𝑛
𝑃2 = 1200 𝑝𝑠𝑖
1.9
𝐶𝑜𝑚𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛 𝑟𝑎𝑡𝑖𝑜 = Compressed to Static
1

Fig. 2 Landing Gear Schematic with Axis System and NLG Layout
Max vertical load picked from TABLE IV
𝑀𝑎𝑥 𝑠𝑡𝑎𝑡𝑖𝑐 𝐿𝑜𝑎𝑑 = 14291.06 𝑁 = 3211.78 𝑙𝑏𝑓
B. Shock Absorber 𝑃3 = 1.9 ∗ 𝑃2 (1)
A shock absorber is a component with a mechanism to 𝑃3 = 2280 𝑝𝑠𝑖
absorb and dissipate the impact energy thus protecting the 𝑀𝑎𝑥 𝑠𝑡𝑎𝑡𝑖𝑐 𝑙𝑜𝑎𝑑
𝑃𝑖𝑠𝑡𝑜𝑛 𝐴𝑟𝑒𝑎 (𝐴) = (2)
aircraft structure from impulsive shock loads. The function of 𝑠𝑡𝑎𝑡𝑖𝑐 𝑝𝑟𝑒𝑠𝑠𝑢𝑟𝑒
a shock absorber is to damp impulsive shock loads which arise 3211.78 𝑙𝑏
A= = 2.67 𝑖𝑛 2
due to the deceleration motion of the aircraft during landing or 1200 𝑝𝑠𝑖
encountering a bump while taxing. [6]. There are several types A∗4 2.67∗4
𝑃𝑖𝑠𝑡𝑜𝑛 𝐷𝑖𝑎𝑚𝑒𝑡𝑒𝑟 (𝐷) = √ =√ (3)
of shock absorbers in a variety of aircraft depending upon the 𝜋 𝜋
requirement. These include a) Leaf Spring b) Rubber spring c) 𝐷 = 1.87𝑖𝑛
Pneumatic Shock Absorber d) Liquid Spring e) Oleo- 𝑉3 = 0.10 ∗ 𝐷𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 (4)
Pneumatic Shock Absorber. The leaf spring shock absorbers
are the most common type used widely in general aviation 𝑉3 = 0.10 ∗ (𝑆 ∗ 𝐴)
aircraft. The strut is made up of spring steel and acts as a 𝑉3 = 0.10 ∗ (12.6 ∗ 2.67) = 3.36 𝑖𝑛 3
shock-absorbing structure and performs its usual function. 𝑉1 = 𝑉3 + 𝐷𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 (5)
Rubber spring shock absorber consists of a bearing, guide rod, 𝑉1 = 3.36 + 33.64 = 37 𝑖𝑛 3
and vulcanized rubber disks stacked. Contrary to these two 𝑃3 ∗𝑉3
types the pneumatic and liquid spring shock absorbers use 𝑃1 = (6)
𝑉1
fluids for the purpose of shock absorption. The pneumatic 2280∗3.36
shock absorber uses only air as a working fluid whereas the 𝑃1 = = 207.04 𝑝𝑠𝑖
37
liquid spring-type uses liquids such as oil for the purpose. A 𝑀𝑖𝑛𝑖𝑚𝑢𝑚 𝑃𝑒𝑟𝑚𝑖𝑠𝑠𝑖𝑏𝑙𝑒 𝑂𝑣𝑒𝑟𝑙𝑎𝑝 (𝑂) = 2.75 ∗ 𝐷 (7)
hybrid of both pneumatic and liquid spring is the oleo-
pneumatic shock absorber. Its design simplicity, 𝑂 = 2.75 ∗ 1.87 = 5.14 𝑖𝑛
serviceability, and efficient mechanism make it an obvious 𝐶𝑦𝑙𝑖𝑛𝑑𝑒𝑟 𝐿𝑒𝑛𝑔𝑡ℎ (𝐿) = 𝑆 + 𝑂 (8)
choice for our application. 𝐿 = 12.6 + 5.14 + 1𝑖𝑛 (𝑚𝑎𝑟𝑔𝑖𝑛) = 18.4 𝑖𝑛
1) Design of Oleo-Pneumatic Shock Absorber
C. Critical Loads
Oleo-Pneumatic shock absorber is the first choice for use
in contemporary aircraft of the present age. It is the most For the sizing, we don’t limit to a single landing condition
reliable and efficient form of shock absorber with efficiency rather we take a comprehensive approach and pick the most
reaching as high as 90% and weighing seven-folds less than critical load in each direction from the TABLE IV. The loads
the earlier type of shock absorbers [6]. Shock loads are transferred to attachment of Landing gear i.e. point where
accompanied resulting during landing are dissipated by maximum loading occurs, then transformed into local landing
forcing the liquid such as oil against the compartment of gas gear axis and identified as Axial, Bending, Shear, and Torsion
i.e. nitrogen. Both fluids are compressed and the kinetic Loads. TABLE V and VI contain the critical and transformed
energy is absorbed by the liquid which is forced out of the loads respectively.
TABLE V CRITICAL LOADS
compartment through the orifice plate holes. The pressure of
gas controls the rebound by pushing the liquid back through LOADS (N)
the recoil orifice to the liquid container, dissipating the energy X Y Z
in the process and preventing the shock loads to reach the 5448.85 4767.74 -14291.64
aircraft structure.
TABLE VI TRANSFORMED LOADS IN LANDING GEAR AXIS

Fx Fy Fz Mx My Mz
Reactions (N) (N) (N) (Nm) (Nm) (Nm)

-10862 3034 1293 119.3 1176.4 -4758.1

Main Strut

1733.
4071 1858.4 -1114.2 1200.1 -2612.8
8
Drag Strut
Fig. 3 Schematic Diagram of Oleo Pneumatic Shock Absorber Axial Shear Shear Torsion Bending Bending

Seventh International Conference on Aerospace Science & Engineering 21

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

D. Material Selection Net Net

Net Net
Material for landing gear strut should be of adequate Ro Ri
Axial Normal Normal
𝝉𝑻,𝑴𝒙
Shear Shear
Strut t Stress Axial+ Axial+ 𝝉𝑭𝒚 𝝉𝑭𝒛
strength to withstand the loads and be easily machineable. (m) (m)
𝝈𝑩,𝑴𝒚 𝝈𝑩,𝑴𝒛 (MPa)
+ 𝝉𝑻
(MPa) + 𝝉𝑻
Titanium Alloys have specific strength i.e. strength to weight (MPa) (MPa) (MPa) (MPa)
ratio and resistant to corrosion. These properties make it 10% 0.043 0.038 9.85 64.65 231.52 2.77 5.11 8.26
suitable for components with high strength requirements Main
15% 0.038 0.032 8.63 65.76 239.72 2.89 4.94 7.69
such as landing gear. Aluminum Alloys constitute a major
portion of material used in aircraft construction due to its low 10% 0.035 0.031 5.57 109.25 231.30 48.13 53.20 52.86
Drag
density and high strength making it a suitable candidate for 15% 0.031 0.0263 4.86 112.22 238.59 49.83 54.25 53.96
consideration for our application. Steel is another material
2) Ti-6Al-6V-2Sn
widely used in high-strength applications but its high density
results in weight penalty which makes it unsuitable for our In a similar manner for Ti-6Al-6V-2Sn using F.O.S=2.3
𝜏𝑑𝑒𝑠𝑖𝑔𝑛 = 330.43 𝑀𝑃𝑎 𝜎𝑑𝑒𝑠𝑖𝑔𝑛 = 526.08 𝑀𝑃𝑎.
application. For our purpose we selected Al 7076-T61 having
yield strength 572 MPa, Shear Strength 300 MPa [7] and Ti-
6Al-6V-2Sn having yield strength 1210 MPa, Shear Strength TABLE X NLG SIZING, NORMAL STRESS AND SHEAR STRESS ANALYSIS
760 MPa [8] TI-6AL-6V-2SN

E. Sizing and Stress Analysis

Net Net Net Net
The strut is a hollow cylinder with a circular cross-section. Axial Normal Normal
𝝉𝑻,𝑴𝒙
Shear Shear
Ro Ri
The outer radius (Ro) is determined by varying the thickness Strut t Stress Axial+ Axial+ 𝝉𝑭𝒚 𝝉𝑭𝒛
(m) (m) (MPa)
(MPa) 𝝈𝑩,𝑴𝒚 𝝈𝑩,𝑴𝒛 + 𝝉𝑻 + 𝝉𝑻
as a percentage of the outer radius. The values of Ro which (MPa) (MPa) (MPa) (MPa)
satisfy both the Design Bending Stress and Design Shear 10% 0.033 0.0297 16.72 137.98 507.16 6.14 10.12 15.47
Stress are selected. The selection focused on finding Ro with Main
thickness of 10% and 15% of Ro. The sizing is done using 15% 0.03 0.0255 13.85 129.97 483.5 5.88 9.17 13.59
absolute values in order to size for maximum stresses. 10% 0.027 0.0204 9.36 239.47 511.82 104.84 114.77 114.3
Drag
1) Shock Absorber Cylinder 15% 0.024 0.0243 8.11 235.21 501.07 107.4 113.37 112.8

During the design of Oleo-pneumatic we already fix the

piston diameter as 47.5mm thus the cylinder internal radius Ri
is fixed as Ri= 24.24mm. We perform sizing and stress F. Final Selection
analysis to find the suitable cylinder size that not only The Finally Selected Size is that of Ti-6Al-6V-2Sn for
withstands the axial, shear, torsion, and bending but also 15% of Ro thickness. The reason for selection is that Ti-6Al-
withstands hoop stress as well. The choice of material for the 6V-2Sn provides better strength to weight ratio as compared
cylinder is Ti-6Al-6V-2Sn as we are in need of a cross-section to Al7076-T61.The total Weight of the NLG struts is 13.789
that fulfills internal radius constraint and doesn’t exceed the kg i.e. 30.4 lb. thus satisfying the weight constraint as well.
cross-sectional size of the main strut. TABLE XI FINAL SELECTED SIZE
TABLE VII TRANSFORMED LOADS FOR CYLINDER SIZING
Strut t Ro (mm) Ri (mm)
Fx Fy Fz Mx My Mz Main 15% 30 25.5
Reactions
(N) (N) (N) (Nm) (Nm) (Nm)
Drag 15% 24 24.3
-15283 -4768 600 7.8 521.5 -4340.9
Value Cylinder 6 mm 30 24.25
Axial Shear Shear Torsion Bending Bending

G. Layout CAD and FEM Model

TABLE VIII CYLINDER SIZING AND STRESS ANALYSIS
The layout consists of several parts however only essential
Net Net components such as Upper mounting brackets for both main
Axial Hoop
Normal Normal and drag strut, Fork Lug, Main strut, and Drag strut go into
Ro Ri 𝝈𝑩,𝑴𝒚 𝝈𝑩,𝑴𝒛 (Axial (Axial+
(m) (m)
Stress
(MPa) (MPa)
Stress +Hoop Hoop
FEM Model for analysis. The drag strut is connected to the
(MPa) (MPa) main strut as a welded connection to ensure proper load
+𝝈𝑩,𝑴𝒚 ) +𝝈𝑩,𝑴𝒛 )
(MPa) (MPa) transfer and simplify the analysis.. The design of the upper
0.03 0.024 15.60 54.80 221.67 66.30 124.84 439.90 mounting brackets, and fork lug are based on the principle of
simplified lug analysis and it is an iterative process. The axial
1) Main Strut and Drag Strut stress at the connection point is calculated, then depending
upon the constraint of available space and weight a 3D space
a) Al7076-T61 is drawn. Different layouts of the upper mounting bracket and
We have limited the design stress by using a F.O.S 2.3 fork lug used in similar UAVs are studied and a layout meeting
therefore the size of the strut should be such that the working the constraints is selected. Using different materials strength
stresses do not exceed the design stresses 𝜏𝑑𝑒𝑠𝑖𝑔𝑛 = 130.43 𝑀𝑃𝑎 analysis is performed until an appropriate size lug fulfilling the
𝜎𝑑𝑒𝑠𝑖𝑔𝑛 = 248.7 𝑀𝑃𝑎. constraints is obtained. FEM analysis is performed and if the
results are within the required range, the design is finalized
otherwise the whole process is repeated. FEM model
TABLE IX NLG SIZING, NORMAL STRESS AND SHEAR STRESS ANALYSIS components dimensions are tabulated in TABLE XII.The
AL-7076-T61 complete Nose Landing Gear consists of several parts as
depicted in Fig 4.
1. Main Strut

Seventh International Conference on Aerospace Science & Engineering 22

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

2. Drag Strut range, the design is finalized otherwise the whole process is
3. Oleo-Pneumatic Cylinder repeated again. The complete Main Landing Gear consists of
4. Oleo-Pneumatic Piston several parts as depicted in the layout.
5. Fork
6. Fork-Piston connection
7. Torque Link
8. Upper mounting brackets for both struts
9. Drag strut lower mounting bracket
10. Female lug for drag strut connection

Fig. 5 NLG Equivalent (von-Mises) Stress 6mm mesh

Fig. 4 Nose Landing Gear Model

TABLE XII FEM Model Component Dimensions

Main Strut Drag Strut

Upper Upper
Parameter Fork Lug
Mounting Mounting
Bracket Bracket
Inner radius
15 12 10.70
Ri (mm)
Outer radius
40 30 22.19
Ro (mm)
Thickness t
17 12 6.1
(mm)

Fig. 6 NLG Equivalent (von-Mises) Strain 6mm mesh

V. FINITE ELEMENTS ANALYSIS
We utilize finite element analysis on the designed landing
gear to verify and further enhance our worked-out design. The
results obtained are more precise and give a better insight. For
the purpose of FEA only the 15% thickness case of Ti-6Al-
6V-2Sn i.e. Ro=30mm for main strut, Ro=24mm for drag strut
is modeled and analysis is carried out. The FEA model
consists of the main strut and drag strut with upper mounting
brackets and the piston fork connection. Only static analysis is
performed using boundary conditions as upper mounting
brackets as fixed support, fork wheel connection takes the
loads. For the Analysis different size mesh starting from 7mm
to 5 mm element size is used for each model to obtain precise
results and study the dependence of results on the mesh size.
The design of the upper mounting brackets and the fork lug
are based on the principle of simplified lug analysis and it is
an iterative process. The axial stress at the connection point is
calculated, then depending upon the constraint of available
space and weight a 3D space is drawn. Different layouts of the Fig. 7 NLG Equivalent (von-Mises) Stress 5mm mesh
upper mounting bracket and fork lug used in similar UAVs are
studied and a layout meeting the constraints is selected. Using
different materials strength analysis is performed until an
appropriate size lug fulfilling the constraints is obtained. FEM
analysis is performed and if the results are within the required

Seventh International Conference on Aerospace Science & Engineering 23

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

TABLE XIII FINITE ELEMENT ANALYSIS RESULTS

Max
Orthogonal Max
Model Nodes Elements Skewness Stress
Quality Strain
(MPa)

7mm 145165 43295 0.6080 0.5247 506 0.00527

6mm 185118 52520 0.5792 0.5600 516.43 0.00538

5mm 243751 65800 0.5162 0.6285 523.87 0.00545

Von-misses stresses arising in the structure should not

exceed the yield stress of the material. The obtained strain
values for different mesh sizes show that the deformation
Fig. 8 NLG Equivalent (von-Mises) Strain 5mm mesh remains within the elastic limit when compared with the yield
strain of Ti-6Al-6V-2Sn i.e. 0.010341mm/mm which means
FEM analysis shows that the main strut takes the majority that linear elastic FEA is valid in our case. With different
of load which results in higher stresses born by the main strut elements size mesh, the number of nodes and elements change
whereas the drag strut takes a part of the load from the main but there is no significant change in stress which means it is
stress thus reducing the level of stress on the main strut hence insensitive to mesh size therefore it is concluded the FEM
providing structural support to the landing gear. Analysis is valid. Results gathered for different mesh sizes in
FEA show that both stress and strain remain within the limits
of material thus satisfying the maximum loading cases for the
landing gear, therefore our designed landing gear is safe.
VI. CONCLUSION
This study provides the simplest procedure to design Nose
Landing gear for a UAV. A desirable design for Nose landing
gear is achieved through parametric study utilizing Aluminum
and Titanium alloy of adequate strengths. Simultaneous sizing
and stress analysis for 10% and 15% of outer radius thickness
cases result in different sizes of the main strut. The selected
size for the main strut is Ro =30 mm and for drag strut Ro=24
mm for Ti-6Al-6V-2Sn. The simplified design of the oleo-
pneumatic shock absorber leads to the sizing of cylinder
Ro=30mm adopting the same method as main strut sizing and
stress analysis. The total weight for the designed landing gear
comes out to be 30.4 lb. thus satisfying the weight constraint
Fig. 9 Close-up view NLG Equivalent (von-Mises) Stress 5mm mesh of 62.52 lb. FEM analysis for the final model at 5 mm element
size mesh shows that maximum equivalent stress is 523.87
MPa which is below the yield strength of Ti-6Al-6V-2Sn
1210MPa.
Landing gear should be as simple as possible because
complexity increases the cost at a higher rate than the weight.
Simpler design leads to fewer structural support members
which in turn leads to a bigger cross-sectional area of the main
strut of the landing gear. The addition of drag strut provides
structural rigidity to the landing gear by dividing the load and
providing resistance to structural deformation. However, if the
design requirement prevents the use of additional structural
components then a bigger cross-sectional main strut with
approximately 10-20% of outer radius thickness will be
adequate to bear the stresses and resist the deformation but a
negative effect is a significant increase in the total weight.
REFERENCES
Fig. 10 Close-up view NLG Equivalent (von-Mises) Strain 5mm mesh
[1] A. Jha, "Landing Gear Layout Design for Unmanned Aerial Vehicle,"
The area of interest for both the main strut and drag strut in 14th National Conference on Machines and Mechanisms
are near to the attachment of strut upper mounting brackets. (NaCoMM09), Durgapur, 2009.
Measuring the level of stress and strain at various points of [2] J. Parmar, V. Acharya and D. Challa, "SELECTION AND ANALYSIS
interest confirms that stress and strain lie within the design OF THE LANDING GEAR," International Journal of Mechanical
values thus there is no risk of failure.

Seventh International Conference on Aerospace Science & Engineering 24

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Engineering and Technology (IJMET), ISSN 0976, vol. 6, no. 2, pp.

10-18, 2015.

[3] A. Dutta, "DESIGN AND ANALYSIS OF NOSE LANDING GEAR,"

International Research Journal of Engineering and Technology
(IRJET), vol. 03, no. 10, 2016.
[4] N. S. Currey, Aircraft Landing Gear Design : Primciples and Practices,
Washington, D.C.: American Institute of Aeronautics and
Astronautics,Inc, 1988.
[5] E. A. S. A. (EASA), "Easy Access Rules for Normal, Utility, Aerobatic
and Commuter Category Aeroplanes (CS-23) (Amendment 4),"
European Aviation Saftey Agency (EASA), 2018.
[6] R. Ganugapenta, S. M. sudhan, M. D. Babu and B. S. R. Satheesh,
"SYSTEM DESIGN AND ANALYSIS OF MAIN LANDING
GEAR," INTERNATIONAL JOURNAL OF CURRENT
ENGINEERING AND SCIENTIFIC RESEARCH (IJCESR), vol. 5,
no. 2, 2018.
[7] "Aluminium 7076-T61," Matweb, [Online]. Available:
http://www.matweb.com/search/datasheet_print.aspx?matguid=8916a
21e7d604d3499f7637fde053930. [Accessed 08 2021].
[8] "Titanium Ti-6Al-6V-2Sn (Ti-6-6-2) STA 910 C/ 540 C," Aerospace
Specification Metals Inc, [Online]. Available:
http://asm.matweb.com/search/SpecificMaterial.asp?bassnum=MTP6
63. [Accessed 08 2021].

Seventh International Conference on Aerospace Science & Engineering 25

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Design and Analysis of An Effective Ventilation System

To Reduce Covid-19 Spread In a Confined Room
Salem Mohammed Alohani Muzaffar Habib Muhammad Anas
Aerospace EngineeringDepartment Aerospace EngineeringDepartment Aerospace EngineeringDepartment
College of Aeronautical Engineering, College of Aeronautical Engineering, College of Aeronautical Engineering,
NUST NUST NUST
Risalpur, Pakistan Risalpur, Pakistan Risalpur, Pakistan
[email protected] [email protected] [email protected]

Abstract— The purpose of this project is to design and analyze system is also known to be an effective remedy in reducing the
an efficient and effective ventilation system for a small room to transmission rate of these viruses [7][8][9]. Ventilation can be
remove maximum COVID germ particles in the shortest described as a disturbance created by the natural or forced air
possible time. Al first, the spread of germs was modeled and into a steady flow of air. A good ventilation system is designed
simulated using ANSYS Fluent software to find the spread of to remove infectious germs particles from the confined and
germ particles inside the confined room. Later, nine different congested areas to open air.
problem cases were made consisting of different exhaust
locations and different outlet exhaust boundary conditions. Fluid flow analysis is performed to track the motion of the
Geometry for different locations of the ventilation sources was infected particles and their probability of infecting others.
created, later CFD analysis was performed using the ANSYS Computational Fluid Dynamic (CFD) analysis is performed to
Fluent software. The transient analysis was performed using the effectively model the fluid motion inside the confined areas of
Discrete Phase Method (DPM) in ANSYS Fluent software. The rooms [10] [11] [12]. The distance traveled by the germs and
Reynolds Averaged K – epsilon turbulence model was used to the residence time are predicted using the CFD analysis. These
capture the spread of germ particles inside the room. The CFD distance and traveling times also depend on the factors like
results for all 9 cases were analyzed and it was observed that coughing velocity, airflow pattern, cough or sneeze droplet
without a ventilation source, the cough particles slowly spread size, etc [13][14]. The movement of the fluid particles is
in the room and then were settled on the floor. With the
dependent on the size of aerosols. Particles having a diameter
ventilation source (exhaust fan), the cough particles merged with
of fewer than 10 microns are light enough to spend more time
the airflow and exited the room through the outlet. As the
velocity of the fan was increased, it was observed that the time
in the air and hence enhance the chances of virus spread [4].
required for cough particles to vacate the room, was reduced. On the other hand, particles having an aerosol diameter of
The research was concluded with the selection of optimal greater than 10 microns settle to the ground relatively quickly
exhaust velocity and the location of the exhaust system inside the [15].
room. The case with an exhaust source near to the ground was This study is aimed at analyzing the transmission and
proved to be most effective in eliminating the germ particles, propagation of aerosol particles emitted for respiratory events,
with an inflow condition of 10 m/s.
like cough or sneeze in an indoor space such as an office or
classroom. To study the spread of cough particles in the area,
Keywords— Discrete Phase Method, COVID Ventilation, the CFD analysis was first performed inside an empty room
cough, RANS. without ventilation and later it was integrated by a ventilation
source to simulate and visualize the particles getting out of the
I. INTRODUCTION room. The first step of the project was to simulate a steady-
Humankind has regularly encountered epidemics over the state analysis of the cough inside a room just to learn how it
timespan of the last decade such as spinach flu, swine flu, etc. spreads. The cough was simulated using a discrete phase
Recently the world has been affected by the COVID-19 method with transient modeling. This puts the simulation
pandemic, which has proved to be more powerful and deadly closer to the real-life scenario where one can visualize cough
as compared to its predecessors. These diseases spread from particles entering inside the room environment through the
mouth and then exiting the room environment with the help of
an infected person’s mouth or nose in the form of small a ventilation system at a specific speed and under a certain
liquid droplets as they cough or sneeze. These droplets can time. The analysis has been done using ANSYS software and
its different modules such as Fluent, ICEM, and design
modular have been used to simulate the cough inside the room
stay in the air or on surfaces for a long time can infect other [16].
persons coming in contact. Studies so far suggest that a
confined room or classroom is one of the most favorable In this research, first, a room geometry of 15 by 15 ft. was
locations for the spread of COVID-19 germs. The importance generated and then a velocity inlet source was added to act as
of an effective ventilation system for removing exhaled a cough source. The cough particles enter the room through
aerosols and preventing cross infections cannot be denied this inlet source. Discrete phase modeling was used to track
[1][2][3][4]. the movement of cough particles inside the room. The spread
of the germs was observed using the velocity path lines. Our
problem deals with liquid-gas phase interaction. Hence, we
Researchers have studied multiple techniques for stopping the have to use both multiphase and models. The volume of the
spread of such pandemics by the use of sanitizers, masks, fluid method was used to capture the multiphase nature of the
isolation, etc. [3][5][6]. Designing an effective ventilation problem. The volume of the Fluid method is an advection

Seventh International Conference on Aerospace Science & Engineering 26

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

scheme that allows to track the shape and position of the pressure forces exerted on the boundaries without any losses.
interface. This scheme couples with the Navier-Stokes The differential form of the momentum equations are as under,
equation to operate properly. The volume of the fraction is a 𝐷𝑢 𝜕 𝑝 𝜕 𝜏 𝑥𝑥 𝜕 𝜏 𝑦𝑥 𝜕 𝜏 𝑧𝑥
discontinuous volume fraction function that allocates the cell 𝜌 = − + + + + 𝜌 𝑓𝑥 (3)
the value between 0 and 1. 𝐷𝑡 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑧
𝐷𝑣 𝜕 𝑝 𝜕 𝜏 𝑥𝑦 𝜕 𝜏 𝑦𝑦 𝜕 𝜏 𝑧𝑦
Time taken by the cough particles to settle down and reach 𝜌 = − + + + + 𝜌 𝑓𝑦 (4)
zero velocity was noted. After the initial simulation, a 𝐷𝑡 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑧
𝐷𝑤 𝜕 𝑝 𝜕 𝜏 𝑥𝑧 𝜕 𝜏 𝑦𝑧 𝜕 𝜏 𝑧𝑧
ventilation (velocity output) source was added. As an initial 𝜌 = − + + + + 𝜌 𝑓𝑧 (5)
case, a rectangular exhaust fan was given a 5 m/s exit velocity 𝐷𝑡 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑧
and movement of germs outside the room was observed. Time
taken to clean the room was noted. Later, different CFD cases B. Eulerian Multiphase Modeling Approaches
were analyzed with different ventilation locations and fan exit
velocities. The results were then compared to find an optimal We are using the Eulerian multiphase approach to
ventilation system. capture the phase interaction. The eulerian multiphase
approach can be accounted in two categories,
II. LITERATURE REVIEW  Dispersed continuous phase interaction
A. Governing equations
 Continuous-continuous phases interaction
The basic technique used in Computational Fluid Dynamic
(CFD) is to replace continuous fluid boundaries into discrete In the dispersed continuous phase interaction, dispersed
fluid boundaries. In a continuous domain, each flow variable phases might be solid particles, liquid droplets, or gas
is defined at every point in the domain. In discrete domain, bubbles. They are absorbed in the continuous phase. The
each flow variable is defined only at the grid points or for 3 D dispersed phase is mostly very small and is in mm or even
at the cell nodes. In a conventional CFD problem, flow smaller. On the other hand, in the continuous phase’s
variables such as velocity, pressure, density, etc are solved for interaction, the two phases form a discrete interaction
each grid point on the mesh. The governing equations and between them. In other words, they are immiscible. For
appropriate boundary conditions are defined as well. CFD example, the interaction of oil and water or interaction of
software approximates these in the discrete domain in terms of water and air. It is mostly used when we are interested to
the discrete variables. Most commercial CFD codes use finite examine this interaction. The full Eulerian model is
volume or finite-element methods as it is better suited for divided into two sub-models. [7]
modeling flow inside complex geometries [6]. The solution of  Mixture models
the following equations is required for solving any flow
problem.  The volume of the fluid model
• Continuity equation (conservation of mass) The mixture model is a simplified version of Dispersed
continuous phase interaction and the Volume of fluid
• Momentum equation (conservation of momentum) method is the simplified version of continuous-continuous
• Energy equation (conservation of energy) phase interaction. In this research, the air was used as static
fluid and water coming out of the injector and there is a
For this research, a finite control volume solution strategy is discrete interaction. [7] Therefore, the volume of the fluid
considered to determine the properties of fluid inlet and outlet method was used in this research [8].
are calculated. The flow is considered incompressible. The
problem is also amused to be adiabatic therefore energy C. ANSYS Fluent Software. Fluent solvers are based on the
equation is also simplified. finite volume method. The domain is discretized into a
finite set of control volumes or cells. FLUENT offers a
1) Continuity equation wide array of physical models that can be applied to a wide
The continuity equation works on the principle of array of industries. The general conservation (transport)
conversation of mass i.e., the mass entering any control
equation for mass, momentum, energy is discretized into
volume must be equal to mass going outside the control
volume. The differential form of the continuity equation is as algebraic equations.
under, III. METHODOLOGY
𝐷𝜌 The room consists of one inlet, out the outlet, and the
+ ∇ . (𝜌 𝑉̅ ) = 0 (1)
𝐷𝑡 ventilation source. The back wall was defined separately
during meshing as the particles will strike on the wall. The
volume of fluid method which is the Eulerian approach and a
In this particular problem, we have assumed our flow to be continuous phases interaction model is used for the continuous
incompressible hence density variations are not taken into inlet flow. The Reynolds Averaged Navier Stokes (RANS)
𝐷𝜌 model k-epsilon was used to capture the turbulence and
consideration therefore becomes zero, the used simplified
𝐷𝑡 fluctuations of the system. The Discrete Phase model (DPM)
version of the continuity equation is as under, is used to capture the germs. This project consists of two major
∇ . (𝜌 𝑉̅ ) = 0 (2) parts, these parts are as follows, the first part is to make the
room and put the cough and sneeze source without the
2) Momentum equation ventilation source. And the second part is to see the spread of
The momentum equation works on the principle of germs in the room. The project has been divided into different
conversation of momentum of the system i.e., the momentum parts. In the first part, we have taken water as a working fluid
of any closed and preserved system should be converted to and a continuous stream of fluid is taken as input. The
continuous inlet will let the flow enter without any pause. But

Seventh International Conference on Aerospace Science & Engineering 27

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

the procedure will be cleared. After that, we will use the

ventilation source with the continuous inlet source of a cough
fluid. The spread will be noticed. The last step will be the use
of the Discrete Phase Model (DPM). The Discrete Phase
Model (DPM) is a utility of ANSYS Fluent which allows the
users to define the discrete phases. It allows the users to enter
the particles in the working domain for a defined timeline. For
our cases, the cough and sneeze consist of small discrete
particles which will act as germs and these particles will enter
the domain for a small amount of time, this small-time will be
the time that a person sneezes. The time was taken as 0.5
seconds as a normal person's cough time is almost 0.5 seconds.
The discrete particles will enter from the inlet interact with the Figure 1. Room CAD Model
air that is present in the room and interacts with the air that is
coming from the ventilation source and the germs are spread B. Meshing
inside the room and concerning the time the particles will
The geometry was imported into ICEM CFD software. The
leave the room from the outlet and the room will be germ-free.
mesh was created using the blocking technique. The mesh
The following steps were taken to complete this project,
was refined at the inlet because the gradient of properties at
 Geometry creation for all cases that point was very high. The named selection was done. The
 Mesh creation of all cases using ICEM CFD software. quality, orthogonal quality, and other parameters of the mesh
were up to standards. The general information of mesh is
 Numerical setup of all cases in ANSYS Fluent. shown as under.
 Simulation for the steady-state analysis of cough Mesh element Information
without ventilation source Total elements 461164
 Simulation for the steady-state analysis of cough with Total nodes 450300
ventilation source at velocity 5 m/s Orthogonal quality Min 0.0270391-- Max 1
 Simulation for the steady-state analysis of cough with Method Triangular
ventilation source at velocity 10 m/s
 Simulation for the transient flow analysis of cough
with ventilation source at velocity 5 m/s using
Discrete phase method (DPM)
 Simulation for the transient flow analysis of cough
with ventilation source at velocity 10 m/s using
Discrete phase method (DPM)
 Simulation for the transient flow analysis of cough
with exhaust at velocity 10 m/s using Discrete phase
method (DPM)
 Simulation for the transient flow analysis of cough
with exhaust at velocity 5 m/s using Discrete phase
method (DPM)
 Comparison of results in all cases
 Final design and conclusion.
IV. CFD ANALYSIS
A. 3D Modeling
The geometry was created in ANSYS design modular. The
general dimensions of the room were taken and the room was
created. The inlet was taken as the source for the cough. The
inlet was very small. The outlet was taken as the window
which was placed on the opposite side of the inlet. The red
area in the figure below shows the outlet and the yellow part Figure 2. Geometry Meshing
shows the walls. The inlet is not visible in the figure below
C. Air and Water fraction
because it is very small. The outer wall is closed for better
visualization. The geometry is shown as under. The phases of the fluids were defined. The air was defined
as the primary phase and the water was defined as the
secondary phase. The sneeze for simplicity was taken as
water, as it contains the denser parts. The domain was
initialized by air and water (cough source) was entering from
the velocity inlet [17] [18].

Seventh International Conference on Aerospace Science & Engineering 28

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

D. Numerical Setup in Fluent computationally expensive due to simultaneous solution of

The numerical details of the simple case are described as two equations for turbulence modeling. The results indicated
under, at most 5 % variation in the results using both turbulence
models and therefore all subsequent analysis was performed
1) Time Step size using the K-ɛ turbulence model.
In this part, we have considered that there is no time
dependency. It is a Steady-state solver for incompressible, V. CFD SIMULATION RESULTS
turbulent flows. Hence to solve the problem, the time-
dependent terms of the governing equations were neglected.
The results obtained during CFD simulations of all 10 cases
2) Compressible or incompressible are described below,
Our problem is incompressible, as described in the
A. Case 1: This is the first case of our simulations. In
problem statement. Hence, the density was kept constant, and
the compressibility effects were ignored. this case, a simple cough source was placed and the Volume
of Fluid (VOF) method was used to capture the interaction of
3) Grid Independence Study sneeze with air inside the room. There was no ventilation
If any particular flow field variable doesnot change source placed in the room. The results obtained are as under.
significantly with the change is grid size, we call the CFD
problem to be grid independent. The purpose of the grid
independence study is to find appropriate mesh size that is
both acceptable computationally and also produce result with
significant accuracy. In this research, a velocity value of any
arbitrary point was chosen inside the grid, and then
simulations were performed with different mesh sizes. Once
the value became stable and no significant change was noeted
in the velocity value, the grid size was finalized. This method
of grid independence study is iterative and time-consuming
and not suitbale for large CFD problems.
In grid Independence a certain parameter is analyzed for
finer and finer mesh sizes till the change in its value becomes
negligible or ignorable There are three methods usually used
for grid independence
i. Global size reduction method Figure 3 Air velocity in domain

ii. Surface or Curve mesh reduction method

iii. iii. Y+ value matching method
4) Convergence Criteria
Convergence is said to be achieved when the force
convergence value which is the square root sum of
squares (SRSS) of the imbalance of the forces is
greater than the criteria which are the product of SRSS
and tolerance value were taken equal to 0.001.

5) Reynolds number
Reynolds number is the ratio of inertial to viscous forces.
The formula of Reynolds number is,
𝑈∗𝐻∗ 𝜌
𝑅𝑒 = (6)
𝑢
where 𝑈 is the free stream velocity in m/s. H is the height Figure 4 Water velocity in domain
in meters. 𝜌 is the density (𝑘𝑔/𝑚3 ) and 𝜇 is the kinematic
viscosity 𝑁. 𝑠/𝑚3 . The inlet velocity for both cases was set to
6.5 m/s. The Reynolds number was calculated to it was in
turbulent regime. For, turbulence we used the k-epsilon RANS
model. Reynold’s Averaged Navier-Stroke model is based on
averaging of fluctuating terms. The fluctuation of Reynold
stresses in the Navier stokes equations causes instability in the
CFD solution.
6) K-ɛ Turbulence model
Initially the Spalart-Allmaras (SA) turbulence model was
used. The SA turbulence model is a low Reynolds number
model and does not require the use of wall functions as such.
This allows use of less dense grid near-walls, which saves
computational cost. The K-ɛ turbulence model is

Seventh International Conference on Aerospace Science & Engineering 29

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 5 Water volume fraction

Figure 8 Water velocity volume rendering

B. Case 2: In this case, a simple inlet was placed and

we used the Volume of Fluid (VOF) method to capture
the interaction of sneeze with air inside the room. In this
case, the ventilation source was placed in form of the
window. The flow entered the domain interacted with the
air and came out due to the presence of a ventilation
source

Figure 6 Water velocity cut plane

Figure 9. Density volume rendering

Figure 7 Water velocity vectors

Figure 10 Volume rendering of cough velocity

Seventh International Conference on Aerospace Science & Engineering 30

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 13 Volume fraction of cough

C. Case 3
In this case, we used a fan as the ventilation source. The
velocity was given to the fan. This case includes further two
analyses. In the first analysis, the velocity of the fan was set
as 5 m/s. In this case, the volume of fluid method is used to
capture the phase interactions.
The results of both cases are presented in this section of the
report.

1) Fan velocity 5 m/s

Figure 11 Contour of cough velocity top view

Figure 14 Vectors of cough velocity

Figure 12 Vectors of cough velocity

Seventh International Conference on Aerospace Science & Engineering 31

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 15 Volume rendering of cough velocity

Figure 18 Vectors of cough and air velocity

Figure 16 Volume rendering of cough velocity with air velocity

Figure 19 Velocity streamlines of cough and air velocity

Figure 17 Velocity streamlines of cough and air flow

D. Case 4 Figure 20 Volume rendering of cough velocity

Fan velocity 10 m/s

Seventh International Conference on Aerospace Science & Engineering 32

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 24 Screenshot at last time step

Figure 21 Volume rendering of air velocity

E. Case-5
In this case, the fan has been working at 5m/s speed to carry
the cough particles outside the room, below has been shown
the particle at different time steps, this simulation is transient.

In the upcoming cases, the DPM technique is used to capture

interactions. In this case, the fan has been working at 5m/s
speed to carry the cough particles outside the room, below has
been shown the particle at different time steps, this simulation
is transient

Figure 25 Total pressure-volume rendering

Figure 22 Screenshot at Initial time step

Figure 26 Velocity streamlines

Figure 23 Screenshot at middle time step

Seventh International Conference on Aerospace Science & Engineering 33

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 30 Screenshot at last time step

Figure 27 Velocity vectors

F. Case-6
In this case, the fan has been working at 10 m/s speed to carry
the cough particles outside the room, below has been shown
the particle at different time steps, this simulation is transient.

Figure 31 Total pressure Streamlines

Figure 28 Screenshot at Initial time step

Figure 32 Velocity volume rendering

Figure 29 Screenshot at middle time step

Seventh International Conference on Aerospace Science & Engineering 34

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 35 Screenshot at middle time step

Figure 33 Velocity streamlines

G. Case-7
In this case, the exhaust has been working instead of a simple
fan to carry the cough particles outside the room, below has
been shown the particle at different timesteps, this simulation
is transient. The speed of exhaust is high in this case 10 m/s)

Figure 36 Screenshot at last time step

Figure 37 Total pressure vectors

Figure 34 Screenshot at Initial time step

Seventh International Conference on Aerospace Science & Engineering 35

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 41 Screenshot at last time step

Figure 38 Velocity streamlines

H. Case-8
In this case, the exhaust has been working instead of a simple
fan to carry the cough particles outside the room, below has
been shown the particle at different timesteps, this simulation
is transient.
The speed of exhaust is low in this case (approx. 5 m/s)

The speed of exhaust is low in this case (approx. 5 m/s) Figure 42 Total pressure Streamlines

Figure 39 Screenshot at Initial time step

Figure 43 Velocity streamlines

Figure 40 Screenshot at middle time step

Seventh International Conference on Aerospace Science & Engineering 36

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 47 Screenshot at last time step

Figure 44 Velocity vectors

I. Case-9
In this case, a simple fan has been working to carry the cough
particles outside the room, below has been shown the particle
at different time steps, this simulation is transient.
The speed of exhaust is high in this case (approx. 10 m/s).
The main change, in this case, is the location of the fan, which
is now at the top of the room instead of on the wall.

The speed of exhaust is low in this case (approx. 10 m/s), in

this case, the location of exhaust has been shifted from wall
to bottom of the room. Figure 48 Total pressure Streamlines

vectors

Figure 45 Screenshot at Initial time step

Figure 49 Velocity volume rendering

Figure 46 Screenshot at middle time step

Seventh International Conference on Aerospace Science & Engineering 37

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 53 Screenshot at last time step

Figure 50 Velocity streamlines

J. Case-10
In this case, the exhaust has been working instead of a simple
fan to carry the cough particles outside the room, below has
been shown the particle at different time steps, this simulation
is transient.
The speed of exhaust is low in this case (approx. 10 m/s), in
this case, the location of exhaust has been shifted from wall
to bottom of the room.

Figure 54 Total pressure-volume rendering

Figure 51 Screenshot at Initial time step

Figure 55 Total pressure Streamlines

Figure 52 Screenshot at middle time step

Seventh International Conference on Aerospace Science & Engineering 38

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 59Velocity vectors

Figure 56 Total pressure vectors

Figure 57 Velocity volume rendering

Figure 58 Velocity streamlines

Seventh International Conference on Aerospace Science & Engineering 39

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

VI. DISCUSSION OF RESULTS to an exhaust. They are the transient flow analysis of cough
The above results show that different cases take different with exhaust at velocities using the Discrete Phase Method
times to remove cough and sneeze particles from the room. (DPM). The trend was the same again as we increase the
An effective ventilation system takes minimum time to velocity the time to remove the germs was reduced. The
remove maximum germs from the room while the least results till now showed that the best case was to use a
efficient design will more time. Table 1 shows the summary ventilation source using the exhaust is the best-case scenario.
of the result of all cases. The best thing to optimize was the location of the ventilation
source. For this, two more cases were simulated, one with the
Results show that case 1 takes the most time to remove the ventilation source above the room and the other with a
germs from the room. Case 1 was the simple analysis of ventilation source below the room. The ventilation source, the
cough without ventilation source, without any ventilation exhaust were placed at the roof and flow of the room. Case
system it takes a large amount of time to remove the germs 8th shows that the case in which the ventilation source is at
particles. The 2nd and 3rd cases were the steady-state analysis the roof. The results showed the sneeze particles enter inside
of cough with ventilation source at velocity 5 m/s and 10 m/s the room and due to ventilation at the top the particles spread
respectively. It is shown that with an increase in the inflow inside the room. This is dangerous in the current situation of
velocity the time to remove the germs is decreasing. But in Covid, we must make the ventilation system such that the
these cases, the ventilation source was a window only. The spread does not affect the people inside the room. Hence, the
germs came out of the room through a window. Hence, a large last case was created in which the ventilation source was at
amount of time is taken here too. In the 4th and 5th cases there the bottom. The particles enter the room and leave the room
is a presence of the fan. The trend is the same, as we increase from the ventilation source in 15 seconds. This was the most
the velocity of the fan the time is reducing from 75 to 55 optimized design for the ventilation system which takes a
seconds only. But in this case, the ventilation source was a small amount of time to remove the germs with the least
fan. In the 6th and 7th cases we changed the ventilation source spread. The results obtained were according to the literature.

TABLE 1: RESULT COMPARISON OF ALL 10 CASES OF CFD SIMULATIONS HAVING DIFFERENT

VENTILATIONS VELOCITIES AND LOCATIONS

Case
Variables Case 2 Case 3 Case 4 Case 5 Case 6 Case 7 Case 8 Case 9 Case 10
1

10 m/s
5 m/s 5 m/s 10 m/s
10 m/s 5 m/s 10 m/s with
with with with 5m/s 10 m/s
Ventilation No with with with fan exhaust
window window window with fan with fan
exhaust exhaust on roof on
only and fan and fan
bottom

Window Window Window

Type of Only Window Window Window Window Window
None and and and
ventilation window and fan and fan and Fan and Fan and Fan
Exhaust Exhaust Exhaust

Position of
- - - - Wall Wall Wall Wall Roof Floor
fan/exhaust

Time to get Didn’t

250 200 200 75 55 20 35 40 15
room clean get
seconds seconds seconds seconds seconds seconds seconds seconds seconds
(average) clean

Seventh International Conference on Aerospace Science & Engineering 40

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

10.1098/rsif.2009.0305.focus.
VII. CONCLUSION
[10] S. Bhattacharyya, K. Dey, A. R. Paul, and R. Biswas, “A
novel CFD analysis to minimize the spread of COVID-
This research uses CFD simulation analysis to study the 19 virus in a hospital isolation room,” Chaos, Solitons
spatial spread of cough and sneeze droplets in a confined and Fractals, vol. 139, p. 110294, 2020, DOI:
room. The project was aimed at designing and modeling an 10.1016/j.chaos.2020.110294.
effective ventilation system for a small room. Different [11] S. Peng, Q. Chen, and E. Liu, “The role of computational
fluid dynamics tools on investigation of pathogen
cases were selected based on different locations of exhausts transmission: Prevention and control,” Sci. Total
and different exhaust velocities. It was observed that when Environ., vol. 746, p. 142090, 2020, DOI:
there was no ventilation source, the cough spread in the 10.1016/j.scitotenv.2020.142090.
room and then was settled on the ground in the room. But [12] Mariam et al., “CFD Simulation of the Airborne
when a ventilation source (fan) was introduced, the cough Transmission of COVID-19 Vectors Emitted during
particles merged with the airflow and exited the room Respiratory Mechanisms: Revisiting the Concept of Safe
through the outlet. As the velocity of the fan was increased, Distance,” ACS Omega, vol. 6, no. 26, pp. 16876–16889,
it was observed that the number of cough particles exiting 2021, DOI: 10.1021/acsomega.1c01489.
the room through the outlet increased. The results obtained [13] L. Bourouiba, “The Fluid Dynamics of Disease
Transmission,” Annu. Rev. Fluid Mech., vol. 53, pp.
show that the cough comes out from the velocity inlet. The 473–508, 2021, DOI: 10.1146/annurev-fluid-060220-
volume fractions show the actual path taken by the cough, 113712.
the velocity vectors show that the germs in the cough come [14] L. Morawska and J. Cao, “Airborne transmission of
down due to gravity and gather at the lower surfaces. This SARS-CoV-2: The world should face the reality,”
is clearly shown in the volume rendering part. The velocity Environ. Int., vol. 139, no. April, p. 105730, 2020, DOI:
is high as it comes out of the mouth. It decreases over time 10.1016/j.envint.2020.105730.
and due to gravity, it settles down. The velocity vectors [15] J. Gralton, E. Tovey, M. L. McLaws, and W. D.
show the path taken by the cough. The showed that the case Rawlinson, “The role of particle size in aerosolised
with the ventilation source exhaust with a velocity of 10 pathogen transmission: A review,” J. Infect., vol. 62, no.
1, pp. 1–13, 2011, DOI: 10.1016/j.jinf.2010.11.010.
m/s and placed at the bottom is the most optimized [16] J. Redrow, S. Mao, I. Celik, J. A. Posada, and Z. gang
ventilation system. Feng, “Modeling the evaporation and dispersion of
airborne sputum droplets expelled from a human cough,”
REFERENCES Build. Environ., vol. 46, no. 10, pp. 2042–2051, 2011,
DOI: 10.1016/j.buildenv.2011.04.011.
[1] X. Gao, J. Wei, H. Lei, P. Xu, B. J. Cowling, and Y. Li, [17] S. Shao et al., “Risk assessment of airborne transmission
“Building ventilation as an effective disease intervention of COVID-19 by asymptomatic individuals under
strategy in a dense indoor contact network in an Ideal different practical settings,” J. Aerosol Sci., vol. 151, no.
City,” PLoS One, vol. 11, no. 9, pp. 1–20, 2016, doi: September 2020, p. 105661, 2021, DOI:
10.1371/journal.pone.0162481. 10.1016/j.jaerosci.2020.105661.
[2] Y. Li et al., “Role of ventilation in airborne transmission [18] M. R. Pendar and J. C. Páscoa, “Numerical modeling of
of infectious agents in the built environment - A the distribution of virus carrying saliva droplets during
multidisciplinary systematic review,” Indoor Air, vol. sneeze and cough,” Phys. Fluids, vol. 32, no. 8, 2020,
17, no. 1, pp. 2–18, 2007, doi: 10.1111/j.1600- doi: 10.1063/5.0018432.
0668.2006.00445.x.
[3] T. Jefferson et al., “Physical interventions to interrupt or
reduce the spread of respiratory viruses: Systematic
review,” BMJ, vol. 339, no. 7724, p. 792, 2009, DOI:
10.1136/bmj.b3675.
[4] R. Tellier, “Review of aerosol transmission of influenza
A virus,” Emerg. Infect. Dis., vol. 12, no. 11, pp. 1657–
1662, 2006, DOI: 10.3201/eid1211.060426.
[5] B. J. Cowling, Y. Zhou, D. K. M. Ip, G. M. Leung, and
A. E. Aiello, “Face masks to prevent transmission of
influenza virus: a systematic review,” Epidemiol. Infect.,
vol. 138, no. 4, pp. 449–456, 2010, DOI:
10.1017/S0950268809991658.
[6] A. D. LANGMUIR, “Epidemiology of airborne
infection.,” Bacteriol. Rev., vol. 25, pp. 173–181, 1961,
DOI: 10.1128/mmbr.25.3.173-181.1961.
[7] R. L. Kitching, “Air-borne Infection,” Br. Med. J., vol.
2, no. 4485, pp. 965–966, 1946, DOI:
10.1136/bmj.2.4485.965-b.
[8] C. B. Beggs, S. J. Shepherd, and K. G. Kerr, “Potential
for airborne transmission of infection in the waiting areas
of healthcare premises: Stochastic analysis using a
Monte Carlo model,” BMC Infect. Dis., vol. 10, pp. 1–8,
2010, DOI: 10.1186/1471-2334-10-247.
[9] C. J. Noakes and P. Andrew Sleigh, “Mathematical
models for assessing the role of airflow on the risk of
airborne infection in hospital wards,” J. R. Soc. Interface,
vol. 6, no. SUPPL. 6, 2009, DOI:

Seventh International Conference on Aerospace Science & Engineering 41

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Development Of Low Cost Wind Tunnel For

Measuring Lift & Drag
Hira Aftab Beenish Batul
Institute of Aviation studies Institute of Aviation Studies
University of Management & Technology Lahore, Pakistan University of Management & Technology Lahore, Pakistan
[email protected] [email protected]

Abstract-- The wind tunnel helps in understanding of complex Generally, there are two types of wind tunnel, that are OCWT
fluid flows. Although commercial wind tunnels are only limited to (Open Chamber Wind Tunnel) and CCWT (Close Chamber
industries and exclusive institutions, this paper discusses the Wind Tunnel). Further, they are categorized into four different
development of low cost subsonic open circuit wind tunnel for types based on the speeds they can produce. These four WTs
educational-purposes. The aim is to develop an open circuit wind are named subsonic, transonic, supersonic, and hypersonic
tunnel with 0.03 Mach to meet the Reynold number of 𝟔. 𝟒 × wind tunnels. The purpose of testing any model in the wind
𝟏𝟎^𝟒. For generating flow, nine electrical fans were used at the tunnel is the data validation of the results obtained from the
diffuser end, and honey comb structure was made by plastic straws relevant model and full scale. Practically wind tunnel is
at the contraction cone for creating laminar flow. For providing classified by its size. For a low-speed wind tunnel having
the visibility of the airflow, smoke system was introduced which
Mach< 0.3 the test section dimensions must be less than
comprises of fog machine and reservoir. Moreover, six copper
tubes were attached ahead of test section for converting smoke into
1m*1m. In the same way, WT is classified according to
linear pattern. For experimentation, wooden airfoil of NACA 0012 different applications, either used to checking forces or for
was used with chord of 0.1092m and a wing span of 0.17m having checking ice formation. (Barlow, Rae, and Pope, 1999).
pressure taps inside airfoil chord wise. For measuring data, digital Different calculations can be taken from the wind tunnel by
manometer were attached to pressure ports on the airfoil, so the using different methods. Calculations include finding forces
readings taken from sensors are used for calculating lift and drag. acting on the wind tunnel. Checking the characteristics of
Then, quantitative measurements of lift and drag were compared boundary layer separation. Also different types of ice
with published literature. formulation and how ice formation can be prevented can be
checked.
Keywords— Wind tunnel, Reynolds number, airfoil, lift, drag,
manometer, subsonic The wind tunnel (UMT wind tunnel) discussed in this
paper is a low cost, open chamber wind tunnel that is used for
I. INTRODUCTION educational purposes. The wind tunnel is developed for
A wind tunnel (WT) is an instrument used for experimenting measuring aerodynamic forces. The wind tunnel is operable at
with the airflow stream over an object. It provides a cheaper a velocity that lies in the subsonic region. For the measurement
way of airflow profile testing over different bodies including of aerodynamic forces the test piece used is a small-scale
the calculation of the aerodynamic forces both in industries rectangular wing made from the symmetrical airfoil. Different
and also for research. The wind tunnel has three parts. The types of sensors are used on WT for taking measurements of
First part is the intake which is mostly like a divergent- air pressure and air velocity. The wind tunnel must be properly
convergent duct. This is accompanied by the straight tunnel sealed for all these measurement. These include an
anemometer, pressure sensors like manometer (Louis, 2010)
that is the second part of it. Likewise, the third part is inverse
of the first part made of a converging-diverging nozzle tunnel. To measure aerodynamic forces like lift and drag we use a
Wind tunnels ease the testing of scale models of ships, rectangular wing made from NACA 0012 airfoil. The wind
aircraft, and vehicles in the basic understanding of tunnel is operable at Reynold number of 6.4*10^4. Nine fans
fundamentals of aerodynamics. Moreover, wind testing is are incorporated that provide the flow of air over the airfoil by
effective as it allows the improvement of scale models of suction. The test section is of glass and glass protector used on
changing the profile to reduce lift and drag coefficient and it for checking the angle of attack of the airfoil. Aerodynamic
also reduce fuel consumption. Wind tunnel also helps in forces of lift and drag can be measured by using pressure tabs
material testing specifically in the estimation of the fatigue with a manometer The specific formulas of lift and drag are
life. Some wind tunnels are large enough that it allows the taken from the book of Anderson's Fundamentals of
Aerodynamics. The values could be extracted by using the
testing of the entire aircraft. Different wind tunnels have
integration of the coefficient of pressure.
different in-built functions. Some are operable by digital
sensors and some by load cells. The medium that provides II. LITERATURE
forces and helps in testing is either air or fluid that passes over
Wind tunnel testing is predominant in aerodynamics and also
the object and sensor measured the required values.
aeromechanics. Wind tunnel experimental analysis of the lift,
The development of WTs gave researchers an practical
drag, boundary, velocity, moment and pressure distribution
idea of how flows perform. Wind tunnels are classified in
plays significant role in fluid mechanics theory. These factors
different ways. Classification is based on operational flow
have remarkable role in performance of the aircraft. All these
regimes. The Reynolds number is used for the classification of
factors are important in designing, for increasing the fuel
incompressible flow. Mach number and Reynolds number are
economy of the automobile or of aircraft and also in the
both used for the classification of compressible flow.
teaching basic aerodynamic principles. It also helps in

Seventh International Conference on Aerospace Science & Engineering 42

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

aerodynamic analysis of sting- mounted test models. Wind

tunnel provides a way to demonstrate principles of the flight
mechanics as well as the application of these in designing.
Recent example of the wind tunnel testing is the measurement
of the inviscid fluid flow and the mechanical energy losses to
wind tunnel design that experiment is done by using
instrument for the pressure distribution measurement. For
finding values of Cl and Cd a set of equations needed that
taken by literature. The work of Abbot et al. helps in the graph
of co efficient of lift and the angle of attack. These are used Fig. 1. Continuity for Fluids Courtesy by nuclear-power.net
for finding pressure distribution on the airfoil. Uncertainty
analysis is done for finding the precision of the used method.
The entire testing area is equipped with glass to allow the
airflow observation over the entire object. This low speed B. Bernoulli's law:
wind tunnel is equipped with digital manometer that calculates Bernoulli’s law (Fig. 2) can be seen as a declaration of the
the rate of pressure variation over the entire section. It was conservation of energy principle applicable for fluids flowing
found that there is adverse pressure gradient in the direction of whose equation is shown below. Scienceworld.wolfram.com.
air due to the increase of the angle of attack. (Nan, L., 2013). (2019).
In 2020, research was done in the mechanical engineering
department and technology workshop center of the University P + ½ ρ v2 + ρ g h =constant
of Bagdad to design and construct a low-speed small model of As P1 + ½ ρ v12 + ρ g h1 = P2 + ½ ρ v22 + ρ g h2
a subsonic wind tunnel. The wind tunnel is designed on a
specific set of standard dimensions that could be used for
laboratory educational experimentation. It is designed for the
calculation of lift, drag, and moment coefficient by using
pressure distribution. The flow speed is measured by using a
digital flow meter. Floe is generated by using five high-speed
fans and these are regulated to control different fan speeds
during experiments. Readings were taken after the calculation
of the deciding factors. Validation of these results showed
accuracy with the theatrical calculated values and it was
approved for laboratory experimentation. Fig. 2. Bernoulli’s law Courtesy by math.stackexchange.com

Modification to the test section of the wind tunnel is done to

improve the visibility and to convert it into the converging
duct flow. The wind tunnel is incorporated with the smoke IV. WIND TUNNEL DESIGN
rake and a low-cost lightning system (low power laser-based)
for improving the visibility of the airflow over the object. The
In OCWT, there is an exhaust and intake. The open type of
paper illustrated different ways of wind tunnel testing by using
WT is the easiest and inexpensive to construct. In open-type
cheap materials along with the best airflow visibility through
a smoke rake injection system. The smoke visualization helps WTs, the air is discharged directly into the workshop and is
in detecting inviscid streamline airflow and the boundary layer usually recycled in the workshop before investment, although
separation some wind tunnels use compressed gas sources. There are two
basic sorts of OCWTs. (i) Blower, (ii) Suck down. The easiest
III. WORKING PRINCIPLE OF WT way to distinguish between the two basic types of WT is the
WT working principle is on the Bernoulli’s principle and position of the fan. The blower tunnel is most flexible because
continuity equation. the fan is located at the entry to the WT; therefore, the test
sector can be replaced. Blower type WTs are so tolerant that
A. Continuity law: the outlet diffusers can usually not be there entirely to provide
The principle of continuity or equation of continuity, the easier contact to test samples and devices, although, this
principle of the fluid procedure, in short, defines the flow in omission usually leads to significant power outages. Suck-
the time volume minus what the stream from the volume must down WTs are generally further prone to low-frequency
accumulate in this volume. This principle is the result of the instability in the coming backflow than blowers, although it
law of conservation of mass. This equation fully describes the claims that the suction vortex is less problematical in these
behavior of fluid motion, plus the second equation based on types of tunnels because it does not permit the fan section
Newton's law of motion, and the 3rd equation created on earlier incoming the test. (Iopscience.iop.org, 2020.)
energy saving. Engineeringtoolbox.com. (2019) The wind tunnel presented in this paper is OCWT, suction
Aplusphysics.com. (2019). type as shown in Fig. 3.

Fluid flow rate = A*v

A1v1 = A2 v2

Seventh International Conference on Aerospace Science & Engineering 43

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

B. Wood for outer borders

For the outer borders of the wind tunnel, low-weight wood is
required with efficient strength. For this purpose, different
woods like pine wood, poplar wood, walnut, rosewood, and
plywood can be used. However, among these woods,
plywood is the basic wood used extensively in all
woodworks. So in this project, plywood ash wood is used for
the outer sides of the main border.
Fig. 3. 2D view of wind tunnel Courtesy by UMT wind tunnel
C. Wood for inner borders
For the inner side of the wind tunnel, smooth wood is required
A. 2D WT measured dimensions for smooth airflow. For this, different woods like pine boards,
The dimensions of the WT are shown in Fig. 4. There is a oak plywood and fiberboard (MDF) can be used. Pine is a
curve in the inner side of the contraction cone but not in the softwood but expansive while the oak plywood becomes
diffuser. This is because high-velocity air is required in the heavy for the inner border. Thus, MDF sheets are used for
test section. As the air is sucked at the honeycomb structure, inner borders which are less expansive and commonly used
it passes through the contraction cone before entering the test in woodworks.
part. Because of the cure in the contraction cone, the velocity D. Air Supply by Fans
of air is increased and it enters the test section. There is no
curve in the diffuser as this section is used as an outlet for the Air supply to wind tunnel is given by using suction
airflow. phenomena with the help of using fans. These fans are usually
available in size of 0.125m which is suitable to use as it
produces enough flow and requires only nine fans to cover
the whole area of the fan section (Fig. 5). These fans require
low power which in turn create a high suction

Fig. 4 2D view along the dimensions of wind tunnel Courtesy by UMT

wind tunnel Fig. 5. Fan used in Wind Tunnel Courtesy by UMT Wind Tunnel

V. RESOURCES FOR WIND TUNNEL DEVELOPMENT E. Pipes for honey comb structure manufacturing
Different resources were used in the development of wind
tunnel. Three types of wood are required to manufacture all It is mandatory to provide a straight airflow inside the wind
the sections. A high strength wood is required to be used as a tunnel. This can only be possible by inserting large amount of
main border, a low weight wood is required to be used at the round pipes in the honeycomb structure. For this purpose,
outer sides of border, and a smooth wood is required to be different types of pipes are available from which Plastic
used on the inner sides of border for smooth airflow. For this Straws having 6.096mm diameter are selected due to their
purpose, different types of woods are analyzed which are low weight (Fig. 6).
discussed below.

A. Wood for main borders

Woods generally used for borders because of their hardness
are Oak, Scottish Wood, Tower Burn Wood, and Partal
Wood. All these woods are hard and suitable to use for the
main borders of any project. However, they have high weight
and are expensive as well as they are not easily available in
the market except for Partal wood. It has a low weight as
compared to others and is readily available in the market. For Fig. 6. Pipe straws used for honey comb structure Courtesy by UMT
Wind Tunnel
this reason, Partal wood was used as the main border in this
wind tunnel VI. MANUFACTURING OF WIND TUNNEL
After all the sections of the wind tunnel are manufactured
separately, they are integrated successfully. These sections

Seventh International Conference on Aerospace Science & Engineering 44

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

are joined with each other with the help of clamps, in such a
way that the honeycomb structure is attached to the
contraction cone. The contraction cone is attached to the test
section, the test section is further joined with the diffuser, and
finally, the fan section is fitted with the diffuser. The wind is
finally painted to prevent them from changing the
environment and then placed on the stand. In Fig. 7, the
unpainted integrated WT is shown.

Fig. 8. Design of smoke system used in WT- Courtesy by UMT wind tunnel.

Fig. 7. Wind tunnel in integrated form Courtesy by UMT Wind Tunnel

VII. MODIFICATION OF WIND TUNNEL

Further modifications were done in the wind tunnel to
improve its performance and taking measurements. These
include sealing of connecting joints, incorporating airflow
visualization, sensors installation, and lift and drag
measurement.
A. Sealing of open joints Fig. 9. Smoke Rake Courtesy by UMT wind tunnel
Sealing is done to avoid pressure loss for the measurement
of lift and drag. The wind tunnel has openings in the joints
that cause air leakage due to which aerodynamic forces
cannot be measured accurately. (Affix. 2019). BISON
silicon-based sealant is used for WT joints sealing. It can seal
air and maintain pressure inside the wind tunnel the sealant is
applied on the joint openings with the help of a gun.
B. Airflow Visualization
Modification of wind tunnel test section is done by
installation of smoke flow duct in the exit section of
contraction section to visualize airflow over the airfoil. The
smoke visualization duct can display inviscid flow over the
object (airfoil) and separation of the boundary layer in the test
section. A smoke duct is designed to see the smooth flow of
air, the smoke collected in the reservoir, and ducts discharge Fig. 10. Smoke system on wind tunnel Courtesy by UMT wind Tunnel
a bunch of uniform smoke in the entry section of the test
section, in the direction against the nozzle or diffuser model. C. Sensors
The whole airflow in the test section brings out the boundary For the measurement of aerodynamic forces dynamic
layer separation and the desire internal flow. A duct attached pressure and speed is required. In today’s aviation industry
to the reservoir and smoke ducts; supplies smoke to the duct. different type of pressure sensors are used to measure
The fans in the wind tunnel suck smoke from the duct and dynamic pressure, ambient pressure and static pressure of the
visualize the flow of air over the object. air. The pressure of the air on airfoil can be determined by
Smoke system is designed for improving airflow visibility. making pressure tapping along the top and bottom of the
The design of the system is shown in Fig. 8, smoke rake for airfoil. These tapping are connected to the digital manometer
flow visualization is shown in Fig. 9. for taking pressure measurements as shown in Fig. 11.

Fig. 11. Sensors used in wind tunnel (left) and sensors installed on WT
(right)- Courtesy by UMT wind tunnel

Seventh International Conference on Aerospace Science & Engineering 45

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

VIII. TEST PIECE (AIRFOIL NACA 0012)

The test piece used is rectangular wing for lift and drag
calculation. The airfoil is NACA 0012 with span of 0.17m. It
is a symmetrical airfoil. It is selected because it has simple
geometry, is best demonstrator for low subsonic flows and is
easily experimented airfoil. The dimension of airfoil was
selected according to the dimensions of the test section of the
wind tunnel. The chord of the selected airfoil is 0.1092m.
The airfoil is manufactured by using Cal wood and
Eucalyptus tree wood and cut according to the selected
dimensions of NACA 0012 airfoil. 8 holes chord wise are
made and narrow plastic tubes are passed from these holes.
The holes are made in such a way that four holes are above Fig. 12. Experimental graphs of NACA 0012- Courtesy by Sauvageat,
2016
the chord and four holes below the chord. These are the points
from where pressure readings are taken, 4 above and 4 below IX. RESULTS AND DISCUSSION
the chord whose location is shown below.
The conditions while running the test are as follows;
For the lift and drag determination, the value of coefficient of
Temperature: 30℃
lift (Cl) and drag (Cd) is required respectively. This is
Density: 1.164 kg/m3
obtained by using the pressure (P) values obtained from the
Velocity: 9.5m/s
WT experiment into the integration formulas shown below.
1 Pref= 0.49 kpa
Cl (xi)= ∫0 𝐶𝑃𝑈 − 𝐶𝑃𝐿 (𝑑𝑥/𝑐)
The WT test is run and pressure readings from digital
𝟐 1 𝑞1 𝑞2 manometers are noted. These reading are used as input to the
Cd (xi) = ∫0 √ − (𝑑𝑦)
𝒄 𝑞0 𝑞0 formulas given in Section VIII along with the test conditions.
Where The readings obtained are shown in the table below;

CP = (P-Pref) /(0.5*density*velocity^2) Table 2 Cl and Cd from WT measurements

Table 1 Pressure tapping on airfoil
Quantity UMT WT data
Cl 0.071
A. Experimental Grpahs of NACA 0012 Cd 0.037

The experimental graphs of NACA 0012 are taken from a

The data obtained in Table 2 was compared with the
research paper at a Reynold number of 4*10^4 (Fig. 12). This
experimental results available for NACA 0012 at Reynold
data is collected for the purpose of comparing the
number close to our WT. The results are shown in Table 3.
measurements taken from UMT wind tunnel, although errors
are expected due to difference in Reynold number as for UMT
Table 3 Comparison of results
wind tunnel its value is 6.4*10^4. collected for the purpose of
comparing the measurements taken from UMT wind tunnel, Quantity UMT WT data Sauvageat et al. 2016 data
Cl 0.071 0.44
Port # Distance Along the Cd 0.037 0.026
Chord From Leading
Edge (mm) The comparison of data shows variation in results which were
1 20 further investigated. The first reason was that the Reynold
number didn’t match with the data used that led to difference
2 35
in values for Cl and Cd obtained from the two sources. This
3 50 being said, the wind tunnel being a low cost development, is
4 65 a good representation of the more accurate WTs used for
although errors are expected due to difference in Reynold aerodynamic measurements from which the students can
number as for UMT wind tunnel its value is 6.4*10^4. benefit and have practical insight into the wind tunnel
measurements .
X. CONCLUSION
The wind tunnel developed can be used to calculate lift and
drag for different airfoils and is a good practical
representative of the procedure used to conduct wind tunnel
tests in the aerospace industry. The wind tunnel has been
developed at a low cost and can be effectively used for
educational purpose. The work can be further extended to
incorporate the setup that can take measurements at different

Seventh International Conference on Aerospace Science & Engineering 46

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

angle of attacks and by taking steps to improve the accuracy [5] Wikiwand. 2021. Aerodynamic Force | Wikiwand. [online] Available
at: [Accessed 18 January 2021].
of coefficient of lift and drag values respectively.
[6] (PDF) Low Subsonic Wind Tunnel – Design and Construction",
ResearchGate, 2020. [Online]. Available:
REFERENCES https://www.researchgate.net/publication/323408885_Low_Subsonic_
Wind_Tunnel__Design_and_Construction. [Accessed: 18- Dec- 2019].
[1] Nan, L., 2013. The Methods Of Drag Force Measurement In Wind
Tunnels. FACULTY OF ENGINEERING AND SUSTAINABLE [7] Sauvageat, E., 2016. Prediction and Comparison of low-Reynolds
DEVELOPMENT Airfoil Performance. MES Reoprt, 1, p.1.
[2] 2008. A Simple Educational Wind Tunnel Setup For Visualization Of [8] 1991. Fundamental Of Aerodynamics. 3rd ed.
Duct Flow Streamlines And Nozzle/Diffuser Boundary Layer [ebook] McGrwa Hill companies Inc., pp.283,284.
Separation. [online] Available at [Accessed 18 January 2021]. Available
at:<http://ftp.demec.ufpr.br/disciplinas/TM045/FUNDAMENTALS_
[3] . Kalagobe, T., 2017. WindTunnel Testing Of A NACA0012Aerofoil. OF_AERODYNAMICS.PDF> [Accessed 9 January 2021].
[online] Available at: [Accessed 18 January 2021].
[9] Lyon,, C., 1997. Summary Of Low Speed Airfoil Data. [online]
[4] Mpoweruk.com. 2005. Aerodynamic Lift And Drag And The Theory Mselig.ae.illinois.edu. Available at: <https://m-
Of Flight. [online] Available at: [Accessed 18 January 2021]. selig.ae.illinois.edu/uiuc (R mukesh, 2014)

Seventh International Conference on Aerospace Science & Engineering 47

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Resolution Improvement in GNSS Based Synthetic

Aperture Radar Using GPS L1 Signals from
Different Look Angles
Muhammad Shoaib Ul Hassan Hassan Sadiq Najam Abbas Naqvi
Department of Space Sciences Department of Space Sciences National Center of GIS and Space
Institute of Space Technology Institute of Space Technology Applications
Islamabad, Pakistan Islamabad, Pakistan Institute of Space Technology
[email protected] [email protected] Islamabad, Pakistan
[email protected]
Mohammad Usman
ABM ASTUMA (PVT) LTD
Islamabad, Pakistan
[email protected]

Abstract—The Global Positioning System (GPS) is used to targets. [6]. The key parameter in a remote sensing system is
calculate position of a user and navigation purposes around the spatial resolution which is characterized by the analysis of
globe. This paper describes a way to use reflected GPS signals point spread function (PSF) [7].
to obtain images of the targeted area of interest with improved
resolution. The GPS signals reflected from different objects on II. MATHEMATICAL MODELING
surface of earth could be used in different remote sensing
applications because they provide valuable information about
This research describes simulations for a BSAR imaging
reflecting surface. Generally, the principle is based on a kind of system that uses reflected GPS signals, reflected from area
synthetic aperture radar (SAR) which image reconstruction is of interest on surface of the Earth. A BSAR system consists
performed by using a technique called matched filter. The of GPS satellites, GPS receiver and its signal detection
process to generate GPS signal and reconstruction of image is components. A GPS satellite that moves onto the orbit may
simulated for different time periods on MATLAB. Different act as the basis for synthetic aperture, even if the continual
point target based scatter images of the interested area are shift in satellite geometry is quite modest and results in a
obtained with the help of SAR signal processing techniques. very long data acquisition period of several hours. The data
Later on these obtained images are combined by image fusion acquisition time may be of order of seconds, in case a
using stationary wavelet transformation (SWT) technique to
improve the resolution.
receiver is placed on a moving platform. Due to the
continuous motion of GPS satellites, the signal received
Keywords—GPS, Bistatic synthetic aperture radar, SAR during integration time at receiver’s position will be a
signal processing, Image Fusion frequency modulated (FM) signal. The basic concept of
radar exhibits that range resolution is opposite of signals
I. INTRODUCTION bandwidth which makes it possible for BSAR imaging
GPS was developed by U.S Department of Defense (DOD) system to get high resolution and enough range at the same
to come up with accurate estimation of position, velocity time. This helps the BSAR system to obtain high quality
and time to U.S military [1]. We use direct GPS signals for images of the target area with low cost [8].
positioning and navigation purposes while reflected GPS
signals that are also called multipath signals can be used for
different remote sensing applications [2]. Many innovative
applications have been proposed in remote sensing field on
the basis of reflected signals including soil moisture analysis
[3], oil slicks detection, ocean altimetry and wind speed and
direction estimation [4][5].

Bistatic synthetic aperture radar (BSAR) is one of the most

developing areas based on GPS transmissions. A lot of work
has been done in this field on both experimental and
theoretical level. A single GPS receiver that is located near
or on the ground could collect signals directly from satellites
as well as reflected from the surface of earth. At any time of
the day 4-8 GNSS satellite are consistently available at any
position on Earth’s surface over horizon of a single
constellation with a chance of increasing this number with

fully operational constellations. A combination of GPS-

based SAR images taken from a variety of satellites can
improve image resolution by working on points such as Fig.1 Imaging Scenario-Static Imaging System

Seventh International Conference on Aerospace Science & Engineering 48

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

different look angles. Each image is capable of pointing the

The transmitter, targets and receiver geometry can be seen in target with different view. In MATLAB, a simple code
Fig. 1. If we perform sampling for longer period for above based on SWT was used to combine these images in order to
mentioned scenario then there is a possibility of generating get more information. A basic concept about image fusion is
an image with stationary receiver. With a stationary receiver shown in Fig. 3. The basic purpose of combining images is
change in geometry is being provided with the help of to extract un-necessary information from images and as an
moving GPS satellites. Although change in the geometry is output it gives complete detailed image which we call as a
relatively a low process and it takes data acquisition time in fused image.
the order of hours. Monitoring landslides and seismic
studies for long times are some of the applications that uses
this type of systems. The possibility of applying this method
for the purpose of imaging has been defined in many
research papers [9].

As the signal reaches to a target on Earth, it is reflected

back. The way of reflection totally depends on the
irregularity of surface and insulating properties of the
targeted surface. For imaging, non-specular reflected signals
are considered as more useful. As a result of this kind of
reflection signal will be dispersed in a wider angle and will
be further attenuated according to travelled distance. We
can consider that some point reflectors comprise the surface
of the target.

To generate the synthetic aperture, moving GPS satellites

completing one orbit in about 11 hours and 58 minutes will Fig. 3 Basic Concept of Image Fusion
provide the change in geometry. GPS data is required in the
order of hours to provide the requisite synthetic aperture by
an orbiting GPS satellite which is not possible in practical There are different types of image fusion that are widely used
scenario. Thus, there was a need to process the data with for different applications. One of those types is wavelet
selected amount only during image reconstruction skipping transform that is specifically used for the fusion of images
other and modernizing phase, frequency and time of week having same type of characteristics. Additionally, this
(TOW) updates appropriately. This divides GPS data into technique has two types: one is discrete wavelet transform
(DWT) and other is stationary wavelet transform (SWT).
portions known as skipping time (Ts) and dwell time (Tdw).
The first is the time interval during which the information is The DWT obeys some defined rules of discrete set of
not processed or omitted to effectively affect the geometry. wavelet scales and translations in order to implement
Second one is the period, when GPS data for the correlation wavelet transform. In these types of transformation, signal is
of direct and reflected signals to construct an image are mutually divided into orthogonal set of wavelets. On the
processed by reconstruction algorithms. A distribution of other hand, SWT is almost very similar to DWT except the
satellites during observation time is shown in Fig. 2 difference that the process of down sampling is suppressed
due to which SWT is translation invariant.

The main purpose of this research is to improve the

resolution of simulated images in a suitable software
environment. In general, image fusion is a process in which
two images are joined together into a solo image holding the
significant structures and topographies from each of the
original images. The significant uses of image fusion are:
medical image processing for better diagnostics and
interpretation of diseases, imaging to analyze microscopic
details, remote sensing, engineering and robotics and
computer sciences including computer vision. From many
types of image fusion techniques, SWT technique was
employed that was most suitable for our images.

Fig. 2 Polar Plot-Showing Visible GPS Satellites III. EXPERIMENT

The simulation work was performed in MATLAB. The
By using the above information, different images were simulation work is intentionally divided into two parts,
generated for using different satellites that were providing signal synthesizer and reconstruction engine. The signal

Seventh International Conference on Aerospace Science & Engineering 49

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

synthesizer helps to synthesize the signal that is coming Ts


2


from analog to digital converter (ADC) of GPS frontend Sij ( ) = SRij (t+  ) SRij  (t) dt (1)
while reconstruction engine helps to generate image. Care Ts

has been made to ensure that a digitized GPS signal was 2

only available to the reconstruction engine along with

satellite’s ephemeris information and receiver’s location. Where Ts represents the time of observation and (*) is the
complex conjugate.
The purpose of signal synthesizer is to determine the orbits
of satellite together with the individual PRN code offsets. It For the development of GPS scenario, a key consideration is
incorporates direct signals from all satellites in view, and to simulate various GPS errors. The errors such as satellite
also signals reflected from the target, as well as receiver clock error, orbital error and atmospheric error behaves
noise. The reconstruction engine divides the intended area mostly in the same manner and thus produces error in the
into square bins and attempts to correlate the incoming measurements of pseudorange and carrier phase [8]. During
signal in a coherent manner with a specific individual model simulation, direct and reflected signals can be affected by
chirp that is unique to each bin. Fig. 4 depicts a search area, these errors and they can prohibit the mechanism of
where targets can be located. Two arrays of size 101 each generating correct images. In our instance, atmospheric
were chosen to approximate the area where a target can be effect errors, receiver clock and position errors can be more
found, while the target has a pixel size of 20m resulting in significant than in comparison to other problems. During
an area of around 2000 m2. The size of 20m was a random MATLAB simulations, the results revealed that the errors
value during initial simulations and the target size was due to atmosphere had a significant effect on our
referred to as one element, square or pixel value in the box. simulations. However, a receiver clock error is introduced as
a result of the quartz clock utilized in the receiver, which
The position (0, 0, 6378050) in Earth Centered Earth Fixed causes drift into the clock and results in incorrect placement
(ECEF) coordinate system is specified at the center of area in a commercial receiver and loss of target in our simulation.
being searched, while target could be at any place in the
IV. RESULTS AND DISCUSSIONS
matrix. Actually, these are the north-pole coordinates that
were assigned to simplify the coordinate transformation. According to simulations, a reasonable spatial resolution
Two targets are settled down at (20, 0, 6378050) and (-20, 0, needs at least one hour of integration time to meet the
6378050) respectively in the ECEF coordinate system. The requirements. Two hours of data, on the other hand, revealed
starting position and velocity of receiver is specified to be at a resolution of the sub-meter level. The quality of the image
(0, 0, 6378050) and (0, 0, 0) respectively. is directly related to the geometry variations of GPS
satellites; hence the image resolution relies on the time of
integration. Different images were constructed by using the
above mentioned information that directly pin points the
target at desired location with different satellite geometries.

Fig. 4 Position of the Target

The processing of matched filters plays an important role in

this simulation. It is done with the help of sample-by-sample
multiplication of two signals in MATLAB, followed by Fig. 5 Images obtained with integration time of 7000
accumulation. This procedure was carried out for each seconds (a) SV-27 (b) SV-25 (c) SV-16
sample and data in two dimensional matrices were stored.
Subsequently, the matrix was scanned over time for each Fig. 5 show the results obtained by integrating over 7000
sample, making the procedure equivalent to the matched seconds for different satellites. The resolution of these
filter processing and each instant of time yielded a new simulated images can be improved by merging them with
correlation value. A uniform maximum correlation value is each other with the help of stationary wavelet transform
expected for the target’s location. The output of matched image fusion.
filter is given by:

Seventh International Conference on Aerospace Science & Engineering 50

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

angles gives better resolution for pin pointing a target’s

position.

CONCLUSION
The entire process of generating GPS signals and
reconstruction of images was simulated during the present
Fig. 6 Basic Block Diagram for Image Fusion research. The simulated images have been fused by means
of stationary wavelet transformation. The results indicate
A block diagram for complete image fusion process is that in high-resolution land and change monitoring
shown in fig. 6. Two images as input 1 and input 2 are applications, this method may be used to obtain detailed
combined together with stationary wavelet transformation information at cheaper cost with fixed area images.
technique in order to form a single image that has detailed
information. REFERENCES

Table (1) shows that images obtained from SV 16, 25 and 27

[1] E. D. Kaplan and C. Hegarty, Understanding GPS/GNSS:
are fused with each other with the following combination: principles and applications, Artech House 2017.
TABLE 1 IMAGE FUSION COMBINATION [2] M. Usman and D. W. Armitage, “Acquisition of reflected
GPS signals for remote sensing applications,” Proceedings
of 2nd International Conference on Advances in Space
Figure Combination Technologies, 2008, pp. 131–136.
7(a) 16,25 [3] A. Egido, G. Ruffini, M. Caparrini, C. Martin, E. Farres,
7(b) 16,27 and X. Banque, “Soil Moisture Monitorization Using
7(c) 25,27 GNSS Reflected Signals” arXiv preprint arXiv:0805.1881
May, 2008.
[4] E. Valencia, A. Camps, H. Park, N. Rodriguez-Alvarez, X.
Bosch-Lluis, and I. Ramos-Perez, “Oil slicks detection
using GNSS-R,” In 2011 IEEE International Geoscience
and Remote Sensing Symposium, 2011, pp. 4383-4386.
[5] R. Notarpietro, S. De Mattia, M. Campanella, Y. Pei, and
P. Savi, “Detection of buried objects using reflected
GNSS signals,” EURASIP Journal on Advances in Signal
Processing, 2014(1), pp. 1-13.
[6] F. Santi, M. Bucciarelli, D. Pastina, M. Antoniou, and M.
Cherniakov, “Spatial Resolution Improvement in GNSS-
Based SAR Using Multistatic Acquisitions and Feature
Extraction,” IEEE Transactions on Geoscience and
Remote Sensing 54, no. 10 (2016): pp. 6217-6231.
[7] M. Soumekh, Synthetic aperture radar signal processing
Vol. 7, New York: Wiley, 1999.
[8] Y. Zhang, X. Qi, H. Li and H. Mu, “Bistatic Synthetic
Fig. 7 Fused images obtained from (a) SV 16 & 25 (b) SV Aperture Radar Imaging With Multi-GNSS
Transmiitters,” IEEE International Geoscience and
16 & 27 Remote Sensing Symposium, pp. 2909-2912, July, 2019.
[9] M. Usman, D. Armitage, “Initial Experiments for a
The fusion results in Fig. 7 Shows that simulated images Remote Imaging System based on Reflected GPS Signals”
obtained from any two satellites having different viewing Journal of Space Technology, Vol 3, No. 1 July, 2013

Seventh International Conference on Aerospace Science & Engineering 51

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Evaluation Of Flight Mechanics Performance Of A

Fighter Aircraft Using Computational Fluid
Dynamics
Asra Imtiaz Hussna Zainab Dr. Hossein Raza Hamdani
Institute of Space Technology (IST) Institute of Space Technology (IST) Institute of Space Technology (IST)
Islamabad, Pakistan Islamabad, Pakistan. Islamabad,Pakistan
[email protected] [email protected] [email protected]

Abstract— F-104 Starfighter has always been considered as the exhaust of engine is a complex procedure therefore
a marvel in the field of engineering because of its design and mathematical modelling of the J79 engine has been carried out
ability to sustain flights at higher speeds. The aim of this at different flow conditions.
research is to examine the viscous, compressible and steady state
flow over this vintage aircraft using modern computational After the analysis of this model the results are compared
techniques and validate the results using flight data and results and validated with existing flight data. Another software,
from Advanced Aircraft Analysis (AAA). The second objective AAA, is used to generate these derivatives. The results
of the research is to carry out the analysis on the scaled down obtained from this software at a specific flight condition are
model and comparing the results to flight conditions thus also used to validate the results.
defining the technique as a substitute for wind tunnel testing.
The next step of the research includes the analysis of
Keywords— F-104, CFD, Mach number, CAD, FLUENT, scaled down model of the Starfighter at the same flight
AAA conditions. This additional part of the research proves the
working mechanism behind a wind tunnel. As we know that
I. INTRODUCTION for wind tunnel testing a scaled down model of an aircraft is
F-104 Starfighter, manufactured by Lockheed Martin, is a used and the analysis is carried out at required flight
supersonic interceptor which was originally used as a bomber. conditions. Its main goal is to predict the free flight conditions
F-104 had set many records and was considered a wonder in of an aircraft. This entire process of predicting the conditions
the field of engineering. Its design was considered is very complex and challenging since it is being done at sub-
phenomenal since it had a light airframe and the most scale conditions. It is quite evident that using CFD instead of
technically advanced and powerful engine available. This wind tunnel testing is very economical. The CFD method is
aircraft became the first operational fighter to sustain flights at also capable of producing results over an extensive data set
speeds greater than Mach 2. Furthermore, it became the first enabling a thorough analysis of flow.
aircraft to hold world speed and altitude record Finally, these results are compared and validated with the
simultaneously. The exterior of Starfighter was also a design previously carried out analysis.
marvel as it had an aerodynamic streamlined fuselage. All the
weapons, avionics, undercarriage an engine were installed in II. AIRCRAFT SPECIFICATIONS
a cramped internal layout. One of the best part of the F104
was the selection of General Electric J79 axial-flow turbojet A. Design Specifications
engine. Before any analysis we need to study the design
specifications. Using these design specifications the CATIA
The aerodynamic efficiency of any aircraft plays a very model has been developed. Following table shows the detailed
vital role in the design and performance. The Computational design specifications.
Fluid Dynamics (CFD) supports the performance analysis of
these aircrafts. The CFD analysis, in the present time, is being TABLE I DESIGN SPECIFICATIONS [1]
used for optimization of problems in aeronautical engineering.
Crew 1
CFD uses numerical scheme approaches and advanced solver
techniques which have been a replacement of wind tunnels in
Length 54 ft 8 in
accurate design and modelling of complex problems.
The main objective of this research is to study the flow Wingspan 21 ft 9 in
analysis around the F-104 surface. This flow is subjected to
different Mach conditions at different angle of attacks. This Height 13 ft 6 in
defined range of Mach numbers includes subsonic and
transonic flow conditions. Clean geometry of the aircraft has Wing area 196.1 sq ft
been generated using CATIA V5 software.
One of the most critical steps of the analysis is finalization Empty weight 14,000 lb
of boundary conditions. Correct choice of the conditions is the
key for accurate numerical analysis. The conditions for engine Power plant 1×GE J79 turbojet
inlet, exhaust, aircraft, inlet and outlet have been defined for
each flight condition. Calculation of boundary conditions at

Seventh International Conference on Aerospace Science & Engineering 52

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Airfoil Biconvex 3.36% root and tip C. Grid Independence

Accuracy of results depends significantly on the size of
grid. A fine grid might contain a lot of cells and thus give
B. Performance Specifications more accurate results but it will require greater
computational power. However if a less dense grid with
TABLE II PERFORMANCE SPECIFICATIONS [1] fewer number of cells is used it will not give accurate
results. Therefore, it is important to wisely choose the
Maximum speed 1528 mph , Mach 2 grid size. For this purpose a grid independence study is
Combat range 420 mi carried out. Three different meshes were generated and
Ferry range 1630 mi
the results for drag, lift and momentum were evaluated.
After comparing the results it was observed that variation
Service ceiling 50,000 ft of results between grid 2 and 3 is quite negligible. On the
Rate-of-climb 48,000 ft/min other hand, grid 1 has less number of cells and could not
Lift-to-drag 9.2 provide accurate results. Finally grid 2 was selected for
analysis.
Wing loading 105 lb/sq ft
Thrust/weight 0.54 TABLE III GRID INDEPENDENCE ANALYSIS

Grid Cells

III. METHODOLOGY Grid 1 5.02 million

A. CATIA Model Grid 2 7.53 million

The first step of the analysis is CAD modeling of F-104 Grid 3 11.29 million
which is done using CATIA V5R20. The flow analysis is done
over this solid model. While modeling we need to be careful
that the parts of the geometry must not intersect or overlap
each other because if the geometry is not perfect it would
result in poor quality mesh. The geometry of F-104 used for
0.25
analysis is shown in the Fig.1 below.
0.2
0.15
0.1
0.05
0
Grid 1 Grid 2 Grid 3

CD CL

Fig. 3 Grid Independence study

D. Turbulence Model

Selection of turbulence model is yet another vital step to

Fig. 1 F-104 Starfighter modeled using CATIA V5
obtain an accurate solution. The turbulence models used for
this analysis is K-ω SST and SA. The solver used is density-
B. Mesh Generation based because all the simulations are dealing with
This geometry is further imported for meshing in ANSYS compressible flow. Energy equation was also selected.
ICEM 18. The size of the surface mesh is defined using curve
setting parameters. A circular domain is defined around E. Materials
the body of the aircraft which is almost ten times greater The fluid used in these simulations is air with ideal gas
than the length of the aircraft. The radius of the domain is properties. It was considered that F-104 was flying at sea level
460 ft. For scaled down model the radius is 100ft. The conditions. Moreover viscosity was solved using Sutherland
volume occupied by the fluid in these bounds is divide equations.
into meshes. Octree method is used for volume meshing.
F. Boundary Conditions
The surfaces of the aircraft were selected as wall so that no
slip condition can be maintained. Inlet of the engine was
selected as pressure outlet and the exhaust was selected as
pressure inlet. For all Mach numbers the static pressure was
calculated using isentropic equations. Mass flow rate for the
engine was selected as 77 kg/s. Its analysis was carried out at
Mach numbers 0.4, 0.7 and 0.9 respectively. The mass flow
for the scaled down model was chosen as 1 kg/s.
Fig. 2. Meshed Geometry of Aircraft

Seventh International Conference on Aerospace Science & Engineering 53

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

G. Solver Setup
The solver with explicit algorithm was used. The courant
number is set to 1. For flow discretization 2nd order upwind
scheme was selected and 1st order upwind scheme for
turbulent viscosity. Convergence of residuals and force
coefficients are monitored for each case.
IV. ANALYSIS OF FLOW ON F-104
The flow over the aircraft has been analyzed at three different
Mach numbers; 0.4, 0.7, 0.9 at four different angle of attacks;
0, 5, 10, 15. Following contours and plots show the flow over Fig. 7 Velocity contours at Mach 0.9 & AoA 5
the aircraft.
V. ANALYSIS USING AAA
A. Pressure Contours
The first step of the analysis as defining the geometry of the
The pressure contours at Mach number 0.4, 0.7 and 0.9 at zero
entire aircraft. Using these defined parameters the software
angle of attack.
develops the geometry of the aircraft. The next step is to
It is evident that the dynamic pressure is varying higher at
define the weights, center of gravity. Other parameters like
certain portions and low at other portion. That is at sharp
propulsions systems are also provided. Using empirical
edges dynamic pressure is more. With the increase in Mach
formulae it defines the stability and control derivatives,
number the pressure also increases.
aerodynamic forces and moments on the provided flight
conditions. The results obtained from here are compared to
the previous results obtained from CFD.
The accuracy of the results through any software depends
majorly on the geometry of the aircraft. In CFD we import an
accurate model that has been designed on CATIA on the other
hand AAA accumulates the information provided and
constructs a geometry which may or may not be very
accurate. The figure below shows the front and side view of
the aircraft created by AAA. It can be seen as compared to
CFD this geometry is not very accurate.
Fig. 4 Pressure Contours at Mach 0.4 & AoA 5

Fig. 8 Front view of Geometry

Fig. 5 Pressure Contours at Mach 0.9 & AoA 9

B. Velocity Contours
The variation of velocity is higher at curved and edge Fig. 9 Top view of Geometry
portions of the surface of the Starfighter. Velocity
increases when Mach number increases. VI. ANALYSIS OF SCALED DOWN MODEL
The model of F104 has been scaled down in ICEM using
geometric scaling by a factor of 4.4. A mesh is then generated
on this scaled down model using the exact technique which
was used earlier. The analysis was carried out on the exact
same cases using the same boundary conditions and solver
setup. Following contours at same conditions are plotted.
A. Pressure Contours
The pressure contours obtained here show the same behavior
as the previous pressure contours.

Fig. 6 Velocity Contours at Mach 0.4 & AoA 5

Seventh International Conference on Aerospace Science & Engineering 54

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

through analysis of the full scale model. These results can

again be refined by using prism mesh.
Another way to get more refined is to take a model with
the complete engine analysis. Since that is an entirely different
task to model the engine we can still consider it as a future
scope of this project.
From the CFD analysis it is evident from the contours that
the flow properties obtained for both the models are quite
comparable with each other.
VIII. CONCLUSION
Fig. 10 Pressure contours at Mach 0.4 & AoA 0
In this research CFD analysis of the F-104 Starfighter is
B. Velocity Contours
conducted, studied and then compared with reference values.
The velocity contours show the patters as observed The design and mesh size played a key role in the accuracy of
previously. The reasons for this behavior has been results. This vintage aircraft which has been an engineering
explained in the analysis of F-104. wonder was analyzed at different flow conditions and the
results obtained were accurate to a great level. The model was
then scaled down. Analysis over this scaled down model in
ANSYS gave us almost the same results as the actual flight
data. This ensures that the testing during design of an aircraft
which is carried out using wind tunnels can be conducted
using numerical equation using certain CFD tools like
ANSYS. This method of testing is not only accurate but also
gives the opportunity to carry out tests on wide range of
conditions.
ACKNOWLEDGMENT
Fig. 11 Velocity Contours at Mach 0.4 & AoA 0 Several people played an important role in the completion
of this research. First we will like to than our parents for their
VII. RESULT AND DISCUSSION complete support. We would like to thank our project
After the simulations are carried out for each case the supervisor and all the seniors that have been always there for
results are validated with existing data of F-104 obtained. us.
TABLE IV RESULTS REFERENCES
Parameter Flight F-104 AAA Scaled [1] J. L. K. Loftin, "NASA SP-468: Quest for Performance The evaluation of
Data Down Modern Aircraft," NASA, Washington D.C., 2004 .
L/D [1] 9.2 9.42 10.42 9.40 [2] R.K.Heffly, "Aircraft Handling Qualitiess Data," NASA, 1972.
CDo [2] 0.0172 0.018 0.0205 0.0185
[3] I. Arif, "Computational Analyis of Integrated Engine Exhaust Nozzel on a
CMα [2] -1.5 -1.64 -1.66 -1.59 sSupersonic Fighter Aircraft," Jornal of Applied Fluid Mechanics , vol. 11, 2018.
CLα [2] 4.44 4.25 4.24 4.01
[4] B. Kodavanla, "CFD Analysis of F-104 Falcon," Internation Journal of
Mechanical and Production Engineering Research and Development , vol. 8, no.
It is evident that the results obtained from CFD analysis of 2, 2018.
the aircraft are very much close to that obtained from the flight [5] M.R, "Role of Computational Fluid Dynamics and Wind Tunnel in Aeronautics
data. However the little difference we observe can be removed R&D," 2012.
by using prism mesh instead of tetrahedral mesh. But in our [6] M. Sharma, "Flow analysis over F-16 Aircraft wing using Computational Fluid
analysis we had limited computational power. Dynamics," International Journal of Emerging Technology and Advanced
Engineering , 2013.
The results obtained from AAA are also very close
however the geometry in AAA is not very accurate this is the [7] P. Bulat, "Comparison of turbulence models in the calculation of Supersonic
Separate flows," World Applied Siences Journal .
reason the drag here is a little more compared to original
results. [8] H.S.Jr., "Flight Testing the F-104 Auomatic Flight Control System".
It is also apparent that the results for scaled down model
are also close to those obtained from flight and those obtained

Seventh International Conference on Aerospace Science & Engineering 55

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

CFD Analysis of Multi-Element Natural Laminar Flow

Airfoil NLF (1)-0115 for a Fixed-Wing UAV
Absar Ahmed Khan
Institute of Space Science and Technology (ISST), University of Karachi
Karachi, Pakistan
[email protected]

Abstract—As the use of composite structures in the I. INTRODUCTION

aerospace vehicles have significantly increased over the years, it
has now become possible to acquire levels of surface smoothness If the flow over the wings is laminar, fast-moving
and tolerances by using the laminar airfoil. Studies have shown Unmanned Aerial Vehicle (UAV) may achieve significant
that the use of Natural Laminar Airfoil (NLF) results in gains in cruise efficiency. Active laminar flow control may
substantial gains in the performance of aerospace vehicle. On avoid turbulent flow across the majority of the wing. However,
the other hand, the deployment of slats and flaps in NLF airfoils the system has certain drawbacks as the take-off weight is
is a fairly novel concept allowing substantial enhancement in the increased, while also increasing the cost and making
aerodynamic performance as compare to a single-element NLF maintenance more difficult [1]. On the other hand, Natural
airfoil. Of late, various attempts have been made to exploit NLF laminar flow (NLF) is a potential alternative to active laminar
airfoils, however many studies have shown that the harnessing flow control. NLF makes use of an airfoil design that promotes
of optimized results has not met notable success. For instance, as laminar flow to extend farther across the airfoil than other
the laminar flow reaches the leading-edge (LE), it faces traditional forms. Over the majority of the chord, the airfoil
significant loss owing to the contamination at the LE that results form must maintain a favorable pressure gradient. While the
in the reduction of the lift coefficient (Cl). This creates hazardous benefits of laminar flow for increasing aerodynamic efficiency
situation during the takeoff and landing. In this paper, the are well recognised, the issue of sustaining natural laminar
NLF(1)-0115 with max thickness 15% at 44.1% chord and max
flow at higher Reynolds numbers and Mach values associated
camber 1.8% at 30% chord is used to design a 2-D multi-element
airfoil with slat and flap for an Unmanned Aerial Vehicle
with fast-moving UAVs has never been research in great
(UAV). The multi-element airfoil is designed and analysed on details.
Ansys Fluent using Spalart-Allmaras (S-A) turbulence model NASA has experienced the NLF for over five decades and
for various Reynolds and Mach numbers where the loss of lift the results suggest that the use of NLF airfoils reduce the
coefficient due to contamination is expected to be minimal, viscous drag on the aircrafts [2]. While there have been classic
simultaneously the low drag coefficient is expected to occur. concerns about the achievability and maintainability of NLF
While the low-drag benefit as a result of the laminar flow could
wings, the later studies conducted by NASA have suggested
be achieved during the cruise flight, the use of NLF airfoil in a
UAV shows rich potential for meeting mid and far-term
that the NLF airfoils have demonstrated satisfactory
objectives of lowering the UAV energy consumption, whether performance on modern metal and composite airframe
fossil or electric. surfaces [3, 4, 5, 6, 7, 8]. These experiments by NASA, more
than 30 in total, were carried out using the varying freestream
Keywords— Natural Laminar Flow, Multi Element Airfoil, conditions with the maximum Mach number achieved was 0.7.
Ansys Fluent, Unmanned Aerial Vehicle, NLF(1)-0115, Unlike the difficulties encountered on older airframe
Aerodynamics, Energy
surfaces during the second world war up to 1950s, NLF is now
NOMENCLATURE possible owing to the modern technology used in the
fabrication and manufacturing of wings. While NLF has been
Cl = Lift coefficient tested to reduce drag by using the modern composites for wing
Cd = Drag coefficient surfaces and with the Mach up to 0.7, this study tests the
aerodynamic performance of NLF(1)-0115 airfoil beyond
Cl,max= Maximum lift coefficient of airfoil
Mach 0.7. While NLF has shown persistent and durable results
∝ = Angle of Attack / Alpha / AoA at high-speed subsonic conditions, this study entails the use of
𝑣 = Fluid velocity
supersonic conditions of Mach 1.18 to understand the
aerodynamic behavior of the airfoil. While NLF airfoils have
𝜌 = Density been tested by NASA for various chord Reynolds numbers as
𝜇 = Kinematic viscosity large as 30 x 106, this research has numerically tested the
airfoil for Re 38.5 x 106 to identify the aerodynamic effects on
LE = Leading Edge the airfoil despite the presence of turbulent layer.
TE = Trailing Edge Other studies by NASA have shown that the usually low lift
UAV =Unmanned Aerial Vehicle coefficients and consequently low generated drag, form drag
generated on wings is the greatest contribution to overall
Ma = Mach Number
aircraft drag under cruising conditions for most aircraft [9].
Re = Reynolds Number For cargo aircraft, wing profile drag accounts for
NLF = Natural Laminar Flow approximately a third of overall drag. With the decrement in
the size of the aircraft, from commercial airplanes to UAVs,
Cp = Pressure Coefficient
the profile drag tends to increase, as shown in the Table 1.
L/D = Lift-to-Drag Ratio / Aerodynamic Efficiency

Seventh International Conference on Aerospace Science & Engineering 56

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Various approaches have been identified to reduce the

form drag on the wings. One of the most known approaches is
the deployment of high-lift system, such as slats and flaps, so
the lift coefficient can be maximised [10]. When compared to
a traditional turbulent-flow wing with no high-lift technology
deployed on the wings, this strategy can cut form drag by
nearly half. While the cost of manufacturing an aircraft rises
as high-lift devices are deployed, the promised returns in terms
of reaching maximum aerodynamic efficiency are more
tempting. The employment of Multi Elements – Slats and
Flaps – with the NLF airfoil results in pressure recovery that a
single-element NLF airfoil cannot achieve [10]. Multi-
Element NLF airfoils have been employed in a variety of Fig. 1. NLF(1)-0115 Airfoil
aircraft types, including business jets [11] and rotorcraft
applications [12], however most of their usage have been in
subsonic conditions. This does not suggest that the NLF
airfoils have never been used in supersonic conditions for
optimum aerodynamic efficiency. The primary idea behind
Multi-Element NLF airfoils is to take use of favorable pressure
gradients to assist stabilize the boundary layer without causing
turbulent separation [13].
Studies have shown that Multi-Element airfoils can attain
a Cl,max > 2, allowing the potential benefit to reduce the wing
planform area or enabling the use of light weight high-lift
system. These benefits for Multi-Element NLF comes
without any trade-offs and compromises, thus allowing the
designer a greater degree of freedom [14].
This work focuses on the design and analyses of the multi-
element NLF(1)-0115 airfoil with max thickness 15% at
44.1% chord and max camber 1.8% at 30% chord. Within our
Fig. 2. C-grid with multi-element NLF(1)-0115 in center
design, a 2-D multi-element airfoil for an Unmanned Aircraft
System (UAV) is being hypothesised. Various studies on the The leading edge slat and trailing edge flap deployed on
NLF airfoils have been conducted [15, 16, 17, 18, 19], whereas the NLF(1)-0115 airfoil are shown in the following Fig. 3.
no study has been found to have conducted analysis on
NLF(1)-0115 by deploying high-lift multi-elements, viz., slats
and flaps.

TABLE I. FORM DRAG ON VARIOUS AIRCRAFTS [9]

Type of Aircraft Form Drag on Wings / Total
Aircraft Drag Fig. 3. Multi-element NLF(1)-0115 airfoil with leading edge slat and
Sailplane > 1/2 trailing edge flap
Unmanned Aerial Vehicle (UAV) 1/3 to 1/2
Transport ~ 1/3 To investigate the effect of leading edge slat and trailing
Low-speed general aviation > 1/3
edge flap, the mesh is generated on Ansys Fluent. The
Business jet ~ 1/3
triangular mesh was generated with varying number of
inflation layers and first layer thickness based on the y+
II. GEOMETRY AND GRID GENERATION values. The y+ value for various cases ranged between 1 and
30. Following Fig. 4a, b, c and d highlight the generated mesh
The airfoil studied for this research is NLF(1)-0115 with on the c-domain. The number of elements generated at which
max thickness 15% at 44.1% chord and max camber 1.8% at Cl,max was obtained were 101,724. The mesh was evaluated for
30% chord (Fig. 1). The main element of the airfoil, as skewness and orthogonal quality to ensure that the
presented in the Fig. 1, has chord length of 1.4 meters. The convergence and results do not falter. The mesh skewness
airfoil is placed in a C-type domain while the whole geometry ranged between 0.11 to 0.78 and show in Fig. 4e, whereas the
is processed on Ansys SpaceClaim. mesh orthogonal quality remained between 0.37 to 0.95 as
To capture the fluid flow behavior as suggested in various shown in Fig. 4f [32, 33]. The average wall y-plus value was
research studies [20, 21, 22, 23, 24], we have used semicircle 3.7 and shown in Fig. 4g.
with the radius of 13.5m and the rectangular grid of 27m as
shown in Fig. 2.20
The primary reason the C-grid is used for 2-D airfoils is
for the fact that the curvature of the mesh grid has to
correspond with the leading edge of the airfoil. Studies have
shown that this allows to capture better separation of flow.
Besides, the mesh does not need to be updated as alpha is
changed [43].

Seventh International Conference on Aerospace Science & Engineering 57

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 4a. Mesh generation over main element airfoil Fig. 4d. Mesh generation over trailing edge flap

Fig. 4b. Mesh generation over the trailing edge Fig. 4e. Mesh Skewness Quality

Fig. 4f. Mesh Orthogonal Quality

Fig. 4c. Mesh generation over leading edge slat

Seventh International Conference on Aerospace Science & Engineering 58

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Outlet was assigned as ‘Pressure Outlet’, whereas the Airfoil

was set to ‘Wall’. At the outset of the Fluent, the ‘Double
Precision’ was selected. The use of Double Precision while
raised the convergence time with more CPU usage, it is
deemed necessary to resolve the pressure differences that
drive the flow [27, 28, 29]. Convergence based on double
precision is believed to be more accurate [30, 31]. Since four
cases with varying Mach numbers were studied and the flow
appeared compressible, therefore the “Pressure-based” solver
was used. The standard sea-level density and dynamic
viscosity of 1.225 kg/m3 and 1.7894e-5 kg/m-s was used
respectively. The increased under-relaxation was applied by
using the SIMPLE-Consistent or Semi-Implicit Method for
Pressure-linked Equations: Consistent (SIMPLEC) that is a
pressure-velocity coupling method. SIMPLEC may frequently
find a converged solution more rapidly for very simple
situations (laminar flows with no extra models active) when
convergence is restricted by the pressure-velocity coupling
Fig. 4g. Wall Y-plus at Ma=0.71 at ∝ = 5°
[32, 33]. The pressure-correction under-relaxation factor in
SIMPLEC is set to 1.0 to enhance the convergence process
III. COMPUTATIONAL METHODOLOGY and get the solution in time. Increasing the pressure-correction
under-relaxation to 1.0, on the other hand, may cause
A. Turbulence Modelling instability in certain situations owing to excessive mesh
In order to study the effect of leading edge slat and trailing skewness [34, 35]. Using the SIMPLEC method was justified
edge flap on the NLF airfoil and evaluate its aerodynamic with the use of Mesh Adapt or Adaptive Meshing approach
performance, the Spalart-Allmaras (S-A) turbulence model that was used to improve the mesh quality, specifically
was used. The Spalart-Allmaras is Reynolds Average Navier- skewness and orthogonal quality [36, 37]. The convergence
Stokes (RANS) one-equation model used for a modified criteria is set to 1e-5 for all the residuals.
turbulence kinematic viscosity (𝑣̃). S-A model allows the use
IV. COMPUTATIONAL RESULTS
of less computational effort as compared to other turbulence
model. However, the S-A model has extensively been used A. Section Characteristics
within the CFD to study the external flow to study various Four different test cases with varying Mach and Reynolds
aerospace problems and the results have also been validated Numbers were used to identify the fluidic and aerodynamic
with the experimental research over the decades [39, 40, 41, behavior of the NLF(1)-0115 multi-element airfoil. The
42]. The standard model is given in the following transport section characteristics for the NLF(1)-0115 airfoil for various
equation (1). Mach and Reynolds numbers are presented in the Fig. 5 and 6.
𝐷𝑣̃ 1
The lowest Mach and Reynolds number tested respectively is
= 𝑃 − 𝐷 + [∇. ((𝑣 + 𝑣̃) + 0.19 and 6.47 x 106. The highest aerodynamic efficient (L/D)
𝐷𝑡 𝜎
𝐶𝑏2 (∇𝑣̃)2 ]   achieved in this low Mach and Reynolds number
configuration and plotted in the Fig. 7 is 10.20 at ∝ = 10° ,
where, while the maximum lift coefficient (Cl,max) of 2.50 was
achieved at ∝ = 17° (Fig. 8). As the Mach and Reynolds
𝑑 = distance between airfoil wall point and solid wall number is increased to 0.42 and 1.37 x 107 respectively, it is
𝑣=
𝜇
= laminar kinematic viscosity identified that the maximum aerodynamic efficiency (L/D)
𝜌 reached to ∝ = 15° , while the maximum lift coefficient
𝑢 = fluid velocity (Cl,max) of 2.63 was achieved at ∝ = 18° . When the Mach and
Reynolds numbers are further increased to 0.71 and 2.31 x 107
B. Test Cases respectively, the maximum aerodynamic efficiency (L/D) of
The test cases studied are presented in the following Table 11.71 was obtained at ∝ = 15° and plotted in the Fig. 11,
II: whereas the maximum lift coefficient (Cl,max) of 2.68 was
achieved at ∝ = 18° (Fig. 12). Further introducing the Mach
TABLE II. TEST CASES SUMMARY and Reynolds numbers to 1.18 and 3.85 x 107, it is noticed that
Case Mach Reynolds the maximum aerodynamic efficiency (L/D) ratio as plotted in
No. No. No. Speed
Fig. 13 has increased to 12.08 but has remained existent at ∝
1 0.19 6.47 x of = 15° , similar to the preceding Mach numbers tested;
106 Sound: simultaneously, the maximum lift coefficient (Cl,max) of 2.72
2 0.42 1.37 x 343m/s
107 Length:
has performed similarly. The Cl vs alpha for Ma=1.18 is
3 0.71 2.31 x 1.39m plotted in Fig. 14.
107
4 1.18 3.85 x
107

C. Boundary Layer and Solver Settings

The airfoil boundary is a solid wall and it was set for ‘no-
slip condition’. The Inlet was assigned was ‘Velocity Inlet’,

Seventh International Conference on Aerospace Science & Engineering 59

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

2.8
2.8 Ma=0.19 Ma=0.19
Ma=0.42 2.6
2.6 Ma=0.71
Ma=1.18 2.4
2.4
2.2
2.2
2.0
2.0
Cl

Cl
1.8
1.8

1.6 1.6

1.4 1.4

1.2 1.2

1.0 1.0

0 5 10 15 20 0 2 4 6 8 10 12 14 16 18 20 22
Alpha (Degree) Alpha (Degree)

Fig. 5. Alpha vs Cl Fig. 8. Cl vs Alpha (Ma=0.18)

3.0
Ma=0.19 12
Ma=0.42
Ma=0.42 11
Ma=0.71
2.5 Ma=1.18 10
Coefficient of Lift (Cl)

8
2.0
L/D 7

6
1.5 5

3
1.0
2

0.2 0.4 0.6 0.8 1

0 2 4 6 8 10 12 14 16 18 20 22
Coefficient of Drag (Cd)
Alpha (Degree)
Fig. 6. Cd vs Cl

Fig. 9. Alpha vs L/D (Ma=0.42)

11
Ma=0.19
10 2.8
Ma=0.42
9 2.6

8 2.4

7 2.2
L/D

6 2.0
Cl

5 1.8

4 1.6

3 1.4

2 1.2

0 2 4 6 8 10 12 14 16 18 20 1.0
Alpha (Degree)
0 2 4 6 8 10 12 14 16 18 20 22
Alpha
Fig. 7. Alpha vs L/D (Ma=0.19)

Fig. 10. Cl vs Alpha (Ma=0.42)

The performance improvements offered by the multi-

element NLF airfoils may be quantified from the Breguet
range equation (equation 2) that significantly influences the
design of modern aircrafts, including engines.

𝑅𝑎𝑛𝑔𝑒 =
𝑀𝐿 1 𝑊𝑖
𝑎∞ ( )( ) 𝑙𝑛 ( )  
𝐷 𝑇𝑆𝐹𝐶 𝑊𝑓

Seventh International Conference on Aerospace Science & Engineering 60

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

For aircrafts using non-NLF airfoils, for instance, Boeing

747-400 at an altitude of 31,000 ft, there is an increase in L/D 13
until the airfoil stalls and the decrement starts happening. At Ma=1.18
12
low Mach numbers, a non-NLF aircraft encounters boundary 11
layer separation at the trailing edge, but at high Mach 10
numbers, shock waves develop on the wing. For these 9
aircrafts having non-NLF airfoil, the aerodynamic 8

L/D
performance (ML/D) is maximum at lift coefficient (Cl) of 7

0.5. But then, the NLF airfoils exhibit a different 6

phenomenon, as the numerical results obtained suggest. The 5

maximum aerodynamic performance (ML/D) has 4

consistently increased as the velocity reaches the speed of 3

sound (343 m/s). It is identified that the maximum ML/D in 2

1
the lowest calculated Ma=0.19 is 2.02. As the Mach number 0 2 4 6 8 10 12 14 16 18 20 22
is increased to 0.42, the maximum ML/D reaches 4.70. When Alpha (Degree)
Ma=0.71, the ML/D gets as high as 11.71. However, at
Ma=1.18, the maximum ML/D is 12.08. The aerodynamic
Fig. 13. Alpha vs L/D (Ma=1.18)
Fig. of merit for the various tested Mach numbers is presented
in the Fig. 15.

12 2.8 Ma=1.18
Ma=0.71
11 2.6
10
2.4
9
2.2
8
2.0
Cl

7
L/D

1.8
6

5 1.6

4 1.4
3 1.2
2
1.0
1 0 2 4 6 8 10 12 14 16 18 20 22
0 2 4 6 8 10 12 14 16 18 20 22
Alpha (Degree)
Alpha (Degree)

Fig. 14. Alpha vs Cl (Ma=1.18)

Fig. 11. Alpha vs L/D (Ma=0.71)

16
15
Ma=0.19
2.8 14
Ma=0.42
Ma=0.71
13 Ma=0.71
2.6 12 Ma=1.18
11
2.4
10
2.2 9
ML/D

8
2.0 7
Cl

6
1.8
5
1.6 4
3
1.4 2
1
1.2
0
1.0 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8
0 2 4 6 8 10 12 14 16 18 20 22 Coefficient of Lift (Cl)
Alpha (Degree)
Fig. 15. Aerodynamic Fig. of Merit of Multi-Element NLF(1)-0115
Fig. 12. Alpha vs Cl (Ma=0.71) Airfoil

The CFD results of boundary layer separation for various

Mach numbers and alpha is presented in the Fig. 16.

Seventh International Conference on Aerospace Science & Engineering 61

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 16a. Boundary layer separation for Ma=0.18 at 5 Alpha Fig. 16f. Boundary layer separation for Ma=0.42 at 18 Alpha

Fig. 16b. Boundary layer separation for Ma=0.18 at 15 Alpha Fig. 16g. Boundary layer separation for Ma=0.71 at 2 Alpha

Fig. 16c. Boundary layer separation for Ma=0.18 at 18 Alpha Fig. 16h. Boundary layer separation for Ma=0.71 at 10 Alpha

Fig. 16d. Boundary layer separation for Ma=0.42 at 2 Alpha Fig. 16i. Boundary layer separation for Ma=0.71 at 18 Alpha

Fig. 16e. Boundary layer separation for Ma=0.42 at 10 Alpha Fig. 16j. Boundary layer separation for Ma=1.18 at 2 Alpha

Seventh International Conference on Aerospace Science & Engineering 62

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 16k. Boundary layer separation for Ma=1.18 at 10 Alpha

Fig. 17c. Effect of Ma=0.71 on various Alpha

Fig. 16l. Boundary layer separation for Ma=1.18 at 18 Alpha

B. Pressure Distribution
The plotting for the pressure distribution on the multi-
element airfoil is presented in the Fig. 17. The effects of
increasing Mach number on the pressure distribution is rather
prominent.
One particular observation of increasing Mach number on
pressure distribution is the increase in lift coefficient (Cl). The
origin of this behavior is seen in the Fig. 15 when the Mach
number is increased from 0.19 to 0.42 and then to 0.71, Fig. 17d. Effect of Ma=1.18 on various Alpha
ultimately reaching 1.18.
In all the Mach number test cases, the increase in pressure
coefficient (Cp) has gone up between 𝛼 = 5° and 𝛼 = 10° .
However, while the pressure has increased beyond up until the
stall angle is achieved, what is critically important to note is
that little influence of increase in Mach number at the slot
between the slat and main element and flap and main element,
except when the pressure distribution beyond 𝛼 = 10° . It is
also identified that the increase in freestream Mach number
creates additional flow acceleration on the upper surface of the
main element of airfoil. This phenomenon helps in stabilizing
the boundary layer while it also promotes delaying transition
[38].
From the pressure distributions plotted in the Fig. 17 for
various alpha and corresponding Mach numbers, one critical
Fig. 17a. Distribution of Pressure at Ma=0.19 on various Alpha observation is made. The pressure coefficient tends to increase
at the aft element (flap) rather than the fore element (slat). This
also suggests that the NLF may perform well only when the
aft element is deployed and fore element is removed from the
current configuration. This theory demands further
investigation.
When the results of the NLF(1)-0115 multi-element airfoil
are compared against the single element NLF(1)-0115 airfoil
using the XFoil [44], the results suggest that at Ma=0.19 and
Re=4.65x106, the Cl at ∝ = 5° is 0.77, which in comparison
with the multi-element airfoil of similar Mach number but
different Reynolds number of 6.47x106 is 100.529% less.
Similarly, at Mach number of 0.4 and Reynolds number of
9.86x106, the Cl at ∝ = 10° is 1.23, whereas for similar Mach
number but Reynolds number of 1.37x107, the Cl is 2.06. The
results are plotted in the following Fig. 18 and 19.
Fig. 17b. Effect of Ma=0.42 on various Alpha

Seventh International Conference on Aerospace Science & Engineering 63

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
Ma=0.19
Ma=0.71
the aft element is deployed and the fore element is removed
65
Ma=0.42
Ma=1.18 from the configuration. Further studies may be required to
60 computationally and experimentally investigate this
60
hypothesis to not only obtain a boundary layer separation
55
55 without turbulence at the trailing edge but also the
50
aerodynamic optimization of the multi-element airfoil itself.
50

45 REFERENCES

L/D
45
L/D

[1] L. Steers, "Natural Laminar Flow Flight Experiment", NASA Dryden

40 40
Flight Research Center, 1981. [Online]. Available:
35 https://ntrs.nasa.gov/api/citations/19810010484/downloads/19810010
35
484.pdf. [Accessed: 03- Sep- 2021].
30
30 [2] B. Holmes, C. Croom, E. Hastings, C. Obara and C. Vandam, "Flight
25
research on natural laminar flow", NASA Langley Research Center,
25 1986. [Online]. Available:
4 5 6 7 8 9 10 11 12 13 14 15 16 https://ntrs.nasa.gov/api/citations/19880005568/downloads/19880005
Alpha 568.pdf?attachment=true. [Accessed: 15- Aug- 2021].
[3] B. Holmes, C. Obara, G. Gregorek, M. Hoffman and R. Freuhler,
"Flight Investigation of Natural Laminar Flow on the Bellanca
Fig. 18. Alpha vs L/D of NLF(1)-0115 [44] Skyrocket II", SAE Technical Paper Series, 1983. Available:
10.4271/830717 [Accessed 4 September 2021].
Ma=0.19
Ma=0.71 [4] B. Holmes and C. Obara, "Observations and implications of natural
1.42 Ma=0.42
Ma=1.18
laminar flow on practical airplane surfaces", Journal of Aircraft, vol.
1.4 20, no. 12, pp. 993-1006, 1983. Available: 10.2514/3.48203 [Accessed
1.40
2 September 2021].
1.3 1.38 [5] B. Holmes, C. J. Obara and L. Yip, "Natural laminar flow experiments
1.36 on modern airplane surfaces", The Pennsylvania State University,
1.2 1984. [Online]. Available:
1.34 https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.731.3131
1.1 &rep=rep1&type=pdf. [Accessed: 18- Aug- 2021].
Cl
Cl

1.32

1.0 1.30
[6] C. J. Obara and B. Holmes, "Flight-measured laminar boundary-layer
transition phenomena including stability theory analysis", Stanford
1.28 University, 1985. [Online]. Available:
0.9
1.26
https://searchworks.stanford.edu/view/5714059. [Accessed: 09- Jul-
0.8 2021].
1.24
[7] D. WARD and S. MILEY, "An evaluation of flight test measurement
0.7 1.22 techniques for obtaining airfoil pressure distributions and boundary
4 5 6 7 8 9 10 11 12 13 14 15 16 layer transition locations", 2nd Flight Testing Conference, 1983.
Alpha Available: 10.2514/6.1983-2689 [Accessed 4 August 2021].
[8] W. Wentz, A. Ahmed and R. Nyenhuis, "Further Results of Natural
Laminar Flow Flight Test Experiments", SAE Transactions, vol. 94, pp.
Fig. 19. Alpha vs Cl of NLF(1)-0115 [44]
508-521, 1985. [Accessed 4 September 2021].
[9] D. Somers, "Design of a Slotted, Natural-Laminar-Flow Airfoil for
V. CONCLUSION Business-Jet Applications", Airfoils Incorporated, Port Matilda,
The multi-element Natural Laminar Flow NLF(1)-0115 Pennsylvania, 2019. [Online]. Available:
https://ntrs.nasa.gov/api/citations/20190031816/downloads/20190031
airfoil for a fast moving UAV was designed and 816.pdf. [Accessed: 11- Jul- 2021].
computationally evaluated in this research. It has been
[10] A. M. O. Smith, "High-Lift Aerodynamics", Journal of Aircraft, vol.
identified that the deployment of multi-element – the leading 12, no. 6, pp. 501-530, 1975. Available: 10.2514/3.59830 [Accessed 2
edge slat and trailing edge flap – allows the improvement of June 2021].
aerodynamic efficiency (L/D) and the aerodynamic [11] Y. Shi, C. Mader, S. He, G. Halila and J. Martins, "Natural Laminar-
performance (ML/D) as the Mach number is increased from Flow Airfoil Optimization Design Using a Discrete Adjoint
0.19 to 1.18. The computational results also demonstrated that Approach", AIAA Journal, vol. 58, no. 11, pp. 4702-4722, 2020.
the trailing edge boundary layer separation at high Mach Available: 10.2514/1.j058944 [Accessed 14 June 2021].
number has little turbulence, inherently owing to the [12] J. Coder, M. Maughmer and D. Somers, "Theoretical and Experimental
Results for the S414, Slotted, Natural-Laminar-Flow Airfoil", Journal
aerodynamic characteristics of the natural laminar flow airfoil of Aircraft, vol. 51, no. 6, pp. 1883-1890, 2014. Available:
NLF(1)-0115. Through the deployment of the slat and flap, the 10.2514/1.c032566 [Accessed 17 June 2021].
separation of boundary layer that create turbulence is also [13] R. T. Whitcomb and L. R. Clark, An Airfoil Shape for Efficient Flight
improved. The multi-element configuration enables at Supercritical Mach Numbers, 1st ed. NASA technical memorandum,
significant improvement in the NLF airfoil. The maximum 1965, pp. 1-57.
aerodynamic performance (ML/D) from the Breguet range [14] . Coder and D. Somers, "Design of a slotted, natural-laminar-flow
equation achieved was 14.26 at Mach number 1.18 and 𝛼 = airfoil for commercial transport applications", Aerospace Science and
Technology, vol. 106, p. 106217, 2020. Available:
15° which when compared to the lowest tested Mach number 10.1016/j.ast.2020.106217 [Accessed 16 June 2021].
at 0.19 where ML/D was highest at 2.02 at 𝛼 = 10° . The [15] M. Fujino, Y. Yoshizaki and Y. Kawamura, "Natural-Laminar-Flow
results show high lift coefficients at relatively low value of Airfoil Development for a Lightweight Business Jet", Journal of
drag coefficient. It was also identified that achieving Aircraft, vol. 40, no. 4, pp. 609-615, 2003. Available: 10.2514/2.3145
significant laminar flow behavior on the lower surface of the [Accessed 4 September 2021].
multi-element airfoil is fundamentally challenging. [16] Y. Shi, C. Mader, S. He, G. Halila and J. Martins, "Natural Laminar-
Flow Airfoil Optimization Design Using a Discrete Adjoint
Results have confirmed that there is an increase in pressure Approach", AIAA Journal, vol. 58, no. 11, pp. 4702-4722, 2020.
on the lower surface of the aft element. This may be suggestive Available: 10.2514/1.j058944 [Accessed 4 September 2021].
of the idea that the NLF airfoils may perform better only when

Seventh International Conference on Aerospace Science & Engineering 64

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

[17] Z. Han, J. Chen, K. Zhang, Z. Xu, Z. Zhu and W. Song, "Aerodynamic https://www.afs.enea.it/project/neptunius/docs/fluent/html/ug/node78
Shape Optimization of Natural-Laminar-Flow Wing Using Surrogate- 5.htm. [Accessed: 17- Aug- 2021].
Based Approach", AIAA Journal, vol. 56, no. 7, pp. 2579-2593, 2018. [35] M. Mohamed, G. Janiga, E. Pap and D. Thévenin, "Multi-objective
Available: 10.2514/1.j056661 [Accessed 4 September 2021]. optimization of the airfoil shape of Wells turbine used for wave energy
[18] M. Selig, M. Maughmer and D. Somers, "Natural-laminar-flow airfoil conversion", Energy, vol. 36, no. 1, pp. 438-446, 2011. Available:
for general-aviation applications", Journal of Aircraft, vol. 32, no. 4, 10.1016/j.energy.2010.10.021 [Accessed 25 June 2021].
pp. 710-715, 1995. Available: 10.2514/3.46781 [Accessed 4 September [36] T. Fondelli, A. Andreini and B. Facchini, "Numerical Simulation of
2021]. Dam-Break Problem Using an Adaptive Meshing Approach", Energy
[19] M. Selig, M. Maughmer and D. Somers, "Natural-laminar-flow airfoil Procedia, vol. 82, pp. 309-315, 2015. Available:
for general-aviation applications", Journal of Aircraft, vol. 32, no. 4, 10.1016/j.egypro.2015.12.038 [Accessed 4 June 2021].
pp. 710-715, 1995. Available: 10.2514/3.46781 [Accessed 18 August [37] T. Michal et al., "Comparing Unstructured Adaptive Mesh Solutions
2021]. for the High Lift Common Research Airfoil", AIAA Journal, vol. 59,
[20] S. Xiao and Z. Chen, "Investigation of Flow over the Airfoil NACA no. 9, pp. 3566-3584, 2021. Available: 10.2514/1.j060088 [Accessed
0010-35 with Various Angle of Attack", MATEC Web of Conferences, 25 July 2021].
vol. 179, p. 03020, 2018. Available: [38] J. Coder and D. Somers, "Design of a slotted, natural-laminar-flow
10.1051/matecconf/201817903020 [Accessed 24 June 2021]. airfoil for commercial transport applications", Aerospace Science and
[21] A. Diab, M. Alaa, A. Hossam El-Din, H. Salem and Z. Ghoneim, Technology, vol. 106, p. 106217, 2020. Available:
"Performance Degradation of Wind Turbine Airfoils due to Dust 10.1016/j.ast.2020.106217 [Accessed 6 May 2021].
Contamination: A Comparative Numerical Study", Volume 9: Oil and [39] V. D'Alessandro, S. Montelpare, R. Ricci and A. Zoppi, "Numerical
Gas Applications; Supercritical CO2 Power Cycles; Wind Energy, modeling of the flow over wind turbine airfoils by means of Spalart–
2015. Available: 10.1115/gt2015-44012 [Accessed 16 August 2021]. Allmaras local correlation based transition model", Energy, vol. 130,
[22] M. Mohamed and R. Yadav, "Flow separation control using the pp. 402-419, 2017. Available: 10.1016/j.energy.2017.04.134
cetacean species nose design", European Journal of Mechanics - [Accessed 12 July 2021].
B/Fluids, vol. 89, pp. 139-150, 2021. Available: [40] S. Benjanirat, L. Sankar and G. Xu, "Evaluation of Turbulence Models
10.1016/j.euromechflu.2021.05.005 [Accessed 7 August 2021]. for the Prediction of Wind Turbine Aerodynamics", ASME 2003 Wind
[23] I. F. Zidane, K. M. Saqr, G. Swadener, X. Ma and M. F. Shehadeh, Energy Symposium, 2003. Available: 10.1115/wind2003-517
"COMPUTATIONAL FLUID DYNAMICS STUDY OF DUSTY AIR [Accessed 18 July 2021].
FLOW OVER NACA 63415 AIRFOIL FOR WIND TURBINE [41] F. Rizzo, V. D'Alessandro, S. Montelpare and L. Giammichele,
APPLICATIONS", Jurnal Teknologi, vol. 79, no. 7-3, 2017. Available: "Computational study of a bluff body aerodynamics: Impact of the
10.11113/jt.v79.11877 [Accessed 15 August 2021]. laminar-to-turbulent transition modelling", International Journal of
[24] A. Meana-Fernández, J. Fernández Oro, K. Argüelles Díaz and S. Mechanical Sciences, vol. 178, p. 105620, 2020. Available:
Velarde-Suárez, "Turbulence-Model Comparison for Aerodynamic- 10.1016/j.ijmecsci.2020.105620 [Accessed 6 August 2021].
Performance Prediction of a Typical Vertical-Axis Wind-Turbine [42] M. Costes and F. Moens, "Advanced numerical prediction of iced
Airfoil", Energies, vol. 12, no. 3, p. 488, 2019. Available: airfoil aerodynamics", Aerospace Science and Technology, vol. 91, pp.
10.3390/en12030488 [Accessed 4 September 2021]. 186-207, 2019. Available: 10.1016/j.ast.2019.05.010 [Accessed 25
[25] N. Sorensen et al., "CFD code comparison for 2D airfoil July 2021].
flows", Journal of Physics: Conference Series, vol. 753, p. 082019, [43] S. Mittal and T. Tezduyar, "A finite element study of incompressible
2016. Available: 10.1088/1742-6596/753/8/082019 [Accessed 14 flows past oscillating cylinders and aerofoils", International Journal
September 2021]. for Numerical Methods in Fluids, vol. 15, no. 9, pp. 1073-1118, 1992.
[26] A. Subramanian et al., "Effect of airfoil and solidity on performance of Available: 10.1002/fld.1650150911 [Accessed 3 August 2021].
small scale vertical axis wind turbine using three dimensional CFD [44] Airfoiltools, "NLF(1)-0115 (nlf0115-il) Xfoil prediction polar at
model", Energy, vol. 133, pp. 179-190, 2017. Available: RE=1,000,000 Ncrit=9", Airfoiltools.com, 2021. [Online]. Available:
10.1016/j.energy.2017.05.118 [Accessed 15 August 2021]. http://airfoiltools.com/polar/details?polar=xf-nlf0115-il-1000000.
[27] V. Anil Hemanth and U. Jyothi, "CFD Analysis of a Solid Propellant [Accessed: 17- Nov- 2021].
Retro Rocket Motor using Ansys Fluent", E3S Web of Conferences,
vol. 184, p. 01054, 2020. Available: 10.1051/e3sconf/202018401054
[Accessed 8 August 2021].
[28] P. Badelwa and V. Kesharwani, "Thermal Simulation and an Analysis
on an Exhaust Gas of a Teg Model using Fluent", International Journal
of I.C. Engines and Gas Turbines, vol. 6, no. 2, pp. 1-19, 2020.
[Accessed 14 August 2021].
[29] D. Sofialidis, "HPC with ANSYS Fluent", PRACE Autumn School
2013 – Industry Oriented HPC Simulations, 2013. [Online]. Available:
https://materials.prace-ri.eu/350/1/hpcAnsys.pdf. [Accessed: 18- May-
2021].
[30] A. Serageldin, A. Abdelrahman and S. Ookawara, "Earth-Air Heat
Exchanger thermal performance in Egyptian conditions: Experimental
results, mathematical model, and Computational Fluid Dynamics
simulation", Energy Conversion and Management, vol. 122, pp. 25-38,
2016. Available: 10.1016/j.enconman.2016.05.053 [Accessed 15
August 2021].
[31] P. Kumar and S. Singh, "Flow past a bluff body subjected to lower
subcritical Reynolds number", Journal of Ocean Engineering and
Science, vol. 5, no. 2, pp. 173-179, 2020. Available:
10.1016/j.joes.2019.10.002 [Accessed 11 August 2021].
[32] P. Mondal, M. Rahman and A. Hasan, "Numerical analysis of bio-
inspired corrugated airfoil at low Reynolds number", 2016. Available:
10.1063/1.4958388 [Accessed 11 August 2021].
[33] L. Kharulaman, A. Aabid, F. Mehaboobali and S. Khan, "Research
onFlows for NACA 2412 Airfoil using Computational Fluid Dynamics
Method", International Journal of Engineering and Advanced
Technology (IJEAT), vol. 9, no. 1, 2021. [Accessed 6 August 2021].
[34] Ansys, "Choosing the Pressure-Velocity Coupling Method", National
Agency for New Technologies, Energy and Sustainable Economic
Development, 2009. [Online]. Available:

Seventh International Conference on Aerospace Science & Engineering 65

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

An Overview of Passive Flow Control Techniques

in Supersonic Flows
1st Zukhruf Liaqat 2nd Aamer Shahzad
Department of Aeronautical Engineering Department of Aeronautical Engineering
National University of Sciences and Technology National University of Sciences and Technology
Risalpur, Pakistan Risalpur, Pakistan
[email protected] [email protected]

Abstract—Fluid flow is characterized by active and passive II. DISCUSSION

controls. Active flow control exploits external energy to control
TABLE I: Passive Flow Control Devices with Their Advantages and
flow externally and internally. However, passive flow control
Disadvantages
used physical devices to regulate the flow of the fluid without
incorporating any auxiliary power. Passive control is inherent to CONTROL ADVANTAGES DISADVANTAGES
air vehicles to achieve desired characteristics smoothly. Most of DEVICE
the times, passive flow control devices are installed from the time
Vortex Reduction of stall Inconvenience in
of the manufacturing. These devices also come as external kits.
generators High controllability at washing and/or
The shape and size of the device being installed determines the
low speeds painting the wing or
pressure regulation, here the boundary layer effects are
surfaces where VG is
prominent. Several techniques are discussed in this paper with
installed
a focus on how they achieve desired flow characteristics by
pressure regulation and ultimately contribute positively towards Leading edge Additional lift creation Additional weight and
required flight characteristics. Some common passive flow slats at different flight structural complexity
devices include vortex generators, tabs, leading edge slats, sequences such as Creation of high stall
cavities, and flow vanes. These physical devices do not need an takeoff, climb, angle
external energy and change pressure distribution to a desirable approach and landing
form through geometric changes of the surfaces. The at low speeds
phenomena of flow separation and stall are of particular Cavities Drag reduction Not preferred for high
significance when passive flow techniques are under Noise reduction speed flows
consideration. The weight and drag may experience slight Flow vanes Efficient control High maintenance
increase but a cost-efficiency trade-off is made to achieve device in turbines
optimum results. The aim behind flow control is to regulate Tabs High aerodynamic Added weight leads to
pressure recovery and remove unsteady pressure changes. The performance drag production
techniques of passive flow control are widely used because of Corrugations Enhancement of Increased surface drag
their simplicity, affordability, and accessibility. aerodynamic leads to increased fuel
Keywords—Auxiliary energy, boundary layer, control performance at gliding consumption
devices, flow characteristics, flow separation, passive flow, flight
pressure fluctuations, supersonic regime, vortex generation Low membrane stress

I. INTRODUCTION
In fluid dynamics, flow control is a necessary tool to regulate A. Vortex generators
pressure fluctuations in order to control the total drag of the TABLE II: Specifications of a Vortex Generator
body [1]. The flow controls techniques are primarily used to
SPECIFICATION DESCRIPTION
achieve improved flow characteristics and reduce the
Used for Flow separation
unwanted disturbances such as unsteady pressure changes. Drag generation
The flow control techniques are divided into two types: active Installation Mostly close to leading edge
and passive. Passive control techniques do not require any Various lifting surfaces
auxiliary energy and are achieved by the geometric changes Shapes Triangular (delta, cropped delta)
most of the time. These flow control techniques are used to Rectangular (stretched square, trapezium)
improve the flight conditions of air vehicles [2]. The passive Height Comparable to boundary layer thickness
flow control manipulates the characteristics of the flow in the (for MVGs, << BL thickness)
boundary layer and eventually affects and controls the mixing
of the flow. The passive flow control is achieved by installing Vortex generator is one of the devices used in techniques
several devices which will be briefly discussed in subsequent employed for passive flow control. In simple terms, it delays
sections. These devices are the mechanical options the flow separation thus reducing the formation of trailing
aerodynamicists use to achieve flow characteristics. In edge vortices that ultimately lead to the generation of drag on
contrast, an active flow control technique makes use of the body. It is an aerodynamic device with varying sizes
external energy and is a paid choice usually. Passive flow according to the requirement of the application that is
control techniques have been extensively used due to their attached to delay the flow separation. In aerodynamics
simplicity, affordability, and accessibility. The paper presents applications, the shape is usually of an airfoil attached to a
a review of several passive flow control techniques in the vane that is placed on the aircraft’s wing [3]. The vortex
supersonic regime. generators are usually triangular or rectangular in shape with
a height of approximately equal to boundary layer thickness

Seventh International Conference on Aerospace Science & Engineering 66

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

and placed on the thickest part of the surface over which they enhance the mixing effect of higher momentum external flow
are employed [4]. and the lower momentum flow around the wall [2].

Figure 1: Various shapes of vortex generators Figure 2: Vortex generators rows [4]

They are attached to the wings and vertical tail of most This figure represents the installation of triangular vortex
airliners. generators throughout the wing structure.
Vortex generators are also used on other aerodynamic control
surfaces such as elevators, ailerons and rudders and flaps. The
application of vortex generators is used in turbine blades as B. Leading Edge Slats
well [5]. Leading-edge slats are one of the earliest passive flow control
In most of the applications, the vortex generators are placed techniques employed on aircraft to delay the flow separation
very close to leading-edge to ensure a smooth fluid flow over and check pressure fluctuations in high-speed flows. These
the surface all the way to the trailing edge [6]. The vortex are attached to the leading edge of the airfoils and manipulate
generators are given an angle of attack with respect to the the flow in the space between the main body and slat to delay
local flow to produce an energetic tip vortex that captures the the flow separation. The vortices are generated due to the
surrounding fluid flow towards itself thus removing the local presence of these slots and vary in size while accelerating the
eddies and directing the flow into slow-moving boundary flow. The larger vortices correspond to the presence of a
layer that sweeps across the surface. The vortex generators lower Reynolds number while smaller vortices indicate the
encourage the creation of the turbulent flow that is less prone presence of high Re flow [11]. Due to the installment of slats,
to flow separation than the laminar flow [7]. the flow gains momentum and transports it to the boundary
Passive flow control is inherent to most of the aircraft and layer thus energizing it enough to delay the flow separation.
these vortex generators are present on aircraft from the time The concept of slats in high speed flows is effective since the
of manufacturing. In addition to that, vortex generators are pressure recovery becomes a significant consideration in
also available in VG kits [8, 9]. One problem that may arise supersonic regimes. The increase in drag can be traded off to
with vortex generators installed on the external surfaces is of the advantages provided by the installment of leading-edge
icing due to local flow. However, since the vortex generator slats [12]. Leading-edge slats are effective passive flow
is placed within the boundary layer, the icing due to local cold control techniques with no requirement of external energy
flow does not usually arise and hence extra measures to de- and best suited to delay the stall of the aircraft. They are
ice the surface are not usually needed. The application of typically used in fixed as well as retractable slats in high-
vortex generators is subjected to weight and lift speed flows [13].
considerations [3]. The one added advantage with the
employment of such devices with the addition of low weight
considerations is that they do increase the maximum takeoff
weight but it can be overcome with a tradeoff with enhanced
delay in flow separation and hence drag reduction.
In supersonic flow, shock wave and boundary layer
interactions are common and not only can change the shock
wave structure but also induce accelerated boundary layer
development [10]. This can affect the total pressure recovery
of the system and hence the need for a technique to control Figure 3: Leading-edge slat [7]
the flow arises. The concept of micro vortex generators
(MVGs), although is novel but it has gained much C. Cavities
appreciation in the recent past. They have shown better results TABLE III: Features of Cavities Over Airfoil Surfaces
than conventional suction systems. Usually vortex generators
have comparable height with boundary layer thickness. SPECIFICATION DESCRIPTION
However, VGs smaller in height than BL thickness have also Preferred flight regime Subsonic
shown effective results with enhanced flow control ability Used for Generation of pressure
gradient due to counter
with minimum resistance and such VGs are conveniently
rotating vortices
named as micro vortex generators. MVGs do not decrease
Installation Over airfoil’s suction surface,
the mass flow rate and have greater potential than suction mostly at leading edge
systems since no power penalties are observed in the former. Added benefit Noise reduction
The vortices generated by the MVGs were experimentally
validated where the main purpose of such vortices was to

Seventh International Conference on Aerospace Science & Engineering 67

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Although more preferred for the subsonic regime, cavities are

used for passive flow control in low supersonic flows as well.
The basic principle behind mounting a cavity on an airfoil
surface is to generate a pressure gradient [6]. The pressure
gradient is generated due to the presence of counter-rotating
vortices trapped inside the cavity. The presence of these
trapped vortices ensures the generation of a low-pressure
region and eventually contributes to reducing the total drag.
The cavity created in the airfoil’s suction surface hosts
counter-rotating flows which delay the stall phenomenon thus
contributing positively to the lift production [14]. Cavitied Figure 5: Flow vane over a simplified wing showing the vane and wing
surface are attached at root of the flow vane [18]
airfoils do not work efficiently since the high speed will
induce higher pressure fluctuations and geometric changes
E. Tabs
will be sensitive to such fluctuations. Near transonic and low
TABLE V: Salient Features of a Typical Tab
supersonic flows, they prove to be effective devices for
passive flow. The cavity can be introduced at leading or SPECIFICATION DESCRIPTION
trailing edge, however, leading-edge modifications have been Installment Trailing edge of lifting
more successful for the transonic regime than their trailing surface
edge counterparts.[15] Not only the cavitied airfoils improve Vortex generation Counterclockwise
flight characteristics through passive flow control, but these Size and position Determine flow mixing in
are also an effective way to reduce noise disturbances in the high and low momentum
flow. Since open cavities in supersonic flows can create regions
further problems of unsteady pressure tones, spoilers are used Height Comparable to boundary
to reduce the amplitudes of such disturbances. [16] layer thickness

Tabs or lift-producing tabs are installed on the trailing

surfaces of the lifting devices in the form of small plates to
provide effective results for flow control. Again, these tabs
do not need an external power and control flow through the
geometric changes of the body they are attached with [1]. The
flap size, height and position relative to the flow determine
Figure 4: Generation of vortices in an airfoil [16]
the mixing of the flow in high and low momentum regions.
D. Flow vanes The height of the tab is measured relative to the boundary
layer to have a perspective about the impact on the boundary
TABLE IV: Specification of a Flow Vane
layer and eventually on flow separation. A typical example of
SPECIFICATION DESCRIPTION tabs providing passive flow control is Gurney flap [19]. After
Commonly used in Turbines installation, counter-rotating vortices are produced behind the
Commonly used for Delaying stall and pressure flap. These flaps encourage flow to stay attached with the
regulation surface thus delaying the flow separation [20].
Size Chord length comparable to The tabs are installed mainly on the wing airfoils but are also
attached surface used on the tail surface and have shown commendable results
Installation Above suction surface of for the increase in the lift at an angle of attack with little
airfoil increase in lift coefficient with respect to baseline position
[1]. These tabs are known as lift-enhancing tabs for the same
A typical passive flow control device for turbines, flow vanes reason since they favor drag reduction due to their
benefit the flow by delaying the stall and improving pressure performance on the main airfoil as well as on the tail flaps.
regulation by catering for excessive and unsteady fluctuations
in pressure. [15, 17]. The flow vanes have small yet
comparable chord length with the main body on which these
are installed [18]. They are placed above the suction surface
of the airfoil and delay the flow separation thus improving the
characteristics of the flow.

Figure 6: Lift enhancing tab on an airfoil along with counterclockwise

vortices [10]

The passive flow control techniques makes use of a variety of

different devices such as grooves and ribs, also of high
significance in supersonic regime [21]. The principle behind
using these devices is to achieve maximum pressure recovery

Seventh International Conference on Aerospace Science & Engineering 68

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

using least external energy. In many cases, this means a trade- applicable for both high and low Re flows. Cavities are
off between cost and efficiency. another passive flow control device that use trapped counter-
rotating vortices to achieve required flow characteristics.
F. Corrugations
Among other devices, flow vanes are another common name
TABLE VI: Features of Corrugation
typically used for turbine flows hence remain in near
SPECIFICATION DESCRIPTION transonic regions [24]. The main aim of flow vanes is to delay
Used for Passive flow control through the stall. Many of leading and suction surface devices have
effective jet mixing in been discussed but trailing surfaces also experience passive
supersonic regime flow control techniques. The tabs attached to trailing edges of
Shape Triangular or semi triangular lifting surfaces provide remarkable results in the lift
Installation Airfoil generation and consequent drag reduction due to geometric
Nozzle exit changes. The constraints of passive flow include weight and
drag increases at certain instances; however, a trade-off is
The flow control at supersonic regime often uses the made between cost and efficiency to achieve desirable flow
mechanism of corrugation to enhance the effectiveness of the characteristics by using only physical devices and without
flow jet mixing. At high speeds semi corrugated triangular employing external energy.
tabs are employed for flow control processes. The
REFERENCES
effectiveness of corrugation is measured through pitot
[1] M. Asadullah, S. A. Khan, W. Asrar, and E. Sulaeman, “Passive
pressure determination where the tabs are fixed at nozzle exit control of base pressure with static cylinder at supersonic flow,”
at the inclination of 180 degrees from each other. The static IOP Conf. Ser. Mater. Sci. Eng., vol. 370, no. 1, 2018, doi:
pressure is measured at different values of back pressure and 10.1088/1757-899X/370/1/012050.
the vortices generated due to the presence of the tabs [2] B. Zhang, Q. Zhao, X. Xiang, and J. Xu, “An improved micro-
enhances flow mixing as high as Mach 2. Corrugation is used vortex generator in supersonic flows,” Aerosp. Sci. Technol., vol.
47, pp. 210–215, 2015, doi: 10.1016/j.ast.2015.09.029.
practically to reduce thrust loss and enhance aerodynamic
[3] A. R. Paul, S. Joshi, A. Jindal, S. P. Maurya, and A. Jain,
performance. The effect of corrugation will deteriorate “Experimental studies of active and passive flow control
propulsive efficiency in low Re flows, yet the advantages of techniques applied in a twin air-intake,” Sci. World J., vol. 2013,
high thrust and lift outweigh the disadvantages. 2013, doi: 10.1155/2013/523759.
[4] T. M. Reedy, G. S. Elliott, J. C. Dutton, and Y. Lee, “Passive
control of high-speed separated flows using splitter plates,” AIAA
J., vol. 50, no. 7, pp. 1586–1595, 2012, doi: 10.2514/1.J051566.
[5] I. Kim and N. Chokani, “Navier-Stokes study of supersonic cavity
flowfield with passive control,” J. Aircr., vol. 29, no. 2, pp. 217–
223, 1992, doi: 10.2514/3.46147.
[6] J. G. Dudley and L. Ukeiley, “Detached eddy simulation of a
supersonic cavity flow with and without passive flow control,”
20th AIAA Comput. Fluid Dyn. Conf. 2011, no. June, pp. 27–30,
2011, doi: 10.2514/6.2011-3844.
[7] Y. V. Pehlivanoglu and B. Yagiz, “Optimization of active/passive
Figure 7: A corrugated airfoil outlining exaggerated pattern inside [22] flow control parameters on airfoils at transonic speeds,” J. Aircr.,
vol. 48, no. 1, pp. 212–219, 2011, doi: 10.2514/1.C031066.
III. LIMITATIONS [8] M. Gad-el-Hak, “Introduction to Flow Control,” Flow Control, no.
1904, pp. 1–107, 2007, doi: 10.1007/3-540-69672-5_1.
While passive flow control techniques enable improved flight [9] Applications of Circulation Control Technology. 2006.
characteristics without using auxiliary power, instantaneous [10] N. Chokani and I. Kim, “Suppression of pressure oscillations in an
increases in drag may be observed in some cases such as open cavity by passive pneumatic control,” AIAA 22nd Fluid Dyn.
cavity flows [20]. Moreover, some devices above mentioned Plasma Dyn. Lasers Conf. 1991, 1991, doi: 10.2514/6.1991-1729.
work best in low supersonic flow, and assumptions have to be [11] K. C. Schadow and E. Gutmark, “Research For Improved Subsonic
And Supersonic Combustion AI AA / ASM E / SAE / ASEE 25th
made for high supersonic flow analysis if the same devices Joint Propulsion Conference,” 1989.
are used for passive flow control at higher Mach. [12] S. J. Lawson and G. N. Barakos, “Assessment of passive flow
control for transonic cavity flow using detached-eddy simulation,”
IV. CONCLUSION J. Aircr., vol. 46, no. 3, pp. 1009–1029, 2009, doi:
Flow control is a necessary requirement for any fluid flow. 10.2514/1.39894.
Two basic control techniques are active and passive flow [13] H. K. and T. S. Youngki Lee, Minsung Kang, “Passive Control
Techniques to Alleviate Supersonic Cavity Flow Oscillation,” J.
controls. Passive flow control is achieved by physical devices Propuls. power, vol. 24.
without an external source of energy while active flow [14] S. B. Verma and A. Hadjadj, “Supersonic flow control,” Shock
requires a continuous supply of energy. Passive flow Waves, vol. 25, no. 5, pp. 443–449, 2015, doi: 10.1007/s00193-
manipulates the flow characteristics and is involved in mixing 015-0587-y.
of flow to achieve desirable characteristics. Vortex generator [15] J. Sivasubramanian and H. F. Fasel, “Numerical investigation of
is one of the widely used device to control flow through supersonic axisymmetric wakes with active and passive flow
control,” 4th AIAA Flow Control Conf., no. June, pp. 1–18, 2008,
passive means. It creates vortices and delays production of doi: 10.2514/6.2008-3999.
trailing edge vortices eventually reducing the drag being [16] A. J. Saddington, V. Thangamani, and K. Knowles, “Comparison
produced. VGs are installed on lifting as well as control of passive flow control methods for a cavity in transonic flow,” J.
surfaces [23]. Leading edge slats are passive flow control Aircr., vol. 53, no. 5, pp. 1439–1447, 2016, doi:
device in high speed and transonic flows that delays the flow 10.2514/1.C033365.
separation and reduces pressure fluctuations. They are [17] A. P. Heffron, “Rotating Stall and Passive Flow Control on Blade
Profiles and in Centrifugal Compressors,” 2017.

Seventh International Conference on Aerospace Science & Engineering 69

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

[18] J. S. X. S. B. Richard M. Wood Floyd J. Wilcox and J. M. Allen, Fluids, vol. 24, no. 12, 2012, doi: 10.1063/1.4772063.
“Vortex Flows at Supersonic Speeds,” no. TP-211950, p. 94, 2003. [22] M. E. Abd El-Latief, K. Elsayed, and M. Madbouli Abdelrahman,
[19] S. A. Khan, A. A. Al Robaian, M. Asadullah, and A. M. Khan, “Aerodynamic study of the corrugated airfoil at ultra-low Reynolds
“Grooved cavity as a passive controller behind backward facing number,” Adv. Mech. Eng., vol. 11, no. 10, pp. 1–18, 2019, doi:
step,” J. Adv. Res. Fluid Mech. Therm. Sci., vol. 53, no. 2, pp. 185– 10.1177/1687814019884164.
193, 2019. [23] J. Sivasubramanian and H. F. Fasel, “LES and DES of high
[20] B. Yagiz, O. Kandil, and Y. V. Pehlivanoglu, “Drag minimization Reynolds number, supersonic base flows with control of the near
using active and passive flow control techniques,” Aerosp. Sci. wake,” Proc. - HPCMP Users Gr. Conf. UGC 2006, pp. 80–88,
Technol., vol. 17, no. 1, pp. 21–31, 2012, doi: 2006, doi: 10.1109/HPCMP-UGC.2006.42.
10.1016/j.ast.2011.03.003. [24] S. Aradag, K. A. Gelisli, and E. C. Yaldir, “Effects of active and
[21] H. Choi, H. Park, W. Sagong, and S. im Lee, “Biomimetic flow passive control techniques on Mach 1.5 cavity flow dynamics,” Int.
control based on morphological features of living creatures,” Phys. J. Aerosp. Eng., vol. 2017.

Seventh International Conference on Aerospace Science & Engineering 70

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Possible Seismo-Ionospheric Anomalies Associated

With the Earthquakes Occurred in 2018 Using
GNSS TEC: A Statistical Analysis
Muhammad Arqim Adil Munawar Shah Asim Khan
Institute of Space Technology Institute of Space Technology Institute of Space Technology
Islamabad, Pakistan Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected] [email protected]

Ejaz Nazir Mohamed Freeshah Hamza Bin Babar

University of Agriculture Wuhan University University of Agriculture
Faisalabad, Pakistan Wuhan, China Faisalabad, Pakistan
[email protected] [email protected] [email protected]

Abstract—The ambient detection of ionosphere momentary produced through lightning strokes during a thunderstorm
changes preceding earthquakes (EQs) is becoming an ever- activity, where they propagate in the Earth’s ionosphere and
growing debate with the excel in the satellite technology. In this further in the magnetosphere along the magnetic lines. These
paper, we study ionospheric irregularities before 11 significant studies provide some insights towards the exploration of the
EQs that occurred in 2018 by analyzing Global Navigation physical processes in the atmosphere and the possible EQ
Satellite System (GNSS) derived Total Electron Content (TEC) forecasting.
in pursuit of possible detection of the seismic precursors. We
carefully choose these EQs so that there should be at least three The ionosphere is a dispersive layer of the upper
International GNSS Service (IGS) permanent stations operating atmosphere comprising large free electrons of different gas
around the epicenter and the geomagnetic activity should quiet molecules, mainly oxygen and nitrogen, that spreads between
(Kp≤3, Dst≥ −𝟐𝟎). We perform a statistical analysis based on an altitude between ~80 km to 1000 km above the mean sea
sliding interquartile (IQR) method over TEC for 15 days before level [3]. The sun is the primary source responsible for the
and 5 days after each EQ that shows abnormal variations within changes of electron concentrations in this layer, however,
10 days before all these EQs. Our analysis shows that the some local hazardous events (volcano, typhoons, EQs) are
possible EQ induced TEC anomalies have deviated averagely ~4 also observed to be transient influencers in this part of the
TECU, with a percentage deviation of about 44%, from the atmosphere [4][5][6]. Furthermore, Freeshah et al. [7] have
corresponding upper bound (UB) during quiet geomagnetic reported the involvement of some anthropogenic processes as
activities (Kp≤3). Whereas the other TEC anomalies induced by
an unexpected outlier that may bring changes in the
moderate geomagnetic activities of ~4 Kp show relatively fewer
ionosphere by studying a recent Beirut explosion that killed
deviations of about 3 TECU (~37%) from the corresponding UB.
Moreover, we did not find any proportionality between TEC
more than 218 people and leaving more than 7000 to be
anomalies and the magnitude of the EQ. However, the sequential injured. However, EQs are known to release several trace
EQs, including fore and aftershocks, exhibit a strong gases which react with the atmosphere and the reaction
proportionality between TEC anomalies and days before EQ. consequences bring changes in the atmosphere that could help
Conversely, our results also explicate the complexness of finding in the future EQ forecasting through monitoring the electron
an ionospheric EQ precursor even in such quiet geomagnetic density variations in the ionosphere through various in-situ
activities. and satellite measurements [8][9][10][11]. These findings
urged us to conduct a statistical analysis to detect seismic
Keywords—GNSS-TEC, Ionosphere, Seismic Precursors, related ionospheric anomalies that may contribute to the
2018 Earthquakes. possible EQ prediction.
I. INTRODUCTION In this paper, 11 > 5.0 Mw EQs are studied in pursuit of
possible detection of the ionospheric anomalies as potential
Natural hazards are the most devastating geological,
seismic precursors. The purpose in not limited to the
hydrological, or meteorological events that severely affect
inspection of the seismo-ionospheric anomalies but also to
human life in terms of fatalities, infrastructure loss, and the
closely monitor the space weather conditions. In Section 2, the
environment across the globe. In recent times, scientists have
criteria of EQ selection and data analysis are briefly discussed.
become able to monitor and forecast some meteorological
Section 3 comprises results and limitations about the detection
hazards (hurricanes, tornadoes, and floods) in near real-time.
of the ionospheric anomalies as precursors. Finally, Section 4
But the most destructive geological hazards remain extremely
and 5 consist of a brief discussion and conclusion of all the
difficult or impossible to predict that includes tsunamis,
results, respectively.
earthquakes (EQs), volcanoes, and typhoons. Therefore, the
timely detection of these natural hazards is a prerequisite to II. CASE STUDIES AND DATA
avoid further human loss. Studies have suggested that the
geological natural hazards (especially EQs) induce acoustic We have studied the ionospheric effects potentially
and gravity waves into the atmosphere that propagates affected by the EQ phenomenon around the local areas of the
vertically upwards and causes variations in the electron corresponding EQs. Generally, the ionosphere is a pleasant
densities of the ionosphere [1]. Besides, some have reported residence for a number of other phenomena other like, space
the involvement of whistler waves during a seismic weather, magnetic storms, travelling ionosphere disturbances,
preparation activity [2]. The whistler waves are usually thunderstorms, and extreme weather events etc. we have

Seventh International Conference on Aerospace Science & Engineering 71

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

exampled a very few of them here. The induced electron of sight from the epicenter to ionosphere, while the IGS
variations from these phenomena are widely reported stations returns slant TEC (STEC) values, where STEC is
previously which depicts the vulnerability of the ionosphere to estimated by the total number of electrons in 1𝑚2 tube along
these extreme events [1][5]. However, some strong evidences the signal ray path. Its unit is TEC, where 1 TECU=1016
have been provided about the seismic influences in the electrons/m2 [13][14]. Now this STEC is reformulated to
ionosphere. Scientists have been continuously warning about estimate the Vertical TEC (VTEC) as:
the limitation if the using ionosphere as a reference field for
EQ forecasting, as the ultimate ionospheric dynamics are 𝑓12 𝑓22
 𝑆𝑇𝐸𝐶 = (𝐿1 − 𝐿2 + 𝜆1 (𝑁1 − 𝑁2 ) − 𝜆2 (𝑁1 −
constructed space weather events [6]. Considering this, we 40.28(𝑓12 −𝑓22 )
plan to study EQ from different regions under the solar 𝑁2 ) + 𝜀)  
minimum year of 2018 (to avoid space weather) and the
thought is: the ionospheric variations should be local. We 𝑓12 𝑓22
 𝑆𝑇𝐸𝐶 = (𝑃1 − 𝑃2 − (𝑑1 − 𝑑2 ) + 𝜀) 
formulate criteria such that there should be three IGS stations 40.28(𝑓12 −𝑓22 )
and geomagnetic conditions should be quiet. Therefore, after
𝑅.𝑠𝑖𝑛 𝑍
applying this filter we can only manage to find 11 such EQs  𝑉𝑇𝐸𝐶 = 𝑆𝑇𝐸𝐶. (𝑐𝑜𝑠 (𝑠𝑖𝑛−1 ( ))) 
(https://earthquake.usgs.gov/earthquakes/search/). It should 𝑅+𝐻

be noted that there are thousands and thousands of EQs occur

each year but the space weather phenomena don’t allow to Where λ, N, P, L, d, 𝜀, 𝑓1 , 𝑓2 , R, H, and Z are wavelength,
monitor every single one, the geomagnetic conditions are ambiguity, pseudo-range, random residual carrier phase
retrieved from http://isgi.unistra.fr/data_download.php to frequencies, radii of the Earth, height of the top ionospheric
check any disturbance during the events. The selection of the and elevation angle of the satellite [15], respectively.
IGS station to monitor ionosphere is subjected to its location B. Statistical Analysis
in EQ preparation zone which is calculated by Dobrovolsky et
The identification of VTEC ionospheric electron density
al. [12] formula:
anomalies endangered by the studied EQs is carried out by
using a total of 18 dual-frequency GNSS receiver stations,
 R  where the VTEC values are calculated through a temporal
resolution of 1-second. We constructed the confidence bounds
where 𝑀 is the magnitude of the EQ and 𝑅 denotes the through calculating the median (𝑋̃ ) and Inter Quartile Range
radius of the critical region. The detailed information about (IQR) for an interval of 10 days before each value over a
these selected EQs and their corresponding IGS stations is period of 15 days before and 5 days after each observational
provided in Fig. 1. day.
III. METHODOLOGY
A. GNSS TEC Estimation
To monitor the ionosphere variations above the epicenter
we need the electron concentrations vertically along the line

Fig. 1: The geographical locations of 11 EQs (red stars) and IGS stations (green dots) along with their respective country names in which they are operating.

The upper/lower bounds (UB/LB) can be extracted as:  𝐿𝐵 = 𝑋̃ − 1.5𝐼𝑄𝑅 

 𝑈𝐵 = 𝑋̃ + 1.5𝐼𝑄𝑅  Eqs. (5-6) demonstrate the formation of the UB and LB,
and the TEC values exceeding from either of the limit bounds

Seventh International Conference on Aerospace Science & Engineering 72

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

will be declared as anomalies with a confidence level of that occurred at 17:29:29 UT on November 30, 2018, while
99.99% [11]. Such a precision in the confidence level makes the other EQs will be briefed through statistical results. The
this method significant one. geographical coordinates of Alaska EQ are 61.346oN,
149.955oW in latitude and longitude, respectively, while
On the other hand, through subtraction method, where the according to the USGS the depth of this EQ was recorded as
actual VTEC values are subtracted from the UB values, 47 km (Fig. 1a). For monitoring the ionospheric VTEC
differential TEC (ΔTEC) is utilized to quantify potential variations before this event, we selected 3 IGS stations
seismic induced anomalies. The observed TEC (𝑇𝐸𝐶𝑜𝑏𝑠 ) is operating closest and falling under the umbra of the EQ
subtracted from UB to get the deviation in TEC from the preparation area whose geographical locations are depicted in
normal distribution as: (Fig. 1a), under observably quiet geomagnetic and solar
conditions. One can see in Fig. 2c-e and Table 1 a deviation in
𝑇𝐸𝐶𝑜𝑏𝑠 − 𝑇𝐸𝐶𝑈𝐵
 𝛥𝑇𝐸𝐶 = {   maximum 𝛥𝑇𝐸𝐶 values is of about 1.6 TECU at FAIR, 1.2
0 𝑖𝑓 𝛥𝑇𝐸𝐶 < 0 TECU at WHIT, and 1.3 TECU at GCGO stations, 3 days
before the EQ, respectively. Furthermore, the percentage
The purpose of using this methodology is separately deviations are also recorded as pretty significant where the
present the geomagnetic and seismic anomalies because the values exceeded about 56.71% at FAIR, 26.7% at WHIT, and
majority of the trend been dismissed. 73.2% GCGO from the corresponding UB. It should be noted
that these variations are under silent geomagnetic conditions.
IV. RESULTS
The geomagnetic indices timeseries during the observational
We present the ionospheric states in terms of GNSS-TEC period is shown in Fig. 2a-b, where a threshold for
before Mw > 5.0 EQs occurred during 2018. The EQs, geomagnetic Kp is represented by a red dashed over the value
belonging to different regions like Alaska, Argentina, Chile, 3. In any case a value touching or exceeding the threshold line
Greece, Honduras, Russia, and Taiwan are explored. will declare an active geomagnetic activity. Such a scenario
Additionally, the diurnal TEC variations along with can be observed in about 3 days after the Mw 7.1 Alaskan EQ
differential TEC values are analyzed. The differential TEC in Fig. 2b on December 03, 2018, where the geomagnetic Kp
values helped a lot to distinguish whether these anomalies are values overlaps the declared threshold triggering a
due to seismic interventions or due to geomagnetic geomagnetic activity. Some positive deviations can also be
disturbances. The ionospheric states before these 11 EQ events observed on 19th and 20th of November which are pretty
that occurred in 6 different regions are discussed below. consistent with the Mw 7.1 EQ variations. Interestingly, these
The word count limitations of the conference don’t allow variations are the result of an EQ as well, but with a relatively
us to give a comprehensive view of the each event, however, smaller magnitude of Mw 5.6, which was a foreshock to the
we present the full analysis of a significant Mw 7.1 Alaska EQ gigantic Mw 7.1 EQ, its anomaly details are given in Table 1.

Fig. 2: For Alaska EQ (a.) Geomagnetic Dst (b.) Geomagnetic Kp Index, while GNSS-TEC from (c) FAIR, (d) WHIT, and (e) GCGO.

Seventh International Conference on Aerospace Science & Engineering 73

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Table 1: Statistical results of the 𝛥TEC deviations for SIAs and GIAs

Ave. Max. Ave. Max. Ave. Max. Max. Ave.

Geomagnetically Value of
Sr. No Event Date 𝜟TEC of %TEC for 𝜟TEC for %TEC for
Disturbed Day Kp Index
SIA SIA GIA* GIA
1 5.6 2018-11-21 1.7193 59.62 % 2018-11-12 2.7431 88.35 % 4-
2 7.1 2018-11-30 1.3756 52.20 % 2018-12-03 2.0459 69.35 % 4-
3 5.4 2018-01-28 6.5606 33.54 % 2018-01-24 5.6546 26.68 % 4o
4 5.2 2018 -02-01 5.4180 36.41 % 2018-01-24 5.6546 26.68 % 4o
5 5.6 2018-02-08 6.1690 33.13 % 2018-01-24 5.6546 26.68 % 4o
6 6.8 2018-10-25 3.3180 26.58 % 2018-10-13 N/A N/A 4+
7 7.5 2018-01-10 7.5472 54.41 % 2018-01-08 4.3924 38.96 % 4-
8 7.3 2018-12-20 1.9925 39.05 % 2018-12-09 1.1133 29.18 % 4-
9 6.2 2018-12-25 1.7730 41.45 % 2018-12-09 1.1133 29.18 % 4-
10 6.1 2018-02-04 6.0543 42.75 % 2018-01-24 2.7532 19.82 % 4o
11 6.4 2018-02-06 9.0018 45.69 % 2018-01-24 2.7532 19.82 % 4o
*Geomagnetically Induced Anomalies

According to Pulinets et al. [17], the possible cause of remarkably less than the percentage deviation of SIA
production of these SIAs is the air ionization process that takes anomalies as shown in Fig. 4. Besides, one should keep in
place around the vicinity of a seismic epicenter some days to mind that in this analysis the geomagnetic conditions are not
few months before the seismic jolt. The emission of radon gas that much active and only moderate to low (Kp 3 – 4), so the
due to the stress effect of rocks along the active faults is largely SIA anomalies are only superior to the moderate geomagnetic
responsible for this air ionization process. This air ionization conditions. However, severe geomagnetic storms (Kp > 5) will
further causes decrease in the relative humidity and increases ultimately have higher influences in to the ionosphere where
the thermal budget of the atmosphere in the form of OLR. they can even mask the seismic anomalies. Furthermore, some
Therefore, the anomaly maps of the relative humidity and reports favors the SIA up to only 20% [18]. In contrast, our
OLR from NOAA are investigated to confirm the air results show deviations of more than 20 % (Fig. 4).
ionization production that caused the seismo-ionosphere Furthermore, the sequential EQs, including fore and
variations on November 19,20, and 28, 2018. Fig. 3 depicts aftershocks, exhibit ionospheric anomalies at the almost same
the relative humidity and OLR variations where the humidity number of days as its predecessor (Fig. 5).
decreased to 20% and consequent OLR anomalies increased
up to 30 W/m2 around the epicenters of corresponding EQs
that fairly demonstrates the production of air ionization over
the epicenters. The Fig. 2 demonstrates an extremely
interesting fact about one after another synchronous
anomalous pattern where both back to back EQs show a strong
proportion in anomalies exhibition and number of days before
their corresponding shocks as depicted by Fig. 2c-e.
We adopted the same methodology for other selected EQs
(Fig. 1) and the final results are summarized in Table 1. These
findings suggested that the SIAs can be useful in the future EQ
prediction, but it will be difficult to find a precursor in high
geomagnetic conditions.
V. DISCUSSION
The purpose of this paper was not limited to describe the
seismic effect on the ionosphere but it was equally devoted to
discuss the geomagnetic effects, and how we can distinguish
both effects. A brief overview of this discrimination is
presented in Table 1 and Fig. 4. This comparison shows that
the TEC deviations may deviate averagely about 4.63 TECU
from UB for possible seismic anomalies as compared to the Fig. 3: The anomaly maps of the relative humidity and OLR depicting air
geomagnetic anomalies (Kp = 3) whose values are averagely ionization over and around the epicenters of both sequential Alaskan EQs
recorded up to 3.39. The production of the air ionization on December 19, 20, and 28, 2018. Top two rows describe the variations
for M 5.6 while the third row represents the anomalous variations on
around the corresponding epicenters is main cause behind the December 28, 2018. The white stars depict the location of the
generation of these ionospheric anomalies [17]. In a similar corresponding epicenters.
analysis, Adil et al. [3] find decreasing humidity with
increasing OLR and TEC anomalies before three New Zealand
EQs. Whereas the geomagnetically induced anomalies have
shown a percentage deviation of 37.47% averagely, which is

Seventh International Conference on Aerospace Science & Engineering 74

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

ACKNOWLEDGMENT
The authors are thankful to the IGS, USGS, OMNIWEB
NASA and NOAA for GNSS-TEC, EQs, geomagnetic and
atmosphere data.
REFERENCES
[1] E. Astafyeva, “Ionospheric Detection of Natural Hazards,” Reviews of
Geophysics, vol. 57, no. 4, pp. 1265–1288, 2019.
[2] M. Parrot and J.-L. Pinçon, “Is There an Earthquake Weather?,” Open
Journal of Earthquake Research, vol. 09, no. 02, pp. 69–82, 2020.
[3] M. A. Adil, A. Abbas, M. Ehsan, M. Shah, N. A. Naqvi, A. Alie,“
Investigation of ionospheric and atmospheric anomalies associated
with three Mw >6.5 EQs in New Zealand,” Journal of Geodynamics,
vol. 145, pp. 101841, 2021.
[4] W. Li, J. Guo, J. Yue, Y. Shen, and Y. Yang, “Total electron content
anomalies associated with global VEI4 + volcanic eruptions during
2002–2015,” Journal of Volcanology and Geothermal Research, vol.
325, pp. 98–109, 2016.
[5] M. Freeshah et al., “Analysis of Atmospheric and Ionospheric
Variations Due to Impacts of Super Typhoon Mangkhut (1822) in the
Northwest Pacific Ocean,” Remote Sensing, vol. 13, no. 4, p. 661, 2021.
Fig. 4: Max. ΔTEC (upper panel) and % deviations (lower panel) for all [6] M. A. Adil, E. Şentürk, S. A. Pulinets, and C. Amory-Mazaudier, “A
EQs. Lithosphere–Atmosphere–Ionosphere Coupling Phenomenon
Observed Before M 7.7 Jamaica Earthquake,” Pure and Applied
Geophysics, vol. 178, p. 3869–3886 2021.
VI. CONCLUSION.
[7] Freeshah, M., Zhang, X., Şentürk, E., Ren, X., Arqim Adil, M., and Xu,
We presented EQ related ionospheric anomalies from G., “Could the Beirut Explosion perturb the Ionosphere? Pre-results
GNSS TEC of 11 global EQs in 2018 during quiet Using TEC-GNSS observations.” EGU General Assembly Conference
geomagnetic conditions. Results suggested the TEC Abstracts, pp. EGU21-8990, 2021.
anomalies during Kp ≤ 3 geomagnetic activities are the [8] M. A. Adil, E. Şentürk, M. Shah, N. A. Naqvi, M. Saqib, and A. R.
Abbasi, “Atmospheric and ionospheric disturbances associated with the
possible seismic anomalies appeared because of the air M > 6 earthquakes in the East Asian sector: A case study of two
ionization produced from the increased radon emanations consecutive earthquakes in Taiwan,” Journal of Asian Earth Sciences,
around the epicenters of the 11 corresponding EQ prone vol. 220, p. 104918, 2021.
regions that induced ionospheric changes through decreasing [9] A. Kiyani et al., “Seismo ionospheric anomalies possibly associated
humidity and increasing the thermal budget of the atmosphere. with the 2018 M 8.2 Fiji earthquake detected with GNSS
For seismo ionospheric anomalies, 4 TECU deviation of daily TEC,” Journal of Geodynamics, vol. 140, p. 101782, 2020.
TEC beyond UB is recorded, while moderate geomagnetic [10] M. Saqib, E. Şentürk, S. A. Sahu, and M. A. Adil, “Ionospheric
anomalies detection using autoregressive integrated moving average
activities (Kp=4) recorded about 3.7878. The percentage (ARIMA) model as an earthquake precursor,” Acta Geophysica, vol.
corresponds to 37.47% for the GIAs, while 42.28% for SIAs. 69, no. 4, pp. 1493–1507, 2021.
Furthermore, during the observed periods before and after the [11] M. Shah et al., “Possible seismo-ionospheric anomalies associated with
respective shocks, the EQ induced ionosphere anomalies have M > 5.0 earthquakes during 2000-2020 from GNSS TEC,” Advances in
shown a better tendency than moderate geomagnetic Space Research, Apr. 2022.
variations. These results indicate the existence of seismo [12] I. P. Dobrovolsky, S. I. Zubkov, and V. I. Miachkin, “Estimation of the
ionosphere anomalies, however, to further clarify this size of earthquake preparation zones,” Pure and Applied Geophysics
PAGEOPH, vol. 117, no. 5, pp. 1025–1044, 1979.
marvelous phenomenon more dedicated ground and space
observations are needed. [13] H. Li, Y. Yuan, Z. Li, X. Huo, and W. Yan, “Ionospheric Electron
Concentration Imaging Using Combination of LEO Satellite Data With
Ground-Based GPS Observations Over China,” IEEE Transactions on
Geoscience and Remote Sensing, vol. 50, no. 5, pp. 1728–1735, 2012.
[14] D. Roma-Dollase et al., “Consistency of seven different GNSS global
ionospheric mapping techniques during one solar cycle,” Journal of
Geodesy, vol. 92, no. 6, pp. 691–706, 2017.
[15] J. Klobuchar, “Ionospheric Time-Delay Algorithm for Single-
Frequency GPS Users,” IEEE Transactions on Aerospace and
Electronic Systems, vol. AES-23, no. 3, pp. 325–331, 1987.
[16] M. Saqib, E. Şentürk, S. A. Sahu, and M. A. Adil, “Comparisons of
autoregressive integrated moving average (ARIMA) and long short
term memory (LSTM) network models for ionospheric anomalies
detection: a study on Haiti (Mw = 7.0) earthquake,” Acta Geodaetica et
Geophysica, vol. 57, no. 1, pp. 195–213, 2022.
[17] S. A. Pulinets, D. P. Ouzounov, A. V. Karelin, and D. V. Davidenko,
“Physical bases of the generation of short-term earthquake precursors:
A complex model of ionization-induced geophysical processes in the
lithosphere-atmosphere-ionosphere-magnetosphere
Fig. 5: The anomaly pattern followed by sequential EQs sharing same system,” Geomagnetism and Aeronomy, vol. 55, no. 4, pp. 521–538,
regions. 2015.
[18] J. M. Forbes, S. E. Palo, and X. Zhang, “Variability of the
ionosphere,” Journal of Atmospheric and Solar-Terrestrial Physics,
vol. 62, no. 8, pp. 685–693, 2000.

Seventh International Conference on Aerospace Science & Engineering 75

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Analysis of The Atmospheric Anomalies Associated

with Mw>6.0 Continental Earthquake in The
Afghanistan and Tajikistan
Daboor Sultana Dr. Munawar Shah
Department of Space Science Department of Space Science
Institute of Space Technology Institute of Space Technology
Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected]

Abstract— The recent advances in space-based atmospheric enhancement in humidity, AOD, and soil moisture exists
precursors of earthquakes (EQ) aid in the development of the before the coastal EQs [7-10]. Nevertheless, the high
lithospheric atmospheric coupling system. In this paper, elevation and heterogeneous geomorphology of the
anomalies associated with the atmosphere are investigated in epicentral region could not influence the abnormality in
the Central Asian region of Mw>6.0 EQs (Afghanistan and these parameters.
Tajikistan) are examined. The atmospheric precursors are
There are different hypotheses about the evaluation
searched in humidity, Aerosol Optical Depth (AOD), and soil
moisture from Giovanni (A Web interface to analyze NASA's mechanism of pre-seismic atmosphere anomalies from
gridded data). Pre atmospheric anomalies are observed remote sensing procedures. Among them, the most
before all three EQs during a 5- 50-day window beyond the compromised theory is the activation of positive holes (p
upper bounds. On the other hand, post atmospheric holes) in the stressed rocks in the seismic preparation zone
anomalies are only observed after the Mw 7.0 Afghanistan EQ [11]. According to the p-hole hypothesis [11], the positive
within 5 days. The overall analysis shows that pre atmospheric charge carriers triggered by tectonic forces in the crustal
anomalies can be monitored more frequently than post rocks propagate out of the compressed rocks to the nearby
atmospheric anomalies; however, more analyses need to clear atmosphere. Furthermore, it causes the air ionization at the
the lithosphere atmosphere coupling. There is a lack of
Earth-atmosphere interface in a thin layer and induced
possible precursors of the EQs, but the need for more analysis
can certainly validate the phenomenon. significant atmospheric anomalies due to the
recombination of p holes. Adil et al. [12] showed from a
ray-trace model that a millimeter disturbance at the Earth’s
Keywords— Earthquake; Atmospheric anomalies; Mw> surface can be amplified to wave amplitude of 100 m at 100
6.0; Lithospheric-Atmospheric Coupling; AOD; km altitude up in the atmosphere.
Humidity; Soil Moisture II. DATA AND METHOD
I. INTRODUCTION The temporal and spatial atmospheric anomalies before the
Satellite remote sensing procedures detected the EQ- large magnitude continental EQs are reported based on
induced atmospheric precursors at different altitudes before remote sensing procedures. United States Geological
the megathrust over the seismogenic zone. These Survey (USGS) provides the EQs information via link
precursors can be of three types depending on the altitude: (https://earthquake.usgs.gov/earthquakes/) [13-14]. The
(1) Lithospheric precursors occurred around the epicenter different parameters of the EQs are listed in Table. 1. We
on the ground, such as land surface temperature and soil studied three different hypocentral depth EQs (e.g. shallow,
moisture [1], [2] air ionization and humidity variations in intermediate, and deep) on the continental part of the
the atmosphere near the Earth surface [2], and [3] the lithosphere in Afghanistan and Tajikistan (Fig. 1). The
ionospheric variations over the seismic prone region. [3-4] terrestrial DEM (Digital Elevation Model) data is used for
According to previous reports, the atmospheric anomalies determining the crustal thickness in the study area. The
from different sources over the epicenter of the mainshock temporal data for DEM is retrieved from NASA via earth
event occurred within one to five months before the explorer (https://earthexplorer.usgs.gov/). Similarly, the
mainshock [5]. Nevertheless, it depends on the magnitude Geoid height and Gravity anomaly are obtained from the
and local terrain of the event. They further explained that ICE
different atmospheric parameters behaved abnormally over (https://www.aviso.altimetry.fr/es/applications/geodesy-
the seismic zone within several days to several months and-geophysics/geodesy.html). [15-18] these datasets
before the impending mainshock. On the other hand, the confirm the existence of high topography of the studied
lack of detection of the right anomaly is the selection of EQs epicenters. In this study, we aim to show the
appropriate statistical parameters [6]. Therefore, the consistency in atmospheric anomalies before the
existence time of the EQ-induced atmospheric anomalies continental EQs, which occurred on high terrain in
became very essential for the prediction of the megathrust. Afghanistan and Tajikistan.
The multi-parameter-based study of the atmosphere is
important for the reliability of the EQ precursor detection
and understanding of the complex mechanism of seismic
related anomalies. According to previous reports, the

Seventh International Conference on Aerospace Science & Engineering 76

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

To identify any significant deviation triggered by the EQ,

all the variables were studied for 3-4 months before and one
month after the mainshock. Furthermore, the anomalies
were detected using µ±σ (median ± sigma), which is
obtained from the data of (2002-2015). We declared the
anomaly as the deviation from normal distribution to a
significant level beyond the confidence levels, as the
method implemented in [9,10].

III. RESULTS AND DISCUSSIONS

In this study, atmospheric anomalies associated with EQs
Fig. 1. Epicenter of the three EQs in Asian regions. before three large magnitudes EQs from remote sensing
The red-filled star is for the EQs epicenter. data in the continental portion in Afghanistan and
Tajikistan are analyzed. The occurrence anomaly over the
epicenter during the EQ preparation period in humidity,
AOD, and Soil moisture before the main shock confirmed
the synoptic blocking of huge energy appeared in remote
sensing procedures. There exist different conclusions in
past about the formation of such anomalies at the
lithospheric-atmospheric interface over the EQ preparation
zone (EPA).

Fig. 2. (a) DEM, (b) Geoid, and (c) Gravity anomaly of Fig. 2. The analysis of different atmospheric indices with
the study area. their confidence bounds before and after the 5th October
The atmospheric data (e.g. humidity, AOD, and soil 2008 Afghanistan EQ.
moisture) were retrieved from NASA through GIOVANNI The atmospheric anomalies are clear on the day (-35)
via (http://www.giovanni.sci.gsfc.nasa.gov/giovanni). before the 2008 Afghanistan EQ (Fig. 2). On the other
GIOVANNI is an online data source of remote sensing to hand, there are no post-seismic atmospheric anomalies after
assist in atmospheric and meteorology research and the mainshock day within the seismogenic zone.
maintained by the National Aeronautics and Space
Administration (NASA).
Table 1. Information for the three EQs of Mw> 6.0

Earthquake Occurrenc Pre- Post-

e Date atmospheric atmosphe
anomalies ric
days anomalies
days

M6.0 2008-10-05 -35 days Nil

Afghanistan before
M6.0 2011-01-24 -50 days Nil
Tajikistan before
M7.0 2005-10-25 -5 days 5 days Fig. 3. (a) Humidity, (b) AOD, and (c) Soil Moisture
Afghanistan before after analyses with confidence bounds associated with the 24th
January 2011 Tajikistan EQ.
The 2011 Tajikistan EQ shows clear anomalies in AOD on
-50 days as a pre-seismic atmospheric perturbation (Fig. 3).

Seventh International Conference on Aerospace Science & Engineering 77

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Which is duly verified by the analyses of humidity and soil [3] M. Shah, A. Ahmed, M. Ehsan, M. Khan, M. A. Tariq, A.
Calabia, and Z. ur Rahman, “Total electron content anomalies
moisture from different remote sensing satellites data (Fig.
associated with EQs occurred during 1998–2019,” Acta
3). As in Figure 4, the -5 day anomalies in the atmosphere Astronautica, vol. 175, pp. 268–276, 2020.
are clear within 5 days before the 2005 M7.0 EQ in [4] M. Ulukavak, “Deep learning for ionospheric tec forecasting
Afghanistan (Fig. 4). The AOD shows more than 10% at mid-latitude stations in turkey,” Acta Geophysica, vol. 69,
variation as an anomaly from the normal distribution of the no. 2, pp. 589–606, 2021.
data (Fig. 4). Similar variations are recorded in humidity [5] E. Astafyeva, “Ionospheric detection of natural hazards,”
Reviews of Geophysics, vol. 57, no. 4, pp. 1265–1288, 2019.
and soil moisture in Figures 4a and 4c, respectively.
[6] A. Kiyani, M. Shah, A. Ahmed, H. H. Shah, S. Hameed, M. A.
Adil, and N. A. Naqvi, “Seismo ionospheric anomalies
possibly associated with the 2018 MW 8.2 Fiji EQ detected
with GNSS TEC,” Journal of Geodynamics, vol. 140, p.
101782, 2020.
[7] K. Davies and D. M. Baker, “Ionospheric effects observed
around the time of the Alaskan EQ of march 28, 1964,” Journal
of Geophysical Research, vol. 70, no. 9, pp. 2251–2253, 1965.
[8] S. Pulinets, “Ionospheric precursors of EQs; recent advances
in theory and practical applications,” Terrestrial Atmospheric
and Oceanic Sciences, vol. 15, no. 3, pp. 413–436, 2004.
[9] S. Jin, G. Occhipinti, and R. Jin, “GNSS ionospheric
seismology: Recent observation evidence and characteristics,”
Earth-Science Reviews, vol. 147, pp. 54–64, 2015.
[10] M. Free shah, X. Zhang, E. S¸ entu¨rk, M. A. Adil, B.
Mousa, A. Tariq, X. Ren, and M. Refaat, “Analysis of
atmospheric and ionospheric variations due to impacts of super
typhoon mangkhut (1822) in the northwest Pacific Ocean,”
Fig. 4. The analysis of different atmospheric indices with Remote Sensing, vol. 13, no. 4, p. 661, 2021.
its [11] J. Ahmed, M. Shah, M. Awais, S. Jin, W. A. Zafar, N. Ahmad,
confidence bounds before and after the 25th October 2005 A. Amin,M. A. Shah, and I. Ali, “Seismo-ionospheric
anomalies before the 2019 Mirpur EQ from ionosonde
Afghanistan EQ. measurements,” Advances in Space Research, 2021.
IV. CONCLUSION [12] M. A. Adil, A. Abbas, M. Ehsan, M. Shah, N. A. Naqvi, and
A. Alie, “Investigation of ionospheric and atmospheric
In this paper, atmospheric anomalies associated with 3 anomalies associated with three mw¿ 6.5 Eqs in New
large magnitude EQs in the continental part of Asia; i.e., Zealand,” Journal of Geodynamics, vol. 145, p. 101841, 2021.
Afghanistan and Tajikistan are examined. All parameters [13] I. Dobrovolsky, S. Zubkov, and V. Miachkin, “Estimation of
the size of EQ preparation zones,” Pure and applied
were investigated for 3-4 months before and one month geophysics, vol. 117, no. 5, pp. 1025–1044, 1979.
after the main EQ event. Furthermore, the anomalies were [14] M. Shah and S. Jin, “Pre-seismic ionospheric anomalies of the
detected using µ±σ (median ± sigma), which is obtained 2013 mw= 7.7 Pakistan EQ from GPS and cosmic
from the data of (2002-2015). We found different observations,” Geodesy and Geodynamics, vol. 9, no. 5, pp.
atmospheric parameters behaved abnormally over the 378–387, 2018.
seismic zone within several days to several months before [15] A. R. Abbass, M. Shah, A. Ahmed, and N. A. Naqvi, “Possible
ionospheric anomalies associated with the 2009 m w 6.4
the impending EQ. Results show the occurrence of Taiwan EQ from Demeter and GNSS tec,” Acta Geodaetica et
atmospheric anomalies within the EPA in humidity, AOD, Geophysica, vol. 56, no. 1, pp. 77–91, 2021.
and Soil moisture before the main shock confirmed the [16] L. He, L. Wu, A. De Santis, S. Liu, and Y. Yang, “Is there a
synoptic blocking of huge energy appeared in remote one-to-one correspondence between ionospheric anomalies
sensing procedures. We declared the anomaly as the and large EQs along longmenshan faults?” in Annales
Geophysical, vol. 32, no. 2. Copernicus GmbH, 2014, pp. 187–
deviation from normal distribution to a significant level 196.
beyond the confidence levels of high terrain at the [17] A. Ali, C. Sheng-Chang, and M. Shah, “Continuous wavelet
lithosphere-atmosphere interface over the EPA. transformation of seismic data for feature extraction,” SN
Applied Sciences, vol. 2, no. 11, pp. 1–12, 2020.
ACKNOWLEDGEMENT [18] H. Li, Y. Yuan, Z. Li, X. Huo, and W. Yan, “Ionospheric
electron concentration imaging using a combination of LEO
The authors are grateful to the Earth Explorer community
satellite data with ground-based GPS observations over
for the availability of atmospheric indices data. We are China,” IEEE Trans. Geosci. Remote Sens vol. 50, no. 5 PART
obliged by the United States Geological Survey (USGS), 2, pp. 1728–1735, 2012, doi: 10.1109/TGRS.2011.21 964.
I CE for providing information about the EQs, Geoid
height, and atmospheric parameters, respectively.

REFERENCES
[1] K. Davies and G. Hartmann, “Studying the ionosphere with the
global positioning system,” Radio Science, vol. 32, no. 4, pp.
1695–1703, 1997.
[2] J. Shuang-gen, J. Wang, Z. Hong-ping, and Z. Wen-yao,
“Real-time monitoring and prediction of ionospheric electron
content by means of GPS,” Chinese Astronomy and
Astrophysics, vol. 28, no. 3, pp. 331–337, 2004.

Seventh International Conference on Aerospace Science & Engineering 78

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Analysis and Comparison of Maximum Power Point

Tracking Algorithms for Electrical Power System of
the CubeSats
Usman Azeem Dr.Rehan Mahmood
Small Satellite Technology and Research Lab, NCGSA Small Satellite Technology and Research Lab, NCGSA
Institute of Space Technology Institute of Space Technology
Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected]

Abstract— The photovoltaic (PV) system has been used for The optimal PV output current to voltage (I-V) characteristic
years as one of the leading renewable energy sources. We are depends nonlinearly on temperature and irradiance.
concerned with reducing cost and improving efficiencies today Photovoltaic arrays are typically designed so that their output
so these natural resources can be tapped and utilized effectively. power rests at a single point on the graphs of its I-V and P-
Using a maximum point tracker (MPT), it is possible to extract
V. If arrays are continually operated at their maximum power
maximum power from a PV panel. PV systems have varying
output powers based on changes in temperature and irradiance. point (MPP), they can generate maximum power. To make
Charger efficiency must be enhanced. Several maximum power cubesats as efficient as possible, it is critical to select a proper
point tracking (MPPT) methods are available to achieve the MPPT when designing EPS [3]. In this study, two MPPT
maximum point, for example, Incremental Conductance techniques are compared: perturbation and observation
algorithm (INC), Perturbation and Observation algorithm technique and incremental conductance technique.
(P&O), and a Constant Voltage technique. To measure and
evaluate the performance parameters and to test the efficiency II. How a Solar cell works
and validity of the proposed controller, an entire PV system will
be modeled and simulated with Simulink. On the design level,
we used Incremental Conductance Method (INC) and Perturb
Light energy is converted into electrical energy by solar cells
and Observe (P&O) MPPT algorithm. For comparison or photovoltaic cells through the photovoltaic effect. PN
purposes, the output data of the MPPT is compared to find a junctions are the basic building blocks of solar cells. The
significantly better MPPT system. photovoltaic effect is responsible for the operation of solar
cells. Through semi conductive silicon materials, light causes,
Keywords— MPPT (Maximum Power Point Tracking), electrons to move from one layer to another. Electrons are
Photovoltic Panel (PV), INC (Incremental Conductance), P&O ultimately responsible for producing an electric current. Cells
(Perturb and Observation), Boost converter, Buck Converter, that convert solar radiation directly into electric current are
Buck-Boost Converter.
called photovoltaic, which are made from semiconductor
I. INTRODUCTION materials.
EPS electronics, or Electrical Power System (EPS), provide
considerable power to nanosats and cubesats.As a long-term
cubesat bus, the EPS needs to be flexible as well as efficient.
As a result, EPS should be designed to meet specific power
requirements, and capable of supporting multiple missions
without having to be completely redesigned [1]. In a
photovoltaic (PV) a system, solar cells are employed. Various
cells are linked in the series to create specific voltage (V) and
are linked in parallel to produce specific current (A). Modules
are also connected parallelly and in the series to create an
array of modules [2].

Figure. 2. Photovoltaic cell structure

PN junctions produce electric current in their zones of
depletion. A p-n junction has depletion zones caused by
electrons diffusing from the n-type to holes of the p-type [3].
By absorption of light by one of these silicon atoms, light will
displace electrons, creating a hole and a free electron. The
result is the formation of free electrons and holes with
sufficient energy to leave the depletion zone. The current will
Figure. 1. PV Cell, Module, Array
flow (electrons) through the wire when connected between
the cathode (n-type silicon) and the anode (p-type silicon). An

Seventh International Conference on Aerospace Science & Engineering 79

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

electric current is created when the electron attracts the

positive charge of the "p-type" material and travels through
the load. If an electron is freed from its charge, the negative
charge of "n-type" material will attract the electron and cause
it to fall towards the back electrical contact. As a result of the
electrons entering p-type silicon from the back electrical
contact, the electrical neutrality is restored.
We can further describe the PV cell characteristics by using
an equivalent circuit as shown in the fig 3. PV models consist
of three components: current sources, diodes, and series
resistances. The parallel resistance results in a very small
leakage resistance in a single module. Under constant
temperatures and a constant incidence of light, the current
source's output is constant.

Figure. 4. Schematic of a solar array

Figure. 3. Equivalent circuit of an ideal PV cell

According to the PV cell model, the equations are as follows:

[1]

𝐼 = 𝐼𝑝𝑣 − 𝐼𝑑 − 𝐼𝑟𝑝 (1)

𝐼 ∗ 𝑅𝑠 𝑅𝑠
𝐼 = 𝐼𝑝𝑣 − 𝐼𝑜 (exp (𝑉 + ) − 1) − 𝑉 + 𝐼 ∗ (2)
𝑎 ∗ 𝑉𝑡 𝑅𝑝
𝑇
𝑉𝑡 = 𝑁𝑠 ∗ 𝐾 ∗ (3)
𝑞
𝐺
𝐼𝑝ℎ = (𝐼𝑝𝑣𝑛 + 𝑘𝑖(𝑇 − 𝑇𝑛) (4)
𝐺𝑛
𝑇 3 𝐸𝑔 1 1 Figure. 5. V-I and P-V curve
𝐼𝑜 = 𝐼𝑟𝑠 ( ) ∗ exp(𝑞 ∗ ( − ) (5) Utilizing the IV curve, we can determine the maximum power
𝑇𝑛 𝑎 ∗ 𝑘 𝑇𝑛 𝑇
for a solar panel in an ideal situation. Below is a graph of
𝐼𝑠𝑐𝑛 current versus voltage. The peak represents the maximum
𝐼𝑟𝑠 = (6) power available [7].
𝑉𝑜𝑐𝑛
exp ( )−1
𝑎 ∗ 𝑉𝑡𝑛

Where:

I = solar cell current (A)

V= solar cell voltage (V)
Ipv = photo current (A)
Id=diode current (A)
Io=reverse saturation current (A)
K= boltzman constant
T=solar cell temperature
q= charge of electron
G= irradiance intensity Figure. 6. Basic V-I curve

Solar cells typically produce the following outputs for a given

A solar panel's maximum power can be determined using the irradiance and operating temperature. According to figure 6.
IV curve in an ideal position. A graph showing how current 1. Short circuit current (Isc) is equal to maximum current at
varies with voltage is shown below. The peak will provide the zero potential. In an ideal world, if V = 0, Isc = IL [4]. There
best power output for the panel. is a direct relationship between Isc and the amount of sunlight
available

Seventh International Conference on Aerospace Science & Engineering 80

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

2. Voltage of an open circuit (VOC) at zero current is defined MPP, therefore the module voltage must be lowered to put the
as the maximum voltage. An increase in sunlight operating point on the left [6]. According to figure8 INC
exponentially increase Voc. In this way, solar cells are method, at the MPP, the PV array slope is zero, it slopes
excellent for charging batteries. positively to the left of the MPP, and it slopes negatively to
the right.
III. MPPT ALGORITHMS 𝑑𝑃
MPP is reached when =0
𝑑𝑉
IF
As the name implies, MPPT is intended to find the maximum 𝑑𝑃
voltage and current when using a PV module, which is > 0 𝑡ℎ𝑒𝑛 𝑉𝑝 < 𝑉𝑚𝑝𝑝 (1)
referred to as the maximum power point. The main MPPT 𝑑𝑉
𝑑𝑃
algorithms are. = 0 𝑡ℎ𝑒𝑛 𝑉𝑝 = 𝑉𝑚𝑝𝑝 (2)
A. Perturb and Observe technique 𝑑𝑉
𝑑𝑃
< 0 𝑡ℎ𝑒𝑛 𝑉𝑝 > 𝑉𝑚𝑝𝑝 (3)
The MPP is commonly determined with the help of this 𝑑𝑉
method. As a result of this method, the controller adjusts
voltage slightly from the array, measures power, and makes
adjustments until there is no longer increase in power. We call
this method "Perturb and Observe”. An important feature of
P&O methods is the comparison of (Pn) recent PV
power with (Pn-1) previous PV power. Voltage and current
are measured to determine a photovoltaic power. This
algorithm tries to determine the optimal point in the left or
right side of recent position when the difference between
previous power and recent power is not equal to zero. When
P (n)-P (n-1) is equal to 0, a maximum power is achieved. A
voltage will be increased or decreased until the MPP is found
[3].

Figure. 8. Incremental Conductance Algorithm

Figure. 7. Perturb and Observe Algorithm

B. Incremental Conductance technique

The incremental conductance method was used to observe a Figure. 9. Incremental Conductance Graph
P-V characteristic curve. Power loss is reduced, PV efficiency
is improved, and system costs are reduced using this method.
The PV module's voltage and current are measured first. IV. POWER CONVERTERS
Next, it calculates incremental changes, dI (change in current)
and dV (change in voltage), by using the current value and In electronics, DC-DC converters are circuits that change
previous voltage value. Based on the assumption that the direct current (DC) from one level to another. Typical
operating point lies on the MPP's left when the inequality applications for these DC-DC converters include dc motor
equation below is satisfied, so the module voltage must be drives and switch-mode power supplies [2]. Their input
increased to move it to the right. In the same manner, voltage is often unregulated. Due to the magnitude changes
whenever the condition is satisfied by the inequality equation, in the line voltage, this is obtained by rectifying the line
we assume that the operating point lies on the right of the voltage. A switch mode DC-DC converter is employed in

Seventh International Conference on Aerospace Science & Engineering 81

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

MPPT to regulate the input voltage at the PV MPP and match

the load to ensure maximum energy transfer. The basic DC to
DC converters considered for analysis are

A. BUCK Converter

Step-down converters, or buck converters reduce voltage

while maintaining a constant level of power. Buck converters
were designed to improve efficiency and extend battery life, Figure. 11. Boost Converter
reduce heat, and reduce the size of the device [6]. As DC-DC
converters, Buck converters are much more energy efficient C. BUCK BOOST Converter
than linear regulators because they act as switching
converters. The diode is in an open circuit condition when the In order to generate stable current (PV cells) from a high or
switch is on, because it is reverse-biased. As a result, an initial low input supply, a DC-DC converter with a low ripple output
current will pass through the load, but the inductor will is needed. An effective Buck-Boost converter transforms DC
restrict the current, so it will start charging as well. In other voltage into a higher or a lower voltage [7]. The Buck boost
words, there is constant charge build-up in the capacitor that converter is especially useful for tracking PV optimum
reflects across the load during on time of the circuit. power. When the switch is on, it creates no resistance to
Energy stored in the inductor L1 decreases when the switch current flow, so all the current flows through it and back to
is in an off state and passes through the diode D1. Due to this the source. An inductor stores energy while the switch is on.
reverse polarity in the inductor, the diode is in forward-biased During the off-state, by reversing the inductor's polarity, it
condition [6]. In this case the inductor’s magnetic field releases its stored energy through the load resistance,
decreases, the current will flow through the load until it runs maintaining the same direction of current, and boosting the
out of the charge. Generally, Buck converters have a 90% output voltage because it serves as both an inductor and a
efficiency rate. source.

Figure. 12. Buck Boost Converter

Figure. 10. Buck Converter
V. RESULTS
B. BOOST Converter
The INC and P&O methods track maximum power points
There are many types of switch mode converters, one of steadily and determine the operating point at which maximum
which is a boost converter. In other words, it boosts or power is achieved using the proposed method. As described
increases an input voltage [7]. This circuit includes an within the method, the controller gives the PWM signal to
inductor (L), a MOSFET switch, a diode (D), and a capacitor regulate the voltage, the controller adjusts the voltage by
(C). A switch represents a short circuit when it is turned ON, Buck converters, and when the voltage increases, further
therefore, there is no resistance in its path, so the DC input modifications are attempted until the voltage remains stable.
source will receive the full amount of current. During the off For the calculations of system performance and to determine
state, the inductor has reverse polarity. The load resistance how well it performs, this model is needed. The PV voltage is
eliminates the inductor energy, thereby ensuring a current stepped down in the DC/DC buck converter model. To drive
flow in the same direction through the load. Inductor energy the converter switch, the pulse width modulation (PWM) a
act as an additional source in the combination with the input model is applied. It depends on the duty cycle.
source, so that the output voltage can also be increased.

Seventh International Conference on Aerospace Science & Engineering 82

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure. 13. Mppt Simulation

A. Perturb and observation method

Figure shows the Irradiance, Temperature, Solar array Vin, Solar array Iin, Solar array Pin, Iout, Vout, and Pout.

Figure. 14. Irradiance, Temperature, solar array Vin, Solar array Iin, Solar array Pin, Vout, Iout, Pout

B. Incremental conductance method

Figure shows the Irradiance, Temperature, Solar array Vin, Solar array Iin, Solar array Pin, Load, Iout, Vout, Pout.

Seventh International Conference on Aerospace Science & Engineering 83

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure. 15. Irradiance, Temperature, Solar array Vin, Solar array Iin, Solar array Pin, Vout, Iout, Pout

VI. Comparison between Perturbation and Observation VII. CONCLUSIONS

Technique and Incremental Conductance
Technique Based on features such as control variables involved and type
of control strategies, circuitry, manufacturing cost, the
For comparison purposes, the output data of the MPPT is algorithm technique is considered one of the best
compared to achieve a significantly better MPPT system. mathematical models in use. A comparison is made between
the incremental conductance technique and Perturbation and
Table I. Comparison between P&O and INC algorithm at Irradiance observation technique in this paper, and we find incremental
1000W/m2 conductance to have a 93% efficiency rate and perturb and
observe a 90% efficiency rate. Compared to perturb and
Perturb and Incremental observe, incremental conductance is cost-effective and more
observation conductance efficient. Therefore, it is demonstrated that incremental
Irradiance 1000W/m2 1000W/m2 conductance provides a better performance. By using these
Temperature 25°C 25°C algorithms, we improve the efficiency of DC-DC converters
Vin 31.93V 33.48V and improve the dynamics of photovoltaic systems.
Iin 29.02A 20.92A
Pin 926.6W 700.5W
Vout 29.02V 31.39V ACKNOWLEDGEMENTS
Iout 29A 20.92A Throughout the thesis, the faculty of Small Satellite
Pout 842.4W 656.7W Technology and Research Lab (SSTRL) / NCGSA, Institute
Efficiency 90% 93% of Space Technology (IST) guided and supported me in
choosing a topic of interest and monitoring my work. I have
Table II. Comparison between P&O and INC algorithm at Irradiance been enlightened about solar cells and power electronics by
500W/m2 them, who has provided various ideas. Throughout this thesis,
they helped and guided me in a way which was invaluable.
Perturb and Incremental
observation conductance
Irradiance 500W/m2 500W/m2 REFERENCES
Temperature 25°C 25°C [1]. Ananya Dutta, Niloy Barua, Aninda Saha. (2016). Design of an arduino
Vin 21.42V 29.39V based Maximum Power Point Tracking (MPPT) solar charge controller.
Iin 19.47A 18.37A [2]. B Abdelilah1, A Mouna1,2, N KouiderM'Sirdi2 and A El Hossain.
Pin 417.1W 539.9W (2018). Implementation of Maximum Power Point Tracking (MPPT) Solar
Charge Controller using Arduino.
Vout 19.1V 27.55V [3]. Hamidullah Kandahari,Rakesh Kumar. (2019). Maximum Power Point
Iout 19A 18.37A Tracking (MPPT) Solar Charge Controller. International Journal for
Pout 379.2W 508.1W Scientific Research and Development , 7, 6.
Efficiency 90% 93% [4]. Jesus Gonzalez-Llorente,Eduardo I.Ortiz-Rivera . (2013). Comparison
of Maximum Power Point Tracking Techniques in Electrical Power Systems
of Cubesats . 27th Annual AIAA/USU Conference on Small Satellites , (p.
6).
[5]. Md. Rokonuzzaman,Md Hossam-E-Haide. (2016). Design and
Implementation of Maximum PowerPoint Tracking Solar Charge Controller.

Seventh International Conference on Aerospace Science & Engineering 84

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
2016 3rd International Conference on Electrical Engineering and Information
Communication Technology (ICEEICT), (p. 5).
[6]. Ponugoti, R. S. (2016). Comparative Study of P&O and Incremental
Conductance MPPT Algorithms for Photovoltic Systems . International
Journal of Innovations in Engineering and Technology(IJIET) , 7.
[7]. Yejee Choi,Monineath Khun,Giselle Verbera. (2017). Maximum Power
Point Tracking.

Seventh International Conference on Aerospace Science & Engineering 85

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Formation Flying Strategy for Attitude

Determination and Control of CubeSat
Muhammad Abubakar Arshad Dr.Rehan Mahmood Dr.Mohammad Tahir
Small Satellite Technology and Small Satellite Technology and Department of Electrical Engineering
Research lab Research lab International Islamic University
Institute of Space Technology Head of Department Islamabad, Pakistan
Islamabad, Pakistan Institute of Space Technology [email protected]
[email protected] Islamabad, Pakistan
[email protected]

Abstract— This paper explains the comparison between

different types of sensors and actuators used to measure and
control the orientation of the spacecraft to determine spacecraft
position in space and direction in which it is pointing its target.
Typically, sensors used for this purpose are Magnetic, Star, Sun,
Earth and Gyro sensors. While, actuator on the other hand is
the reaction of the sensor which cause motion or action to keep
satellite in the right orientation. For possible two types of
controls, active and passive controls [1], we will discuss
actuators like reaction wheel, magnetic torque bars and
thrusters. The major constraints that take place for small
satellite configuration are mainly, size, cost and power Fig. 1. CubeSat with reaction wheel installed [2]
specification. Considering all of these, appropriate sensors
(magnetometer, sun sensor and earth sensor) are used for low A satellite mission requires a complete satellite
earth orbit (LEO). Presently, use of conical scan earth sensors orientation control system. The specific sensors and
are being genuinely studied and found to be effective for the use actuators altogether make up this whole system. CubeSat, a
of attitude determination (AD) in small satellite. While, small satellite used for mission, requires not too large and
controlling the satellite, it shall concentrate both; an effective expensive controls systems. Commercial off the shelf and
instantaneous communication and attitude controls method e.g., low-cost designs are installed in this small satellite. The
magnetic actuation (magnetorquer, permanent magnet and sensors (sun, star, earth, gyro, magnetic) and actuators (torque
reaction wheels). All in all, the paper holds discussion of
rods, thrusters, reaction wheels) with a detailed study have
different types of sensors and actuators that can be used for
small satellite missions, following; strategies of implementing a been implemented to design a more credible attitude control
sensor less Formation Flying method for satellites in (LEO) with system.
minimum budged and improved reliability in the field of small An Attitude Determination and Control Systems, is
satellite (CubeSat) communication. divided into two parts as stated by [3], the ADCS1 and
ADCS2. The ADCS1 performs the simple task of Detumbling
Keywords— Attitude determination, Attitude controls, of the satellite. The second part, ADCS2, performs the simple
CubeSat, Conical Scan Earth sensors, CubeSat Formation Flying, task of pointing control. The purpose of introducing these two
Gyro sensors, Magnetorquer, Magnetic sensors, Magnetic
systems is for reliability and accuracy. It also helps to support
Actuation, Reaction wheels, Sun sensors, Star trackers, Thrusters
in case, if one system fails, the other is present as a backup.
I. INTRODUCTION The main idea of this paper is to determine the most suitable
sensors and actuators that could be used for reliability and
The aim of introducing Nano-satellites has evolved over accuracy.
the years. One of the examples of Nano-satellites is CubeSat. The paper discusses the comparison of different types of
The idea of CubeSat processing and developing was proposed sensing (determination) and reaction (control) methods, their
by i.e., California polytechnic state university and Stanford differences and drawbacks for the satellite mission. First, in
University in 1999. They planned to test small satellite flights section 1, we briefly describe the specification and model of
for low earth orbit communication. CubeSat are typically the CubeSat for the mission. Section 2, describes the sensor
used by mass research stations for their research and students specification following the section 3 which shows the
for their training missions. They are generally of mass suitability and credibility. Section 4 explains the list of
ranging to a kilogram and dimensions of 10 x 10 x 10 cm3 [2]. actuators leading to section 5 detailing the pros and cons of
Such a CubeSat with reaction wheels installed is shown in different types of actuators used for spacecraft mission.
Fig. 1. Section 6 includes the future proposed plan and theory of new
ideas. Section 7 concludes the paper.
II. SENSOR AND ACTUATORS SPECIFICATION
The feedback in the control system has no purpose if no
suitable sensors are employed. The attitude of the satellite is
necessary to be estimated before the deployment and after it
is released in space. For this purpose, single or number of
sensors are placed to attain the accurate estimation of satellite.

Seventh International Conference on Aerospace Science & Engineering 86

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

The sensors have two classes in which they are categorized; estimation. The compact design of this sensor makes it
the reference one and the other are inertial [5]. These both are convenient use for small satellite missions. A CES sensor had
sensors but the difference is between the frames of references been considered a great source of satellite attitude
in which they are operated. Sun, star, and earth sensors determination application with cost saving and its wide field
incorporate the field of reference for their output estimation of view (FOV) [13].
whereas the inertial sensor includes gyros which work
because of the inertial frame. As for the point of convenience
TABLE I. ADVANTAGES AND DISADVANTAGES OF SENSOR TYPES
of mission, both frames are important for the attitude USED IN NANO SATELLITE [[5] [6] [7]]
estimation process.
A. Sun Sensor Sensors Advantages Disadvantages
Sun sensor or a reference sensor, often a measurement is High power
estimated from a particular reference for the given instant of
consumption
time. The basic work of a sun sensor is simple, to determine
Large in size
the angle between the base and the incident light falling upon
a small opening (slit). They are typically used nowadays in Star sensor High accuracy Cost expensive
spacecraft missions, however, needs an un-obstructed (clear Spinning may cause
view of visible light) field of view. As LEO has to face eclipse the sensor to not
periods, then somehow AD shall counter that problem and work properly
provide a clear reading on the loss of reference (sun). During eclipse it
Typically, sun sensors have an accuracy ranging between does not work
Simple
0.05-4 degrees [11]. Disturbed by earth
Low power
Sun sensor albedo
B. Star Sensor consumption
Some two axis
The best and the far more accurate method are using the Cheap
sensors are
star sensors for the attitude estimation of the space satellite. expensive
The sensor works in a coordinate system, the position of star Simple sensor Not much accurate
present at a distance is measured with respect to object Smaller in size Magnetic field not
(satellite) and then this estimation is fed into system to locate Magnetic
Cheaper fully known
the observed star position from its (star) history of behaviour sensor
Available in huge Only useful for low
(movement). This method is accurate (arc second to arc
minutes) [11], while its costly, for such small satellite space amount earth orbits (LEO)
missions. Large size
High power
C. Magnetic Sensor Helps in during
Gyro sensor consumption
the eclipse
The principal object used for the measurement of earth’s Some of them are
magnetic field is magnetometer. The use of magnetometer in expensive
the space satellite defines the measurement of earth’s Simple
Accuracy affected by
magnetic field and comparing it with the observed value for Earth Earth, the
sun and moon
the attitude estimation of the satellite. The method is quite sensor reference is
convenient to use but some internal and external disturbances always present
during mission makes it inaccurate and falls below from other
F. Credibility of combined Sensor Techique
methods (star and earth references model) used for this
particular attitude estimation technique [5]. According to the availability, a reference sensor makes it
a lot easier for the estimation procedure, however during
D. Gyroscope (Gyro Sensor) eclipse time the reference sensor may fail to achieve the goal,
The word gyro involves the inertial measurement of the in that case inertial frame of sensor can be used; in conclusive,
satellite; it is basically a mass spinning around which to perform the Attitude Estimation process of the satellite.
determines the speed and angle of the spacecraft orientation. Reference and inertial frame of sensors are used altogether
The process is entirely blank regarding external references [12] e.g., a sun sensor (reference sensor), because of its low
involved around the spacecraft environment. The mechanism cost and less space occupancy, they are highly used sensor for
is durable with a good accuracy rate if this mechanism the Attitude Determination process and along using a gyro
coexists with external reference or point. sensor (inertial sensor) can help the estimation process during
eclipse. Similarly, using magnetic actuation with momentum
E. Earth Sensor (Horizon Sensor)
wheels benefits the system controls in providing high
Earth sensors like sun are also known as reference pointing accuracies [14].
sensors, as they use the reference object to determine the
attitude of the space satellite. They are usually an infrared G. List of Actuators
sensor, study the difference of temperatures between earth After studying and intelligent formation from a suitable
and space environment. Many spacecraft pointing nadir sensor application the use of actuators comes in hand to
directions have a wide FOV (field of view) which enables control the satellite’s attitude. As actuator is nothing but an
them to view entire earth and stabilize the spacecraft assembly of mechanical hardware to produce suitable torque
according to it [4]. Earth sensor e.g., Conical Scan earth in response to the attitude control of the satellite.
sensor (CES) used in small satellite uses earth as a reference
to point roll and pitch angles for the process of attitude

Seventh International Conference on Aerospace Science & Engineering 87

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

H. Reaction Wheels reaction control thrusters were used to control the satellite;
Reaction wheels mainly involve the moving mechanical the mechanism used, was the switching one of those thrusters
parts, which rotate at higher speed [8], they work on the which were needed to control the stability [9]. Thrusters on
principle of Newton’s third law. The mechanical part moves one side are found to be effective, by reverting satellite back
with a great speed and produces torque in the opposite on track (original in which it was moving) however it
direction of the rotation. However, increasing momentum and occupies mass in satellite which reduces the mission life.
small satellite specification this cannot be used extendedly. Thrusters were officially made an unreliable approach by
NASA as the uses of magnetorquer were compared to be
I. Control Moment Gyro effective in orbits closed to earth e.g., LEO [10].
Control moment gyros (CMGs) basically provide the
M. Spin Stabilization
active control mechanism for the process of AC of the
spacecraft. CMG, composed of a spinning wheel in the center The method proposes the rotation of the satellite on its
(reaction wheel) attached to one or more gimbals which tilt axis; it’s particularly a movement which causes a gyroscopic
the angular momentum of the rotor in reaction to certain error effect to take place by such movement in order to control the
in the control systems (Fig. 2). They are considered very attitude of the satellite. Usually spinning satellites are divided
quickly, with high response time, however some limitation to into three categories; passive spin, spin with precession
their weight and higher cost they are disregarded [8]. control and dual spin [11]. Satellite which are independent of
control authority are known as passive spin, contrast to those
having spin authority, spins with precession control, however
the most effective of them is the dual spin, which is a spinning
central mass and a platform of de-spun tracking the target
[11]. Although, spin stabilization techniques are found in
great number but this kind of method is generally unreliable,
as such movement and satellite itself revolving around LEO
with great speed would cause disturbance in communication.

TABLE II. DIFFERENT METHODS FOR SATELLITE CONTROL [12]

Degrees of
Method Performance
Accuracy
Response time is
Reaction Wheels 0.01 quick
High precision
Fig. 2. Control Moment Gyros (CMG)
Quick response
Thrusters 0.1 time
J. Parmanent Magnet High authority
Permanent magnet stabilization is similar to the Are light in
magnetorquer but the difference which accounts, is about weight
Magnetorquer 1-2
power consumption. The permanent magnet needs no power Low cost
from the satellite’s ADCS (attitude determination and control Used near Earth
system), this benefits the two-axis stabilization of the satellite Simple in use
Spin
for its entire life. Whereas during flight time, oscillation 0.1 Cheaper
Stabilization
accomplished, cannot be removed easily by this permanent Inertially oriented
magnet control which makes it unsuitable for the satellite Very fast
Control Moment
mission requirement [3]. 0.1 response time
gyro
High authority
K. Magnetorquer
The method uses the comparison of produced magnetic
field with the earth’s magnetic field and finally works on N. Assets and Liabilities
rotating the satellite in particular position until the two get According to the flight configuration and CubeSat
aligned. This is useful and much reliable method to control compatibility every control has its pros as well as cons;
the satellite in regions with a strong field like, LEO (low earth however, for a smooth mission more than one control
orbit) because the mechanism involves no mechanical altogether is required. Like magnetorquers used with
rotation or momentum, where it only compares the measured permanent magnet make it sufficient for the attitude control
magnetic field with the reference field (one present in the of a small satellite.
reference model) [5], however it should be perpendicular to Based on structure and the type of flight, some controls are
the geomagnetic field of the earth. left behind; as because of the structure of reaction wheels, that
are totally incompatible with CubeSat ADCS, similarly on the
L. Thrusters other hand considering spin stabilization as the source for
Thrusters are the key approach in basic satellite attitude control actuator makes it difficult for the satellite to complete
control; the method uses Newton’s third law in such a way its mission (communication) because of such high-speed
that, expulsion of propellant at high velocity gives a rotation along its axis. Thruster is considered to be the typical
momentum (push) in the opposite direction. In past 2010, control actuator for satellite, but as their propellant needed

Seventh International Conference on Aerospace Science & Engineering 88

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

some space to be stored, so, later it could be expelled when A. Formation Flying of CubeSat
needed during attitude control; which deliberately occupies GNSS (Global Navigation Satellite System) is the kind of
space of ADCS which significantly shortens the life of small navigation system which provides autonomous Geo position
satellite. with global coverage. A good precision and accuracy for
position and satellite orientation could be obtained from this
III. DISCUSSION AND ANALYSIS
GNSS. This area of research is evolving and could later be
For a particular instant, key is to choose the right sensors used for many future experiments by many research students
and actuators which can be effectively used in the limited and and scientists. Similarly, an approach being studied is use of
constraints of a small satellite, which involves; power GNSS in Formation Flying of small satellites and who’s
consumption, size and cost of the sensor and actuator used to computational efficiency is still part of active research i.e.,
construct a reliable source of system for attitude use of interferometric principle for the attitude determination
determination and its control. using GNSS phase measurements [15]. This method uses a
As discussed earlier, main sensors which are used for pair of GPSs (global positioning system) antenna, user can
determination process are Star, Sun, Earth, however, cost of determine the attitude of the satellite using the phase
choosing the right sensors for this process depends upon difference between like signals of that antenna pair, which
structure, power consumption, size of the sensor, cost and makes this technique more accurate and precise. A similar
how much efficient it is during communication e.g., the use approach to Satellite Formation Flying (SFF) was the use of
of gyroscope is not a good choice for the system as its weight line of sight (LOS) and Ranges between the satellites (Deputy
and large power consumption makes it inappropriate for the and Chief). The method uses a relative state estimation
flight. Where using a star tracker or horizon sensor (scan between deputy and chief satellite vehicle using the data
wheel) occupies space and is cost effective, which makes it acquired from, LOS and ranges between them [16].
inconvenient for the process of attitude determination (AD). For active research, students and many researchers are
With limited budget and size (space), leaving choice for only attracted towards the Formation Flying Strategies of CubeSat,
magnetometer, sun sensor and currently most used which enables users to build low-cost and higher success rate
application of conical scan earth sensor. Sun sensor; cheap design for determination process of satellite’s orientation like
and smaller in size makes them suitable for attitude i.e., Relative Attitude and position determination of satellite in
determination. Magnetometer, surpassing every sensor, e.g., Formation Flying of CubeSat, however, this was for the time
during eclipse, sun sensor fails to deliver the information for being only helped for one Satellite (Deputy Vehicle) [17].
attitude where magnetometer copes with this abnormality Later, in the same research field, successive results lead in
effectively [3]. There is advancement in earth sensors; where determining relative state of multiple small satellites using
conical scan earth sensor has found to be effective in the field array signal detection; the method was quite adoptive in sense
of attitude determination of small satellites. With a wide of Formation Flying, it explains how one satellite (deputy
conical field of view (FOV) and a tilt angle of 40 degrees, use satellite) follows and estimates its attitude with respect to
of such two sensors in small satellite covers almost all the another satellite (chief satellite). The procedure follows,
antenna arrays present on both satellite, deputy as well as
surface to which it’s pointed. Another important aspect,
chief. Furthermore, using angle of arrival and angle of
where many satellites use different sensors for three axis
departure attitude knowledge is calculated by deputy satellite
spacecraft and transfer orbit operation; conical scan earth relative to chief satellite [18]. This use of Formation Flying in
sensors can perform these two-operation making it very cost multi-CubeSat enables a sensor free design (free from large
effective for a budgeted satellite system [13]. volume, size and power consuming sensors) and an approach
To control the satellite, actuators system is installed and which is also more extensive in providing better accuracy than
for firm communication, combination of actuators is used GPS based navigation system.
altogether. Using technique, in which magnetorquer and
permanent magnet; with low power consumption and a
lifespan of lifetime for the permanent magnet make them ideal
for the mission. Where, accuracy and light structure for V. CONCLUSION
magnetorquer (does not involve any mechanical movement) This paper contributes in overviewing the basic idea of the
makes them a wide source of application, collectively, these sensors and actuators that could be use in the process of
both make a reliable actuator system for the control of small attitude determination and control system (ADCS). The
satellites [3]. In another approach of controls deals in magnetic different types of sensors, actuators are outlined extensively
actuation, use of magnetorquer with momentum wheels (a with their use and drawbacks for AD and CS respectively.
hybrid system) make a better system for controls of satellite; This paper can be helpful in the field of designs and
helping both, damping of mechanical torque and providing implementation of ADCS for Formation Flying Concept for
better pointing accuracy for satellite [14]. small satellite. The purpose for building this approach is to
IV. FUTURE PROPOSAL counter the use of conventional sensors reducing the size,
weight and cost. With this report, researchers could use this
The paper states the types of sensors used for the attitude approach in implementing various designs, flights, and novel
estimation and control actuators for the possible control approach to determine the orientation of a small satellite
system for small satellite. As with successive research and
using this advance study of GNSS for the Formation Flying
study in the field of astronomy; students and scientists are
working on the new details to make satellite communication of CubeSat [17].
easier with limited constraints. ACKNOWLEDGMENT
The contribution of this paper shows the unique idea
of formation flying strategy and it’s use in Nano satellite
technology, CubeSat in IV. Furthermore, this paper also

Seventh International Conference on Aerospace Science & Engineering 89

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

shows and explain this different method which briefs how Design Division Rept. NASA/TP-2015-216648/REV1, Moffett Field,
CA.
they are different from random conventional sensors
[11] Larson, W. J., & Wertz, J. R. (1992). Space mission analysis and
technique which wholly makes system crowded and design (No. DOE/NE/32145-T1). Torrance, CA (United States);
unreliable for specification like, CubeSat. Microcosm, Inc.
[12] CubeSat Design Specification. Tech. rep. California Polytechnic State
REFERENCES University, 2009.
url:http://www.cubesat.org/images/{\%}0Ddevelopers/cds {\_}
rev12.pdf
[1] P. Hendrik, "Development of attitude controllers and actuatores for
solar sail cubesat," March 2011. [13] Fowler, R. Z. (1981). THE USE OF THE CONICAL SCAN EARTH
[2] M. P. T. W. R. L. Junquan Li, "Design of Atitude control systems for SENSOR IN COMMUNICATION SATELLITE APPLICATIONS.
International Foundation for Telemetering.
CubeSat-Class nano satellite," Journal of control science and
engineering, 2013. [14] Tebbani, S., Makarov, M., Maniu, C. S., & Hinostroza, I. (2016, May).
Nanosatellite attitude control using electromagnetic actuation. In 4th
[3] Jensen, K. F., & Vinther, K. (2010). Attitude determination and control
SONDRA Workshop.
system for AAUSAT3. Master's Thesis, Aalborg University.
[4] Reigber, C., (Ed.) (1996). Champ, Phase B. Scientific technical report [15] Garrick, J. 1996. Investigation of Models and Estimation Techniques
STR96/13. champ project office. for GPS Attitude Determination. In Flight Mechanics/Estimation
Theory Symposium 1996 (p. 89).
[5] T. Bak, "Spacecraft attitude determination- a magnetometer approach,"
Aalborg University, 1999. [16] Wang, X., Qin, W., Bai, Y., & Cui, N. (2016). A novel decentralized
relative navigation algorithm for spacecraft formation
[6] J. R. Wertz, "Spacecraft aatitude determination and control," Kluwer flying. Aerospace Science and Technology, 48, 28-36.
academic publishers, 1994.
[17] J. Wang, J. Yuan, R. Zhang, and H. Zhang Relative position and
[7] Arbinger, C., & Enderle, W. (2000, September). Spacecraft attitude attitude determination for CubeSats and drones using signal direction
determination using a combination of GPS attitude sensor and star based on antenna arrays. In Proc. Int. Astronaut. Congr., Guadalajara,
sensor measurements. In Proceedings of the 13th International Mexico, Sep. 2016, pp. 1–7.
Technical Meeting of the Satellite Division of The Institute of
Navigation (ION GPS 2000) (pp. 2634-2642). [18] Wang, J., Zhang, R., Yuan, J., & Luo, J. (2019). Multi-CubeSat relative
position and attitude determination based on array signal detection in
[8] Paula Sorolla Bayod “ Study and Design of the Attitude control of a formation flying. IEEE transactions on aerospace and electronic
CubeSat 1U based on reaction wheels”, Bachelor’s Degree Thesis, systems, 55(6), 3378-3393.
September 30, 2019
[19] Naqvi Najam Abbas and Dr. Li Yan Jun, “The Quest for Optimal
[9] Curti, F., Romano, M., & Bevilacqua, R. (2010). Lyapunov-based Spacecraft Attitude Determination Filtering Algorithm Using GNSS
thrusters' selection for spacecraft control: analysis and experimentation. Phase Measurement: An Overview”, 2nd International Conference on
Journal of guidance, control, and dynamics, 33(4), 1143-1160. Signal Processing Systems (ICSPS), 2010.
[10] Agasid, E., Burton, R., Carlino, R., Defouw, G., Perez, A. D.,
Karacahoglu, A., ... & Shimmin, R. (2015). Small spacecraft
technology state of the art, NASA, Ames Research Center, Mission

Seventh International Conference on Aerospace Science & Engineering 90

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
Overview of ICUBE-N: National C ube Sat of Pa kista n

Overview of ICUBE-N: National CubeSat of

Pakistan
Usman Imran
National Centre of GIS and Space Muhammad Asfandyaar Arshad Arooj Shafique
Applications National Centre of GIS and Space National Centre of GIS and Space
Institute of Space Technology Applications Applications
Islamabad, Pakistan Institute of Space Technology Institute of Space Technology
[email protected] Islamabad, Pakistan Islamabad, Pakistan
asfandyaar @mail.ist.edu.pk arooj.shafique @mail.ist.edu.pk

Rehan Mahmood Rizwan Mughal Hayat Muhammad Khan

Department of Electrical Engineering Department of Electrical Engineering Department of Aerospace Engineering
Institute of Space Technology Institute of Space Technology Institute of Space Technology
Islamabad, Pakistan Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected] [email protected]

Abstract— The miniaturized version of satellites encouraged Institute of Space Technology (IST) started the development
cheap access to space at the university level for research purposes. of ICUBE-1 [6], a low-cost small spacecraft back in 2009 that
The miniaturization of electronic components and advancements in was successfully launched in 2013. ICUBE-1 carried an
engineering technologies have paved the way for developing and imaging payload that had a low-resolution CMOS camera
launching small, low-cost spacecraft. IST initiated the satellite responsible for taking images of earth. The ground station
program back in 2010, ICUBE-1 a 1U CubeSat, and was launched facility exists already at IST along with a clean box and
in 2013. The second satellite of this program is ICUBE-N, a 3U vibration testing facility.
CubeSat and is in development. The mission’s primary goal is to
launch ICUBE-N and establish communication with the ground IST with the support of the Higher Education Commission,
station successfully. The secondary goal is the technology Pakistan initiated the project of an experimental nanosatellite
demonstration and data acquisition from different payloads, named ICUBE-N which is a 3U CubeSat. The different
including a camera. In this work, preliminary design and analysis of universities in Pakistan will collaborate in the design and
different subsystems of ICUBE-N are discussed in detail, along with development of the payload. ICUBE-N will carry a high-
its required testing and technical support facilities. The presented resolution camera as a payload along with other payloads
agile systems engineering approach and components selection will which are yet to be finalized. The development of a clean room
help develop a small budget and educational CubeSats missions. is underway and other facilities like thermal vacuum chamber,
vibration testing facility and the ground station would be
Keywords— CubeSat, Space Technology, Global Positioning
System (GPS), Onboard Computer, Payload, Tracking, tele-
upgraded too.
command, telemetry (TTC), ADCS
ICUBE-N project aims to facilitate the students in gaining
experience and skills which are essential for the development
I. INTRODUCTION of nanosatellites. The main objective of ICUBE-N is to design
Launching satellites for research and space exploration had and develop indigenous modules instead of Commercial off-
always been carried out by national space organizations due to the Shelf (COTS) systems which could help later in capacity
huge funding requirements which constrained initiating such building. The secondary mission objective is to successfully
projects at the university level. The standardized collect payload data and communicate with the ground station.
specifications of CubeSat were defined for educational and The knowledge and experience gained during the operation
research institutes for easy access to space. CubeSat concept and development of ICUBE-1 would be utilized in the
containing certain specifications was first proposed by development of complex subsystems for ICUBE-N. ICUBE-
professors at California Polytechnic State University [1] in N will enable the involvement of IST students and graduates
1999 which later became a standard for the development of in all stages of CubeSat development, operation, and launch.
education satellite. Until now, around 1600 CubeSats have
been launched since its inception. CubeSat is a small satellite Section II provides the mission overview consisting of ground
consisting of one or more units where one unit (1U) is of size track and operational modes for ICUBE-N. The detailed
10 cm × 10 cm × 10 cm and mass 1.33 kg. CubeSat can be system design for the structure, power, communication,
arranged in 2U, 3U, 6U, or multiple unit configuration by onboard computer, attitude determination, and control system
following the specifications [2]. CubeSat missions can range and payload are provided in Section III. Section IV explains
from educational projects to technology demonstration [3], the engineering facilities consisting of a ground station, clean
science missions like measuring electron precipitation [4] and room, thermal vacuum chamber, Helmholtz coil, and vibration
providing commercial services. The Poly Picosatellite Orbital testing facilities. Section V summarizes and concludes the
Deployer (P-POD) [5] built by California Polytechnic paper followed by acknowledgment and references.
University ensures the safety, proper deployment of CubeSats
and reduces the concerns of the launch providers. P-POD has II. MISSION OVERVIEW
gone through various improvements and modifications since
its inception. The objective of ICUBE-N is to design and develop the
subsystem modules indigenously instead of using commercial

Seventh International Conference on Aerospace Science & Engineering 91

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

off-the-shelf components. In addition to that, ICUBE-N aims the power saving mode in which only the onboard computer
to indulge the other universities of Pakistan in the will be in a power-on state. ICUBE-N will shift to normal
development of payload subsystems as it is the national mode when the power of batteries reaches the required
CubeSat of Pakistan. Apart from this, the aim is to threshold and performs the check on housekeeping tasks and
demonstrate the capability of low-cost nanosatellites to communication of the beacon signal. The block diagram of
observe the properties of the earth’s atmosphere and space operational modes is shown in Figure 2.
using scientific payload subsystems. The other objective of
ICUBE-N is to bring the onboard components, subsystems, ICUBE-N will execute the payload tasks like image
and instruments to the technology readiness level so that they capturing, transferring to memory storage, etc. The payload
can be implemented in future scientific missions with more and onboard data are saved on the memory storage device
reliability. which is later transferred to the ground station. While
contacting the ground station, ICUBE-N shifts to transmitting
ICUBE-N is anticipated to be propelled in a low earth orbit mode to send the data to the ground. It can also receive
(LEO). The altitude of orbit will range from 550-650 km but commands and instructions from the ground station to
for investigation, an altitude of 600km is chosen. The LEO is perform accordingly. In case of any anomaly or undesirable
selected so that ICUBE-N is accessible from everywhere on situation, ICUBE-N will switch to emergency mode. ICUBE-
earth. The slant of ICUBE-N is around 97° concurring with N will restart onboard devices in emergency mode and try to
the selected parameters. The orbital period of ICUBE-N is bring the system in the power-saving mode.
96.62 minutes with a maximum obscure time of 34.27
minutes. The thermal and power generation for ICUBE-N are
analyzed considering the worst-case scenario of dawn and
midnight. The STK simulation is used for obtaining the
ground track of ICUBE-N as shown in Figure 1. ICUBE-N
can communicate with the IST ground facility 4-5 times a day
with a communication period of 53 minutes per day
approximately.

Figure 2 – Operating Modes of ICUBE-N

III. SYSTEM DESIGN

The draconian space environment and payload onboard

ICUBE-N demands highly efficient system design
considering power, attitude, and control. ICUBE-N consists
of various subsystems just like its predecessor [7] ICUBE-1.
Each subsystem is developed in-house instead of using COTS
modules. The subsystems are described in detail in the
following sections and shown in Figure 3.
Figure 1 – 3D view of ICUBE-N ground track

There are five operating modes of ICUBE-N which are

Normal, Transmit, Payload, Emergency, and Power saving
mode. The modes can be changed manually and
automatically. It can be shifted manually by sending the
telecommand from the ground station. The onboard computer
is responsible for automatic switching of the modes
depending upon the situation and requirements.

The deployment cycle is initialized instantly after deployment

by following the guidelines [2] given in the CubeSat
specifications. When the CubeSat is deployed from P-POD,
the system bus is connected to the batteries after the release
of the deployment switch. A delay of 10-25 minutes is
essential before deploying any structure, such as boom or Figure 3 - Subsystems of ICUBE-N
antenna, so that the CubeSats in the same P-POD have enough
separation. After deploying successfully, ICUBE-N will be in

Seventh International Conference on Aerospace Science & Engineering 92

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Structure: The structure for ICUBE-N is designed by rate of deterioration over time. Moreover, each battery has a
considering the space material [8] and structural dimensions thermostat device that can be operated via I2C bus, ensuring
for the standard CubeSats [2]. After reviewing the available that the temperature remains within the working range for
structures, it is determined to develop an in-house structure optimal efficiency.
that would help later in the capacity building. The structure
for ICUBE-N is a 3U structure of dimensions 340mm X Communication: One of the most important subsystems of
100mm X 100mm and mass less than 0.3kg which is very ICUBE-N is the communication subsystem. A satellite is
lightweight. useless without dependable communication. For uplink and
downlink communication in ICUBE-N, amateur radio bands
The structure is portable, and all the faces of structures can be would be employed. Amateurs and hobbyists are welcome to
removed easily. The structure consists of rails, ribs, standoffs, utilize the amateur frequency band, which has few
and L-pieces as shown in Figure 4. The PC104 stacks [9] can coordination requirements. In addition, several ground
be easily placed anywhere within the structure using internal stations currently use this frequency range, allowing ICUBE-
ribs. The choice of materials is critical in any discussion of N to interact with other ground stations. After consulting with
tiny spacecraft construction. Mechanical qualities like the International Amateur Radio Union, the downlink and
modulus, stress, and toughness, as well as physical features uplink frequencies would be determined. The uplink for
like thermal properties, and radiation resistance, must all be ICUBE-N will be in the VHF band, while the downlink will
satisfied. These are focused on payload designs, as well as be in the UHF band. Commands would be sent to the satellite
mission readiness and deployment procedures, with via uplink rate of 1200bps using Audio Frequency Shift
knowledge of rideshare safety issues. Protection switches Keying. The downlink would employ Binary Phase-shift
which include the remove before flight pins, footswitch, keying with a transmission rate of 1200 bits per second to
and the need for spacecraft to be shut off when placed in the collect telemetry data from the satellite. The system of
deployer, are all considered. The structure is made up of communication can send out a CW beacon in Morse code at
Aluminum 6061-T6 and the surfaces are anodized to prevent regular intervals. It is used to determine the CubeSat and to
the ICUBE-N from external electrical field and radiations. send data on the spacecraft’s fundamental health.

The output power of the transceiver utilized will be around

350mW. For packet communication, the AX25 connection
protocol would be utilized. Uplink and downlink will be
served by monopole and dipole antenna. The antennas would
be positioned at the launch sequence.

On-Board Computer (OBC): The primary goal of an onboard

computer is to process and control all the data and connect
with other satellite subsystems. ICUBE-N is developing in-
house OBC instead of using a commercial off-the-shelf
module. The onboard computer (OBC) will consist of EFM
32 microcontroller with external RAMs and ROMs. The
Figure 4 - Structure of 3U ICUBE-N block diagram of OBC for ICUBE-N is shown in Figure 5.

Power: The power generating capabilities of CubeSat is The EFM 32 microcontroller is a 32-bit ARM Cortex M4
always restricted due to its compact physical size. With controller operating at 72MHz. This controller has the
advancements in solar cell fabrication techniques, it is now
feasible to produce Ultra Triple Junction Cells that are more
than 30% efficient, nearly twice the efficiency of ordinary
Silicon solar cells. Advanced Triple Junction cells will be
used in ICUBE-N. ICUBE-N has solar panels on its faces.
Each cell has an efficacy of more than 28 percent at room
temperature.

The in-house power module has been designed and can offer
both regulated and unregulated power at +3.6V and +5.5V.
The power supply can also be used with the OBC and other
ICUBE-N subsystems. Battery charge regulators are also
included. The voltages and currents of the battery and buses,
as well as the temperature of the different EPS components,
make up the health data of the electrical power system. The
data is sent to the OBC through the I2C protocol.

To keep all the systems active during eclipses, rechargeable

Li-ion batteries are linked in series with the EPS. Each battery Figure 5 – Block diagram of OBC
has a voltage of 3.7 volts. Lithium polymer batteries are
popular because they have a high storage density and a low capability of SPI interface, SD card, and CAN bus. Its

Seventh International Conference on Aerospace Science & Engineering 93

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

features include 144 GPIO pins, 2 MB of flash program gyroscope by measuring angular rate, translation with help of
memory, and 512KB of data RAM. It also hosts an integrated accelerometers by measuring specific acceleration
DC buck converter. The serial interfaces consist of UART,
Ethernet, I2C, and CAN controllers. The survey of different Payload: The ICUBE-N is going to have several payloads.
electrical bus interfaces is provided in [10]. The One of which is a high-resolution camera. The other payloads
microcontroller also hosts two Analog to Digital converters are yet to be finalized. The camera with long-range and high-
and two Digital to Analog converters. It supports a 16-bit resolution would result in taking better images of the earth for
timer and there is a total of 7 timers available. It has a low later observation. The images would be saved on the SD card
power consumption of around 80μA in energy mode. memory and would be later sent to the Ground Centre. The
special lens would be considered for the draconian space
The connections between OBC and other subsystems of environment. The image will initially be compressed using
ICUBE-N are made through the PC-104 stack. The PC-104 the image compression technique because of the
stack consists of 104 pins which are assigned different lower downlink data rate. The picture compression is carried
functions based on the standard so that pins are compatible out by a dedicated microcontroller According to the
with other commercially available subsystems. The Micrium requirements, the information will be transferred to
operating system which is a real-time operating system the Ground Centre.
(RTOS) is used for programming the EFM32 controller using
Simplicity studio.
IV. ENGINEERING FACILITIES
Some of the libraries from FreeRTOS would also be used in Sophisticated engineering labs are required to effectively
software development. The tests will be performed to validate design, build, and incorporate the ICUBE-N. To increase the
the working of OBC in the extremely harsh environment of reliability of the spaceship and its components, appropriate
space bearing all the radiations, temperature changes, and test processes are specified [11]. A satellite’s certification test
vibrations. verifies that each of the satellite parts and systems is
operational and sustainable at launch in the environment of
Attitude Control: Attitude Determination and Control System space during its lifespan. Vibration, electromagnetic
(ADCS) is designed and developed instead of purchasing a and thermal tests are among the checks that may be
commercial module as ICUBE-N is following the in-house necessary. Moreover, the construction of a base station
development strategy. Active control is based on the feedback capable of communicating with the satellite is critical. IST is
principle, in which the spacecraft’s attitude is sensed and working to build and enhance the engineering and
torque is supplied to adjust it. The reaction wheel is used as a maintenance infrastructure needed for the ICUBE-N project.
common actuator for active attitude control (AAC). The Most of the laboratory equipment will be provided by IST.
reaction wheel has high power and weight but results in fast However, other universities and organizations’ assistance and
and continuous feedback resulting in more efficient control. testing resources would be used as well.
The control block diagram for AAC is shown in Figure 6.
Ground Station: As part of its space development effort,
IST constructed a spacecraft ground station during the
development of ICUBE-1. The same ground station would be
upgraded to match the needs of ICUBE-N. The ground
station is capable of independently getting remote sensing
data. With independent Yagi-Uda cross antennas for both
lateral and vertical orientation, the base station can send and
accept data in the VHF and UHF frequencies. The image of
the ground station is shown in Figure 7.

Figure 6 - Control block diagram for active attitude control

The number of reaction wheels equals the number of axes of

torque. Usually, three wheels are required for three axes. The
vibrations which are induced by imbalances of the reaction
wheel are also considered while designing ADCS. Several
sensors are also installed on the ADCS module which
includes Global Positioning System (GPS), magnetometers,
and inertial measurement unit (IMU). The GPS receives data
from a number of satellites revolving around the earth and
gives an accurate position for our system. The magnetometers
get attitude by measuring the magnetic field and comparing it
Figure 7 – Ground Station at IST
to the modeled magnetic field.
For receiving S-Band transmissions, a tiny dish is also
The IMU is an integrated unit that consists of hardware installed. The alignment of antennas is controlled through
sensors and software. IMU measures rotation with the help of computerized software by rotor assembly. A coaxial cable
producing 100W of electricity connects the satellite ground

Seventh International Conference on Aerospace Science & Engineering 94

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

station transceiver to the antennas. It was created specifically gained from this research will serve as a foundation for the
for satellite communications. To encode, decode, and institute’s subsequent space missions.
packetize digital data, many TNCs are linked to the ICUBE-N design emphasizes the development of in-house
transceiver. The ground station can follow spacecraft modules for capacity building instead of the utilization of
autonomously. space-proven components and extensive COTS components.
The in-house development includes the design and
Clean Room: Satellites are built in a clean room to minimize
development of ICUBE-N structure, OBC, ADCS with active
dirt and surface dust, particularly on parts like solar panels.
attitude control system and electrical power system
The spacecraft’s endurance is improved because of this
module. The predicted environmental assessments, as well as
advancement procedure. IST already has a clean box that was
the orbital attitude, have been accomplished. Using these
used at the time of ICUBE-1 and is now working on
simulation settings and analysis findings, an
constructing a class 10,000 cleanroom for ICUBE-N.
initial engineering model of ICUBE-N is being developed
Thermal Vacuum: The configuration of a CubeSat's thermal and will be integrated later in the flying model. Subsequent
subsystems and the trajectory in which it operates influence space missions will benefit from the in-house development of
its thermal requirements. To guarantee the spacecraft’s ICUBE-N.
correct operation, all parts and subsystems must withstand
high temperatures. The modules and the entire CubeSat will ACKNOWLEDGMENT
be placed in a TVC to assess and qualify them according to The authors would like to thank the IST administration for all
the specified thermal environment. The CubeSat is subjected their support and the Higher Education Commission of
to a thermal cycle and evaluated in accordance with the Pakistan, and the National Centre for GIS and Space
requirements [12]. Applications for initiating and funding this project for the
development of ICUBE-N, the national CubeSat of Pakistan.
Vibration Testing: ICUBE-N subsystems and elements must
all endure the shipping and launch procedure. During its REFERENCES
flight, the launch vehicle creates severe vibrations and
[1] J. Puig-Suari, C. Turner, and W. Ahlgren, “Development of the
accelerates by many g’s. ICUBE-N structure must be able to standard CubeSat deployer and a CubeSat class picosatellite,” in
withstand these severe vibrations and remain unaltered not IEEE Aerospace Conference Proceedings, 2001, vol. 1, pp. 1347–
just for the mission’s success, but also for the health of the 1353. doi: 10.1109/aero.2001.931726.
other cargo and the launch vehicle. Depending on the launch [2] CalPoly, “Cubesat design specification,” The CubeSat Program,
vehicle, vibrations might vary from 20Hz to 2000Hz. California Polytechnic State …, vol. 8651, no. June 2004, p. 22,
2009.
Helmholtz Coil: ICUBE-N has an active attitude control [3] M. R. Mughal et al., “Aalto-1, multi-payload CubeSat: In-orbit
system that comprises a reaction wheel and a number of results and lessons learned,” Acta Astronautica, vol. 187, pp. 557–
568, Oct. 2021, doi: 10.1016/j.actaastro.2020.11.044.
sensors like IMU, GPS. The control of ADCS must be tested
[4] FORESAIL collaboration et al., “FORESAIL-1 cubesat mission to
in a homogeneous magnetic field. Helmholtz coils can be measure radiation belt losses and demonstrate de-orbiting,”
used to create a homogeneous magnetic field. The Helmholtz Journal of Geophysical Research: Space Physics, vol. 124, no. 7,
coil facility is located at IST, and it will be used to evaluate pp. 5783–5799, May 2019, doi: 10.1029/2018JA026354.
the attitude control system. [5] W. Lan, R. Munakata, R. Nugent, and D. Pignatelli, “Poly
Picosatellite Orbital Deployer Mk III Rev. E User Guide,”
Air Bearing: A spherical air bearing for testing the attitude California Polytechnic State University T, vol. The CubeSa, pp. 1–
control of ICUBE-N is in the development phase. The air 21, 2014.
bearing can test the three axes control, pitch, roll, and yaw, [6] R. Mahmood, K. Khurshid, A. Zafar, and Q. ul Islam, “ICUBE-1:
First Step towards Developing an Experimental Pico-satellite at
for attitude determination and control system module. The Institute of Space Technology,” Journal of Space Technology, vol.
whole CubeSat is placed in the air bearing system and the 1, no. 1, pp. 5–10, 2011.
ADCS is tested along all three axes. The spherical air bearing [7] R. Mahmood, K. Khurshid, and Q. Ul Islam, “Institute of Space
is frictionless, so it can mimic the control of roll, pitch, and Technology CubeSat: ICUBE-1 subsystem analysis and design,”
yaw in the space environment. IEEE Aerospace Conference Proceedings, 2011, doi:
10.1109/AERO.2011.5747229.
Launch: Small satellites are usually launched alongside the [8] “ECSS-E-ST-32-08C – Materials | European Cooperation for
major payload as a piggyback. Furthermore, the primary Space Standardization.” https://ecss.nl/standard/ecss-e-st-32-08c-
payload sets the orbital parameters, and the secondary materials/ (accessed Sep. 10, 2021).
payloads must adapt their missions to fit inside these limits. [9] PC/104 Embedded Consortium, “PC/104 Specification Version
2.6,” 2008. Accessed: Sep. 10, 2021. [Online]. Available:
Due to increased involvement in the CubeSat program from www.pc104.org
international institutions, the CubeSat association has begun [10] J. Bouwmeester, M. Langer, and E. Gill, “Survey on the
to coordinate CubeSat launches. The launch of CubeSats is implementation and reliability of CubeSat electrical bus
also being coordinated by other parties. This lowers total interfaces,” CEAS Space Journal, vol. 9, no. 2, pp. 163–173, 2017,
launch and procedural costs, letting mission planners to focus doi: 10.1007/s12567-016-0138-0.
only on satellite construction. [11] J. Richard. Wertz and W. J. Larson, “The Space Mission Analysis
and Design Process,” in Space Mission Analysis and Design,
Microcosm, 1991, pp. 1–18. doi: 10.1007/978-94-011-3794-2_1.
V. CONCLUSION [12] DNEPR, “Safety Compliance Requirements,” no. August 2004,
The initial design and subsystem analysis of ICUBE-N is pp. 2–4, 2004.
presented in this paper. ICUBE-N is a 3U CubeSat and will
carry scientific payloads. The scientific value intended to be

Seventh International Conference on Aerospace Science & Engineering 95

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Design and Development of a Portable Dish

Antenna for SATCOM Man-Pack Terminal
Shujaat Hussain Awais Munawar Qureshi
College of Aeronautical Engineering, NUST College of Aeronautical Engineering, NUST
Risalpur, Pakistan Risalpur, Pakistan
[email protected] [email protected]

Rizwana Kausar Mohsin Javaid

College of Aeronautical Engineering, NUST Air University
Risalpur, Pakistan Islamabad, Pakistan
[email protected] [email protected]

Abstract— In this paper, design analysis of a portable dish II. LITERATURE REVIEW
antenna for Satellite Communication (SATCOM) application
has been presented. SATCOM is of great significance as it can Parabolic antenna consists of a parabolic reflector which is
provide coverage over entire globe. RF signals in Ku-band are of curved shape design utilized to direct the radio signals. The
transmitted over longer ranges after reflection by a parabolic most popular of parabolic reflector is the dish antenna or
dish antenna. These signals are utilized to send and receive parabolic dish. Reflector antenna can provide extremely high
information from remote areas where other communication gain at microwave frequencies (300 MHz to 300 GHz). In Ku-
technologies may fail. The high-gain and high-efficiency offered band (12 - 18 GHz), it can provide more than 30 dB gain. This
by a parabolic dish antenna make it suitable for SATCOM, high gain is only achievable if the antenna is properly designed
Telecommunication, broadcast and reception applications. and installed at proper angle towards the satellite. The other
However, the procurement of a customized parabolic dish radiation characteristics (polarization, directivity, efficiency)
antenna is expensive for developing countries. This paper of a reflector can be improved by properly designing the
focuses on the development of a low-cost parabolic dish antenna structural configuration The reflector of the parabolic dish
for SATCOM Man-Pack terminal. The design and construction antenna collects the radio waves from satellite and converges
of a parabolic dish antenna used to intercept RF signals from the beams of parallel waves at a spot called the focal point
satellite have been discussed in detail. The proposed design
(foci). The Low Noise Block (LNB) placed at the focal point
approach offers a number of advantages such as low complexity,
transforms the concentrated radio waves into required signal
high gain and cost-effectiveness. Based on our proposed
approach, parabolic dish antenna can be fabricated in a time
level and shape. After conversion, the signals are forwarded to
efficient manner using easily available materials and local the receiver circuitry of SATCOM for postprocessing [2]. The
resources. The simulation results validate the high-efficiency diagram of a portable dish antenna is shown in Fig. 1
and high-gain characteristics of our proposed parabolic dish
antenna design. A link budget analysis (uplink and downlink)
has also been conducted to investigate the atmosphere effects on
RF signals propagation in the air. The attenuation factors
include atmosphere clouds, gases, rain, noise temperature and
beam spreading.

Keywords— Block Up Converter, Intermediate Frequency,

Local Oscillator, Low Noise Block, Portable Dish Antenna,
Radio Frequency, SATCOM

I. INTRODUCTION
Portable dish antenna plays a major role in
telecommunication, television, military command and control, Fig. 1. Portable dish antenna
radio, computer communication and maritime navigation
applications. This research is based on design and construction A. Portable Antenna
of a parabolic dish antenna for RF signals transmission and Parabolic dish antenna has a transducer that is capable of
reception in Ku-band (12 to 18 GHz) with high gain. Portable transmitting and receiving the electromagnetic (EM) waves.
dish antenna can provide a high gain in Ku-band and operate The transmitter side transforms the electric current into EM
in exactly similar fashion as of optical telescope. The parabolic waves while the receiver side converts the EM waves into
dish antenna can transmit and receive the RF signals to or from electrical current. Dish antenna is a type of unidirectional
any desired location in a satellite orbit. This type of antenna antennas as it can efficiently radiate or receive radio waves in
comes in different sizes and shapes, and most commonly use a particular direction. The designing of parabolic dish antenna
to transmit and receive the satellite guided television signal. is a function of its need to the society as compared to other
Its working principle is similar to that of a television antenna types of antennas and it is the best selection when it comes to
with the exception that it is intended to pick up the weak ultra- the reception of satellite signals. Moreover, parabolic dish
high frequency satellite signals. Thereafter, it transforms the antenna is known for its high gain and fast data
signals frequency and passes them through the coaxial cable communication as it can reflect the radio and communication
to the television. One of its main roles is to provide wide range signals very efficiently. EM waves due to their shorter
of television signals to the consumers. [1].

Seventh International Conference on Aerospace Science & Engineering 96

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

wavelength require a reasonable small size reflector that is amplification. The block diagram of LNB is shown in Fig. 3.
more effective at both the transmitter and receiver end. It comprises an input waveguide, followed by the receiver
circuitry. There is a vertical pin at the input waveguide that
Usually, a parabolic dish antenna comprises of a parabolic extracts the electrical current with vertical polarization. The
reflector along with a small antenna. The reflector is made up bandpass filter allows only the required MW frequency range
of metallic material curved into a parabolic shape and then and attenuates all other unnecessary frequencies. The low
rotated into circular shape with proper diameter. The curving noise amplifier amplifies the signal and feeds to the mixer
is in such a way that the light source produces a light beam where it is combined with the local oscillator. Finally, the
parallel and aligned with the curving axis. The small antenna intermediate signal in Ku-band is filtered and amplified before
is placed at the foci of parabolic reflector in order to have transmission through cable [5].
better transmission and reception of radio waves through
coaxial cable or wave guide. The gain of a portable dish
antenna can be improved to a greater extent with proper
designing and precise fabrication [3].
B. Materials for Dish Refelector
The most prominent aspect in achieving the stable gain of
parabolic dish antenna is the selection of proper material. As
the demand for better performance increases for radio waves,
the selection of appropriate materials becomes more vital. A
suitable material gives higher structure stability to the antenna
and provides a better support. In Table I, Aluminium, Steel
and Carbon Fibre have been compared based on the price,
density, Young’s modulus, specific modulus, thermal Fig. 2. Operating principle of portable dish antenna
expansion and heat conduction properties of each material.
Based on the availability and cost effectiveness, the
Aluminium material has been selected for reflector design.

TABLE I. MATERIAL SELECTION

Carbon
Properties Aluminium Steel
Fibre
Price per kg in USD 1.79 1.59 - 1.79 27 - 51
Density 2,710 8,050 1,800
Young’s Modulus 69 200 90.5 - 190
Specific Modulus (Mm2/s2) 26 25 56 - 120
Thermal Expansion (in/F) 13 07 02 Fig. 3. LNB block diagram
Heat Conduction (W/mK) 210 50 5-7
F. Block Up-converter
C. Dish Antenna Design
Block Up-converter (BUC) is utilized for the transmission
The portable dish antenna works in the same way as a of satellite signals (uplink). The main function of BUC is to
reflecting optical telescope. The microwave (MW) signals convert a band of lower frequencies into higher frequencies. It
arriving on parallel paths from distant source, like satellite in is used in combination with LNB. Since the BUC steps up the
our case, are reflected by the reflector to the focal point. As lower frequencies to higher, it is installed at the transmission
the curve is parabolic, all rays are meet at one focal point. side. Whereas LNB steps down the higher frequencies to
Some difficulties are faced in real-life dish antennas design lower, is mounted at the receiver side.
such as the placement of LNB for receiving MW beams and
BUC for transmitting radio signals. The main difficulty is to G. Aperture Gain and Efficiency
find the focus point position such that maximum energy may The aperture gain of dish antenna is calculated using (1).
be collected from the receiving signals, transmitted by the The efficiency for well-designed dish antenna is often of the
satellite, as shown in the Fig. 2 [4]. order 0.6 – 0.7 [4].
D. Antenna Feed
Dish antenna reflector consists of a radiating element, (1)
BUC or LNB, which acts as a feed system for the antenna. It
is placed at the focal point of the parabolic reflecting surface. H. Satellite Segments
The energy from the BUC or LNB is focused in such a way In a satellite communication framework, different tasks are
that dish antenna transmits or receives electromagnetic signals performed. Among these, the principle tasks are circle
with minimum attenuation. controlling, elevation of satellite, and checking and controlling
E. Low Noise Block of different subsystems. It comprises of two main portions,
namely; the Space Segment Subsystem and the Ground
Low Noise Block (LNB) is utilized for the reception of Segment Subsystem.
satellite signals (downlink). It consists of a low noise
amplifier, a frequency mixer, a local oscillator, band pass 1) Space Segment Subsystem: It comprises of following:-
filters and an intermediate frequency amplifier. LNB collects a) Satellite
the signals in form of microwaves from the satellite as b) Transponder
received by the dish. Thereafter, it amplifies the received
signals and converts into intermediate frequency signals and c) Tracking, Telemetry and Control System (TT&C)
forwards to indoor receiver through transmission cable after

Seventh International Conference on Aerospace Science & Engineering 97

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

The satellite is a body that rotates in a specific orbit around

the other body. The communication satellite system acts a
receiver for receiving the radio waves in the space. The criteria
selected for satellite signals transmission and reception is that θ = 24.34◦
the frequency range is kept different for both uplink and the
downlink so as to avoid interference between the transmitted For effective reception of signals in Ku-band, the focal
and received signals. The transponder has two functionalities, length (f) and diameter (D) ratio must lie between 0.65 – 0.75.
the frequency up-conversion of signals at the receiver side and The ‘f’ value has been taken as 0.70 times the optimum value
translation of received signals. of ‘D’. Since the value of D = 90 cm, the ‘f’ value comes out
to be 63 cm. This is the focal point where all incoming satellite
2) Ground Segment Subsystem: The ground station signals are focused to LNB. The depth of antenna is the
provides a communication link to the satellite. The ground deepest point of the dish and has been calculated using (3).
segment comprises two subsystems; the receiving earth station
and the transmitting earth station.
(3)
III. METHODOLOGY
In case of satellite communication system, the most
important factor to be taken into consideration is the satellite
antenna. The better choice of the satellite antenna leads
towards an interference free communication between the
satellite space stations and the earth units, with high efficiency IV. ANTENNA CHARACTERISATION PARAMETERS
and the least possible interference of the noise. Therefore, for
a feasible communication, there must be a well-defined set of Following are the simulation results of important antenna
parameters for antenna design. For efficient antenna design, characterization parameters that have been considered while
various techniques are adopted such as frequency reuse designing and developing a parabolic dish antenna.
configuration, multiple beams amount and the circular A. Radiation Pattern
polarization. The dish antenna model that has been proposed This is the best quantitative description of an antenna that
here is efficient enough that fulfills all necessary requirements defines how much energy it can concentrate in one direction
for a better, noise-free and effective communication between with reference to other directions. The radiation pattern also
the space stations and the ground units. informs about the bore-sight maximum gain in a particular
A. Design of Offset Reflector Antenna direction which in our design is equal to 38.8 dB. The 3D
The designed antenna operates in Ku-band of microwave radiation pattern of designed dish antenna is shown in Fig. 5
region. In this band, the frequency range is 10.9 – 12.75 GHz
for downlink and 13.75 – 14.5 GHz for the uplink. The
commercial offset dishes used for satellite reception at homes
generally have major axis dimension of 90 cm. It is designed
with an offset parabolic reflector antenna of dimensions 90 cm
x 82 cm. This particular design enables to track a number of
satellites with low Equivalent Isotropic Radiated Power
(EIRP) values or footprints. The our design of a reflector dish
antenna with horn antenna at offset point is shown in Fig. 4.

Fig. 5. 3D radiation pattern of designed dish antenna

B. Polar Plot
This plot displays the magnitude and specific response in
a particular direction. In this plot, the radiation pattern is taken
in a particular plane. For dish antenna, it is on 90-degree angle
as shown in Fig. 6. Polar plots provide us the information
Fig. 4. Reflector design with horn antenna about Sidelobe Level (SLL) which indicates how large or
small are side lobes with reference to main lobe in the
B. Calculation for Offset Dish Antenna Design radiation pattern of portable dish antenna. Usually, it is the
Based on the theoretical and practical studies, following difference between the maximum level at a bore sight of dish
calculations have been made. The tilt Angle is defined as the antenna and the maximum value of side lobe, expressed in dB.
arcsine of the ratio of short axis to the long axis (2).

(2)

Seventh International Conference on Aerospace Science & Engineering 98

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

reflection co-efficient of our designed dish antenna is plotted

in Fig. 9. It is highlighted that S11 value of our antenna is less
than -19 dB for both uplink and downlink frequency range.

Fig. 9. Reflection coefficient (S11) of designed dish antenna

Fig. 6. Side lobe level of designed dish antenna
E. Voltage Standing Wave Ratio (VSWR)
SidelobeLevel (SLL) = Max Side lobe Value – Max Gain
SidelobeLevel (SLL) = -1:2069 - 38:7561 = -39:963dB VSWR is the ratio of maximum to the minimum voltage
value in a standing wave pattern along the transmission line.
C. Half Power Beamwidth (HPBW) According to the given requirements, VSWR parameter value
The polar plot can be represented as a 2D plot. The should be less than 1.25 for both uplink and downlink link
radiation pattern along vertical and horizontal axis are shown frequency range. The VSWR of our designed dish antenna is
in Fig. 7 and 8, respectively. The elevation HPBW (-3 dB) plotted in Fig. 10. It is highlighted that VSWR value of our
along elevation angle is given in Fig. 7. The azimuth (-3 dB) antenna is less than 1.25 for both uplink and downlink
HPBW along azimuth angle is given in Fig. 8. According to frequency range.
the required specification, the -3 dB HPBW is expected to be
equal or less than 1.58 degree. In our design, its value is 0.93
degree along elevation and 1.1 degree along azimuth.

Fig. 10. VSWR of designed dish antenna

V. LINK BUDGET ANALYSIS

Fig. 7. Elevation (-3 dB) beamwidth of designed dish antenna A. Link Budget
The link budget is the summation of all gains and losses
HPBW = 90.49 – 89.56 = 0.93 degree from the ground segment to the satellite segment. There are
two types of link budget.
1) Uplink
In uplink of a satellite circuit, the earth station is
transmitting the signal and the satellite is receiving it.
2) Downlink
In downlink of a satellite circuit, the satellite transmits the
signals and the earth station receives it.
B. Equivalent Isotropic Radiated Power
The important factor in the link budget analysis is the
concept of Equivalent Isotropic Radiated Power (EIRP). It is
Fig. 8. Azimuth (-3 dB) beamwidth of designed dish antenna the sum of transmitted power and the gain of antenna (4),
where Pt is the transmitted power in dBW and Gt is the
HPBW = 90.62 – 89.52 = 1.1 degree transmitted antenna gain in dB.
D. Reflection Coefficient EIRP = Pt + Gt (4)
Reflection Coefficient (S11) is the ratio of reflected to C. Transmission Losses
incident wave signal. This parameter explains the reflection of Basically, EIRP is the concept of power input to a system
waves due to mediums impedance mismatch. According to at one end and calculating the power at other end while taking
given requirements, the S11 value should be less than -19 dB into consideration the transmission lines losses as the signal
for both uplink and downlink link frequency range. The proceeds. Fig. 11 shows the PAKSAT-1R satellite EIRP level.

Seventh International Conference on Aerospace Science & Engineering 99

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

increase in temperature of physical surrounding but not always

equal to one.
PN = kTNBN (9)
E. Carrier-To-Noise Ratio
In order to evaluate the performance and efficiency of a
communication link, Carrier to Noise Ratio (CNR) is utilized.
It is defined as the ratio of carrier power to the noise power at
the receiver side (10) and (11). It is the most important
parameter in determining the receiving system performance.

(10)

(11)

Fig. 11. PAKSAT-1R satellite EIRP F. Satellite and Ground Segment Specification
The satellite and ground segment parameters are presented
1) Free-space transmission: The first thing in calculating in Table II.
the losses is the estimation of losses due to the spreading of
signals in free-space. The power delivered to a receiver is TABLE II. SATELLITE AND GROUND SEGMENT SPECIFICATIONS
given as (5), where PR is the received power in dBW and GR Parameters Specifications
is the receiver antenna gain in dB. Satellite PAKSAT-1R
𝐸𝐼𝑅𝑃𝐺𝑅 𝜆2 Reflector Diameter 0.6-0.9 m
𝑃𝑅 = Transmitter/Receiver Frequency 10.9-14.5 GHz
4𝜋𝑟 2
Data Bit Rate 10 Mbps
Symbol Rate (QPSK) 5 MPs
𝑃𝑅 = [𝐸𝐼𝑅𝑃] + [𝐺𝑅 ] − 10𝑙𝑜𝑔 (4𝜋r/𝜆)2 (5)
Bandwidth 6.25 MHz
Transponder Bandwidth 72 MHz
The signal received at the receiver side can be expressed
EIRP of Satellite 53 dB
as the sum of EIRP, the receiver antenna gain and the Free-
Space Loss (FSL) as (6), where FSL component is defined as G/T 5 dB
(7) in dB. G. Data Bit Rate and Symbol Rate
𝑃𝑅 = [𝐸𝐼𝑅𝑃] + [𝐺𝑅 ] − 10𝑙𝑜𝑔 (4𝜋r/𝜆)2 It measures the rate at which number of bits are transmitted
per second. It is also specified by the term baud rate which is
𝑃𝑅 = [𝐸𝐼𝑅𝑃] + [𝐺𝑅 ] − [𝐹𝑆𝐿] (6) the transmission of the number of the symbols per unit time.
1) Quadrature Phase-Shift Keying: This model converts
FSL = 10log (4𝜋𝑟/𝜆)2 (7)
the binary data into two-bit symbol which is further utilized to
Generally, the wavelength is known instead of frequency modulate the carrier signal (12). Against the binary states of
and the frequency can be determined using equation (8), where zeros and ones, there are four possible states for a symbol.
‘c’ represents the speed of light and is constant, frequency (f) QPSK and 8PSK are used for high speed communications.
is in GHz and the distance is in Km. QPSK modulation is selected to compare the required CNR
𝑐 value with the actual CNR.
𝜆= (8)
𝑓

2) Feeder losses: The losses which are suffered between Symbol Rate (Rs) = (Data bitrate)/2 (12)
the antenna and receiver path are known as the feeder losses. = (10)/2 = 5 Msps
Such losses arise from the filters, couplers and the waveguides
that connect the receiver antenna to high power amplifier. The operating bandwidth can be determined from the
3) Antenna misalignment losses: For an ideal case, once symbol rate as (13), where alpha is the roll-off factor which is
the link has been established, the earth and the space station often around the 0.25 value and Rs is 5 Msps. Under these
are aligned in such a way that the gain is maximum. In reality, assumptions, it comes out to be 6.25 MHz. According to Table
III, the minimum value of CNR for QPSK is 6.42 dB, if 9 bits
there arise off-axis losses due to misalignment between the
out of 10 are used. The CNR value must exceed from 6.42 dB
satellite and earth. The off-axis losses at the space station are for availability.
considerable and are taken into account while designing the
antennas and the link operations. However, the earth off-axis Bandwidth = Rs (1 + α) (13)
losses are minimum and are usually of the order of tens of dB.
TABLE III. DVB-RCS-S2 ADAPTIVE MODEM CHARACTERISTICS
D. System Noise
The noise power can be determined from the thermal noise
sources and is presented as (9), where PN is the noise power in
dBW, TN is the noise temperature, BN is the noise bandwidth
and k is the Boltzmann constant. The noise increases with the

Seventh International Conference on Aerospace Science & Engineering 100

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

VI. RAIN ATTENUATION

In ITU Radiocommunication (ITU-R) model, the rain is
related to the input as rain intensity with 0.01% probability
level [6-8]. The link budget analysis for PAKSAT-1R in Ku A. Impact of Rain on Wireless Communication
band has been presented in this paper. Rain attenuation for Ku- Precipitation causes serious distortion to amplitude level
bands has been calculated. Here, the location of earth station of received signals above 10 GHz frequency. This debasement
for Islamabad, Pakistan is (33.7167° N, 73.0667° E). causes the recurrence of radio waves. Therefore, rainfall is one
of the main causes of degradations in received signals. The
degradation increases with the increase in radio frequency. As
the rain increases the attenuations of desired signal increases.
The main factors that contribute to the attenuation are the size
of rain drop and the signal frequency. The larger rain sizes
increase the scattering and absorption of radio signals. In any
kind of communication, either the satellite or terrestrial, the
main component is prediction of rain. The mathematical
approximation of attenuation due to rain is given as (14),
where ‘k’ is the constant (kH or kV) known as the rain
coefficient and α is the rain drop size / shape, and is frequency
dependent. Their values can be calculated using (15) and (16).
ar = k (R0.01) α (14)

Fig. 12. Pakistan map from the satellite

The elevation angle for the dish antenna is 35.9◦, latitude

is 33.68◦ and the Rain Rate for 0.01% of average year, is taken (15)
from the Table IV. As per the “ITU” the whole Pakistan is
classified into a K-rain region, with having almost 148 mm/hr
intensity of the rain. The rain is varied from one city to another
city in Pakistan due to which the different cities is classified (16)
into different zones. Islamabad is labelled as L-zone with rain
amount of 60mm per hour, whereas the cities like Peshawar The suggested values for all the constants are mentioned
and Islamabad is identified as K-region with rain intensity of in Table V to VIII. Values for coefficient kH for horizontal
about 42mm per hour. polarization are given in Table V and values for coefficient kV
for vertical polarization are given in Table VI. Table VII
In order to have a complete availability that is about 99.99 provides the values for coefficient αH for horizontal
percent for the required length, there is only one requirement polarization and Table VIII gives the values for coefficient αV
of the microwave hop. But if the link is designed with wrong for vertical polarization [9-11].
assumptions of the rain, then in such conditions there is need
of more than one hop to achieve the objectives. TABLE V. COEFFICIENT FOR KH

TABLE IV. ITU CLIMATIC ZONES

TABLE VI. COEFFICIENT FOR KV

Seventh International Conference on Aerospace Science & Engineering 101

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

noise temperature and Tc is the composite noise temperature

of the receiver equal to 10 K. Ta is calculated using (20), where
Tm is the medium temperature and according to ITU its value
ranges from 270 K to 290 K. By putting the values of Tm, Tc
and attenuation, we found the value of Ta equal to 66.844 K.
TS = Ta + Tc (19)
TABLE VII. COEFFICIENT FOR αH
Ta = Tm (1 − 10) −A/10 + Tc −A/10 (20)
We have performed the calculations for series
NJR2835U/36U/37U/39U of LNB. The Noise Factor (NF)
value of this series ranges from 0.8 to 1 dB. For our
convenience, we have taken the mean value of NF = 0.9 dB.
Using Table X, we have found the corresponding temperature
value of LNB at NF = 0.9 dB as 66.78 K [11]. According to
TABLE VIII. COEFFICIENT FOR αV ITU-R, the temperature of clear sky is 30 K and total losses
temperature is 35.39 K. Therefore, the antenna clear sky
temperature is the sum of clear sky temperature and the LNB
temperature (21).
T = TCS + TLNB (21)
T = (30 + 35.39) + 66.78 = 132.17 K
TdB = 21.211dB
Using coefficient values provided in Table V to VIII, Eq
(15) and (16) gives the value of k =0.622 and α =0.2937. After TABLE X. NOISE FACTOR AND TEMPERATURE
putting these values in (14), the value of specific attenuation
is calculated as 0.014391dB/km. Thereafter, rain attenuation
using Slant Range = 40.15 km and Elevation Angle = 35.9°
has been calculated as 0.9855dB (18).
At = ar (LS); Ls = Slant Range/(sin(θ))
At = 0.014391(40.15/sin (35.9))
At = 0.9855dB (18)
VII. KU-BAND LINK BUDGET ANALYSIS FOR SATCOM
A. Uplink Budget Analysis
The uplink budget analyses have been presented in Table
IX. It provides different parameters values for both the clear
sky and rain conditions. During rain, the value of temperature increases and we
have to add the clear sky antenna temperature to the rain
TABLE IX. UPLINK BUDGET ANALYSIS FOR SATCOM IN KU-BAND temperature (22).
Parameters Clear Sky Rain Units TS = TCS + TRain (22)
Earth Station EIRP 74.8 74.8 dB
Satellite G/T 05 05 dB/K TS = 132.17 + 66.844 = 199.014K
Path Loss 205.51 205.51 dB TSdB = 22.986 dB
Rain Attenuation 00 0.9855 dB
Cloud Attenuation 00 01 dB Since, the temperature of antenna for clear sky and during
Gaseous Attenuation 01 01 dB rain is different, it will affect the value of G/T as following.
Transmitter Waveguide Loss 0.4 0.4 dB
(G/T) clear sky = 38.8 − 21.211
Atmospheric Loss 02 02 dB
Polarization Loss 0.1 0.1 dB (G/T) clear sky = 17.588 dB
Additional Losses 0.6 0.6 dB
Ionospheric losses 0.7 0.7 dB
Moreover, the attenuation of radio signals also increases
Pointing loss 0.5 0.5 dB
with rain. This results in temperature increase for scattering of
communication signals. Subsequently, this affects the G/T
Feed Loss 0.5 0.5 dB
factor too.
Total Losses (Lt) 211.81 213.795 dB
Boltzmann Constant -228.6 -228.6 dB (G/T) Rain = 38.8 − 22.988
Noise Bandwidth 67.95 67.95 dB
(G/T) Rain = 15.812dB
CNR 28.64 26.6515 dB

B. Downlink Budget Analysis The downlink budget analyses have been presented in
Table XI. It provides different parameters values for both the
The downlink budget analysis is directly dependent upon clear sky and rain conditions.
the (G/T) value of receiver. The total temperature of ground
segment can be calculated from (19), where Ta is the antenna TABLE XI. DOWNLINK BUDGET ANALYSIS FOR SATCOM IN KU-BAND

Seventh International Conference on Aerospace Science & Engineering 102

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
[7] X. Yang and Z. Lidong, “Research on rain attenuation for ka-band
Parameters Clear Sky Rain Units satellite communication system,” in 2008 ISECS International
Satellite EIRP 53 53 dBw Colloquium on Computing, Communication, Control, and
Management, vol. 2. IEEE, 2008, pp. 412–415.
Ground G/T 17.588 15.812 dB/K
[8] J. S. Ojo, M. O. Ajewole, and S. K. Sarkar, “Rain rate and rain
Path Loss 204.87 204.87 dB attenuation prediction for satellite communication in ku and ka bands
Rain Attenuation 00 0.9855 dB over nigeria.” Progress in Electromagnetics Research, vol. 5, pp. 207–
Cloud Attenuation 00 01 dB 223, 2008.
Gaseous Attenuation 01 01 dB [9] P. Series, “Propagation data and prediction methods required for the
Transmitter Waveguide Loss 0.4 0.4 dB design of earth-space telecommunication systems,” Recommendation
ITU-R, pp. 618–12, 2015.
“Atmospheric Loss” 02 02 dB
[10] Z.-w. Zhao, M.-g. Zhang, and Z.-s. Wu, “Analytic specific attenuation
Polarization Loss 0.1 0.1 dB
model for rain for use in prediction methods,” International Journal of
Additional Losses 0.6 0.6 dB Infrared and millimeter waves, vol. 22, no. 1, pp. 113–120, 2001.
Ionospheric losses 0.7 0.7 dB [11] M. Zubair, Z. Haider, S. A. Khan, and J. Nasir, “Atmospheric
Pointing loss 0.5 0.5 dB influences on satellite communications,” Przeglad Electrotechnics, vol.
“Feed Loss” 0.5 0.5 dB 87, no. 5, pp. 261–264, 2011.
Total Losses (Lt) 210.67 212.655 dB
Boltzmann Constant -228.6 -228.6 dB
Noise Bandwidth 67.95 67.95 dB
Ground Stn. C/No(10Mbps) 20.568 16.8065 dB
Required C/No (QPSK) 6.42 6.42 dB
Link Margin 14.048 10.3865 dB

VIII. CONCLUSION
In this paper, we have presented the design analysis of a
portable dish antenna for SATCOM man-pack terminal
application. The designed parabolic dish antenna is composed
of aluminum material with 90 x 82 cm size. It can be
efficiently utilized for satellite signals transmission / reception
in Ku-band. To achieve high gain, offset dish reflector type
has been selected that offered more aperture efficiency than
other types of portable dish reflector. A horn antenna for feed
system of portable dish reflector has also been designed. A
maximum gain of 38.8 dB in a particular direction towards the
satellite has been successfully achieved. Space and ground
segments of SATCOM have also been studied in-depth that
helped in performing an accurate link budget analysis. The
effect of various atmosphere conditions on RF signals
propagation in the air has also been explained. For this
purpose, ITU-R models have been efficiently utilized. The
attenuation factor due to rain during uplink and downlink
communication has been calculated to design an all-weather
dish antenna. The CNR value validates that proposed design
of portable dish antenna is feasible for SATCOM applications.
It can be easily and time-efficiently constructed using
available materials and local resources.
REFERENCES
[1] C. Onah, K. Arinzechukwu, and F. Elekwachi, “Designing,
constructing and testing a 90 cm parabolic satellite dish using fiberglass
material,” Research & Reviews: Journal of Space Science &
Technology, vol. 2, no. 2, pp. 12–26, 2019.
[2] G. R. Adewale, “Design and construction of a cost-effective parabolic
satellite dish using available local materials,” International Journal of
Advanced Materials Research, 2019.
[3] B. J. Feder, “” taylor howard, 70, pioneer in satellite tv for the home”.
new york time. retrieved 19 july 2014,” 15 November 2002.
[4] H. A. Syed Muhammad J. S Ammar N., “A survey paper on security
issues in satellite communication network infrastructure,” October-
November 2014.
[5] Mahmudul, HM and Sirajum, M. and Syed, N.S. and A.K.M.F.H., and
Haque, Shahina and Tasnim, I.R., “Design and development of an lnb
based dvb-s receiver,” International Journal of Engineering and
Technology (UAE), 2018.
[6] I. Abubakar, J. Din, and H. Y. Lam, “Disparities in the induced rain
attenuation between beacon (narrowband) and broadband satellite links
in tropical zones,” Indonesian Journal of Electrical Engineering and
Computer Science, vol. 10, no. 2, pp. 687–695, 2018.

Seventh International Conference on Aerospace Science & Engineering 103

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Design and Simulation of Cost-efficient Reaction

Wheel for 1U CubeSat
Pawan Kumar Oscar Lionel
Department of Space Science GNSS & Space Education Research Lab
Institute of Space Technology National Center of GIS & Space Applications
Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected]

Abstract— Rapid progress in CubeSat technology has made wheels are used in configuration of three to provide
them feasible to be used for different scientific research stabilization in three directions whilst 4-wheels configuration
purposes but the cost associated with different subsystem used are also used for redundancy and more torque generation in
in the CubeSat are relatively high which limits student each axis. In this research, 3-wheel orthogonal configuration
researcher to limit their scope. The attitude control mechanism is employed for attitude control. The cost-feasibility of the
is an important aspect of a satellite, and for this purpose, this reaction wheel is done by comparing totals cost incurred with
research presents the cost-effective development of a reaction the commercially available reaction wheels and also by
wheel alongside with three-reaction wheels actuation model for performance analysis of the developed hardware.
1U CubeSat. In the presented setup, the reaction wheel is
developed using a commercial-off-the-shelf (CoTS) component, The primarily objective of this research is to introduce a
based upon hard drive disk (HDD). HDD based brushless DC COTS component for the development of reaction wheel and
motor and flywheel are utilized for this research. The simulation study its feasibility to be used as an actuator in CubeSat
of the developed reaction wheel for 1U CubeSat is carried out in missions. For this, a simulation modelled is developed with
the Simulink environment using a three-wheel setup and PID the parameters of the used COTS component and this model
based controller with space environment disturbance torques is subjected to various space-environment disturbances to
values as initial parameters. The output of the simulation analyze the reaction wheel response. The mathematical model
presented the working and response of the purposed reaction
of reaction wheel and attitude control system are defined in
wheel against the disturbance torques. The HDD based reaction
MATLAB.
wheel with one flywheel has potential to produce maximum
torque of 37.6 𝑚𝑁.𝑚.s at the angular momentum of 6200 rpm II. SATELLITE ATTITUDE DYNAMICS AND
and has power consumption ratings from 0.4 watts to 0.7 watts
ACTUATION MODELLING
and an operating temperature of -40o C to 75o C. The final
design output confirms the use of HDD based components for A. Satellite Attitude Dynamics
reaction wheel development. Moreover, the overall cost
associated with hardware is comparatively less than the
In space environment, satellite is subjected to different
commercially available systems. torques, which can be divided into two groups of internal and
external torques. Internal torques are produced due to different
Keywords— Reaction wheel; Attitude Control; Disturbance payload and system inside a satellite whereas external
Torques; Brushless DC Motor; 1U CubeSat, Simulink Model disturbances and the control torque provided by actuators
produce external torques. In this research, internal torques are
I. INTRODUCTION out of scope and will be assumed zero and equations
1U CubeSat comes in the category of miniaturized satellite governing the attitude dynamics will only include external
with dimensions of 10 cm × 10 cm × 10 cm and with mass of torques.
1.33 kg and are often developed for research purposes using Let MT be the total external torques affecting the satellite
COTS components[1]. Attitude system of satellite refers to the and it comprises of sum of disturbance torques MD and sum of
satellite orientation and position with respect to an object, command torques MC. Thus, total torque on satellite body can
mostly Earth and the control of this attitude has fundamental be given as:
importance for payloads, which require orientation in correct 𝑀𝑇 = 𝑀𝐷 + 𝑀𝐶 (1)
direction[2]. Rotation of satellite around Earth subjects it to
various kind of disturbance torques, which results in This research focuses on the satellite at the altitude of 600
disorientation of satellite, and to counter the disturbances
km, which subjects satellite to disturbance torques due to
torques, active attitude control systems are employed[3]
gravity gradient, aerodynamic torque, Earth’s magnetic field
Advancement in CubeSat technology and its low cost has
extended the extent of purposes for which it is being used and Solar pressure. The dynamic modeling is based upon
where many of the missions require high pinpoint accuracy. Newton-Euler formulation where angular momentum deviates
To provide attitude control of satellite, different mechanisms relating to the applied torque and is given as:
are utilized which are divided into active and passive attitude 𝑑𝐻⃗𝐼
control mechanisms. 𝑀𝑇 = (2)
𝑑𝑡
Reaction wheels are used in active attitude control system
of a satellite to provide pinpointing accuracy of orientation the In this research, reaction wheels are the only source for
satellite. For two decades, this method of ACS is employed in commanded torque, thus angular momentum HB of satellite
different CubeSat’s design specifically where the pinpoint will become:
accuracy is mission requirement[4]. To achieve 3-axis 𝐻𝐵 = 𝐼𝑠 𝑤𝑏 + ℎ𝑤 (3)
stabilization for a satellite (pitch, roll and yaw), reactions

Seventh International Conference on Aerospace Science & Engineering 104

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Here hw denotes the angular momentum produced by

reaction wheels which is defined by:
ℎ𝑤 = 𝐶𝑟𝑤 𝐼𝑤 𝑤𝑤 (4)

Here Iw denotes reaction wheel diagonal inertia matrix

where ww represents reaction wheel angular velocity and C
represents the distribution matrix associated with the reaction
wheel distribution in satellite body. Figure 1 Orthogonal arrangement of
By using Newton’s Euler method, and taking equation of reaction wheels
moment Hb, we will get: 1 0 0
𝑑𝐻𝐵 𝐶𝐵𝑅𝑤 = [𝐶 1 𝐶 2 𝐶 3 ]𝑇 = [0 1 0] (11)
= −𝑤𝑏 (𝑡) × 𝐻𝐵 (𝑡) + 𝑀𝑇 (5)
𝑑𝑡 0 0 1
The use of distribution matrix provides solution to the
By using equation of HB and hw from (3) and (4): problem associated with control allocation between reaction
wheel and Euler angles. When satellite is affected by external
𝑑𝐻𝐵 𝑑𝑤𝑏 torque, then this torque command is distributed to all reaction
= [𝐼 ] + ℎ𝑤̇ (6)
𝑑𝑡 𝑑𝑡 𝑏 wheels respective to their axis. This is done by using inverse
of distribution matrix as shown by below equation[3].
𝑏𝑜𝑑𝑦 −1
𝐵
By comparing (5) and (6), we will get: 𝐻𝑟𝑤 = [𝐶𝑟𝑤 ] 𝐻𝑟𝑤 (12)
𝑑𝑤𝑏
[𝐼 ] = −𝑤𝑏 (𝑡) × 𝐻𝐵 (𝑡) + 𝑀𝑇 − ℎ𝑤̇ (7) The overall angular momentum (Hrw) produced by
𝑑𝑡 𝑏 actuation system equals to the sum of individual angular
momentum provided by each reaction wheel and can be
Now, as reaction wheel are functioned to provide opposite written as,
torque to oppose disturbing torque and by using principle of
exchange of momentum then commanded torque will became: 𝑤𝑟𝑤 = [𝑤𝑟𝑤1 𝑤𝑟𝑤2 𝑤𝑟𝑤3 ]𝑇 (13)
Thus, by using principle of momentum exchange and by
ℎ̇𝑤 = − 𝑀𝑐 (8) use of (1) to (13), total Mc in each axis comes out to be:
𝑑𝑤𝑏
= 𝐼 −1 (−𝑤𝑏 (𝑡) × 𝐻𝐵 (𝑡) + 𝑀𝑇 + 𝑇𝑐 ) (9) 𝐻̇𝑅𝑊𝑥
𝐵 ⃗ 𝑅𝑊𝑧
+𝐻 𝐵 𝐵
𝑤𝐼𝐵𝑦 ⃗ 𝑅𝑊𝑦
−𝐻 𝐵 𝐵
𝑤𝐼𝐵𝑧
𝑑𝑡 𝑀𝐶𝑟𝑤 = 𝐻̇𝑅𝑊
𝐵 𝐵
+ (𝑤𝐼𝐵 𝐵 )
+ 𝐻𝑅𝑊 = [ 𝐻̇𝑅𝑊𝑦
𝐵 ⃗ 𝑅𝑊𝑥
+𝐻 𝐵 𝐵
𝑤𝐼𝐵𝑧 ⃗ 𝑅𝑊𝑧
−𝐻 𝐵 𝐵 ] (14)
𝑤𝐼𝐵𝑥
̇ 𝐵 ⃗ 𝐵 𝐵 ⃗ 𝐵 𝐵
𝐻𝑅𝑊𝑧 + 𝐻𝑅𝑊𝑦 𝑤𝐼𝐵𝑥 − 𝐻𝑅𝑊𝑥 𝑤𝐼𝐵𝑦
Equation (9) gives time derivate of angular velocity in
body frame whereas commanded torque Tc is also in body
frame.
C. Reaction Wheel Electrical and Mechanical Model
In this research, the reaction wheel hardware and software
B. Attitude Control Model
implementation utilize BLDC motor which speed is
In this research, three reaction wheels control each axis of controlled electronically by commutation. This commutation
satellite. Whenever satellite is subjected to disturbance torque becomes source of ripple in torque command, which creates
in any axis, the aligned reaction wheel of that axis provides instability, thus use of speed and torque controller are
opposition to overcome that disturbance torque and continues implemented in the model.
to do so until it reaches saturation[6].
To control reaction wheel and to deal with instability,
In the case of satellites, four-wheel configuration is used speed and torque control are implemented in the model. The
for redundancy but due to mass and size constraint of CubeSat,
this model will only employ 3-wheel reaction model. The idea
behind choosing the 3-wheel system is due to constraint of
CubeSat and relative size and mass of used reaction wheel. In
orthogonal formation of reaction wheels, required distribution
matrix is given as[7];
𝑟𝑥1 𝑟𝑥2 𝑟𝑥3
𝐶𝐵𝑅𝑤 = [𝐶 𝐶 𝐶 ] = [𝑟𝑦1
1 2 3 𝑇 𝑟𝑦2 𝑟𝑦3 ] (10)
𝑟𝑥1 𝑟𝑧2 𝑟𝑧3
Above matrix gives matrix for three-reaction wheel in \
three axis. Number of columns represent the total number of Figure 2 Speed and Torque PID control model
reaction wheel and number of rows represent the axis. In the
case of orthogonal configuration, 𝐶𝐵𝑅𝑤 becomes[8],
commanded torque is achieved by providing current to the
motor hence, torque can be controlled by controlling current
supplied to electronic which is done through torque control.
In contrast to torque control, speed control is in cascade loop
which controls the voltage supplied to the inner loop. This is
necessary, because without cascade control more ripple
torque is generated which makes system unstable[9].

Seventh International Conference on Aerospace Science & Engineering 105

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Earth’s z-axis, and Ix and Iy are satellite’s moment of inertia

A simplified model for PID controller is implemented in along x and y axis respectively.
above model. It can be seen that speed control is in cascade In this research, maximum torque is required which will be
loop whereas toque controller is inside the loop. The above in the case when sin(2θ) = 1. Hence, torque is calculated by
model is based upon model described by Andresen in[7]. This putting the parameters in the equation (18)[1].
model is linearized and commutation effect is not
represented. The bldc motor of reaction wheel has been B. Magnetic Torque
modelled in under given equation in which armature voltage Earth’s magnetic field produce the torque on the outer
v is related to armature current I and BLDC motor’s angular space bodies, which also includes the different satellites
velocity w. orbiting around Earth. This field give rise to the disturbance
𝑑𝑖 torque acting on body. It decreases as we move farther from
𝑣 = 𝑅. 𝑖 + 𝐿 + 𝐾𝑠 . 𝑤 (15) the Earth nonetheless, it produce noticeable torque in LEO
𝑑𝑡
satellites, which can be stated as[4]:
In above equation, R represents the stator’s resistance, L ⃗ (19)
𝑀𝑚 = 𝑚
⃗⃗ × 𝐵
represents stator’s inductance, and Ksw represents
electromotive force when conductor is motioned in magnetic Where m is satellite residual dipole that in our case can be
field. The torque Mrw is represented by following equation[6], estimated from reference case and which comes out to be
which is directly proportional to the torque produced, by ~0.01 A.m2 [1], [12] and B is the Earth’s magnetic field, which
motor Kt.i and friction b.w. can be stated as[13]:
𝑑𝑤 2𝑀
𝑀𝑟𝑤 = 𝑗𝑟𝑤 = 𝑘𝑡. 𝑖 + 𝑏. 𝑤 (16) ⃗ =
𝐵 = 4.7016 × 10−5 𝑇𝑒𝑠𝑙𝑎 (20)
𝑑𝑡 𝑅2
This equation is used to model reaction wheel in Simulink.
Whereas M is magnetic moment of Earth (M = 7.96 1015
III. DISTURBANCE TORQUES. Tesla.m3) and R is orbit radius.
Satellites orbiting and revolving around the earth become C. Aerodynamic Torque
subjected to different torques, which alter the attitude along Satellites in LEO orbit is subjected to aerodynamic drag
the course of satellite time in orbit. These disturbance torques due to their low altitude. This drag results in producing
are excellent for modeling attitude control system as it is disturbance torque on satellite body. The aerodynamic torque
foremost functionality of reaction wheels to overcome these depends upon the difference of satellite’s center of pressure
disturbances[3]. and center of mass. The formulation of aerodynamic drag
In this research, the selected alattitude lies in LEO orbit force on satellite is given as[1], [2]:
and in this orbit, satellite is generally affected by four kind of 1 2
torques, this torques are[2], [10]; 𝐹𝑎𝑒𝑟𝑜 = 𝜌𝑉 𝐶𝑑 𝐴𝑖𝑛𝑐𝑒 (21)
2
1. Gravity gradient (Mg) Whereas ρ represents atmospheric density at altitude, V is
2. Aerodynamic torque (Ma) orbital velocity of the satellite, Cd is co-efficient of drag and
for CubeSat it values lies between 2 and 2.5 and Aince is
3. Solar radiation (Ms) satellite surface area subjected to drag.
4. Earth’s magnetic field (Me) Putting ρ = 3.725 10-12 kg/m3[12], V = 7558.5 m/s for
For modeling the system, worst-case scenario will be altitude of 600 km and Aince = 0.1 m2 in (21), the Fg comes out
assumed which will be maximum disturbance torque in all to be:
axis, by summing up all the torques and applying them to each 𝐹𝑎𝑒𝑟𝑜 = 2.660 × 10−6 𝑁 (22)
axis.
For aerodynamic torque,
𝑀𝑑 = 𝑀𝑔 + 𝑀𝑎 + 𝑀𝑠 + 𝑀𝑒 (17)
𝑀𝑎 = 𝑅𝑐𝑝 × 𝐹𝑎𝑒𝑟𝑜 (23)
Where Rcp is the difference of satellite’s center of pressure
A. Gravity Gradient Torque (cp) from center of mass (cp) and this comes to 0.05 m for
Gravitational force is a long-distance force and satellites in maximum torque[12].
Earth’s orbit are susceptible to planetary gravitational force,
which decreases by 1/r2 factor as we move away from center D. Solar Radiation Torque
of planet. The attraction due to this force will be more on In outer space, due to continuous activity in solar region
satellite’s lower side as compared with its upper side thus it high energetic particles strike with outer space bodies, which
will produce disturbance torque affecting satellite’s attitude. in result give rise to disturbance torque alongside with
The mathematical expression of gravity gradient for multiple other factors. In contrast to gravity gradient, solar
maximum torque can be stated as[6]: radiation affect maximizes as we move away from center of
Earth. The amount of power per meter2 emitted by Sun is
3μ⊕ known as solar flux density and is taken as constant. Solar
𝑀𝑔 = |𝐼 − 𝐼𝑦 | sin(2𝜃) (18)
2𝑅3 𝑧 radiation force is given as:
In this equation, μ⊕ is the gravitational parameter of Earth 𝐹𝑠𝑟 = −𝑝𝑠𝑟 𝑐𝑟 𝐴𝑠 cos(𝑖) (24)
and is multiple of Earth’s mass and gravitational constant
(3.986*1014 m3/s2), R denotes orbit radius which in our case is Where psr is sun pressure force applied per unit area which
Re + 600 km, θ denotes maximum deviation of satellite’s from can be calculated by dividing solar flux with speed of light (c)
which is given as[14]:

Seventh International Conference on Aerospace Science & Engineering 106

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
𝑆𝐹 𝐻𝑚𝑖𝑛
𝑝𝑠𝑟 = (25) 𝐼𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 = 𝜋 (28)
𝑐 𝑤𝑚𝑎𝑥 × 𝑥𝐼𝑚𝑜𝑡𝑜𝑟
30
The torque due to this force can be calculated by[15]: This equation implies that total moment of inertia of disk
𝑀𝑠 = 𝐹𝑠𝑟 × (𝑐𝑝𝑆𝑅 − 𝑐𝑔 ) (26) should be,
𝐼𝑑𝑖𝑠𝑘 = 𝐼𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 − 𝐼𝑚𝑜𝑡𝑜𝑟 > 0 (29)
cpSR in above equation is the point of where solar pressure
acts and for maximum torque it should be located at farthest Considering the requirements, in this research flywheel
point from the satellite. Its value for CubeSat comes out to be selected are the aluminum disks of HDD mounted on spindle
0.05[14]. motor. The disks are made up of aluminum or ceramics, and
The total disturbance torque for each axis is calculated is comes up with the coating of magnetic material and are
shown in table 1. perfectly balanced thus, negating the horizontal and vertical
motion of disk. The flywheel used in this research has inner
TABLE 1 DISTURBANCE TORQUES
radius of 25 mm with the outer radius of 95 mm and has
Symbol Disturbances Torque Magnitude in Nm thickness of 1 mm. The mass of one disk is 22.9 g. Thus, the
Mg Gravity gradient 7.8171 × 10−8 𝑁𝑚
disk moment of inertia is given by (30) and dimensions can
Mm Magnetic torque 4.706 × 10−7 𝑁𝑚
Ma Aerodynamic torque 1.3301 × 10−7 𝑁𝑚 be visualized in fig 3[20];
Ms Solar radiation torque 3.645 × 10−9 𝑁𝑚 𝑁𝑑 𝑚𝑑 2
𝐼𝑧𝑧 = (𝑟2 + 𝑟12 ) (30)
Mmax Total Torque 6.850 × 10−7 𝑁𝑚 2
In (30), Nd is number of disks, md is mass of disk, and r1
and r2 are disk I inner and outer radius respectively. Putting
IV. SELECTION AND COMPUTATION OF HARDWARE up the parameters, where Nd =2, md =22.9 g, r1=12.5 mm and
A. Electric Motor Selection r2=47.5 mm. the moment of inertia comes out to be:
𝐼𝑧𝑧 = 5.5 × 10−5 𝑘𝑔. 𝑚2 (31)
The ideal motor for designing reaction wheel should meet
satellite’s power and size requirement while also generating The inertia of motor is approximated through literature
enough torque to provide command over attitude. Given the review[17] and is given as:
requirements, the BLDC motor is selected over simple DC 𝐼𝑚 = 3 × 10−6 𝑘𝑔. 𝑚2 (32)
motor. The selection of bldc motor can be justified as due to
no brushes, they provide more efficiency and generate more In addition, total moment of inertia is:
torque on the consumption of small power[16]. Even though 𝐼𝑇 = 𝐼𝑧𝑧 + 𝐼𝑚 = 5.8 × 10−5 𝑘𝑔. 𝑚2 (33)
the complexity increases while controlling bldc motor as it
does not follow the same commutation principle as DC motors The nominal angular momentum of hdd reaction wheel at
but BLDC motor enable to precisely control the torque nominal rpm of 3100 is given by:
generated and are more durable than DC motor. Therefore, our 2𝜋. 𝑟𝑝𝑚
primarily consideration for hardware is BLDC motor. 𝐻𝑟𝑤 = 𝐼𝑇 . = 18.8 𝑚𝑁. 𝑚. 𝑠 (34)
60
Taking into consideration the research scope of this paper,
reaction wheel is developed using hard disk drive (hdd) motor In addition, the maximum angular momentum at 6200 rpm is
of personal computer. HDD spindle motor and its disks will given by using (34) is:
be utilized for purpose of developing reaction wheel. In HDD, 𝐻𝑚𝑎𝑥 𝑟𝑤 = 37.6 𝑚𝑁. 𝑚. 𝑠 (35)
the BLDC motor is used for spindle motor and uses ball
bearings. Due to special characteristics of hard disk drives, the
motors are built in such a way that they have very low leakage
of magnetic field to have no impact on data reading/writing
process. Furthermore, they are built to run for a long time and
provide high rotations rate on the consumption of less power.
Additionally, from literature review it was noted that working
temperature of hard drive spindle motor lies between -40o C to
75o C. This makes them perfect to be used in reaction wheel
assembly, as they will be subjected to harsh space conditions.
The moment of inertia corresponding to hdd spindle motor is
[17], [18]. Figure 3 HDD Dimensions
B. Flywheel Selection and Theoertical Calculation
Flywheel are important part of reaction wheel assembly.
They are selected to meet the torque generation requirement
of a satellite and in our case, for CubeSat. size and material
for flywheel is selected by considering constraints, which are
size, mass and required storage of angular momentum. In this
research[19], the needed moment of inertia has been derived
by using needed angular momentum change, this can be
expressed as:
𝜋
𝐻𝑠𝑡𝑜𝑟𝑒 = 𝑤𝑚𝑎𝑥 × 𝑥𝐼 (27)
30 𝑚𝑜𝑡𝑜𝑟
In addition, total needed moment of inertia is given as,

Seventh International Conference on Aerospace Science & Engineering 107

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Furthermore, the total torque that can be generated from

reaction wheel depends upon rate of change of angular
TABLE-2 HDD CALCULATED PARAMETERS
momentum, and depends upon the time taken for reaction
wheel to reach nominal angular momentum from rest Parameter Symbol Value Unit
position. Figure 4 and 5 shows the selected motor with Inductance Lo 0.48 H
flywheel respectively. Resistance R 7 Ω
Voltage V 3.3 - 5 volts
Motor Field Ke 0.922*10-3((60)/2π)) V/(rad/s)
constant
Motor Torque Kt 8.8*10 ((60)/2π))
-3
N.m/amp
constant
Damping b 8.85 * 10-3 Nm/(rad/s)
Coefficient
mass m 0.12 Kg
RW Radius r 0.047 m
RW height h 0.012 m
Moment of inertia IT 5.8 * 10-5 kg m2
Nominal Hrw 18.8 * 10-3 Nm.s
momentum
Figure 4 HDD BLDC Motor Max. momentum Hmaxrw 37.6 * 10-3 Nm.s
Power consumption P 0.4 – 0.7 W

C. Control Method for Reaction Wheel

The most common method of controlling BLDC motor is
Pulse Width Modulation (PWM) signal. In PWM, current
signals are provided to the stator to generate commanded
torque. The electronic circuit consists of electronic speed
controller; Arduino Uno based micro controller and power
source. The required angular momentum signal is given
through laptop using Arduino language. The schematics of
circuit are given in figure 6 and the hardware implementation
of circuit is shown by figure 7.
Figure 5 HDD Flywheel

The calculated and approximated[17], [20] parameters for

hard drive spindle motors alongside with its electrical
parameter are summarized in following table 2.

Figure 6 ESC speed control circuit for BLDC motor

Figure 7 RW Testing System

V. RESULTS AND DISCUSSIONS

This simulation modelling is done for 3-axis stabilization
using three reaction wheels aligning with the CubeSat’s
principle axis. In figure 8, a complete model of actuation
model is presented. In this model, input command is 1 X 3

Seventh International Conference on Aerospace Science & Engineering 108

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

matrix of the maximum disturbance torque in all axis, due to TABLE 3 HDD HARDWARE PARAMETERS
space environment and the output of the model is the response HDD Reaction wheel Parameters
angular momentum and change in angular momentum
Parameter Symbol Value Unit
produced by reaction wheels in their aligned axis. Moreover,
figure 9 shows the algorithm used in the modelling of the Inductance Lo 0.48 H
system. Resistance R 7 Ω
Motor Field Ke 0.922 * 10- V/(rad/s)
constant 3((60)/2π))
Motor Torque Kt 8.8 * 10- N.m/amp
constant 3((60)/2π))
Damping b 8.85 * 10-3 Nm/(rad/s)
Coefficient
mass m 0.12 Kg
RW Radius r 0.047 m
RW height h 0.012 m
Moment of IT 5.8 * 10-5 kg m2
inertia
Nominal Hrw 18.8 * 10-3 Nm.s
momentum
Max. momentum Hmaxrw 37.6 * 10-3 Nm.s
Power P 0.4 – 0.7 W
consumption
Sampling time S(t) 0.1 second

TABLE 4 CUBESAT ENVIRONMENT PARAMETERS

Parameter Value
CubeSat Type 1U
CubeSat Mass 1.33 kg
Surface Area 0.01 m2
C.G. Offset <0.02, 0.02, 0.02> m
Inertia (Ixx,Iyy,Izz) <0.01, 0.01, 0.01> A.m2
Figure 8 Actuation Control Model - 3-axis Stabilization Dimensions 2.22 10-3 kg
Orbit Altitude 600 km
Earth Angular velocity 7.291 10-5 rad/sec
Drag coefficient 2.2

Conversion from
Command torque to
Input torque body frame to
reaction wheel
comman reaction wheel
system
reference frame
In figure 10, it is visible that reaction wheel has generated
the same torque as commanded torque in all three axes but in
Output of response
DC motor conversion opposite direction. This evidently proves that this model of
RW reference fame of command torque
torque, and change
angualr
to body reference in response change reaction wheel is successfully negating the disturbance
frame in angular
momemnrum
momentum torques. Furthermore, total response time taken to negate
disturbances is 10 seconds. The values are steeper during
startup of reaction wheel which shows that maximum change
Figure 9 Block diagram of Simulink model implementation
in torque occurred during initial time and this change become
The initial parameters for the simulation model are gradual with passage of time.
calculated from the HDD based reaction wheel alongside the Evidently, the response torque provided by reaction wheel
conditions used for model are defined by the conditions faced will be continuous and last until the reaction wheel reach the
by 1U CubeSat satellite at LEO orbit. The tables for both HDD saturation. For three reaction wheels placed in orthogonal
reaction wheel and 1U CubeSat satellite are shown in the table structure, the problem of saturation is met by using magneto
3 and 4 respectively, whereas the initial torque offset is the torquers rod which reset the reaction wheels when they reach
disturbances torques that are summarized in table 1. saturation.
In figure 11, angular momentum response of reaction
wheel is represented for 3-axis stabilization and response time
of angular momentum is shown to the command torque. It is
evident from the figure that the total angular momentum of
system is in reverse direction of command torque. The
linearity of time with angular momentum shows that the
controller used for these simulations are working efficiently
with minimum settling time.

Seventh International Conference on Aerospace Science & Engineering 109

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 12 shows the angular velocity change. In figure, the

linearity of time with angular velocity can be seen. This
linearity is due to less settling time of PID cascade speed
controller. Angular velocity also obeys the same change
principle as angular momentum and torque as it also showed
response in negative direction to the commanded torque. The
reaction wheel response to disturbance torque is linear for the
most part of simulation.
The response of model is susceptible to the input provided
and model of DC model implemented. The smooth curve of
graph represents that the plant function used in the model are
well tuned according to the parameters of HDD based RW
motor model.

Figure 12 Angular velocity response to command torque

A. Total Budget
The cost effectiveness study involves comparing the
developed product with the commercially available set of
reaction wheels. The major cost cutting is due to the reduce
benchmarking tests as the current setup involves already-
tested component that has working parameters This research
work focused on in-house development of low-cost reaction
wheel using hard drive-based BLDC motor and flywheel.
Different kind of hard drive are available commercially with
different sizes and parameters and most common ones are 2.5”
inches and 3.2” inches. For our research, 3.2” inch variant was
used for one-dimensional control. Furthermore, for controlling
the reaction wheel the drive circuitry is based upon Arduino
Uno controller with electronic speed controller/ the total cost
Figure 10 RW response change in angular momentum to incurred in given in table. The cost break down of reaction
torque
wheel is for one dimensional. For three dimensional, total cost
comes to be $330 which is economical as compared to the
commercially available reaction wheels.
TABLE 5 REACTION WHEEL BUDGET DISTRIBUTION
S. No Component Cost
1. 3.2” inch Hard drive $50
2. Arduino Controller $10
3. Drive Circuitry $40
4. Miscellaneous $10
Total Cost $110

VI. . CONCLUSION
Since the inception of CubeSat technology, many space-
based applications shifted to this relatively new technology.
Due to cost effectiveness and easy to develop nature, many
students led projects are based on CubeSat. Many COTS
components are employed for developing subsystem of
CubeSat, which includes development of satellite attitude
control systems. Moreover, different mechanisms are used for
ACS.
The ACS model of satellite was developed using
orthogonal formation of three-reaction wheel system. Model
design used three reaction wheel model which. Each reaction
Figure 11 Angular momentum response to command torque wheel has a defined motor, which is parametrized in
accordance with HDD BLDC motor. The speed and torque are
controlled with PID. Furthermore, the model was simulated on
the disturbance torque values which were calculated through

Seventh International Conference on Aerospace Science & Engineering 110

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

the relevant equations. Maximum value of disturbance torque [8] D. G. Diez, “Study and Design of the attitude control of a cubesat 1U
was selected as initial parameters, which was given manually based on reaction wheels,” Politician de Catalunya University, BArcelona,
Spain, 2019.
to model. In response, opposing torque was generated by
[9] B. Wie and C. M. Raithmayr, “Integrated Orbit, Attitude, and Structural
reaction wheel, which suggests that hdd motor-based reaction Control Systems Design for Space Solar Power Satellites,” Langley Research
wheels store enough torque to oppose the maximum torque Center, Hamptom, Virginia, Technical Memorandum NASA/TM-2001-
and can be used to maneuver satellite to required attitude. 2108554, 2001.
Saturation problem associated with reaction wheel is not dealt [10] W. F. Hilton, “Spacecraft Attitude Dynamics. P. C. Hughes. John
in this research work. Wiley & Sons, Incorporated, New York. 1986. 564 pp. Illustrated. £48.95.,”
The Aeronautical Journal, vol. 90, no. 894, pp. 152–152, Apr. 1986, doi:
Secondly, prototype of reaction wheel is also developed in 10.1017/S0001924000015578.
accordance with the constraints of 1U CubeSat. The COTS [11] M. D. Griffin and J. R. French, Space vehicle design, 2nd ed. Reston,
components are hdd based BLDC motor and flywheel. The VA: American Institute of Aeronautics and Astronautics, Inc, 2004.
torque and angular momentum storage of components were [12] juan molina, victor ayerdi, and L. Zea, “Attitude Control Model
calculated using the motor formulas and the calculated values forCubeSats,” in Techn Session VII: Attitude Determination and Control
Systems, Florianopolis, Brazil, Jan. 2016, vol. 2nd, p. 22.
were found out to be accurate enough to control the attitude of
satellite precisely. [13] J. Gießelmann, “Development of an Active Magnetic Attitude
Determination and Control System for Picosatellites on highly inclined
Conclusively, this research work provides cost efficient circular Low Earth Orbits,” School of Aerospace, Mechanical and
Manufacturing Engineering Science, Engineering and Technology Portfolio
method to develop reaction wheel system for 1U CubeSat RMIT University, Melbourne City, 2006.
alongside with actuation model that simulates the working of [14] D. A. Vallado and W. D. McClain, Fundamentals of astrodynamics
developed reaction wheel. From the simulations, it can be and applications, 2nd ed. Dordrecht; Boston: Kluwer Academic Publishers,
assessed that reaction wheel system are providing suitable 2001.
torque to control the attitude of 1U CubeSat. [15] J. Zorita, “Dynamics of small satellites with gravity gradient attitude
control,” KTH, School of Electrical Engineering (EES), Zurich, 2011.
REFERENCES [16] N. Muruganantham, “STATE SPACE MODELING AND
SIMULATION OF SENSORLESS PERMANENT MAGNET BLDC
MOTOR,” International Journal of Engineering Science and Technology,
[1] J. Li, M. Post, T. Wright, and R. Lee, “Design of Attitude Control
vol. 2, no. 10, p. 8, 2010.
Systems for CubeSat-Class Nanosatellite,” Journal of Control Science and
Engineering, vol. 2013, p. e657182, May 2013, doi: 10.1155/2013/657182. [17] H. RUOYU, “Moment of inertia and torque performance sensorless
measurement for HDD used spindle motors,” Thesis, Natioanl University of
[2] J. R. Wertz and W. Larson, Eds., Space Mission Analysis and Design.
Singapore, Singapore, 2004.
Springer Netherlands, 1991.
[18] M. Feng and L. Qu, “Torque Performance Analysis of HDD Spindle
[3] H. Polat, “PROTOTYPE DESIGN AND MISSION ANALYSIS FOR
Motors Based on Geometrical Interpretation,” IEEE Transactions on
A SMALL SATELLITE EXPLOITING ENVIRONMENTAL
Magnetics - IEEE TRANS MAGN, vol. 47, pp. 1906–1910, Jul. 2011, doi:
DISTURBANCES FOR ATTITUDE STABILIZATION,” Naval
10.1109/TMAG.2011.2131122.
Postgraduate School, Monterey, California, 2016.
[19] T. Vergoossen, “Design, Integration and Testing of the Delfi-PQ
[4] J. R. Wertz, D. F. Everett, and J. J. Puschell, Space mission engineering:
Satellite Reaction Wheel,” Delft University of Technology, Delft, 2017.
the new SMAD. Hawthorne, CA: Microcosm Press : Sold and distributed
worldwide by Microcosm Astronautics Books, 2011. [20] L. Sahar, E. Edlerman, H. Agalarian, V. Balabanov, and P. Gurfil,
“Hard Disk Drive Based Reaction Wheels for CubeSat Attitude Control,”
[5] “small_spacecraft_technology_state_of_the_art_2014.pdf.” Accessed:
Journal of Spacecraft and Rockets, vol. 55, no. 1, pp. 236–241, 2018, doi:
Sep. 10, 2020. [Online].
10.2514/1.A33866.
[6] S. S. Nudehi, U. Farooq, A. Alasty, and J. Issa, “Satellite attitude control
using three reaction wheels,” in 2008 American Control Conference, Seattle,
WA, Jun. 2008, pp. 4850–4855. doi: 10.1109/ACC.2008.4587262. .
[7] Z. Ismail and R. Varatharajoo, “A study of reaction wheel configurations
for a 3-axis satellite attitude control,” Advances in Space Research, vol. 45,
pp. 750–759, Mar. 2010, doi: 10.1016/j.asr.2009.11.004.

Seventh International Conference on Aerospace Science & Engineering 111

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Comparative Study of Image to Image Translation

Models for Abnormal Weather Effects Removal
Maria Siddiqua Naeem Akhter Javaid Khurshid
Department of Computer and Information Department of Computer and Information Department of Computer and Information
Sciences Sciences Sciences
Pakistan Institute of Engineering and Pakistan Institute of Engineering and Pakistan Institute of Engineering and
Applied Sciences Applied Sciences Applied Sciences
Islamabad, Pakistan Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected] [email protected]

Abstract—Contamination of images by abnormal weather

environments such as haze, rain, fog, snow is inevitable as they II. RELATED WORK
are naturally occurring weather effects. These weather effects Restoration and enhancement procedures for weather-
degrade the effectiveness of various computer vision
affected images are classified into two main groupings
applications like traffic sign detection, audiovisual surveillance,
aerial and ground navigation. Therefore, a unified image-to-
namely prior-based and deep data-driven methods.
image transformation model is proposed that is Multi-Modal In prior-based methods, assumptions or additional
Contextual Conditional Generative Adversarial Network information is provided such as layer, channel, shape, natural,
(MC2GAN) which is capable of eliminating several weather or color prior to our model to generate meaningful outputs.
effects that are snowflakes, haze, rain streaks, and fog. The Anparasy Sivaanpu and Kokul Thanikasalam [1] proposed a
addition of multiple bottleneck layers amid the encoder and defogging model which computes the depth map, dark
decoder of the generator makes the construction of contextually channel, and transmission for the dense fog presence in local
alike output images to the ground truth image given together scene-specific regions of an image. Yuanyang Zou et al. [2]
with the input image as a constraint. The usefulness of MC2GAN
used logarithmic guide filtering to estimate the general
is compared with the existing image-to-image transformation
networks specifically, pix2pix, cycleGAN, unsupervised image
lighting and proposed a self-adaptive method to overcome the
translation (UNIT), contrastive unpaired translation (CUT), defects of dark and bright channel prior. Qiaosi Yi et al. [3]
and fastCUT visually and numerically. After a thorough used a method with residual channel prior to preserve the
examination of supervised and unsupervised networks, it is structure in rainy images. Similarly, semantic and geometric
concluded that our proposed model MC2GAN outpaces other priors are exploited for snow removal by Kaihao Zhang et al.
models. [4]. Although prior-based methods have produced remarkable
results, in some cases, where the images are heavily degraded
Keywords—Street views, image restoration weather effects, by weather effects, the prior-based methods do not guarantee
image-to-image transformation, conditional adversarial network, to perform well, the assumptions might disrupt the restoration
context encoders. process.
Owing to the promising results of deep neural networks,
I. INTRODUCTION the recent focus from prior-based methods is shifted towards
In recent times, research in the image restoration domain deep learning models. For specifically working on images,
has gained a lot of consideration due to its immense deep convolutional neural networks (CNN) and generative
requirement in various machine vision applications. With the adversarial networks (GAN) are commonly applied. As
increased commercialization of autonomous vehicles, the compared to the former, the latter approach learns the
need to have clear images for accurate navigation is essential representation from input to the target distribution, thereby,
but the weather effects like fog, haze, rain, and snow invade making it an image-to-image translation problem.
the performance by corrupting its visual quality which results Various algorithms are presently based on CNN for the
in poor scene visibility. Similarly, images with improved restoration of weather-affected images. R. Suganya and R.
visibility are a requisite for many tasks such as criminal Kanagavalli [5] proposed a hybrid gated recurrent unit and
detection and recognition from CCTV footages, CNN-based scheme for defogging. Mayuri Dongare and Jyoti
computational photography, traffic sign, and lane detection. Kendule [6], introduced a CNN-based dense haze removal
In this paper, a comparative study of our proposed network network. Similarly, P. Musafira and K. S. Shanthini [7] used
named MC2GAN with different image-to-image translation the CNN model for removing rain within a single image.
networks is presented for weather effects removal. The Minghan Li et al. [8] introduced an online multi-scaled
subsequent section of this paper provides a brief outline of convolutional light coding model to remove real-time snow.
previously proposed techniques for the restoration of weather Recently, many remarkable results are achieved by GAN-
degraded images. Section 3 explains the proposed work. based methods for improving visibility. Dong Li et al. [9]
Experimental analysis of the proposed work is specified in performed image defogging using conditional GAN. Amrita
section 4. At last, concluding statements are specified in Sanjay et al. [10] worked on haze removal using GAN.
section 5. Yanyan Wei et al. [11] used cycleGAN an unsupervised
model with rain attentive detecting for deraining images.
Similarly, Aiwen Jia et al. [12] used an attention-based GAN

Seventh International Conference on Aerospace Science & Engineering 112

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

model for single image snow removal. image transforming models: pix2pix, cycleGAN, UNIT,
An enormous amount of work is accomplished in the field CUT, and fastCUT is reported.
of image enhancement and restoration, but most of the The proposed model MC2GAN is a supervised variant of
methods are confined to single weather effect removal the GAN model with a generator and discriminator. The
whereas, our proposed model MC2GAN is capable of generator is made up of eight encoding and decoding blocks
handling four weather effects: fog, haze, rain streaks, and with skip connections. Each encoder block has convolutional,
snowflakes. A solo model is proposed which is compared to batch normalization, and leaky ReLU activation layers, while
five image-to-image translation models. The comparative the decoder block involves leaky ReLU, batch normalization,
analysis shows the reliability of our proposed model. and deconvolution layers together with two bottleneck layers
amidst as shown in Fig. 1, for better encoding the image
III. PROPOSED WORK background. The discriminator network has five encoding
The goal in this paper is to have weather affected image blocks alike generator, as shown in Fig. 2, which takes the
with either fog, haze, rain, or snowflakes effect and clean it affected and ground truth images as input, and outputs a
by removing the weather effect. This is considered as an guess, 1 if the degraded image is restored and 0 if it is
image-to-image translation problem and such a model which degraded. The condition to produce an output image is alike
does competent image-to-image translations are generative a ground truth image, and for contextual similarity, bottleneck
adversarial networks (GANs) which consist of generator and layers are applied to make the proposed model effective.
discriminator models. Generator and discriminator work in
adversary, each one tries to outperform the other. The The loss function of MC2GAN is stated as
generator generates an image as close to the ground truth
image as possible to fool the discriminator, whose task is to LGAN (G, D) =(m,n )[log D(m, n)]+𝔼(m)[log[1-D(m, G(n))]] (1)
identify the generated image as real or fake. The output image
is generated via GAN in two ways either unsupervised or Here m represents input image, n output image, and
supervised. In supervised GAN models, paired images are generator G put efforts in dropping this loss for an adversarial
provided to the network for generating output images, while, discriminator D that tries to surge it, G* = minG maxD LGAN(G,
the unsupervised model takes unpaired images lacking details D). G* together with L1 distance prevent images to get blurry
of the ground truth image corresponding to the input. in the restored images. The discriminator work remains the
Pix2pix [13] is a supervised GAN-based model which same, while the generator is held responsible to deceive the
takes the ground truth image together with the input image as discriminator by producing an output image analogous to the
a condition. Many alternatives of GAN exist from which ground truth image.
cycleGAN [14], UNIT [15], CUT [16], and fastCUT [16] are
a few of the prominent unsupervised GAN-based models. LL1(G) = 𝔼(m, n)[||J-G(m)||1] (2)
CycleGAN model works cyclically by generating the output
image from an input image and vice versa. UNIT is a coupled Therefore, the final loss function is
version of GAN where the hidden space is shared for learning
the target domain. The CUT model tries to compare the small G* = minG maxD LGAN(G, D) + λLL1(G) (3)
input image areas to its ground truth image areas precisely to
achieve maximally similar output images, whereas, fastCUT Here G* signifies lowest with respect to G of the highest
is a modified version of the CUT model analogous to the with respect to D. This loss function is improved by training
cycleGAN model but simpler and has fewer computations as the model with a batch size of 1, learning rate 0.0002, and
identity loss is not used in it, unlike the CUT model. In this decay rate 0.99.
paper, a comparative study of MC2GAN with five image-to-

Fig. 1. Block diagram of MC2GAN generator.

Fig. 2. Block diagram of MC2GAN discriminator.

Seventh International Conference on Aerospace Science & Engineering 113

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

IV. RESULTS AND EVALUATIONS removal. A paired dataset for training MC2GAN and pix2pix
The proposed model MC2GAN is evaluated and is used, while, cycleGAN, UNIT, CUT, and fastCUT models
compared both qualitatively and quantitatively with other are trained by an unpaired dataset.
models. The outcomes are examined and reported in the next
sections. B. Training
A. Dataset All the architectures in this work are trained on an
The emphasis of this work is to enhance the visibility of NVIDIA GeForce GTX 1070 GPU system.
images with street views. Google street view dataset [17] is C. Qualitative results
used to accomplish this goal. Moreover, a self-made image
The Qualitative comparison of MC2GAN with pix2pix,
dataset is created by embedding snow, fog, and rain effects cycleGAN, UNIT CUT, fastCUT is shown in Fig. 3.
within it using an image augmentation library [18]. A
synthetic dataset OTS RESIDE [19] is utilized for haze
Fog Haze Rain Snow

Input

Pix2Pix

CycleGAN

UNIT

FastCUT

CUT

MC2GAN
(ours)

Ground
truth

Fig.3. Visual evaluation of MC2GAN results with pix2pix, cycleGAN, UNIT, CUT, and fastCUT model results.

The first row shown in Fig. 3 contains input images properly. The results of snowflakes removal are better with
provided to our generator model and the latter row shows the MC2GAN, in the rest of the results some of the snowflakes
ground truth target images. The first column of fog results are still present in the output. Though, from the visual
shows how inefficient pix2pix, cycleGAN, and UNIT are in comparison, it can be concluded that the output of MC2GAN
removing fog from the sky and street, moreover, as compared is maximally analogous to the ground truth image as the
to the target image, pix2pix, cycleGAN, UNIT, CUT and effects are removed completely and have improved visual
fastCUT are unable to color the leaves properly on the left quality than the results of other models. The output of other
side of the image. From column three and four, it can be seen models is incapable to eliminate the weather effects properly,
that the haze is still present in all model's results except specifically in the case of snowflakes. Therefore, our model
MC2GAN, which have restored the hazy image properly and performs well for all input modalities.
also enhanced the image contrast. In case of rain, UNIT has The proposed model is also tested on a few dust and
underperformed the most by not removing the rain streaks lightning images, though the model is not trained on these

Seventh International Conference on Aerospace Science & Engineering 114

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

images but is still able to improve its visibility as shown in dust, the brownish shade is removed thereby increasing the
Fig. 4, this shows the generalization power of the proposed clarity of the scene, and the lightning streak in the sky part is
model. From Fig. 4, it can be determined that the overall also taken out successfully.
image contrast and scene visibility is boosted. In the case of
Dust

Input

Output

Lightning

Input

Output

Fig.4. Qualitative results of dust and lightning effect removal.

improving the weather degraded images both visually and
D. Quantitative results
numerically. In the future, more weather effects can be added
Two performance measures: Peak Signal to Noise Ratio to the unified proposed model. Training and testing can be
(PSNR) and Structural Similarity Index Measure (SSIM) are done on more images. For comparing the performance of the
used for quantitative comparison of MC2GAN with other proposed model, more models can be used.
models. PSNR tells the ratio of an image to get corrupted by
noise to its maximum power. SSIM is a measure that tells the
extent to which the two images are structurally similar. A ACKNOWLEDGMENT
higher value of PSNR and SSIM is desired. The numerical I would like to thank the Data Science Lab of Pakistan
values of PSNR and SSIM for all models are mentioned in Institute of Engineering and Applied Sciences (PIEAS) for
Table I. giving an outstanding informative atmosphere and up-to-date
lab services. This research work was backed by PIEAS
TABLE I. PSNR AND SSIM VALUES COMPARISON
Information Technology and Telecom Endowment Fund.
Models PSNR SSIM
MC2GAN 26.2924 84.2333
Pix2Pix 24.3770 80.0911
REFERENCES
CycleGAN 23.6015 81.3006
UNIT 22.3177 79.8123 [1] He, K., Sun, J., Tang, X., “Single Image Haze Removal Using Dark
FastCUT 21.0206 72.1580 Channel Prior. IEEE Transactions on Pattern Analysis and Machine
Intelligence,” Vol. 33 No. 12, 2341-2353, 2011.
CUT 20.6191 70.4681 [2] Tan, Z., Bai, X., Wang, B., Higashi, A., “Fast Single-image
Defogging,” FUJITSU Sci. Tech. J., Vol. 50 No. 1, 60-65, 2014.
Table I. specifies that MC2GAN has outpaced the other [3] Li, Y., Tan, R.T., Guo, X., Lu, J., Brown, M.S., “Rain Streak Removal
models numerically likewise in PSNR and SSIM values. Using Layer Priors,” In: Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition (CVPR), 2736-2744, 2016.
V. CONCLUSION [4] Lu, Y., Xu, M., Jia, S., Huang, W., “Fast snow removal algorithm based
on the maximum value of the degree of polarization and angle of
To assess the usefulness of our proposed model in polarization,” Physica Scripta, Vol. 94 No. 4, 2019.
restoring the weather degradations, its comparison with [5] Qingjiang, C., Xue, Z., “Image defogging algorithm combined with full
various image-to-image transforming models is presented. convolution neural network,” Journal of Applied Optics, Vol. 40 No. 4,
Our proposed model MC2GAN is proved to be optimal in 596-602, 2019.

Seventh International Conference on Aerospace Science & Engineering 115

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
[6] Li, B., Peng, X., Wang, Z., Xu, J., Feng, D., “AOD-Net: All-in-One
Dehazing Network,” In: Proceedings of the IEEE International
Conference on Computer Vision (ICCV), 4780-4788, 2017.
[7] Fu, X., Huang, J., Ding, X., Liao, Y., Paisley, J., “Clearing the Skies:
A deep network architecture for single-image rain removal,” In:
Proceedings of the IEEE Transactions on Image Processing, Vol. 26,
2944-2956, 2017.
[8] Liu, Y., Jaw, D., Huang, S., Hwang, J., “DesnowNet: Context-Aware
Deep Network for Snow Removal,” In: Proceedings of the IEEE
Transactions on Image Processing, Vol. 27, 3064-3073, 2018.
[9] Liu, W., Yao, R., Qiu, G., “A physics based generative adversarial
network for single image defogging,” Image and Vision Computing,
Vol. 92, 2019.
[10] Anvari, Z., Athitsos, V., "Dehaze-GLCGAN: Unpaired single image
dehazing via adversarial training," ArXiv, abs/2008.06632, 2020.
[11] Matsui, T., Ikehara, M., “GAN-Based Rain Noise Removal from
Single-Image Considering Rain Composite models,” IEEE Access,
Vol. 8, 40892-40900, 2020.
[12] Jaw, D., Huang, S., Kuo, S., “DesnowGAN: An Efficient Single Image
Snow Removal Framework using Cross-resolution Lateral Connection
and GANs,” IEEE Transactions on Circuits and Systems for Video
Technology, 2020.
[13] Isola, P., Zhu, J., Zhou, T., & Efros, A.A., "Image-to-Image
Translation with Conditional Adversarial Networks," IEEE Conference
on Computer Vision and Pattern Recognition (CVPR), pp. 5967-5976,
DOI: 10.1109/CVPR.2017.632, 2017.
[14] Liu, M., Breuel, T., & Kautz, J., “Unsupervised Image-to-Image
Translation Networks,” ArXiv, abs/1703.00848, 2018.
[15] Zhu, J., Park, T., Isola, P., & Efros, A.A., “Unpaired Image-to-Image
Translation Using Cycle-Consistent Adversarial Networks,” IEEE
International Conference on Computer Vision (ICCV), 2242-2251,
2020.
[16] Park, T., Efros, A.A., Zhang, R., & Zhu, J., “Contrastive Learning for
Unpaired Image-to-Image Translation,” ArXiv, abs/2007.15651, 2020.
[17] Zamir, A., Shah, M., “Image Geo-Localization Based on Multiple
Nearest Neighbor Feature Matching Using Generalized Graphs,” IEEE
Transactions on Pattern Analysis and Machine Intelligence, Vol. 36,
1546-1558, 2014.
[18] Image Augmentation Library. [online] https://github.com/aleju/imgaug
Last Accessed 6th July 2021.
[19] Li, B., Ren, W., Fu, D., Tao, D., Feng, D., Zeng, W., Wang, Z.,
“Benchmarking Single-Image Dehazing and Beyond,” IEEE
Transactions on Image Processing, Vol. 28, 492-505, 2019.

Seventh International Conference on Aerospace Science & Engineering 116

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

PICK AND PLACE HAND-GESTURED-ROBOT

FOR PHYSICALLY CHALLENGED PEOPLE
Muhammad Ihtisham Babar Muhammad Ibtisam Babar Azeema Azhar
Department of Aerospace Engineering Department of Mechatronics Department of Mechatronics
CAE,National University of Sciences Engieering Engieering
and Technology, NUST Air University, AU Air University, AU
Islamabad, Pakistan Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected] [email protected]

Abstract— Robots, which are used to pick up an object and robotic vehicle and arm. The movement of the vehicle will be
transfer it to the desired place, are known as pick and place controlled via hand gesture wrest. The movement of the arm
robots. It can be of any shape and design e.g., a cylindrical robot,
will be controlled the motion of fingers. The successful
spherical, articulate or scara robots providing movement in
horizontal, vertical and rotational axes. In this article, transmission is desired between the transmitter and receiver.
development of pick and place robot is discussed. The robot is Gesture recognition is the main task in the design. The
controlled via NRF communication. The robot comprises of two movement will act as an input for the motion of robot. Robot
main parts, robotic vehicle and arm. The robotic vehicle is will identify the movement of human hand to complete the
capable to move in all possible four directions. The movement of certain tasks assigned to it. The sensors on the glove will
the robotic vehicle is controlled by gesture human hand wrest record and convert the raw mechanical data into electrical
using accelerometer. The movement of the arm is controlled by
the gesture of human hand fingers using optical sensors. The
signals. This data will be further utilized to make it in
controlled movement of robotic arm is such that it can pick and understandable format by the system.
place objects up to weight of 1 kg from any source place to
desired destination. The prime aim of the designed robot is to II. LITERATURE REVIEW
make it feasible for physically challenged people to transfer and B. A. Malar [2], designed a basic robot which will be controlled
get things using hand motion. by accelerometer rather than button control. The robot moves
according to the motion of the hand. Arduino Uno is used in
Keywords— NRF communication, pick and place robot,
the design of the system.
accelerometer, optical sensor, hand motion.
Wu. [3], implemented hand-gesture control on a smart robot to
I. INTRODUCTION take data of hand path with the help of three-axis
accelerometer. The hand data is sent to computer wirelessly
Robots are one kind of electromechanical device which is
with the help of an RF module. The received paths are then
operated by its internal circuitry. Robotics is one of the new sorted to one of six control commands for navigating a robot.
emerging field of engineering in the upcoming years. The The dynamic time warping (DTW) algorithm is assumed by
demand of the robot’s is increasing significantly from the classifier to classify hand path. With the help of hand paths,
education sector to transportation, restaurant and many other robot can be controlled by the user.
sectors. The traditional robot controlled via wire are very Thivagar [4], developed a voice-based information controlled
common now a days. But lot of wires make it very difficult robot with the help of Arduino. With the help of “hand
to handle and limits the distance covered by the robot. The gesture”, feature user can move robot wirelessly depending on
wireless robot can solve these problems. The wireless robot the Radio Frequency, placed on the hand.
will be controlled by human hand gestures. The movement of Kumar [5], designed a system to record wireless movement
robot is controlled by the hand and finger gestures using and controlled the movement of robot using respirators.
microcontroller and optical sensors, which makes it a Robotics and part of body are connected by a wireless
beneficial robot system for user friendly support, bomb connection. Mainly the robot is controlled by Arduino Uno.
disposal, material handling and to enhance the independence The aim was to improve the components and programming for
of physically impaired persons [1]. It is a kind of efficient a smart arm controlled by an accelerometer, have been
alternative for wired control robot that aids the impaired achieved. From the idea that has been made, it obviously shows
people which often suffers from physical problems. By that development and production of this smart robot is easy,
making the wireless interface using transmitter and receiver friendly and easy to control.
of specific frequency, makes the robot to be controlled very Herkariawan [6], developed hand gestured controlled Arduino
easily. Gesture control is so far a better alternative to voice- based warfare robot. MPU6050 gyroscope was used to
controlled devices for controlling the robots as voice controls determine the orientation of the robot. Nathan [7], concluded
that robots are viable options to replace manual operations and
mostly suffers from interruption of unwanted noise and
integrations of vision and touch sensors in robots can change
interface.
the work efficiency significantly.
The main aim of this paper is to design and fabricate a hand
Ifeoma [8], designed a robot which is able to handle real time
gesture-based pick and place robot controlled by NRF instructions using gesture control using Arduino mega and
communication for physically challenged people. The robot Baron [9], designed single application-controlled NXT robot
should be capable to move in all possible four directions by tracking calibration using gesture and voice command.
(forward, backward, right and left). The robot comprises of

Seventh International Conference on Aerospace Science & Engineering 117

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Bodiroza [10], discussed the generalized design method of the b. Optical Sensor
gesture vocabulary (GV) for intuitive and natural two-way Optical sensor is used here instead of flex sensors,
robot-human dialog. The designed is based on motion gestures normally used in gesture control systems. Usually Flex
using robot waiter (RW) dialog. Zhao [11], discussed about sensors are mounted on finger to detect the amount of
the gesture control technology and its impact on the education bend. The change in the internal resistance of flex sensor
sector. gives appropriate output to microcontroller. But these flex
sensors are very costly. The similar work out can be
III. SYSTEM ARCHITECTURE achieved from a low-cost optical sensor as shown in
The complete system comprises of transmitter and receiver Figure 3.
end. The gesture sensing glove is the transmitter end while the
robot is the receiving end.

1. The Gesture Sensing Glove (Transmitter End)

The block diagram of the transmission end is shown in Figure
1. The main components of the gloves are the accelerometer,
optical sensor, Arduino nano microcontroller and NRF with
2.4GHz transmission module. A separate power bank is used
to excited all the components of the gesture sensing glove.
Figure 3 Optical Sensor

The optical sensor includes bright LED and LDR (Light

Dependent Resistor). The LED and LDR are fixed in at
each end of a fixable rubber pipe [13]. When a user bends
his finger, the intensity of the light from LED on the LDR
changes, the output of the LDR changes accordingly. This
Figure 1 Block diagram of the transmission end change in the output of LDR can be processed by the
microcontroller.
Following are the details of the each component in the
transmission end (gesture sensing glove) of the system:- c. Arduino NANO
Arduino Nano is used on the glove while Arduino UNO is
a. GY-61 Accelerometer used on the robotic vehicle. Arduino Nano can be
Gy-61 accelerometer is used to detect the exact orientation considered as a self-sufficient system with a processor,
of glove. It detects the orientation both axes and gives its storage and programmable input/output device and can be
equivalent analog values to the microcontroller. Gy-61 is used as an implanted system. The Arduino Nano is a small
a low power, thin, small three axis accelerometer gives full package circuit based on the ATmega328P; that gives
easy way to the X, Y and Z axis analog outputs. It is the same specifications of the Arduino UNO but little less.
capable of detecting 3g forces in all three axes. It is used
commonly in the tilt detecting applications for the shock
or vibration evaluation due to dynamic acceleration and
static acceleration of gravity [12]. The bandwidth of the
accelerometer is selected by the user pin capacitance X
out, Y out and Z out. The elective suitable bandwidth
depends on the application, X-axis and Y-axis bandwidth
is between 0.5 Hz and 1600 Hz, while Z-axis of the
bandwidth is between 0.5 Hz and 550 Hz. Accelerometer
takes 5V supply for its operation. However, 3.3V on-board
regulator is also available letting it to be powered from a
3.3 to 5V power supply source. Its sensitivity is about
300mV/g and is dimension is 20.3mm x 15.7mm. Figure
2 shows the pin configuration of the Gy-61 accelerometer. Figure 4 Detail diagram of Arduino NANO

Arduino Nano is another member of the facilely with similar,

lesser or more features than Arduino UNO. Arduino Nano
Pinout contains 14 digital pins, out of 14; five can be used for
built in PWM feature, 8 Pins are for anlog input, 2 Pins are for
reset purpose as shown in Figure 4.

d. Wireless Transmission
In the design of our glove system, NRF24L01 wireless
transmitter module is used. It runs on frequency band of
Figure 2 Pin configuration of G-61 accelerometer 2.4 GHz. Its optimal range is 100m which can be increased
if the antenna is active. It can work within baud rates of

Seventh International Conference on Aerospace Science & Engineering 118

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

250 kbps to 2 Mbps. In open area, its range can rise up-to d. Rechargeable battery
100 meters, with lower baud rate. The module uses variety The receiver circuit (robotic vehicle) is operated by a 9.4
of channels which gives a opportunity to have an V rechargable battery.
independently working network modems at the same time.
Only 12mA power is consumed by this module during this e. DC Buck Regulator
transmission. The operational voltage of the module is The servo can operate only on 5V while the receiver
from 1.9 to 3.6V, but the main advantage is that the other cricuit require 9.4V, for this purpose a LM2596 DC-DC
pins can endure 5V logic, so it can be easily attached with buck convertor is used.
an Arduino without using any logic level converters. Other
characteristics of the NRF radio frequency module is that f. NRF24L01 with 2.4GHz Receiver Module
it is of low cost, easily available and able to make Two NRF transceiver are used one as a transmitter in the
communication possible even through lot of obstacles. glove and another as a receiver in the robot. NRF24L01
is an 8-pin board. The pin configuration of NRF24L01 pin
2. Robot (Receiver End) 1 and pin 2 will be connected to GND and VCC
The receiver end consists of following components:- respectively. Pin 3 is the chip enable used to enable the
SPI communication. Pin4 CSN must be high fir SPI
a. L298N motor driver and DC motors communication. Pin5 is serial clock used to provides the
The speed of the DC geared motor used here is 125 RPM. clock pulse using which the SPI communication. Pin6 and
Its operating voltage ranges from 5 V to 15 V and draw pin7 are MOSI (Master Out Save In) MISO (Master in
current of 80 mA to 100 mA. Motors and wheels are Slave Out) respectively. MOSI and MISO are used to
shown in Figure 5. transmit or receive data in SPI communication. The pin
configuration of NRF24L01 is shown in Figure 7.

Figure 7 Pin configuration of NRF24L01

The block diagram of the receiver end is depicted in the

Figure 5 DC Motor mounted on the wheels Figure 8.

b. Servo motors
The servomotor used in the robot is known as Tower Pro
MG 945. It has an operating voltage range of 4.8 V to 7.2
V and is capable to provide 10 Kg/cm. The robotic arm
will be controlled via Arduino UNO microcontroller and
it will be able to grasp, lift and translate objects.
Figure 8 Block diagram of receiver end
c. Arduino UNO
Arduino UNO boards consists of USB interface, 14
digital I/O pins, out of 14 ports, 6 pins can be used for IV. ELECTRICAL WIRING
PWM output and 6 for analog signals [14]. Arduino UNO Electrical and electronics components handling is very
comes with integral LED, which is attached through pin13 tedious job in the project. Any miss handling of wires polarity
as shown in Figure 6. can leads to serious damages. The electrical part of the project
is consisting of two circuits i.e., the transmitter circuit and the
receiver circuit.

1. The Transmitter Circuit

The movement of the robotic vehicle is controlled by the
accelerometer mounted on the glove. The movement of hand
that comes from accelerometer is feed into Arduino nano
microcontroller board for further processing. The outputs of
the accelerometer X-out, Y-out and Z-out are connected to the
analog pin A0, A1 and A2 of the microcontroller respectively.
The gripper of the robotic arm is controlled by the optical
sensor mounted on the two fingers of the glove. The two
sensors are connected to the analog pin A3 and A4 of the
Figure 6 Detail diagram of Arduino UNO microcontroller. The analog outputs of the accelerometer and

Seventh International Conference on Aerospace Science & Engineering 119

 Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

optical sensors are processed by the microcontroller and the 3. Methodology of Control Motion
output of the microcontroller are transmitted by the NRF Serial Peripheral Interface (SPI) is an interface bus is used to
module to receiver circuit. The MOSI pin and MISO pin of the send data between two microcontrollers. The two NRF
NRF module is connect to SPI pins; pin D11 and D12 modules communicates to each other by the SPI protocol. The
respectively. The chip enable CE pin of the NRF is connected direction of hand movement is given by the accelerometer. As
to pin D7, CSN pin is connected to pin D8 and the serial clock hand moves, the accelerometer reading varies. This varying
SCK is connected to pin D13 of the microcontroller. The signal is sent to Arduino analog input. Arduino board has an
transmitter circuit is operated by 5V power bank provided for integrated 8 bit analog to digital converter. Signal is conveyed
the glove. Figure 9 gives the wiring and circuit diagram of the to the receiver by the NRF module of the transmitter. The
transmitter. signal is trabsfered to the NRF module of the receiver and is
sent to the Arduino digital input for further process. This signal
is then transferred to the driver of the motor to rotate the motor
such that the robot moves in the desired direction as of hand
movement.
The optical sensors also send determinant values to the
microcontroller. As the user moves his fingers, the LDR
reading varies accordingly. These values are sent to
microcontroller analog inputs and a code is generated based on
the value of the LDR by the microcontroller which is
transmitted to the receiver in order to control the two-servo
motor.
V. HARDWARE IMPLEMENTATION
1. Robotic Vehicle
Figure 9 Circuit diagram of transmitter end The robotic vehicle comprises of 4 wheels which are attached
to the DC geared motors and is controlled by the
2. The Reciever Circuit accelerometer. The vehicle movement have the four motions
The signal transmitted by the NRF of the transmitter is received (Forward, Backward, Left and Right) respectively. The vehicle
at the NRF of the receiver on the robotic vehicle. The MOSI is controlled by the Arduino UNO on the receiver end through
pin and MISO pin of the NRF module is connect to SPI pins; the signal received from the accelerometer on the glove. Four
pin D11 and D12 respectively. The chip enable CE pin of the wheels coupled with DC geared motors can be mounted on a
NRF is connected to pin D7, CSN pin is connected to pin D8 wooden base. The forward and backward movement can be
and the serial clock SCK is connected to pin D13 of the achieved by rotating all the four wheels in same direction. For
microcontroller. The code received is processed by the the right turn the RPM of the right-side wheels should be
microcontroller and the outputs are generated to operate the reduced by PWM and for the left turn the RPM of the left side
motors and servos. wheels must be lower than RPM of the right-side wheels.
The logical outputs for the motor on the right side of the vehicle The first part to implement is to design a strong mechanical
are generated from pin9 and pin10 while the left side motor of base for the gesture control robotic vehicle. Base can be made
the vehicle is connected to pin2 and pin3 of the microcontroller. from low weight acrylic with an appropriate dimension of
The logical inputs of the two servo motors are connected to the 30cm x 30 cm. Slots of 7 cm are made in order to make space
pin5 and pin6 of the microcontroller. Figure 10 shows the circuit for all the four tires inside the base. Slots are moved 4 cm
diagram of the receiver. inward for safety and stability purposes. Wheels can be directly
coupled with the DC geared motor. The finally mount all the
four DC geared motor to base. Figure 11 shows the CAD
model of four wheels coupled with DC geared motor mounted
on the base of the robotic vehicle.

Figure 11 Base of the robotic vehicle

The diameter of the wheel is 6.5 cm while the diameter of the
shaft is 0.5 cm.
Figure 10 Circuit diagram of receiver end

4
Seventh International Conference on Aerospace Science & Engineering 120
Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

2. Robotic Arm The complete model of the hand gesture control robot is shown
Robotic arm is mechanical arm which is statically mounted on in Figure 14. The base is made in such a way that in future
the receiver part of the circuit. The gripper on robotic arm can modification, objects can be place on the base with the help of
revolve up to 180 degree, which is fixed to the servo motor. gripper which will make the robot to carry more objects.
Specific load is applied to test the arm for lifting the objects of
desired weight (up to 1 kg). The servo motor is used for the VI. TESTING
purpose of picking and placing the object, because of its the The accelerometer reading varies between maximum and
low torque applications. Optical sensor in the transmitter part minimum values, as the user moves his hand. These two values
controls the gripper. are used to specify the range for each task of robot. If the
The length of the arm is 12.7 cm. The jaws of the gripper are received value falls between this range, then the equivalent
capable to pick and place an object of width 5cm. Figure 12 determinant is generated. Then the value of the determinant is
shows the CAD model of the robotic arm with opening and received by the microcontroller, which identifies the
closing mechanism. The movement of the arm and micro corresponding movement after processing, and emits signals to
gripper can be controlled by the servo motor. move the robot accordingly.
Figure 15 shows the four specific hand gestures that are used
to control the robot’s motion. When the hand tilts in forward
direction, the gesture is identified as the forward motion, and
the robot moves forward. The speed of the robot is controlled
by the tilt angle. Tilt angle is calculated by the difference
between the angle of tilt of hand and the threshold value of
forward motion. When the hand tilts in the right direction, the
robot identifies the gesture and moves rightward.
Similarly, when hand tilts in the left direction, the robot moves
Figure 12 Robotic arm with (a) opened gripper position, (b) in the left direction. The angle of the tilt of hand defines that
closed gripper position whether it is a normal or sharp turn. In a sharp turn a car varies
the direction without getting slower. When the user tilts his
The schematic model of the robotic arm is shown in Figure 13. hand backwards, the robot moves in the reverse direction. To
Several Test is performed to find out the exact limit of the move the robot in the diagonal direction the movement of hand
weight that can be easily handled without slipping by the will be in somewhere between two gestures.
gripping. It is concluded that it can only handle the weight of
1kg, due to the limitation imposed by the gripper and motor.

Figure 13 Robotic arm attached with servomotor Figure 15 Hand Gestures to control Robot Motion

Two specific gestures to control the movement of the robotic

arm and gripper is shown in Figure 16. When the user bends
his finger, the gesture is recognized and the arm go in forward
direction, when he straightens his finger, the arm goes
backward. The gripper is controlled by second finger; the jaws
of the gripper closes when the user bends his finger and the
jaws of the gripper opens when the finger is straightened.

Figure 16 Finger Gestures to control Arm Motion

Figure 14 Final model of robot

5
Seventh International Conference on Aerospace Science & Engineering 121
 Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

VII. RESULTS AND DISCUSSION designed robots as discussed in literature. Secondly by using
a. Overall performance analysis the Arduino NANO, and modern receiver, the hand glove
The main purpose of this test is to evaluate the movement system has been made more light weight for the easiness of
response of the pick and place robot. Table 1 shows the the user. Thirdly the gripper is controlled only with the
movement response of the robot w.r.t different gestures movement of finger instead of hand, so that rest of the fingers
given command. Good means 100 percent accuracy and can be used for multi-tasking in the future version of the robot
less depicts accuracy below than 90%. whereas in previous robots both gripper and robot were
controlled by hand. Fourthly, the robot can pick and place the
objects up to 1kg easily whereas in previous robots the weight
Table 1 Movement Testing Results carrying capacity was up to 500g.
No Movement Functions Responses IX. CONCLUSION
1 Hand-Down Forward Good
The purpose of research is to design a pick and place robot
2 Hand-Up Reverse Good
using accelerometer sensors attached to a hand glove for the
3 Tilt Hand Left Left Direction Good
physically challenged people. The designed robot allows the
4 Tilt Hand Right Right Direction Good
user to control the motion of vehicle and robotic arm by only
5 Close Finger Close Gripper Good using the hand and finger. The robot is designed using the
6 Open-Finger Open Gripper Good accelerometer and optical sensor which is controlled by the
Arduino. Finally, a test ready model is developed to implement
b. Distance testing in open and closed space area the idea. The model performed the desired tasks up to the mark.
The range testing of the robot is performed to find out It is concluded that this type of pick and drop robot can help
actual capability of the robot. Table 2 and 3 shows the the disable person so that he can use only his hand and finger
behavior of the robot in open and closed space area. movement to control the robot for picking the objects.
Table 2 Distance Testing in Open space Area X. FUTURE WORK
No Distance (m) Response Feedback  The robot can only pick the weight up to 1 kg, more
1 200 Move Good powerful motors can be used to pick more weight.
2 400 Move Good  Camera can be mounted, which could be attached via
3 600 Move Good wireless connection to mobile so that person can pick the
4 800 Move Less things from out of sight places.
5 1000 No Less  For now, the gripper can only move along single axis,
multi axes gripper can be mounted to improve the design.
Table 3 Distance Testing in Closed space Area
No Distance (m) Response Feedback REFERENCES
1 50 Move Good [1] S. K. Talakayala, “Design and fabrication of pick and place
2 100 Move Good robotic arm. 2nd National Conference on Recent Trends in
Mechanical Engineering, GIST”, Nellore, 2020.
3 200 Move Good [2] B. A. Malar, "Hand Gesture Control Robot. International Journal
4 300 Move Less of Innovative Technology and Exploring Engineering (IJITEE)",
5 400 No Less Volume 9, Issue 2,2019.
[3] X. H. Wu, "A Hand-Gesture-Based Control Interface for a Car-
Robot. International Conference on Intelligence Robotics and
The results shows that robot has a range of 800m in open System”, IEEE, Taipei, Japan:2010.
space and 300m in closed space. [4] T. Thivagar, "Hand Gesture and Voice Controlled Smart Vehicle.
International Journal of Modern Science and Technology", 164-
c. Weight Testing 167,2020.
[5] C. P. Kumar, "Gesture Control Robot with Arduino. Material
The Table 4 shows the tests results of the weight that Science and Engineering", IOP Publishing, 2018.
gripper can carry and hold easily. The results shows [6] C. Herkariawan “Design control system using gesture control on
that gripper can carry and displace up to 1 kg of the Arduino-based warfare”. IOP Conf. Series: Material Science
weight. and Engineering,2021.
[7] N. Nathan, “A gesture-controlled system”. International Journal
Table 4 Weight Testing results of Science & Engineering Research, 2015.
No Weight (g) Pick Displace [8] U. Ifeoma, “Design and Implementation of Arduino-Based
1 200 Yes Yes Gesture- Controlled System with accelerometer”. International
Journal of Recent trends in Engineering and Research,2019.
2 400 Yes Yes [9] G. Baron “Gesture and voice driven mobile tribot robot using
3 600 Yes Yes Kinect sensor”. IEEE International Symposium on
4 800 Yes Yes Electrodynamic and Mechatronic Systems (SELM), 2013.
[10] S. Bodiroza, “Dynamic gesture vocabulary design for intuitive
5 1000 Yes Yes human-robot dialog”. IEEE 7th ACM/IEEE International
6 1200 No No Conference on Human-Robot Interaction (HRI), 2012.
[11] L. Zhao “Gesture Control Technology: An investigation on the
potential use in Higher Education”. IT Innovation Centre,
VIII.COMPARISON WITH PREVIOUS ROBOTS University of Birmingham, 2016.
[12] R. Grepl, Selection of MEMS Accelerometers for Tilt
The designed robot can be controlled within the range of Measurements. Journal of Sensor,2017.
800m, which is slightly greater than the conventionally

6
Seventh International Conference on Aerospace Science & Engineering 122
 Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

[13] N. Sabri, "Toward Optical Sensors: Review and Applications.

Journal of Physics Conference Series"2013.
[14] L. Louis, Working Principle of Arduino and Using it as a Tool
for Study and Research. International Journal of Control,
Automation, Communication and Systems (IJCACS),2016.

7
Seventh International Conference on Aerospace Science & Engineering 123
Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

A Review of Vulnerabilities and Cyber-Security

Threats Associated with Overseas Telecom
Companies’ Operations
M Naveed Tahir Asim Shahzad Areeba Azhar
School of Interdiciplinary Enginerring and Federal Urdu University of Arts, Sciences and King Edward Medical University
Sciences (SINES), National University of Technology Lahore, Pakistan
Sciences and Technology, H-12 Campus Lahore, Pakistan [email protected]
Islamabad, Pakistan [email protected]
[email protected]

Abstract— Growing technological advancements in Mobile of population in the country. As this telecom sector is
Telecommunication Companies’ connectivity services have led responsible for provisioning of the wide spread services, thus
them to be a vital component of our daily lives enabling their is prone to security and privacy challenges. Majority of the
users to perform routine tasks on the go. On the other hand, this operations being on the Internet necessitates ensuring security
rapid advancement in technology, broadened bandwidth along and safety of data and information.
with tremendous usage have made these companies vulnerable
to malware and various security breaching attacks. In addition, This paper contributes a brief yet comprehensive
vast options of connectivity, device software platforms and assessment of cyber security threats, security breaches, and
telecom services have also made mobile stations critical vulnerabilities associated with the telecommunication sector
regarding security and privacy. The current study briefly in Pakistan linked directly or indirectly to their operations and
discussing the telecom infrastructure, reviews vulnerabilities classified widely based on a number of key variables.
associated with overseas mobile companies, resulting security
threats, attacks and their solutions for both mobile stations and A. Scope and Design of Current Research
switching centres. Vulnerabilities and resulting attacks are The critical data and information gathering required for the
classified widely considering various variables. With this study vis-à-vis the access limitation to the telecom companies’
categorization, the paper presents an easy and brief view of
infrastructure mainly limited the data acquisition to qualitative
major vulnerabilities, sources associated with these
vulnerabilities, their classification and types which concludes to
research techniques focused on review of pervious researches.
a critical analysis focusing on why they occur, their effects, As is obvious that no setup or specialized industry would
possible solutions, contributing internal and external factors neither share its data highlighting their weak areas and
and key recommendations for their apprehension and vulnerabilities nor would they allow access into their
suppression. infrastructure for such sensitive organization level data
collection as sharing the knowledge of the vulnerabilities
Keywords — cyber-attack, encryption, passive threat, would indirectly mean an in-depth evaluation of their
vulnerability, data security and privacy technical infrastructure to an extent so as to identify possible
loop holes. Therefore, indirect data collection techniques
I. INTRODUCTION through literature review remains the method of choice to
Modern communities have dependent on cyberspace that access the vulnerabilities in form of various cyber security
offers valuable and necessary services for performance of attacks associated with telecom companies. Survey and
majority of functionality of human life. With the increasing comparison of telecom security services vis-à-vis the telecom
adoption of services based on internet aimed to increase the service providers was carried out. The main aim remains
interactions and connections, the government offices are assessing the vulnerabilities, attacks which can occur as a
moving towards transformation of majority of their services result of vulnerabilities, classification of these attacks and any
online. However, this has led to increasing ratio of cyber- possible solutions already worked upon.
attacks and cyber criminals have been able creating an Researchers considered following few questions during
availability, integrity or confidentiality attack upon networks the course of research: (a) What is the infrastructure of
as well as services. These types of cyber activities may Telecom sector in Pakistan? (b) What are the security policies
influence or even hinder the essential ICT services. Though and regulatory authorities governing the telecom industry in
research is being done in the field of cyber security, however, Pakistan? (c) What is meant by the risks, threats and
it needs to be backed by an authentic assessment of the vulnerabilities associated with telecom sector? (d) What is the
vulnerabilities which may be associated with various attacks basic security architecture? (e) What previous researchers
[1]. It is, therefore crucial to analyze these vulnerabilities have studied about the vulnerabilities? (f) What are the main
associated with the operations of Telecom Companies in order sources of vulnerabilities? (g) What are the key vulnerabilities
to protect our cyber space and further design and implement associated with the telecom companies’ operations? (h) What
effective cyber security policies. kind of attacks are associated with these vulnerabilities? (i)
Telecom industry in Pakistan is growing at a rapid pace How can these attacks be categorized and classified? (j) What
with Pakistan’s growing economy being the key factor for its are the implications and possible solutions of the attacks? (k)
growth. With more than a billion users and four mobile What could be the way forward to minimize these attacks?
companies, the sector provides employment to a major chunk

Seventh International Conference on Aerospace Science & Engineering 124

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

B. Organization of the Paper C. Pakistan’s Telecom Sector Overview

The paper is divided into five sections. Section II Telecom industry in Pakistan is growing at a rapid pace
highlights earlier studies related to the infrastructure of with Pakistan’s growing economy being the key factor for its
telecom sector, Pakistan’s current telecom industry’s setup, growth. With more than a billion users and four mobile
cyber security issues and vulnerabilities, mobile companies, the sector provides employment to a major chunk
telecommunications security and vulnerabilities, the of population in the country. The Pakistan Telecommunication
authentication issues, data security and privacy. Section III Act, 1996 is the key legislation governing the telecom industry
presents an analysis of the related works studied in Section II [3][4].
and summarizes the sources and major vulnerabilities. Section The Act proposed the creation of:
IV documents discussion; we consider some findings and
propose some recommendations to minimize the  Pakistan Telecommunication Authority (PTA) for
vulnerabilities and associated cyber-attacks. Section V purpose of regulating telecom segment.
concludes the paper.
 Frequency Allocation Board (FAB) to sort out the
II. VULNERABILITIES RELATED TO TELECOM INDUSTRY allocation and further, the assignment of frequency
operation band to various telecom operators, broadcasters
A. Threat and private wireless users in Pakistan.
A security threat may be called as “a potential violation of
 National Telecommunication Corporation for
security” [2]. For instance: (a) unlawful leakage of data (b)
provision of telecommunication services to government
unauthorized obliteration or alteration of information or
organizations such as defense services, provincial
related resources (c) deletion or damage of data (d) disruption
governments and other institutions.
or rejection of services (e) impersonation, or masked as a legal
entity.  Pakistan telecommunication Company Limited to
A security threat may be classified as accidental or provide basic telecom facilities to users.
intentional and further can be active or passive. PTA being the regulatory authority is accountable for
formation, operation, maintenance and provisions of
 An accidental threat is one with no premeditated
telecommunication services in country. In addition,
intent such as a system or software malfunction or a
provisioning and reinforcement of licenses to telecom
physical failure.
industry’s operators is also responsibility of PTA. It provides
 An intentional threat is one that is realized by guidelines, regulates tariffs, investigates complaints and
someone committing a deliberate act. Intentional threats authorizes the approvals for use of various equipment to the
may range from casual examination, using easily- telecom companies and service providers [4][5][6].
available monitoring tools, to sophisticated attacks using
1) Policies and Deregulation: Mobile cellular policy was
special system knowledge.
launched by Government of Pakistan in 2004 with an
 An active threat is one that results in some change to objective to promote competition among the cellular
the state or operation of a system, such as alteration of companies in Pakistan.it further focused on fair use of
data or destruction of physical equipment. frequency bands, encouraging private investments and lay
down the rights and obligations of the companies along with
 A passive threat involves no change of state.
Eavesdropping and wiretapping are examples of passive customer’s facilitations [7].
threats.
Broadband Policy was also formulated in the same year to
B. Vulnerability encourage growth of the private sector and appreciate civil
Vulnerability is “a flaw or weakness that could be service providers to initiate their services. The policy also
exploited to violate a system or the information it contains. If aimed the provision and development of a high-speed broad
a vulnerability exists, then it is possible for a threat to be band network which would be available without interruption
realized successfully unless effective countermeasures are in [8][9].
place” [2]. 2) Foreign Investment and its Impact on Telecom Sector:
The vulnerabilities associated with the telecom sector may Above stated policies coupled with deregulation policy not
be classified into undermentioned four types : only encouraged the foreign companies to invest in the
(a) Threat model vulnerabilities resulting from failure to telecommunication sector but also enhanced these
foresee possible future threats. companies’ operations manifold. The Pakistan Board of
Investment (BoI) is offering due incentives to these foreign
(b) Design and specification vulnerabilities resulting operators to establish and flourish. As per the Pakistan
from errors or oversights in the design of a system or
Telecommunication Act PTA is required to [10][5];
protocol and make it inherently vulnerable.
(c) Implementation vulnerabilities Introduced by errors “Promote the availability of wide range of high quality,
or oversights during system or protocol implementation. efficient, effective and competitive telecommunication
services throughout Pakistan”
(d) Operation and configuration vulnerabilities
originating from improper usage of options in As a result, in current situation of security and instability
implementations or weak deployment policies and in the country, the massive involvement, operation and data
practices. handling being carried out by the foreign telecom companies
has led to a number of vulnerabilities and cyber security
concerns for their users. Large amount of confidential data is

Seventh International Conference on Aerospace Science & Engineering 125

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

being transferred or stored via various wireless services, from our current situation. An excellent collaboration between
internet connections and direct dialling calls provided by these the technical service providers and security service providers
companies [11][12]. is essentially required to curtail the sensitive situation and the
security providers need to be more focused and modified of
D. Telecom Infrastructure, Cyber Security Issues and the security concerns and the possible vulnerabilities to the
Vulnerabilities users and organizations [16][8].
Cyberspace offers many services for functioning of
modern industrial world and the environment, thereby making E. Vulnerabilities associated Mobile Telecommunications
the modern communities dependent on it. On the other hand, Security
it has posed a number of challenges. Cyber security is Mobile telecommunications are developing from agility
concerned on protecting networks and databases from any that is restricted to a particular technological aspect to cross
cyber security breach, destruction or mutation [13][14][4]. It networks flexibility with incorporation of IP. Networks of the
aims at proposing and providing the capability of influencing future may encompass incorporation of wired as well as
the actions of cyberspace thus requiring in-depth knowledge wireless networks to provide an extensive new services that is
of limitations and vulnerabilities associated with ICT. In not possible by a sole prevailing network. By utilizing true
today’s world, there arises a requirement of thorough and in- Fixed Mobile Convergence (FMC), roaming across
depth involvement at various levels of development and heterogenous networks is possible. The requirements for the
operations of the cyber world, including the telecom sector as security of each type of access are to be met in various ways,
a major chunk, by the developing countries [9][10]. however, all these requirements are be fulfilled for protecting
users, assessable networks and applications[4].
Although the organization related threats have drawn the
attention of researchers and think tanks, however, required Security problems can generally be classified as : -
attention is not being paid to the vulnerabilities’ assessment
and cyber security particularly when creating the national (a) those arising because of the usage of IP in mobile
security strategies. Cyber security is undoubtedly considered wireless;
a challenge nowadays for a number of countries including (b ) because of usage of various multi-technology
Pakistan [4]. networking.
A number of cyber security issues are being faced by Cyber-attacks put a hazard to wireless networks which
Pakistan. Though cyber security issues are threatening the utilize IP for the transport protocol. Further, vulnerabilities are
world, security providers in Pakistan have also become alert prone to appear as a result of the nature of wireless network
and concerned due to continuous security breaches and itself i.e., its mobility. The two most important factors while
cybercrimes. People and telecom companies’ lack of considering the smartphone based threats and vulnerabilities
knowledge regarding the importance of digital security is the are [29]: -
key reason making them vulnerable to cyber criminals. The
two broader aspects of the vulnerabilities associated with the (a) Authentication Issues.
users in telecom sector may be categorized as [5][15]; (b) Data Protection / Privacy.
(a) External Attacks. Cyber security attacks on the 1) Authentication Issues: Refrence [17] suggests a mutti-
companies unaware or unable to enforce strict security sensors-based protocol for mobile stations for achieving
measures. Ultimately, the company and users, both fall a
inherent authentication. This system persistently understands
victim to cyber criminals.
patterns of human behaviour as well as the setting by
(b) Internal Attacks. Risks and vulnerabilities for end permitting users to utilize smartphone without unsettling user
users created by the company through various means. actions. The method can also continuously update and
After a number of incidents, the situation has improved improve the user model. Experiments conducted prove that
and the companies are trying harder to provide digital security effectiveness of the model only needs only 10 seconds to
to their users. However, the obstacle for a flawless security is execute the it and 20 seconds to sense irregular and bogus
the diversity of attacks and old tackling approaches. A rapid requests. Thus, level of accuracy attained by utilizing this
advancement is essential because as the technological world is model can be up to 90-95% .
progressing with each passing day, cybercriminals are also
getting smarter. Telecommunication companies’ major Zahid et al. [18], suggested an identification system to
concern is to provide good security setup to be safe from observe the key users of smartphone in order to distinguish
external attacks to allow their end users utilize the services authentic users and quacks. They utilized the data set acquired
without the concern to digital security [3][4][6][8]. from twenty-five users for establishing the recommended
A number of companies are working in Pakistan on threat system. The proposed system has a lesser than 2% fault rate
recognizing systems in order to provide data security. In which was observed to be nearly zero once PIN authentication
addition, there are exclusive ways to get better data has been done. The authors also coupled their technique with
surveillance technology to prevent mishaps. It appears that state-of-the-art existing detection methods of the users’
digital security companies are currently the only source in keystrokes.
Pakistan to hold the vulnerable situation and take the country Ref [19] proposed “Typing authentication and protection”
out of both external and internal threats associated with (TAP), which is basically a virtual key grounded in the typo
telecom sector and its operation in the country. system of smart phones / devices. The security of smartphone
The current cyber security situation can become worsen if using TAP can be improved with a twostep process i.e. in the
the authorities remain unable to get better cyber security login phase and secondly in the post login phase. During first
through various means as discussed earlier. Acceptance of this stage, TAP works by controlling the biometric information
vulnerable state by the country / users may make us fall back and morphology of users’ hand for protecting the identity

Seventh International Conference on Aerospace Science & Engineering 126

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

while in the second stage, it controls the TAP virtual key. The understanding of telecom industry and finally summarize the
conducted experimental trials established the utilization of major attacks on a client’s smartphone resulting due to the
TAP to conserve security as well as usability of the smart vulnerabilities associated with telecom companies operation.
phones.
A. Sources Associated With Vulnerabilities In The Telecom
Reference [20] suggests a non-intrusive validation process Sector
using the data collection through orientation sensors in smart Deployment of internet and telecom-based services in
phones. Being a new approach, it is functioned by smart various telecom companies operating in Pakistan has
devices in a distinctive way whereby orientation sensor is used increased the incidents of systems’ digital security sabotage
to store these behavioral biometrics. The authors have both from external and internal sources. Though these attacks
concluded that this non-intrusive mechanism may be and incidents are not device or network-specific but
employed with some intrusive instrument. For instance, using commonly used devices and networks are: -
Password or PIN alongside the biometric may enhance the
security of the device. 1) Mobile Phones: In the current decade, the smart
Morris studied on merging qualities of the three different phone usage has increased significantly as these have become
approaches including, hand-written signature, biometrics , a cheap and convenient source of various services such as
face recognition, and voice based systems. Consequently, the sharing and storage of data, online / offline entertainment and
author reports authentication accuracy of smartphones which games, internet services, phone calls and numerous other
is acceptable for a large variety of applications [21]. applications. Though smartphones provide diverse services,
they are burdened with challenges; security and privacy being
Reference [22] state that accessing the network, making
user registration is allowed by SIM. As without a SIM and most important. Most of the smart phones’ operations are
successful verification, mobile phones have no access, so the associated with internet, so security and safety of data is
traffic on the infrastructure of these mobile phones cannot always on the risk. Passwords, PIN and biometric recognition
inject. Thus, the enabling SIM is first utilized to conduct an systems are being employed for smart phone authentication,
attack and afterwards the usage of such a device by a naïve however these measures do not remain reliable with a higher
user makes them involved in the attacks being botnet nodes of ratio because of guessing and brute forcing can easily counter
the network. these measures[28].
2) Data Protection and Privacy: Authors in Ref [23],
studied data protection issue from the perspective of users and Considering the vulnerabilities and cyber security threats
associated with the use of smart phones, there are a number of
investigated need for data security systems for smart devices.
ways of cyber-attacks; a lot of viruses, Trojans and various
Types of the data that are aought to be protected by the users malware based on smart phone APIs’ operations are
are delineated. Also, the current users practices for data continuously being developed and mostly they appear as if the
protection are also examined. On the basis of this thorough user is using a safe software and hence go undetected. Some
analysis they showed how different types of data require of these application based tools gather sensitive information of
different security arrangements. The authors have reported the users for example geo-location without knowledge of the
the results of an interview based study conducted on a 22 users through GPS based services in smart phone[29][30].
individuals which concludes that although all of the users Many smart phones operating systems have been
wish to protect their data, however, the availbe solutions were developed such as Microsoft Windows, IOS, Android,
not found convienient by the users, thus highlighting the need Symbian and blackberry. Amongst these Android is the most
for adaptive and efficient solution. common due to improved performance. It is based on Linus
system architect. It’s important to understand that there is a
In his research, Muslokhlove highlights data protection difference between the Desktop OS and the versions created
issues against physical threats. The study summarizes the for smart phones. Smart phones enable instant internet
user’s requirements for data protection based on surveys and connection and communication with friends or gather data and
interviews. Requirements for data protection are highlighted information. As the powerful hardware coupled with advanced
after interviews and surveys. It concludes that recognizing the operating systems and smart applications increase the
malicious approaches for data access do not cover the capabilities and functionalities of the smart phones, the users
requirements of data security and thus remains a number of become more vulnerable to data security threats since the
vulnerabilities [24]. However, authors in Reference [25] weaknesses of these systems are continuously being identified
believe that privacy of smart devices and the users under their and exploited by cyber criminals. These attacks include both;
designed framework is protected through semantic policies the ones generated by software developer itself and the ones
dependent on settings and privacy of users. most commonly launched by external agencies in the form of
external attacks[31].
Kodeswaran et al. [26] present a framework for
implementation of privacy policies in smart phones as well as Similarly, features including broad band width, multiple
for data protection. The authors show their framework interfaces and internet accelerators etc., become a source of
designed for Android based devices and the results viruses spread on the network. Multiple connectivity is
accomplished through that framework. associated with high risks and makes the transmission of
viruses easier which may further aggravate the threats. Thus,
III. ANALYSIS : ASSESSMENT OF VULNERABILITIES security challenges for smart phone environment may be
AND RESULTING ATTACKS considered similar to those observed in personal computer
In this section, we first identify the main sources of environment. “Threat means possible breach of smart phone’s
vulnerabilities associated with the telecom sector, then discuss security” [32][33]. In view of above discussion, smart phone
the major vulnerabilities extracted out of the literature and security issues may be summed up into undermentioned
categories;

Seventh International Conference on Aerospace Science & Engineering 127

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

 Authentication problems 3) Unsecure networks: WIFI technology is being used

most commonly by a large number of users. Similarly, other
 Data protection and privacy
services such as GPS, Cellular Network and Bluetooth are
 Vulnerabilities also being used to connecting to internet or any network.
 Cyber Attacks There can be two possibilities; either some network device
hacker can steal network data packets or the telecom company
2) Internet Services: As a result of rapid adoption and itself can ‘sniff’ into the sensitive information. So it is a
application of internet-based services in pursuit of increased vulnerability which the end user can overcome by using
interactions / connectivity, mobile companies are encryption and related tools in their communication and data
continuously transforming their majority of the services to / information transfer [37][9].
online. Implementation of internet-based services to such an
extent has resulted in an increase in the possibility of security 4) Lack of user’s awareness: A potential vulnerability
breaches and data thefts despite of the fact that a high-end associated with the telecom companies’ operations is the lack
security measures are also being employed by these mobile of awareness of the user to security which must be of prime
operators. Again, the systems are vulnerable to attacks from importance. When the user gets connected to network
internal as well as external sources [34][13]. services provided by the company for downloading data or
installing any unknown application, the save button may be
3) Digital Applications: In a developing country like linked to some malicious codes thereby being a cause of
Pakistan where majority of telecom industry are employing external security breach or otherwise, the sensitive
the modern IT services extended by telecommunication information passed on to the network for this purpose may be
companies into their infrastructure and the higher authorities hacked by the service provider [38][15].
are also taking keen interest to use these technologies. These
digital applications lead to severe security breach of Various triggers are set which automatically detect if the
customers and mobile operators making these companies not required data is being transmitted to steal it for the application
secure for their customers. developer or the telecom service provider. Activating wireless
or Bluetooth can be executed secretly without an indication to
B. Major Vulnerabilities Associated with Telecom the user. So, this is a critical vulnerability associated with
Companies telecom sector which can only be overcome with the users’
There are several cyber security attacks and vulnerabilities awareness and employing strict security measures [39][6].
associated with mobile phone companies operating in
C. Vulnerabilities And Attacks On Smartphones Associated
Pakistan. The vulnerability can cause insecurity or make the
With Telecom Service Providers’ Operations
companies a victim to malicious attackers for creating the
attacks. These vulnerabilities include the following: - In previous sections we have discussed the sources of
attacks on users’ data and information and the vulnerabilities
1) System fault/ defects: This is implausible for any associated with the telecom companies’ operations which
telecommunication company to circumvent defects related to when coupled with the identified sources become a cause of
software and hardware. Some companies have faults in their security breach by cyber criminals. All computing devices and
mobile phones. Most of these defects only reveal once the smart devices are prone to these attacks. These attacks can be
device is put in use with some of these observed earlier and broadly grouped into; old attacks and new attacks.
some identified later. Though software related defects can be 1) OLD ATTACKS
rectified easily however hardware defects can cause
a) Malware: Companies and cyber criminals may
irregularities and rectification is possible through change of launch malwares which attack the users’ device by creating
hardware or the device architechture. Exploitation of these an application and providing it for download. Malware is a
defects by attackers thus remains inevitable to initiate the critical security threat which is slowing the development of
attacks on smart phones as well as central data bases[35][15]. wireless applications. It may result into crashing of the data
once it is accepted or installed by the user.
2) Insufficient management of applications: Smart b) Radio and Wireless Network Attacks: Services
devices utilizing the services extended by the provided by the telecom sector can become prone to the
telecommunication companies employ a distinctive feature of wireless communication intruders who, then, can simulate
utilizing flexible APIs in order to develop a number of and launch wireless network based attacks which can be
applications. Nevertheless, if these APIs are not managed classified as:
properly, it can become a main reason for a variety of cyber  Active attacks. These include corruption, spoofing,
crimes. Therefore, mismangement of API appears to be a blockage and modification techniques for data theft.
common reason for code based malicious infections.
 Passive Attacks. Passive attacks include
APIs can be categorized as Third-Party App Development eavesdropping and sniffing.
Open APIs, and Control APIs. Amongst the listed The most common technique employed by the intruders or
classification of APIs, controlled APIs generally have higher the telecom company onto the vulnerable user in this category
privileges in terms of updating the system, creating files and of attacks is detecting the information by listening to the
destroying and fetching data. If somehow, the attackers are communication and data transfer while being broadcasted
able to gain control of these APIs of the companies, attacks wireless through the use of malicious nodes. Sometimes a fake
can easily be initiated and cyber criminals can exploit the Wi-Fi network is created by the intruders to which other users
privileges that these APIs can extend to them [36][5][11]. connect and fall a victim of data theft [40][13].

Seventh International Conference on Aerospace Science & Engineering 128

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

So, it is vulnerability associated with services provided by c) USSD Attacks: USSD stands for Unstructured
the Telecom Company which can be dealt by generating Supplementary Service Data and serves as a protocol utilized
encrypted communication and beware of what networks our by the operators to run specific functions between user and
devices are connected to. the operator which include functions such as credit inquiry,
c) Virus: Viruses are software generated by the sending a prepaid call back etc. USSD mainly consists of
attackers that infect boot sectors, executable files and even these components: interceptor service, permission testing,
the common files for instance word documents and PDFs etc. main activity and boot service.
Virus actually makes duplication of files which consume Smart phones are vulnerable to mainly two types of USSD
system’s capacity and capability. They also link the device to based attacks by telecommunication operators. The first one is
unknown source like installation software. sending an SMS and push notifications network to snip credit.
The second one uses the mobile dialer application utilizing
Users remain vulnerable to viruses. A system that can
USSD protocol being in the background [41][5].
detect and alert smart phones user regarding this vulnerability
has been introduced. Using such systems which can study d) USB Connection Attacks: Weaknesses can be
unusual behaviors of the device as well as communication exposed in the service provider’s system that may be used to
actions taking place within the device to create an alert for the pose the vulnerability. Advanced features may be intended to
users can help the users to save them from falling a prey to this configure and control device.
vulnerability.
e) Denial of Service Attack: This category of attacks is
d) Worms: These are the programs created by the usually called DoS attack as it prevents the use of service by
attacker that replicate the copies from one device to others by operation of data management protocol connection in order to
employing various transport mechanisms throughout network locate the smart phone having Wi-Fi antenna. The attacker
requiring no interference of the user[7][37]. creates a false eye Wi-Fi internet access with the help of a
e) Trojan: Trojan is mostly some useful program that device such as laptop, then asks for a Wi-Fi smart phone to
has hidden malicious functions. Intention is to sneak in terminate the existing mobile broad band connection and link
system by breachng authentication requirements of the to the created bogus Wi-Fi connection. Once connected, it
administrators. Concerns regarding the security threats refrains the targets to connect to other internet links till target
pertaining to smart phones users’ security are growing as the user identifies the malicious activity and disables the Wi-Fi.
mobile companies are becoming more complex and dominant Even the most popular smart phones with IOS and Android
for provision of a variety of functions. They are becoming are also susceptible to denial of service [42][6]
susceptible to the Trojans and viruses of the same class as of f) Camera based Vulnerabilities and Attacks: Most
personal computers[5][13]. of the smart phones have camera and touch screen as essential
f) Spam: Spam is a malware attachment that can be features. Attacker and service providers can use these features
appended to the digital messages conveyed to the smart as a tool to attack on the smart phone. Installation of third-
phones via electronic mail etc. The attachments appear party applications from Application stores makes the users
normal but when a user opens such an attachment, his device vulnerable to attackers. Users fall a victim to the attacks
gets infected by malicious software such a Trojan or worm. which intend to steal personal information, data and even
Smartphone zombies are manipulated by the cyber criminals establish root access to the smartphone or device.
who send junk messages which are used a porthole by the g) Control Flow Attacks: Control flow and
attackers for security breach of smart phones. runtime attacks include injection of milacious codes or return
2) NEW ATTACKS oriented programming and are one of the most severe threats
and critical vulnerability coupled with the services provided
a) SMS Based Attack: Another vulnerability associated
by telecom companies. Being common attacks, they have
with services provided by the telecom companies is that a
been applied to smart phone applications downloaded by
cyber criminal advertises and distributes phishing linkages
many users [43][5][31].
through SMS. They may also use text messages to feat
vulnerabilities. However, applications can be designed to IV. CRITICAL ANALYSIS AND RECOMMENDATIONS
detect this malware and avoid abuse by this vulnerability.
In today’s world of Information Technology, data privacy,
vulnerabilities and attacks caused by these vulnerabilities are
Telecom operators provide a service to their clients to the key issues related to digital security both for the service
transfer credit / unit through SMS. Subsystem “permission” providers, mobile stations and the end users. A good and
and ordering mechanism helps establishing haven for SMS feasible solution to the burning issue of authentication
based malware and gives them a complete authority over problems can save the users from security attacks. As every
receipt, transmission and hiding of SMS messages. Therefore, user wishes for data protection, therefore data privacy is
malicious confirmations may be kept hidden from telecom another key apprehension of clients. Though a considerable
operators by the application and have the potential to affect a breakthrough into the data privacy issue can chiefly be
large set of users and telecom operators. achieved by employing trusted sources and applications, the
b) Escalation Attack: These attacks aim to acquire root cause of most of the attacks is associated vulnerabilities.
complete access to all the information in a particular device. If these vulnerabilities are addressed, most of the attacks on
They may be avoided by employing soft wares that have a smartphones and digital data can be saved. In rapidly
scheme to detect and prevent these attacks. If the system or advancing Information Technology world, achieving 100 %
data is called by special components of Andriod based security is difficult, however, continual design and
systems, these programs prevent it from happening and thus development of the digital security systems keeping in view
the associated vulnerabilities can lead to secure mobile
detection and blockage of the malware is possible.
stations and networks.

Seventh International Conference on Aerospace Science & Engineering 129

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

The root cause of most of the attacks is vulnerabilities investigates vulnerabilities associated with telecom industry
associated with telecom companies’ operations. Based on the operating in Pakistan and attacks that can occur in client’s
study following recommendations are made : - smart phones as a result of these vulnerabilities. Further,
sources associated with these vulnerabilities, their broad
(a) Mobile companies must introduce active firewalls. classification and smartphone vulnerabilities associated with
(b ) Encryption and authentication algorithm for telecom service providers’ operations have been discussed.
communication to be improved. Vulnerability can cause insecurity or make the companies a
victim to malicious attacks. We have also identified and
(c) Techniques to avoid denial of service attacks to be categorized these attacks focusing on why they occur, their
introduced. effects, possible solutions and the contributing internal and
(d ) Proper authentication mechanism to be adopted at external factors. Consistent design and development of the
mobile switching centers digital security systems keeping in view the associated
vulnerabilities, however, can lead to secure mobile stations
In view of above stated recommendations, the general and networks.
cyber-security objectives for ICT networks may be : -
(a) Access to, and use of networks and services to be
restricted to only authorized users. REFERENCES
(b) Only authorized users to access and operate on assets [1] Hamid, Bushra, N. Z. Jhanjhi, Mamoona Humayun, Azeem Khan, and
Ahmed Alsayat. "Cyber Security Issues and Challenges for Smart
they are authorized to access. Cities: A survey." In 2019 13th International Conference on
(c) Networks to support confidentiality to the prescribed Mathematics, Actuarial Science, Computer Science and Statistics
(MACS), pp. 1-7. IEEE, 2019.
level in the security policies.
[2] Telecommunication Standardization Bureau, ITU, “Security in
(d) All network entities to be held accountable for their Telecommunications and Information Technology”, September 2015.
own actions. [3] N. Yildirim, R. Das, and A. Varol, “A Research on Software Security
Vulnerabilities of New Generation Smart Mobile Phones,” in 2nd
(e) Networks to be protected against unsolicited access. International Symposium on Digital Forensics and Security
(ISDFS’14), 2014, pp. 6–16.
(f) Plans to be set in place to address how security [4] A. Jawad and M. Shahzad, “Threats of Cyber Security and Challenges
breaches are to be handled. for Pakistan”, Institute of Information and Communication
Technology, University of Sindh, Jamshoro, 76200, Pakistan,
(g) Procedures should be there for restoration of normal Conference Paper · March 2016
operation once a security breach is detected. [5] A. Awais, H. shabbier, “Pakistan Telecom Sector Overview”
Haidermota BNR, Insight, February, 2017.
(h) The network architecture is to be designed in such a
[6] "Telecom Indicators| PTA"
way so that it may support different security policies and (http://www.pta.gov.pk/index.php/en/telecom-indicators/1).
various up gradations of different strengths in security www.pta.gov.pk. Retrieved 2018-11-15.
mechanisms. [7] R. van der Meulen, “Gartner says 821 million smart devices will be
purchased worldwide in 2012; sales to rise to 1.2 billion in 2013,” 2012.
A. Research Limitations
[8] R. van der Meulen and J. Rivera, “Gartner Says Smartphone Sales
Though extensive efforts have been made to conduct the Grew 46.5 Per Cent in Second Quarter of 2013 and Exceeded Feature
study and to explore the existing literature as well as telecom Phone Sales for First Time,” 2013.
infrastructure, however, the study posed a few limitations such [9] “Global Smartphone Use Continues to Climb, Studies Show.” [Online].
as (a) Limited literature particular to Pakistan’s telecom Available: http://www.eweek.com/mobile/global-smartphone-use-
continues-to-climb-studies-show.html. [Accessed: 30-Oct-2015].
industry. (b) Lack of access as well as information sharing by
telecom companies due to sensitive and confidential nature of [10] H. Luo, G. He, X. Lin, and X. Shen, “Towards hierarchical security
framework for smartphones,” in Communications in China (ICCC),
the data. (c) Vulnerabilities present the weakness of service 2012 1st IEEE International Conference on. IEEE, 2012, pp. 69–72.
providers and security providers. Moreover, they give a [11] M. Ahmad and N. Musa, “Comparison between android and iOS
thorough insight into the technical infrastructure of these Operating System in terms of security,” in Information Technology in
companies which made the research dominantly qualitative Asia (CITA), 2013 8th International Conference on. IEEE, 2013, pp.
literature review. 1–4.
[12] M. Goadrich and M. Rogers, “Smart smartphone development: iOS
B. Ethical Considerations versus Android,” in Proceedings of the 42nd ACM technical
This research is primarily of qualitative nature and symposium on Computer science education. ACM, 2011, pp. 607–612.
primary as well as secondary sources of information have been [13] L. Ma, L. Gu, and J. Wang, “Research and Development of Mobile
Application for Android Platform,” Int. J. Multimed. Ubiquitous Eng.,
used with a main chunk of information gathering from primary vol. 9, no. 4, pp. 187–198, 2014.
sources; wherein the primary sources being the existing [14] J. Liu and J. Yu, “Research on Development of Android Applications,”
literature and conducted surveys and studies. These have been in Fourth International Conference on Intelligent Networks and
properly cited wherever used throughout the research. Intelligent Systems. IEEE, 2011, pp. 69–72.
Similarly, the secondary sources remained the professionals [15] T. Grønli and J. Hansen, “Mobile application platform heterogeneity:
and experts of the telecom industry who were ensured that Android vs Windows Phone vsiOSvs Firefox OS,” in Advanced
their identity would not be shared with anybody and any type Information Networking and Applications (AINA), 2014 IEEE 28th
International Conference on. IEEE, 2014, pp. 635–641.
of information related to their maintenance setups would be
kept confidential. [16] 14. D. Tilson, C. Sørensen, and K. Lyytinen, “Change and control
paradoxes in mobile infrastructure innovation: the Android and iOS
mobile operating systems cases,” in System Science (HICSS), 2012
V. CONCLUSION 45th Hawaii International Conference on. IEEE, 2012, pp. 1324–1333.
In this paper, we first highlight the recent authentication [17] W. Lee and R. Lee, “Multi-sensor authentication to improve
issues, data protection and the evolving privacy problems vis- smartphone security,” in Conference on Information Systems Security
à-vis presenting the telecom infrastructure. The study and Privacy., 2015, pp. 1–11.

Seventh International Conference on Aerospace Science & Engineering 130

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
[18] S. Zahid, M. Shahzad, S. Khayam, and M. Farooq, “Keystroke-based [31] D. Padmavathi and M. Shanmugapriya, “A survey of attacks, security
user identification on smart phones,” in Recent Advances in Intrusion mechanisms and challenges in wireless sensor networks,” arXivPrepr.
Detection. Springer Berlin Heidelberg, 2009, pp. 224–243. arXiv0909.0576, vol. 4, pp. 1–9, 2009.
[19] T. Feng, X. Zhao, B. Carbunar, and W. Shi, “Continuous Mobile [32] M. A. Dar and J. Parvez, “Smartphone operating systems: Evaluation
Authentication Using Virtual Key Typing Biometrics,” in 2013 12th & enhancements,” in 2014 International Conference on Control,
IEEE International Conference on Trust, Security and Privacy in Instrumentation, Communication and Computational Technologies
Computing and Communications, 2013, pp. 1547–1552. (ICCICCT), 2014, pp. 734–738.
[20] C.-C. Lin, C.-C. Chang, D. Liang, and C.-H. Yang, “A New Non- [33] J. H. Choi and H.-J. Lee, “Facets of simplicity for the smartphone
Intrusive Authentication Method Based on the Orientation Sensor for interface: A structural model,” Int. J. Hum. Comput. Stud., vol. 70, no.
Smartphone Users,” in 2012 IEEE Sixth International Conference on 2, pp. 129–142, Feb. 2012.
Software Security and Reliability, 2012, pp. 245–252. [34] W. Lee and R. Lee, “Multi-sensor authentication to improve
[21] A. Morris, “Multimodal person authentication on a smartphone under smartphone security,” in Conference on Information Systems Security
realistic conditions,” in Defense and Security Symposium. and Privacy, 2015, pp. 1–11.
International Society for Optics and Photonics, 2006, p. 62500D– [35] Reddy, G.N. & Reddy, G. 2014, "A Study of Cyber Security Challenges
62500D. and Its Emerging Trends on Latest Technologies", arXiv preprint
[22] N. Gobbo, A. Merlo, and M. Migliardi, “A denial of service attack to arXiv:1402.1842, .
GSM networks via attach procedure,” Secur. Eng. Intell. Informatics. [36] Sebastian Bortnik 2012, Trends for 2013: astounding growth of mobile
Springer Berlin Heidelb., vol. 8128, pp. 361–376, 2013. malware, WeliveSecurity:Security news, views and insight from the
[23] I. Muslukhov, Y. Boshmaf, C. Kuo, J. Lester, and K. Beznosov, ESET experts, ESET Latin America.
“Understanding Users’ Requirements for Data Protection in [37] Ten, C., Manimaran, G. & Liu, C. 2010, "Cybersecurity for critical
Smartphones,” in 2012 IEEE 28th International Conference on Data infrastructures: attack and defense modeling", Systems, Man and
Engineering Workshops, 2012, pp. 228–235. Cybernetics, Part A: Systems and Humans, IEEE Transactions on, vol.
[24] I. Muslukhov, “Survey: Data Protection in Smartphones Against 40, no. 4, pp. 853-865.
Physical Threats,” in Term Project Papers on Mobile Security. [38] Threat Track Security 2014, December 12, 2014-last update, Spammers
University of British Columbia., 2012. Accelerate Dyre Distribution [Homepage of Threat Track
[25] D. Ghosh, A. Joshi, T. Finin, and P. Jagtap, “Privacy Control in Smart Security],[Online]. Available: http://www.threattracksecurity.com/it-
Phones Using Semantically Rich Reasoning and Context Modeling,” in blog/ [2015, January12, 2015].
2012 IEEE Symposium on Security and Privacy Workshops, 2012, pp. [39] Z. Farhan, M. Shah, K. Muhammad, J. Qiaser and Z. Sejing, “A Survey
82–85. on Security for Smartphone Device”, International Journal of
[26] P. Kodeswaran, V. Nandakumar, S. Kapoor, P. Kamaraju, A. Joshi, and Advanced Computer Science and Applications, Vol. 7, No. 4, 2016.
S. Mukherjea, “Securing Enterprise Data on Smartphones Using Run [40] Telecommunication Standardization Bureau, ITU, “Security in
Time Information Flow Control,” in 2012 IEEE 13th International Telecommunications and Information Technology”, September 2015.
Conference on Mobile Data Management, 2012, pp. 300–305.
[41] A. Agrawal and A. Patidar, “Smart Authentication for Smart Phones,”
[27] V. Remenar, S. Husnjak, and D. Peraković, “Research of Security Citeseer, vol. 5 (4), pp. 4839–4843, 2014.
Threats in the Use of Modern Terminal Devices,” in 23rd International
[42] P. Schulz and D. Plohmann, “Android security-common attack
DAAAM Symposium Intelligent Manufacturing & Automation: Focus
on Sustainability, 2012. vectors,” in Rheinische Friedrich-Wilhelms-UniversitatBönn,
Germany, Tech. Rep, 2012.
[28] Maji and K. Hao, “Characterizing failures in mobile oses: A case study
with android and symbian,” in Software Reliability Engineering [43] C. Pettey and R. van der Meulen, “Gartner says worldwide sales of
(ISSRE), 2010 IEEE 21st International Symposium on. IEEE, 2010, pp. mobile phones declined 3 percent in third quarter of 2012; smartphone
249–258. sales increased 47 percent,” Gartner),[Online], 2012.
[29] O. Riva and C. Qin, “Progressive Authentication: Deciding When to
Authenticate on Mobile Phones.,” in Presented as part of the 21st
USENIX Security Symposium (USENIX Security 12)., 2012, pp. 301– .
316.
[30] F. Al-Qershi, “Android vs. iOS: The security battle,” in Computer
Applications and Information Systems (WCCAIS), 2014 World
Congress on. IEEE, 2013, pp. 1–8.

Seventh International Conference on Aerospace Science & Engineering 131

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Sentiment Analysis of Social Media Text for

Selective World Events and Geographical
Segregation
Muhammad Ahsan khan
Institute of Space Technology
Islamabad, Pakistan
[email protected]

Abstract— This document covers the domain of natural practices in machine learning/deep learning in text
language processing and the use of natural language processing classification, so we will put forward the popular neural
(NLP) to make a filter on social media (in our case Twitter) text architectures we came across in our literature review and
which in return will give us feedback about how a specific describe them briefly theoretically and practically.
topic, entity or scenario is seen or felt by different and huge
population. In this paper we are trying to apply sentiment II. WORD RERESENTATIONS
analysis on Twitter. Our final model is going to give us amount
of positive and negative tweets about a given input word that A. Word Vectors
we are interested in, such as, event, entity etc. This paper will
go briefly through popular neural networks that have been Since words are unstructured data, before 2013, generally
used till 2018 for sentence classification. This paper is not words were represented as one-hot encodings and that was a
explanation of all the neural network architectures such as massive flaw in the development of neural architecture, as
those found in surveys rather it will share the author’s our input vector of words would not have semantics’
thoughts regarding implementation of Sentiment Analysis information in its vectors. The problem was two folds. First
based on practical constraints. was that one-hot encodings (each word’s vector length
equaled to the number of words in the vocabulary) had a
Keywords— NLP, Artificial Intelligence, Text Classification sparse vector of zeros which carried no meaning in the first
place. Other problem was the complex computations done in
I. INTRODUCTION neural network which did not favored/allowed computation
Sentiment analysis (SA) is a technique in which the aim due to constraint of computation power by high-end systems
(obvious by its name as well) is to detect/know sentiments back in the days.
of individual(s) regarding anything that may be a person, In 2013, Google research team made a ‘bottle neck’ solution
place, scenario etc. [2]. In this study we will divide for obtaining a lower-dense vector representation of every
individual’s emotions in two categories, that are positive and word in the vocabulary [1]. The concept is equivalent to
negative. ‘Auto-encoders’ neural architectures in which the whole aim
SA helps us to have insights on large amount of audience, is to find a small dimensional vector representation of a
say for example Apple users all from around the world, given input.
which were 1.4 billion in 2020[11]. And hearing views from Google team published two papers in the year 2013, both
all of 1.4 billion people is impossible for a human or finding lower dimensional vectors which are dense
organization. So, we train our system to diagnose the representations.
audience and in return hope that it gives us the crux of what The whole intuition behind these architectures is to develop
are the demands/wants/sentiments of individuals. a way in which a word can be expressed in regards to its
The concept of our study lies in the fact that if those users surrounding/context words. If the word gets expressed in
are kept under consideration and a mechanism is engineered this manner, then it is likely that the word’s vector is
which filters this enormous useless data (comments) into enriched with its semantic nature. This idea also suffices the
useful data (which our machine can understand), then a idea of how we (humans) interpret language/words because
machine can quickly analyze the data and can give us how according to a famous American linguist, John Firth, whom
the audience are seeing a particular thing/character/scenario says, “You shall know a word by the company it keeps.”
etc. By this, we can save an enormous amount of time and These architectures have been implemented on very large
can quickly have our analysis which can prove very corpus which has 100 billion of words [3]. This huge
beneficial for an organization/company/institution/person. amount of data has given us an opportunity to gather the
This paper aims to analyze Twitter data. For this analysis, ‘connections between the words’ or semantics of all words
users from United States of America (USA) are kept under via just few equations.
consideration. The final shape of analysis will be to have a 1: Continous Bag Of Words (CBOW)
GUI (map) that will show the amount of percentage of
In this a context is given as an input and a two-layer neural
tweets that are positive and negative about any given input
architecture tries to predict the specific word from whose
entity/scenario etc. with respect to the States. Since our
surroundings the context was taken. In Figure 1, left side,
work’s genre is purely based upon research of past popular
shows a representation of this architecture taken from their

Seventh International Conference on Aerospace Science & Engineering 132

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

paper [1]. In the input, one-hot encoded words are added and Fig. 2. Formula for negative-sampling (from the paper [2])
linearly transformed in the first hidden layer and then the
hidden layer is linearly transformed again to an output layer III. NEURAL ARCHITECTURES
which is followed by a hierarchical soft-max (nonlinear
function). In the soft-max layer, theoretically, we are A. Convolutional Neural Architecture
intending to maximize the probability of the required word
In 2014 a one-layer architecture of CNN was proposed by
(in one-hot encoding) among all words in the vocabulary.
Yoon Kim [3]. The simplicity of the architecture with
After training/back-propagation, the weights between
promising results is the main hype of this paper. It was one
hidden/projection layer and input layer, and projection layer
of the first paper that used semantic representation of words.
and output layer are added or concatenated. And these are
Fig. 3 shows its accuracy on different datasets.
the new representation of the output word. If the same word
In this input word vectors of each sentence were given and
comes again as an output (in future iterations) then the
100 filters each of size 2,3 and 4 were applied. It is then
resultant matrices are added to its previous dense vector.
passed through a Max-pooling layer which takes out the
most significant value out of the vector/matrix. After that
they are passed through a fully connected and sigmoid layer.
The diagram is shown in Fig. 4.
Though, its results are promising in regards to its simplicity
but this architecture is not fit to be used in NLP because this
architecture do not full fill the intrinsic property of how a
sentence must be read as we humans do. Generally, this
architecture is finding a sequence of words in our training
data that generally corresponds to our binary classification.
It is analogous to how CNN first detect edges in a picture in
Fig. 1. CBOW (left side) and Skip-gram (right side) its early layers and then start to detect minute details such as
eyes etc., as the picture go deeper in the network. We can
generally say that though the architecture has learned
2: Skip-gram general combination of words but this does not mean that its
This model is the opposite of CBOW. Here we give an input ‘general word’s combination’ will work on Test data
word and the neural network tries to predict the surrounding because the amount of vocabulary and the style of
words, it is shown in Fig. 1, right sided. In this research interpretation is quite vast in human beings.
paper [2] they have implemented everything similar to
CBOW except that the inputs and outputs are now
completely opposite in number of words, and a new concept
called ‘Negative sampling’ is also introduced. Previously
[1], in the computation of soft-max, the algorithm would Fig. 3. Datasets accuracy from the paper
take out soft-max of all vocabulary which was very
computationally intensive as vocabulary would reach in
millions [1]. In Negative-sampling (NS) [2], the algorithm
deals this task as supervised classification task rather than
unsupervised approach in [1].
For NS, 2 sets are generated. One has proper context and
word pairs. The other set has context and word pairs that do
not occur in the corpus (they are random pairs). Once
context and word are considered from the corpus, the
algorithm randomly selects other words that are not pairs.
The randomly selected words range from 2 to 5 for large
Fig. 4. Architecure of Yoon-Kim’s CNN model(taken from its paper [3])
corpus and 5-20 for small corpus [2], these numbers have no
proper mathematics behind them as they are taken out by hit
and trial method. B. Recurrent Neural Network Architectures (RNN)
So, what it means is that now the neural network has to just We humans absorb a text document or a sentence in a
change and update weights for the word(s) that are actually sequential manner because we generally go from the first
the answer(s) and for the random samples. Through this word to the last word of the sentence and try to find out the
approach, computation power is saved hugely at every meaning of the sentence. The meaning of the sentence in our
iteration because before NS, all of the probability of words brain evolves from time to time (from start to finish) while
were computed (using soft-max function), and now through we are reading the text. This sequential manner is sort of
NS, only word(s) plus ‘k’ number of bad/incorrect pairs are mimicked in RNNs, as inputs are computed in a sequential
computed. The formulae which the paper gives for log- manner.
likelihood is given in Fig. 2. RNN is a dynamic neural network which means that it
changes its hidden state with respect to input. Theoretically,
it is used in places where a memory or sequential/temporal
information is required. Every time step learns an input
given at that instant in regards to the previous given
inputs/temporals. It is shown in equation 1 and diagram is

Seventh International Conference on Aerospace Science & Engineering 133

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

shown in Fig. 5. Equation 1 gives the mathematical

interpretation of maximum likelihood. In this ‘x’ terms are
the inputs words. Equation 2 represents a generic
computation done at each temporal step in which ‘g’
represents a nonlinear function, x(t) represents input word,
a(t-1) represents information from the past temporals, rest of
terms are the learning weights. The other most popular
architectures used in NLP in sequence modelling (other than
RNN) are Long Short-Term Memory (LSTM) [6]and Gated
Recurrent Unit [4] which will be discussed later.
The drawback of this (RNN) architecture is the inability to
capture long-term-dependencies due to vanishing gradients
and exploding gradients [5] as either the architecture does Fig. 5. Basic RNN architecture. Our analysis result will be only on the last
not learn or resist to learn data respectively. During back output. The other outputs are shown for the reader’s understanding to let
propagation, the inputs/weights of inputs that are far away them know that we can take output at each temporal (if we want to).
from required loss functional node tend to produce either no
effect or extreme high effect thus making our learning C. Long Short Term Memory(LSTM)
extremely slow or impossible to learn respectively. This is There are other solutions to the vanishing and exploding
due to the derivative’s contribution at each time step, gradient problem rather than introducing new architectures
because when moving from right to left (in back such as gradient clipping and using skip connections in
propagation) each temporal step requires its derivative to get scenario of gradient exploding and vanishing respectively.
multiplied on its temporal step before going further to the In gradient clipping we just deduct a user defined value from
left (chain rule in derivative). the slope (gradient) if our gradient gets larger from a
Equation 3 shows one derivative effect of the previous specified value in order to retain our learning rate rhythm
temporal towards the present temporal. One can see that with respect to our loss curve [4] [5], while in skip
how weight is multiplied at the very end. The same connection we add a direct input from past layers which
multiplication process is done for other previous temporals makes the past layer(s)’s information directly accessed by
which means that same weight will be multiplied, but now, the future time steps [12]. The other approach to overcome
with the power of the number of previous temporals from the gradient problem is to devise a new neural architecture.
which we are calculating their effect on the required LSTM [6] uses gates to control and propagate the flow of
temporal (which is the last layer in our case). information. It has Input gate, Forget gate, Output gate.
The weights of all temporal steps are the same (RNN has Equations 4,5,6,7,8 and 9 are showing the processes done at
this property because we are trying to capture general every temporal. LSTM has the advantage to store
representation of data from first temporal to the very end information in each time step and carry it forward in time
temporal), so if the matrix has the values less than 1 then its steps (if it realizes that the information is necessary for
multiplication will make the derivative value close to zero upcoming temporals).
which will make the gradient vanish, On the other hand, we Theoretically, they are designed in such a way that if an
will get a very large value if values of the matrix are greater information is seen important at any time step by our
than 1. machine, then the Forget gate will remember it and carry it
RNNs gets addition in layers as input (word) is incremented. for a longer number of time steps thus ensuring long-term-
This drawback in RNN architecture does not appreciate us dependency capture (which was the main flaw in RNN), or
to opt it as the ‘gradient problem’ can be encountered more vice versa.
because the more the temporal steps i.e., input words in a In the equations the memory is called a Cell. Forget gate (F)
sentence are, the greater is the chance of vanishing or decides what shall be kept from previous cell state or
exploding gradients. changed i.e., updated with new variation. Input gate (I)
To our knowledge there is no research paper which has decides what to write from current time step into the
given analysis about the number of words that would work Cell/memory. Output gate (O) decides what part of the
efficiently on RNNs. On the other hand, we are hoping to overall memory (after addition of the information from the
have long sequenced texts and if those long sequences are present temporal to the cell state) must be written to the
effectively quantified/analyzed then our system would cater output or hidden state. ‘c`’ is the new information computed
all and different kind of Tweets effectively in a practical from the current temporal’s information for memory. On
environment. this intermediate cell state i.e., ‘c`’, we apply Input gate
which takes the necessary information from ‘c`’ vector,
p(x1,…,xT)=p(x1)p(x2|x1)p(x3|x1,x2)p(xT|x1…xT-1) (1) which then gets added to the memory/Cell. ‘C’ is the overall
combination of previous cell state and current cell state,
a(t) = g(Waaa(t-1) + Waxx(t)+ba ) (2) after the application of gates. ‘H’ is the output given to the
next time step which is taken from ‘C’ or cell state which it
d(a(t) )/d(a(t-1) )= (inner function derivative) *Waa (3) thinks is necessary for the next temporal’s computation. ‘g’
corresponds to sigmoid function and the rest of the variables
are learning weights.
One can think of it as equivalent to conversating with a
person. In which one does not tell everything (what the
individual has stored in his mind while hearing the other

Seventh International Conference on Aerospace Science & Engineering 134

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

person) every time whenever they speak to the other person, of semantic nature really makes it hard on sequence models
rather, the individual just decides what information is to make them understand on a test set. As one sentence can
necessary to tell from all of the information which the be written in multiple ways reflecting the same meaning, so
individual has stored in mind. intuitively, it seems to us that there is less chance of making
Here contrary to RNN, we will not have the problem of our machine understand well as compared to pictures or
vanishing or exploding gradient because none of the weights videos where clues like edges and borders tend to be same
will be repeated with the same value. Another advantage is regardless of where an object is in the picture (in the case of
that there are more weights, so this gives the architecture CNN) [13].
more expressivity or gives more opportunity to learn the
data at each temporal. U(t)=g(Wuh(t-1)+Uu x(t) +bu) (10)

F(t)= g(Wf h(t-1) + Uf x(t) +bf) (4) R(t)=g(Wr h(t-1) + Ur x(t) +br) (11)

I(t)= g(Wi h(t-1) + Ui x(t) +bi) (5) h`(t)=tanh(Wh(r(t).h(t-1))+Uhx(t)+bh) (12)

O(t)= g(Wo h(t-1) + Uo x(t) +bo) (6) H(t)= (1-u(t)). h`(t-1) + u(t). h`(t) (13)
E. Transformers/Bi-directional Encoded Representations
c`(t)= tanh (Wc` h(t-1) + Uc` x(t) +bc`) (7)
from Transformers (BERT)
C(t)= F(t) . C(t-1) + I(t) . c`(t) (8) Till now our study summed up to three important
constraints. CNN has the ability to capture the sequence for
H(t)= O(t) . tanh (C(t)) (9) both positive and negative polarities but it is limited to the
vocabulary and sequence of vocabulary used in the training
D. Gated Recurrent Unit (GRU) data. Sequential architectures have the problem of vanishing
GRUs [9] are equivalent to LSTM but has a smaller number gradient even in LSTM [16] as well if input length gets
of computations [5]. Here in the explanation, contrary to larger. On the other hand in sequential architectures, there is
Cell state/memory (in explanation of LSTM) we will use the also a lack of parallel computation i.e., ensuring the
term ‘hidden state’. maximum use of GPU for efficient computation.
It has an Update gate (U) which decides what part(s) of These above important constraints were all catered in
current and previous temporal’s hidden states are to be Transformers architecture back in 2017.
retained or deleted or changed. It has a Reset gate (R) which The Transformers architecture [7], Fig. 13 came in 2017 and
decides which information of previous hidden state is BERT [8] architecture came in 2018. The Transformers
necessary to compute the current temporal data. ‘h`’ is new architecture is composed of Encoder and Decoder block.
hidden state of current time/temporal step and ‘H’ is the Our study’s interest is in their Encoder block as BERT
combined combination of previous and current hidden model is established only on Encoders.
states. The general computation for one temporal is shown Encoder block is based on parallel mechanism (which the
in Equations 10,11,12 and 13. sequential neural architectures lacked) that takes in an input
Our analysis on temporal based networks found two flaws. of a sentence and tries to induce information of all the words
According to the paper [5] and to our research there is no in every word with respect to semantic nature and their
concrete evidence (theoretically) in the literature that LSTM respective position in the sentence.
or GRU would suffice the problem of long-term-dependency Difference between Transformers and BERT is generally in
if our input sentence is very long. By ‘very long’ we mean two ways. BERT has an array of Encoder blocks whereas
more than 100 words because we are assuming that our Transformers architecture has arrays of both Encoder and
comments of Twitter are not always supposed to be small in Decoder. BERT takes in information in a sentence in both
length. The other reason is the ‘sequential process’ of all direction whereas Transformers architecture computes
temporal architectures. This means that second input will information from left side only (in translation such as from
start its computation once first input computation is done French to English).
and so on. In the architecture, first comes the Positional Encoding (PE)
The main reason we see for this is that at the time of RNN block which is required to give a ‘signature’ of position to
and LSTM, the world was not familiar with the concept of each word because all of the words are processed parallelly.
parallel processing done with GPUs and researchers did not The formulas are shown in Equations 14 and 15. Equation
keep this constraint in the design of their architectures. We 14 and Equation 15 corresponds to words’ dimensions that
in our research are aiming to take maximum advantage of are at even number and odd numbers respectively. Here ‘i’
GPU and also find an architecture that ensures a better represents the sequence number of the dimension of the
memory based processing of information of a sentence. word’s vector and ‘d’ corresponds to output’s total
One theoretical intuition we found was that though GRU dimension that the architecture produces after performing
tries to mimic the same process of LSTM with less computation. In Transformer architecture [7] they use ‘d’ as
computation, it will have a less accuracy on dataset due to 512.
less weight matrices learning at each time step as we Once the PE is calculated it is then added to the original
generally know that the more weight matrices are, the word’s vector respectively. Through this we infuse the
greater would be the learning of the neural network. position information of every word. Figures from Fig. 6 to
One thing we concluded from our study is that due to the Fig. 9 show the effect of how sequence of positional
vast expressivity of our language, the dense information encoding changes as position of the word gets further from
with respect to temporal (sequence) and the vast vocabulary the start of the sentence. We can see more ‘1’s’ addition to

Seventh International Conference on Aerospace Science & Engineering 135

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

the dimension of the word’s vector as the word gets further dependent on one another, then we concatenate their results
from the starting word. and multiply it by a matrix Wo to bring it down to the
dimension of the words that were given at the input.
PE(pos,2i) = sin ( pos /10000 2i/ d ) (14) After that our resultant matrix goes through position wise
PE(pos,2i+1) = cos( pos /10000 2i/ d) (15) fully connected layer which means each word is linearly
transformed separately in the layer and none of the words
After addition of the sequence information, the next step is are sharing their information with each other.
to multiply each word with a weighted matrix of Query, Key The above-described whole process is replicated 11 times
and Value. These matrices are responsible to convert each more in BERT. For our classification purpose we then added
word (which we gave in the input) to three new and different a classification layer (that is sigmoid layer) which gets us
representations. The intuition behind Query, Key and Value our final output of whether an input sentence is positive or
is the format of a database structure. Key might be a negative.
person’s name and value will be the person’s phone-number. The model we use is BERTBASE model which has 12
Query on the other hand can be information asked regarding Encoder layers, with 768 hidden size and 12 Attention
Key and Value such as whether a specific person name with heads. Its total parameters are 110Million.
golden phone-number is present or not. We used a pretrained model of BERT BASE and fine-tuned it.
Once the multiplication (linear transformation) is done by Its training is done on two tasks. Firstly, it is trained on
the weighted matrices, we will have 3 different Masked language modelling which means that a big
representations of one word. For instance, let us assume we corpus/document is given to BERT and 15% of the words
have 4 words (of input dimensions of 3) in a sentence shown are made hidden/masked from BERT. In training, BERT job
in Fig. 10. As an example, we are considering the first word. is to predict the word that is hidden/masked by looking at
What we do is that we take the first word’s Query vector and the words that are present at either side of the hidden word
have dot product multiplication with all the words’ Key in the sentence.
vector in the sentence (including itself). Which means that The other training approach done on this model is to give 2
first word’s Query vector gets multiplied (dot product) with sentences from a big corpus. The model is then required to
all of the keys’ vector separately. Now, after that when all of predict whether the second sentence is likely to follow after
the dot product multiplications are computed parallelly then the first sentence.
we make them normalize and apply SoftMax function so The miracle of this model is that this paper is the proof that
that our dot products answers are brought in range from 0 to if a big/huge model is trained well enough then it can work
1. Normalization is done by dividing it by the dimension of on small problems in that relevant field with just a little fine-
the vector of key in order to avoid operating very large tunning [8]. A table from BERT paper is given in Fig.15
numbers because there is a possibility that our dot product which shows its performance compared to other neural
might result into a very large number [7]. architectures.
Then we multiply each answer with its own word’s Value The reason to pick BERT is not only for best-accuracy-
vector. From this we see that how much each word was chase rather this algorithm gives us State of the art (SOTA)
contributing individually in regards to the Query vector. results on other 11 NLP tasks [8] as well.
Then after that we add all the vectors of all the words. And
this is now the final/new representation of the first word
(Query word) with respect to all 4 words.
From the above process the first word (Query word) now
has all the contextual/surrounding information, achieved
parallelly. The diagrams of all processes are shown in Fig.
10, Fig. 11 and Fig. 12. Here we considered ‘I’ word from
the sentence ‘I study at school’. The last column in Figure
12 will be summed up to have a contextual representation of
word ‘I’. This above process will be repeated to all the
words and they all will get the same dimensional
representation.
Fig. 6. Sine wave for position 1
The above given process is known as a single ‘Head’ in the
BERT paper. There are a total of 12 heads which means that
there are 12 different Queries, Values and Keys for each
word. It is referred to as ‘Multi-Head’ attention, Fig.14.
The intuition behind it is that since the machine can not
properly concentrate (or figure out similarity of all words
with one another) on each word and find the context which
is necessary or important with in a sentence in one go. So, it
is wise practice to give our machine room to concentrate on
different combination of words in different attention heads.
The reason for doing is to enable our model to detect
different attentions on different things such as nouns with
verbs, nouns with preposition etc.
When all of the attention heads perform their computation
parallelly, because none of the processes (heads) are

Seventh International Conference on Aerospace Science & Engineering 136

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 11. ‘I’ computation with repsect to all words[14]

Fig. 7. Sine wave for position 20 Fig. 12. Soft-max coloumn will now get multiplied with the ‘Value’ vector
of each word. And then all vectors will be added together which will result
in vector representation of the word ‘I’ in that specific sentence[14]

Fig. 8. Cosine wave at position 1

Fig. 13. A Transformer’s Encoder block (taken from original Transformers
paper)

Fig. 14. Left side shows attention mechanism and Right side shows the how
Multi-Head phenomena takes place (taken from original Transformers
paper)

Fig. 9. Cosine wave at position 20

Fig. 15. Accuracy score of BERT on different datasets(taken from original

BERT paper)

Fig. 10. Linear Transformation of each word into Query, Key and
IV. TRAINING STRATEGY
Value[14]
As a social media test experiment, it was hard to choose the
dataset to train/fine-tune our model on. First problem was
the approximate 110 million parameters learning time
because we are training all the weights (due to TensorFlow
documentation), as our one epoch of 200000 sentences took

Seventh International Conference on Aerospace Science & Engineering 137

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

around 2 hours on Google Colab. The other problem was

how to effectively tackle every kind of Tweet as some
Tweets would use standard english, some would use slang
language etc.
We decided to use SENT140 dataset that contain 1.6 million
tweets of binary classification, Movie reviews and SST-2
data set that contains 10662 and 50000 binary classified
sentences respectively.
We took 100000 sentences each from both of the classes of
Fig. 16. A ‘part’ of our full map of United states. One can see the percent of
SENT 140 dataset. We took all of the sentences from SST2 tweets regarding an event from the state of Illinois
dataset for the training. We took 4000 sentences from Movie
Review dataset only (that had equal number of sentences
with respect to polarity). We merged them and shuffled VI. CONCLUSION
them. We also made the combined dataset lower case.
We hope that our paper will ensure to save a lot of time of
For testing, we used the remaining sentences in Movie
the young researchers by giving a bird’s-eye view of popular
Review dataset.
neural architecutures’ theoretical framework with respect to
This strategy was induced to see how robust our proposed
practical constraints. We also showed how a practical
neural model is in practical terms after Fine-tuning. First
application can be made through sentiment analysis i.e.
reason is that these differnet datasets were taken in order to
working on real time Tweets. By our findings we conclude
train our model on different genre of sentences such as
that why it is necessary to study BERT model, additionaly it
personal opinions, accusations, general comments, informal
will also help young researchers to not only have promising
language etc, because this combined dataset will have data
results on sentiment analysis rather on other 11 NLP tasks as
that will be similar to real-time Tweets on which we are
well [8]. On the other hand, BERT model will also help
going to build our application. Second reason is that we are
young researchers grasp the concepts of other architectures
testing on that dataset whose training data is in a low
‘a bit more easily’ that came after BERT such as
proportion in training phase. The test results will depict our
Transformer-XL, GPT-2 and GPT-3 etc because they all use
neural model authenticity of understanding binary classified
Transformers architecuture’s concept in some sort of way.
sentiment analysis. Though all three datasets correspond to
the same classification but we are taking account the
different platforms (twitter and movie reviews) and their REFERENCES
data-collection’s time as well, because there is a chance that [1] K. C. G. C. J. D. Tomas Mikolov, “Efficient Estimation of Word
at different platforms the ways/style of writing can be Representations in Vector Space,” arXiv preprint arXiv:1301.3781,
different, and even the time can also play its part as fashion 2013.
of writing changes [15]. [2] Mikolov, T., Sutskever, I. and Chen, K., 2013. Distributed
Representations of Words and Phrases and their Compositionality.
Thus we belief that our analysis has important meaning in Cornell University.
analyzing how effective the understanding of our model is in [3] Y. Kim, “Convolutional Neural Networks for Sentence
regards to the nature of a sentence because in standard Classification,” Association for Computational Linguistics, vol.
practice of machine learning, we gather a dataset from a Proceedings of the 2014 Conference on Empirical Methods in Natural
certain controlled environment, may that be a specific topic, Language Processing (EMNLP), p. 1746–1751, 2014.
platform, group of selected indivduals, or even the person(s) [4] Chung, J., Gulcehre, C., Cho, K. and Bengio, Y., 2014. Empirical
Evaluation of Gated Recurrent Neural Networks on Sequence
whom is/are collecting the raw dataset(s). What we are Modeling. Cornell University.
trying to point out is that there is a chance that there could [5] Y. Bengio, P. Simard and P. Frasconi, "Learning long-term
be a certain ‘biasness’ of some sort of nature in the datasets. dependencies with gradient descent is difficult," in IEEE Transactions
If these constraints are kept under consideration then it must on Neural Networks, vol. 5, no. 2, pp. 157-166, March 1994, doi:
help our model correctly analyze a sentence in a practical 10.1109/72.279181.
environment. [6] S. Hochreiter and J. Schmidhuber, "Long Short-Term Memory," in
Neural Computation, vol. 9, no. 8, pp. 1735-1780, 15 Nov. 1997, doi:
The test results on this dataset gave us a 84.6 percent of 10.1162/neco.1997.9.8.1735.
accuracy. From this we can interpret that how robust our [7] vaswani, A., Shazeer, N., Parmer, N., Uszkoreit, J., Jones, L., N.
new model is. Gomez, A., Kaiser, Ł. and Polosukhin, I., 2017. Attention is all you
We think that this is a novel idea that we have established in need. nips.cc.
our strategy. We, to our best knowledge do not think that [8] Devlin, J., Chang, M., Lee, K. and Toutanova, K., 2018. BERT: Pre-
training of Deep Bidirectional Transformers for Language
this strategy is used in any of the literature before in the field Understanding. arXiv.org > cs > arXiv:1810.04805.
of sentiment analysis. [9] K. C. B. van Merrienboer Caglar Gulcehre and D. Bahdanau,
“Learning phrase representations using RNN encoder–decoder for
statistical machine translation,” Arxiv.org. [Online]. Available:
V. GEOGRPAHICAL OUTPUT http://arxiv.org/abs/1406.1078v3. [Accessed: 31-Aug-2021].
[10] S. Albawi, T. A. Mohammed, and S. Al-Zawi, “Understanding of a
We used TWINT library in Python to filter out Tweets with convolutional neural network,” in 2017 International Conference on
respect to a keyword, from co-ordinates of States of Engineering and Technology (ICET), 2017.
America. We then used Folium library to create GUI map of [11] M. Potuck, “Apple hits 1.5 billion active devices with ~80% of recent
America. After testing the tweets on our model we then iPhones and iPads running IOS 13,” 9to5Mac, 28-Jan-2020. [Online].
Available: https://9to5mac.com/2020/01/28/apple-hits-1-5-billion-
counted how much percent were tweets postive and negative active-devices-with-80-of-recent-iphones-and-ipads-running-ios-13/.
and then we represented them on our map. The picture of a [Accessed: 18-Nov-2021].
GUI is given in Fig. 16.

Seventh International Conference on Aerospace Science & Engineering 138

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
[12] Arxiv.org. [Online]. Available: http://arxiv.org/abs/1512.03385v1. [15] F. Can and J. M. Patton, “Change of writing style with time,” Comput.
[Accessed: 21-Nov-2021]. Hum., vol. 38, no. 1, pp. 61–82, 2004.
[13] L. Zhang, S. Wang, and B. Liu, “Deep learning for sentiment analysis: [16] M. O. Topal, A. Bas, and I. van Heerden, “Exploring Transformers in
A survey,” Wiley Interdiscip. Rev. Data Min. Knowl. Discov., vol. 8, Natural Language Generation: GPT, BERT, and XLNet,” arXiv
no. 4, p. e1253, 2018. [cs.CL], 2021.
[14] “Transformer (Attention is all you need),” 28-Apr-2019. [Online].
Available:
https://www.youtube.com/watch?v=z1xs9jdZnuY&list=RDCMUCxP
77kNgVfiiG6CXZ5WMuAQ&index=26. [Accessed: 21-Nov-2021].

Seventh International Conference on Aerospace Science & Engineering 139

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Autonomous Drone Navigation using Deep

Convolutional Neural Network
Muhammad Arif Arshad Saddam Hussain Khan Muhammad Waleed Khan
Pakistan Institute of Engineering and Pakistan Institute of Engineering and Pakistan Institute of Engineering and
Applied Sciences Applied Sciences Applied Sciences
Pattern Recognition Lab, DCIS Pattern Recognition Lab, DCIS Pattern Recognition Lab, DCIS
Islamabad, Pakistan Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected] [email protected]

Suleman Qamar Asifullah Khan*

Pakistan Institute of Engineering and Pakistan Institute of Engineering and
Applied Sciences Applied Sciences
Pattern Recognition Lab, DCIS Pattern Recognition Lab, DCIS
Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected]*

Abstract— Drones are unmanned aerial vehicles utilized for CNNs have emerged as one of the most potentially fruitful
a broad range of functions, including delivery, aerial avenues in the computer vision field in recent years, delivering
surveillance, traffic monitoring, architecture monitoring and outstanding results in a variety of vision recognition tasks,
even in War field. Indeed, drones confront significant obstacles including pose estimation [1], face recognition [2], digit
while navigating independently in unstable and highly dynamic recognition [3][4], image classification [5][6] medical
environments. In comparison with the standard “map-localize- analysis[7], content-based image retrieval [8][9] and object
plan” approaches, this research paper investigates an and pedestrian detection [10][11].
information-driven strategy to address the issues mentioned
above. For this purpose, a convolutional neural network (CNN) Due to its usage and suitability on commercially feasible
is proposed to pilot a drone across city streets safely. Every input drones that are often implemented with a fourth-looking
image generates two outputs: a steering angle for the drone's camera and lack additional power-hungry or heavyweight
movement whilst dodging the hindrances and the possibility of a sensors, vision-based navigation has recently received great
collision for alerting the Unmanned Aerial Vehicle (UAV) far attention. Simultaneously, advances in machine learning have
away from spotting risky situations and responding quickly. improved visual navigation capabilities, like Deep Neural
However, gathering enough data in an unorganized outside area Networks (DNNs) have enabled the creation of tail-to-tail
such as a metropolis is a challenge. It is not easy to collect data learning methods [12], learn-able parameters are used for
for the sake of the pilots, so data is used, which is collected by
feature extraction in it. While the previous method has limited
mounting a camera on the cars and bikes. This paper suggests
generalization capabilities, DNNs offer visual navigation in
that data collected by automobiles and bicycles can be used to
train a UAV that would not threaten other vehicles or
real-world contexts where visual appearances are inevitably
pedestrians. Dataset is divided using the train-test split method. diverse.
Despite being trained on urban roads from the perspective of Recently, new deep learning techniques have been
urban cars, the navigation strategy of the proposed architecture producing excellent results in a variety of domains
is very generalized. The high recall (99.38%), great accuracy [13][14][15]. Reinforcement learning (RL)-based techniques,
(97.78%) and good f-score (94.81%) of the proposed technique in particular, suffer from a significant rise in sample
suggest that it can be used for autonomous driving of drones.
complexity, making them unsuitable for usage by UAVs in
The proposed technique masters a navigation strategy that
safety-critical settings. Successful flying policies, on the other
accomplishes a relatively higher level of accuracy, recall, f-score
and precision, which shows that it outperforms previous hand, maybe learned using supervised-learning approaches.
approaches in the literature. [13][15][16] however, it hasn't worked out how to collect
enough expert trajectories from replicating yet. In addition, as
Keywords— Residual Network, Drone, Convolutional Neural mentioned by [16] drones must learn how to react in risky
Network, Perception and Autonomy circumstances just like human pilots.

I. INTRODUCTION Clearly, a UAV effectively flying in the streets must

follow the road and react to risky circumstances in the same
Unmanned Aerial Vehicles (UAVs) are one of the most manner that any other manned ground vehicle would. As a
significant disciplines in recent technology, with autonomous result, we introduce employing the information acquired from
drones being a key study focus. Self-flying, also called self- ground automobiles incorporated in the above-mentioned
piloting, refers to a drone's capacity to conduct aerial settings already. The dronet architecture is customized by
movements without the assistance of a person. In this case, modifying the initial and final layers to be compatible with the
autonomy is defined as the drone's decision to run the input feature map dimensions and output multi-class challenge
aforementioned self-flying activities without the need for (2 classes). Comprehensively, contributions made by this
human intervention. Manually operating UAVs is challenging, work are as followed:
as is deploying many sensors on a drone. As a result, it is
planned to build drones that will be controlled by a single front  We proposed a model named as customized modified
and rear camera, from which the drone would get real-time residual-Dronet architecture that can predict the
visual information and act independently. steering angle and collision probability

Seventh International Conference on Aerospace Science & Engineering 140

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Table 1: Quantitative findings on regression and classification tasks: For the steering regression task, EVA and RMSE are
calculated, while for the collision prediction job, Average accuracy and F1-score are assessed. In comparison to the considered
baselines, our model performs admirably. Despite having a small number of parameters, Modified Residual-Dronet performs
admirably on both tasks.
Model EVA RMSE Avg. F-Score Recall Num. Num.
Accuracy Parameters Layers
Random Baseline −1 ± 0.022 −0.3 ± 0.001 50 ± 1% 0.3 ± 0.001 - - -
Constant Baseline 0 0.34 75.6% 0.00 - - -
VGG-16 [17] 0.712 0.097 96.6% 0.921 - 2.6 ∗ 107 50
6
Resnet 50 [18] 0.795 0.119 92.7% 0.847 - 7.5 ∗ 10 16
Dronet [12] 0.737 0.109 95.4% 0.901 - 3.2 ∗ 105 8
4
Modified Residual- 0.99 0.48 97.78% 0.9481 0.9938 8.2 ∗ 10 8
Dronet

(a) (b)

Figure 1: (a) Modified Residual-DroNet is a branched CNN which anticipates the steering angle and collision probability
from a single 150 x 150 grey-scale frame. A ResNet-8 consists of three residual blocks (b) forms the common component
of the architecture, which is accompanied by dropout and non-linearity ReLU. Following that, the architecture is divided
in two fully connected layers and out of those two, one for prediction of steering angle and other for inferring a probability
of collision
which helps in safely navigation of a drone in urban vision methods. There are numerous techniques for drone
situations. navigation and obstacle avoidance in the existing literature.
The sensory input and processing required for the controlling
 The proposed model has been employed on
“Udacity” and “Collision” datasets, collected from of flying platforms vary greatly across these methods.
automobiles and bicycles. The “Udacity” dataset is The challenges of obstacle detection and avoidance are
basically used to calculate the steering angle, and the inextricably intertwined with those of autonomous navigation
“Collision” dataset is used to predict the probability in terms of their complexity. Object detection approaches for
of collision. identifying barriers are based on either machine learning
algorithms or computer vision techniques.
 The proposed model outperforms the existing state of To assess the device form, figure out the existence of
the art CNN models. hindrances, after that execute course devising, a UAV flying
Despite our system's impressive outcomes, we don't want outside is generally equipped with GPS field, and optical
to change the standard "map-localize-plan" drone navigation sensing devices [19]. However, due to construction, massive
approaches; instead, we want to explore if a likewise task can rushes, and dynamic states in metropolitan regions, these
be performed with a single shallow neural network. types of activities are still likely to suffer, resulting in
Traditional and learning-based techniques, we believe, will significant unobserved inaccuracy in the assessment of
eventually complement one another. system state.
DNN has recently attracted a lot of attention for their ability
II. RELATED WORK to teach control rules directly from raw sensory input. There
Obstacle detection and avoidance challenges are comparable
to autonomous navigation. Obstacle detection systems are
centered on either machine learning algorithms or machine

Seventh International Conference on Aerospace Science & Engineering 141

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 2: (a) Images from Udacity were utilized for learning the steering angles. (b) Images were collected in order to
determine the likelihood of a collision. No-collision frames are in the green box, whereas collision frames are in the red box.
are two types of approaches: reinforcement learning (RL) performance, we tweaked this design choice. Tasks stop
methods [15][20] and supervised learning methods [14], sharing parameters at the last ReLU layer, and then the path
[16][17][21][22]. is separated for both steering angle and probability of
Antonio Loquercio developed a Deep Convolutional Neural collision using two separate, fully connected layers. The
Network model named Dronet in his paper [12] which can steering angle is output by the first, while the collision
safely fly a drone across a city's streets. For every input probability is output by the second. The latter, strictly
image, Dronet produces two outputs a steering angle which speaking, is an index that quantifies the network uncertainty
controls the directions of the UAV and encounter possibility in prediction rather than a Bayesian probability. We still refer
that allows the UAV to identify risky conditions and respond to it as a probability.
quickly. For training a deep CNN, the proposed technique To train the collision and steering predictions, we use binary
leverages footage captured by a drone-mounted camera cross-entropy (BCE) and mean square error (MSE),
during real or simulated flights that approximately follow the respectively. Despite the fact that the network design is
intended path. suitable for reducing perplexity and time taken to process, a
Maria Tzelepi proposed a human crowd detection technique naive joint optimization faces severe convergence issues due
[24] which uses deep Convolutional Neural Networks (CNN) to the extremely varied gradient magnitudes produced by
for driving the autonomous UAV. The goal of the research is each loss. Using no weighting among the two losses during
to develop light architectures that can successfully the learning phase, in particular, leads towards convergence
discriminate between swarmed and non-swarmed images to unsatisfactory and weak results. That can perhaps be
taken by drones, as well as crowd heat maps which could be described due to the magnitude differences in gradients
useful for semantically improved flight blueprints by between classification and regression tasks towards the
designating no-fly zones. beginning of training, which could make optimization
challenging [25]. Because the MSE gradients' standards are
III. METHODOLOGY relative to the absolute steering error, the regression job's
Our learning approach seeks to anticipate the probability of gradients are initially much higher. As a result, as we progress
collision and predicting a steering angle using the drone’s through the training process, we place more emphasis on
forward-looking camera in real-time. These may be translated classification loss. Applying a constant loss weight between
into control flying instructions, allowing a UAV to fly safely the two losses or not will almost need considerably prolonged
around objects while avoiding collisions. The resulting optimization duration and possibly leads to sub-optimal
Architecture is shown in Figure 1. efficiency for the reasons described above. This is a form of
For the reduction of complexity of the architecture, the curricular learning [25]. This may be regarded as a sort of
proposed methodology is mutually used by two jobs to some curricular learning. More specifically, the BCE weight
extent, but it is then split into two branches at the end. One of coefficient is (1), but the MSE weight coefficient is always 1.
the two jobs is steering angle which is a regression problem, We used decay = 1/10 and epoch0 = 10 for our trials.
while the other one is collision probability prediction which
is a classification problem. Because the two objectives are so 𝑳𝒕𝒐𝒕 = 𝑳𝑴𝑺𝑬 + 𝐦𝐚𝐱(𝟎, 𝟏 − 𝒆−𝒅𝒆𝒄𝒂𝒚(𝒆𝒑𝒐𝒄𝒉𝟎 −𝒆𝒑𝒐𝒄𝒉) ) ∗ 𝑳𝑩𝑪𝑬 (1)
dissimilar, we recommend separating the last fully connected
layer of the network. We exclusively use images acquired by The Adam optimizer[26] is employed with 0.001 learning
manned vehicles throughout the training. Images collected rate and decay rate of 10−5 . For optimization, hard mining is
from bicycles are used to understand the likelihood of a used for learning the most difficult samples. In particular, for
collision, while images captured from cars are used to learn each epoch, we choose the k samples with the greatest loss
steering angles. and calculate the overall loss using Equation 1. We set the
A. Learning Methodology value of k to decrease with time.
The ResNet-8 design is followed by a 0.5 dropout and a ReLU
non-linearity in our network that the two jobs share. Figure 1
shows the blocks of the residual network.
The success of Resnet can be attributed to its learning idea,
which enables it to handle vanishing/exploding gradient
problems in extremely deep networks. To improve model

Seventh International Conference on Aerospace Science & Engineering 142

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

B. Dataset 0 for steering angle always and "no collision," as well as a

random estimator. The above-mentioned findings imply this
We utilize one of the freely accessible datasets from Udacity's
while being 317 times smaller than the previous best design.
project [27] to learn steering angles. Nearly 60,000 images of
Our method outperforms it in terms of prediction accuracy.
car driving are split across five experiments, 1 for testing and
Additionally, a favorable contrast to the VGG-16 network
4 for training. Every experiment saves images from three
[17] demonstrates a utilization for the residual learning
cameras (left, centre, and right) and data from the GPS, IMU,
scheme in terms of generalization, as described in Section III-
brake, steering angles, throttle, gear, and speed. To learn the
A. As demonstrated in Table 1, Figure 4, Figure 5 and Figure
probability of collision from images, We utilize a dataset that
6. Our design achieves a great performance as compared to
is freely available online from Dronet's project [12]. They
other models in the literature.
collected the dataset by mounted the GoPro camera on the
bicycle's handlebar. And drive the bicycle in multiple city B. Performance Metrics
areas, attempting to mix up the sorts of barriers (vehicles, Various common performance indicators are used to
people, plants, construction sites) and appearance of
surroundings. As a result, the drone can generalize in a variety
of settings. This dataset has 32000 images spread across 137
sequences. In these sequences, frames are labelled as 0 (no
collision) if the distance from the obstacle is enough, which
means there is less chance of collision, and those frames in
which there are some obstacles are labelled as 1 (collision).
Frames labelled as 1 (i.e. collision frames) are that kind of
data that is very difficult for a drone to gather yet required for
the development of a reliable as well as safe technique.
C. Drone Control Figure 3: Model performance: Anticipated vs Actual
The UAV is instructed to fly with a forward velocity of 𝑉𝑘 steering of Udacity testing sequence dataset using the
and a steering angle of 𝜃𝑘 using the outputs. The network uses Probability Density Function (PDF).
the probability of collision to regulate the forward velocity:
when the collision probability is zero, vehicles are assess the performance of the implemented models. Accuracy,
commanded to move with the maximum velocity i.e. 𝑉𝑚𝑎𝑥 ,
and it stops when the probability of collision is close to 1. The
forward velocity is filtered using a low pass filter 0 < 𝛼 < 1
as shown in Equation 2.
𝑽𝒌 = (𝟏 − 𝜶)𝑽𝒌−𝟏 + 𝜶(𝟏 − 𝑷𝒕 )𝑽𝒎𝒂𝒙 (2)
Similarly, the predicted steering angle is also converted into
a yaw angle (rotation around the z-axis). We transform 𝑠𝑘
from the [-1, 1] range to the required yaw angle 𝜃𝑘 in the
−𝜋 𝜋
[ , ] range and low pass filter it as shown in Equation 3.
2 2
𝝅
𝜽𝒌 = (𝟏 − 𝜷)𝜽𝒌−𝟏 + 𝜷 ( ) 𝑺𝒌 (3)
𝟐
Lastly, in conclusion, a novice dynamic navigation strategy Figure 4: Confusion matrix for collision categorization based
that will operate a drone accurately with just only a single on the obtained dataset’s testing images. The percentage of
forward-looking camera is developed. Our method has the samples in each group is shown by the numbers in this matrix.
advantage of being able to calculate a collision probability
using one image, excluding any prior knowing the platform's recall, and F-score are examples of these measures. Equation
velocity. Our belief is that the proposed architecture will be 4 measures the total number of accurate assignments and is
making decisions based on the range between noticed items used to determine accuracy. The proportion of actual collision
in the sphere of vision [18].
IV. EXPERIMENTAL SETUP
The experimentation conducted to evaluate our suggested
technique is presented in this portion, along with the set of
performance measures to assess the correctness of the model.
A. Regression and Classification Results
We initially analyze our model's regression performance
using the Udacity dataset's testing sequence [28]. We utilize
two measures to measure steering prediction performance: Figure 5
Root Mean Square Error (RSME) and Explained Variance
Ratio (EVA). We utilize F-score and average classification
accuracy to evaluate collision prediction performance. Table
1 compares Modified Residual-Dronet architecture with
several known architecture from the literature [16][27][29].
Weak baselines include a constant estimator that anticipates

Seventh International Conference on Aerospace Science & Engineering 143

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Through extensive testing, it is proved that a drone can be

trained for navigation in urban areas by simulating manned
automobiles. Furthermore, intriguing abstraction skills across
a broad range of circumstances is exhibited. It might be a
useful supplement to standard "map-localize-plan"
techniques in navigation related activities such as airborne
delivery.

Figure 6 REFERENCES
probability predictions is measured by recall Equation 5. [1] A. Toshev and C. Szegedy, “Deeppose: Human pose estimation via
deep neural networks,” in Proceedings of the IEEE conference on
Equation 6 specifies the F-score.
computer vision and pattern recognition, 2014, pp. 1653–1660.
𝑻𝒓𝒖𝒆 𝑷𝒐𝒔𝒊𝒕𝒊𝒗𝒆 + 𝑻𝒓𝒖𝒆 𝑵𝒆𝒈𝒂𝒕𝒊𝒗𝒆 [2] Y. Taigman, M. Yang, M. Ranzato, and L. Wolf, “Deepface:
𝑨𝒄𝒄𝒖𝒓𝒂𝒄𝒚 = (4) Closing the gap to human-level performance in face verification,”
𝑻𝒐𝒕𝒂𝒍 𝒏𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒔𝒂𝒎𝒑𝒍𝒆𝒔 in Proceedings of the IEEE conference on computer vision and
𝑻𝒓𝒖𝒆 𝑷𝒐𝒔𝒊𝒕𝒊𝒗𝒆 pattern recognition, 2014, pp. 1701–1708.
𝑹𝒆𝒄𝒂𝒍𝒍 = (5)
𝑻𝒓𝒖𝒆 𝑷𝒐𝒔𝒊𝒕𝒊𝒗𝒆 + 𝑭𝒂𝒍𝒔𝒆 𝑵𝒆𝒈𝒂𝒕𝒊𝒗𝒆 [3] D. Ciregan, U. Meier, and J. Schmidhuber, “Multi-column deep
𝑷𝒓𝒆𝒄𝒊𝒔𝒊𝒐𝒏 ∗ 𝑹𝒆𝒄𝒂𝒍𝒍 neural networks for image classification,” in 2012 IEEE
𝑭 − 𝑺𝒄𝒐𝒓𝒆 = 𝟐 ∗ (6) conference on computer vision and pattern recognition, 2012, pp.
𝑷𝒓𝒆𝒄𝒊𝒔𝒊𝒐𝒏 + 𝑹𝒆𝒄𝒂𝒍𝒍 3642–3649.
V. DISCUSSION [4] Y. LeCun et al., “Learning algorithms for classification: A
comparison on handwritten digit recognition,” Neural networks
Both conventional and learning-based techniques have Stat. Mech. Perspect., vol. 261, no. 276, p. 2, 1995.
advantages and limits, and our methodology is no exception. [5] C. Szegedy et al., “Going deeper with convolutions,” in
The benefits include allowing a drone to securely explore Proceedings of the IEEE conference on computer vision and
pattern recognition, 2015, pp. 1–9.
previously unexplored areas utilizing our easy learning and
[6] S. H. Khan, A. Sohail, and A. Khan, “COVID-19 detection in chest
control approach. Unlike prior techniques, there is no need to X-ray images using a new channel boosted CNN,” arXiv Prepr.
be given or construct an online map of the environment, pre- arXiv2012.05073, 2020.
define collision-free destinations, or find inside it. [7] S. H. Khan, A. Sohail, M. M. Zafar, and A. Khan, “Coronavirus
Furthermore, we exhibited intriguing generalization skills disease analysis using chest X-ray images and a novel deep
across a wide range of circumstances. It might be a useful convolutional neural network,” Photodiagnosis Photodyn. Ther.,
p. 102473, 2021.
supplement to standard "map-localize-plan" techniques in
[8] A. Babenko, A. Slesarev, A. Chigorin, and V. Lempitsky, “Neural
navigation related activities such as airborne delivery. Indeed, codes for image retrieval,” in European conference on computer
unlike CNN-based methods [21], the flying strategy that we vision, 2014, pp. 584–599.
offer may successfully fly in significant unknown [9] J. Wan et al., “Deep learning for content-based image retrieval: A
circumstances, excluding the need for any re-training or fine- comprehensive study,” in Proceedings of the 22nd ACM
tuning. Furthermore, our technique may be suitable for international conference on Multimedia, 2014, pp. 157–166.
platforms with limited resources because of the relatively [10] R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich feature
hierarchies for accurate object detection and semantic
basic and efficient network design. The drawbacks include segmentation,” in Proceedings of the IEEE conference on
the nimble dynamics of drones that aren't completely utilized. computer vision and pattern recognition, 2014, pp. 580–587.
So this is not feasible to unambiguously assign the robot an [11] P. Sermanet, K. Kavukcuoglu, S. Chintala, and Y. LeCun,
objective for achievement, such as in previous CNN-based “Pedestrian detection with unsupervised multi-stage feature
controllers [16][22][30]. There are various options for dealing learning,” in Proceedings of the IEEE conference on computer
vision and pattern recognition, 2013, pp. 3626–3633.
with the restrictions described above. When a high chance of
[12] A. Loquercio, A. I. Maqueda, C. R. Del-Blanco, and D.
collision is expected, one may construct 3D collision-free Scaramuzza, “Dronet: Learning to fly by driving,” IEEE Robot.
paths to make use of the drone agility, as shown in [29]. Autom. Lett., vol. 3, no. 2, pp. 1088–1095, 2018.
It is not difficult to estimate the distance to the target [13] S. Ross et al., “Learning monocular reactive uav control in
[30] if the 2D map is known. one may give the distance to the cluttered natural environments,” in 2013 IEEE international
target or use the new learning approaches developed for conference on robotics and automation, 2013, pp. 1765–1772.
ground vehicles [31] for the generalization of the goal-driven [14] T. P. Lillicrap et al., “Continuous control with deep reinforcement
learning,” arXiv Prepr. arXiv1509.02971, 2015.
problems. Furthermore, as shown in [32], a measure of
[15] A. Giusti et al., “A machine learning approach to visual perception
uncertainty might be added to our system to make it more of forest trails for mobile robots,” IEEE Robot. Autom. Lett., vol.
resilient. 1, no. 2, pp. 661–667, 2015.
[16] D. Gandhi, L. Pinto, and A. Gupta, “Learning to fly by crashing,”
VI. CONCLUSION in 2017 IEEE/RSJ International Conference on Intelligent Robots
In this paper, we introduced a CNN model capable of safely and Systems (IROS), 2017, pp. 3948–3955.
piloting a drone across city streets. Because gathering data in [17] K. Simonyan and A. Zisserman, “Very deep convolutional
networks for large-scale image recognition,” arXiv Prepr.
such an unregulated environment using UAV is both risky arXiv1409.1556, 2014.
and time-consuming, our model learns to fly by emulating [18] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for
automobiles and cyclists, which already obey the laws of the image recognition,” in Proceedings of the IEEE conference on
road. This model forecasts the collision probability and computer vision and pattern recognition, 2016, pp. 770–778.
required steering angle simultaneously with the high accuracy [19] S. Scherer et al., “River mapping from a flying robot: state
(97.78%), precision (99.38%), and f1-score (94.81%) estimation, river detection, and obstacle mapping,” Auton. Robots,
vol. 33, no. 1, pp. 189–214, 2012.
compared to other architectures in the literature, allowing a
[20] S. Shen, Y. Mulgaonkar, N. Michael, and V. Kumar, “Multi-sensor
UAV to react quickly to unexpected occurrences and hurdles. fusion for robust autonomous flight in indoor and outdoor

Seventh International Conference on Aerospace Science & Engineering 144

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
environments with a rotorcraft MAV,” in 2014 IEEE International [26] A. KingaD, “A methodforstochasticoptimization,” Anon. Int.
Conference on Robotics and Automation (ICRA), 2014, pp. 4974– Learn. Represent. SanDego ICLR, 2015.
4981. [27] M. Mancini, G. Costante, P. Valigi, T. A. Ciarfuglia, J. Delmerico,
[21] Y. Zhu et al., “Target-driven visual navigation in indoor scenes and D. Scaramuzza, “Toward domain independence for learning-
using deep reinforcement learning,” in 2017 IEEE international based monocular depth estimation,” IEEE Robot. Autom. Lett., vol.
conference on robotics and automation (ICRA), 2017, pp. 3357– 2, no. 3, pp. 1778–1785, 2017.
3364. [28] A. Kazakova, Y. Cho, and C.-G. Lee, “OSCAR: An Open-Source,
[22] G. Kahn, T. Zhang, S. Levine, and P. Abbeel, “Plato: Policy Self-Driving CAR Testbed,” 한국정보과학회 학술발표논문집,
learning using adaptive trajectory optimization,” in 2017 IEEE
pp. 720–722, 2018.
International Conference on Robotics and Automation (ICRA),
2017, pp. 3342–3349. [29] S. Yang, S. Konam, C. Ma, S. Rosenthal, M. Veloso, and S.
Scherer, “Obstacle avoidance through deep networks based
[23] N. Smolyanskiy, A. Kamenev, J. Smith, and S. Birchfield,
intermediate perception,” arXiv Prepr. arXiv1704.08759, 2017.
“Toward low-flying autonomous MAV trail navigation using deep
neural networks for environmental awareness,” in 2017 IEEE/RSJ [30] M. Pfeiffer, M. Schaeuble, J. Nieto, R. Siegwart, and C. Cadena,
International Conference on Intelligent Robots and Systems “From perception to decision: A data-driven approach to end-to-
(IROS), 2017, pp. 4241–4247. end motion planning for autonomous ground robots,” in 2017 ieee
international conference on robotics and automation (icra), 2017,
[24] M. Tzelepi and A. Tefas, “Human crowd detection for drone flight
pp. 1527–1533.
safety using convolutional neural networks,” in 2017 25th
European Signal Processing Conference (EUSIPCO), 2017, pp. [31] W. Gao, D. Hsu, W. S. Lee, S. Shen, and K. Subramanian,
743–747. “Intention-net: Integrating planning and deep learning for goal-
directed autonomous navigation,” in Conference on Robot
[25] Y. Bengio, J. Louradour, R. Collobert, and J. Weston, “Curriculum
Learning, 2017, pp. 185–194.
learning,” in Proceedings of the 26th annual international
conference on machine learning, 2009, pp. 41–48. [32] C. Richter and N. Roy, “Safe visual navigation via deep learning
and novelty detection,” 2017.

Seventh International Conference on Aerospace Science & Engineering 145

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Parameter Identification and Optimization of

Permanent Magnet DC Motor using Nonlinear Least
Square Error Method
1st Raja Muhammad Sohail Khan 2nd Hamza Malik 3rd Hammad Munawar
College of Aeronautical Engineering College of Aeronautical Engineering College of Aeronautical Engineering
National University of Sciences and National University of Sciences and Nautional University of Sciences and
Technology, Technology, Technology,
Risalpur, Pakistan Risalpur, Pakistan Risalpur, Pakistan
[email protected] [email protected] [email protected]

Abstract— Parameter identification and optimization of any identification method while, [7] implemented the inverse
system play an important role in system modeling and theory technique and [8] applied the moments method. Least
subsequent control system design. DC motors are frequently Square Error (LSE) is the most common technique for
employed in industry for their ease of use and control. Dynamic parameter estimation as it involves the minimization of
models of motors based on motor parameters are used to design squared error between the simulated and measured
the control systems. However, for designing sophisticated (experimental) responses of DC motor. The same technique is
control systems accurate motor parameter values are required. employed by [9] whereas, a recursive approach to LSE is
This paper presents a technique for the identification and employed by [10]. DC motor parameters can also be estimated
optimization of these parameters for any permanent magnet DC
using Artificial Neutral Network [11], where an added
motor using the non-linear least square method. Parameters and
advantage is that parameter estimation could be achieved by
speed response are measured using experimental readings from
data acquisition equipment and exploiting analytical utilizing adaptive learning method in an online scenario. The
relationships of PMDC motor dynamics. These experimental accurate DC motor parameters identified using the mentioned
values are used to optimize nominal motor parameters using the approaches are used to model the DC motor which is then
non-linear least square error method. A computer-based subsequently utilized for accurate controller design as
algorithm is employed to minimize the error in several iterations implemented in [6], [11] and [7]. An approach like this paper
based on the non-linear least square error function. The results is implemented in [9] utilizing the MATLAB constraint
show that the non-linear least square error method was able to optimization technique which also involves the curve fitting
minimize the relative error below the set tolerance which method based on LSE technique. Another very important
resulted in the close matching of simulated and measured aspect to look for, before selection of parameter estimation
responses of the motor. These optimized motor parameters can techniques, is the identification through inexpensive and
be utilized to design a sophisticated control system for various readily available basic testing apparatus. In addition to this, the
applications. employed technique should consist of a short testing cycle.
This is especially useful when quick prototyping of controller-
Keywords— estimation, nonlinear least square error based systems is performed. In this paper, a DC motor
method, parameter optimization, PMDC motor parameter identification approach based on Simulink Design
Optimization technique is presented. The motor used was a
I. INTRODUCTION part of actuator (shown in “Fig. 1”) required for linear
In industrial control systems, DC motors have various actuation in Stewart Platform. The measured parameters and
applications as they can be easily modeled and controlled. speed response of DC motor are used to estimate and find the
However, in certain scenarios, a precise and accurate model optimum parameter values based on nonlinear LSE technique
of the DC motors is required to achieve optimum control in parameter estimation tool of Simulink.
system design. In such a case, the nominal values for motor
specifications provided by the manufacturer in its data-sheet A. DC Motor Background Theory
are not sufficient to achieve the desired optimum performance The schematic diagram for a permanent magnet (PM) DC
of the required control system design. This is especially the motor is shown in “Fig. 2”. The electro-mechanical schematic
case when cheaper DC motors are used where values provided for a DC motor system can be used to develop DC motor’s
in the manufacturer's data-sheet have considerably large transfer function after applying the basic circuit analysis
tolerances making it difficult to achieve the desired goal. In techniques. The key DC machine relationships used to obtain
addition to this, another downside of selecting cheap motors is the required transfer function are shown as (1). The transfer
that no specifications with respect to electrical and mechanical function equates input voltage (V) to the output motor speed
parameters is provided especially if selected DC motor is part (𝜔) as shown in (2). The nomenclature of all variables and
of some other product as a whole e.g. a linear actuator. Various constants is shown in “TABLE I”.
techniques have been identified and implemented to model the
accurate DC motor model. System identification methods [1]–  𝑇 = 𝑘𝐼, 𝑉𝑐 = 𝐾𝑏 𝜔, 𝑉 = 𝐸 + 𝐼𝑅 
[4] usually employing MATLAB System Identification
toolbox can be utilized for determining motor’s parameters.
Other than System Identification, [5], [6] utilized the algebraic

Seventh International Conference on Aerospace Science & Engineering 146

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 1. Linear actuator used as legs of Stewart platform

𝜔𝑚 (𝑠) 𝐾𝑚
 =   Fig. 2. Schematic diagram for permanent magnet DC motor (adapted from [12])
𝑉(𝑠) 𝐽𝑚 𝑅𝑠+𝑅𝐵𝑚 +𝐾𝑏 𝐾𝑚

B. Armature Resitance (R)

B. Methodology
Motor stall test is usually employed to determine the DC
The overall methodology for implementing an optimum
Motor’s armature resistance. The two common methods used
control design of linear actuator of Stewart platform is divided
for measuring the armature resistance are as follows:
into four sections i.e. experimental measurement of motor
parameters, parameter estimation (optimization) using 1) Multimeter: After noting any zero offset in the
Simulink MATLAB, controller design using optimized ohmmeter, measure the resistance across DC motor positive
parameter values, and finally implementation and verification and negative terminals. Removing the zero offset will give
of results by implementation on actual system. Initially, the the correct measurement reading. Due to the chance of DC
electrical and mechanical parameters (Jm; Bm; R; L; Km; Kb)
motor brush to be in-contact with commutator, the measured
of DC motor of the linear actuator were determined from DC
motor model and experimental measurements on DC motor. resistance may vary at different rotor positions. To cater for
In addition to this, Motor speed response was also measured this, repeated measurements were taken at different rotor
at its rated voltage using required experimental setup. This was positions and averaged out to get the final armature
followed by estimation and optimization of measured resistance.
parameters using MATLAB Simulink constrained 2) Voltage and Current Measurement: The measured
optimization technique. The constraint optimization technique values obtained using multi-meter will tend to be slightly
compared the measured speed response with the simulated inaccurate due to capacitive and inductive losses of the
speed response obtained using nominal values of DC motor sensors inside multi-meter. To avoid this, another method can
parameters. The optimum parameters achieved using this be used to determine near-accurate resistance. In this method,
technique can be used to design an accurate controller for the
a DC voltage supply is used to excite the motor at its rated
actuators for further integration on a Stewart platform.
voltage while clamping the motor’s rotor in a fixed position.
II. EXPERIMENTAL MEASUREMENT OF DC MOTOR The current flowing through the motor’s terminals is
PARAMETERS measured to find the armature resistance. As specified
previously, repeated measurements are taken at different rotor
A. Experimental Setup
positions and averaged out. As the resistance is sensitive to
The acquisition hardware to carry out parameter rotor position, a number of rotor positions should be tested
measurements and finding the speed step response consisted
and the average taken.
of DC motor (disassembled from linear actuator), quadrature
encoder, DC power supply, Digital Oscilloscope, Digital C. Armature Inductance (L)
Multi-meter, National Instruments Data Acquisition Device DC motor’s armature inductance can be measured by
(NI-DAQ), HBridge motor driver (L298) and LABVIEW / exciting the end terminals of the motor with a low voltage AC
MATLAB installed CPU. supply while measuring the voltage drop across the series
resistance such as 1K Ohm added to motor’s path [13]. The
TABLE I. NOMENCLATURE
circuit diagram showing the given configuration is shown in
State Variable (unit) Description “Fig. 3”. The impedance (Z) can then be calculated using the
equation shown in (3) while equivalent series resistance
𝑉𝛼 (V) Supply voltage (RESR) and total phase (α) can be calculated using (4) and (5),
𝑅 (Ω) Armature resistance respectively. The voltages 𝑉𝐴 , 𝑉𝐵 and phase difference 𝜃
𝐿 (H) Armature inductance between the two voltages can be measured using a digital
oscilloscope. The armature inductance can then be found using
𝐼𝛼 (A) Armature current (6) where f denotes the frequency of the AC supply voltage.
𝜔 (rad/s) Back EMF of the motor
𝑉𝐵
𝐽𝑚 (𝑘𝑔/𝑚2 ) Inertia of the motor  𝑍=  
√𝑉𝐴2 −2𝑉𝐴 𝑉𝐵 cos(θ)+V2
B
𝐵𝑚 (𝑁𝑚𝑠/𝑟𝑎𝑑) Damping coefficient
𝐾𝑚 (𝑁𝑚/𝐴) Torque constant of the motor  𝑅𝐸𝑆𝑅 = 𝑍 cos(𝛼) 
𝐾𝑏 (𝑉/𝑟𝑎𝑑/𝑠) Back EMF constant
−𝑉𝐵 sin(𝜃)
𝜏𝑚𝑒𝑐ℎ (s) Mechanical time constant  𝛼 = 𝜃 − tan−1  
𝑉𝐵 −𝑉𝐴 cos(𝜃)
𝜏𝑒𝑙𝑒𝑐 (s) Electrical time constant
𝑍𝑠𝑖𝑛(𝛼)
 𝐿=  
2𝜋𝑓

Seventh International Conference on Aerospace Science & Engineering 147

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

D. Motor Torque Constant (𝐾𝑚 ) / Back EMF Constant (𝐾𝑏 ) G. Speed Step Response
Both these motor constants are based on physical and The speed response measured using experimental
geometrical properties of the DC motor. Factors such as procedure acts as the reference (final) response to be achieved
number of turns and magnetic flux density are the main in parameter estimation and optimization algorithm
contributors. By neglecting any non-ideal electrical and implemented using Simulink Design Optimization technique
mechanical losses, we can assume that the motor torque in MATLAB. “Fig. 4” shows the equipment setup for speed
constant and back EMF constant are equal. Based on this, only measurement. In order to obtain the speed step response, 12
back EMF constant will be measured. Under steady state VDC was given manually as step input repeatedly till 10
conditions, the voltage to current relationship is reduced to the seconds and the resultant speed response of DC motor was
(7). Using this relationship under no-load, DC motor is excited acquired using NI-DAQ (LABVIEW) attached through
quadrature encoder with DC motor. The obtained response is
Fig. 3. Circuit diagram for measuring inductance
shown in “Fig. 5”.
using a DC power supply while incrementing the voltage from III. PARAMETER ESTIMATION AND OPTIMIZATION
zero to its rated voltage in small steps. The current and speed
of the motor are measured during each iteration. The speed of In order to optimize the DC motor model parameters based
the motor was measured using a quadrature encoder via on the measured speed response, a toolbox of MATLAB
National Instruments Data Acquisition Device (NI-DAQ) Simulink known as Design Optimization is utilized. It is a
while current readings were recorded from DC power supply. numerical optimization technique that estimates and tunes the
The back EMF voltage (𝑉𝑐 ) was calculated using (2) at each Simulink model parameters and performs two specific tasks,
iteration and was plotted to get a trend line with speed (rad/s) which are:
on x-axis and voltage (volts) on y-axis. The back EMF 1) Calibration of DC motor model parameters with
constant can be calculated using (7). experimentally measured data to enhance the accuracy of the
model.
 𝑉 = 𝐼𝑅 + 𝐾𝑏 𝜔  2) Find the model parameter values that fulfill the design
requirements. This is achieved using a user interface called
E. Damping Coefficeint (𝐵𝑚 ) parameter estimation tool, which is used to estimate and
Based on the electrical and mechanical relationships of optimize the parameters of the model.
motor torque shown in (8) and (9), the damping coefficient The significance of this method is that, it can even find the
(𝐵𝑚 ) can be simplified under steady state conditions after unknown parameters based on the physically measured speed
canceling the derivative term as shown in (10). When excited data. The simulated response must match the measured
under its rated voltage, the measured current, speed and torque response to find the optimum estimates of the model
constants can be used to calculate the damping coefficient. parameters. The overall process after collection of

 𝜏𝑚 = 𝐾𝑚 𝐼 
𝑑𝜔
 𝜏𝑚 = 𝐽𝑚 𝐼 + 𝐵𝑚 𝜔 
𝑑𝑡

𝐾𝑚 𝐼
 𝐵𝑚 =  
𝜔

F. Motor Inertia (𝐽𝑚 )

The motor inertia ( 𝐽𝑚 ) can be measured by simply
measuring the mechanical time constant (𝜏𝑚𝑒𝑐ℎ ) and using the
relationship given by (11). The mechanical time constant can
be measured by switching off the running motor by opening
the electrical circuit while recording the time series data of
speed of motor using quadrature encoder data acquired using
NI-DAQ (LabView software).
Fig. 4. Motor speed measurement experimental setup
experimental data is as follows:-
a) Specification of parameters to be estimated in the
user interface (inital and final ranges, initial guesses (if
known))
b) Build the simulink model of DC motor based on
nominal values
c) Find optimum estimates based on suitable algorithm
(non-linear least square)
d) Repeat with other experimental data-sets for
verification
𝐽𝑚
 𝜏𝑚𝑒𝑐ℎ =  
𝐵𝑚

Seventh International Conference on Aerospace Science & Engineering 148

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

A. Parameter Estimation Tool in Simulink

Fig. 5. Measured speed response of DC motor

Fig. 7. Initial Measured and simulated responses

Based on the optimization problem, parameter estimation b) DC motor parameters used for simulation were
tool in Simulink MATLAB is used to find the optimum selected in Select Parameters option. Meanwhile, the lower
parameters for DC motor. The parameter estimation tool bounds were also set to zero.
requires specifying certain variables and options which are as c) A New Experiment was created and the measured
follows:- speed data was imported. The Plot and Simulate option was
a) Design parameter variables (𝑥: 𝑅, 𝐿, 𝐵𝑚 , 𝐽𝑚 , 𝐾) of DC selected to plot the measured and simulated response on the
motor to be estimated same window as shown in “Fig. 7”.
b) Defining variables bounds (𝑥𝑚𝑖𝑛 ≤ 𝑥 ≤ 𝑥𝑚𝑎𝑥 ) - d) After specifying the objective function SSE,
limits of design parameters to be estimated, as parameter stimation was run using Estimate option.
values cannot be negative, so lower limits were set to zero. After several iterations, the estimation stops after the
c) Selection of suitable objective function: calculates selected objective function has converged and the simulated
difference between measured and simulated response of DC response closely matched the measured response giving the
motor model. It iteratively tunes the parameter variable final estimated parameter values.
values to obtain the simulated speed response (𝑦𝑠𝑖𝑚 ) that
tracks the reference signal ( 𝑦𝑟𝑒𝑓 ) which in this case is IV. RESULTS AND DISCUSSIONS
measured speed step response. It minimizes the estimation After 28 iterations, the optimization stopped because the
error (𝑒(𝑡)) defined as, relative sum of square of errors (SSE) (5.76e-4) minimization
function came below its defined tolerance of 1e-3. Moreover,
 𝑒(𝑡) = 𝑦𝑟𝑒𝑓 (𝑡) − 𝑦𝑠𝑖𝑚 (𝑡)  the simulated response came very close to the measured plot
as shown in “Fig. 8”. The changing parameter values as
Simulink design optimization tool consists of various estimation algorithm was iteratively executing to reach the
algorithms such as nonlinear least square (NLS), Gradient optimum point are shown in “Fig. 9”. Whereas, the initial and
Descent (GD), simplex search (SS) etc. The algorithm tuned parameters are shown in “TABLE II”. It can be seen in
implemented in this paper is NLS as it is considered to be most “Fig. 8” that the simulated response is closely following the
suitable algorithms for parameter estimation of DC motor measured response other than a few milliseconds delay in
model. [14] some points. The parameter values achieved using this
optimization technique can be used in accurate controller
B. Estimation Procedure design for DC motor. The same optimized model along with
The Simulink model for DC motor with nominal / its controller design can then be used in actual controller
measured parameter values was developed to get the simulated implementation of the linear actuator and verified. The
response. The Simulink model with input set as voltage and achieved controller gains and overall system response can then
output set as speed is shown in “Fig. 6”. Afterwards, following be compared with the previously implemented model for
procedure is adopted to carry out the estimation: - further analysis.
a) In DC motor Simulink model, Parameter Estimation
from Analysis tab was selected to open a new session of
parameter estimation tool for DC motor.

Fig. 8. Final measured and simulated responses

Fig. 6. Simulink model

TABLE II. MEASURED / NOMINAL AND TUNNED PARAMTERS OF DC

MOTOR

Seventh International Conference on Aerospace Science & Engineering 149

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Parameter Measured / Tunned

[1] A. Simpkins, “System identification: Theory for the user, (ljung, l.;
(unit) Nominal 1999)[on the shelf],” IEEE Robotics & Automation Magazine, vol. 19,
𝑅 (Ω) 31.64 0.3342 no. 2, pp. 95–96, 2012.
Fig. 9. Iterative plot of parameter values H. Unbehauen and G. Rao, “A review of identification in
𝐿(𝐻) 0.301 1.1098 [2]
2 continuoustime systems,” Annual reviews in Control, vol. 22, pp. 145
𝐽𝑚 (𝑘𝑔/𝑚 ) 1.66e-14 3.289e-11 171, 1998.
𝐵𝑚 (𝑁𝑚𝑠𝑟𝑎𝑑) 2.54e-7 3.9675e-7 [3] J. C. Basilio and M. V. Moreira, “State-space parameter identification
in a second control laboratory,” IEEE Transactions on Education, vol.
𝐾𝑚 (𝑁𝑚/𝐴) 9.916e-4 0.0018 47, no. 2, pp. 204–210, 2004.
[4] G. F. Franklin, J. D. Powell, M. L. Workman et al., Digital control of
dynamic systems. Addison-wesley Menlo Park, CA, 1998, vol. 3.
V. CONCLUSION [5] G. Mamani, J. Becedas, V. Feliu-Batlle, and H. Sira-Ramirez,
The main parameters governing the dynamic equations of “Openloop algebraic identification method for a dc motor,” in 2007
the PMDC motor were measured and subsequently used to European Control Conference (ECC). IEEE, 2007, pp. 3430–3436.
generate a Simulink model of DC motor. These values also act [6] G. Mamani, J. Becedas, and V. Feliu-Batlle, “On-line fast algebraic
parameter and state estimation for a dc motor applied to adaptive
as the initial estimates in Simulink Design Optimization control,” in Proceedings of the World Congress on Engineering, vol. 2,
algorithm. In addition to this, the actual speed response of 2008.
PMDC motor was measured experimentally which also act as [7] M. Ruderman, J. Krettek, F. Hoffmann, and T. Bertram, “Optimal state
the final reference response to be achieved in the optimization space control of dc motor,” IFAC Proceedings Volumes, vol. 41, no. 2,
algorithm. After executing the parameter estimation tool in pp. 5796–5801, 2008.
Simulink, it was observed that the resultant simulated response [8] M. Hadef, A. Bourouina, and M. Mekideche, “Parameter identification
was very closely following the measured response. The tuned of a dc motor via moments method,” 2008.
parameters achieved during this optimization are the accurate [9] S. Adewusi, “Modeling and parameter identification of a dc motor
values based on which accurate and precise control of DC using constraint optimization technique,” International Journal of
Mechanical and Civil Engineering, ISSN, pp-46-56, vol. 13, 2016.
motor can be achieved.
[10] R. Krneta, S. Antic, and D. Stojanovi ´ c, “Recursive least squares
method ´ in parameters identification of dc motors models,” Facta
VI. FUTURE WORK universitatisseries: Electronics and Energetics, vol. 18, no. 3, pp. 467
Coupled model of DC motor and linear actuator can be 478, 2005.
obtained by augmenting rotary to linear dynamics in DC motor [11] A. Rubaai and R. Kotaru, “Online identification and control of a dc
transfer function. This complete model can be simulated in motor using learning adaptation of neural networks,” IEEE
Transactions on industry applications, vol. 36, no. 3, pp. 935–942,
MATLAB using optimized DC motor parameter values 2000.
obtained in this paper. This can lead to accurate controller
[12] M. W. Spong and M. Vidyasagar, Robot dynamics and control. John
design, which can finally be implemented in actual Stewart Wiley & Sons, 2008.
platform equipment. [13] “Kde-dms kde direct device manager software for uas electronic speed
controller (esc) series.” [On line]. Available: https://www.
ACKNOWLEDGMENT kdedirect.com/collections/uas-multirotorelectronics/products/kde-
The authors would like to thank the anonymous reviewers dms?page=specs
and the editors for their time and efforts in reviewing the paper [14] W. Koech, S. Rotich, T. Rotich, and F. Nyamwala, “Dynamic model of
a dc motor-gear-alternator (mga) system,” Asian Research Journal of
and giving valuable suggestions which greatly improved this Mathematics, vol. 1, pp. 1–16, 2016.
paper.
REFERENCES

Seventh International Conference on Aerospace Science & Engineering 150

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Threats to the Navigation of Drone Based Delivery

Services
Imtiaz Nabi, Salma Zainab Farooq Sunnyaha Saeed
National Center of GIS and Space Institute of Space Technology National Center of GIS and Space
Applications, Islamabad, Pakistan Applications,
Institute of Space Technology, [email protected] Institute of Space Technology,
Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected]

Abstract—Drone-based delivery services are expected to have Air, employing drone based service model, would reduce to
massive commercial deployment in the upcoming years. The 8.6% of the cost of conventional delivery services. Drone-
potential for drones to deliver goods across different terrains, based delivery services are expected to reduce time and cost
at an unprecedented speed, has been attracting both investors of delivery and assist in environmental protection as they use
and customers alike. Medical supplies, pesticides, ammunition,
eco-friendly Lithium batteries as power source [10].
and other variety of goods can be transported more quickly
and efficiently by using fixed-wing and rotary-wing unmanned However, there are some challenges in the way of UAV
aerial vehicles (UAVs). These delivery drones, however, require based delivery system. In literature, different issues have
a sound navigation system that can guide and navigate them been discussed regarding, the battery life, design of the
properly from one place to another. Global navigation satellite vehicle [11]– [13] and path planning of delivery drones [14].
systems (GNSS) have been playing a vital role in drone-based However, these aren’t the only problems in the way of UAV
delivery services as they provide navigation services by based delivery network. One of the most important issues is
employing size, weight, power, and cost efficient hardware. the navigational threats. These threats include GPS/GNSS
However, satellite-based navigation systems are vulnerable to jamming and spoofing. These navigational threats, if not
several threats, which turn drone-based delivery into a
catered for, pose physical, loss and financial risks [8].
complete disaster causing collateral damage to goods and
human infrastructure. This paper highlights some of the main Commercial UAVs depend on GNSS/GPS for their
threats to the GNSS-based navigation with regards to drone guidance and navigation [2]. Nonetheless, GNSS is
delivery services. vulnerable to interference from different sources, posing
threats to drone-based delivery services, endangering goods
Keywords—Global Navigation Satellite Systems (GNSS), and human infrastructure. This paper highlights the main
Global Positioning System (GPS), Zipline, Amazon Prime Air, threats to GNSS-based navigation of delivery drones, namely
Drone Based Delivery Services, GNSS Jamming, GNSS jamming and spoofing, and their impact on the service. The
Spoofing
flow of the paper is as follows; Section II of the paper
introduces jamming and a brief description of the
I. INTRODUCTION AND BACKGROUND countermeasures that can be adapted to counteract these
Unmanned Aerial vehicles (UAVs) or drones are known effects. Section III of the paper details spoofing, its threats to
for their military applications. They have proven themselves GNSS based navigation of delivery drone and finally the
as solemn force multipliers by showing effectiveness in countermeasures. Finally, conclusion is drawn in section IV
reconnaissance, surveillance, and armed combat e.g., “MQ after reviewing all the threats.
8b Fire Scout” currently used by the United States of
America [1]. In recent years, with the emergence of the II. GNSS JAMMING
Internet of Things (IoT), the capabilities of UAVs have One of the challenges towards the navigation of
extended to numerous civilian sectors. One major application autonomous drone-based delivery services is GNSS signal
is agriculture, where drones are used to spray pesticides in jamming. Jamming is the generation and transmission of
the fields [2] as well as monitoring the crops [3]. Other signals, with enough power, towards the victim receiver that
applications include, forest fire monitoring [4], emergency leave it unable to track GNSS signals [11] (Figure 1). There
and health management services [5], traffic monitoring [6], are multiple jamming incidents reported by literature that
search and rescue (SAR) operations [7] and even fast food confirm the effect of intentional and unintentional RF
industry [8]. The company “Zipline” operating in Rwanda, interference on GNSS navigation based devices. One of the
East Africa, helps to provide blood and other emergency most well-known jamming incidents in the GNSS
medical supplies to distant hospitals using UAVs [9]. Zipline community is in January, 2007 in San Diego, California
has proven itself quite helpful by timely supplying blood and during a naval exercise. The technicians jammed radio
medicines to patients in need, and thus saving thousands of signals and unknowingly also jammed GNSS signals,
lives. resulting in disruption of GNSS service As a result, Naval
The concept of drone-based delivery service has attracted Medical Center emergency pager stopped working. Besides
different companies, such as Amazon (Prime Air drone), that, the harbor traffic management system also failed and
Google (Project Wing drone), and German logistic company the Air Traffic Control (ATC) had to use a backup to
Deutsche Post DHL (Parcelcopte drone) to adopt it to maintain the flow of air traffic. The problem of GNSS
improve their quality of service and customer satisfaction [8]. disruption was unknown for three days [11]. Moreover,
According to [10], the net shipping cost of Amazon Prime several other reports confirm that GNSS signals are prone to

Seventh International Conference on Aerospace Science & Engineering 151

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

signal jamming leading to unwanted situations during signal proves to be a lot more dangerous than GNSS signal
outage [12]. jamming. It is because in case of spoofing the attacker can
Authors of [2] concluded that jamming signals depend transmit signals with false information that can even allow
upon two main parameters; the power of signals and distance the attacker to gain full control off the target. In the case of
of the jamming source from the victim receiver. In addition delivery drones, they can easily be landed and captured by
to this, jamming signals have two main characteristics; the attacker without risk of damaging the package that the
jamming to signal ratio (J/S) and central frequency. The UAV was carrying. A GNSS spoofing attack can be of two
nearer the GNSS signal jammer to the victim receiver, the main types; “meaconing” or “fabrication” [2]. Both
greater will be power of the jamming signal [13]. Thus, it can meaconing and fabrication are discussed in detail in the
overpower the carrier to noise ratio (C/N0) of GPS/GNSS following sections after the mathematical description of
receivers which is the basic acquisition parameter of these spoofing attacks. During GNSS Spoofing attacks the victim
receivers [14]. receiver receives both authentic and fake signals
Delivery drones that have safety-coded firmware can simultaneously. The combined signals can be modeled at the
easily be landed when exposed to GNSS/GPS jamming input of the victim receiver as follow:
signals. It is because jamming of both GNSS signals and N s 1

command and control link (C2 link) triggers their built-in

a
yRF , l (t )  s
l 0
RF , l (t )  nl (t )  a(t ) (1)
safety feature [15]. This increases the risk of stolen goods
that the delivery drone was carrying to its destination. Here a in superscript indicates that the receiver is under
Besides that, jamming signals can also cause these unmanned attack while sRF ,l is the authentic signal. The term nl (t ) is
aerial vehicles (UAVs) to crash, thus endangering humans noise component of the signal and a (t ) is the component
and animal lives and such an incident has been reported by from the attacker. The attacker’s component can be modeled
[16]. as:
N a 1
a(t )   s'
l 0
RF ,l  na (t ) (2)

In equation (1.2) na (t ) is the number of counterfeit

replicas s 'RF ,l is the signal generated by the spoofer and
na (t ) represents additional noise contribution. Signal
generated by spoofer can be affected by demodulation or
estimation errors and during imperfect synchronization, the
counterfeit signal includes a residual modulation term due to
difference in frequency f l and phase  l between both the
real and counterfeit signal. Due to this reason, the signal can
Fig. 1. GPS Jamming be modeled as.
s 'RF , L (t )  Al (t ) sˆRF ,l (t   l (t )) cos(2f l (t )  l ) (3)
With regards to jamming, figure 2 illustrates some of the In the above equation Al (t ) represents the ratio between
countermeasures that can be adapted [11].
amplitude of fake replica signals and authentic signals, sˆRF ,l
is the estimated signal, and  l (t ) is the relative delay
between them. The values of both Al (t ) and  l (t ) can be
either constant or vary with time depending upon the type of
attack.
Spoofing attacks can be of different types and levels
depending upon the objectives of the attacker [1]. Several
types of spoofing attacks are discussed in the following
section. Although different review papers can be found on
spoofing attacks, however, in this paper we have summarized
all the attacks to provide an understanding and show
spoofing as a potential threat to drone-based delivery
networks
A. Meaconing
Meaconing can be defined as the recording, delaying,
Fig. 2. Classification of GNSS Jamming mitigation techniques and rebroadcasting of the GNSS signals [2]. It is the most
basic type of spoofing attacks, however, it still can be
III. GNSS SPOOFING dangerous in some situations. It requires a signal repeater or
hardware that can record and replay signals. In the case of
GNSS spoofing or structured interference can be defined meaconing, the signal received from visible satellites is not
as the act of transmission of falsified signals towards the separated from the reception phase [3]. In addition, a
victim receiver to achieve one of the four main objectives; relative delay is added to the signal due to which the
distraction, destruction, hijacking, or apprehension [2]. Due
to the transmission of counterfeit signal, GNSS spoofing

Seventh International Conference on Aerospace Science & Engineering 152

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

meaconing signal arrives at the target or victim receiver with

a positive delay.
During a successful meaconing attack, the number of
fake signals matches the number of available satellites in
view. The main feature of meaconing attacks is that the
attacker generates an exact replica of the signals with a
constant amplification factor and delay. In addition, a noise
figure is introduced into the signal due to the meaconer
hardware component. The noise figure from the meaconer
hardware components is not negligible and it increases the
noise power. However, no residual modulation is introduced
in case of spoofing attacks [3] i.e. f l (t )  0 and l  0
Fig. 3. Simplistic Spoofing Attack: A signal simulator based spoofing attack
B. Simplistic Spoofing Attack
A simplistic GNSS spoofing attack is considered Intermediate spoofing attacks are harder to detect using
fabrication of GNSS signals, using commercial GNSS signal basic anti-spoofing algorithms due to the synchronization of
simulators, to broadcast signals towards target receiver to the code phase. However, they can be detected using
distract it from its true path (Figure 2). Signal simulators are complex measures such as angle of arrival defense. A
easily available on GitHub or can be created using possible implementation of a portable civilian GPS spoofer
Matlab/Simulink environment. Simplistic attacks may be is described in [6] which can be implemented for the
easier to implement, however, they are not consistent and development of a sound counter spoofing algorithm.
time-synchronized with the real signals. Signals generated
by the GNSS simulators usually have higher power as
compared to real signals. In addition, slight discrepancy in
the navigation message such as old ephemeris data, etc. can
affect the counterfeit signal. Also, the attack generated by
simplistic spoofer may represent satellites that are not
visible on the scene.
Due to these shortcomings, such types of attacks can
easily be countered with basic anti-spoofing algorithms.
Apart from that, in most cases, the attacker has to jam the
signal first to lose lock on the real signal and then inject the
stronger signal which is then locked by the receiver [3]–[5].
Though it can lead to disastrous situations once successfully Fig. 4. Intermediate Spoofing Attack: A receiver spoofer architecture
carried out but this technique is still a naive approach to
GNSS spoofing and can easily raise an alarm about signal D. Sophisticated Spoofing Attack
spoofing due to jamming of the signals first. The sophisticated spoofing attack is the most advanced
C. Intermediate Spoofing Attack technique of spoofing attack. It requires sub-centimeter-
In contrast to the simplistic attack, the intermediate level knowledge about the attacker's antenna phase center
attacks are far more advanced in nature. In these types of and the target antenna phase center. In addition to this, the
attacks, the attacker is able to determine the time, location, fake signals are accurately synchronized with the real ones
and satellite information from the received signals, and then due to which the phase and frequency differences becomes
makes believable counterfeit signals by synchronizing the negligible. Such types of attacks can easily distract a
local codes and carriers (Figure 3). For carrying out a receiver with advanced anti-spoofing algorithms due to the
successful attack, the counterfeit signals must be nature of the attack as they can easily suppress the authentic
synchronized with the authentic ones by code-phase within signals as well.
at least half a chip. In addition to this, the knowledge about Sophisticated spoofing attacks can be extremely hard to
the relative positions of the spoofer antenna and target detect even using advanced countermeasures such as angle
antenna and the receiver dynamics must also be known for of arrival defense. However, they can be countered through
successful attacks. In case of intermediate spoofing, each cryptographic techniques such as signals authentication [3].
tracking channel is attacked by first code-phase alignment, The spoofing attacks and their countermeasures are listed
and then signal power is increased gradually. It must be kept in Table I and II respectively
in mind that the signals are synchronized due to which the
f l parameter is negligible.

Seventh International Conference on Aerospace Science & Engineering 153

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

delivery to various locations. Due to which many marketing

giants are in a race to adopt this system. Amazon, Google,
DHL have already conducted tests and are highly interested
in such a system. However, there are many challenges that
drone-based delivery services have to face and overcome.
Apart from the weight of the payload & package, battery
life, structure of drones, and weather condition, navigation is
the most important aspect of such services which must be
reliable and fool-proof. Since drones rely on GNSS for their
guidance and navigation and GNSS can easily be
compromised with signals jamming and spoofing which can
Fig. 5. Sophisticated Spoofing Attack: A chain of spoofers synchronized
result in unwanted situations. Therefore, they must be
together and has information about the victim’s antenna phase center equipped with solid anti-jamming and anti-spoofing
algorithms to fight against several GNSS threats and make
IV. CONCLUSION drone-based delivery more reliable and easy.
Drone-based delivery networks can be a major productive
innovation as they can easily reduce the cost and time of

TABLE I –SPOOFING ATTACKS AND THEIR SPECIFICATIONS

Type Of Attack Hardware Requirement Context Requirement Difficulty of
Detection
Meaconing GNSS antenna, Low noise amplifier Must be located in proximity to target receiver, LNA gain, and Easy [11]
(LNA), RF Transmitter power level must be adapted according to the distance from the
target.
Simplistic Spoofing Signal simulator, power amplifier, The transmitting antenna must be close to the target Very Easy [11]
Attack RF transmitter
Intermediate Spoofing Modified receiver with transmitting Target receiver antenna position and velocity must be accurately Hard [11]
Attack front end known
Sophisticated Spoofing Multiple phase-locked receivers- Sub-centimeter level knowledge about target antenna phase center Extremely Hard
Attack spoofers position and velocity [11]

TABLE II – SPOOFING MITIGATION TECHNIQUES

Countermeasure Type Complexity Robustness Limitation

Navigation Message System: Low Robust against simplistic attacks Fails against meaconing and
Authentication Receiver: Low intermediate spoofing

Spreading Code Authentication System: Medium Robust against simplistic and Fails against meaconing
Receiver: Medium intermediate

Navigation Message Encryption System: Medium Robust against simplistic attacks Fails against intermediate,
Receiver: Medium sophisticated, and meaconing
attacks

Spreading Code Encryption System: high Robust against simplistic, Fails against meaconing
Receiver: High intermediate, and sophisticated

[5] J. C. Rosser, V. Vignesh, B. A. Terwilliger, and B. C. Parker,

REFERENCES “Surgical and Medical Applications of Drones: A Comprehensive
Review,” JSLS : Journal of the Society of Laparoendoscopic
[1] K. Wesson and T. Humphreys, “Hacking Drones,” Sci. Am., vol. 309,
Surgeons, vol. 22, no. 3. NLM (Medline), Jul. 01, 2018. doi:
no. 5, pp. 54–59, Oct. 2013, doi: 10.1038/scientificamerican1113-54.
10.4293/JSLS.2018.00018.
[2] S. Z. Khan, M. Mohsin, and W. Iqbal, “On GPS spoofing of aerial
[6] F. Outay, H. A. Mengash, and M. Adnan, “Applications of unmanned
platforms: a review of threats, challenges, methodologies, and future
aerial vehicle (UAV) in road safety, traffic and highway infrastructure
research directions,” PeerJ Comput. Sci., vol. 7, p. e507, 2021, doi:
management: Recent advances and challenges,” Transp. Res. Part
10.7717/peerj-cs.507.
Policy Pract., vol. 141, pp. 116–129, Nov. 2020, doi:
[3] F. Veroustraete, “The Rise of the Drones in Agriculture,” EC Agric.,
10.1016/j.tra.2020.09.018.
vol. 2, no. 2, pp. 325–327, 2015.
[7] S. Mayer, L. Lischke, and P. W. Woźniak, “Drones for Search and
[4] C. Yuan, Y. Zhang, and Z. Liu, “A survey on technologies for
Rescue,” p. 7.
automatic forest fire monitoring, detection, and fighting using
[8] M. Knobloch, “What impedes Consumers’ Delivery Drone Service
unmanned aerial vehicles and remote sensing techniques,” Can. J.
Adoption ? A Risk Perspective Arbeitsberichte aus dem Fachbereich
For. Res., vol. 45, no. 7, pp. 783–792, Apr. 2015, doi: 10.1139/cjfr-
Informatik,” no. 1, 2020.
2014-0347.

Seventh International Conference on Aerospace Science & Engineering 154

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
[9] E. Ackerman and E. Strickland, “Medical delivery drones take flight [14] D. Silva, “GPS Jamming and Spoofing Using Software Defined
in east Africa,” IEEE Spectr., vol. 55, no. 1, pp. 34–35, 2018, doi: Radio,” 2017.
10.1109/MSPEC.2018.8241731. [15] GNSS — Global Navigation Satellite Systems. Springer Vienna, 2008.
[10] Jung, Sunghun and kim hyunsu, “Analysis of Amazon Prime Air doi: 10.1007/978-3-211-73017-1.
UAV Delivery Service,” J. Knowl. Inf. Technol. Syst., vol. 12, no. 2, [16] P. Tedeschi, G. Oligeri, and R. Di Pietro, “Leveraging Jamming to
pp. 253–266, 2017, doi: 10.34163/jkits.2017.12.2.005. Help Drones Complete Their Mission,” IEEE Access, vol. 8, pp.
[11] T. Cokyasar, W. Dong, M. Jin, and İ. Ö. Verbas, “Designing a drone 5049–5064, 2020, doi: 10.1109/ACCESS.2019.2963105.
delivery network with automated battery swapping machines,” [17] C. G. L. Krishna and R. R. Murphy, “A review on cybersecurity
Comput. Oper. Res., vol. 129, p. 105177, May 2021, doi: vulnerabilities for unmanned aerial vehicles,” in SSRR 2017 - 15th
10.1016/j.cor.2020.105177. IEEE International Symposium on Safety, Security and Rescue
[12] T. Cokyasar, “Optimization of battery swapping infrastructure for e- Robotics, Conference, Oct. 2017, pp. 194–199. doi:
commerce drone delivery,” Comput. Commun., vol. 168, pp. 146– 10.1109/SSRR.2017.8088163.
154, Feb. 2021. [18] J. Bhatti and T. E. Humphreys, “Hostile Control of Ships via False
[13] T. Cokyasar, “Delivery drone route planning over a battery swapping GPS Signals: Demonstration and Detection,” Navig J Inst Navig, vol.
network,” in Procedia Computer Science, 2021, vol. 184, pp. 10–16, 64, no. 1, pp. 51–66, 2017, doi: 10.1002/navi.183.
doi: 10.1016/j.procs.2021.03.013. [19] L. Dobryakova and E. Ochin, “On the application of GNSS signal
[14] H. Huang, A. V. Savkin, and C. Huang, “Reliable Path Planning for repeater as a spoofer,” Zesz Nauk Akad Morska W Szczecinie, vol. nr
Drone Delivery Using a Stochastic Time-Dependent Public 40 (112, no. 112, pp. 53–57, 2014.
Transportation Network,” IEEE Trans. Intell. Transp. Syst., pp. 1–10, [20] T. E. Humphreys, B. M. Ledvina, M. L. Psiaki, B. W. O’Hanlon, and
2020, doi: 10.1109/TITS.2020.2983491. P. M. Kintner, “Assessing the spoofing threat: Development of a
[12] F. Dovis, GNSS Interference Threats and Countermeasures. Artech portable gps civilian spoofer,” 21st Int Tech Meet Satell Div Inst
House, 2015. Navig ION GNSS 2008, vol. 2, pp. 1198–1209, 2008.
[13] T. Westbrook, “The Global Positioning System and military
jamming: The geographies of electronic warfare,” J. Strateg. Secur.,
vol. 12, no. 2, pp. 1–16, 2019, doi: 10.5038/1944-0472.12.2.1720.

Seventh International Conference on Aerospace Science & Engineering 155

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Investigation of Pre-Seismic Ionospheric and

Atmospheric Anomalies before Mw 6.1 El Salvador
Earthquake on October 28, 2018
Abdur Rafeh Abbasi Abdul Hanan Hamid Muhammad Arqim Adil
Institute of Space Technology Institute of Space Technology Institute of Space Technology
Islamabad, Pakistan Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected] [email protected]

Daboor Sultana Saqib Mehdi Erman Senturk

Institute of Space Technology Institute of Space Technology Kocaeli University
Islamabad, Pakistan Islamabad, Pakistan Kocaeli, Turkey
[email protected] [email protected] [email protected]

Abstract— Seismo-ionospheric anomalies (SIA) from space 6.1 in the southwest of the Acajutla region in El Salvador
and ground instruments can provide deep insights over the on 28 October 2018.
epicenter of future earthquakes (EQ). An earthquake may Researchers introduced many theories and hypotheses
disrupt the lower as well as upper ionosphere through several that indicated ionospheric anomalies before EQs
coupling mechanisms during the EQ preparation periods. This
[5][6][7][8]. However, it is not yet confirmed, and no one
paper analyzes the temporal variations in Total Electron
Content (TEC) from Global Positioning System (GPS), culminates this topic with an explicit definition. As far as
Outgoing Longwave Radiation (OLR), and Air Temperature the researchers are concerned, many theories have been
(AT) from GIOVANNI-AIRS and MODIS, respectively, introduced like comparing local TEC with global TEC; solar
before El Salvador EQ that occurred on October 28, 2018. The activity, and analyzing with variations of f0F2 and NmF2
temporal TEC observations retrieved from the International [9] [10]. Recent studies also observed numerous pre-seismic
GNSS Service (IGS) stations of Managua and Guatemala ionospheric abnormalities in TEC, Nmf2 (F2 peak electron
cities, laying within the earthquake preparation area (EPA), density), and hmF2 (F2 peak height) [11][12][13], yet, the
showed a positive ionospheric response from 2 to 3 TECu from underlying mechanism is unclear and more research is still
the corresponding upper bound about 2-6 days before the EQ.
required in this field. The relationship between lithosphere-
The variations are found based on the confidence interval of
median and associated interquartile range for 10 days before ionosphere coupling is not pursued entirely and is largely
and 5 days after the main shock. The bounds are calculated open for research [14]. Outgoing Longwave Radiation
from continuous analysis of 10 days before and after the (OLR) is also very important when observing the anomalous
observed day. Additionally, sudden increments in atmospheric behavior of LAIC, as it measures the amount of energy
AT and OLR were also found consistent with ionospheric TEC released from the ground to outer space and atmosphere
variations that indicate an EQ-induced atmospheric which is considered another important source of finding
environment. Furthermore, to monitor the space weather seismic related anomalies in the ionosphere [15]. Ouzounov
conditions during the observation period for the EQ, we et al., [16] use multi-sensors and thermal infrared radiations
examined the solar wind speed (VSW), solar radio flux (F10.7)
(TIR) i.e., OLR to investigate atmospheric response and
solar activity, disturbance storm-time (Dst), Kp geomagnetic
activity indices to distinguish SIA. The mutual enhancements process before impending EQ in China from 2008 to 2014.
in atmospheric and ionospheric parameters might be due to the A study indicated ionospheric perturbations developed
abrupt energy released by this EQ that exhibits a lithosphere- within the few days before the happening of the seismic
atmosphere-ionosphere coupling (LAIC). activity that was monitored by DEMETER and GNSS-TEC
[17]. Many researchers work in the LAIC mechanism to
Keywords—Earthquake, GPS-TEC, Ionosphere, OLR, Space investigate a chain mechanism during the EQ phenomena.
Weather A researcher examined the seismic precursors by analyzing
the LAIC phenomena using the positive holes or p holes [2].
I. INTRODUCTION According to Kuo et al., [2] stress in the lithosphere
Advancement in space science and research intrigues emanates the positive charge carriers that result in a
many modern-day researchers to study the space disturbance in the ionospheric and atmospheric parameters.
atmosphere and its impact on human life and planet Earth. Seismo-ionospheric irregularities over the EQ
Scientists analyze space-based ionospheric observations to preparation zone can be conferred by two different
study the earth's atmosphere and their impact on techniques; positive holes (p hole) emission [18] [19] and
lithosphere-atmosphere-ionosphere coupling (LAIC) Radon emanation from EQ-prone regions [20] [21]. Radon
related to seismic activity [1][2][3][4]. Particularly emanation is argued by Freund et al., [18]. According to a
forecasting of EQs via seismo-ionospheric anomalies is now study [18], stress caused in the rocks in the epicenter of EQ
getting debate among researchers. This paper analyzes releases the radon that emanates the Earth’s surface, that
ionospheric and atmospheric anomalies that occur during result tends to vary the air ionization near the lithosphere.
the EQ period. The analysis shown in this research is one of Consequently, these phenomena are responsible to cause
the EQ events that occurred in the El Salvador region of Mw abnormalities in the atmospheric and ionospheric behavior
over the EQ breeding zone. Radon discharge from the EQ

Seventh International Conference on Aerospace Science & Engineering 156

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

breeding zone due to future seismic preparation period,

initiates processes like air ionization and ion recombination
subsequently propagating to the ionosphere by variating air
temperature (AT) and relative humidity [18] [19].
Solar radiation and geomagnetic activity are also
driving forces that cause trouble in the ionosphere.
Therefore, it is mandatory to exclude the factor of storms
before examining SIA. Therefore, researchers are quite
reluctant to examine SIA under storms. Gonzalez et al., [22]
characterized that if the value of Dst is between -50nT to -
30nT, it is considered a weak storm. While from -100nT to
-50nT, a moderate storm is considered, and below -100nT
storm is considered as a strong geomagnetic storm.
Furthermore, positive Dst is considered by the consequence
of solar wind pressure.
This paper analyzes pre-seismic and post-seismic
GNSS-TEC data that occurred on 28 October 2018 at 22:23
UTC in the Southwest region of Acajutla, El Salvador. The
magnitude of the EQ was 6.1 (Mw) and the depth of the EQ
was 22km from the surface of the earth. Furthermore, this
research also investigates geomagnetic indices i.e., Dst, Kp,
and solar index VSW, f10.7 to discriminate seismo- Fig. 1. Shake Map of Mw 6.1 El Salvador EQ
(https://earthquake.usgs.gov/earthquakes/eventpage/us1000hib5/shakema
ionospheric anomalies with solar or geomagnetic activities. p/intensity).
Additionally, AT and OLR are also examined to analyze the
atmospheric abnormalities and justify GNSS-TEC. OLR is an important constituent in the prediction of EQ
that depends on AT. As OLR is taken from the top of the
II.DATA AND METHOD atmosphere, it is considered an important source of
predicting lithosphere-ionosphere anomalies as energy
In this paper, TEC is being analyzed and anomalies
radon raises from ground to atmosphere and will provide a
were observed concerning EQ. The epicenter of the EQ was
better justification for the observed anomalies. The change
13.0350 N and 90.3690 W, 85 km from the southwest region
in OLR data is dependable with changes in other
of Acajutla, El Salvador (Table 1). TEC is a very important
atmospheric parameters [24]. Therefore, we analyze the
constituent in EQ forecasting studies, but earlier researchers
OLR as well as AT by collecting the data for 16 days from
were found to be very selective as they always associated
October 18, 2018, to November 02, 2018. The anomaly can
the TEC anomalies with EQ without checking the nature of
be detected in OLR by applying the limit bounds on OLR
geomagnetic storms during that EQ happening [23].
data. OLR data is freely available in
Therefore, before observing TEC, solar and geomagnetic
(https://giovanni.gsfc.nasa.gov/giovanni/).
indices are analyzed in this study. Space weather data are
The ionosphere primarily consists of Total Electron
retrieved from the OMNI website
Content (TEC), which plays a very vital role in the
(https://omniweb.gsfc.nasa.gov/).
ionosphere. Most researchers use TEC to analyze and
observe anomalies in the ionosphere before the EQ
[25][26][6]. Therefore, this research also uses TEC data to
analyze the behavior in the ionosphere for 10 days before
and 5 days after the impending EQ. To examine TEC
variations, TEC data is utilized from two different
International GNSS Service (IGS) stations located in
Managua and Guatemala cities. These stations are available
near the epicenter and within the earthquake preparation
area (EPA). We used the Dobrovolsky formula to calculate
the EPA [27].

𝜌 = 100.43𝑀 (1)

In (1), M indicates the scale of an EQ where ρ is

considered as the area of the affected region, which is in km
unit.
Slant TEC (STEC) from the GNSS ground-based
stations is acquired alongside the ray-path from the satellite
to the GNSS receiver [28][29]. To examine vertical TEC
(VTEC), STEC has been derived from the IGS ground-
based stations using the equation.

Seventh International Conference on Aerospace Science & Engineering 157

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
𝑓22 𝑓12 ℎ uses the information of pre ten days of the seismic activity
𝑆𝑇𝐸𝐶𝑎ℎ = − [𝑃4,𝑎 − 𝑐 (𝐷𝐶𝐵𝑎 + 𝐷𝐶𝐵ℎ )](2)
(𝑓1 −𝑓22 )40.3
2
and five days after the seismic activity. Discrepancies were
found based on the confidence interval of median (X)̃ . Now
From the above (2), f1 and f2 are the GPS dual for the limit bounds, we use the following equations to
h
frequencies, c is the speed of light, P4,a is the differences define the Upper Bound (UB) and Lower Bound (LB):
between smoothed code measurements while
DCBa and DCB h are differential code biases for the satellite 𝑈𝐵 = 𝑋̃ + 𝐼𝑄𝑅 (5)
and receiver [30]. Furthermore, to convert STEC into VTEC
following equation has been applied to STEC: 𝐿𝐵 = 𝑋̃ − 𝐼𝑄𝑅 (6)
𝑅𝑠𝑖𝑛𝑧
𝑉𝑇𝐸𝐶 = 𝑆𝑇𝐸𝐶 𝑥 𝑐𝑜𝑠 [𝑎𝑟𝑐𝑠𝑖𝑛 ( )] (3) For the detection of abnormalities in the time series, a
𝑅+𝐻
confidence interval of about 95% is declared, it is suggested
From the eq. (3), H is defined as the altitude of the that if the observed TEC gets a higher peak than normal
ionospheric layer from the surface of the Earth, R is distribution is anomalous. Hence, we can visualize the
considered as the radius of Earth while “z” concerning the abnormality in the time series by plotting (5-6). A similar
Earth is considered as the zenith angle of the satellite [31]. median-quartile methodology was used to examine the
The station-based TECs are calculated by weighting or changes in the OLR and AT. In addition, we used (5-6) for
surface interpolation of vTECs [32]. OLR data to generate UB and LB as well.
Detrended TEC (dTEC) is applied on vTEC to further
∑ 𝑤(𝜀)𝑉𝑇𝐸𝐶 scale the GPS-TEC anomalies [33]. dTEC extracted from
𝑇𝐸𝐶 = ⁄∑ 𝑤(𝜀) (4) vTEC using the following equation:

TEC is denoted as TECu, where is denoted as a unit 𝑇𝐸𝐶 − 𝐿𝐵; 𝑤ℎ𝑖𝑙𝑒 𝑇𝐸𝐶 < 𝐿𝐵
16 2
(therefore, 1 TECu is equal to 10 el/m ). Ciraolo et al., 𝑑𝑇𝐸𝐶 = { 0; 𝑤ℎ𝑖𝑙𝑒 𝐿𝐵 ≤ 𝑇𝐸𝐶 ≤ 𝑈𝐵 (7)
[29] use TEC data extracted from the TEC stations that lie 𝑇𝐸𝐶 − 𝑈𝐵; 𝑤ℎ𝑖𝑙𝑒 𝑇𝐸𝐶 > 𝑈𝐵
within EPA.
To find the TEC variations, a running median upon the
basis of quartile is performed. Inter-quartile range (IQR)
Table 1. Information about the EQ

Information about the EQ

Magnitude Country Location Date and Time Preparation Zone [km]

13.0350N
6.1 Mw El Salvador 2018-10-28 22:23:53 (UTC) 564.94
90.3690W

Information about the IGS stations

Distance from Epicenter Max. dTEC Max. OLR Max. AT

Station Name Location
[km] [ TECU] [ W/m2] [ oK]
14.590oN
GUAT 136.73 3.886
90.520oW
299 279
12.149oN
MANA 417.47 3.597
86.249oW

III. RESULTS AND DISCUSSION f10.7 solar activity indices. The Solar wind speed ranges
show a low-speed Vsw (i.e., 300-500 km/s) [35][36][37]; 1-
This study examines morphological features associated 2 days before the EQ, the index values reached 421 km/s.
with the El Salvador EQ that occurred on 28 October 2018 f10.7 values vary between 66-71 sfu before the seismic
by utilizing GPS-TEC, OLR, and AT datasets. Space activity, which indicates quiet or low-solar activity. The low
weather phenomena play a decisive role in ionospheric ranges of the space weather indices supported the indication
dynamics [25]. Therefore, before the investigation of the that the abnormalities noticed before the El Salvador EQ
ionospheric seismic interventions, the space weather may be associated with the impending seismic activities.
conditions should be monitored carefully. Fig. 2a and 2c
show the time series of Dst and Kp magnetic activity
indices. These indices suggest the weak magnetic activity
before the main seismic shock, where the Dst index ranges
within ±20 nT and Kp (Planetary K-index) range is below
than 3 [34]. Fig. 2b and 2d show the time series of V SW and

Seventh International Conference on Aerospace Science & Engineering 158

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

before the EQ) that lasted for three continuous days till 24
October 2018 and went back to normal afterward. Then a
severe change in AT was observed on 26 October 2018 (2
days before the EQ) where the temperature raised to 279 0K
(maximum during the observation period). [21] explained
that radon released by a tectonic movement initiates several
processes in the atmosphere such as air ionization due to the
dropped relative humidity. These ionizations lead to
changes in the air temperature (depending on the region)
which further variates the OLR. From Fig. 3, one can
observe that OLR measurements have undergone sudden
increments between 22-26 October 2018 (2-6 days before
the EQ) that may be induced by increased AT due to
increased radon around the EQ preparation zone. Similar
findings were reported by [24], where they find mutual
anomalies in the OLR and AT within 5 days before M 7.8
Nepal EQ. These mutual enhancements indicate an
increased radon activity around the EQ periphery that might
lead to ionosphere disturbances.
the ionospheric variations were monitored by two IGS
stations falling within the EPA named GUAT and MANA.
The details about the locations of these stations and the
recorded measurements are depicted in Table 1. The
temporal TEC time series for 15 days within the seismic
period, along with the dTEC values, are represented in Fig.
4. Significant anomalies in the terms of dTEC can be
observed between 22-26 October 2018 (2-6 days before the
EQ), where a deviation of about 3 TECU can be observed at
both stations, simultaneously. It can be observed that
Fig. 2. Depicts the space weather indices 10 days before and 5 days anomalies exhibited by OLR, AT (Fig. 3), and ionospheric
after the El Salvador EQ. TEC (Fig. 4) correlate perfectly with each other which
indicates a LAIC coupling phenomenon. In another study,
The quiet geomagnetic and solar indices provide a great Ouzounov et al., [3] observed different aspects of
opportunity to monitor the ionosphere conditions that might synchronized TEC and OLR abnormalities inside 5 days
experience seismic interferences. Before analyzing the period while examining the Tohoku EQ of Mw 9.0. Jianyong
ionosphere directly, we prefer to start analyzing the lower et al., [38] also analyzed similar abnormalities in
atmospheric conditions by considering the OLR and AT ionospheric GPS-TEC as impending seismic forerunners
variations. Fig. 3 provides the temporal variations observed using a compact system of GPS receivers around the region
during the observed period. It can be seen that the positive of the Wenchuan seismic zone. Hence, we suggest that
variations in the AT started on 22 October 2018 (6 days mutual ionospheric-atmospheric analysis may provide
significant contributions to future EQ forecasting.

Seventh International Conference on Aerospace Science & Engineering 159

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 3. Time series datasets for AT and OLR over the EQ's region. The red dotted line represents EQ day.

Fig. 4. Analysis of TEC and dTEC values recorded at (a) GUAT and (b) MANA stations, respectively, during 15 days

that showed consistency with the TEC anomalies, may be

IV. CONCLUSION attributed to the increased radon emanations around the El
Salvador EQ preparation zone that consequently disturbed
The motivation of this research is to investigate the the ionosphere. Interestingly, the anomalous response in the
seismic-induced ionospheric-associated abnormalities that atmospheric-ionospheric parameters was enhancing
occurred before the El Salvador EQ. LAIC is the focus of concerning days to the EQ. As the maximum response was
the analysis throughout this study. Before processing the observed just 2 days before the EQ which was initiated 6
data, solar and geomagnetic indices are being analyzed. To days ago. Henceforth, we can say from the derived results
analyze the data, we opted for two different GNSS stations, that observed anomalies can be a possible precursor to the
named GUAT and MANA, within the EPA. The ionospheric impending EQ. Even though, these results propose the
TEC observations were treated under a statistical analysis relation between EQ and SIA yet more analysis require to
based on confidence bounds where considerable deviations indicate a persuasive part in calculating EQ.
of 2 to 3 TECu were observed on October 23-26, 2018. The
anomalous AT and OLR observed on October 23-26, 2018,

Seventh International Conference on Aerospace Science & Engineering 160

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

ACKNOWLEDGMENT [19] F. T. Freund, I. G. Kulahci, G. Cyr, J. Ling, M. Winnick, J. Tregloan-

Reed, M. M. Freund, “Air ionization at rock surfaces and pre-
earthquake signals,” Journal of atmospheric and solar-terrestrial
The authors are very thankful to NASA, GIOVANNI,
physics, 71, 1824–1834, 2009.
and IGS for sharing the GPS-TEC, OLR, and Space weather
[20] S. A. Pulinets, D. Ouzounov, L. Ciraolo, R. Singh, G. Cervone, A.
indices data. We also pay our gratitude to the USGS for Leyva, M. Dunajecka, A. V. Karelin, K. A. Boyarchuk, A. Kotsarenko,
providing the seismic catalog and regional EQ information. “Thermal, atmospheric and ionospheric anomalies around the time of
the Colima M 7.8 earthquake of 21 January 2003,” Annales
REFERENCES Geophysicae 24(3), 835–849, 2006.
[21] S. A. Pulinets, and D. Ouzounov, “Lithosphere – Atmosphere –
[1] M. Hayakawa, Y. Hobara, “Current status of seismo-electromagnetics Ionosphere Coupling (LAIC) model – An unified concept for
for short-term earthquake prediction. Geomatics,” Natural Hazards and earthquake precursors validation,” Journal of Asian Earth Sciences 41
Risk, 1(2): 115-155, 2010. (4–5), 371–382, 2011.
[2] C. L. Kuo, L. C. Lee, and J. D. Huba, “An improved coupling model for [22] W. D. Gonzalez, L. A. Joselyn, Y. Kamide, et al., “What is a
the lithosphere‐atmosphere‐ionosphere system,” Journal of geomagnetic storm?” Journal of Geophysical Research Space Physics.
Geophysical Research: Space Physics, 119(4), 3189-3205, 2014. 99 (A4), 5771-5792, 1994.
[3] D. Ouzounov, S. A. Pulinets, A. Romanov, K. Tsybulya, D. Davidenko, [23] J. Ahmed, M. Shah, W. A. Zafar, M. A., Amin, T. Iqbal,
et al., “Atmosphere-ionosphere response to the M9 Tohoku earthquake “Seismoionospheric anomalies associated with earthquakes from the
revealed by multi-instrument space-borne and ground observations: analysis of the ionosonde data,” Journal of Atmospheric and Solar-
Preliminary results,” Earthquake Science, 24(6), 557-564, 2011. Terrestrial Physics, 179, 450-458, 2018.
[4] X. Zhang, K. Hattori, T. Tramutoli, “Preface to the special issue on [24] M. Shah, M. Khan, H. Ullah, S. Ali, “Thermal Anomalies Prior to the
electromagnetic phenomena related to seismic and volcanic activities 2015 Gorkha (Nepal) Earthquake from Modis Land Surface
from EMSEV in 2016,” Earthquake Science 30(4): 165-166, 2017. Temperature and Outgoing Longwave Radiations,” Geodynamics and
[5] D. Ouzounov, M. Kafatos, G. Guiliani, S. A. Pulinets, J. Y. Liu, et al., Tectonophysics, 9, 123-138, 2018.
“Observing of pre-earthquake transients features in the Earth
atmosphere-ionosphere environment associated with M6.4 and M7.1 [25] E. L. Afraimovich, E. I. Astafieva, M. B. Gokhberg, V. M. Lapshin,
Ridgecrest, California, Earthquakes of July 2019,” American V. E. Permyakova, G. M. Steblov, S. L. Shalimov, “Variations of the
Geophysical Union, 2019. total electron content in the ionosphere from GPS data recorded during
[6] A. R. Abbasi, M. Shah, A. Ahmed, N. A. Naqvi, “Possible ionospheric the Hector Mine earthquake of October 16, 1999, California,” Russian
anomalies associated with the 2009 M w 6.4 Taiwan earthquake from Journal of Earth Sciences, 6(5), 339-354, 2004.
DEMETER and GNSS TEC,” Acta Geodaetica et Geophysica, 1-15, [26] E. Senturk, M. S. Cepni, “Performance of different weighting and
https://doi.org/10.1007/s40328-020-00325-1, 2020. surface fitting techniques on station-wise TEC calculation and
[7] A. Kiyani, M. Shah, A. Ahmed, H. H. Shah, S. Hameed, M. A. Adil, N. modified sine weighting supported by the sun effect,” Journal of
A. Naqvi, “Seismo ionospheric anomalies possibly associated with the Spatial Science, 64(2), 209-220, 2019.
2018 Mw 8.2 Fiji earthquake detected with GNSS TEC,” Journal of [27] I. P. Dobrovolsky, S. I. Zubkov, V. I. Miachkin, “Estimation of the
Geodynamics, 140, 101782, 2020. size of earthquake preparation zones,” Pure and applied geophysics,
[8] M. Parrot, V. Tramutoli, J. Y. Liu, S. A. Pulinets, et al., “Atmospheric 117(5), 1025-1044, 1979.
and ionospheric coupling phenomena related to large Earthquakes,” [28] J. Y. Liu, Y. J. Chuo, S. J. Shan, Y. B. Tsai, Y. I. Chen, S. A. Pulinets,
Natural Hazards and Earth System Sciences Discussions, 2016. S. B. Yu, “Pre-earthquake ionospheric anomalies registered by
[9] E. M. Ovalle, M. A. Bravo, C. U. Villalobos, A. J. Foppiano, continuous GPS TEC measurements,” In Annales Geophysicae, 22(5),
“Maximum electron concentration and total electron content of the 1585-1593, 2004.
ionosphere over Concepción, Chile, prior to the 27 February 2010 [29] L, Ciraolo, F. Azpilicueta, C. Brunini, A. Meza, S. M. Radicella,
earthquake,” Advances in Space Research, 52(7), 1274-1288, 2013. “Calibration errors on experimental slant total electron content (TEC)
[10] F. Basciftci, C. Inal, O. Yildirim, S. Bulbul, “Comparison of regional determined with GPS,” Journal of Geodesy, 81(2), 111-120, 2007.
and global TEC values: Turkey Model,” International Journal of [30] A. Sekertekin, S. Inyurt, S. Yaprak, “Pre-seismic ionospheric
Engineering and Geosciences, 3(2), 61-72, 2018. anomalies and spatio-temporal analyses of MODIS Land surface
[11] J. Y. Liu, Y. I. Chen, Y. J. Chuo, C. S. Chen, “A statistical temperature and aerosols associated with Sep, 24 2013 Pakistan Earth-
investigation of pre earthquake ionospheric anomaly,” Journal of TEC findingsquake,” Journal of Atmospheric and Solar-Terrestrial
Geophysical Research, 111, A05304, 2006. Physics, 200, 105218, 2020.
[12] K. Heki, “Ionospheric electron enhancement preceeding the 2011 [31] J. A. Klobuchar, “Ionospheric time-delay algorithm for single-
Tohoku-Oki Earthquake,” Geophysical Research Letters, 38, L17312, frequency GPS users,” IEEE Transactions on aerospace and electronic
2011. systems, (3), 325-331, 1987.
[13] S. Kalita, B. Chetia, “A novel approach for ionospheric total electron [32] E. Senturk, H. Livaoglu, M. S. Cepni, “A Comprehensive Analysis of
content earthquake precursor and epicenter detection for low-latitude,” Ionospheric Anomalies before the Mw 7·1 Van Earthquake on 23
International Journal of Engineering and Geosciences, 5(2), 94-99, October 2011,” The Journal of Navigation, 72(3), 702-720, 2019.
2020. [33] M. A. Tariq, M. Shah, M. Hernández-Pajares, T. Iqbal, “Pre-
[14] E. Senturk, M. S. Cepni, “A statistical analysis of seismo-ionospheric earthquake ionospheric anomalies before three major earthquakes by
TEC anomalies before 63 M w≥ 5.0 earthquakes in Turkey during GPS-TEC and GIM-TEC data during 2015–2017,” Advances in Space
2003–2016,” Acta Geophysica, 66(6), 1495-1507, 2018. Research, 63(7), 2088-2099, 2019.
[15] N. Venkatanathan, B. Kaarthick, C. Priyadharshini, “OLR anomalies [34] S. Mehdi, M. Shah, N. A. Naqvi,” Lithosphere atmosphere ionosphere
prior to big earthquakes (Mw> 6.0)–A case study on earthquakes of coupling associated with the 2019 Mw 7.1 California earthquake using
India’s neighboring region occurred during the year 2012,” New GNSS and multiple satellites,” Environmental Monitoring and
Concepts Global Tecton, 1(3), 34-44, 2013. Assessment 193(501), 2021.
[16] D. Ouzounov, S. A. Pulinets, et al., “Atmosphere response to pre- [35] S. R. Cranmer, “Coronal holes and the high-speed solar wind,” Space
earthquake processes revealed by satellite and ground observations. Science Reviews, 101(3-4), 229-294, 2002.
Case study for few strong earthquakes in Xinjiang, China (2008- [36] J. Zhang, M. W. Liemohn, J. U. Kozyra, M. F. Thomsen, H. A. Elliott,
2014),” Annals of Geophysics, 63(5), 2020. J. M. Weygand, “A statistical comparison of solar wind sources of
[17] K. Ryu, M. Parrot, S. G. Kim, K. S. Jeong, J. Chae, S. A. Pulinets, K. moderate and intense geomagnetic storms at solar minimum and
I. Oyama, “Suspected seismo‐ionospheric coupling observed by maximum,” Journal of Geophysical Research, 111(A01104), 2006.
satellite measurements and GPS TEC related to the M7.9 Wenchuan [37] D. H. Brautigam, M. S. Gussenhoven, D. A. Hardy, “A Statistical
earthquake of 12 May 2008,” Journal of Geophysical Research: Space Study on the Effects of IMF Bz and Solar Wind Speed on Auroral Ion
Physics, 119(12), 10,305–10,323, 2014. and Electron Precipitation,” Journal of Geophysical Research, 96 (A4),
[18] F. T. Freund, A. Takeuchi, B. Lau, A. Al-Manaseer, C. Fu, N. Bryant, 5525-5538, 1991.
et al., “Stimulated infrared emission from rocks: assessing a stress
indicator,” eEarth, 2, 1–10, 2007.

Seventh International Conference on Aerospace Science & Engineering 161

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
[38] L. Jianyong, M. Jianyong, Y. Xinzhao, Z. Rui, S. Hongbo, H. Yufei,
“Ionospheric total electron content disturbance associated with May
12, 2008, Wenchuan Earthquake,” Geodesy and Geodynamics, 6 (2),
126–134, 2015.

Seventh International Conference on Aerospace Science & Engineering 162

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Uplifting Agriculture through Information and

Communication Technology and Geospatial
Applications: A Review of Potential Tools and
Techniques
Muhammad Adnan Shahid Saif Ullah Khan Subhe Seeme
Department of Irrigation and Drainage, Agricultural Remote Sensing Lab Agricultural Remote Sensing Lab
University of Agriculture, Faisalabad University of Agriculture, Faisalabad University of Agriculture, Faisalabad
Faisalabad, Pakistan Faisalabad, Pakistan
[email protected] Faisalabad, Pakistan
[email protected] [email protected]
(Correspondent Author)
Muhammad Danish Majeed
Agricultural Remote Sensing Lab Muhammad Saqlain
University of Agriculture, Faisalabad Agricultural Remote Sensing Lab
Faisalabad, Pakistan University of Agriculture, Faisalabad
[email protected] Faisalabad, Pakistan
[email protected]

Abstract - In Pakistan, agriculture is the back bone of field of agriculture can be utilized beneficially for the good
country’s economy and largest contributor in the GDP, i.e. will of the farming community, as it can be helpful for
19.2%, providing employment to a large chunk of labor force, natural resource management (Kumar et al., 2015). Every
as around 70% of population depends on agriculture for its farmer aims at producing the food in a cost-effective
livelihood. Therefore, the role of the agriculture is
manner. The usage of GIS and remote sensing in agriculture
indispensable to the wellbeing of the nation, poverty
is helping the farmers to increase their production, reducing
alleviation, employment generation, and food security. Large-
scale population and labor shift from rural to urban areas
input costs, and achieving more efficient utilization of their
along-with other constraints including shrinking of arable land land. In the country like Pakistan, where agriculture is major
due to horizontal urban advancement, water shortage and source of country’s economy, government can use these
climate change have resulted in shrinking agricultural growth. applications in order to get information about types of the
Innovations and involvement of latest technologies in crops grown, area under cultivation, and expected yield, as
agriculture is the need of hour to increase the agricultural well as monitoring and mitigating natural hazards including
productivity. Agricultural data collection by using Remote floods and droughts.
Sensing (RS) techniques is an efficient method. Radar imagery
The major aspect of remote sensing is the spectral
and satellite based optical technology can be used for data
information, which can be employed to estimate the canopy
acquisition, analysis, monitoring, and mapping. Various factors
can be studied with greater accuracy through digital image
parameter in crop modeling that ultimately determines the
processing (supervised or un-supervised classification, image crop growth stages and crop health. Moreover, GIS and
enhancement, and spectral slicing). Similarly, in-situ Remote Sensing can be used to analyze land use/ land cover
observations can be integrated using ICT based tools to and assess any significant changes in it due to drought,
different geospatial and crop models for flood monitoring, floods and other extreme weather events (Bill, 2018).
drought assessment, crop production forecast, and estimation Similarly, remote sensing can be used for identification of
of crop water requirement and acreage under different crops. pest infestation and diseases. Water resources can also be
These advances may prove to be the foundations for the future
monitored by the use of geospatial techniques, as well as
research
using ICT based tools. Thus, the major uses of GIS and
Keywords:- GIS and RS, ICT, digital agriculture, hazards
remote sensing in the agriculture may be summarized as
mitigation.
identification of crops, area estimation under a certain crop
I. INTRODUCTION and assessment of its condition, biomass and yield
Agriculture is one of the major pillars in the economy of estimation, water resources management, soil moisture
Pakistan contributing 19.2% in the GDP and providing monitoring and management, agro meteorology, and
employment to 38.5% of the labor force. More than 65-70% irrigation management by scheduling irrigations precisely
of the population depends upon agriculture for its livelihood (Nain, 2017).
(PES 2021). Application of Remote Sensing (RS) and Areas of food insecurity can also be identified by using
Geographic Information System (GIS) techniques in the various tools in GIS. This may help ending the global

Seventh International Conference on Aerospace Science & Engineering 163

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

hunger by eliminating its root causes. Data collection chance of negligence of different features, synoptic view of
through remote sensing either directly or by installation on the area can be captured through imagery or aerial
the farm implements allows the farmers to perform different photography, and nothing can be remained unobserved in
farm operations precisely. For example, if combine this case. Figure 1 shows the LULC map developed for a
harvesters are equipped with GPS tracking units, it could canal command area.
help measuring crop yield and moisture content at the exact
locations in the field at the time of harvesting. Similarly,
crop growth monitoring along-with different kinds of stress
identification and their management may help enhancing
agricultural productivity and natural resources sustainability.
II. PRECISION AGRICULTURE
With the introduction of precision agriculture, focus has
been made to develop a system in which the management
practices are selected according to the site condition. All this
has become easy with the introduction of satellite based
image analysis. Similarly, for the assessment of real-time
agricultural land-scape, the usage remote sensing has been
increased. The basis of the system is data acquisition using
modern technologies. With the integrated use of
technologies including Global Positioning System (GPS),
GIS and RS, yield monitoring, assessment of soil related
problems, plant health, pest population, and incorporation of
different inputs can be done as per requirement (Seelan et
Figure 1. LULC Classification for a Canal Command (Work
al., 2003, (Gassner et al., 2012, Fountas and Gemtos, 2016,
conducted under Ignite funded project of the authors).
Davis et al., Shafi et al., 2019)
IV. RAINFALL ESTIMATION AND ANALYSIS
III. LAND USE LAND COVER CLASSIFICATION
Rainfall measurement is necessary in order to determine
Existing natural surface features (forests, hills, rivers etc.)
water requirements of the crops. Moreover, rainfall is one of
are termed as land cover, whereas, the land use features are
the major input parameters of hydrological models. Climate
those modified by human beings (urban, rural settlements,
and meteorology highly influence the agriculture. Several
canal, orchards, etc.) For identification of different surface
agro-meteorological parameters e.g. surface albedo, surface
features, their hierarchical classification, and its role in the
temperature, solar radiation, and evapo-transpiration can be
study of global change is known as land use/ land cover
accurately and frequently estimated through satellite
mapping. It is largely affected by human activities such as
metrology. To develop correlation and weighted regression
environmental problems that lead to deforestation and
model, minimum and maximum temperature and rainfall
biodiversity loss. Therefore, management and future
data are the prime agro meteorology inputs, as crop water
planning for critical decision making can be provided by
demand assessment can be done by subtracting the amount
available data on land use and land cover. Unplanned and
of useful rainfall from ET (Kumar et al., 2015). Satellite-
uncontrolled changes in land use and land cover are due to
based rainfall estimations have been made possible through
growing population and increasing socio-economic
different missions including Tropical Rainfall Measuring
problems (Kaul and Sopan, 2014).
Mission (TRMM) and Global Precipitation Measurement
Mismanagement of agricultural, range and forest lands leads
(GPM). In addition, GIS can be successfully used to develop
to severe environmental problems such as floods, landslides
rainfall raster datasets based on point rainfall data, as well as
that alter the land use and land cover. Satellite imagery can
for rainfall data management by developing interpolated
be used to detect pixel by pixel change for determination of
rainfall maps. Such rainfall datasets and graphs can be
land use/ land cover maps. In different geographical studies,
developed on daily as well as monthly basis to perform
land use and land cover mapping has always remained very
different kind of analysis. Figure 2 presents an interpolated
important. It is composed of basic information of features
rainfall map for Faisalabad for the month of July, 2018.
and their locations such as area, location, shape, and pattern.
For identification of land use and land cover properties
throughout the year, geo-registered multi temporal remote
sensing data is very beneficial, and in extensive regions in
terms of cost, accuracy, and manual errors, it is far better
than manual surveys.
In comparison to the manual surveys, where there is a great

Seventh International Conference on Aerospace Science & Engineering 164

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 3. Flood Vulnerability Map for District Jhang;

activity performed in Agricultural Remote Sensing Lab.
In addition to direct assessment of flood affected areas from
imagery, different RS data like Digital Elevation Models
Figure 2. Monthly rainfall interpolation map of Faisalabad integrated with flow data and other parameters in hydraulic
for July, 2018 (Work conducted under PARB funded project models like HEC-RAS may also be used to develop flood
No.947 of the authors). inundation and velocity maps for a river floodplain. Figure 4
V. FLOOD MONITORING presents such flood inundation map for Kalabagh to
World population is facing great threats from the hydro- Chashma reach of Indus river developed using HEC-RAS
meteorological natural hazards like floods. These threats and model.
their impacts on human beings must be assessed in order to
combat these issues timely. From agriculture point of view
flood, the monitoring is necessary in order to avoid the huge
damages to crops and livestock especially during monsoon
rainfall spell. Moreover, flood water harvesting can be
beneficially used for different applications either agriculture
or domestic, but in order to do that flood monitoring and
forecasting is essential. The flood monitoring has become
necessary due to increased frequency and intensity of such
events under changing climate. Remote sensing technology
has made it easier to obtain data at much larger scales than
the area which could be covered through field-based
instruments and methods. Information about flooded areas at
different temporal and spatial scales can be gathered through
satellite imagery. For management of natural disaster
extensive multi-temporal spatial data is needed. In this Figure 4. Flood inundation map for Kalabagh to Chashma
regard, usage of GIS can be helpful for early warning, Reach (Work conducted under the NRPU project No.7431
monitoring and disaster management (Bannert, 2000). In of the authors).
order to understand the effects of flooding in a particular Geospatial techniques integrated with hydrologic models
area such as roads, airport, buildings, streets, and railways, may also be used for real time monitoring or prediction of
flood inundation mapping is done. Flood inundation or any possible flood event and providing flood warnings. For
vulnerability mapping can be performed by geo-statistical this purpose, the geospatial data in the form DEM, real time
analysis of a historic archive of satellite imageries. Figure 3 rainfall and other parameters including soil and land cover
presents such flood inundation/ vulnerability map for characteristics have integrated to develop hydrologic models
District Jhang. for specific river catchments to predict downstream flows
and any possible flood situations. Figure 5 presents the
hydrographs developed for Poonch and Chenab rivers using
such approach in HEC-HMS.

Seventh International Conference on Aerospace Science & Engineering 165

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

requirements estimation using remote sensing, an irrigation

advisory has already been developed (www.uafimis.com)
and being employed to provide ET-based irrigation
scheduling guidelines to the farmers, which is helpful in
reducing water losses by shifting towards precision
irrigation to increase per acre yield and water productivity.
In addition to precise scheduling, use of High Efficiency
Irrigation System (HEIS) such as drip and sprinkler
techniques may help reducing major portion of water
application losses. Drip irrigation is more efficient than
traditional methods of irrigation including flood, basin and
furrow irrigation systems. However, real benefits of these
HEIS may be achieved by automation, which is possible by
integrating ICT-based tools to perform irrigation need
assessment, as well as automatically apply the required
amount.
VII. CONCLUSION
With increase in population, the pressure on the natural
resources is increasing globally. Sustainable agriculture is
necessary to meet the challenges of food security and
geospatial data can be successfully used for this purpose, as
Figure 5. Flood hydrographs for Poonch and Chenab rivers it is necessary in the current scenario as well as for future
developed using HEC-HMS (Work done under TDF02-117 generations (Joshi et al., 2002). In the developing countries,
project of the authors). where resources are limited, remotely sensed data is a cheap
VI. CROP INVENTORY source for their management. Tremendous opportunities are
Remote sensing and GIS can be effectively used for available for the scientists and data analysts in the form of
identification of crops to determine cropping patterns and application of RS and GIS in modelling and mapping and
perform crop mapping. For a country having agriculture employing ICT based tools. A brief review of potential
based economy, it is important to have knowledge of the opportunities and options has been presented in this paper,
various kinds of crops grown, expected yield, and area under which are already being employed or may be a great
the certain crops. Crop modeling is done by using the breakthrough in future. Integrated use of these geospatial
spectral information, and strongly related to the canopy techniques and ICT based tools may be greatly helpful to
parameter on the basis of which crop growth stages and crop utilize the resources in efficient manner and uplift the
health is determined and yield can be estimated (Singh et al., agriculture sector in Pakistan.
2002; Ferencz et al., 2004; Zhang et al., 2004; Prasad et al., REFERENCES
[1] Ahmad, I., Saeed, U., Fahad, M., Ullah, A., ur Rahman, M. H.,
2006; Ahmad et al., 2018; Ahmad et al., 2020). Various Ahmad, A., & Judge, J. (2018). Yield forecasting of spring maize
maps can be created by a combination of satellite images using remote sensing and crop modeling in Faisalabad-Punjab
Pakistan. Journal of the Indian society of remote sensing, 46(10),
and survey data, which would ultimately give the evidence 1701-1711.
to the agribusinesses such as seed and fertilizer companies, [2] Ahmad, I., Singh, A., Fahad, M., & Waqas, M. M. (2020). Remote
about the land owners and area under cultivation. Thus, in sensing-based framework to predict and assess the interannual
variability of maize yields in Pakistan using Landsat imagery.
inventorying data base on different crops, remote sensing Computers and Electronics in Agriculture, 178, 105732
can play a significant role. It aids in higher precision of data [3] Ara, M.J., and S.A. Chief. 2019. EFFECT OF FLOODS ON
estimation by reducing the amount of field data to be FARMER ’ S LIVELIHOOD : A CASE STUDY FOR BUILDING
AGRICULTURE RESILIENT TO FLOODS IN. 8(2): 334–344.
collected (Kingra et al., 2016).
[4] Bannert, D. 2000. XXXIII: 113–117.
6 IRRIGATION MANAGEMENT [5] Bill, R. 2018. Remote sensing and GIS. gis.Science - Die Zeitschrift
Agriculture is one of the major water consuming fields and fur Geoinformatik (2): I. doi: 10.1201/9780429326349-15.
more than 90% water is used in agriculture sector for [6] Ferencz, C., Bognar, P., Lichtenberger, J., Hamar, D., Tarcsai, G.,
Timár, G., ... & Ferencz- Árkos, I. (2004). Crop yield estimation by
irrigating the crops. Therefore, irrigation management by satellite remote sensing. International Journal of Remote Sensing,
proper scheduling and precise application is necessary. 25(20), 4113-4149.
Remote sensing data integrated with ICT tools may play a [7] Glenn, E. P., Neale, C. M., Hunsaker, D. J., & Nagler, P. L. (2011).
Vegetation index‐based crop coefficients to estimate
key role in the former part, i.e. determining exact water evapotranspiration by remote sensing in agricultural and natural
demands and scheduling irrigations as per needs (Hunsaker ecosystems. Hydrological Processes, 25(26), 4050-4062.
et al., 2005; Sultan et al., 2008; Singh et al., 2009; Glenn et [8] Hunsaker, D. J., Pinter, P. J., & Kimball, B. A. (2005). Wheat basal
crop coefficients determined by normalized difference
al., 2011; Kamble et al., 2013; Yameen et al., 2019) Based vegetation index. Irrigation Science, 24(1), 1-14.
on the concept of ET and crop coefficients and water

Seventh International Conference on Aerospace Science & Engineering 166

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
[9] Joshi, C., J. De Leeuw, and I.C. Van Duren. 2002. REMOTE [17] Prasad, A. K., Chai, L., Singh, R. P., & Kafatos, M. (2006). Crop
SENSING AND GIS APPLICATIONS. (Graph 1). yield estimation model for Iowa using remote sensing and surface
[10] Israel, D.C. 2012. Economic and Welfare Impacts of Disasters in East parameters. International Journal of Applied Earth Observation
Asia and Policy Responses. (December): 553–599. and Geoinformation, 8(1), 26-33.
[11] Hunsaker, D. J., Pinter, P. J., & Kimball, B. A. (2005). Wheat basal [18] Seelan, S.K., S. Laguette, G.M. Casady, and G.A. Seielstad. 2003.
crop coefficients determined by normalized difference Remote sensing applications for precision agriculture : A learning
vegetation index. Irrigation Science, 24(1), 1-14. community approach. 88: 157–169. doi: 10.1016/j.rse.2003.04.007.
[12] Kamble, B., Kilic, A., & Hubbard, K. (2013). Estimating crop [19] Singh, R. A. N. D. H. I. R., Semwal, D. P., Rai, A., & Chhikara, R. S.
coefficients using remote sensing-based vegetation index. Remote (2002). Small area estimation of crop yield using remote sensing
sensing, 5(4), 1588-1602. satellite data. International Journal of Remote Sensing, 23(1), 49-56.
[13] Kaul, H.A., and I. Sopan. 2014. Land Use Land Cover Classification [20] Singh, R. K., & Irmak, A. (2009). Estimation of crop coefficients
and Change Detection Using High Land Use Land Cover using satellite remote sensing. Journal of irrigation and drainage
Classification and Change Detection Using High Resolution Temporal engineering, 135(5), 597-608.
Satellite Data. [21] Sultan, S., & Ahmad, I. (2008). Determination of daily regional scale
[14] Kingra, P.K., D. Majumder, U. Banga, K. Vishwavidyalaya, and S.P. actual evapotranspiration for Indus sub-basin using Landsat ETM+.
Singh. 2016. Application of Remote Sensing and Gis in Agriculture Pakistan Journal of Meteorology (Pakistan).
and Natural Resource Management Under Changing Climatic [22] Yameen, Q. (2019). Normalized Difference Vegetation Index as a
Conditions. (October). doi: 10.5958/2395-146X.2016.00058.2. Tool for Wheat Crop Coefficient and Evapotranspiration Estimation:
[15] Kumar, N., S.. Yamaç, and A. Velmurugan. 2015. Applications of A case study of Nankana Sahib District Pakistan. Acta Scientific
Remote Sensing and GIS in Natural Resource Management. J. Agriculture, 3, 32-39.
Andaman Sci. Assoc. 20(1): 1–6. [23] Zhang, F., Wu, B. F., & Luo, Z. M. (2004). Winter wheat yield
[16] Nain, A.S. 2017. Application of remote sensing and GIS in predicting for America using remote sensing data. Journal of
agrometeorology. J. Agrometeorol. 19(Special Issue): 21–25. remote sensing, 8, 611-617.

Seventh International Conference on Aerospace Science & Engineering 167

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Atmospheric Monitoring Using Artificial

Intelligence Models & Applications
Muhammad Mubeen Dr. Shahid Mahmood Moosa Bin Saif
PMAS AAUR, Rawalpindi PMAS AAUR, Rawalpindi PMAS AAUR, Rawalpindi
Islamabad, Pakistan Rawalpindi, Pakistan Islamabad, Pakistan
[email protected] [email protected] [email protected]

Abstract—in this paper we discuss Artificial intelligence pollution. Air pollution is a severe danger to society and the
and the role it plays in atmospheric monitoring through the environment, from haze hanging overhead cities to smoking
use of different models and sensors. With the discovery of within the house. Dementia, cardiovascular disease, lung
greenhouse gases and global warming atmospheric disease, and acute and chronic pulmonary illnesses are
monitoring has become an essential requirement. With the among the approximately 4.2 million fatalities caused by
help of Artificial intelligence we can fulfill that requirement. exposure to air pollution each year.[3]
AI tools revolutionize the monitoring process through
accurate data gathering and forecast information. In the use II. LITERATURE REVIEW
of AI there are risks and negative effects but its applications In the US, the EPA separates the most often employed
are vast. AI can help in Atmospheric monitoring through its models under three categories: dispersal models, photo
integration in machine and sensors. PSO-SVM model (The catalytic models, and receptor models. Air pollution models
Support Vector Machine (SVM) would be a methodology are a crucial technological tool for determining temporal and
for categorizing data that could be split linearly that uses spatial development, internal mechanisms, and root causes.
fundamental classification concepts) and MODIS model Simultaneously, it aids in the establishment of a quantitative
(This Moderate Resolution Image Spectroradiometer relationship among "pollution reduction" as well as "quality
(MODIS) is indeed a payloads imaging sensor. The improvement," and even the transfer between environmental
movement of aerosol-borne contaminants could now be management and planning toward a more quantitative, more
tracked overland, Atmosphere and ocean surfaces thanks to refined approach. [4]
NASA's Moderate-resolution Imaging Spectroradiometer) Ambient air models are an essential tool for analyzing and
were discussed in this project. resolving environmental issues. Air quality modelling
technology has advanced rapidly in recent years, as well as a
Keywords—Atmospheric Monitoring, AI Models, Air Quality, variety of ambient air models have already been widely used
Artificial Intelligence in the research of environmental air pollution. [5]
Various air quality systems have also subsequently become
I. INTRODUCTION
widely utilized in various fields, including necessary
Exposure to dirty and polluted air quality for a long scientific research, environmental management, especially
duration of time has long term health consequences, environmental impact assessment and decision-making. [6]
including the capacity of Lungs will be lost, and the function
of the lungs will be affected and decreased—development of
some diseases including such, emphysema, asthma, possibly III. ATMOSPHERIC MONITORING AND AI
cancer and bronchitis. Poor governance, a lack of research  Monitoring is done to maintain record the air
competence, ignorance, incompetence, national/international quality throughout order to collect data and
wars, and political turmoil have all contributed to inattention improve it.
to air pollution, despite millions of people dying as a result of  SO2, smoke, and suspended particles seem to be
it across the world. [1] the most refined indications.
Air pollution has a range of effects on agriculture.  The findings are gathered by either a central
Ground-level ozone could lower agricultural crop and organization and monitored on some kind of daily
commercial forest production, stunt tree seedling basis.
development and survival, and make plants more vulnerable  Spectrometry, conductance, impact and effect, and
to disease, insects, and some other environmental stresses aerometry are used to assess SO2, the primary
like harsh weather.[2] pollutant in urban and industrial environments.

Annually, an average of seven million people is going to Artificial Intelligence

die of air pollution throughout the world. According to WHO The replication of human intelligence by machines,
statistics, nine out of ten persons breathe air exceeding WHO particularly computer systems, is generally known as
recommended limits and contains high concentrations of artificial intelligence (AI). Computer algorithms, natural
pollutants, particularly low- especially middle-income language processing (NLP), voice recognition, and machine
nations bearing the brunt of the burden. The World Health vision are examples of major AI applications. Computer
Organization (WHO) is assisting countries in combating air software programs that analyze the audio and pictures

Seventh International Conference on Aerospace Science & Engineering 168

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

through surveillance inability to track persons, cars, objects, Through Air Quality System, the EPA keeps track of air
and situations would be used in artificial intelligence-based quality data.
sensor networks. Purple Air employs laser particle detectors PMS5003 as well
Atmospheric Monitoring as PMS1003. 0.3, 0.5, 1.0, 2.5, 5.0, and 10um small particles
Everywhere else on the planet, this same Copernicus are counted using all these sensors. This sensor uses a
Atmosphere Monitoring Service (CAMS) offers uniform as sophisticated algorithm to determine the PM1 from these
well as quality-controlled data concerning air pollution particulate counts. [8]
levels and health, solar power, greenhouse gas emissions, Electrochemical Sensors
and climate change. You must monitor environmental air The electric current passes through conductive material
whether you're to know precisely what would be in the when a magnetic field moves. This generated electrical
atmosphere you are inhaling. Scientists studying air current may be measured with an electromagnetic sensor.
pollution, including transportation, will benefit from a Electromagnetic (EM) (possibly magneto elastic) stressed
system comprising scientists and researchers monitoring sensor technology with steel cable stress monitoring.
pollution levels. [7] Microstructure, as well as temperature, have essential roles
 NASA monitors air pollution using satellites for magnetization but also magneto elasticity. The EM
circling the Earth. Aerosol data by NASA's Terra, sensor's construction and setup are depicted.
Aqua, as well as Suomi-NPP satellites are used by Optical particle sensors measure the scattered light by
air quality forecasts. particles to measure particulate pollution. Electromagnetic
 A new sensor dubbed known as Multi-Angle sensors measure the absorbance from infrared light to detect
Imaging system for Aerosols, called MAIA, has chemicals, including carbon monoxide and Carbon dioxide.
been being developed by NASA for use on a [9]
prospective spacecraft expedition. MAIA will aid NDIR Sensor
scientists in determining the size, composition, and When infrared radiation reacts with gas particles, infrared
amount of aerosols throughout our atmosphere. light at specific wavelengths is captured mainly through gas
molecules, producing these gas molecules to vibrate. Gas
AI to Measure Air Pollution sensors that use NDIR (Non-Dispersive Infrared) identify a
We can measure air pollution through different machines reduction in infrared transmission light that is proportional
and sensors through AI. by gas concentration.
Machine: The infrared source, sensor, optic filter, gaseous cell, and
 We can use machines to measure air pollution in machine learning circuits make up an NDIR sensor module.
different ways. We can teach machines by machine Two detectors and two optical filters having varying
learning how to measure air pollution, and AI does frequencies are positioned in the front of the detector in
the process of machine learning. some kind of a single measurement, multiple wavelengths
 Researchers can utilize support vector regression gas sensor monitor. For detecting even a target gas, infrared
(SVR), a common machine learning approach to light captured mostly by target gas travels through with an
anticipate pollutant and particle levels and estimate active filter with such a specific bandwidth. [10]
the overall air quality index. Operating Principle
Sensors: Mid-infrared radiation passing through with a sample gas
 We can use different sensors to detect air pollution causes gas molecules to resonate with infrared light at their
through AI algorithms. natural frequency inside this spectrum region in which the
intensity level of infrared seems to be equal to the natural
IV. AI APPLICATIONS USED FOR ATMOSPHERIC MONITORING intensity with gas molecules, actually results throughout
This Air Quality Monitoring challenge necessitates infrared absorption besides gas molecules inside the form
thorough, specific instance knowledge and perhaps even the with molecular vibration.
creation of practical, environmental format string event This Lambert-Beer law describes the relationship between
forecasting, as mandated by law. Artificial Intelligence (AI) infrared transmittance with gas concentration:
methodologies and related computer algorithms to supply 𝑇 = 𝐼 / 𝐼0 𝑒 − 𝑒𝑐𝑑
reliable results could be used to achieve both goals. VOC Sensor
Air pollution could then be managed more successfully These are the sensors that measure the VOC present in the
attributable to Artificial Intelligence (AI). Artificial atmosphere. VOCs are a broad category of carbon-based
intelligence (AI) can help cities monitor, communicate, (organic) compounds (chemicals) present in various man-
persuade, measure, identify, forecast, advise, decrease, and made & naturally present solids and liquids. Compounds are
control air pollution. called volatile because they evaporate quickly at ambient
Air Pollution Sensors temperature.
An air quality sensor is an electronic device that measures  Benzene/BTEX Monitor
the amount of polluted air around them. They are suitable  Series-900 VOC Fixed Monitor
for both indoor as well as outdoor use. Through all the Air
Quality System, the EPA keeps track of air quality data. Aeroqual VOC Sensors
An air quality sensor is an electronic device that measures Photoionization Detector (PID) and Gas Sensitive
the amount of polluted air around them. They are Semiconductor (GSS) are two main types of VOC
appropriate for intended as well as environmental use. measuring sensors offered by Aeroqual. Both sensors

Seventh International Conference on Aerospace Science & Engineering 169

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

generally non-selective, which means sensors can measure a MODIS

wide range of VOCs. This Moderate Resolution Image Spectroradiometer
They were calibrated against isobutylene to achieve this, (MODIS) is indeed a payloads imaging sensor. The
recognizing and responding to such a wide variety of VOCs movement of aerosol-borne contaminants could now be
with variable degrees and sensitivity. tracked overland Atmosphere and ocean surfaces thanks to
Each kind has a measuring range that has been standardized NASA's Moderate-resolution Imaging Spectroradiometer
against a subset of VOCs. VOC detectors of the PID type (MODIS). As little more than a result, CSR and MOD
seem to be more durable and can detect a broader range of employees successfully adapted MODIS statistics to
VOCs. [11] numerous types of pollutants and have an influence on
Measuring Target VOC Sensors Texas air quality on even a regular basis. The findings show
Targeted VOC measures are helpful in regulated contexts that MODIS data-related products can identify and monitor
(e.g., laboratory, research) when the targeted chemical is pollution movement.
employed, but control process and safety and health MODIS offers the platform for creating advanced analytical
concerns are present. Any defined correction factor has been products that might establish a cost-effective yet accurate
hired to assess a specific VOC to find the exact value using environmental monitoring system used in concert with
isobutylene used during sensor calibration. Furthermore, no ground-based measurements. [15]
additional VOCs should really be introduced in the
V. AI MODELS USED FOR ATMOSPHERIC MONITORING
monitoring area to avoid the border. The sensor may be
adjusted towards the targeted VOC and use a calibrated gas The goal for the air pollution model is always to estimate
of both the target compound for increased accuracy. [12] environmental air pollutants generated by industrial
Measuring Mixed VOC Sensors operations, unintentional releases, or transportation, for
To evaluate air pollution, mixed VOC measurements are example. Air quality modelling has been used to determine
widely employed to indicate the existence of VOCs within the pollutant's higher concentrations and even the source of
uncontrolled situations (e.g., workplaces, outdoor). In just unusually significant concentrations. The extra contribution
this case, the assessment is utilized as a proxy for the toward the current load may be predicted in advance with
presence and concentration of VOCs within the atmosphere. projects throughout the planning process, and emission
Because some VOCs might impact the readout more often circumstances can be optimized.
than others, this phenomenon known as cross-interference Mean annual measurements and short-time maximum values
assessment is just a guide for air quality. have been of interest depending upon these air quality
criteria established for both the contaminant in
CO2 Sensor
consideration. The proposed methodology can be used to
The carbon dioxide sensor monitors the quantity of infrared
simplify it:
(IR) radiation absorbed per carbon dioxide molecule can
determine atmospheric carbon dioxide concentrations.  There is no variation in emission circumstances
Carbon dioxide measurement is essential for monitoring throughout time.
various industrial operations as well as indoor  During transportation, the weather conditions do
environmental quality. not alter.
NDIR sensors seem to be spectroscopic detectors that use  Winds speed exceeding one m/s.
carbon dioxide's unique absorption to detect this in a Stochastic Model
gaseous phase. The infrared detector, the interfered filter, a Stochastic modelling is a type of financial modelling and is
lighting tube, and an infrared emitter are all part of it. This used to aid in the decision-making process. Utilizing random
gas is injected or dispersed into another light pipe, and also variables, this form of modelling estimates the likelihood of
the electronics detect the wavelengths and optical emission. possible possibilities under various situations.
Chemicals carbon dioxide sensor is composed up of polymer Data is presented, and results are predicted using stochastic
and heteropolysiloxane based sensitive sheets. They modelling, accounting for a certain amount of volatility or
consume extremely little energy and may be made smaller to unpredictability. Companies across a wide range of sectors
fit within microelectronic devices.[13] can use stochastic modelling to enhance their business
Wireless Sensor Network operations and boost profits. Planners, strategists, and
Current air pollution measuring methods are costly and portfolio managers within the financial sector utilize
spatially constrained. Atmospheric monitoring in either a stochastic modelling to control their rights and obligations
larger area is indeed not possible due to these constraints. and optimize their portfolios. [17]
However, combining a contemporary low-cost sensor with A Case Study in which stochastic model in atmospheric
either a wireless sensor network (WSN) is necessary to monitoring is described given below:
receive actual data from multiple sites and generate a Applying AI-enabled methodology IOT pollution detectors,
comprehensive pollution map. the stochastic model from this work was effectively used in
The primary objectives of this work are to create a low-cost environmental dynamics.
multi-sensor module for measuring air pollution, along with The results show that the suggested technique is well
WSN protocols allowing data collecting and data adapted to determining air quality in various parts of a city
aggregation. Another board like this has been designed and and forecasting PM concentration. Its resulting model
manufactured by us. Standardization from each sensor is employs previous atmospheric data to determine the causal
performed by comparing data transmission utilizing some of relationship between atmospheric variables and PM
the numerous radio technologies alongside fault-tolerant concentrations and correctly forecast micro pollutant
topology control to preserve data correctness. [14] concentrations.

Seventh International Conference on Aerospace Science & Engineering 170

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

MACC Model pollutant concentrations in the atmosphere must be

MACC (Monitoring Atmospheric Composition and Climate) examined.
monitors the structure of the Upper atmosphere that predicts
air quality using a weather prediction technique based upon To establish a new PSO method for obtaining the best SVM
modelling and data integration. parameters. The automated optimization technique is
It offers data that is critical for understanding environmental possible with our suggested PSO–SVM model, which also
issues as well as verifying and refining climate models. solves the SVM's precursor concentration problem. [19]
There is also knowledge on how to preserve one's health and VI. IMPACT OF ARTIFICIAL INTELLIGENCE (AI)
make the most of solar power generating. MACC is working
on creating real atmospheric environmental benefits for Artificial intelligence has the potential to significantly
operational use. [16] enhance the efficiency of our workplaces while also
augmenting the job that humans can accomplish. Whenever
AI takes away mundane or risky activities, it frees
understand the more significant labor to focus on tasks that
need creativity as well as empathy, amongst many other
things.
Environmental Impact of AI
By monitoring fuel reduction emissions, CO2 removal,
assisting in the development of greener transportation
systems, monitoring deforestation, as well as forecasting
extreme weather events, AI can speed global efforts to
preserve the environment and save resources. [20]
Increase of CO2
AI would be a huge carbon emitter in and of itself. The
scientists calculated that perhaps the carbon footprint from
training any single large language model is about 300,000
Figure 1
kilograms of carbon dioxide emissions when they converted
this energy usage towards approximate carbon emissions
Some Points of MACC Model and power expenses.
 The acquisition and pre-processing collecting Not Sustainable Use of Energy
observational data and the supply of predictions The development in acceptance of machine learning, a field
regarding surface emissions from essential species of artificial-intelligence methods that relies on processing
are the two primary components for MACC. massive quantities of data, is primarily to blame for the
 Data assimilation and modelling systems that offer current increase throughout AI's power usage. "Deep
fundamental monitoring and warning products and learning models, which are also substantial computational
technologies that can provide derived materials are formalism with thousands of millions or billions more
included in regional and global processed chains. features, would be used in modern machine learning
 Providers of downstream services and some other techniques.
consumers' interfaces.
VII. CONCLUSION
PSO-SVM Model In this world where technology is evolving at a high rate in
The Support Vector Machine (SVM) would be a order to match the needs and demands of people there is a
methodology for categorizing data that could be split necessity to keep check and balance of what these demands
linearly that uses fundamental classification concepts. By are costing us. With high demand rate there is high risk
integrating kernel principles within high dimensional space, involved. This risk could have disastrous effects in the near
SVM was intended to operate with non-linear situations. or upcoming future so we need to monitor this risk with
The researchers employed Particle Swarm Optimization most accurate and efficient technology. Artificial
(PSO) and improved SVM performance to overcome their intelligence is that technology. In assessing different
drawbacks. [18] models and applications of Artificial Intelligence for
Case Study atmospheric monitoring it was observed that AI integration
 That is, taking these meteorological factors into in monitoring system can revolutionize the monitoring
account would increase the SAPCF accuracy process. With AI monitoring system we can collect accurate
substantially. data of atmospheric gases and their concentration with
 However, different climatic conditions may have minimum error possibility. We can also predict any changes
varying impacts on the concentrations of in the atmospheric composition based in the data collected.
contaminants in the atmosphere. Heavy winds, for AI application in monitoring system can help achieve
example, may transport pollutants across vast areas. atmospheric stability.
 As a result, wind speed and direction are more
essential when compared to many other REFERENCES
meteorological parameters. The same climatic
element might have distinct impacts in different [1] "Spare The Air," 2021. [Online]. Available:
parts of the world. http://www.sparetheair.com/health.cfm.
 As a result, the influence of climatic factors on [2] "Frotiers In Public Health," Frotiers In, 2021. [Online]. Available:

Seventh International Conference on Aerospace Science & Engineering 171

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

https://www.frontiersin.org/articles/10.3389/fpubh.2020.00014/full. [13] "AZO Sensors," 2021. [Online]. Available:

https://www.azosensors.com/article.aspx?ArticleID=234.
[3] World Health Organization, 2021. [Online]. Available:
https://www.who.int/health-topics/air-pollution#tab=tab_1. [14] 2. Bathiya et al., "Air pollution monitoring using wireless sensor
network," IEEE, 2017.
[4] W. e. al, "Air pollution terrain nexus: A review considering energy
generation and consumption," Renewable and Sustainable Energy [15] 2. Hutchison et al., "Applications of MODIS satellite data and
Reviews, 2019. products for monitoring air quality in the state of Texas," Atmospheric
Environment, pp. 2403-2412, 2003.
[5] A. e. al, "Airborne PM10 and lead concentrations at selected traffic
junctions in Khyber Pakhtunkhwa, Pakistan: Implications for human [16] 2. Pour et al., "Spatiotemporal regionalization of atmospheric dust
health," Atmospheric Pollution Research, pp. 1320-1325, 2019. based on multivariate analysis of MACC model over Iran,"
Atmospheric Research , 2021.
[6] Bei et al., "Impacts of local circulations on the wintertime air pollution
in the Guanzhong Basin, China," Science of The Toral Environment, [17] 2. Lee et al., "A stochastic model of particulate matters with AI-
2017. enabled," Microelectronic Engineering, 2020.
[7] 2. Bei et al., "Impacts of local circulations on the wintertime air [18] 2. Chen et al., "A hybrid PSO-SVM model based on clustering
pollution in the Guanzhong Basin, China," Science of The Total algorithm for short-term," Technological Forecasting & Social
Environment, 2017. Change, 2019.
[8] Ilocis, [Online]. Available: [19] 2. Chen et al., "A hybrid PSO-SVM model based on clustering
http://www.ilocis.org/documents/chpt55e.htm. algorithm for short-term," Technological Forecasting & Social
Change, 2019.
[9] Microsoft, 2021. [Online]. Available: https://www.microsoft.com/en-
us/education/hackingstem/electromagnets. [20] "PC," 2020. [Online]. Available: https://www.pcmag.com/news/ai-
could-save-the-world-if-it-doesnt-ruin-the-environment-first.
[10] "FIGARO," 2018. [Online]. Available:
https://www.figaro.co.jp/en/technicalinfo/principle/ndir-type.html.
[11] Aeroqual, 2021. [Online]. Available: https://www.aeroqual.com/voc-
sensors-monitors.
[12] "Aeroqual," 2021. [Online]. Available:
https://www.aeroqual.com/voc-sensors-monitors.

Seventh International Conference on Aerospace Science & Engineering 172

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Unsupervised Classification of Hyperspectral

Images using PCA and K-Means
Ayesha Malik Mamoona Waheed
Department of Electrical Engineering Department of Electrical Engineering
Institute of Space Technology Institute of Space Technology
Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected]

Abstract—The visualization of hyperspectral images in that will help find the number of regions in hyperspectral
display devices, having RGB colour composition channels is image [6].
quite difficult due to the high dimensionality of these images.
Thus, principal component analysis has been used as a The remainder of this paper is broken down as follows.
dimensionality reduction algorithm to reduce information Section 2 presents the literature review of work done in the
loss, by creating uncorrelated features. To classify regions in field. Section 3 defines the problem statement.
the hyperspectral images, K-means clustering has been used Methodology is explained in Section 4. Dataset used and
to form clusters/regions. These two algorithms have been classification results form Section 5. Finally, concluding
implemented on the three datasets imaged by AVIRIS and remarks alongside future work are given in Section 6.
ROSIS sensors.
II. LITERATURE REVIEW
Keywords—Hyperspectral image, HSI, PCA, K-means Prior to remote sensing, geological surveys were done
clustering, unsupervised, classification, bands, satellite, ROSIS, manually. Now, technological advancements help capture
AVIRIS
hyperspectral images via satellites and enable algorithms to
I. INTRODUCTION help classify the data in a matter of moments [4].
Remote Sensing is the data collection of earth’s surface, Colour images have intensity values in three bands i.e.
by measuring the reflected signal at various wavelength RGB, whereas, hyperspectral images contain more than
bands to form a spectral signature from objects at long hundred bands in discrete intervals from visible to infrared
distances, via either of the two techniques: infrared sensing region [7]. HSI obtains the spectrum for each pixel in the
or optical remote sensing (via satellite’s optical sensor) [1]. image frame to identify target areas.
Based on the number of bands, satellite images are classified Hyperspectral images are characterized by two
into: panchromatic, multispectral and hyperspectral images. resolution types, namely [5]: spatial and spectral. Spatial
Attaining hyperspectral image datasets depends on resolution is the relationship among image pixels and is
imaging spectrometers that image in continuous, narrow inversely proportional to patch size i.e. smaller the patch,
regions of electromagnetic waves [2]. Due to which, they greater the details. Whereas, spectral resolution is measured
have high spectral resolution that describes boundaries and as the number of bands and range of electromagnetic
composition of an object and can help in image spectrum, by the sensor.
classification, which is the most active part of research. High dimensional hyperspectral data lacking labelled
Classifying hyperspectral images may seem difficult due to samples can result in Hughes phenomenon i.e. reduced
their high dimensionality, missing labels, variation in data accuracy in classification. This can be overcome by
due to spatial locality and poor image quality due to noise techniques of dimensionality reduction e.g. Principal
or background interference etc. component analysis (PCA), Discrete wavelet transform and
Hyperspectral image classification methods are as such Independent component analysis (ICA) [1]. The first few
[3]: supervised, unsupervised and semi-supervised. principal components of PCA result in 70 percent correct
Supervised classification makes use of prior data classification rate [6].
knowledge. An example of this method is the support vector Previously, only spectral information was being used to
machine (SVM). On the contrary, unsupervised achieve hyperspectral image classification upon high
classification is not based on any prior data knowledge. A dimensional data via support vector machine (SVM) , neural
common method is K-means clustering and its aim is to networks etc. [3]. Whereas, some classification approaches
obtain a meaningful output from the set of unlabelled data work on reducing high-dimensional data first and then using
via clustering/grouping [4]. clustering methods.
Hyperspectral imaging (HSI) is used for [5]: agricultural Supervised deep classification of hyperspectral images
and water resources control (e.g. identifying crops, lakes returns almost 100 percent accurate results [8].
etc. [4]), document imaging and mineralogical mapping of Classification methods are based on parametric and non-
earth surface. parametric classifiers or supervised and unsupervised
The large hyperspectral data, increases processing classifiers [9]. Each having its own limitations [10].
complexity and time. The solution is to effectively reduce Classifying regions help study the geographic location and
data to apply various processing methods e.g. classification, boundaries of a certain patch that is imaged by the satellite.

Seventh International Conference on Aerospace Science & Engineering 173

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Existing work is related to either PCA related schemes Fig. 1. Explained Variance and Elbow graph for Pavia dataset
or discriminant analysis-based schemes. One of the
prominent work done is based on both of these schemes, that  matplotlib: to create figures and plots
is by using PCA for dimensionality reduction and SVM for  spectral: to use functions for hyperspectral data
classification [11]. processing
SVM, a supervised learning method, is effective in high
 import h5py: to import data of HDF5 binary format
dimensional spaces and is accurate for classification. It
requires a smaller number of training samples and labels [1]. B. Dimensionality Reduction via PCA
However, SVM is computationally expensive. Thus, PCA is a classical, unsupervised dimensionality
unsupervised classification methods such as K-means reduction method which does not require labels [14] and
clustering , K- nearest neighbour, Neural networks are a preserves most of the spectral information in a compact
better solution where pixels are to be assigned to clusters number of principal components [1]. Its governing principal
without using prior knowledge [12]. K-means accounts for is that the neighbouring bands of hyperspectral images are
78.3% accuracy for classification [13]. highly correlated, which would mean that they often convey
III. PROBLEM STATEMENT almost the same information. Most of the information may
only be contained in the first few bands [6].
During hyperspectral data analysis, the following
challenges are faced: The basic principle used in PCA is the eigenvalue
decomposition of the covariance matrix [1]. The
 High computational cost eigenvalues represent the variance in the direction of the
 Storage of a few hundred megabytes required eigenvector.
It must be remembered that PCA is a dimension
 Redundancy due to correlation of neighbouring
reduction tool, not a classifier. Thus, a classifier is fit onto
image bands. This high dimensional redundant data
the PCA-transformed data.
needs to be catered for, to efficiently apply
processing techniques [14]. To choose the number of principal components required
to describe a dataset, explained variance ratio graph was
The aim of this research paper was to implement
plotted (number of components vs. cumulative explained
dimensionality reduction and K-means classification, in
variance). Components which had greater than 95 percent
order to find the number of regions present in each
cumulative explained variance were chosen i.e. 8 principal
hyperspectral image and colour label them to differentiate
components in this case for all three hyperspectral datasets,
those classified regions.
as shown in Fig. 1 for Pavia. The curve depicts how much
IV. METHODOLOGY of the total, 102-dimensional variance is contained within
the first N components. The same approach was used to
To achieve the goal of this paper, two-stage process was select principal components for Cuprite and Moffett.
implemented i.e. first, data dimensionality was reduced,
followed by K-means clustering. PCA is a standard C. K-means clustering
dimensionality reduction algorithm for hyperspectral K-Means being an unsupervised method, groups data
images [12] that aims to remove the correlation among into ‘k’ number of user defined clusters. To help decide the
bands [1]. number of clusters, elbow method was used i.e. for each
Clustering works by grouping pixels and iteratively value of k (e.g. in the range 1 to 10) , the sum of squared
updating the centroid of a group. The centroid represents distances from each data point to its centre(distortion) was
mean spectral signature of all pixels in a cluster [8]. calculated.

A. Importing libraries in Python to load and process data Elbow is the point of inflection on the curve which is the
optimal value of k. According to this method, k=5 for Pavia
 numpy: for multidimensional arrays and matrices (as shown in Fig. 1) and Cuprite data sets, k= 6 for Moffett
 scipy: to load .mat file (hyperspectral dataset) [3].

 sklearn: to import PCA and K-means

Seventh International Conference on Aerospace Science & Engineering 174

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

V. EXPERIMENTATION AND RESULTS

In accordance with the problem statement, three
hyperspectral images were utilized [1]. The classification of
different regions was done for the following datasets
acquired by their respective sensors:
A. ROSIS(Reflective Optics System Imaging
Spectrometer) Pavia Centre: covers Pavia, northern
Italy having 102 spectral bands.
B. AVIRIS (Airborne Visible Infrared Imaging
Spectrometer) datasets:
1) Cuprite: covers Cuprite in Las Vegas, U.S having
224 channels.
2) Moffett Field: covers Moffett airfiled, CA, U.S.
having 224 channels.

Unsupervised classification was applied on the above (a) (b)

stated hyperspectral images using python libraries. The
datasets have the following dimensions i.e. rows, columns Fig. 2. Pavia results (a) 8 principal components after PCA. (b) K-means
and bands, respectively: clustering result.

 Pavia: 1096*715*102
 Cuprite: 512*614*224
 Moffett: 753*1923*224
The reshape command was used to reshape the data, by
multiplying the actual rows and columns to form the current
rows and the original bands were retained as columns e.g.
Pavia was reshaped to 783640*102.
To visualize the explained variance ratio, its cumulative
sum was plotted which helped select the number of principal (a) (b)
components to be retained as shown in Fig. 2 (a), Fig. 3 Fig. 3. Cuprite results (a) 8 principal components after PCA. (b) K-means
(a), Fig. 4 (a), that display the images with 8 principal clustering result.
components.
After reducing the dimensions of all three datasets, k-
means was applied using the elbow method to assign ‘k’
optimal clusters. Fig. 2 (b), Fig. 3 (b) and Fig. 4 (b) are the
colour-labelled clusters formed for Pavia, Cuprite and
Moffett, respectively. The axes of Fig. 2 to Fig. 4 represent
the rows and columns of the respective images in 2D.
The ‘KMeans’ command used in python formed clusters
with equal variance to minimize ‘within cluster sum of
(a)
squares’ (WCSS). For data with high dimensionality,
Euclidean distances increase. Thus, implementing a
dimensionality reduction algorithm like Principal
component analysis (PCA) before K-means clustering can
help overcome this problem.
The number of clusters/regions formed for Pavia are 4
after 15 iterations, whereas, for Cuprite and Moffett there
are 5 clusters after 15 iterations. In simple words, the
algorithm segments the image by clustering pixels into
classes which have spectral similarity e.g. roads, trees, water
bodies, soil etc.

(b)
Fig. 4. Moffett results (a) 8 principal components after PCA. (b) K-means
clustering result.

Seventh International Conference on Aerospace Science & Engineering 175

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

VI. CONCLUSION AND FUTURE WORK [10] I. A. Huqqani and K. Khurshid, “Comparative study of supervised
classification of urban area hyperspectral satellite imagery,” J. Sp.
Hyperspectral images contain wide-ranged spectral Technol., vol. 4, no. 1, pp. 1–9, 2014.
information that is used to identify and differentiate objects. [11] A. Rehman, “Project Report: Hyperspectral Image Principal
Classification is the process of assigning objects to classes Component Analysis and Classification.” 2018.
with homogeneous characteristics. This paper is firsthand [12] F. Masood, I.-H. Qazi, and K. Khurshid, “Saliency-based
examination of: dimensionality reduction via PCA and visualization of hyperspectral satellite images using hierarchical
unsupervised classification via K-means to cluster the three fusion,” J. Appl. Remote Sens., vol. 12, no. 04, p. 1, 2018, doi:
10.1117/1.jrs.12.046011.
hyperspectral datasets with the goal of finding number of
[13] S. A., “Hyperspectral Image Classification Using Unsupervised
regions based on some similarity in the datasets. Initially, Algorithms,” Int. J. Adv. Comput. Sci. Appl., vol. 7, no. 4, pp. 198–
SVM was being used for clustering. However, due to 205, 2016, doi: 10.14569/ijacsa.2016.070425.
computational expenses, it was dismissed. With K-means, [14] K. Kotwal and S. Chaudhuri, “Visualization of hyperspectral images
clusters have been formed without any prior knowledge i.e. using bilateral filtering,” IEEE Trans. Geosci. Remote Sens., vol. 48,
labels. no. 5, pp. 2308–2316, 2010, doi: 10.1109/TGRS.2009.2037950.
[15] H. A. Khan, M. M. Khan, K. Khurshid, and J. Chanussot, “Saliency
One of the limitations of PCA for dimensionality based visualization of hyper-spectral images,” Int. Geosci. Remote
reduction is the loss of spectral information, as a result of Sens. Symp., vol. 2015-Novem, pp. 1096–1099, 2015, doi:
selecting principal components by only considering 10.1109/IGARSS.2015.7325961.
variance and not the spectral position [15]. Another
approach of efficiently clustering data and finding number
of regions can be: supervised classification, by breaking the
dataset into training and test sets.

DATA SET

[1] M Graña, MA Veganzons, B Ayerdi, “GRUPO DE

INTELIGENCIA COMPUTACIONAL(GIC),” [Online].
Available:
http://www.ehu.eus/ccwintco/index.php/Hyperspectral_Remo
te_Sensing_Scenes. [Accessed 8 May 2021].

REFERENCES
[1] T. Subba Reddy and J. Harikiran, “Hyperspectral image
classification using support vector machines,” IAES Int. J. Artif.
Intell., vol. 9, no. 4, pp. 684–690, 2020, doi:
10.11591/ijai.v9.i4.pp684-690.
[2] B. Rasti et al., “Feature Extraction for Hyperspectral Imagery: The
Evolution from Shallow to Deep: Overview and Toolbox,” IEEE
Geosci. Remote Sens. Mag., vol. 8, no. 4, pp. 60–88, 2020, doi:
10.1109/MGRS.2020.2979764.
[3] W. Lv and X. Wang, “Overview of Hyperspectral Image
Classification,” J. Sensors, vol. 2020, 2020, doi:
10.1155/2020/4817234.
[4] C. R. Addanki, S. A, and V. R. A, “Study of the Clustering
Algorithms for Hyper Spectral Remote Sensing Images,” J.
Hyperspectral Remote Sens., vol. 10, no. 2, p. 117, 2020, doi:
10.29150/jhrs.v10.2.p117-121.
[5] M. J. Khan, H. S. Khan, A. Yousaf, K. Khurshid, and A. Abbas,
“Modern Trends in Hyperspectral Image Analysis: A Review,”
IEEE Access, vol. 6, no. June, pp. 14118–14129, 2018, doi:
10.1109/ACCESS.2018.2812999.
[6] C. Rodarmel and J. Shan, “Principal component analysis for
hyperspectral image classification,” Surv. L. Inf. Sci., vol. 62, no. 2,
pp. 115–122, 2002.
[7] S. Ranjan, D. R. Nayak, K. S. Kumar, R. Dash, and B. Majhi,
“Hyperspectral image classification: A k-means clustering based
approach,” 2017 4th Int. Conf. Adv. Comput. Commun. Syst.
ICACCS 2017, pp. 0–6, 2017, doi:
10.1109/ICACCS.2017.8014707.
[8] H. Yadav, A. Candela, and D. Wettergreen, “A Study of
Unsupervised Classification Techniques for Hyperspectral
Datasets,” IGARSS 2019 - 2019 IEEE Int. Geosci. Remote Sens.
Symp., pp. 2993–2996, 2019, doi: 10.1109/igarss.2019.8900501.
[9] S. Sattar, H. A. Khan, and K. Khurshid, “OPTIMIZED CLASS-
SEPARABILITY IN HYPERSPECTRAL IMAGES,” in 2016 IEEE
International Geoscience and Remote Sensing Symposium
(IGARSS), 2016, pp. 2711–2714, doi:
10.1109/IGARSS.2016.7729700.

Seventh International Conference on Aerospace Science & Engineering 176

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Implementation of the Maximum Cross Correlation

(Mcc) Method In Python for Ocean Current
Estimation Over the Arabian Sea
Sheeza Khalid Waqas A. Qazi
Geospatial Research and Education Lab Geospatial Research and Education Lab
Dept. of Space Science Dept. of Space Science
Institute of Space Technology, (IST) Institute of Space Technology, (IST)
Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected]

Abstract— Ocean currents are an important parameter in which influence the sea current. Moreover, salinity, moon’s
physical oceanography because they act to neutralize the gravitational pull and winds also influence the ocean currents.
temperature difference between different areas in the oceans. The Earth’s rotation influences currents through the Coriolis
These parameters are important for reduction of shipping cost, Effect acts perpendicular to the direction of motion of the
and also the dispersal of many life forms also depends on them.
ocean current and to the earth’s rotational axis [2].
The Arabian Sea switches direction annually due to strong
monsoon winds and results in changing rainfall patterns This study uses Maximum Cross Correlation Method (MCC)
seasonally. A few studies have mapped ocean currents in the to estimate ocean currents over the Arabian Sea. In 1986,
Arabian Sea using radar altimetry data, however these methods Emery et al.[3] developed the technique to compute advective
generate only geostrophic currents, have low resolution, and also sea surface velocities using cross correlations between
cannot function well near the coasts. A relatively higher sequential infrared satellite images. In 1990, Tokmakian et al.
resolution total ocean currents measurement from satellites does [4] evaluated the method of estimating sea surface velocities
not exist for the Arabian Sea over a long time-series. The from sequences of Advance Very High Resolution
Maximum Cross-Correlation (MCC) method has been used with Radiometer (AVHRR) and Coastal Zone Color Scanner
Advanced Very High Resolution Radiometer (AVHRR) data in
(CZCS) images using MCC Method. In 1992, Kelly &
various studies to extract currents in different regions of world.
For this study, we have implemented the MCC method in Strub[5] conducted the research in which two methods of
Python with the aim of utilizing AVHRR optical remote sensing estimating surface velocity vectors from AVHRR data were
image sequential pairs over the Arabian Sea for ocean current applied to the same set of images and the results were
generation. The AVHRR level-1B images are first corrected for compared with in situ and altimeter measurements. In 2002,
brightness temperature and then re-projected, to be ready for Bowen et al.[6] performed research for ocean surface
ingestion in the MCC algorithm, which is implemented in circulation estimation by automated tracking of thermal
Python as a combination of different functions on pair of images infrared features in pairs of sequential satellite imagery. In
to extract ocean currents. Results represent the velocity vector 2007, Crocker et al. [7] on the basis of many existing studies
of ocean currents in the study area. The estimated velocity
on MCC analysis developed the comparison analysis for the
vectors are then masked with a threshold for the correlation
factor and filtered for removal of outliers, for the generation of currents retrieval but the main focus of this study is to
the final ocean currents map. evaluate the potential of technique which uses velocities of
the current acquired from thermal imagery as compared the
Keywords— Advanced Very High Resolution Radiometer technique of MCC. In 2016, Warren et al. [8]attempted to
(AVHRR), Arabian Sea, Brightness Temperature, Maximum automatically estimate surface current velocities from
Cross Correlation (MCC) Method, Ocean Currents. satellite-derived thermal or visible imagery but faced
limitations in the availability of cloud free sequential images.
I.INTRODUCTION Published in 2017, the work of Heuze et al. [9] analyses the
Satellite oceanography is the field of remote sensing in which accuracy of such an estimation of surface current fields
monitoring and study of ocean is focused. Platforms of retrieved with the maximum cross correlation (MCC)
satellite give advantage to sensors in viewing oceans. These method, here used to track patterns of Advanced Very High
advantages include wide synoptic coverage at higher Resolution Radiometer (AVHRR) brightness temperature
resolution, and daily sampling, which can produce good time between 224 pairs of consecutive images taken between
series analysis over a period of many years. These advantage January and December 2015 in the western Mediterranean
can help in differentiating the remote sensing from all other Sea. Another 2017 study by Heuze et al. [10] focused on using
such techniques. Satellite sensors are used to monitor oceans model outputs instead of satellite images, to assess biases
are of two types active and passive. Visible sensors provide induced by this method, for four different seas of Western
various types of data like suspended particulates, upwelling Europe. In October 2017, Liu et al. [11] studied the
and chlorophyll. Various parameters like salanity, surface computational results of data analysis of two satellites with
temperature and its roughness are measured by active the help of MCC technique. One of these satellites is known
microwave sensors. Whereas, sea surface temperature is as Visible Infrared Imaging Radiometer Suite (VIIRS),
measured more accurately by passive infrared sensors. [1] whereas, the other one is known as Moderate Resolution
In the field of satellite oceanography, the study of ocean Imaging Spectroradiometer (MODIS).
currents is an important aspect to deal with oceanic Rainfall patterns in the Arabian Sea changes annually due to
phenomena. The earth’s rotation is one of the parameter various parameters; and directional change of surface

Seventh International Conference on Aerospace Science & Engineering 177

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

circulation is the most important of all these. A few studies as This correction is performed using band math of SNAP
Kalyani et al.[12] and Eremeev et al.[13] have mapped ocean software. Further we perform the reprojection of the pair
currents in the Arabian Sea using radar altimetry data, images. As shown in Fig.2.
however these methods generate only geostrophic currents,
have low resolution, and also cannot function well near the
coasts. A relatively higher resolution total ocean currents
measurement from satellites does not exist for the Arabian
Sea over a long time-series. Hence for this study, we have
implemented the MCC method in Python with the aim of Fig. 2. Data preprocessingn (Non-corrected image, corrected image and
utilizing AVHRR optical sequential pairs images in the the reprojected image (from left to right)).
region of the Arabian sea for ocean current generation.
The study area map of the Arabian Sea is shown in Fig.1. III.OCEAN CURRENT ESTIMATION FROM MCC
METHOD
The pre-processed pair of images are exported in GeoTiff
format for the further processing in Python where various
steps are performed for the extraction of velocity vectors
whose detail sequence is shown in the flow chart in Fig.3..

Selected bands
Read GEOTIFF Read the metadata Compute lag
named as img1 and
images to select the band difference
img2

Apply padding for

Perform image sub Read Map
Perform cloud mask the optimize image
division coordinates
size

Initialization of Compute Store the vectors

velocity arrays Localize the center correlation and maxcorr matrix

Plot the ocean Return the current Regularize the

Plot Maxcorr plot
current velocities outlier

Fig. 1. Study Area Map. Import the vectors,

coordinates and
plots

According the change of the season the winds of monsoon

change their directions from north-east to south-west. It is
noted that the direction of wind in summer season is usually Fig. 3. Sequence of MCC Method in Python.
from sea towards land in a clockwise manner, whereas in
winter season these winds blow from south-west parts of Asia For the estimation of ocean currents first we select the
towards the sea. These winds cause heavy showers in the corrected brightness temperature band. We need to compute
regions of countries of sub-continent like Pakistan, India and the time difference between the pair images, so after
Bangladesh. These are the cyclic effects which not only computing time lag we apply padding to optimize the image
influence the climate of land but they have impacts on the size. Further we extracted the auxiliary data then develop the
Arabian Sea as well. Ocean temperatures are also influenced window filter to divide images in sub section for the ease of
by the local winds due to the landlocked property of the processing. After that for cloud masking the cut off threshold
Arabian sea in its north. [14]. of value -1 for normalized brightness temperature is executed
with reference to Heuze.et al [9].
II.DATA PRE-PROCESSING METHOD
After initializing the velocity arrays to store values of vector
The Channel 4 data which is related to the temperature of we performed the normalized cross correlation function to
black body is used in this study, which is also known as the compute the velocity vectors for the plotting of ocean
brightness temperature with a wavelength of 10.8 µm of currents.
Local Area Coverage-1.1km (LAC) format level 1B product The MCC method utilizes two images. It cross-correlates one
of AVHRR data. The sequential pair of 27th May, 2019 with image by the other image through a specified template
time of capturing of first image is 01:03:56 GMT and that of window and corresponding search window. Displacement
second image is 03:04:28 GMT are firstly corrected for the vector is calculated in this technique, which is divided by the
brightness temperature using the equation: time lapse of these two images, and hence it yields the
𝑇 = 𝐴 + 𝐵. T* (1)
velocity of the ocean currents. The computed vectors are
*
Where T is the Brightness temperature band (channel 4) finally regularized to remove any outlier in computation and
while A and B are correction factors. regularized velocity vectors are at last used to plot the
estimated ocean currents over the Arabian Sea.

Seventh International Conference on Aerospace Science & Engineering 178

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

IV.RESULTS AND DISCUSSIONS

The parameters whose setting effect the accuracy of the MCC

method are template window size, search window size and
maximum retrieved velocity; in this work we used template
window and search window size as 10m while the maximum
velocity is set by 1ms-1. The estimated ocean current map for
the sequential pair image of 27th May, 2019 (time of capturing
of first image is 01:03:56 GMT and that of second image is
03:04:28 GMT) is shown Fig.4 and the average maximum
correlation co-efficient value for the result is 0.33 ms-1
whereas the 2D correlation map is also given below depicting
the value of maximum correlation co-efficient for each
vector. The range of the co-efficient is 0-1 so higher the
correlation co-efficient value better the derived current and is Fig. 4. Derived ocean current over the Arabian Sea and correlation plot of
more accurately directed. derived ocean currents (from top to bottom).

In this study the major difficulties faced are that first the V.FUTURE RECOMMEDATIONS
selection of properly registered images and then to convert Further in future the same algorithm with few loops’
the MATLAB code into PYTHON was the challenge because execution can be implemented for the large number of
both languages have different syntax and have different kind sequential images even for the years of data.
of function for same task, for example in MATLAB reference Also, the automated application can be designed using this
matrix is read just through command while in PYTHON it has code as the base.
specific single function to perform all information extraction
of image at once. SUMMARY
This study has shown that the correlation between two
Also, in this study we observed that for automated code the sequential images can be utilized to yield the ocean currents.
PYTHON is more efficient and quicker as compared to Brightness temperature data provided by AVHRR for the
MATLAB as in MATLAB multiple functions are executed Arabian Sea is used in this study. A high level of correlation
for the image information extraction while in PYTHON of sequential images is demonstrated through MCC algorithm
single function performed that task. Also once designed the as shown in the results. This technique is sensitive to some of
code in PYTHON is completely automated only manual input the parameters which effect the results of this technique to a
needed is the pressing of run command. large extent. Setting of these parameters is of prime
importance. These parameters are:
This algorithm consumed approx. 5 minutes to execute for the  Template window size (first image).
one pair image on I3 processor computer so it is understood  Max. velocity.
that more advanced the processor is more efficient and  Search window size (second image).
quicker will algorithm will work.  Correlation cutoff.
Retrieval of surface currents in almost real-time manner is
Hence, this study conclude that PYTHON is more efficient possible by the AVHRR provided data with the help of MCC
for execution of MCC Method as compared to MATLAB. technique. It is important to mention here that the cloud free
nature of the images are very important for highest level of
correlation. The average maximum correlation observed for
the pair is 0.33ms-1.

REFERENCES
[3] W. J. Emery, A. C. Thomas, M. J. Collins, W. R.
Crawford, and D. L. Mackas, “An objective method for
computing advective surface velocities from sequential
infrared satellite images,” Journal of Geophysical Research:
Oceanography, vol. 91, no. C11, pp. 12865–12878, Nov.
1986.
[4] Tokmakian, R.T., P.T. Strub, and J. McClean-Padman,
Evaluation of the maximum cross-correlation method of
estimating sea surface velocities from sequential satellite
images, Journal of Atmospheric Oceanic Technology, 7, 852-
865, 1990.
[5] K. A. Kelly and P. T. Strub, “Comparison of velocity
estimates from advanced very high resolution radiometer in
the Coastal Transition Zone,” Journal of Geophysical
Research: Oceanography, vol. 97, no. C6, pp. 9653–9668,
Jun. 1992.

Seventh International Conference on Aerospace Science & Engineering 179

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

[6] M. Bowen, W. Emery, J. Wilkin, P. Tildesley, I. Barton,

and R. Knewtson, “Extracting Multiyear Surface Currents
from Sequential Thermal Imagery Using the Maximum
Cross-Correlation Technique,” Journal of Atmospheric
Oceanic Technology, vol. 19, pp. 1665–1676, Oct. 2002.
[7] R. Crocker, D. Matthews, W. Emery, and D. Baldwin,
“Computing Coastal Ocean Surface Currents From Infrared
and Ocean Color Satellite Imagery,” IEEE Transactions on
Geosciences and Remote sensing vol. 45, pp. 435–447, Mar.
2007.
[8] M. A. Warren, G. D. Quartly, J. D. Shutler, P. I. Miller,
and Y. Yoshikawa, “Estimation of ocean surface currents
from maximum cross correlation applied to GOCI
geostationary satellite remote sensing data over the Tsushima
(Korea) Straits,” Journal of Geophysical Research:
Oceanography, vol. 121, no. 9, pp. 6993–7009, Sep. 2016.
[9]C. Heuzé, G. Carvajal, L. Eriksson, and M. Soja-Wozniak,
“Sea Surface Currents Estimated from Spaceborne Infrared
Images Validated against Reanalysis Data and Drifters in the
Mediterranean Sea,” Remote Sensing, vol. 9, p. 422, May
2017.
[10] C. Heuzé, G. K. Carvajal, and L. E. B. Eriksson,
“Optimization of Sea Surface Current Retrieval Using a
Maximum Cross-Correlation Technique on Modeled Sea
Surface Temperature,” Journal of Atmospheric Oceanic
Technology, vol. 34, no. 10, pp. 2245–2255, Aug. 2017.
[11] J. Liu, W. Emery, X. Wu, M. Li, C. Li, and L. Zhang,
“Computing Coastal Ocean Surface Currents from MODIS
and VIIRS Satellite Imagery,” Remote Sensing, vol. 9, p.
1083, Oct. 2017.
[12] C. Kalyani Devasena, S. P. Subramaninan, and M. M.
Ali, “Study of seasonal current variability in the Arabian Sea
using Geosat altimeter data,” International Journal of Remote
Sensing, vol. 16, no. 14, pp. 2691–2701, Sep. 1995.
[13] V. N. Eremeev, A. N. Zhukov, A. S. Zapevalov, N. E.
Lebedev, and A. A. Sizov, “Seasonal Variations of the
Arabian Sea Level Anomaly Fields,” Oceanology, vol. 54,
no. 3, pp. 289–297, 2014.
[14] V. F. Banzon, R. E. Evans, H. R. Gordon, and R. M.
Chomko, “SeaWiFS observations of the Arabian Sea
southwest monsoon bloom for the year 2000,” Deep.-Sea
Research Part 2.

Seventh International Conference on Aerospace Science & Engineering 180

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Landslide Hazard Assessment And Monitoring In

Azad Kashmir Using Sentinel-1 InSAR Data
Labeeqa Zaheer Waqas A.Qazi
Geospatial Research & Education Lab Geospatial Research & Education Lab
Institute of Space Technology Institute of Space Technology
Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected]

and rivers, resettlement, and also damage natural resources.

They are primarily triggered by earthquakes, intense rainfall,
Abstract— Landslides are frequent pervasive geological
changes in groundwater, a disturbance or change of slope and
hazards that pose a serious threat to people, infrastructure, and volcanic activity. Satellite remote sensing techniques offer
the natural environment in mountainous regions with steep the capability for mapping and characterization of landslides
and assist in the development of guidelines for sustainable
slopes. Azad Kashmir lies in the Western Himalayan range and land-use planning.
is characterized as a subtropical highland climatic zone.
Varying mountainous terrain slopes, relief, and heavy rainfall Spaceborne Synthetic Aperture Radar (SAR) remote sensing
has advantage over optical remote sensing due to availability
during the monsoon season cause frequent landslides and of both day and night time images regardless of cloud
interrupt human activities, damage roads, and cause displaced coverage. Moreover, the coherence of microwave signals
buildings, and houses. In such steep topographic areas with
helps to detect topographical and deformation changes by
processing the phase signals. This forms the basis for the SAR
vulnerable conditions, information about slope stability is an Interferometry (InSAR) and Differential Interferometric
important task for landslide risk assessment. Differential Synthetic Aperture Radar (DInSAR) techniques, which have
now become been widely used in the field of earth and
Synthetic Aperture Radar Interferometry (D-InSAR) is a
environmental sciences for detecting and measuring surface
useful InSAR technique for detection, analysis, and monitoring displacements, land topography, land changes, land
of slow-moving landslides. Differential InSAR measures line of subsidence, ground water levels, soil moisture, snow
accumulation, etc. up to millimeter precision. Differential
sight (LOS) displacement by exploiting information relating to
InSAR method measures displacement in the line of sight
phase difference. In this study, single look complex (SLC) SAR (LOS) direction by exploiting phase information and compute
imagery from Sentinel-1 C-band sensor product has been used a phase difference image called interferogram[1]. External
DEM is used to remove or minimize phase contribution
to evaluate landslides over Azad Kashmir region. Both
introduced by surface topography.
ascending and descending pass image pairs were used to
generate interferograms for the period of October 2014 to Azad Kashmir region is highly susceptible to landslide events
due to its complex geological setting combined with active
November 2015. Ascending and descending pairs were
seismicity, easily erodible lithology, coarse terrain,
processed separately in European Space Agency Sentinel monsoonal climate, deforestation, and accelerated erosion
Application Platform (ESA SNAP). Displacement values of due to construction on steep slopes[2]. The region is deeply
dissected by small and large river valleys and people live on
highly unstable areas that keep high coherence over a period the fragile mountain slopes. In such environments,
are calculated. Low-coherence areas and decorrelating land- monitoring and analysis of landscape is essential to reduce
cover types such as agriculture and dense forest are removed any preventable loss of life and damage of property. In the
past, a number of studies using optical remote sensing data
from the analysis. Results are then displayed in the form of have been carried out to measure land instability in the study
displacement maps showing unstable areas that may be at risk area. Optical remote sensing shows greater dependency on
for further landslides.
atmospheric conditions. Shehzad [3] implemented repeat pass
Differential InSAR (DInSAR) in the study area by using
Envisat ASAR imagery, and analyzed the long-term impact
Keywords— Interferometric Synthetic Aperture Radar of the Kashmir 2005 earthquake on major landslide areas by
(InSAR), Landslide generating displacement maps for the period of 2007-2010.
With the launch of Sentinel-1 in 2014, systematic and regular
I. INTRODUCTION acquisitions of SAR data is made possible. Sentinel-1 has a
Landslides are amongst the most common natural disasters shorter temporal baseline characterized by high spatial
worldwide, resulting in loss of human life, blockage of roads

Seventh International Conference on Aerospace Science & Engineering 181

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

coverage and a greater ground resolution with respect to II. METHODOLOGY

Envisat ASAR and other previous C-band radar systems[4]. Sentinel-1 operates in C-band (5.405 GHz central frequency
f, and 5.547 cm wavelength λ) onboard the two satellites
The current study illustrates the potential of Sentinel-1 data, Sentinel-1A and Sentinel-1B. Single look complex (SLC)
with the aim of detecting regions of relatively high products in Interferometric Wide (IW) swath mode were
displacement rates in the study area. In future, these unstable downloaded from the European Space Agency (ESA)
areas will be prone to landslides and associated destruction. Copernicus Open Access Hub[7]. IW mode captures three
sub-swaths i.e. IW1, IW2, IW3. Each sub-swath contains one
A. STUDY AREA image and one polarization channel.
The footprint of both ascending and descending pass SLC
Azad Kashmir is located in northwestern part of the images over the study region is shown in Fig.2 below
Himalayas range, Pakistan. Geologically, the Himalayan
mountain range is the youngest and most dynamic system in
the world and is specifically famous for the recurrent seismic
activities, varying slopes, land sliding, earthquakes and flash
floods[5]. The study area terrain is mountainous and has a
subtropical highland climate.

The dominating land cover type is agriculture and dense

forest[6]. Muzaffarabad is the capital city of Azad Kashmir
and is situated at the confluence of the Neelum and Jhelum
rivers. Khagan, Neelum and Jhelum are three principle
valleys in study area. In winter season, snowfalls occurs
above 1500m elevation. Snow accumulates in steep
mountainous slopes and then shed in late spring. High
temperature occurs in the summer months. Monsoon season
Fig. 2. Spatial foot prints of the ascending and descending passes over
starts in late June and ends in August. Heavy precipitation the study region
occurs during monsoon season, results in landslides, and
flooding. The study area map with some of the active The acquired data is in right-looking direction and has
landslide areas are shown in Fig.1 below singlePolarization (VV). Relative orbit of ascending pass
images is 100 while for descending pass images relative orbit
is 107. The data was downloaded for the period of October
Fig. 1. Study area map of Azad Kashmir
2014 to November 2015. After data acquisition, InSAR image
pair selection has been done. Interferometric pairs must be of
the same path number and must cover the same area. Images
with different flight directions (ascending vs descending)
cannot serve as interferometric pairs. Moreover, image pair
selection should consider seasonal variations, atmospheric
conditions, and perpendicular and temporal baseline values.
It is better to have an image pair of same season, low
perpendicular baseline and temporal difference to avoid
decorrelation and coherence loss in the area of interest (AOI).
Complete list of descending and ascending pairs used in this
study is shown in the table I and II below

TABLE I. SENTINEL-1 DESCENDING INTERFEROMETRIC PAIRS USED

IN THE STUDY
Descending Pairs
Sr Temporal Perpendicular
No Reference Secondary
Baseline Baseline
Image Image
(Days) (Meters)
1 13-Oct-14 06-Nov-14 24 52

2 06-Nov-14 30-Nov-14 24 8

2 30-Nov-14 17-Jan-15 47 -67

4 17-Jan-15 10-Feb-15 23 -59

Landslide areas and locations from various sources like State
of Disaster Management Authority (SDMA), Planning and 5 17-Jan-15 30-Mar-15 72 20
Development Department (PND), Global Landslide 6 17-May-15 10-Jun-15 24 -13
Catalogue (GLC) by NASA are displayed in the study area
7 17-May-15 28-July-15 72 4
map. These datasets were taken from [3].
8 28-July-15 21-Aug-15 24 -4

9 21- Aug-15 25-Nov-15 96 20

Seventh International Conference on Aerospace Science & Engineering 182

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
TABLE II. SENTINEL-1 ASCENDING INTERFEROMETRIC PAIRS USED
IN THE STUDY
Since the study area lies in IW2 sub-swath of SLC image, so
Ascending Pairs processing has been only applied on IW2 sub-swath for
Sr Temporal Perpendicular
No Reference Secondary
Baseline Baseline
analysis.
Image Image
(Days) (Meters)
1 24-Oct-14 17-Nov-14 24 -55 Phase unwrapping has been done using SNAPHU algorithm.
Low coherence areas need more computational time in phase
2 24-Oct-14 11-Des-14 47 49
unwrapping and it is quite useless to unwrap them. So, only
2 11-Des-14 04-Jan-15 24 -82 those areas were unwrapped which have coherence > 0.4.
28 Dense forest and agriculture areas were masked out for final
4 04-Jan-15 28-Jan-15 23
displacement map because dense vegetation leads to
5 04-Jan-15 21-Feb-15 47 -1
coherence loss, and agriculture areas are mostly on plain
6 04-Jan-15 17-Mar-15 71 2 slopes and soil changes in agriculture areas can also lead to
96
coherence loss.
7 28-Jan-15 10-Apr-15 72
8 04-May -15 21-Jun-15 48 -8 Absolute phase calibration was performed on the
9 21- Jun-15 08-Aug-15 48 -36 displacement maps. Muzaffarbad city urban areas were
considered as stable points for this purpose and used as zero
10 08-Aug-15 01-Sep-15 24 92
displacement reference point for InSAR displacement
measurements. Landslide inventory data taken from[9] and
A total of 19 interferometric pairs were formed for the data collected from SDMA, PND and GLC (shown in Fig.1)
landslide analysis using Alaska satellite facility (ASF) were overlaid on final displacement maps and further analysis
baseline tool[8]. Perpendicular and temporal baseline was performed on these selected areas.
distribution of all granules within a given InSAR stack is
visualized through ASF baseline tool. Multiple reference
images were used in order to decrease the chance of
coherence loss. InSAR pairs with perpendicular base line less
than 100m and temporal baseline less than 100 days were
shortlisted for the analysis.

Both ascending and descending interferometric pairs were

processed separately in ESA’s Sentinel Application Platform
(SNAP). The processing steps are shown in the methodology
flowchart in Fig.2 below.

Fig. 3. Methodology flow chart

Seventh International Conference on Aerospace Science & Engineering 183

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

III. RESULTS AND ANALYSIS In ascending pairs, it has been found that the interferogram of
pair no 9 (21-Jun-15 - 08-Aug-15) and pair no 10 (08Aug-15
Interferograms along with coherence were generated for each -01-Sep-15) shows highest displacement values as compared
interferometric pair for both ascending and descending pairs. to other ascending pairs. Displacement values on the selected
Only landslide areas which were highlighted by the sources landslide areas ranges between -3.0 to 8.0 cm in pair no 9 and
of SDMA, PND, GLC and landslide inventory data in the ranges between -24.0 to 32.0 cm in pair no 10 and coherence
study area were considered for analysis. is also relatively high in both the pairs. Pair 9 and pair 10
displacement and coherence are shown in fig.3 and fig.4
Also it was very challenging to consider small landslide areas below respectively. In both fig. left panel (a) shows
for analysis because of the non-availability of confirmed displacement while right panel (b) shows coherence.
ground truth data for other landslide areas. The areas with low
coherence show noise and one cannot be sure about the
displacement values on that areas. Overall, coherence for
areas under investigation for landslide analysis did not show
very high values especially in ascending pairs.

(b)

(a) (b)

Fig. 4. Ascending pair no 9 (21-Jun-15 - 08-Aug-15)

(a)

In descending pairs, it has been found that the

Fig. 5. Ascending pair no 10 (08Aug-15 -01-Sep-15) interferogram of pair no 2 (06-Nov-14 - 30-Nov-14) and pair
no 8 (28July-15 - 21-Aug-15) shows highest displacement
values as compared to other descending pairs. Displacement
values on the selected landslide areas ranges between -5.0 to

Seventh International Conference on Aerospace Science & Engineering 184

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

23.0 cm in pair no 2 and ranges between -2.0 to 6.0 cm in pair

no 8 and coherence is also relatively high in both the pairs.
Pair 2 and pair 8 displacement and coherence are shown in
fig.5 and fig.6 below respectively. In both fig. left panel (a)
shows displacement while right panel (b) shows coherence.

Fig. 6. Descending pair no 2 (06-Nov-14 - 30-Nov-14)

Fig. 7. Descending pair no 8 (28July-15 – 21-Aug-15)

(a) (b)
(b)

(a)

IV. CONCLUSION

Both ascending and descending pairs of Sentinel -1 SLC

images were utilized to monitor slope instability in selected
areas of Azad Kashmir by applying DInSAR technique. The
ascending and descending pairs were treated separately and
only those areas for landslides were selected that were
highlighted by the SDMA, PND, and GLC landslide
inventory data. Displacement maps were generated for the
period of October 2014 to November 2015 and areas were

Seventh International Conference on Aerospace Science & Engineering 185

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

highlighted in displacement maps that showed significant vol. 118, pp. 68–80, 2016, doi:
displacement high values (areas of instability). 10.1016/j.jseaes.2016.01.002.
[3] K. Shehzad, “Landslide Identification , Assessment and
Areas selected for landslide analysis mainly include Quantification in Azad Kashmir Using Repeat-Pass
Muzaffarabd, Chatter, Lohar Gali, Garhi Dupatta and Hattian Synthetic Aperture Radar ( SAR ) Interferometry,” M.S.
Bala. Three landslide locations are found significant in a Thesis, Institute of Space Technology, 2017.
sense of destabilization with respect to other locations. These [4] M. Del Soldato et al., “Monitoring ground instabilities
locations were selected on the basis of displacement and using SAR satellite data: A practical approach,” ISPRS Int.
J. Geo-Information, vol. 8, no. 7, 2019, doi:
coherence values in both ascending and descending pairs. The 10.3390/ijgi8070307.
results shows highest destabilization in Hattian Bala and
[5] A. ur Rahman, A. N. Khan, and A. E. Collins, “Analysis of
Lohar Gali as compared to any other landslide locations in the landslide causes and associated damages in the Kashmir
study area. Garhi Dhupatta also shows instability in some of Himalayas of Pakistan,” Nat. Hazards, vol. 71, no. 1, pp.
the pairs but relatively less in comparison with Hattian Bala 803–821, 2014, doi: 10.1007/s11069-013-0918-1.
and Lohar Gali. June- September period are found to be more [6] F. M. Qamer et al., “Mapping deforestation and forest
favorable for instability. Heavy precipitation and the degradation patterns in Western Himalaya, Pakistan,”
monsoon period is the main reason for landslide instability in Remote Sens., vol. 8, no. 5, pp. 1–17, 2016, doi:
the study area. 10.3390/rs8050385.
[7] Scihub.copernicus.eu. 2021. [Online] Available at:
<https://scihub.copernicus.eu/dhus/#/home>
V. REFERENCES [8] Search.asf.alaska.edu. 2021. ASF Data Search. [Online]
Available at: <https://search.a (ASF Data Search,
[1] X. Zhou, N. Chang, and S. Li, “Applications of SAR 2021)sf.alaska.edu/#/?searchType=Baseline%20Search>
Interferometry in Earth and Environmental Science [9] S.U. Shah, “Spatio-Temporal Change and Hazard
Research,” Sensors, vol. 9, pp. 1876–1912, 2009, doi: Detection of the 2005 Kashmir Earthquake-Induced
10.3390/s90301876. Landslides,” M.S. Thesis, National Centre of Excellence in
[2] M. Shafique, M. van der Meijde, and M. A. Khan, “A Geology , University of Peshawar, 2015.
review of the 2005 Kashmir earthquake-induced landslides;
from a remote sensing prospective,” J. Asian Earth Sci.,

Seventh International Conference on Aerospace Science & Engineering 186

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Modeling & Simulation of BSAR (Bi Static

Synthetic Aperture Radar) system using Reflected
BeiDou B1I Signals
Hassan Sadiq Muhammad Shoaib ul Hassan Mohammad Usman
Department of Space Science, IST Department of Space Science, IST, ABM SATUMA (PVT) LTD Islamabad
Islamabad, Pakistan Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected] [email protected]

Abstract—The BeiDou Navigation Satellite System is the

recent addition in the GNSS (Global Navigation Satellite
Signals) community. Direct BeiDou signals are used for
navigation and positioning purposes. This research endeavor
intends to utilize the reflected signals as they can provide an
opportunity for passive radar applications. A bi-static radar
system is formed by the BeiDou satellites, a modified BeiDou
receiver, and its signal processing components (antennas). The
Bi-static radar (BSAR) can be utilized for passive detection of
targets and /or any movement or changes. Firstly, power budget
analysis will be calculated for BeiDou B1I signals and compared
with GPS L1 signals. Later the whole Bi-static Radar scenario Fig. 1. The BSAR scenarion
will be simulated as closely and faithfully as possible and a
matched filter technique will used to improve the image
resolution. The simulation will be carried out under different As BeiDou has become fully operational and thus the
conditions and in a detailed manner, and efforts will be made to availability of BeiDou signals it is a great opportunity for
validate that BeiDou signals can be used as an illuminator or researchers. BeiDou system is a transmitter of opportunity
“transmitter of opportunity” for passive BSAR applications and can be used in this research to illuminate the target in an
with acceptable image resolution. area of interest. It eliminates the requirement of dedicated
transmitter for this purpose[9]. With compared to hefty
Keywords— Analysis of Power Budget, BSAR, BeiDou B1I
signal, specialized transmitters, the power consumption of BeiDou-
based bi-static radar is quite low. As the system cross section
I. INTRODUCTION is differ from the mono static radar for this system too, has
The BeiDou is a Chinese Global navigation satellite system the ability of detection of stealthy targets.
and recently it became fully operational [1]. Primarily, the However, the power of BeiDou direct signal is low and after
system provides Global, highly precise, reliable and stable the reflection their power is further reduced by the reflecting
position and velocity information to the user. Secondarily, the surface and difficult to detect. These low power signals can
system signals after reflection can be an opportunity and be used in many applications of remote sensing after applying
utilized for passive remote sensing applications[2]. In some signal processing techniques [10]. During this research
numerous application of remote sensing BeiDou reflected endeavor, the image of target will be generated in the area of
signals can be used like measurement of soil moisture content interest by using reflected signals as the reflected signals has
over the land, bio mass, and bi-static imaging [3]–[6]. valued information about the reflecting surface. Thus, the
BEIDOU-3 design comprises of three GEO, three IGSO, and BeiDou reflected signals have great potential and can be used
24 MEO satellites, with backup satellites in orbit. [6], [7]. in various remote sensing applications like measuring soil
A BeiDou modified receiver with its components of signal moisture content, ice age/thickness, wind speed, passive
detection(antennas) and BeiDou satellites, formed a BeiDou remote sensing, and ocean remote sensing [3][11][12][8].
based bi-static radar system. In applications of remote sensing The use of BeiDou as an opportunity transmitter eliminates
BSAR (bi-static synthetic aperture radar) systems can be the requirement for a specialized transmitter. Thus, the
used[8]. In BSAR scenario both receiver and transmitter are BeiDou system gives an excellent chance for this type of
separated and the distance between them is called based line study; fortunately, such an expensive system has already been
and the angle is called bi-static angle. The complete scenario constructed and maintained, and it covers a large region and
is depicted in Fig. 1. is available 24 hours a day. By selecting appropriate BeiDou
satellites and optimizing for geometry in terms of signal
power and visibility, direct and reflected signals may be
received.
BeiDou radio navigation signals are right-hand circularly
polarized (RHCP), and their polarization changes when they

Seventh International Conference on Aerospace Science & Engineering 187

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

reach the Earth's surface or any other object after reflection. 1561.098MHz, 2.046MHz is the chip-rate of the signal and
These signals may become circularly polarized on left-hand duration is 1milli second has bandwidth of 4.092MHz. The
(LHCP). BeiDou B1I signal Ranging length is 2-times longer than the
Received and locked direct signals from a certain satellite are GPS L1 signal, due to higher bandwidth BeiDou B1I signal
used to acquire and identify reflected signals. The has better signal to noise ratio and spatial resolution then GPS
commercial receivers available in the market have an RHCP L1 frequency signal.
antenna that can only receive direct signals and not able to These signals, in practice, travel a greater distance, but have
receive reflected signals. The BeiDou signals must be a lower power spectral density than noise. A code sequence
captured and received in digital format, and the signals must generated locally that is longer than conventional receivers
be stored in order to do post-remote sensing and imaging of and correlated with the received signals from this efficient
the selected area. To receive the BeiDou signal after and better processing gain can be achieved. Correlation may
reflection, a high gain antenna with circularly polarized on be done in two ways: coherent and incoherent.
the left hand is used. This information is obtained by the use The reflected signal is weaker than the direct signal as a result
of a circuit with the proper reference signal and sampling of the scattering process and the extra route travelled. The
frequency. Receiving a weak signal with a poor SNR at a output may be lower than the noise level even after coherent
continuous data rate (minimum of 20 M samples/s) using correlation. A sequence of waveforms can be gathered and
suitable techniques is a significant problem. This received non-coherent summation can be performed to speed up the
data is then sent to the processing system. processing even further. However, some doppler or doppler
To create image of a certain selected area BeiDou direct and ambiguity removal will be required in this instance, or else
reflected signals will be manipulated using technique of the correlation peak would grow and become comparatively
digital signal processing. The SAR methodology is employed flat, most likely due to drift in-phase and doppler related to
in which it is required to maintain the movement of receiver temporal and spectral variation.
or Bi-static radar in spatial sphere of influence. As mentioned Power budget analysis has been performed in this paper for
above, with the use of suitable hardware the BeiDou reflected scenario depicted in Fig. 1. The simulation results for GPS L1
signal from the object will be acquired, on a mobile object and BDS B1I signals are analyzed.
(UAV or aerial platform) the hardware will be arranged for
the synthetic aperture effect as demonstrated in Figure 1 to II. ANALYSIS OF THE GPS L1 SIGNAL POWER
achieve the needed change in geometry. Range Resolution is BUDGET
a very important parameter for the image reconstruction from The GPS L1 signal is analyzed first for the bi-static radar
the reflected BeiDou signals based on bi-static radar power budget, followed by the BeiDou B1I signal. The signal
architecture. On the same orientation, the Radar system may transmitted power is represented with P t and the gain of
identify two or more objects at different ranges. transmitter is represent with Gt, R1 represent the range from
The range resolution is determined as. transmitter to target and it is around 22,000 Km. From
𝐶 receiver to target the range is represented as R2 and it is the
𝛿𝑟 = (1)
𝛽 slant range [14] and is given by.
2𝐵𝑐𝑜𝑠 ( )
2
𝑃𝑡 𝐺𝑡𝐺𝑟⅄2 𝜎
Where B denotes the transmitted signal's bandwidth and β is 𝑅2 = √ 𝑘𝑚 (2)
the bistatic angle. As a result, range resolution is affected by (4𝜋)3 𝑅12 𝑃𝑟
signal bandwidth and bi-static angle. For better resolution we The power received at the receiver's antenna is represented by
improve the bandwidth of the signal. BeiDou signal is 𝑃𝑡 𝐺𝑡
transmitted at higher bandwidth, as compared to GPS signals, 𝑃𝑟 = ∗ 𝜎𝐴𝑒𝑓 (3)
4𝜋𝑅12 4𝜋𝑅22
therefore range resolution is improved. Where Ae.f is receiver antenna effective area
BeiDou B1I signal is transmitted at frequency of ⅄2 ∗ 𝐺𝑟
1561.098MHz with a ranging code. The ranging code is 𝐴𝑒𝑓 = (4)
4ᴫ
unique for each satellite and freely available for all civilian The strength of the reflected signal is determined by the
users. When compared to GPS 'L1' signals, these new signals receiver's antenna gain and the radar's cross-section. A target
will be more accurate and compatible with remote sensing. can be easily detected with a greater RCS, and other factors
This system has become fully operational[1]. on which detection depends is the size of target, material, and
For verification of their suitability for remote sensing shape.
application, accuracy and compatibility, the power budget For the receiver and Bi-static radar the signal to noise ratio is
analysis is necessary. The analysis for power budget and the given by [14].
influence of clutter on passive target detection and imaging 𝑃𝑡 𝐺𝑡𝐺𝑟⅄2 𝜎
has been interpreted[9] [13] 𝑆𝑁𝑅 = (5)
(4𝜋)3 𝑅12 𝑅22 𝐾𝑇𝐵𝑛
At 1572.42 MHz, the GPS transmits the L1 civilian signal Where receiver antenna gain is Gr, power received at receiver
together with the coarse acquisition (C/A) code. The chip rate
of the BPSK modulation of C/A code is 1.023 and the 𝜆
is Pr, the noise of receiver is Nr =kTBn and is GPS L1 signal
duration is one millisecond. Because of the narrow bandwidth wavelength. When R1=R2 the SNR is maximum however
and low L1 signal strength, the GPS L1 signal has a GPS satellite is transmitting signal at height of 22,000 Km.
bandwidth of 2.046 MHz, made it difficult to detect the The target to receiver range is limited to 1000 m. The graph
reflected signal using conventional signal acquisition in figure (2) shows GPS L1 frequency SNR verses Range plot
techniques. The Chinese navigation system BeiDou B1I with 10 m2 cross-section target. As expected, Radar of passive
signal is design for civilian purpose and freely available microwave imaging of this type has the basic limitation, poor
publicly. BeiDou send B1I Signal at frequency of

Seventh International Conference on Aerospace Science & Engineering 188

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

SNR even for short-range. However, the SNR can be TABLE I GPS L1 Variables values for analysis
increased significantly by correlating the signal for longer
period resulting in more processing gain and improvement in No. Variables Value Unit
the SNR. With the coherent and non-coherent summation, the (Pt) is power of
correlation value can be increased, as noise is uncorrelated 1 transmitted signal
40 Watt
therefore SNR will be enhanced. (Gt) is antenna gain
SNR Vs Range for GPS L1 signal without processing gain 2 of transmitter
13 dB
-30
(Gr) is antenna gain
-40 3 of receiver
30 dB
-50
4 Bandwidth 2.046 MHz
-60 (Nr) is level of
5 -131 dB
-70 noise of receiver
SNR (dB)

-80 (TQ) is interval of

6 0.1 Second
-90
processing
-100
III. BEIDOU B1I SIGNAL POWER BUDGET ANALYSIS
-110

-120
The results of using the BeiDou B1I signal for remote sensing
have been compared to those of using the GPS L1 signal.
-130
0 100 200 300 400 500 600 700 800 900 1000 BeiDou B1I power budget study revealed signal, in same
Target to receiver Range R2(m) scenario as shown in section II, the parameters employed for
Fig. 2. Without processing gain SNR Vs Range for GPS L1 signal BeiDou B1I signal strength budgeting are listed in Table II.

Fig. 2. shows GPS L1 signal simulation result for SNR verses TABLE II BeiDou B1i Variable values for Analysis
range. By simulation, the SNR will be calculated at different
ranges from 0-1000m. The SNR is -70dB for the range of No. Variable Value Unit
100m for GPS signal. With such a low SNR, recovery of
1 (Pt) is power of transmitted 43 Watt
signal is extremely difficult. The SNR attained by comparing
signal
the signal over a prolonged period is given by [15].
2 (Gt) is antenna gain of 15 dB
𝑃𝑡 𝐺𝑡𝐺𝑟⅄2 𝜎
𝑆𝑁𝑅 = 𝐺 𝑁 0.8 (6) transmitter
(4𝜋)3 𝑅12 𝑅22 𝐾𝑇𝐵𝑛 𝑠𝑝 (Gr) is antenna gain of
3 30 dB
Where N represents the maximum non-coherent number of receiver
samples and Gsp represents the processing gain. The non-
4 Bandwidth 4.092 MHz
coherent summing is useful for obtaining the desired gain
since it requires a larger number of samples (N) than the 5 (Nr) is level of noise of -131 dB
coherent summation. receiver
6 (TQ) is interval of 0.1 Second
45
SNR Vs Range for GPS L1 signal with processing gain processing
SNR Vs Range for BeiDou B1i signal without processing gain
40 -20

35 -30

-40
30

-50
SNR (dB)

25
-60
SNR (dB)

20
-70
15
-80
10
-90
5
-100
0
0 100 200 300 400 500 600 700 800 900 1000 -110
Target to receiver Range R2(m)
-120
Fig. 3. with processing gain, SNR Vs Range for GPS L1 Signal 0 100 200 300 400 500 600 700 800 900 1000
Distance (m)

Fig. 4. Without processing gain SNR Vs Range for BeiDou B1I Signal
The SNR for the GPS L1 signal is 43dB, as shown in Figure
3. SNR will be improved with non-coherent summation while
correlating the signal for a longer period, the SNR is 22dB at
the range of 100m and therefore detection is possible.
Therefore, the reflected signal can be utilized for the
application of remote sensing. The parameters and their
values are summarized in Table 1 that will be used for “the
GPS L1 signal's power budget analysis”.

Seventh International Conference on Aerospace Science & Engineering 189

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
TABLE III Comparison of Range and SNR L1 signal. A comparison of Target to receiver ranges and
SNR has been summarized in Table 3 for reflected GPS L1
Without With signal and B1I signal of BeiDou.
Target
processing processing
to REFFERENCE
gain gain
S/N receiver [1] M. Li and Y. Yuan, "Estimation and Analysis of BDS2 and BDS3
L1 B1I L1 B1I
range Differential Code Biases and Global Ionospheric Maps Using BDS
SNR SNR SNR SNR Observations." Remote Sensing 13.3 (2021): 370.
R2 (m)
(dB) (dB) (dB) (dB) [2] M. Martin-Neira, "A passive reflectometry and interferometry
system (PARIS): Application to ocean altimetry." ESA
1 10 -31 -27 41 45 journal 17.4 (1993): 331-355.
2 50 -63 -59 27 31 [3] S. Yan, H. Zhou and J. Gong, "GNSS imaging: A case study of tree
detection based on BeiDou GEO satellites." Progress In
3 100 -77 -73 21 25 Electromagnetics Research C 84 (2018): 227-240.
4 200 -91 -87 15 15 [4] T. Zeng, T. Zhang, W. Tian, C. Hu, and X. Yang, “Bistatic SAR
5 500 -109 -105 7 11 Imaging Processing and Experiment results Using BeiDou-2 /
Compass-2 as Illuminator of Opportunity and a Fixed Receiver,”
6 800 -118 -114 3 7 2015 IEEE 5th Asia-Pacific Conf. Synth. Aperture Radar(APSAR),
7 1000 -123 -119 1 5 pp. 302–305, 2015.
[5] S. H. Yan, F. Zhao, N. C. Chen and J. Y. Gong, "Soil moisture
estimation based on BeiDou B1 interference signal
Fig. 4. shows BeiDou B1I signal simulation results for SNR analysis." Science China Earth Sciences 59.12 (2016): 2427-2440.
vs Range plot. Through simulation, the SNR will be [6] C. Huang, Z. Li, J. Wu, Y. Huang, H. Yang and J.
calculated at different ranges from 0-1000m. The SNR is Yang,"Multistatic Beidou-Based Passive Radar for Maritime
Moving Target Detection and Localization." IGARSS 2019-2019
-25dB for the range of 100m for BeiDou B1I signal which is IEEE International Geoscience and Remote Sensing Symposium.
better than the GPS L1 signal. IEEE, 2019.
By comparing the signal over a longer period of time, the [7] “BeiDou Navigation Satellite System Signal In Space Interface
SNR can be improved and the results are shown in Fig. 5. As Control Document,” no. February, 2019.
[8] F. Xuezhen, "Passive SAR with GNSS transmitters: Latest results
a result, by correlating for a longer time, the SNR may be and research progress." 2017 IEEE International Geoscience and
enhanced with processing gain. As Gsp=BTQ, the processing Remote Sensing Symposium (IGARSS). IEEE, 2017.
gain is calculated by multiplying B by TQ, where B is the [9] B. Mojarrabi, J. Homer, K. Kubik and I. D. Longstaff, "Power
bandwidth and TQ is the processing interval. budget study for passive target detection and imaging using
secondary applications of GPS signals in bistatic radar
SNR Vs Range for BeiDou B1i signal with processing gain
50 systems." IEEE International Geoscience and Remote Sensing
Symposium. Vol. 1. IEEE, 2002.
45 [10] M. Usman, and D. W. Armitage. "Acquisition of reflected gps
40
signals for remote sensing applications." 2008 2nd International
Conference on Advances in Space Technologies. IEEE, 2008.
35 [11] T. Yang, W. Wan, X. Chen, T. Chu, and Y. Hong, “Using BDS
SNR observations to measure near-surface soil moisture
SNR (dB)

30
fluctuations: Results from low vegetated surface,” IEEE Geosci.
25
Remote Sens. Letters 14.8(2017):1308-1312.
[12] T. Zeng, T. Zhang, W. Tian, C. Hu, and X. Yang, “Bistatic SAR
20 imaging processing and experiment results using BeiDou-
2/Compass-2 as illuminator of opportunity and a fixed receiver,”
15
2015 IEEE 5th Asia-Pacific Conf. Synth. Aperture Radar, APSAR
10 2015.
[13] V. Behar, H. Kabakchiev and C. Kabakchiev, Detectability of Air
5
0 100 200 300 400 500 600 700 800 900 1000
Targets using Bistatic Radar Based on GPS L5 signals.” 2011 12th
Distance (m) International Radar Symposium (IRS). IEEE, 2011.
Figure 5 with processing gain SNR Vs Range for BeiDou B1I Signal [14] E. P. Glennon, A. G. Dempster, and C. Rizos, "Feasibility of air
target detection using GPS as a bistatic radar." Positioning 1.10
(2006).
[15] S. Erker, S. Thölert, J. Furthner, and M. Meurer "L5–The New GPS
The SNR for the BeiDou B1I signal is 25dB with processing Signal." Proceedings of IAIN (2009): 27-30.
gain at a target range of 100 m, which is better than the GPS

Seventh International Conference on Aerospace Science & Engineering 190

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Flood Mapping and Crop Damage Assessment

using Sentinel-1 SAR GRD and Optical Data
through Google Earth Engine

Tariq Abbas Waqas A. Qazi

Geospatial Research and Education Lab, Geospatial Research and Education Lab,
Institute of Space Technology, Islamabad, Institute of Space Technology, Islamabad,
Pakistan. Pakistan.
[email protected] [email protected]

I. ABSTRACT Optical remote sensing can be used for flood mapping on

basis of high reflectance in RGB (Red, Green, Blue) and low
This study deals with the flood mapping and crop damage reflectance in Near Infrared (NIR) bands [5,6,7]. Although
assessment for the District of Mirpur Khas, Sindh using optical remote sensing can also be used for flood mapping
microwave Synthetic Aperture Radar (SAR) and optical but only in good weather condition. Generally, in-practice,
satellite remote sensing datasets. Floods are one of the flood occurs in bad weather condition with rainfall and
most disastrous events across the world and their impact heavy clouds which make deficient the use of optical sensor
needs to be evaluated and analyzed timely to alleviate and at such situation when we need cloud-free imagery. In
manage the associated damages. The main objective of addition, optical sensors are not well-suited for detecting
this study is to delineate the flood extent and assessment inundated vegetated areas [8].
of associated damaged crop for the flood that occurred
SAR is widely used in monitoring flood events because SAR
due to monsoon season from August to September 2020 transmits microwave signals which remain unchanged by
using SAR-based imagery provided by ESA Sentinel-1 the weather conditions like cloud, fog, smog, etc and can
satellites and Moderate Resolution Imaging
image the area at any time of day. SAR is used for urban
Spectroradiometer (MODIS) global landcover data on a flooding because its longer wavelengths have high
cloud-based Remote Sensing and Digital Image
penetration tendency. In SAR imagery built-up area gives
processing platform i.e., Google Earth Engine (GEE). high backscatter and appears brighter due to corner
SAR imagery is a more appropriate choice than optical for reflection and water appears dark due to its smooth surface
flood mapping because it has low dependency on weather
backscattering mechanism [9,10]. Therefore, SAR has been
conditions and can provide day/night images in the used in such situation by many researchers. Although SAR
situation of emergency because it is independent of solar
data has great potential for flood mapping but it is
illumination. MODIS data can be used for crop damage challenging to assess the damages with SAR data due to
assessment because it gives the global landcover data with limitation of temporal resolution. In such situation,
a temporal resolution of 1 day. SAR data in VH-
Moderate Resolution Imaging Spectroradiometer (MODIS)
polarization and Otsu thresholding provide reliable results sensors can play a vital role for crop damage assessment due
for flood mapping. to its temporal resolution of producing imagery twice a day.
Keywords— Flood Mapping, Flood Extent, Damage Assessment, It gives data two times a day. Though its spatial resolution
SAR, GEE (Google Earth Engine) is coarser as compared to Sentinel-1 SAR data i.e., 250 m to
1000 m. MODIS uses Shortwave Infrared (SWIR) band
II. INTRODUCTION
which is sensitive to soil and vegetation [24].

F
loods are one of the heaviest natural disastrous events Generally, there are two types of techniques that can be used
and bring many casualties with them [1]. Floods are for flood mapping. The first technique uses machine
needed to be managed timely by the concerned learning [11,12], and the second uses a threshold-based
authorities to mitigate the hazardous impact. But for this mechanism [13,14]. The 1st technique is not used widely for
purpose, we require early information to estimate the spatial near real time flood mapping because it requires the field
extent of the flood to deal with the emergency response [2]. survey for training samples which is hard to collect at the
Since 1928, Pakistan has faced many heavy floods and most time of co-event (flood) [15,16]. The second technique
of the floods come in the monsoon time-period from July to (threshold-based) can be performed manually or automatic
September. Matee et al. [20] has mentioned some renowned and it seems more effective as compared to the prior
floods of 1928, 1929, 1955, 1973, 1976, 1980, 1988, 1992 mentioned technique. The manual threshold approach is a
and 2010 [18]. Since 1948 to 2008, floods in Pakistan have bit difficult because it needs knowledge of some number of
claimed 7,000 lives and caused many infrastructure and crop training sampling [14,17]. This may also cause inaccurate or
losses with 7.7 million acres inundated zones. Spaceborne less accurate flood mapping due to wrong classification of
remote sensing plays a key role in flood mapping [3,4]. water pixels and non-water pixels [14]. However, the

Seventh International Conference on Aerospace Science & Engineering 191

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

automated threshold technique does not use training data

and is used for automated separation of water and non-water
pixels [18,19].
There are certain automated threshold techniques and the
Otsu image threshold is one of them [21,22] and is widely
used for flood mapping [19]. Otsu method requires less
interaction and is less complex and computes the threshold
from a histogram of an image. Otsu method provides a more
precise flood map for large areas using Sentinel-1 GRD SAR
data [19].
Although, many researchers like [19,22] used Sentinel-1 and
manual threshold method and those studies showed
appropriate results but those analysis were limited due to
time demand, to download and process the datasets and their
execution on desktop-based tools. These issues limit the
early response for flood mapping to decision makers and
other concerned authorities. That is why sometimes these
types of tools do not suit for an early and timely response.
On the other hand, Google Earth Engine (GEE), a cloud
based Remote Sensing and Digital Image processing tool,
enables such type of analysis for an early response over large Figure 1: Study Area Map
areas. Google Earth Engine has ingested and makes
available a large variety of datasets of different satellites for Mirpur Khas is one of the largest cities of Sindh Province
user utilization [23]. and is located in the south-east of the province Sindh. Mirpur
The main objective of this study was to demarcate the Khas city comes at the 33rd number in the list of largest cities
flooded areas combining the automatic threshold method across the country and at 5th number in the list of largest
using Sentinel-1 SAR data and crop damage assessment cities of Sindh Province. There are 7 sub-division and 3
using MODIS optical data in the GEE cloud-based platform tehsils in Mirpur Khas division. It has 233,916 population
for an early and precise flood mapping for the area of Mirpur according to census 2017 [26].
Khas, Sindh, Pakistan for the flooding event of August to
September, 2020. IV. DATASETS
The following datasets were used to execute this research
III. STUDY AREA analysis.
A. Sentinel-1
The PMD (Pakistan Meteorological Department) had Sentinel-1 is one of 5 missions of Copernicus by the
reported in late August 2020 that monsoon’s heavy rains in European Space Agency (ESA), operating in all weather,
parts of the Sindh province have affected the districts of day, and night conditions and produces Synthetic Aperture
Mirpur Khas, Hyderabad, Sukkur, and Larkana. The Radar (SAR) imagery up to 10 m in resolution in the
districts of Karachi, Hyderabad, and Mirpur Khas were Interferometric Wide Swath (IW) mode, in C band. The
declared “calamity-affected areas” [25]. Keeping the above- Sentinel-1 constellation consists of 2 sensors (“A” & “B”)
mentioned scenario in mind it was decided to choose Mirpur and its revisit period or temporal resolution is 12 days for
Khas, Sindh, Pakistan as study area. one satellite and 6 days for both combined. Sentinel-1 data
is delivered by ESA in two formats: Single Look Complex
(SLC) & Ground Range Detected (GRD). SLC data is used
mostly for interferometric SAR applications. Also, to be
noted that GEE has ingested only GRD SAR data [27].
Sentinel-1 SAR GRD data was used to highlight change
detection and to delineate the flood map.

B. JRC Global Surface Water Mapping Layers

JRC Global Surface Water Mapping Layers is basically a
derived product of Landsat datasets. This dataset is
generated using Landsat 5, 7, 8’s different scenes from 16-
03-1984 to 31-12-2020. It classifies each pixel into water or
non water area [28]. This dataset was used to classify water
and non-water areas on the basis of pixel values.

C. SRTM DEM
The Digital Elevation Model (DEM) which is used in the
analysis is Shuttle Radar Topography Mission (SRTM)
DEM at 3 arc-second resolution. It was used to remove

Seventh International Conference on Aerospace Science & Engineering 192

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

signals that were weak backscattered from the topographic

area and to extract slope information.

D. MODIS
Moderate Resolution Imaging Spectroradiometer (MODIS)
was launched by NASA in 1999 on board Terra Satellite and
later in 2002 on Aqua. MODIS gives a moderate spatial
resolution from 250 m to 1000 m with a temporal resolution
of 1 day [29]. MODIS land cover product is the only dataset
that gives global coverage and is available in GEE. MODIS
was used to estimate the damaged cropland of the study area.
Figure 3: Comparison of VH-pol, VV-pol & Ascending, and
E. Vector data Descending orbit pass
The Study area’s administrative boundary shapefile was
used to extract the exact area of interest. The next step was to choose polarization between VV and
VH. VV-pol, causes volume scattering in inundated
V. METHODS AND MATERIALS vegetated areas because it is sensitive to rough water, and
The following methodology for delineation of flood extent may provide less information or may cause misclassification
and assessment of damaged crop was performed as shown [32]. Schumann et al. [31] suggested that in the absence of
in Fig.2. 3 criterions (Polarization, Orbit pass, and filter) HH-pol data the other preferable choices are VH and HV
were defined to analyze that which are suitable for said polarized data, HV-pol data was unavailable for the study
study. Total 08 images were analyzed. First of all, 8 images area therefore, VH was chosen.
of before and after flood were accessed in Google Earth When Sentinel-1is in ascending pass (Northward) it looks at
Engine (https://code.earthengine.google.com/) considering the area towards west and when it descends (Southward) it
the following mentioned parameters. is observing towards east [33]. The study area falls in
southeast of Province Sindh. Therefore, for this study
descending pass was chosen. In Fig.4 flood can be observed
more precisely in Descending orbit pass than Ascending
orbit pass.

Figure 4: Ascending & Descending Orbit Pass comparison

After selecting VH-pol and descending pass, 2 filters were

applied on the images of before and after flood to reduce the
speckle-noise. A comparison was also made to highlight the
Figure 2: Flowchart efficiency of which filter can suppress more noise, shown in
Fig.5. Observed areas of Deh No. 150 are on the West side,
Polarization (VH, VV) & Orbit Pass Deh No. 155 are on North side, and Deh No. 160 are on
(Ascending/Descending) as shown in Fig.3. Polarization South side. Dark areas are depicting sandy areas due to
plays a vital role in backscattering mechanism. Hess et al. smooth surface backscattering mechanism. There is also bog
analyzed the best results in co-pol channel and suggested its area and agriculture land features that exist there and
use for flood mapping i.e., HH-pol [30]. However, HH and highlighted areas are showing difference between the focal
HV polarizations are available just for some maritime and median and focal mean filters results. Focal mean filter can
polar regions and HH-pol data for the said study area was suppress more noise than that of focal median filter.
unavailable [26].

Seventh International Conference on Aerospace Science & Engineering 193

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 5: Comparison of Mean & Median Speckle-reducing

filters
Figure 8: Damaged crop layer overlaying on SAR image after
flood
After selecting VH-pol, descending pass and mean filter
slope were calculated using NASA’s SRTM DEM 3 arc-
second resolution. Then using JRC water layer, a difference
layer was created in which the water bodies or water pixels
existing for time greater than 10 months were given
permanent water label and were masked from flood area. An
automated thresholding value of -20.95 dB was generated
using Otsu algorithm in GEE and a histogram was generated
as shown in Fig.6. Succeeding auto-threshold value, change
detection was calculated and flood map was prepared. In the
histogram it can be seen maximum value is falling at -20.95
dB. And finally, using MODIS global coverage data crop
layer was extracted and was overlayed on SAR after flood
image as shown in Fig.7. An assessment of damaged crop
Figure 9: Damaged crop layer overlaying on SAR image after
due to flood was made which can be seen in Fig.9, flood with flooded areas layer
highlighted with red circle in green color with flooded area
layer on SAR after flood image. VI. RESULTS AND DISCUSSION
Here is the result of the said analysis.

Figure 6: Histogram
Figure 10: Study Area Figure 11: Image before flood

Figure 12: Image after Figure 13: Difference layer

Figure 7: Crop layer overlaying on SAR image after flood
flood

Seventh International Conference on Aerospace Science & Engineering 194

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig.14 is showing flood layer in blue overlaying on

difference layer, Fig.15 is showing flood layer (in blue)
overlaying on the image of pre-event.
MODIS data was used because it gives global land coverage,
selected ‘LC_Type1’ band and cropland was extracted and
then overlayed on flood layer. Fig.16 showing is MODIS
global landcover coverage, Fig.17 is showing the final result
of the analysis in which water in blue and damaged crop
detected due to flood in green can be seen.

Figure 14: Flood layer Figure 15: Flood layer VII. CONCLUSION
overlaying on overlaying on image before National Disaster Management Authority (NDMA) had
difference layer flood reported that disastrous rainfall affected the districts of
Mirpur Khas, Karachi, Hyderabad, Badin etc. The flood of
September 2020 in Mirpur Khas caused many casualties in
the form of lives, infrastructure, and crops. Flood mapping
and associated damages should be evaluated quickly to
provide near real time flood information to concerned
authorities to mitigate the losses. It is somehow challenging
to compute flood mapping on a desktop-based system that
could not be much helpful for timely response. This study
might be helpful for flood mapping using SAR data
integrated with automated threshold technique (Otsu) and its
associated damages with MODIS data on a cloud-based
platform i.e., GEE. While starting an analysis using SAR
Figure 16: MODIS Figure 17: Flood & damage data one must be aware of basic SAR principles and
landcover assessment layers overlaying
backscattering parameters i.e., frequency & polarization etc.
on image before flood
HH-pol gives the best result of inundated areas but in the
absence of HH-pol, the next better option is VH-pol and that
During the flooding season, concerned authorities i.e., can provide reliable results. Otsu algorithm can be used for
NDMA, PMD etc, require updated and timely information automated thresholding in GEE. In case of emergency, GEE
to demarcate the affected area, to manage the situation, and can be used as key tool to access the spaceborne data.
mitigate the associated losses. Regarding this situation, a MODIS data has its own benefits, in regard of resolutions
spatio-temporal analysis was made in which Spaceborne although it has coarser spatial resolution with a good
SAR GRD data was used to locate the flooded areas and temporal resolution and damages can be assess using it.
MODIS data was used to assess the associated damages.
Fig.10 is showing the study area administrative boundary,
Fig.11 is showing the Sentinel-1 SAR GRD image of pre- REFERENCES
event taken on 21/09/2019, Fig.12 is showing the Sentinel-
1 SAR GRD image of post-event taken on 21/09/2020, [1] J.-M. Martinez and T. Le Toan, "Mapping of flood dynamics
Fig.13 is depicting difference layer (DL) that was obtained and spatial distribution of vegetation in the Amazon
by “(1)”. floodplain using multitemporal SAR data," Remote sensing of
Environment, vol. 108, no. 3, pp. 209-223, 2007.
[2] J. García-Pintado, D. C. Mason, S. L. Dance, H. L. Cloke, J.
𝐷𝐿 = 𝐼𝑚𝑎𝑔𝑒 𝑎𝑓𝑡𝑒𝑟 𝑒𝑣𝑒𝑛𝑡 ÷ 𝐼𝑚𝑎𝑔𝑒 𝑏𝑒𝑓𝑜𝑟𝑒 𝑒𝑣𝑒𝑛𝑡 (1) C. Neal, J. Freer and P. D. Bates, "Satellite-supported flood
forecasting in river networks: A real case study.," Journal of
Then JRC Global Surface Water Mapping layer was Hydrology, vol. 523, pp. 706-724, 2015.
accessed and its ‘Seasonality’ band was chosen to classify [3] L. Landuyt, A. V. Wesemael, G. J.-P. Schumann, R.
the permanent or seasonal water. Values were assigned on Hostache, N. E. C. Verhoest and F. V. Coillie, "Flood
mapping based on synthetic aperture radar: An assessment of
the basis of the following “(2)” and “(3)”. Permanent water established approaches," IEEE Transactions on Geoscience
was given a value of ‘0’ and seasonal or flood water was and Remote Sensing, vol. 57, no. 2, pp. 722-739, 2018.
given a value of ‘1’. After classifying the permanent and [4] Y. Li, S. Martinis, S. Plank and R. Ludwig, "An automatic
flood water slope was calculated using SRTM DEM of 3 arc change detection approach for rapid flood mapping in
resolution to further refine the flood results. Sentinel-1 SAR data," International journal of applied earth
observation and geoinformation, vol. 73, pp. 123-135, 2018.
[5] A. Twele, W. Cao, S. Plank and S. Martinis, "Sentinel-1-
𝑊𝑎𝑡𝑒𝑟 𝑒𝑥𝑖𝑠𝑡𝑎𝑛𝑐𝑒 > 10 𝑚𝑜𝑛𝑡ℎ𝑠 = 𝑃𝑒𝑟𝑚𝑎𝑛𝑒𝑛𝑡 (2)
based flood mapping: a fully automated processing chain,"
International Journal of Remote Sensing, vol. 37, no. 13, pp.
𝑊𝑎𝑡𝑒𝑟 𝑒𝑥𝑖𝑠𝑡𝑎𝑛𝑐𝑒 < 10 𝑚𝑜𝑛𝑡ℎ𝑠 = 𝑠𝑒𝑎𝑠𝑜𝑛𝑎𝑙 (3) 2990-3004, 2016.
[6] S. Long, T. E. Fatoyinbo and F. Policelli, "Flood extent
mapping for Namibia using change detection and
thresholding with SAR," Environmental Research Letters,
vol. 9, no. 3, p. 035002, 2014.

Seventh International Conference on Aerospace Science & Engineering 195

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

[7] P. A. Townsend and S. J. Walsh, "Modeling floodplain [23] N. Gorelick, M. Hancher, M. Dixon, S. lyushchenko, D. Thau
inundation using an integrated GIS with radar and optical and R. Moore, "Google Earth Engine: Planetary-scale
remote sensing," Geomorphology, Vols. 21 (3-4), pp. 295- geospatial analysis for everyone," Remote sensing of
312, 1998. Environment , vol. 202, pp. 18-27, 2017.
[8] N. G. Pricope, "Variable-source flood pulsing in a semi-arid [24] X. Xiao, S. Boles, J. Liu, D. Zhuang, S. Frolking, C. Li, W.
transboundary watershed: the Chobe River, Botswana and Salas and B. MooreIII, "Mapping paddy rice agriculture in
Namibia," Environmental monitoring and assessment, vol. southern China using multi-temporal MODIS images,"
185, no. 2, pp. 1883-1906, 2013. Remote sensing of environment , vol. 95, no. 4, pp. 480-492,
L. L. HESS, J. M. MELACK and D. S. SIMONETT, "Radar 2005.
[9]
detection of flooding beneath the forest canopy: a review," [25] PMD, "Flood Forecasting Division," Pakistan Meteorological
International Journal of Remote Sensing, vol. 11, no. 7, pp. Department, 2021. [Online]. Available:
1313-1325, 1990. https://ffd.pmd.gov.pk/homepage/.. [Accessed April 2021].
[10] D. C. Mason, I. J. Davenport, J. C. Neal, G. J.-P. Schumann [26] Google Earth Engine, "Synthetic Aperture Radar (SAR)
and P. D. Bates, "Near Real-Time Flood Detection in Urban Basics," Google, 2021. [Online]. Available:
and Rural Areas Using High-Resolution Synthetic Aperture https://developers.google.com/earth-
Radar Images," IEEE transactions on Geoscience and Remote engine/tutorials/community/sar-basics. [Accessed April
Sensing, vol. 50, no. 8, pp. 3041-3052, 2012. 2021].
[11] H. Shahabi, A. Shirzadi, K. Ghaderi, E. Omidvar, N. Al- [27] Google, "Wikipedia," Google, 2021. [Online]. Available:
Ansari, J. J. Clague, M. Geertsema, K. Khosravi, A. Amini , https://en.wikipedia.org/wiki/Mirpur_Khas. [Accessed April
S. Bahrami , O. Rahmati , K. Habibi , A. Mohammadi, H. 2021].
Nguyen , A. M. Melesse , B. B. Ahmad and A. Ahmad, "Flood Google Earth Engine, "Earth Engine Data Catalog," Google,
Detection and Susceptibility Mapping Using Sentinel-1
[28]
2021. [Online]. Available:
Remote Sensing Data and a Machine Learning Approach: https://developers.google.com/earth-
Hybrid Intelligence of Bagging Ensemble Based on K- engine/datasets/catalog/JRC_GSW1_3_GlobalSurfaceWater.
Nearest Neighbor Classifier," Remote Sensing, vol. 12, no. 2, [Accessed April 2021].
p. 266, 2020.
[29] NASA, "MODIS," NASA, 2021. [Online]. Available:
[12] X. Shen , D. Wang , K. Mao, E. Anagnostou and Y. Hong , https://modis-land.gsfc.nasa.gov/. [Accessed April 2021].
"Inundation Extent Mapping by Synthetic Aperture Radar: A
Review," Remote Sensing, vol. 11, no. 7, p. 879, 2019. [30] . L. L. Hess and M. J. M. , "Mapping wetland hydrology and
vegetation with synthetic aperture radar," International
[13] E. Psomiadis, "Flash flood area mapping utilising Journal of Ecology and Environmental Sciences, vol. 20, no.
SENTINEL-1 radar data," in Earth Resources and 1-2, pp. 74-81, 1994.
Environmental Remote Sensing/GIS Applications VII (Vol.
10005, p. 100051G), Eidenburgh, UK, 2016. [31] G. Schumann, R. Hostache, C. Puech, L. Hoffmann, P.
Matgen, F. Pappenberger and L. Pfister, "High-Resolution 3-
[14] P. Manjusree, L. P. Kumar, . C. M. Bhatt, G. S. Rao and V. D Flood Information From Radar Imagery for Flood Hazard
Bhanumurthy, "Optimization of threshold ranges for rapid Management," IEEE transactions on geoscience and remote
flood inundation mapping by evaluating backscatter profiles sensing, vol. 45, no. 6, pp. 1715-1725, 2007.
of high incidence angle SAR images.," International Journal
of Disaster Risk Science, vol. 3, no. 2, pp. 113-122, 2012. [32] R. Manavalan, "Review of synthetic aperture radar frequency,
polarization, and incidence angle data for mapping the
[15] R. Kumar, "Flood Inundation and Hazard Mapping of 2017 inundated regions," Journal of Applied Remote Sensing, vol.
Floods in the Rapti River Basin Using Sentinel-1A Synthetic 12, no. 2, p. 021501, 2018.
Aperture Radar Images," Applications and Challenges of
Geospatial Technology. Springer, Cham, pp. 77-98, 2019. [33] P. Tarikhi, "INSAR OF AQUATICBODIES," ISPRS-
International Archives of the Photogrammetry, Remote
[16] L. Peter, M. Matjaž and O. Krištof, "Detection of Flooded Sensing and Spatial Information Sciences, vol. 39, pp. 85-90,
Areas using Machine Learning Techniques:Case Study of the 2012.
Ljubljana Moor Floods in 2010," Disaster Advances, vol. 6,
no. 7, pp. 4-11, 2013.
[17] R. Malinowski, G. Groom, W. Schwanghart and G. Heckrath
, "Detection and Delineation of Localized Flooding from
WorldView-2 Multispectral Data," Remote Sensing, vol. 7,
no. 11, pp. 14853-14875, 2015.
[18] M. Ul Haq, M. Akhtar, S. Muhammad, S. Paras and J.
Rahmatullah, "Techniques of Remote Sensing and GIS for
flood monitoring and damage assessment: A case study of
Sindh province, Pakistan," he Egyptian Journal of Remote
Sensing and Space Science, vol. 15, no. 2, pp. 135-141, 2012.
[19] H. Cao , H. Zhang, C. Wang and B. Zhang, "Operational
Flood Detection Using Sentinel-1 SAR Data over Large
Areas," Water, vol. 11, no. 4, p. 786, 2019.
[20] M. S. Sardar, . M. A. Tahir and M. I. Zafar, "Poverty in
Riverine Areas : Vulnerabilities, Social Gaps and Flood
Damages," Pakistan Journal of Life and Social Science, vol.
6, no. 1, pp. 25-31, 2008.
[21] J.-H. Xue and Y.-J. Zhang, "Ridler and Calvard’s, Kittler and
Illingworth’s and Otsu’s methods for image thresholding,"
Pattern Recognition Letters, vol. 33, no. 6, pp. 793-797, 2012.
[22] C. Vishnu, K. Sajinkumar, T. Oommen, R. Coffman, K.
Thrivikramji, V. Rani and S. Keerthy, "Satellite-based
assessment of the August 2018 flood in parts of Kerala,
India," Geomatics, Natural Hazards and Risk, vol. 10, no. 1,
pp. 758-767, 2019.

Seventh International Conference on Aerospace Science & Engineering 196

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Prototype of an Autonomous Farming Robot:

A Smart Farming System

Moaz Rao Muhammad Shaarif Younas Nauman Latif

Dept. of Avionics Engineering Dept. of Avionics Engineering Dept. of Avionics Engineering
The Superior University Lahore The Superior University Lahore The Superior University Lahore
Lahore, Pakistan Lahore, Pakistan Lahore, Pakistan
[email protected] [email protected] [email protected]

Ahmad Hammad Sadia Ilyas Mahmood Qureshi

Dept. of Avionics Engineering Dept. of Avionics Engineering Dept. of Avionics Engineering
The Superior University Lahore The Superior University Lahore The Superior University Lahore
Lahore, Pakistan Lahore, Pakistan Lahore, Pakistan
[email protected] [email protected] [email protected]

Abstract— The automated household products in recent the benefits of precision agriculture. By placing the tooling
technology are adding day by day, we have proposed an equipment on tracks rather than driving a tractor, Farming
autonomous CNC farming project that works on a robot takes a completely new approach to precision
microcontroller-based Cartesian coordinate system. We present
agriculture. A similar method has been used for decades in
you a Farming Robot that is designed similar to the 3D printer
printers, machine tools, and, more recently, 3D printers, CNC
or milling machine in shape and similar precise to perform
operation that includes irrigation, seeding, weeding and
machines, and more. It incorporates Arduino as well as other
monitoring soil moisture. This paper briefly describes our components.
project that will assist farmers in monitoring and controlling Farming robot is an autonomous machine designed as a
various activities of farming via Arduino and other components.
CNC machine. It performs various acts related to farming such
The status of agricultural parameters is notified to the farmer,
as watering, weeding, seeding and it monitors and grows your
allowing him to make appropriate decisions on time. The major
benefit of this system is that it is semi-automated, meaning that crops.
the farmer makes the choice rather than a completely automated
II. LITERATURE REVIEW
system that results in precision agriculture. It also gets around
existing traditional procedures that demand a lot of money, Crop data can be managed by agricultural management
energy, labor, and time. As it’s the demand of the environment system in such a way that results are orchestrated to address
to have such a product that can grow healthier food in this customized solution for each farm by data driven management
gradually variant environment and climate change. [1]. Watering system can be managed by automatic plant robot
with scheduling system. In which various tests and tools are
Keywords— computer numerical control (CNC), farming, bot,
including like humidity test, soil moisture test and testing of
moisture [2]. Smart farming system using sensors such as
I. INTRODUCTION infrared sensor which is used to regulate the speed of motor.
And opto coupler converts electrical signal into light and
The Agricultural production has kept up with food activates the switches [3]. The farming robot provides
production due to the industrial and petrochemical revolution, information via the internet, as well, which makes the system
but only at the cost of soil, the environment, and human health. remotely and flexible. In Pakistan 40% people depend on
Technological advancements and economies of scale have agriculture for sustainable living [4]. AgroTutor is a mobile
been driving the increase, but this drive is plateauing and a application that can take data from framers and give the
paradigm shift is needed. Conventional agriculture methods appropriate advice related to their farming. Also, it provides
are unsustainable and need to be re-evaluated. weather information to farmers. And for business, it also
The farming industry has seen a gradual increase in provides the data of historical yield potential and local
accuracy and precision due to technologies such as GPS and benchmarking information [5]. Advanced technology, such as
short-range radio. Autopilots, digital mapping, computer the internet of things (IOT), plays an important role in
vision, and variable rate devices are among the latest streamlining the process for producers. Harvesters, robotic
innovations. Through the application of these technologies, weeding machines, unmanned aerial vehicles (UAVs), and
free-driving tractors are gradually adapted to realize more of other automatic machinery are currently helpful to growers in

Seventh International Conference on Aerospace Science & Engineering 197

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

terms of crop monitoring [6]. Existing research on digitization an IOT application system that goes through the creation,
in agriculture and rural areas does not provide a complete processing, and distribution of agricultural products. [15].
picture of current and future effects. As a result, additional
III. PROBLEM STATEMENT
digitized equipment are needed to fill in the gaps in advanced
farming techniques' outputs [7]. A method assisted by a In this era of automation, there should be such a
mobile object for detecting crop disorder outbreaks and technology as a smart home device that can automate the
managing them using permitted control strategies. The system farming services on one click for the people who have no
ask for the soil and field selection and suggest the ideal farming background so that the agriculture gets promoted as
condition in which color, shape, distribution pattern etc. are well as healthy food production keeps growing.
included through which disorder can identify and overcome
IV. METHODOLY
timely [8]. Capacitive sensors may be used to assess soil
moisture efficiently, and a cost drop is possible while Farming robot is an open source robot which works on 3D
maintaining sufficient accuracy. Analog or digital SDI-12 CNC based operations. We have taken this open concept and
outputs are available for use with the sensor. The sensor's two try to make it in our own way. In our design circuitry
primary components (oscillator and microcomputer) were controllers and two stepper motors for x-axis are fixed unlike
chosen because they are general-purpose products with a high to the farming robot which is an open source machine. Also,
temporary availability for manufacturing [9]. Wetting and farming robot they are using thread rod for the z-axis
drying events were detected by the sensors TDR315, CS655, movement unlike to this, we have used belt and pulley
GS1, SM100, and CropX. Based on root mean square error mechanism to achieve the vertical linear actuation. These are
(RMSE), the TDR315, CS655, and GS1 sensors had sufficient the two major differences in the mechanical structure of our
accuracy rate for guided watering at the site with low salt farming robot.
concentration and low clay component (LSLC), but not at the Also, we are designing our own coded firmware which will
site with high salt concentrations and high clay component further implement on cloud computing in future.
(HSHC) [10]. Sensor data is needed as input in the automation
system application. Sensors in automation systems are used to A. Material selection
create a system that works autonomously to regulate As we are designing the machine for both open and closed
performance of the system. Because there are various aspects environment, therefore, we have taken the material
to consider while constructing an industrial automation selection seriously.
system. Water temperature control, irrigation control, For mechanical structure, we are using anodized
fertilization, and oxygen saturation in the water must all be aluminum made extrusions as they have high strength
taken into account [11]. There is a screening system in against humidity and corrosion. Also, it’s not such
addition to the automation system used in agriculture. This 48- thermal conductive which makes it almost immune to
hour screening system makes use of a wireless sensor network thermal expansion so that it cannot creates any dimension
(WSN). This screening system monitors conditions in real error in mechanical structure. It’s also weather proof and
time [12]. Farmers will benefit from the device since it will no oxidizing property [16].
provide them with real-time data (temp, humid, soil humidity, For the farming tools and parts, we are using ultraviolet
Ultraviolet index, and Infrared index) from their farm, (UV) protected plastics so that it cannot harm or cause any
allowing them to take the required actions to increase crop kind of contamination to the food.
production while conserving natural resources (water,
B. Cad model
fertilizers) [13]. Management of water resources is by far the
most sub-vertical (28.08%), followed by agricultural
production (14.60%), and finally smart farming (14.60%).
(10.11 %). The measuring of surrounding temperature (24.87
%) and ecological moisture (24.87%) yielded the best results
in terms of sensor data collecting (19.79%). Other sensor data
includes moisture levels (15.73%) and soil pH. (7.61 %).
According to statistics, Wi-Fi is the most widely used
technology in IOT application development (30.27 %),
followed by digital technology (21.10 %) [14]. Smart
agricultural designs based on Internet - Of - things
technologies may effectively implement the function of real- Figure 1: Computer Aided Design CAD model of Farming Robot
time data transfer and information processing while also
advancing smart agriculture development. Agriculture IOT is

Seventh International Conference on Aerospace Science & Engineering 198

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

C. Hardware selection These are flexible hollow structures used to guide,

support, and protect cables, and water lines when in
For our farming robot, we have selected the material from
motion
open builds open source parts for CNC as they are the
 wires
pioneers of the CNC building.
4 core and 3 core and single core wire is used to
There is a total of 28 components are used.
supply power to stepper motors, water pump, limit
 1.20x40mm V slot aluminum extrusion (17) ft. in switches and encoders.
total
D. General layout & working
 2.20x20mm V slot aluminum extrusion (3 ft.)
After burning the code into Arduino, Arduino will call for the
 90-degree L-plates (4 in total)
input coordinates from user. After getting the coordinates and
 The two of these 90-degree L-plates are used to computing the required movement, it runs the stepper motors
mount the V slots arms of X-axis on the X-axis small according to the loops as stepper motor achieves one rotation
carriage plates. And the other two 90-degree L plates in 200 steps because it covers 1.8 degrees in each step so that
are used to mount Y-axis gantry V slot with arms of the PWM (pulse width modulated) signal should have to be
X-axis. sent in a loop in a condition to achieve the given coordinate
using motor drivers accordingly through which the robot will
 V slot Wheels (24 in total)
go on to the specified coordinate and perform the farming task
 These wheels roll over the extrusion to give the assigned to it.
smooth movement for actuation. 16 of them are used
on 4 carriage plates on two V-slot of X-axis, 4 of
them are used on the universal gantry plates on V
slot of Y-axis and 4 of them are used another
universal gantry plate mounted with main gantry
plate to give the vertical actuation to Z-axis V-slot.
 Carriages plates
 These plates are used to carry the whole moveable
frame in which arms are directly mounted to these
plates with help of L-bar.
 Gantry plates
 5M bolts
 M3 bolts
 These are used to mount motors with connecting
plates over the extrusions and universal gantry
plates. Figure 2: coordinate system and connection layout of Farming
Robot
 M5 nuts with washers
 There are used to mount frictionless pulleys, L bars E. Block diagram
and rollers.
 M5 T nuts
 These are used to mount the extrusions with each
other, to mount the connecting plates for stepper
motors with V slots, L-plates, belts etc.
 GT 2 belt
 These belts are used to move X, Y, and Z-axis
extrusions with the help of stepper motors.
 20 teeth GT2 pulley
 It mounts on the stepper motor D shaft to push and
pull the belt.
 Bearing pulleys
 These are used to roll the belts around the v slots.
 Motor connecting plates
 These are used to mount the motors with a V slot
and gantry plates. Figure 3: Block Diagram of info flow in Farming Robot
 Tool kit
 Drag chain cable carriers

Seventh International Conference on Aerospace Science & Engineering 199

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

This is the big picture of our project in which our future work H. Software Selection
is also included. For now we are giving input from PC but in
ON SHAPE
future we can give input from mobile, laptop or wirelessly by
making a cloud storage. From cloud storage, input will go to We have used on shape online software to design the CAD
Arduino then it will burn the input. Input will go to motor model of Farming Robot.
drivers and drivers will give the input to motors. ARDUINO IDE (integrated development environments)
We have used Arduino IDE software for Programming and
PC/Smartphone motor synchronization.
FRITZING
We have design the schematic diagram on fritzing software.
Aurdino V. DISCUSSION
Our project is design and development of CNC based farming
robot. So, we design and developed the mechanical system of
Motor Drivers Watering tool the farming robot. In which we used the aluminum material.
Then we develop the coordinate system and control
parameters in which x-axis movement via two synchronized
Two X Axis stepper motors. Y-axis movement via single stepper motor.
Motors
And z-axis movement via single stepper motor. We designed
and develop the spraying mechanism of our robot which is
successfully working. Agriculture sprayers are complete
Y Axis motor spray systems engineered for generating pressure to drive
spray fluid from a tank out to the sprayer's nozzle to crop or
soil. Sprayers are most often used for the administration of
Z Axis motor water, insecticides, pesticides, herbicides, and fertilizer.
VI. FUTURE WORK/CONCLUSION
Figure 4: General Flowchart of central idea of Farming The problem was that there should have to be a product that
Robot can automate the farming services as a household machine
and our Farming Robot is doing it perfectly but the software
F. Schematic diagram framework of the robot is still in progress due to more Future
In block diagram it shows the flow of the input. PC/ smart plans but the research work has been done and prototype
phone gives the input to the Arduino which goes to motor made and testing is under process. In future, we are designing
drivers and watering tool. From motor drivers it goes to all our own coded firmware which will further implement on
cloud computing in future. Also, in open source farming
the motors. First it goes to two X Axis motors then to Y and
robot, the farming tools are changed by the electromagnetic
Z axis motors. system which are few of them. Unlike to this, we are working
on to design 4 in 1 fixed tool which rotates at 90 degrees to
change the tool which is already embedded on it. In this way,
our design is more sophisticated, and price controlled than the
open source farming robot. Also, farming robot is
environment suitable machine as its produce healthier food
than the mass production of the locally available food as it
monitors your farming 24 hours. It works best even in
different climates as well.
VII. REFERENCES
1. M. Chetan Dwarkani, R. Ganesh Ram, S. Jagannathan, and R.
Figure 5: schematic diagram of Farming Robot Priyatharshini, “Smart farming system using sensors for
agricultural task automation,” Proc. - 2015 IEEE Int. Conf.
G. Electronics Technol. Innov. ICT Agric. Rural Dev. TIAR 2015, no. May, pp.
 Arduino Mega 49–53, 2015, doi: 10.1109/TIAR.2015.7358530.

 Motor Drivers L298N 2. S. Suakanto, V. J. L. Engel, M. Hutagalung, and D. Angela,

 Stepper motors NEMA 17 “Sensor networks data acquisition and task management for
decision support of smart farming,” 2016 Int. Conf. Inf. Technol.
 Limit Switches Syst. Innov. ICITSI 2016 - Proc., 2017, doi:
 HEX Inverting IC 10.1109/ICITSI.2016.7858233.

 LED strip light 3. M. Chetan Dwarkani, R. Ganesh Ram, S. Jagannathan, and R.

 Water pump Priyatharshini, “Smart farming system using sensors for
agricultural task automation,” Proc. - 2015 IEEE Int. Conf.
 Relay switch Technol. Innov. ICT Agric. Rural Dev. TIAR 2015, no. Tiar, pp.
 Motor encoders 49–53, 2015, doi: 10.1109/TIAR.2015.7358530.

 Power supplies 12V and 5V 4. M. M. Galgalikar, “Real-time automization of agricultural

Seventh International Conference on Aerospace Science & Engineering 200

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
environment for social modernization of indian agricultural Soil Moisture Sensors under Irrigated Field Conditions in
system,” 2010 2nd Int. Conf. Comput. Autom. Eng. ICCAE 2010, Oklahoma." https://doi.org/10.3390/s18113786
vol. 1, no. March, pp. 286–288, 2010, doi:
10.1109/ICCAE.2010.5451949. 11. Rajaoarisoa, L. H., M’sirdi, N. K., & Balmat, J. F. (2012).
"Micro-climate optimal control for an experimental greenhouse
5. Bayas, J. C. L., Gardeazabal, A., Karner, M., Folberth, C., automation". 2nd International Conference on Communications
Vargas, L., Skalský, R., Balkovič, J., Subash, A., Saad, M., Computing and Control Applications, CCCA 2012.
Delerce, S., Cuaresma, J. C., Hlouskova, J., Molina-Maturano, J., https://doi.org/10.1109/CCCA.2012.6417903
See, L., Fritz, S., Obersteiner, M., & Govaerts, B. (2020).
"Agrotutor: A mobile phone application supporting sustainable 12. Momirović, N., Vasić, B., Raičević, D. and Oljača M.: "Technical
agricultural intensification". Sustainability (Switzerland), systems for microclimate control in greenhouses", Agricultural
12(22), 1–10. https://doi.org/10.3390/su12229309 Engineering, Faculty of Agriculture, Universities in Belgrade,
Institute of Agricultural Engineering, No.4, 2007

6. Khan, N., Ray, R. L., Sargani, G. R., Ihtisham, M., Khayyam, M., 13. Doshi, J., Patel, T., & Bharti, S. K. (2019)." Smart Fanning using
& Ismail, S. (2021). "Sustainability Current Progress and Future IOT, a solution for optimally monitoring fanning conditions".
Prospects of Agriculture Technology: Gateway to Sustainable Procedia Computer Science, 160, 746–751.
Agriculture". https://doi.org/10.3390/su13094883 https://doi.org/10.1016/j.procs.2019.11.016

7. Rolandi, S., Brunori, G., Bacco, M., Scotti, I., & Michailidis, A. 14. A. A Raneesha Madushanki, Malka N Halgamuge, W. A. H.
(2021). "The Digitalization of Agriculture and Rural Areas: Surangi Wirasagoda and Ali Syed, “Adoption of the Internet of
Towards a Taxonomy of the Impacts". Things (IoT) in Agriculture and Smart Farming towards Urban
https://doi.org/10.3390/su13095172 Greening": A Review International Journal of Advanced
Computer Science and Applications(IJACSA), 10(4), 2019.
8. [Sivagnanasundaram, J., Goonetillake, J., Buhary, R., http://dx.doi.org/10.14569/IJACSA.2019.0100402
Dharmawardhana, T., Weerakkody, R., Gunapala, R., & Ginige,
A. (2021). "Digitally-Enabled Crop Disorder Management 15. Li, C., & Niu, B. (n.d.). "Design of smart agriculture based on
Process Based on Farmer Empowerment for Improved big data and Internet of things". Research Article International
Outcomes: A Case Study from Sri Lanka." Journal of Distributed Sensor Networks, 2020(5).
https://doi.org/10.3390/su13147823 https://doi.org/10.1177/1550147720917065

9. González-Teruel, J. D., Torres-Sánchez, R., Blaya-Ros, P. J., 16. Technical datasheet anodized aluminum (n.d.)
Toledo-Moreo, A. B., Jiménez-Buendía, M., & Soto-Valles, F. https://www.trotecmaterials.com/downloads/dl/file/id/555/produ
(n.d.). (2021). "Design and Calibration of a Low-Cost SDI-12 ct/0/technical_datasheet_anodized_aluminium.pdf
Soil Moisture Sensor". https://doi.org/10.3390/s19030491
17. Aronson, R. landon. (2013)." Humanity’s open-source automated
10. Datta, S., Taghvaeian, S., Ochsner, T. E., Moriasi, D., Gowda, P., precision farming machine". https://farm.bot/blogs/news/the-
& Steiner, J. L. (n.d.). "Performance Assessment of Five Different farmbot-whitepaper?_pos=5&_sid=048d7d7b4&_ss=r

Seventh International Conference on Aerospace Science & Engineering 201

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

DESIGN AND CFD ANALYSIS OF OUTER AERODYNAMICS

OF 10KW HORIZONTAL-AXIS WIND TURBINE
Noman Ahsan Saqlain Abbas* Muhammad Ahmad
Department of Mechanical Engineering Department of Mechanical Engineering Department of Mechanical Engineering
University of Engineering and University of Engineering and University of Engineering and
Technology Lahore (Narowal Campus), Technology Lahore (Narowal Campus), Technology Lahore (Narowal Campus),
Pakistan Pakistan Pakistan
[email protected] [email protected] [email protected]

Ijaz Ali Zulkarnain Abbas Muhammad Umair Shahid

Department of Mechanical Engineering Department of Mechanical Department of Electrical Engineering
University of Engineering and Engineeering Khawaja Fareed University of
Technology Lahore (Narowal Campus), NFC, Institute of Engineering and Engineering and Information
Pakistan Technology Multan, Pakistan Technology
[email protected] [email protected] Rahim Yar Khan, Pakistan
[email protected]

Abstract— During past years, wind power has been widely Table 1. Design parameters
recognized and used as a renewable energy resource. However, Type Permanent magnet 3-
there exists sufficient difference in the available wind Generator phase AC electrical
perspective and the overall power production. Pakistan has
Max Power 14 KW
very good wind potential, especially in the provinces of Sindh
and Baluchistan. In the current research study, a 10,000-watt Configuration Horizontal Axis
horizontal axis wind turbine (HAWT) is designed for the Length of Blade 4.61 m
Taftan, Baluchistan region to meet the energy requirements of Rotor Rotor Diameter 3.2 m
Pakistan. For designing the blade profile, the Blade Element Swept Area 66.76 m2
Momentum method is used and CFD is utilized for analysis of Rotor Speed 126.8 rpm
the designed profile. Tip Speed Ratio 7.2
Keywords— Aerodynamics, Wind turbine, TSR, Warlock, Wight Rotor 1,000 kg
CFD analysis Hight 50 m
Tower Material Stainless steel
I. INTRODUCTION
Wind Upwind
It is very difficult for developing countries, including
Pakistan, to fulfill the whole population's energy needs. To
extract energy from renewable resources has become the By using Warlock design parameters, the final shape of the
need of the hour [1]. Energy consumption, energy wind turbine with tower height is drawn on SOLIDWORKS
conversion, and utilizing energy in a sustainable way have 2020 by using the loft command.
been used and recognized a lot for rotating machinery [2, 3].
There are different renewable resources such as solar, wind,
biomass and hydropower potential [4]. The government of
Pakistan is struggling hard to overcome the energy crisis by
initiating mega power projects to achieve sustainable
development goals [5, 6]. The method to extract the energy
from wind is getting mature and reliable all over the world
[7-12]. Several parameters are required to obtain the optimal
design of a wind turbine, including the number of blades,
blade material, blade length, rotor diameter, tip speed ratio,
the wind cut in and cut out speed, rated power, tower height,
etc.
In this research, a computational analysis has been
conducted and presented to study the profile of designed
10KW wind turbine for backward areas in Pakistan. The
proposed work provides a pollution-free, decentralized, and
reliable energy framework. Fig.1 Warlock calculator.

II. DESIGN PARAMETERS

Analyzing the wind speed in different districts of
Pakistan, the Taftan, Baluchistan region is selected in this
study for the design of 10KW HAWT. The gearing method
is the planetary gearbox.

Seventh International Conference on Aerospace Science & Engineering 202

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Pitch Angle 2
0.5 0.406 0.411
0.4
0.3

Cp
0.2 0.102
0.1
0
12.57 6.28 4.19

TSR

Fig.4 Relation between TSR and for , n= 1,2 & 3

Fig.2 Design with tower height Pitch Angle 3

0.5
The power of turbine is calculated by using following 0.3996 0.401
formula; 0.4
P actual = ½ 𝞺 v3 A * Cp 0.3
P actual = 10045.57 W

Cp
Density of Air = 𝞺 =1.225 kg/m3, 0.2
0.082
wind speed = v = 8.5 m/s 0.1
swept area = A = 66.76 m2
Power coefficient=Cp = 0.398 0
12.57 6.28 4.19
Rotational speed= w=12.27 rad/s or 126.8rpm
TSR
The power coefficient is characterized as the ratio of the
examination between genuine power produced by the rotor Fig.5 Relation between TSR and for , n= 1,2 & 3
edge with the power following up on the liquid. The
following relation is used to calculate power coefficients at
different TSR on the different number of blades using III. SIMULATION ANALYSIS
different pitch angles. To justify the proposed design, computational fluid
dynamics (CFD) analysis is also conducted in the current
study. Meshing is performed on ANSYS 19.2, with element
Where, size 845.4mm, while the number of elements and nodes are
1055282 and 185556, respectively. The wind turbine design
is accomplished by using given conditions and the K-epsilon
turbulence model.

Pitch Angle 0
0.5 0.411 0.417
0.4
0.3
Cp

0.2 0.116
0.1
0
12.57 6.28 4.19

TSR

Fig.3 Relation between TSR and for , n= 1,2 & 3 Fig.6 Dissection part of meshing

The lift coefficient (Cl) of the wind blade is calculated 0.26

from Ansys, while the value of drag coefficient (Cd) for
airfoil is 0.04.

Seventh International Conference on Aerospace Science & Engineering 203

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

As torque on a wind turbine is calculated (774.24 Nm) from

Ansys, so power will be
P = 774.24*12.27
P = 9499.92W
Flow distributions are analyzed using the whole model as
shown in Fig.11, It can be seen that the blade generate
motion due to the high lift coefficient.

Fig. 7 lift coefficient

Fig.11 Wind distribution on the whole turbine

The pressure distribution is being taken from the front view

and it shows the cut section at the middle of HAWT. It
represents that the tip of the wind blade shows a decrease in
pressure. An airfoil is a 2-dimensional wing segment that
addresses basic wing execution qualities. The pressing factor
circulation and lift coefficient are significant boundaries that
portray the conduct of airfoils. The pressing factor
circulation is straightforwardly identified with the lift
Fig. 8 drag coefficient created via airfoils.

Fig. 9 Lift Force

Fig.12 Front view of pressure distribution on the whole

turbine

Further, the pressure distribution is also analyzed by using a

side view plot distribution to observe the pressure variation
on the blade of the designed model.

Fig. 10 Drag Force

The power can be calculated as;
P = T*w

Seventh International Conference on Aerospace Science & Engineering 204

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

REFERENCES

1. Asif M. Sustainable energy options for Pakistan. Renewable and

Sustainable Energy Reviews, 2009, 13: 903-909.
2. Xu Q, Lin A, Cai Y, Ahmad N, Duan Y, Liu C. Numerical analysis
of aerodynamic characteristics of exhaust passage with
consideration of wet steam effect in a supercritical steam
turbine. Energies, 2020, 13(7):1560.
3. Fawzy H, Zheng Q, Ahmad N, Jiang Y. Optimization of a Swirl
with impingement compound cooling unit for a gas turbine blade
leading edge. Energies, 2020, 13(1):210.
4. Abbas T, Bazmi A. A, Bhuttoo A. W, et al.
Greener energy: Issues and challenges for Pakistan solar energy
prospective. Renewable and Sustainable Energy Reviews, 2014, 31:
258-269.
Fig.13 Pressure distribution on the blade of the designed
5. Ghaffar M. A. The energy supply situation in the rural sector of
model. Pakistan and the potential of renewable energy technologies.
Renewable Energy, 1995, 6(8): 941-976.
It can be observed from Fig. 13 that in the upper red region 6. Mallawi K.Y, Badr M.A. A practical approach for selecting
of the blade the pressure is high, while the lower side of the optimum wind rotors. Renewable Energy, 2003, 28: 803-822.
blades shows the lower pressure so that it generates motion 7. Reiche D, Bechnerger M. Policy differences in the promotion of
in the wind blade. renewable energies in the EU member states. Energy Policy, 2004,
32: 39-52.
IV. CONCLUSION 8. Brondstead P, Holmes J .W, Sorensen B. F.
Wind rotor blade materials technology. European Sustainable Energy
The current research has proposed a design for HAWT to Review, 2008, 2: 36-41.
satisfy the energy needs at the microscopic level in Pakistan.
9. Stol K. A, Zhao W.X, Wright A.D. Individual blade pitch control
All the determined outcomes have been verified utilizing for the controls advanced research turbine (CART). Journal of Solar
computational analysis. The wind turbine design is Energy Engineering, 2006, 128: 498-505.
accomplished by using the specified conditions and the K- 10. Reiche D, Bechnerger M. Policy differences in the promotion of
epsilon turbulence model. From simulation results, it is clear renewable energies in the EU member states . Energy Policy, 2004,
that there exists difference in pressure between the lower 32: 39-52.
and upper end of turbine blade. Hence, these computations 11. Brondstead P, Holmes J W, Sorensen B F. Wind rotor blade
validate the proposed wind turbine design to provide a materials technology. European Sustainable Energy Review, 2008,
2: 36-41.
decentralized, sustainable and reliable energy framework.
12. Stol K A, Zhao W X, Wright A D. Individual blade pitch control for
the controls advanced research turbine . Journal of Solar Energy
ACKNOWLEDGEMENT Engineering, 2006, 128: 498-505.
The authors of this paper acknowledge the support received
from UET Lahore, Pakistan to complete this study.

Seventh International Conference on Aerospace Science & Engineering 205

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Autonomous Unmanned Aerial Vehicle for

Agricultural, Security, Emergency and Other Uses
Dr. Amir Rashid Chaudhry Javid Rashid Ch.
COMSATS UNIVERSITY FAMBZZHH
Islamabad, Pakistan Islamabad, Pakistan Dr. Imran Tahir
[email protected] [email protected] FAMBZZHH
Islamabad, Pakistan
[email protected]
Tayyaba Shafaq Imaan Shahid Raja
COMSATS UNIVERSITY COMSATS UNIVERSITY
Islamabad, Pakistan Islamabad, Pakistan Salman Amir
[email protected] [email protected] COMSATS UNIVERSITY
Islamabad, Pakistan
[email protected]
Zohair Anjum Nabeel Uddin Sheikh
COMSATS UNIVERSITY COMSATS UNIVERSITY
Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected]

Abstract-- This research is the extension of the work [1] been researched upon to offer solutions to multiple problems.
revolves around exploration of possibilities of using technology to To accomplish this, use of existing technology coupled with
solve some major problems, with minimum capital and immediate
results. These problems include, but are not limited to, Security,
Health Emergencies, Agricultural limitations, Cutting down of innovation to create a product with a diverse set of features
Trees and Lack of Mobility in hard-access regions. The primary has been explored. The consideration for the technology was
focus was laid on mobility and on an all-in-one solution. In this
to be modular.
regard, the use of Unmanned Aerial Vehicles (UAVs), also
known as Drones was investigated. It was observed that a Multi- This aspect is primarily what has been missing in the
Purpose Drone, designed and built upon a Modular Platform, previous work on this technology. A lot of research has been
can prove to be an innovative solution to the problems mentioned performed to explore different uses of drones, however,
above. A drone can access areas which might never be possible to minimal thought has been put towards making this
access on foot otherwise, and it can infiltrate hostile regions with technology all-purpose. Many excellent drones have been
no threat of loss of life. Essentially, a drone can be an eye in the developed, but all of them are usually only good for one
sky and prove crucial in many different scenarios. Building upon single purpose. Our research explores the use of one single
this idea further, our team explored the possibilities of this. As
drone for multiple purposes.
discussed above, the platform must be modular. That is, the
aircraft is to have different swappable modules which can be It has to have a fixed basic structure with additional
conveniently swapped-on or swapped-off, depending on the swappable modules which can be swapped on or off
required use on a given instance. depending on the nature and intended purpose of the flight.
As part of the research, a total of four unique modules were
Keywords— UAV, Drone, Autonomous, Aerial, Security, explored: Agricultural Spray module, Seed-firing module,
Surveillance, Patrolling, Facial Recognition, Agriculture, Health and Emergency module and Security and Surveillance
Forestry, Tree Plantation, Reforestation, Seed Plantation,
module.
Emergency Response, Payload, Remote Controlled
The rest of this article is structured as follows: Section 2
offers the literature review of drone technology, its
I. INTRODUCTION
evaluation and architecture. In Section 3, we describe three
Drones as unmanned aerial vehicles which can either modules of the technology. The basic assembly and software
operate anonymously once they have been programmed with is discussed in Section 4 and Section 5 respectively . Section
a certain set of commands, or they can be operated remotely 6 contain the working mechanism Section 7 provides the
by an operator-also known as a drone pilot. Drones come in result and some discussions on future research directions.
different sizes and forms. They can be large enough to carry Finally, Section 8 draws the conclusion
missiles, or small enough to fit inside the palm of a hand.
The upside of using an aerial drone platform, as opposed to II. LITERATURE SURVEY
robots or land drones, is the significantly higher level of The world is moving at a fast pace and it keeps changing.
mobility it comes with. With a drone, the terrain is never a With this change, the technology in use also keeps evolving
limitation. Drones can reach spots which might otherwise rapidly. Products and services are being vastly improved with
never be accessible on foot or on wheels. This ability can be time, a process which is as essential as technology itself. The
crucial in a series of possible events, i.e., when a hiker on a agenda behind this wave of improvement is clear, it is a race
hillside needs emergency first-aid or when a high-security towards efficiency, practicality, and sustainability for the
sensitive area needs to be patrolled [2]. This technology has greater good of the planet. Old ways are being discarded for

Seventh International Conference on Aerospace Science & Engineering 206

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

modern, more efficient, and practical solutions to problems balls out of this opening. The opening has strategically been
surrounding us. Under these circumstances, directing efforts designed to be barely wider than the seedballs. This allows
towards innovation to help make the world a better place is a only one seedball to pop out of the opening at a time. Below
responsibility which befalls every researcher. Our aim, too, is the opening is placed a 3D-printed flap, controlled by a servo
to develop technology which can cater to a wide range of motor. This servo is connected to an Arduino using a motor-
problems with minimum resources required. In this regard, driver IC.
the primary candidate is a UAV (Unmanned Aerial Vehicle),
more popularly known as a Drone. As the servo receives a signal from the Arduino, it rotates
its shaft, which then opens the flap. This sequence is timed to
Because of their adaptability and cost adequacy, drones are only allow one seedball pass into the loading chamber, after
becoming solid partners in battling against natural issues and which the flap closes again. After passing through the flap,
contentions, for example, crop assessment [3] and study the seedball then passes through an open Arduino-controlled
work for the investigation of flood protections. In light of valve. Once the seedball passes, the valve closes. Now, the
these realities, organizations, benefit and non-benefit seedball is in the loading chamber, between two valves. In
establishments, and solitary substances are attempting to this loading chamber is another junction which connects the
foster ventures that can take on the obligations also, benefits chamber to the compressor.
that the UAV bring to the table. Yearningly, this may
altogether affect a change on the debates also, confusions on Both the valves are closed and the compressor builds up
environmental change. In spite of the fact that there is pressure inside the loading chamber. When the pressure
actually just one of numerous most completed chips away at reaches the required value, the second valve is opened
seed-apportioning airplane models, the writing audit on this immediately using the Arduino. This pushes the seedball into
work will zero in on concentrates on that utilization drones the barrel at a high speed, essentially firing it straight into the
on farming fields[4]. Most different ventures center around soil below.
making mechanical developments that might conceivably
diminish or add to the worldwide soundness of forests and Agricultural Spray:
reforestation strategies. AirBoard Agro [5] is an industrial This is the second of the two Agricultural modules, after
agriculture drone built for precision agriculture and Seed Gun. This module has been developed keeping in mind
environmentally responsible farming. This drone has a the rising demand for food with which the agriculture sector
default spray system length of 10 meters. needs to keep up. Methods used in agriculture are often
traditional and usually have models which require a large
As mentioned above, a lot of research has been done with working force. This makes the process, from sowing to
regard to drone technology and exploration of different harvesting, time-consuming and inefficient. This project aims
possible uses of drones. However, the main limitation of this to solve this problem by minimizing the amount of time and
previous work done is that all the researches are limited to labor force required. Using this drone-mounted spray
very few uses of drone technology at a time. The innovative module, area of one acre would be covered in about 10
point of our research is to develop a drone technology which minutes. On multiple batteries, one drone alone would cover
is multi-purpose and can serve to multiple uses hundreds of acres in one day. While at it, the drone may also
simultaneously. The upside of using this technology fly autonomously over an area pre-defined on the map using
platform, as opposed to robots or land drones, is the Mission Planner software. This significantly reduces the
significantly higher level of mobility it comes with. With a human input required.
drone, the terrain is never a limitation. Drones can reach
spots which might otherwise never be accessible on foot or It involves a tank with a 5-liter capacity to hold the
on wheels. This ability can be crucial in a series of possible spraying liquid. This tank has a plastic build which is
events, i.e., when a hiker on a hillside needs emergency first- preferred due to its lightweight. Moreover, the tank has a
aid or when a high-security sensitive area needs to be compact design which can easily fit between the drone’s
patrolled. This drone will be designed to offer solutions to landing gear. This tank is coupled with a 4-nozzle system
multiple problems. To accomplish this, existing technology which helps efficiently spray the liquid onto the surface/crop
will be coupled with innovation to create new technology below. The nozzles should be as far as possible from each
with a diverse set of features. The research is to be modular. other on the landing gear edges, to ensure maximum spray
It would have a fixed basic structure with additional area with minimum spray overlap regions. The nozzle system
swappable modules which can be swapped on or off is controlled with an Arduino board and both these
depending on the nature and intended purpose of the flight. components are powered by the onboard battery.

Security Module:
The Security module caters to the security aspect. In this,
III. METHODOLOGY
the drone would be equipped with a high-resolution camera
Seed Gun: coupled with an FPV (First Person View) kit [6]. This
equipment gives the capability to run facial-recognition
The Seed-Gun module is one of the two agricultural algorithms on to the feed from the camera, in real time. This
modules. As the name suggests. This module hosts a data can be sent back to an operator who can analyze it.
purpose-built gun capable of firing seedballs into the soil. In
this regard, a mechanism was designed specifically for this Emergency/Delivery Module:
purpose. This mechanism consists of a seedball container, an
air compressor, a servo motor, air valves, an Arduino, a 3D- This module has been designed with disaster
printed flap, Motor-driver IC (Integrated Circuit), a loading management in consideration. In this module, a carrying
chamber and a barrel. Seedball container is placed with an mechanism has been designed with an air-drop functionality.
opening at the bottom. The gravitational force forces the seed That is, the drone would be able to carry a payload to a
location and drop it while it is still airborne, without having

Seventh International Conference on Aerospace Science & Engineering 207

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

to land. This is crucial for first-aid missions when the drone

may have to do back-to-back flights without wasting time on
landing or take-off repeatedly. The technology has been
designed with a payload capacity of 10 kilograms, which is
ample for carrying light medical equipment and first-aid kits,
or even blood bags. This whole functionality can additionally
be also utilized in the logistics sector for a speedy delivery of
couriers and mail.

IV. BASE ASSEMBLY

Chassis:
For the chassis, ZD850 frame was explored. This is a hex
copter frame [7] which would allow to mount up to twelve Fig. 1: Drone with arms folded spray tank attached
motors, two on each arm. This would play a significant part
in obtaining a high payload capacity. The frame is made up
of carbon fiber which is a strong yet extremely light material.
This weight-saving is important for a higher flight time. The
frame is foldable, which provides easier mobility when not in
flight.
Propellers:
Propellers play a vital role in the flight and thrust of the
aircraft. For this purpose, twelve high-strength carbon fiber
2255 propellers can be used. These propellers have a length
of 22 inches and pitch of 5.5 inches. These are fully
compatible with multi rotor disc brushless motors. The
positioning of propellers is crucial. Hex copter has alternate
positions of propellers, clockwise and counterclockwise.
Motors:
Each motor has one propeller. 360kv motors would be used Fig. 2: Arms open
which can revolve up to 360 times per minute each. Each
motor produces a thrust of 800 to 900 grams.
V. SOFTWARE
Electronic Speed Controller (ESC):
ESCs allows the flight controllers to control and adjust the Software is a crucial part of this project. It is here where
speed of motors, so the aircraft can move in any direction, at all sorts of calibrations are performed, such as transmitter
variable speeds. controls calibration, ESC (Electronic Speed Control)
calibration, GPS (Geo-Positioning Satellite) calibration and
Flight Controller: compass calibration. In addition to this, flight of the drone
Flight Controller receives the signal from the pilot and also has major dependability on the software. Software
transmits it to the ESCs to control the drone. would be used to determine flight path and to set waypoints
along the path. All of these functions to be configured on the
Radio Receiver: Mission Planner software.[8] For surveillance, a continuous
Radio receiver only receives the signals from the pilot. path would be defined around a perimeter to patrol it. For the
agricultural modules, the aim would be to fly the drone
Battery: within a defined area, with waypoints scattered all over it so
5200mAh Lithium polymer battery was used to ensure ample that maximum area can be covered by spray or seed-gun.
flight time.

Seventh International Conference on Aerospace Science & Engineering 208

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

eliminates the redundancy of the image characteristics.

Facial Expression Extraction identifies the information of
facial expression like motion, shape of basic facial features.
It is considered an interesting part as it extracts useful
information from the image. The information extracted is on
the premises of facial and appearance features.
Facial expression depends on mood and sentiments. Many
facial specifications were extracted from a human face
image and used to train specialized neural networks. The
best performing specialized convolution neural networks
were recruited into decision-making classes which form an
integrated Convolution Neural Network (CNN). The
integrated CNN is then analyzed using data extracted from
the image used during training and testing.
The classification part assembles the information extracted,
gathers it together according to certain parameters and helps
in their efficient identification. It is a composite procedure
as it may be pretentious by many factors. At last, emotions
are recognized by looking at the outcome separated by
various parameters and estimations.
Fig. 3: Waypoint configuration in Mission Planner

Fig 5: Working of Identification of emotions

While the procedure is to some degree complex, face
recognition calculations regularly start by examining human
eyes. Eyes comprise of what is known as a valley area which
Fig. 4: Path for Agricultural needs is one of the least demanding highlights to recognize. When
eyes are identified, calculations are the initiated to recognize
Also, facial recognition is integration is an important part of facial locales including eyebrows, lips, nose, nostrils, and
the software section. the iris. When the calculation construes that it has
distinguished a facial area, it then applies extra tests to
Processing of the image: approve whether it has, truth be told, identified a face.

Figure 6 shows the processing of an image to improve its The main working of facial expression detection [9] will
quality. Although color images can carry more information work as follows.
about facial expression, 2-D Grey images are more popular
in terms of emotion detection. For capturing images, we use  Detect mouth landmarks
a high-definition camera. This stage detects information of  Collecting the data set
interest by smoothing out the image, by removing unwanted
noise and enhances the quality of the image captured. It  Feature extracting
includes normalizing and filtering of the image. It also
 Training the algorithm

Seventh International Conference on Aerospace Science & Engineering 209

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

 Statistic evaluation of algorithms of pre-handling necessary in a CNN is significantly less.

While channels are hand-designed in primitive approaches,
 Building program CNNs can achieve expertise with these channels/attributes
with enough preparation.
The part on which we have focused is face-detection. for
that, we have used Haar-cascade classifiers [10]which are CNN's are frequently utilized in image recognition on
used for object detection, for which it has been trained. databases utilizing CNN and picture classification learning
While detecting the face, we obtain facial landmarks procedure is shockingly quick.
detection which will help us to tell the machine about
expressions. As seen in fig.6. these numbers are landmarks At the point when applied to facial recognition, CNNs
of facial expressions. Landmarks are the areas of the focus accomplished a huge reduction in mistake rate. CNN's were
around fiducial facial parts. Facial form captures the utilized to survey video quality in a target path after manual
inflexible and unbending facial disfigurements because of training of data set.
head developments and outward appearances.
The Haar Cascade classifier uses the Haar Wavelet approach
to examine pixels in a picture and divide them into squares
based on their function. The features observed in any image
are extracted using integral image concepts. Haar Cascades
employs deep learning algorithms that choose several key
features from a broad number of parameters to get an
efficient outcome. Classifiers then employ cascade methods
to recognise faces and expressions in images. Instead of
calculating 2,500 highlights for each window, we considered
cascading. We analyse 2,500 highlights in x different
cascadings. In many cascading directly, we can now tell if
there is a face or not. If I find a face in a photograph during
the cascade, the photograph is passed on to the next course.
In the unlikely event that no face is discovered, we can
proceed to the next window. As a result, the time complexity
is reduced. [11 ].

Fig 6: Facial landmarks detection

The next step is to make a dataset of lots of images, we will
use a data set of images available on the internet and by
using them we’ll generate our data set and then it will sort
the correct expression from those images in the data set,
we’ll use some machine learning algorithms to detect facial
expression. As shown in the following figure (fig 7) we’ve
approximately more than 40,000 pics for the database and
training of the model.

Fig 7: graph of pictures used in the database

Fig. 8: Path for patrolling
Algorithm used:
A Convolutional Neural Network (CNN) is a Deep Learning
algorithm that can take in a large amount of data, assign
VI. WORKING
significance (learnable loads and predispositions) to distinct
perspectives/questions in the data, and then separate them. The mainlworking mechanism of this drone is the operation
When compared to other grouping calculations, the amount of the connected different modules on the drone. All the
modules are controlled by a DC servo motor, connected to a

Seventh International Conference on Aerospace Science & Engineering 210

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

5V power supply. Figure 9 shows the working mechanism

of drone and its components.
At first, the operator should know the exact location of
the target by relying on pictures, videos, or the satellite-
scanned images. After that, the operator defines the area for
the main operation based on the collected information.
Secondly, the operator uses the mission planner software
to generate the waypoints for the drone and defines the
distance between those waypoints as well. The software
saves those waypoints and generates the coordinates as a
rectangular area to assign the path for the drone, and the
mission path is defined as we can see in figure 8.
Lastly, SD card connected to the Arduino stores all the
coordinate by using the Mission Planner software. The
software sends all the data to the flight controller by using
telemetry radios. The Arduino on the ground station is also
connected to the Arduinos placed on the drone by using
XBee’s radios for communicating with each other. DSMX Fig. 9: Working Mechanism
protocol manages all the commands from the ground station.
The ground station uses a telemetry radio system over the
MAVLink to get the information from the flight controller VII. RESULTS
board. This setup allows a strong connection between the
ground station and the drone with all its modules. The This technology can be deployed for several uses.
ground station can be used to control and track the
movement and position of the drone. The connection I 1-Law Enforcement Agencies:
seamless with minimum delay, so there is no disruption or
risk. The ground controller can control the flight of the LEAs can utilize such a drone to perform Surveillance
drone as well as its connected modules. This includes and Patrolling operations with its built-in facial recognition
remote operation of seed dispensing module, agricultural and path-following functionality. Deployment can also be
spray module, surveillance module and delivery module. done on the battlefield to assist troops, to watch out for any
Additionally, the connection between the radio transmitter potential threats. This can significantly lower the risk faced
and the drone is based on the DSMX protocol which is a by the personnel on the ground.
very reliable connection. This allows the drone to fly upto a
5 kilometer range of the transmitter. This range is crucial for
operation in far-off and hard-access areas where the operator
is forced to maintain a distance from the drone due to 2- Agriculture/Reforestation:
unfavorable conditions of safety or terrain.
The protocol used for the connection between the ground Such a drone will also be equipped with a seed-planting
controller and the drone is the MAVLink protocol. It module which can effectively shoot seed-balls in the soil to
connects a computer with the drone using a telemetry maximize the output of agricultural land and to help in
module which is based on two hardware parts. One part is planting trees. Furthermore, a spraying kit can also be added
connected to the laptop, while other connected directly to to enable crop-spray.
the drone. On the first switch on, both these parts are kept in
close proximity. This enables them the kick off an automatic
pairing exercise. After a few seconds, both the parts of the 3-Health Emergency:
telemetry module are paired and start to blink in sync. This
telemetry module then provides a stable connection between The third module is to allow rapid transport of medical
the computer and the drone and helps transfer location and aid, such as first-aid kits, medicines and blood bags to areas
position data, such as gps and compass co-ordinates to which otherwise may have a limited access on-foot or on-
determine the position and altitude of the drone. wheels.

4- Delivery of Goods:

In addition to medical equipment, the drone may also be

installed with a heavy payload capacity which will enable it
to carry heavy loads over long distances. This can be
utilized by courier and food delivery services to significantly
speed up their networks.

VIII. CONCLUSION
We strongly believe that this research work will help the
humanity in an efficient by using extremely state of the art

Seventh International Conference on Aerospace Science & Engineering 211

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

technology, moreover, its usability and utility in many conference on advanced computing and communication systems
(ICACCS) (pp. 215-220). IEEE.
different areas can mark the difference by comparing it with
old fashioned methods already available. The core emphasis [3] Na, S. I., Park, C. W., So, K. H., Ahn, H. Y., & Lee, K. D. (2019).
of the research is on surveillance and Patrolling operations Photochemical Reflectance Index (PRI) mapping using drone-based
with developed facial recognition and path-following hyperspectral image for evaluation of crop stress and its application to
functionality along with seed-planting module which can multispectral Imagery. Korean journal of remote sensing, 35(5_1),
637-647.
effectively shoot seed-balls in the soil to maximize the
[4] Hovhannisyan, T., Efendyan, P., & Vardanyan, M. (2018). Creation
output of agricultural land and to help in planting trees and of a digital model of fields with application of DJI phantom 3 drone
also spraying kit also added to enable crop-spray. The and the opportunities of its utilization in agriculture. Annals of
developed drone can do rapid transport of medical aid, such agrarian science, 16(2), 177-180.
as first-aid kits, medicines and blood bags to areas which
otherwise may have a limited access on-foot or on-wheels, [5] AirBoard (2019, October 15). Agricultural Drone.
https://www.airboard.co/agro/.
furthermore, it’s designed in a way that enables to carry
[6] Smolyanskiy, N., & Gonzalez-Franco, M. (2017). Stereoscopic first
heavy payload capacity over long distances. This can be person view system for drone navigation. Frontiers in Robotics and
utilized by courier and food delivery services to significantly AI, 11.
speed up their networks. This is triumph proof that with the [7] Arief, U. M., Subiyanto, S., Andrasto, T., Sukamta, S., Sulistyawan,
rising global population and the ever-increasing congestion, V. N., Sarwono, E., ... & Putra, A. D. H. (2021, March). Design of
hexacopter UAV system for disinfectant spraying. In IOP Conference
going airborne is the future and a significant step towards Series: Earth and Environmental Science (Vol. 700, No. 1, p.
prosperous and peaceful world in future. 012023). IOP Publishing.
[8] https://ardupilot.org/planner/

[9] Zeng, Z., Fu, Y., Roisman, G. I., Wen, Z., Hu, Y., & Huang, T. S.
(2006). Spontaneous emotional facial expression detection. J.
REFERENCES Multim., 1(5), 1-8.
[1] Chaudhary, A. R., Ch, J. R., Tahir, I., Shafaq, T., & Rajja, I. S. (2021,
March). UNMANNED AUTONOMOUS ABABEEL SYSTEM FOR [10] Cuimei, L., Zhiliang, Q., Nan, J., & Jianhua, W. (2017, October).
SEED PLANTATION UTILIZING SPECTRAL CAMERA Human face detection algorithm via Haar cascade classifier combined
ANALYSIS USING DRONE TECHNOLOGY. In BOOK OF with three additional classifiers. In 2017 13th IEEE International
ABSTRACTS (p. 319).Erico Pinheiro Fortes (2017). Seed Plant Conference on Electronic Measurement & Instruments (ICEMI) (pp.
Drone for Reforestation. The Graduate Review Article 7, Volume 483-487). IEEE.
2:13-26. [11] Clarke, Roger. The Regulation of Civilian Drones’ Impacts on
Behavioural Privacy. 30(3) COMPUTER LAW & SECURITY REVIEW
[2] Sanjana, P., & Prathilothamai, M. (2020, March). Drone design for 286–305. June 2014. https://lccn.loc.gov/2009217103
first aid kit delivery in emergency situation. In 2020 6th international

Seventh International Conference on Aerospace Science & Engineering 212

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Simulations For Boundary Layer Flows And Heat

Transfer Of Cross Fluid Over A Moving Flat Plate
Mazhar Shehzad Ammar Mushtaq
Research Centre For Modelling and Simulation Research Centre For Modelling and Simulation
National University of Sciences and Technology National University of Sciences and Technology
Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected]

Abstract—During the past century, rheology has emerged as were carried out for Newtonian fluids. We need non-
a new science which deals with the deformation of matter. Metal Newtonian fluids Because of its broad spectrum of industries
spinning, wire drawing, polymer extrusion, blood circulation, applications. Various studies have been carried out on
food industry, and pressure-sensitive adhesion, etc are a few of momentum and heat transfer and of non-Newtonian fluids [7-
the potential application areas. Now in most of these applications 8]. Today many studies are being carried out for many other
materials are non-Newtonian and involve rotation, extrusion, physical investigations of micropolar fluids, power law,
and heat exchange. Therefore, understanding the rheology is Maxwell fluids, second-grade fluids, etc. So the complex
critically important to improve the quality of product fluids particles like suspensions, turbulent shear flows, can be
development, methodology, and resource utilization. The goal of
described using the theory of micropolar fluid flows. [9-14]
this research is to present the boundary layer equations for three
dimensional fluid flow and heat transfer of cross fluid over a
worked on micropolar and power-law fluids. The power-law
shretching sheets. Systems of governing partial differential model refers to non-Newtonian fluids which are time-
equations are converted into highly non-linear ordinary independent. Power-law model is sufficient for many non-
differential equations by introducing suitable similarity Newtonian fluids to explore the shear stress and shear rate
transformations. By using the bvp4c process, the governing measurements. In spite, the power-law model has a strong
ODEs are solved numerically, and the influence of the related engineering influence, but it cannot explore the normal stress
parameters of practical importance such as skin friction differences. In this regard, we will consider a special type of
coefficient and Nusselt number are calculated. The momentum differential equations of fluids for which the analytical
boundary layer demonstrates the elevation impact of the solutions are available. These differential fluids are the class
growing local Weissenberg number. The contrary phenomenon of second-grade fluids [15-17]. This model refers to
for the thermal boundary layer was found. The temperature determining the normal stress results. A significant amount of
function has an exceptional S-shaped profile indicating the research on the physical analysis of the laminar boundary layer
existence of an adiabatic case for the large enough wall to flows across solid surfaces has been published [18-25]. When
ambient temperature ratio. Structures of the momentum a moving surface is subjected to suction or blowing. As
boundary layer along with Velocity fields demonstrated the polymer is a flexible substance, the geometry of the filament
identical enhancement tendency for the rising Weissenberg surface may stretch during ejection, causing the surface
number. velocity to disturb from being uniform.. The thermal radiation
over a continually stretching surface has taken on excellent
Keywords—non-Newtonian flows, Cross fluid, Numerical
Simulations
significance Due to its use in a wide range of manufacturing
and scientific uses, such as the manufacture of plastics,
I. INTRODUCTION metallurgy, crystal processing, and many others. Plastic sheet
drawing. The performance of the final product relies on
An American chemistry professor first coined the term cooling and stretching rates. The boundary layer flows over a
“Rheology” which he inspired from the Greek word continuously moving surface First mentioned by Sakiadis [4].
“πανταρει” meaning “everything flows”. During the past Rees and Bassom [26] predicted the flow of a micro-polar
century, rheology has emerged as a new science which deals fluid flow over a flat surface(Blasius boundary layer) while
with the deformation of the matter. The rheological studies are
Soundalgekar and Takhar proposed a similarity evaluation of
focused on dealing with applied stress and strain. As all
flow and heat transfer over a continuously flowing semi-
materials have rheological properties thus rheology is relevant
infinite plane of micro-polar fluid [27]. The above section
in many fields. Examples include metal spinning, wire
shows that although a huge amount of research has been
drawing, concrete technology, plastic processing, polymer
carried out on the boundary layer flows for Newtonian and
extrusion, paint flows, blood circulation, cosmetics, food
non-Newtonian fluids. However, to investigate the shear-
industry, and pressure-sensitive adhesion, etc. Now
thinning/thickening region most popular non-Newtonian
understanding rheology is critically important in order to
Model used is the Power Law model. Despite its ability to
improve the quality of product development, methodology,
present a vast class of non-Newtonian fluids, it has its
and resource utilization. Particularly to the food industry
limitations in presenting fluids with severe high and low shear
where most materials are Non-Newtonian, Rotating flows in
rate. In 1965 Cross [28] introduce a model that caters to a
mixtures, extruders, and heat exchangers are widely common
wider sub-section of generalized Newtonian fluids which can
[1]. Blasius [2], pioneered in studying the motion of flow on a
detect both the flow rate in the power-law region as well server
stationary flat plate. Then later, Leslie Howarth proposed the
high and low shear rate range. Now, with the available
theory of turbulence in aerodynamics as well as in fluid
computing resources and advancement in numerical
dynamics [3]. The influence of the boundary layer over a
techniques to tackle highly nonlinear equations such as arising
moving plate was then scrutinized by Sakiadis [4-5]. By
in modelling of Cross fluid, researchers have started working
extending the work of Sakiadis, Crane [6] was able to work
on problems with many folds complexity. Khan et al [29] in
for the problem of stretching flat plate and found the exact
2016 worked on momentum and the heat transfer of the cross
solution of boundary layer equations. But all these studies

Seventh International Conference on Aerospace Science & Engineering 213

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

fluid over the stretch surface. In 2017 Khan et al [31] reported Putting (8) in (1) and (2) while keeping in mind (7) and Eq
on cross-axisymmetric flow and heat transfer across a radially
stretched surface. In 2019, Sultan et al [32] explored (9) a straightforward calculation gets the following form.
theoretical elements of thermophoresis and Brownian motion 𝜕𝑢 𝜕𝑣
with the activation energy for the 3-D flow of the cross fluid. + =0 (10)
𝜕𝑥 𝜕𝑦
For the validations purpose the results of the power-law index
against temperature profile were compared with Kumari
nath[33]. Several researchers have carried out experimental
𝜕𝑢
studies on the Cross model over the previous two centuries. 𝜕𝑢 𝜕𝑢 1 𝜕𝑝 𝜕 𝜕𝑦
Escudier et al. carried out an experimental study and the 𝑢 +𝑣 =− +𝜈 (11)
results on the fluid flow were presented by applying the Cross 𝜕𝑥 𝜕𝑦 𝜌 𝜕𝑥 𝜕𝑦 𝜕𝑢 1−𝑛
1 + {Γ ( )}
model to the generalized Newtonian fluid. Xie and Jin studied [ 𝜕𝑦 ]
the Cross Rheology equation to get the free surface flow of
non-Newtonian fluids. An experimental method, namely the
WC-MPS technique, was used to determine the Cross model's III. PROBLEM FORMULATION
four rheology parameters for the numerical application of the
Consider a two-dimensional incompressible flow and heat
Cross equation Therefore, still, a lot of potential fluid
transfer of cross fluid over a moving flat surface. Plate (located
scenarios can be modelled and solved for an in-depth
at y=0) moving outside the boundary layer with the constant
understanding of flow and heat transfer of the Cross fluid
velocity 𝑈𝑤 and 𝑈∞ be the fluid velocity. A study on the
model.
analysis of heat transfer is carried out due to nonlinear heat
II. GOVERNING EQUATIONS flucuations. Also the combined effects of Joule heating and
viscous dissipation were considered. The plate is at constant
The conservation equations of mass, linear momentum, temperature 𝑇𝑤 , while 𝑇∞ implicitly denotes the temperature
and energy for the flow of an incompressible fluid of the ambient liquid.
∇⋅ 𝑉 =0 (1) 𝜕𝑢 𝜕𝑣
+ =0 (12)
𝜕𝑉 (2) 𝜕𝑥 𝜕𝑦
𝜌 = ∇ 𝜏
𝜕𝑡
𝜕𝑢
𝑑𝑇 (3) 𝜕𝑢 𝜕𝑢 𝜕 𝜕𝑦
𝜌𝑐𝑝 = 𝜏. 𝐿 − ∇𝑞 𝑢 +𝑣 =𝜈 , 𝑈𝑤 < 𝑈∞ (13a)
𝑑𝑡 𝜕𝑥 𝜕𝑦 𝜕𝑦 ∂u 1−𝑛
Where V is the velocity vector, 𝜏 is the Cauchy stress 1 + {Γ ( )}
[ ∂y ]
tensor, 𝑐𝑝 is the specific heat constant, 𝜌 is the density, 𝑞 is the
𝑑
heat flux, L and are the material derivatives, T is the fluid 𝜕𝑢
𝑑𝑡
𝜕𝑢 𝜕𝑢 𝜕 −
temperature. The Cauchy stress tensor for the fluid is defined 𝜕𝑦
𝑢 +𝑣 = −𝜈 , 𝑈𝑤 > 𝑈∞ (13b)
as, 𝜕𝑥 𝜕𝑦 𝜕𝑦 ∂u 1−𝑛
1 + {Γ (− )}
𝜏 = −𝑃𝐼 + 𝜇 ∗ 𝐴1 (4) [ ∂y ]
∗
Where 𝜇 is the cross model, 𝑃 is the pressure, 𝐼 is the 𝜕𝑇 𝜕𝑇 𝜕2𝑇 1 𝜕𝑞𝛾
𝑢 +𝑣 =𝛼 2− ( ) (14)
identity tensor, 𝐴1 is the first Rivlin-Ericksen tensor. 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜌𝐶𝑝 𝜕𝑦

𝐴1 = 𝐿 + 𝐿𝑇 , 𝐿 = ∇𝑉 (5) Due to parallel free stream and nonlinear thermal radiation

the following boundary conditions are produced.
𝑢 = 𝑈𝑤 , 𝑣 = 0, 𝑇 = 𝑇𝑤 at 𝑦 = 0
1
(6) (15)
𝛾̇ = √ 𝑡𝑟(𝐴1 )2 𝑢 → 𝑈∞ , 𝑇 → 𝑇∞ as 𝑦 → ∞
2
where 𝑥, 𝑦 are the coordinates along and normal to the
Due to this consideration, (4) will get the following form.
1 plate. 𝑢 , 𝑣 are the velocity components along the 𝑥 and
𝜏 = −𝑃𝐼 + 𝜇0 [ ]𝐴 (7)
1 + (Γ𝛾̇ )1−𝑛 1 𝑦 directions, 𝜈 is the kinematic viscosity, 𝑇 is the fluid
temperature, 𝛼 is the thermal diffusivity, 𝐶𝑝 is the specific
Through Cartesian coordinates for two-dimensional flow, we heat at constant pressure and 𝑞𝑟 is the radiative heat flux.
consider the velocity field of the form The radiative heat flux 𝑞𝑟 is given by the Rosseland
𝑉 = [𝑢(𝑥, 𝑦), 𝑣(𝑥, 𝑦)] (8) approximation as :
Keeping in view the shear rate (6) can be written as 4𝜎 ∗ 𝜕𝑇 4
𝑞𝑟 = − (16)
1 3𝑘 ∗ 𝜕𝑦
𝜕𝑢 2 𝜕𝑢 𝜕𝑣 2 2
(9)
𝛾̇ = [4 ( ) +( + ) ] This is further simplified
𝜕𝑥 𝜕𝑦 𝜕𝑥

Seventh International Conference on Aerospace Science & Engineering 214

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

16𝜎 ∗ 3 𝜕𝑇 In the illumination of the Eq. (19) we obtain dimensional

𝑞𝑟 = − 𝑇 (17)
3𝑘 ∗ 𝜕𝑦 forms of the local skin friction coefficient and the local
So (14) will get the final form Nusselt number as:
𝜕𝑇 𝜕𝑇 𝜕 16𝛼 ∗ 𝑇 3 𝜕𝑇 1 1 𝑓 ′′ (0)
𝑢 +𝑣 = [(𝛼 + ) ] (18) 𝑅𝑒 2 𝐶𝑓𝑥 =
𝜕𝑥 𝜕𝑦 𝜕𝑦 3𝜌𝐶𝑝 𝑘 ∗ 𝜕𝑦 2 1−𝑛
1 + (𝑊𝑒𝑓 ′′ (0))
Where σ* denotes Stefan-Boltzman constant and and 𝑘* 1
denotes the mean absorption coefficient. −𝑅𝑒 −2 𝑁𝑢𝑥 = (25)
Now introduce the corresponding similarity transformation [1 + 𝑅𝑑{1 + (𝜃𝑤 − 1)𝜃} 3 ]𝜃′(0)

for the Eq (13-15).

IV. NUMERICAL SOLUTION
𝑈 In order to solve the higher-order differential problems
𝜂 = 𝑦√ , 𝑢 = 𝑈𝑓 ′ (𝜂),
2𝜈𝑥
numerically. They are converted into a system of 1st order
(19)
𝑈𝜈 𝑇 − 𝑇∞ equations using new variables. Therefore, let 𝑦1 = 𝑓; 𝑦2 =
𝑣 = √ (𝜂𝑓 ′ − 𝑓), 𝜃(𝜂) =
2𝑥 𝑇𝑤 − 𝑇∞ 𝑓 ′ ; 𝑦3 = 𝑓 ′′ ; 𝑦4 = 𝜃 and 𝑦5 = 𝜃′. The equivalent first-order

Where 𝑓(𝜂) represent the dimensionless stream function and equations are as follows:

𝜃(𝜂) represent the non-dimensional temperature. The

𝑦1′ = 𝑦2 ; 𝑦2′ = 𝑦3 ; (26)
required continuity (12) is satisfied while (13) and (14) with
boundary conditions (15) are transformed into the following
−𝑦1 𝑦3 [1 + (𝑊𝑒𝑦3 )𝑛 ]2
form. 𝑦3′ = ; (27)
[1 + sign(𝛾 − 0.5)(1 − 𝑛)(𝑊𝑒𝑦3 )𝑛 ]
𝑓 ′′′ [1 + 𝑠𝑖𝑔𝑛(𝛾 − 0.5)(1 − 𝑛)(𝑊𝑒𝑓 ′′ )𝑛 ]
(20)
+ 𝑓𝑓 ′′ [1 + (𝑊𝑒𝑓 ′′ )𝑛 ]2 = 0 𝑦4′ = 𝑦5 ; (28)
𝜃′′[1 + 𝑅𝑑(1 + (𝜃𝑤 − 1)𝜃)3 ]
+ 3𝑅𝑑[(+(𝜃𝑤 − 1)𝜃)2 (𝜃𝑤 (21) −𝑃𝑟𝑦1 𝑦5 − 3𝑅𝑑[(1 + (𝜃𝑤 − 1)𝑦4 )2 (𝜃𝑤 − 1)𝑦52 ]
2 𝑦5′ = (29)
− 1)𝜃 ′ ] + 𝑃𝑟𝑓𝜃′ = 0 [1 + 𝑅𝑑(1 + (𝜃𝑤 − 1)𝑦4 )3 ]
𝑓(0) = 0, 𝑓 ′ (0) = 1 − 𝛾, 𝜃(0) = 1,
(22) with boundary conditions
𝑓 ′ (∞) → 𝛾, 𝜃(∞) → 0
𝑈∞
𝑦1 (0) = 0; 𝑦2 (0) = 1 − 𝛾; 𝑦4 (0) = 1;
Where 𝛾 = is the velocity ratio parameter, 𝑊𝑒 = (30)
𝑈 𝑦2 (∞) = 𝛾; 𝑦4 (∞) = 0
1
𝑈Γ𝑅𝑒 2 𝑇𝑤
is the Weissenberg number, 𝜃𝑤 = is the
𝑥√2 𝑇∞

temperature ratio parameter, 𝑃𝑟 = 𝛼𝜈 is the Prandtl number The solution of Eq. (26-30) is obtained using the fifth order
3
16𝛼𝑇∞
and 𝑅𝑑 = are the radiation parameter. The expression Runge-Kutta integration technique with the following initial
3𝐾𝐾 ∗
conditions
provided for the local skin friction coefficient and the local
𝑦1 (0) = 0; 𝑦2 (0) = 1 − 𝛾; 𝑦3 (0) = 𝑢1
number Nusselt are: (31)
𝜏𝑤 𝑥𝑞𝑤 𝑦4 (0) = 1; 𝑦5 (0) = 𝑢2
𝐶𝑓 = , 𝑁𝑢 = (23)
1 2 𝑘(𝑇𝑤 − 𝑇∞ ) This solution can agree with the original BVP for suitable
𝜌𝑈
2
values of 𝑢1 and 𝑢2 .
Where 𝜏𝑤 is local wall shear stress and 𝑞𝑤 is described as
surface heat flux. V. RESULTS AND DISCUSSION
𝜕𝑢 Using MATLAB code the boundary value problem is
𝜕𝑦 solved. To make sure that the accuracy of obtained
𝜏𝑤 = 𝜏𝑥𝑦| = 𝜂0 , computations and validation of the code we have reproduced
𝑦=0 𝜕𝑢 1−𝑛
1 + {Γ ( )} (24) the results of [18]. After code validation now we will examine
[ 𝜕𝑦 ]𝑦=𝑜
the impact of various evolving parameters on the velocity and
𝜕𝑇 temperature flow field. The obtained results are tabulated in
𝑞𝑤 = −𝑘 | Tables 1-3. Table 2 is designed to provide a skin friction
𝜕𝑦 𝑦=0

Seventh International Conference on Aerospace Science & Engineering 215

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

coefficient and Table 3 provides heat transfer rate numerical 0 0.6 0.8 - 0.443939
values for various values of relevant parameters.
0.4 - 0.106581
Fig 3 indicates variation in horizontal velocity with an
improvement in the velocity ratio γ. It is observed that the
1 0.625574
velocity function 𝑓 ' rises and that the thickness of the 0.2 0.1 0.2 - 0.87292
boundary layer declines when 0 < γ< 0.5 increases. There is 1.2 - 0.357323
no boundary layer formation at γ=0.5. In this region, the free
stream velocity is less than the velocity of the plate. The 1 0 0.664115
impact of the power-law index '𝑛' on the velocity profile and 1 0.496975
fluid temperature profile is shown in Fig.1 and Fig.2 for two
different values of γ=0.2 and γ =1. The examination of these 0.3 0.8 0.5 - 0.213935
graphs demonstrates that the velocity profiles and momentum 1.5 - 0.184016
boundary layer thickness exhibit a progressive trend for both (
γ<0.2 and γ>1) with increasing power law of index. The 2 - 0.171461
effect of the power-law index 'n' on the temperature profile is
observed in Fig 4. In Fig.4 the thermal boundary layer reveals Table 3:: Numerical findings for Local Nusselt number against different
a declining tendency while rising the power law of the index. values of power law index n, Weissenberg number We, velocity ratio γ,
The physical explanation for this behaviour is that the fluid is Temperature ratio θw, Prandtl number Pr, Rd Radiation parameter.
facing less resistance because of low viscosity which causes
the fluid's velocity to increase and its temperature to fall. 𝑷𝒓 𝜸 𝒏 𝑾𝒆 𝜽𝒘 𝑹𝒅 𝜽′ (𝟎)
Fig 5 and 6 show the behaviour of the velocity profiles and 6.2 0.3 0.2 0.8 1.5 1 - 0.610192
temperature profile corresponding to a change in
Weissenberg's local number. The study of the figures(5-7) 0.8 - 0.644514
reveals that 𝑊𝑒 trigger the velocity of the fluid to increase and
decaying in the temperature of the fluid. Physically, an 1.2 - 0.640061
enhancement in the Weissenberg number value We causes the
velocity of the fluid to boost and the temperature of the fluid
0 0.8 - 0.738588
to decrease. 0.7 -0.545214
Fig 7 Displays the effect of Prandtl number Pr on the
1 -0.485732
temperature profile for fixed 𝑛 and 𝑊𝑒 values. Prandtl
number is the resultant of viscosity, Specific heat, and thermal 0.3 0.5 - 0.641106
conductivity. Furthermore, 𝑃𝑟 regulates the relative velocity
and thermal boundary layers thickness. Keeping in view all the 1.5 - 0.647132
above aspects that smaller Prandtl liquids lead to speedier heat
diffusion in thicker thermal boundaries as compared to higher 2.5 - 0.651992
Prandtl liquids in thinner boundaries. Through the increase of 0.7 0.6 0.8 - 0.191498
𝑃𝑟 the thermal boundary layer thickness declines and the
thermal transfer rate at the plate is increased. Perhaps it can be 2 - 0.347346
assured that the Prandtl number can be used to upgrade the
cooling rate in the conductive liquid flow. The influence of 7 - 0.690414
temperature ratio 𝜃𝑤 on temperature is shown in Fig.8. The
higher temperature ratio parameter 𝜃𝑤 indicates a higher plate 1 - 1.37613
temperature compared to ambient temperature. Due to which 1.6 - 0.842045
the temperature of the fluid rises because of the hotter surface.
Fig 9 indicates that, regardless of the range of values chosen 2 - 0.603513
for the velocity ratio γ the temperature 𝜃 rises with a growing
trend in γ.

Table 1:Numerical validation for Newtonian Fluid Case

γ Cortell [18] Present

0 −0.627547 −0.627555
1 0.469602 0.469600

Table 2:Numerical findings Local skin friction for various values of velocity
ratio γ, power law index n and Weissenberg number We.

𝜸 𝒏 𝑾𝒆 𝒇′′(𝟎)

Seventh International Conference on Aerospace Science & Engineering 216

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig 1: Velocity curve 𝑓 ′ Variation against 𝜂 for Fig 4: Temperature curve 𝜃 Variation against 𝜂 for different values of
different values of 𝑛 for Skiadas 𝛾 = 0 𝑛

Fig 2: Velocity curve 𝑓 ′ Variation against 𝜂 for different :values of 𝑛 Fig 5: Velocity curve 𝑓 ′ Variation against 𝜂 for different values of 𝑊𝑒.
for Blasius 𝛾 = 1.

Fig 3: Velocity curve 𝑓 ′ Variation against 𝜂 for different values of γ Fig 6: Temperature curve 𝜃 Variation against 𝜂 for different values of
𝑊𝑒.

Seventh International Conference on Aerospace Science & Engineering 217

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

 The increasing power-law index values culminated

in the velocity of fluid elevation while a decrease in
the thermal boundary layer was observed.
 The velocity profiles reach the velocity of the free
stream at a great distance from the plate,
demonstrating that the momentum boundary layer
thickens as the velocity ratio increases.The heat
transmission rate and the temperature of the plate
increased.
 The momentum boundary layer and Velocity fields
are demonstrated the elevation impact of the
growing local Weissenberg number. The contrary
phenomenon for the thermal boundary layer was
found.
 The Prandtl number's elevated values reduced the
temperature profile and the thermal boundary layer
thickness.
Fig 7: Temperature curve θ Variation against η for different values of
Pr.  The temperature rises in the region and from the
plate the heat flux reduces as 𝜃𝑤 is increasing.
ACKNOWLEDGMENT
I would first like to thank my thesis advisor Dr. Ammar
Mushtaq (RCMS, NUST). The expertise was invaluable in the
formulating of the research topic and methodology.
REFERENCES
[1] Steffe, Steffe, J.F. Rheological Methods in Food Process Engineering,
Freeman Press (1992)
[2] Blasius, H. Grenzschichten in Flüssigkeiten mit kleiner Reibung. Zeits.
Math. Phys, 56, 1-37 (1908).
[3] Howarth, L. On the Solution of the Laminar Boundary Layer
Equations. Proceedings of the Royal Society of London. Series A,
Mathematical and Physical Sciences 164 547–579 (1937).
[4] Sakiadis, B. C. Boundary ‐ layer behavior on continuous solid
surfaces: I. Boundary ‐ layer equations for two ‐ dimensional and
axisymmetric flow. AIChE J. 7, 26–28 (1961).
Fig 8: Temperature curve θ Variation against η for different values of [5] Sakiadis, B. C. Boundary ‐ layer behavior on continuous solid
𝜃𝑤. surfaces: II. The boundary layer on a continuous flat surface. AIChE J.
7, 221–225 (1961).
[6] Crane, L. J. Flow past a stretching plate. Zeitschrift für Angew. Math.
und Phys. ZAMP 21, 645–647 (1970).
[7] Acrivos, A., Shah, M. J. & Petersen, E. E. Momentum and heat transfer
in laminar boundary ‐ layer flows of non ‐ Newtonian fluids past
external surfaces. AIChE J. 6, 312–317 (1960).
[8] Tashtoush, B., Kodah, Z. & Al-Ghasem, A. Heat transfer analysis of a
non-Newtonian fluid on a power-law stretched surface with suction or
injection for uniform and cooled surface temperature. Int. J. Numer.
Methods Heat Fluid Flow 10, 385–396 (2000).
[9] Schowalter, W. R. The application of boundary ‐ layer theory to
power‐law pseudoplastic fluids: Similar solutions. AIChE J. 6, 24–28
(1960).
[10] Eringen, A. Theory of Micropolar Fluids. Indiana Univ. Math. J. 16, 1–
18 (1966).
[11] Pakdemirli, M. Boundary layer flow of power-law fluids past arbitrary
profiles. IMA J. Appl. Math. 50, 133–148 (1993).
[12] Hassanien, I. A., Abdullah, A. A. & Gorla, R. S. R. Flow and heat
transfer in a power-law fluid over a nonisothermal stretching sheet.
Fig 9: Temperature curve θ Variation against η for different values of γ Math. Comput. Model. 28, 105–116 (1998).
[13] Jalil, M., Asghar, S. & Mushtaq, M. Analytical solutions of the
VI. CONCLUDING REMARKS boundary layer flow of power-law fluid over a power-law stretching
This study is seminal work in the development of cross fluid surface. Commun. Nonlinear Sci. Numer. Simul. 18, 1143–1150
(2013).
flow boundary layer equations. Modeled partial differential
[14] Andersson, H. I., Bech, K. H. & Dandapat, B. S.
equations of cross fluid were Simplified by implementing Magnetohydrodynamic flow of a power-law fluid over a stretching
appropriate similarity transformations and numerical sheet. Int. J. Non. Linear. Mech. 27, 929–936 (1992).
technique using the bvp4c scheme. The key conclusions can [15] Dunn, J. E. & Fosdick, R. L. Thermodynamics, stability, and
be summarized as boundedness of fluids of complexity 2 and fluids of second grade. Arch.
Ration. Mech. Anal. 56, 191–252 (1974).

Seventh International Conference on Aerospace Science & Engineering 218

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
[16] Dunn, J. E. & Rajagopal, K. R. Fluids of differential type: Critical [26] Rees, D. A. S. & Bassom, A. P. The Blasius boundary layer flow of a
review and thermodynamic analysis. Int. J. Eng. Sci. 33, 689–729 micropolar fluid. Int. J. Eng. Sci. 34, 113–124 (1996).
(1995). [27] Soundalgekar, V. M. & Takhar, H. S. Flow of micropolar fluid past a
[17] Hayat, T. & Khan, M. Homotopy Solutions for a Generalized Second- continuously moving plate. Int. J. Eng. Sci. 21, 961–965 (1983).
Grade Fluid Past a Porous Plate. Nonlinear Dynamics. 42 (2005). [28] Cross, M. M. Rheology of non-Newtonian fluids: A new flow equation
[18] Cortell, R. Flow and heat transfer in a moving fluid over a moving flat for pseudoplastic systems. J. Colloid Sci. 20, 417–437 (1965). [5
surface. Theor. Comput. Fluid Dyn. 21, 435–446 (2007). [29] Khan, M., Manzur, M. & Rahman, M. Boundary-layer flow and heat
[19] Gupta, P. S. & Gupta, A. S. Heat and mass transfer on a stretching sheet transfer of cross fluid over a stretching sheet. Therm. Sci. 23, 307–318
with suction or blowing. The Canadian Journal of Chemical (2019).
Engineering. 55 744–746 (1977). [30] Hayat, T., Khan, M. I., Tamoor, M., Waqas, M. & Alsaedi, A.
[20] Salleh, M. Z., Mohd Rohni, A. & Amin, N. Boundary Layer Flow Due Numerical simulation of heat transfer in MHD stagnation point flow of
to a Moving Flat Plate in Micropolar Fluid. J. Teknol. 43, (2005). Cross fluid model towards a stretched surface. Results Phys. 7, 1824–
[21] Siekman, J. The laminar boundary layer along a flat plate. Z. Flugwiss. 1827 (2017).
10, 278–281 (1962). [31] Khan, M., Manzur, M. & ur Rahman, M. On axisymmetric flow and
[22] Tsou, F. K., Sparrow, E. M. & Goldstein, R. J. Flow and heat transfer heat transfer of Cross fluid over a radially stretching sheet. Results
in the boundary layer on a continuous moving surface. Int. J. Heat Mass Phys. 7, 3767–3772 (2017).
Transf. 10, 219–235 (1967). [32] Sultan, F., Khan, W. A., Ali, M., Shahzad, M., Irfan, M., & Khan, M.
[23] Hayat, T., Iqbal, Z., Qasim, M. & Obaidat, S. Steady flow of an Eyring Theoretical aspects of thermophoresis and Brownian motion for three-
Powell fluid over a moving surface with convective boundary dimensional flow of the cross fluid with activation energy. Pramana -
conditions. Int. J. Heat Mass Transf. 55, 1817–1822 (2012). J. Phys. 92, 1–10 (2019).
[24] Mustafa, M., Hayat, T., Pop, I. & Aziz, A. Unsteady boundary layer [33] M. Kumari and G. Nath, Bangalore, India MHD boundary-layer flow
flow of a Casson fluid due to an impulsively started moving flat plate. of a non-Newtonian fluid over a continuously moving surface with a
Heat Transf. - Asian Res. 40, 563–576 (2011). parallel free stream. Acta Mechanica 146, 139 150 (2001)
[25] Ellahi, R., Alamri, S. Z., Basit, A. & Majeed, A. Effects of MHD and
slip on heat transfer boundary layer flow over a moving plate based on
specific entropy generation. J. Taibah Univ. Sci. 12, 476–482 (2018).

Seventh International Conference on Aerospace Science & Engineering 219

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Design, Analysis and Fabrication of Composite

Curing Compressive Mold for Aerospace
Applications
Muhammad Ayaz Ahmad Sabih Ahmad Khan Syed Irtiza Ali Shah
Aerospace Engg Dept, Aerospace Engg Dept, Aerospace Engg Dept,
College of Aeronautical Engineering, NUST College of Aeronautical Engineering, NUST College of Aeronautical Engineering, NUST
Risalpur, Pakistan Risalpur, Pakistan Risalpur, Pakistan
[email protected] [email protected] [email protected]

Hamza Akbar Khan Taimur Ali Shams

Aerospace Engg Dept, Aerospace Engg Dept,
College of Aeronautical Engineering, NUST College of Aeronautical Engineering, NUST
Risalpur, Pakistan Risalpur, Pakistan
[email protected] [email protected]

Abstract—An undeniably significant development in the cycle, single monomers, and oligomers, blended in with or
aviation business is the utilization of composite materials, as without a curing agent, react to frame a tridimensional
these empower designers to conquer the boundaries made by polymeric network. In the initial segment of the reaction,
utilizing metals. Composite materials have assumed a significant branched atoms with different structures are formed, and their
part in weight reduction, and consequently they are utilized for
sub-atomic weight increments in time with the degree of the
both primary applications and segments of all space apparatus
and airplane from UAVs and sight-seeing hot air balloon to response until the organization size is equivalent to the size
military aircraft, space transport and traveler carriers. During of the framework [1]. The leftover particles begin to exist
this research, an extensive literature review has been performed together with the plainly visible organization until they
related to composite curing techniques and processes. As it is respond with the organization making other crosslinks. The
known that many of the aircraft components including skin crosslink thickness increments until the framework arrives at
panels, wing-box and even landing gear struts are being made the finish of the synthetic reaction. Restoring can be started
up of composite materials, therefore, specimen manufacturing by heat, radiation, electron shafts, or compound added
for testing and qualification of defined curing process and substances. To cite from IUPAC: relieving "may or probably
material becomes the paramount. Composite curing using
won't need blending in with a substance restoring agent."
compressive molds is considered to be the most widely used and
economical process in this field. In the same context, a Thus, two wide classes are a) curing actuated by synthetic
compressive mold was designed and developed to study the added substances (likewise called restoring specialists,
effect of temperature and pressure for finalization of the curing hardeners) and (b) curing without added substances. A
cycle for various thicknesses of composite specimen. CAD halfway case includes a combination of resin and added
assembly of the designed mold was modeled in substances that requires outer upgrade (light, heat, radiation)
SOLIDWORKS®. The design was kept simple for ease of to instigate curing. The relieving technique relies upon the
manufacturing and cost effectiveness. Static structural followed resin and the application. Specific consideration is paid to the
by transient thermal analysis for different materials were shrinkage actuated by the curing [2]. Normally little
performed in ANSYS® to validate the structural integrity of
estimations of shrinkage (2-3%) are desirable. The aim of this
manufacturing materials under high temperature and pressure
environment required for curing of composite materials. Based research was to design, analyze and fabricate a compressive
on these analysis, final design and material needed for mold for curing of different composite based pre-preg
manufacturing of compressive mold were finalized followed by materials. Specimen needed to validate mechanical properties
fabrication of mold. Use of this fabricated mold is aimed to of cured composite materials as per ASTM standards would
define curing cycle for different composite based pre-preg be manufactured using the fabricated mold under different
materials of various thicknesses being used in the ongoing and temperature and pressure environment. Qualified material
future research projects for aerospace applications. and curing cycle would then be recommended for different
aerospace applications including landing gear struts.
Keywords— Composite Curing, Compressive mold, Design,
ANSYS, Structural Analysis, Transient Thermal Analysis, II. METHODOLOGY
Structural Integrity.
In the first phase of this research, various methods for
I. INTRODUCTION curing were studied and the pros and cons of each were
Composite curing is an essential process to enhance the compared. Curing process involving compressive mold was
strength of composite products being formed. Over the years, selected based on cost effectiveness, material availability and
researchers have developed various methods for curing availability of fabrication facility. A reference mold used for
composite materials. Each method has its own benefits and manufacturing of main landing gear strut of a lightweight
are mainly used to improve the strength of products being aircraft was taken as a base model.
manufactured by composite materials. During the curing

Seventh International Conference on Aerospace Science & Engineering 220

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

That mold consists of a temperature regulator b3sed on

four thermocouples and a pressure mechanism based on five
screw jacks pressure mechanism having maximum capacity of
15 tons. Glass fiber prepreg material is used to manufacture
main landing gear struts of a lightweight aircraft using
reference mold. The process of curing begins by carefully
heating and applying pressure simultaneously. The heat is
increased gradually and is held for several minutes at several
different temperatures for the cross linking to take place. The
resin is removed simultaneously as the pressure is applied till
the required thickness of the final product is achieved. After
the curing process is completed, final product is allowed to
cool for 48 hours and is finally removed from the mold. Two
test specimens are obtained from the final product and
subjected to shear and bending tests as per ASTM standards.

On qualification of these tests, the final product (main

landing gear struts) is declared fit for installation on aircraft.
Reverse engineering approach was used to develop a small
size compressive mold capable of making specimen of
composite materials of required length, width, and height for
testing as per ASTM standards. Glass fiber-based specimen
were required to be manufactured and qualification of small
size compressive mold was subjected to clearance of bending
and shear strength tests of these specimen [3]. After
qualification of these tests, fabricated compressive mold was Fig. 1. Research methodology for the design and development of
cleared for future research work related to various Composite curing mold.
thicknesses, materials, and finalization of curing cycles.
Even though it is most strongly associated with the
This project required the use of SOLIDWORKS® and development of thermosetting polymers, the term curing can
ANSYS® software. The initial models were developed in be applied to any method that results in a solid product from
SOLIDWORKS® and the analysis were carried out using a liquid solution. The curing of composite resins is divided
ANSYS® software. ANSYS® was used for the structural and into three stages: pre-gel, gel point, and post-gel. To
thermal analysis. Our research methodology involved compensate for shrinkage forces, the material can flow and
following steps and is depicted as: undergo molecular rearrangement during the pre-gel process.
There is a predominance of linear polymer chains during this
1. The Study of different techniques conventionally phase [5]. The resin then transitions from the flow state (pre-
used for composite curing. gel) to the viscous state (post gel), establishing its gel stage.
2. Selection of suitable manufacturing techniques During the post gel process, the resin loses its flowing
based on various parameters studied earlier. capacity and transmits the stress generated by polymerization
3. To select base line design for reference and detailed shrinkage to the tooth-restoration interface. At this point,
study of the reference mold to understand the cross links predominate in the polymer structure.
governing design principles.
4. Detailed design and assembly of different parts of The curing process can be carried out by using various
mold in SOLIDWORKS®. techniques. Each of these techniques have their own
5. Selection of the most appropriate design with ease advantages and disadvantages. The technique used for
of manufacturing and useability. composite curing depends on the cost and strength
6. Structural and thermal analysis at different requirement of the designer. The various methods for
temperature and pressures using higher fidelity composite curing are as follow:
tools.
7. Finalization of mold for fabrication. 1. Autoclave Curing
8. Development of the composite curing mold. 2. Resin Transfer Molding
3. Compressive Molding
III. LITERATURE REVIEW
A thorough literature review was carried out for better A typical curing profile contains two ramps and two
understanding of methods involved in composites curing and isothermal holds. The first ramp and isothermal hold, which
controlling physical and chemicals changes occurring during are typically in the 240-280oF range, are used to enable the
the curing process. Composite curing is a chemical process resin to flow (bleed) and volatiles to escape. The viscosity
used in polymer chemistry and process engineering to drops dramatically during this process due to the fracturing of
toughen or harden a polymer material through the cross- cross polymer chains by heating. The polymerization
linking of polymer chains [4]. component of the cure cycle is represented by the second ramp
and hold. During this point, the resin viscosity drops slightly
due to the addition of heat and then rises dramatically as the
resin's kinetics initiate the cross-linking process. The resin gels

Seventh International Conference on Aerospace Science & Engineering 221

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

into a solid, and the cross-linking process continues product. Some of the frequently utilized materials are
throughout the second isothermal hold, which is usually aluminum, mold steel, die steel and stainless steel for making
between 340-370oF for epoxy resin systems. The resin is compressive molds. Material properties must be provided for
usually kept at this cure temperature for 4-6 hours to enable the analysis as it is nearly impossible to simulate or analyze
the cross-linking process to complete [6-8]. the model without the definition of material properties.

IV. DESIGN OF COMPOSITE CURING MOLD

The design requirements of a compressive mold required
for curing of composite materials are as follow:

1. The product should be light weight for ease in

transportation.
2. The product should have high strength and thermal
conductivity.
Fig. 2. Schematic diagram of the Autoclave [9]. 3. The product should be able to sustain high pressure
and temperature environment.
4. The product should be able to accommodate varying
thicknesses of composite specimens.
5. The product should be economical.
6. The product should be easily manufacturable.
7. The design should be compact and easy to use.

A baseline design of compressive mold used to cure thick

composite beams (landing gear strut) of fiber glass based
composite material was selected to study and understand the
complete curing process. This thick composite beam is used
as a full-scale main landing gear strut of a lightweight aircraft.
Fig. 3. Working Principle of Resin Transfer Molding (RTM) [10]. However, for research purpose a small-scale compressive
mold was aimed to be manufactured to cure test specimens
before manufacturing of a full-scale composite product.

Fig. 4. Working Mechanism of Compression Molding [11]. Fig. 6. The side walls and heating element of the reference mold in
disassembled form.

A. CAD Modelling of Mold

Computer aided model is the prerequisite of any design
project. 3D CAD allows engineers to not only design in 3D
but also animate aspects to better visualize the working of a
concept. It also helps the customers to better understand the
features of their required product[[12, 13]. Various models
were designed as per the required dimensions of specimens to
be cured. All these designs were modeled in
SOLIDWORKS® and were then imported to ANSYS® for
further computational analysis. The static structural analysis
under high temperature was carried out for all these models
using different materials. The upper plate of the mold was
subject to a pressure of 15 ton using screw jack mechanism.
A uniform controlled temperature was provided by heating
Fig. 5. Composite Curing profile for temperature and pressure during rods placed in upper and lower plates. The structural analysis
manufacturing [8].
was conducted to qualify the mold design and material for
given pressure and thermal loads. Final design of the mold is
For curing of composite materials, use of compressive
shown in figure 7.
mold is considered to be the most economical solution.
However, selection of material for manufacturing of
compressive mold is very critical as curing cycle involves
high temperature and pressure environment. Material
Selection is one of the critical issues of any designing project
as it affects the weight, performance, and efficiency of the

Seventh International Conference on Aerospace Science & Engineering 222

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 9. Heating elements to provide the required heating temperature.

Heating Plates
Rod

Fig. 7. Completed assembly of the mold to be developed designed in

SOLIDWORKS®.
Fig. 10. Heating rods enclosed in plates for curing through both ends.
B. Mold Component design
The design is kept remarkably simple for ease of 3) Thickness Measurement Gauge
manufacturing. It has the following components: A manual thickness gauge was set on the sides of the rod
1) Screw Jacks which was used to read the thickness of the composite
The pressure is applied on the mold through a screw jack material when pressure is applied to it. The main aim of the
mechanism for the purpose of compression of the composite pressure jacks is to achieve the required thickness of the
pre-pregs sheets and are placed on the top plate. It was composite material being cured. The thickness gauge
estimated that maximum of 17.5 tons force which comes out determines the thickness of the composite material under
to be 33.18 ton-force/ ft2 pressure (32 bar approx.) can be curing. alongside the thickness gauge are side plates which
applied on baseline model. However, approximately 40 % of holds prepreg sheets in place and do not let them slip
this force is applied. The screw Jack mechanism shown in sideways.
figure 8.

Fig. 11. Thickness plates and side plates.

V. COMPUTATIONAL ANALYSIS OF THE COMPRESSIVE MOLD

Fig. 8. Manual screw jack mechanism of pressure application.

Analysis of the designed molds was carried out in
ANSYS®. At first, the mesh was defined followed by
boundary conditions. A structural analysis and transient
2) Heated Pressing Plates thermal analysis were carried out. The Analysis was carried
A minimum of 6 heating rods were placed at the top and out using different materials which are used in mold
bottom of the mold for uniform heating throughout the mold manufacturing and are readily available in the market. Some
at temperatures up to 250oC. Six heating rods were used to of these materials are as follows [5]:
cater for the failure of any rod during curing process. These
rods are operated through a Control Box having magnetic 1. Aluminum
connectors, breaker, temperature controller, and other items 2. Carbon steel
which are required for application of temperature under 3. Titanium Alloy
control environment. The controller measures up to 400oC. 4. P-20 pre-hardened mold steel
Such a high temperature is kept in case of studying curing
methods of other composite materials. At least two A. Mesh Convergence
thermocouples are placed in each pressing plate at edges and Mesh convergence was carried out and hence the
center of the mold to measure and show a uniform temperature following parameter has been finalized after the mesh
in digitally. convergence study as depicted in Table :

TABLE I. MESHING PARAMETERS OF CONVERGED SOLUTION

Seventh International Conference on Aerospace Science & Engineering 223

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Nodes 1292296
Elements 509139
Element type Tetrahedral
Element size 5 mm

Fig. 14. Boundary conditions for transient thermal analysis.

1) Transient Thermal Analysis

Fig. 12. Meshing of the mold. The mold underwent a variable change in temperature
over time while pressure was applied simultaneously. The
B. Boundary Conditions maximum temperature which the heating rods could provide
Boundary conditions are defined for the assembly prior to was around 400oC but the working temperature was around
the analysis. For this purpose, mechanism of the mold (the 200oC. The transient thermal analysis was carried out at a
upper tool and the lower plate) were kept fixed to the four temperature of 200oC. The results are as shown below as
rods. The heating rods were placed inside the heated pressing figure 15.
plates. The upper heated pressing plate was attached to the
base of the screw jack pressure mechanism while the lower
heated pressing plate was fixed to the lower plate. The screw
Jack mechanism could apply a pressure of 33.18 ton-force/
ft2 (32 bar approx.). However, approximately 40 % of this
force as applied on baseline model was applied. There was no
mechanism to calculate the pressure as the use of pressure
was only to compress and achieve the desirable thickness of
the specimen. The maximum temperature which could be
conducted through heating rods was 400oC but the working
temperature for the mold was around 200oC which is
considered sufficient for the most of composite material. The Fig. 15. Aluminum Transient Thermal Analysis.
Boundary Condition for Static Structural Analysis and
transient thermal are depicted in figure 13 and 14, The results indicated that the Aluminum material is a good
respectively. conductor of heat and the thermal stresses on it are negligible.
2) Total Heat Flux
The heat flux for the mold using aluminum as the material
was carried out and results are as shown:

Fig. 13. Boundary conditions for static structural analysis.

C. Analysis with Aluminum

Aluminum alloys, with their flexibility, versatility, and
minimal effort, offer numerous favorable circumstances over
steel and different materials for mold makers. Results of
Fig. 16. Aluminum Heat Flux.
analysis using aluminum material are as follows:
3) Maximum Principal Stress
The result for maximum principal stresses occurring on
the assembly is as follows:

Seventh International Conference on Aerospace Science & Engineering 224

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 19. Transient Thermal Analysis.

Fig. 17. Maximum Principal Stress of Aluminum.

The results indicated that the material carbon steel is a
good conductor of heat and the thermal stresses on it are
4) Factor of Safety (FOS) negligible.
The results indicated that a max stress of 152.33 MPa
occurs on the mold. The Ultimate Tensile Strength of the 2) Total Heat Flux
Aluminum alloy is 310 MPa. Hence the Factor of Safety is Heat flux for the mold using carbon steel was carried out
calculated as: and results are as shown in figure 20:
𝐹OS = 2.035

Fig. 20. Carbon Steel Total Heat Flux.

Fig. 18. FOS for Aluminum.
3) Maximum Principal Stress
The Aluminum Alloy material qualifies for the compressive The result for maximum principal stresses occurring on
mold. the assembly is as follows:
D. Analysis With Carbon Steel
Carbon steel is an exceptional kind of steel that, as the
name proposes, has a higher convergence of carbon than
different sorts of steel. Other forms of steel have a generally
low carbon substance of about 0.05% to 0.3%. In correlation,
carbon steel has a carbon substance of up to 2.5%. Over two
percent carbon may sound irrelevant, however it presents a
few alluring advantages that are not found somewhere else.
Results of analysis using carbon steel material are as follow:

1) Transient Thermal Analysis

The mold underwent a variable change in temperature
over time while pressure was applied simultaneously. The Fig. 21. Maximum Principal Stress.
maximum temperature which the heating rods could provide
was around 400oC but the working temperature was around 4) Factor of Safety (FOS)
200oC. The transient thermal analysis was carried out at a The results indicated that a max stress of 932.5 MPa
temperature of 200oC. The results are as shown. occurs and the min stress -416.09 MPa occurs. The negative
sign indicates compression. The Ultimate Tensile Strength of
the Carbon Steel is 635 MPa. Hence the Factor of Safety is
calculated as 4.81 and is also shown in figure.22.

Seventh International Conference on Aerospace Science & Engineering 225

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

steel was finalized as the material was easily accessible and

cheaper. Basic design layout, modeling, analysis, and
fabrication of the compressive mold was carried out. The
fabricated mold was tested by curing a composite specimen
to verify the application of its pressure mechanism and
conduct of heat through heating rods. These specimens were
qualified for bending and shear strength tests as per ASTM
standards which ultimately qualified the fabricated mold as
well.
The mold was fabricated in a way that it has the following
Fig. 22. FOS of Carbon Steel.
advantages:

5) Principal Elastic Strain 1. The height of the mold was kept flexible by
The results for principal elastic strain occurring on the changing the vertical rods.
assembly is as follows: 2. The mold was designed to accommodate around 200
pre impregnated sheets, but the fabrication is done
such that we can accommodate more than 200 sheets
by changing the height as mentioned above.
3. The control box is designed such that it can work on
both single phase and 3 phase power sources.

The fabricated mold is as shown:

Pressure Jacks
Fig. 23. Carbon Steel Maximum Principal Elastic Strain.

The results show that the maximum strain to be.015007 and the Side
minimum strain to be 0.00165. plates

6) Total Deformation
The total deformation of the carbon steel assembly while Control
undergoing pressure and temperature is shown as follows: Box

Frame

Fig. 25. Fabricated Compressive Mold Front View.

Pressure Jacks

Fig. 24. Carbon Steel Total Deformation. Thermocouple

The results for carbon steel indicate that the assembly will
be fully functional without any sort of damage to the
assembly during the curing process. All the other materials Heating Heating
i.e., Titanium alloy and P-20 pre-hardened mold steel will plates rods
also qualify but the analysis was not carried out as these
materials are very expensive and difficult to machine.

A comparison of the aluminum and carbon steel shows

that carbon steel is less costly, and it has particularly good
manufacturability. Moreover, it can be heat treated and
hardened to get more strength and it has a good durability and
finishing ability. Therefore, carbon steel is preferred for this
Fig. 26. Fabricated Compressive Mold Side View.
Mold assembly.
VII. CONCLUSION
VI. FABRICATION OF DESIGNED MOLD The basic design layout, modeling, analysis, and
After the design and analysis phase, fabrication of the fabrication of compressive mold for curing of composite
mold was undertaken. On the basis of results obtained, carbon materials were carried out. The purpose of this mold is to

Seventh International Conference on Aerospace Science & Engineering 226

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

conduct experiments and research work related to curing [3] H. Thuis, "The development of composite landing gear
components for aerospace applications," 2004.
cycle of different composite materials. The compressive mold
[4] P. Sormunen and T. Kärki, "Compression molded thermoplastic
provides a controlled environment for the curing of composite
composites entirely made of recycled materials," Sustainability,
pre-pregs materials at elevated temperature and pressure. vol. 11, no. 3, p. 631, 2019.
Structural analysis was carried out prior to the fabrication of [5] A. Trende, B. Åström, and G. Nilsson, "Modelling of residual
mold. The material selected for fabrication is based on the stresses in compression moulded glass-mat reinforced
availability in market, price, ease in manufacturing and thermoplastics," Composites Part A: Applied Science and
Manufacturing, vol. 31, no. 11, pp. 1241-1254, 2000.
qualification through computational analysis. After the
[6] R. Upadhyay, "Compression molding of composites: Mold
experimental verification of the mold through a series of heating system design," Advances in Polymer Technology:
shear and bending stress tests as per ASTM standards (D790 Journal of the Polymer Processing Institute, vol. 8, no. 3, pp. 243-
and D2344), the mold is qualified for further studies 264, 1988.
involving use of various composite materials and varying [7] R. A. Tatara, "Compression molding," in Applied plastics
thicknesses. engineering handbook: Elsevier, 2017, pp. 291-320.
[8] F. Campbell, "Curing: It’sa Matter of Time (t), Temperature (T)
VIII. ACKNOWLEDGEMNT and Pressure (P)," Manufacturing Processes for Advanced
Composites, pp. 175-221, 2003.
Authors acknowledge receipt of a grant from Higher [9] P. J. Halley, "Rheology of thermosets: the use of chemorheology
Education Commission (HEC) of Pakistan under TDF 03-332 to characterise and model thermoset flow behaviour," in
project for completion of this research project. Thermosets: Elsevier, 2012, pp. 92-117.
[10] A. Ahmadova, "Numerical modelling of porosity generation,
REFERENCES movement, and compaction during the RTM process."
[1] N. Ismail, A. Sulong, N. Muhamad, D. Tholibon, M. MdRadzi, [11] C. Park and W. Lee, "Compression molding in polymer matrix
and W. Wanlbrahim, "Review of the Compression Moulding of composites," in Manufacturing techniques for polymer matrix
Natural Fiber-Reinforced Thermoset Composites: Material composites (PMCs): Elsevier, 2012, pp. 47-94.
Processing and Characterisations," Pertanika Journal of Tropical [12] M. Barone, D. Caulk, and M. Panter, "Experimental verification
Agricultural Science, vol. 38, no. 4, 2015. of an optimal thermal design in a compression mold," Polymer
[2] I. Baran, K. Cinar, N. Ersoy, R. Akkerman, and J. H. Hattel, "A composites, vol. 7, no. 3, pp. 141-151, 1986.
review on the mechanical modeling of composite manufacturing [13] M. A. Kuczmarski and J. C. Johnston, "Improved thermal design
processes," Archives of computational methods in engineering, of a compression mold," Advances in Polymer Technology, vol.
vol. 24, no. 2, pp. 365-395, 2017. 26, no. 2, pp. 86-99, 2007.

Seventh International Conference on Aerospace Science & Engineering 227

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Liberation Points of Space

Ms. Hadia Batool Ms. Sumaira Abbasi

Fazaia Bilquis College of Education, PAF Base Nur Khan Fazaia Bilquis College of Education, PAF Base Nur Khan
Rawalpindi, Pakistan Rawalpindi, Pakistan
[email protected] [email protected]

Abstract— In the world of space, in the region of zero gravity, points.

some places are found to have balanced forces and can be proved
as stable areas of space. On the basis of three-body problem, in
any system of two celestial massive bodies, these points are III. MACHANISM
present by Nature and act as parking places of space. By years,
this phenomena is used by agencies to explore many areas of
The mechanism of Libration points actually works of the pace
space. They are total 5 in number and categorized as stable and of space object relative to the pace of Earth. Objects at L-
unstable Libration points. Their stability depends upon how points remain in the same position relative to the Earth. The
much an object, placed there, can change its initial position after orbital period of Earth is 365 days per year and so that is for
experiencing small perturbation. These points act as point of satellites at L-points. [1]
equilibrium where spacecraft may be parked to make Simply the orbital period of L-points are all same about 365
observations.
days but at different places. This is how the benefit of these
Keywords— Balanced Gravity, Gravitational Law, Imaginary points is that there is no need for much fuel for continuing the
and Real Eigen Values, Linear Stability Analysis, Three-body motion of spacecraft or telescope in space and also no much
Problem need to have some controlling office on Earth, because
naturally the objects place there with move with the relative
I. INTRODUCTION speed of the Earth.
Gravity, as an attractive force, is always pulling you down, The proposal of using the system of “Libration points” of the
but there are places in the solar system where gravity Earth-Sun system is for various purposes, including to design
balanced out so nicely and agencies use them as “Parking spacecraft orbits which explores a large region of space in the
places” of space [3]. Nature is on them and has already been vicinity of Earth, using low-fuel control.
using them for billions of years. These are named as
“Libration points”. They are actually stable location in space IV. STRUCTURE OF L-POINTS
where the combined effect of force of gravity of two large Libration points are total 5 in number which are naturally
bodies, equalizes the centrifugal force felt by a much smaller present in system of any two celestial bodies.
third body. All matches up in a way that causes a smaller First 3 points are located on the same line with
object to maintain a stable position relative to larger bodies. L1 approximately 1.5 million km from Earth towards the Sun.
These are points in space where forces of gravity of larger Also L2 is approximately 1.5 million km from Earth in the
bodies and orbital motion of a smaller body balanced each direction opposite the Sun. And L3 is located on the side of
other and result motion with zero velocity. They are also the Sun opposite to the Earth, as being completely hidden by
known as “Lagrangian points” or “L-points” [2]. If a the Sun. L1, L2 and L3 are termed as “Collinear Libration
spacecraft uses a Libration point close to Earth, there are points”. While, L4 is located close to the Earth's orbit at 60°
many benefits to the locations with low fuel controls. As ahead of Earth and L5 is located close to the Earth’s orbit at
Mainzer said, “A spacecraft placed at a lagrangian point Figure (1): Structure of L-points
would be able to do more”.
60° behind the Earth figured as creating the apex of two
II. LITERATURE REVIEW
In the light of astrophysics theory, many studies have been
made about Galaxy’s objects. From the special case of n-body
problem, we have considering the three-body problem and
demonstrating its specific solutions for both collinear and
triangular three masses. Some of the contributions are:
Isaac Newton (1643 - 1727), Newton’s laws of motion make
bases for the three body problem, by taking the initial
positions and velocities of three point-masses.
Leonard Euler (1707 - 1783), solved the first three-body
problem considering the system of Sun, Earth and moon triangles. That’s why L4 and L5 are termed as “Triangular
during collinear eclipse motion. Libration points”.
Joseph-Louis Lagrange (1736 - 1813), contributed for the
solution of equilateral triangle in 1772.
Both these concept (collinear and equilateral problems) V. STABILITY ANALYSIS
constitutes the base for the trace of discovery of Libration To examine the stability of these points, we will assume some

Seventh International Conference on Aerospace Science & Engineering 228

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

structural composition of this system. This concept consist of

primary mass, secondary mass and an object at some of L- (11)
points. Let the mass of primary, Sun, is (1-μ) and secondary Now, by multiplying (10) by cos (t) and (11) by sin (t);
mass, Earth, is denoted by μ. The position of the object at L- We get (a=c), as:
point is denoted under the circular restricted three body
problem, by the reference of center of mass on a line between (12)
primary and secondary masses. The distance of primary to the Also, by multiplying (10) by -sin(t) and (11) by cos(t);
center is μ and the center of mass from secondary is (1-μ). We get (b=d), as:
These distances between the masses is sums to unity. The
collinear points are lie on a straight line having y-axis points (13)
as zero [6]. And the triangular points are based in the form of These are our required two equations of motion for a satellite.
equilateral triangle, placed at a position of its vertex, with the Now for simplifying these equations we will use an effective
length equals to unity. The base of these equilibrium point is potential for approximate method, in order to simplify
the `Law of Universal Gravitation'. complex systems, called as “Pseudo-Potential”.
Using the Law of Gravitation: And we introduce it by a function U as:
−𝑚 𝑚 𝑟 −𝑚0 𝑚1 𝑟2
𝐹0 = 0 3 1 1 + 3 (1)
𝑟1 𝑟2

Where; , Derivate U with respect to x0:

and . (14)
Proceeding (1); Derivate U with respect to y0:

(2) (15)
Also, from 2nd Law of motion F=ma, Now equation (12) can be rewrite as:
Equation (2) become:
(16)
And also equation (13) can be rewrite as:
(3)
Further we will use inertial frame of reference with ᶓ and η. (17)
Equation of motion in inertial frame are: By comparing (14) and (15) with (16) and (17) respectively,
we have;
(4)
and (18)

(5) (19)
Using inertial axis components as: Now, Multiplying (18) and (19) by derivative of x0 and y0 ,
and then adding both equations we get:
Now by taking double derivative of both ᶓ and η, we get:
(20)
(6) Since, U is the function of both x0 and y0:
and So,
(7)
Putting values of ᶓ and η in (4) and (5), we are: Also,
(8)
and
(9) Putting these 3 expressions in (20) and multiplying by ‘dt’
Comparing (6) and (7) with (8) and (9) respectively with we have:
similar terms taken as:

On integrating this expression, we get:

It is required equation to represents the region of satellite with zero

velocity.
Now,
It is to be determined that all of these Libration points are not fully
stable. Some of them are stable while others are unstable points. In
Now the resulting equations are: order to analyze their stability we will further perform some of the
calculations in order to find out the Eigen values for each L-point.
(10)
For this, we will see the behavior of object at those places by slightly
displaced them from their position. If the position of object at any L-

Seventh International Conference on Aerospace Science & Engineering 229

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

point is displaced by the smaller amount X and Y. In other words, L5 8.95×1027 𝑖 9.46×1013 𝑖
the object's position is perturbed by its initial value (xL , yL) into Table (2): Resulting Eigen values
(xl+X , yL+Y).
By proceeding (18) and (19) and solving the following determinant: Thus, we get a no. of real and imaginary Eigen values for all L-
points.

VI. RESULTS
The resulted equation we have is: If all 𝛌-values are real, the exponential terms containing 𝛌
will increase with time and the object will drift away from the
(21)
equilibrium point toward infinity. Thus real Eigen values
Solving terms for (21) using values of U, r1 and, r2 we have:
represent unstable L-points. The resulted values shows that as
being ‘unstable’ L1, L2 and L3 are placed in saddle nodes and
(22)
small perturbation would set them ‘falling’ in either direction
and they need station keeping. Thus ‘Collinear Libration
(23) points’ are “Unstable L-Points”.
However, if 𝛌-values are imaginary, the position of the object
(24) will oscillate about the equilibrium points and then become
Now, for “collinear-points” (along x-axis at the line joining Sun and stable. Hence, imaginary Eigen values represent the stable L-
Earth), all the y-points are zero: y0 = y1 = y2 = 0. This means the points. From the above resulted values L4 and L5 being stable,
terms (y1 - y0) and (y2 - y0) are zero in all equations. Further solving feel attractional forces of gravity from both bodies toward
(31), (32) and (33) and putting the resulting values in (21), the
them in the same ratio and the resultant acceleration will not
required equation in terms of 𝛌 is:
tend them fall out of orbits and it will not drift away and
(25) maintain its equilibrium position [4]. Thus, Triangular L-
Similarly, for ‘triangular-points’ (in the form of equilateral Points are Stable L-Points.
triangle), having all sides of equal length, so; r0 = r1 = r2 = 0. Also
using x1 = -μ, y1 = 0, x2 = 1-μ and y2 = 0, the resulting equation (21) VII. CONCLUSION
in terms of 𝛌 becomes: Being unstable places means under the effect of small
perturbation the object doesn’t resist its initial position and so
(26) there is chances to leave position at L-point and attain
By using quadratic formula for solving (25) and (26), we get the another. For this reason, to control objects at collinear L-
following expressions for Eigen value 𝛌: points, there is need of some station on Earth to keep that
object in its place.
For collinear points (27) But being stable, triangular L-points doesn’t need any sort of
station-keeping. This interesting phenomena is attractive to
For triangular points (28) all space missions with low-fuel contain.
Where;
VII. SPACE MISSION
Space missions are formally the sending of manned or
Now, if we specify our results for the system of Sun and unmanned vehicles into space, beyond the Earths
Earth. gravitational effect, by setting its specific orbital velocity for
We can find out exact value of 𝛌 for all Libration points, with exploring the facts of Universe, for gathering data about
using following values: activities in space, for the structural views of other planets,
Mass of Sun; Primary mass = 1-μ = 1.989×1030 and for many other purposes.
Mass of Earth; Secondary mass = μ = 5,972×10 24 To set a satellite for space missions, some controlling centers
Also the values of r1 and r2 for all the libration points as been have to be created at Earth so that to controls its pace and its
determined by using the restricted three-body problem and orbital location. Along with this controlling system, a huge
gravitational law between Sun and Earth. The resulted values amount of fuel is needed by satellite or spacecraft to keep
are: itself into a stable orbital position, to prevent any
destructions.
L-Points r1 (km) r2 (km) ‘A’
L1 1.505×108 1.5×106 2.35×106 The position of Libration point is servable with more benefits
L2 1.535×108 1.5×106 2.319×106
because the vehicle at such point doesnt need any controlling
L3 152.09×106 304.18×106 5.65×105
station and it can work using low-fuel control. This makes L-
Table (1): Calculated values of r1 and r2 points beneficial for this reason and thus there are many space
Putting all these values in (27) and (28), to get Eigen values missions working at L-points. As per structural position of
for each L-point. The resulting values for stability analysis Libration Points, the space missions working are as followed:
are: o Space mission at L1:
 Deep Space Climate Observatory (DSCOVR);
L-Points 𝛌2 𝛌 Launched in 2015, for the observation of solar
L1 469500 6.95×102 wind for affecting many events of space
L2 2.3×106 1.516×106 weather like geomagnetic storms, by NASA.
L3 1125000 1.06×103  Solar & Heliospheric Observatory (SOHO);
L4 8.95×1027 𝑖 9.46×1013 𝑖 Launched in 1996, to study internal structure

Seventh International Conference on Aerospace Science & Engineering 230

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

of the Sun, its outer atmosphere and origin of There are also some of the proposed missions toward L4 and L5.
Solar wind. SOHO marks 20 years watching
the Sun, by ESA. VIII. LIBRATION POINTS AS LIBRATION AREAS
Libration points are indeed points, of infinitesimal in size,
they can be named as Libration areas. Surrounding all L-
point, is an extended zone where a spacecraft can
conveniently put itself in an orbit that require little fuel to
Figure (2): SOHO
maintain
o Space mission at L2:
 Wilkinson Microwave Anisotropy Probe
(WMAP); Operating from 2001 to 2010, to
measure temperature differences across sky in
the cosmic microwave background, by NASA.
 James Webb Space Telescope (JWST); Started
in 2002, and will be launched in 2021. Its
function include observing some of the distant .
events and objects in the Universe and direct Figure (5)
imaging of exoplanets, by ESA. It is successor Each object form halo orbit around the Libration points [7].
to Hubble Telescope [5]. Halo orbit is possible due to the gravitational interpolation
between two massive bodies, and that creates a small area of
stability for the third smaller body.

On the basis of this work, we will extend this concept for

Mars-Moon system. As many of the space mission has been
Figure (3):JWST
starting to reach Mars an exploring life there. So the discovery
o Space mission at L3:
As per structure of L3, it is always hidden by the Sun. So of Mars-Moon Libration points will enhance the possibility
there is no space mission for the working at L3 till now of reaching Mars with more ease.
because it has no direct approach [8]. It can be supposed
that L3 is the living place for aliens hidden by human
beings
o Space mission at L4 and L5: REFERENCES
It is observed that some of the natural objects, such as [1] Oscar Perdomo, Relativistic Lagrange points, May 15, 2016.
asteroids and meteoroids, are positioned stably at [2] Thomas Greenspan, Stability of Lagrangian points, January 7, 2014.
Triangular L-points. [3] Elizabeth Howell, Lagrangian points: Parking places in space, August
22, 2017.
[4] Markeev A. P, Libration points in celestial mechanics and space
dynamics, 1978.
[5] Cacolici and Kyle Alexander, ‘Stability of Lagrange points: James
Webb Space Telescope’.
[6] Sean P. Kemp, Stability at Triangular Lagrange point for a Three-Body
Problem, May 2015.
[7] Klarreich, Lagrangian points of the Earth-Sun system.
Figure (4): Asteroids [8] Elena Fantino, Spacecreaft trajectories to L3 point of sun-earth system,
November 2010.
Some of asteroids found at L4 and L5 of Sun-Earth system
are:
 Asteroid 2010 TK7, is the first discovered
"tadpole" orbiting L4, companion to Earth
 Asteroid 2010 SO16, is a “horseshoe”
companion to Earth, orbiting L5 with high
inclination.
 STEREO A, in 2009, made its closest pass to L4,
slightly faster than Earth.
 STEREO B, in 2009, made its closest pass to
L5, slightly slower than Earth.

Seventh International Conference on Aerospace Science & Engineering 231

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Deinococcus Radiodurans As A Chassis For

Synthetic Biology Experiments In Space Biology
Muneef U. Rehman Khudeja Salim Faisal F. Khan
Institute of Integrative Biosciences, CECOS-RMI Precision Medicine Lab, CECOS-RMI Precision Medicine Lab,
CECOS University of IT and Emerging Rehman Medical Institute, Phase 5, Rehman Medical Institute, Phase 5,
Sciences, Phase 6, Hayatabad, Hayatabad, Peshawar, Pakistan Hayatabad, Peshawar, Peshawar
Peshawar, Pakistan Peshawar, Pakistan Institute of Integrative Biosciences,
Peshawar, Pakistan [email protected] CECOS University of IT and Emerging
[email protected] Sciences, Phase 6, Hayatabad,
Peshawar, Pakistan
Department of Bioengineering,
University Of Engineering and Applied
Sciences, Swat, Pakistan
Peshawar, Pakistan
[email protected]

Abstract—The design and construction of new biological parts,

devices, and systems, including the re-design of existing, natural B. Details of Martian Conditions
biological systems for beneficial purposes, is referred to as Space and Martian conditions including Cosmic Ionizing
synthetic biology. Engineered systems are then deployed in living Radiation (CIR), Ultraviolet Radiation (UV), desiccation,
host cells referred to as a chassis or cell-free systems. One such low temperatures, variable pressures and vacuum have an
promising candidate for a strong chassis is Deinococcus adverse effect on living organisms [2]. Space and Mars
radiodurans owing to its viability in extreme conditions, efficient
environment both have a high flux of cosmic and solar
proteomic shielding, production of novel carotenoids and high
transformability. Interplanetary space travel and the colonization particle events (SPE’s) [1], combined solar radiation on Mars
of Mars come with several key challenges due to the extreme varies seasonally, in summer the radiation calculated on the
environment. Various microbes have been reported to have southern hemisphere of Mars has been shown to be roughly
capabilities that can help them survive harsh Martian and space 13000 kJ m-2 of wavelength > 400nm substantially decreasing
conditions such as Deinococcus geothermalis, Bacillus subtilis, in winter [3].
Pseudomonas aeruginosa, Candida albicans, Escherichia coli and
Micrococcus luteus. In this study, we explored the natural C. Concept Explaining Transfer of Life
potential of D. radiodurans by the identification of individual Studies have been performed to understand the concept of a
features it possesses as an optimal chassis. An analysis of rocky transfer of life (microorganisms) now termed as
transcriptomic datasets from the literature, revealed 14 Lithopanspermia [5]. One such study demonstrated the
Differentially Expressed Genes (DEGs) that confer survival and
increased survivability of Deinococcus geothermalis biofilm
ideal chassis features to D. radiodurans. This data can be used to
construct a desirable synthetic biology device (chassis) that could cells under UV-radiation, when supplemented with S-MRS
be used for experiments in space and on Mars. (A Martian soil Analog) as compared to the non-
supplemented biofilm [2]. In other experiments, spores and
Keywords—Deinococcus radiodurans, synthetic biology, chassis, cells of Bacillus subtilis and Chroococcidiopsis genus were
LEO, Martian and space conditions used in simulations by encapsulating them inside the
meteorites and exposing them to different stress conditions
I. INTRODUCTION [5].
A. Introduction to Space conditions and Interplanetary D. Model Organisms in Space and their Negative Effects
travel
Microbes accompany humans to the space-craft environment,
Interplanetary space travel and the colonization of Mars different studies have demonstrated the effect of microbial
encapsulates certain challenges, featuring extreme communities on crew health and craft components [43].
environments. Primarily, two prominent stressors of concern Space shuttle flight STS-95 mission aboard the ISS illustrated
are the solar radiation and microgravity. Space radiation P. aeruginosa, which formed biofilms when exposed to
consists of high-charge particles (HZE’s) and Galactic polycarbonate filters and microgravity [44]. Reference [45]
Cosmic Rays (GCR’s) having high penetrability, ionization showed that the air and surface contaminants of the LEO
and oxidation strength [1]. Also, Earth’s gravity in the Low stations can affect crew health and equipment integrity. The
Earth Orbit is not in the downward direction, like on Mars importance of microbes surviving in a confined environment
(1/3rd G) and Earth’s moon (1/6th G) which causes raises concerns like antimicrobial resistance and increased
physiological changes, that can include less bone and muscle virulence among pathogens [43]. Staphylococcus spp. have
density because of the shifting of the blood to the head region been shown to develop increased antibiotic resistance and
[47]. virulence when infected human subjects were living on the
International Space Station (ISS) [46]. These analogs provide

Seventh International Conference on Aerospace Science & Engineering 232

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

substantial data which can help space missions and more I. Repair Mechanisms of Deinococcus radiodurans and In
successful. Space
E. Synthetic Biology and Chassis Reference [11] described the efficient DNA repair in D.
radiodurans owing to its ability to repair thousands of DSB’s
Synthetic Biology is a field encapsulating living organisms
[12] and exploiting a mechanism named as the Extended
and how we can engineer them to perform different functions,
Synthesis-Dependent Strand Annealing (ESDSA). It contains
determine their applications and solve various problems
multiple copies of its chromosomes during its exponential
related to a specific industry. Synthetic biologists study,
phase (log) and stationary phase enabling it to perform the
design, re-design whole new organisms and existing
annealing process repeatedly [13]. D. radiodurans
genomes.
ultrafiltrate (without proteins) enabled it to prevent protein
In Synthetic biology, a term, referred to as a chassis which is
oxidation and carbonylation, which was also able to protect
a host organism genetically designed to maintain a stable
human Jurkat T-cells with no inflicted viability damage up to
configuration and perform functions specified by genetic
16kGy, which signifies the importance of Mn complexes in
constructs. The selection of a suitable chassis usually follows
the organism [14]. Furthermore, because of this resistance, D.
two approaches, the first one is the nearest wild type approach
radiodurans has been subjected to simulated Low Earth Orbit
and the second is standard chassis approach. The nearest wild
(LEO) conditions and stress conditions onboard the ISS to
type method involves the selection of a chassis on the basis
assess its survival [17, 18, 19].
of specific functions that it performs, these abilities of the host
are then honed in the lab to maximize its performance, the J. Aim of the Research
standard chassis approach enables the synthetic biologist to The main aim of this research is to identify the genes that are
choose from the already well-understood and highly alterable associated with the different features that can make D.
chassis. [20]. radiodurans an ideal chassis which can be used for synthetic
F. Common prokaryotes and eukaryotes Used as Chassis biology experiments. Additionally, we want to further our
understanding regarding the genes that can be used to build
Other microbes, like E. coli which was engineered as a an entirely new genetic circuit (chassis) of D. radiodurans
chassis to express type IV pili system and a synthetic peptide capable of performing novel functions and applications.
to produce Electrically Conductive Protein Nanowires (e-
PNs), ultimately producing electricity using the atmosphere II. MATERIALS AND METHODS
[32]. S. cerevisiae was also used as a chassis for the
A. Studying features of an optimal chassis
production of a wide array of biochemical and biochemistry
products [48]. Literature review was performed to identify the reported
features of an optimal chassis. A total of 3 papers were
G. Synthetic Biology and Deinococcus radiodurans selected which were screened to identify the features
D. radiodurans consists of key features that are necessary for fundamental for a chassis. These identified features were then
a functioning chassis [21, 22, 23, 7]. Genes associated with used as reference to identify relevant features in D.
these features are also well-known [22, 24, 25, 26]. radiodurans that qualify for a synthetic biology chassis.
Considering D. radiodurans, an IGEM team from Wuhan, Reported methods that could be used to incorporate the
China (XHD-Wuhan-China) synthesized a recombinant D. missing features in the microbes were also identified. [22, 21,
radiodurans, two engineered plasmids (gabY-pRADK and 23].
gcd-pRADK) were used to promote the dissolution of B. Identification of the genes associated with the features
phosphates in the Martian soil to enable optimal plant growth.
Regarding other projects, Team Botchan.Lab, Tokyo did not For comparative analysis, gene IDs, protein IDs and gene
use D. radiodurans but introduced three of its genes (recA, annotation were added to the list of chassis features.
pprM, pqqE) into E. coli to make it more resistant, although, Information regarding the genes was obtained from the
the results of the project are not recorded yet. literature [7, 22, 27] and protein knowledge databases like
Uniprot (release 2021_03), String dB (V.11.5), Kyoto
H. Introduction to Deinococcus radiodurans Encyclopedia of Genes and Genomes (KEGG) (release 99.1)
Deinococcus radiodurans was first discovered in irradiated and Panther (V.16.0).
cans of spoiled meat [6]. The Genome size of D. radiodurans
C. Comparative and Differential Expression Analysis
is 3,284,156 bp present as 2 chromosomes, 1 plasmid, 1
Megaplasmid (223kb) and 3195 genes [7]. D. radiodurans The single dataset was filtered on the basis of molecular
cells form non-sporulating colonies and are red in color due response of D. radiodurans under Low Earth Orbit (LEO)
to the presence of unique pigment known as deinoxanthin [8]. conditions. Gene expression data analysis was performed
The Deinococcaceae family consists of more than 60 using Maxquant, Fragments Per Kilobase of transcript per
recorded species [7, 9]. It is a polyextremophile bacterium Million mapped reads (FPKM) while comparing LEO and
that is extremely tolerant to desiccation, harsh temperatures, control cells in the transcript dataset, respectively [17].
ionizing radiation and variable fluences of UV-radiation, its Comparative analysis of the chassis genes with the chosen
extreme radio-resistance enables it to survive up to 15000 dataset (transcripts) was conducted in R (V.4.1.0) using plyr
kGy of radiation during stationary phase with no cell damage and dplyr packages.
or loss of viability [10]. Differential expression of extracted genes corresponding to
each chassis feature was analyzed in the transcript dataset.
Ground control (ctrl) and LEO FPKM values contained in the
dataset were used, these normalized values were divided in
the order LEO/Ctrl to obtain log2 fold change values for each

Seventh International Conference on Aerospace Science & Engineering 233

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

of the common genes (obtained through comparative After screening, the non-molecular aspects of the chassis
analysis). For the visualization of the Differentially feature list were removed and ten significant molecular
Expressed Genes (DEGs), Q-values (from the dataset) and characteristics were selected (Figure. 1). These features were
log2 fold change values calculated were used to construct selected on the basis of genes pertaining in D. radiodurans.
volcano plots, upregulated and downregulated genes were Comprehensive literature review revealed the processes that
visualized by log2FC values (Table. 1) and the corresponding can be used to incorporate the missing relevant features in D.
Gene_Locus_IDs. The data was analyzed and was used for radiodurans, multiple strains of the same microbe can be
construction in R (V.4.1.0) utilizing plyr, dplyr, generated by introducing specified deletions in the genome
EnhancedVolcano and ggplot2 packages. by a process called λ red type homologous recombination,
other methods include comparative analysis of stipulated
D. Enrichment Analysis for study of the genes
sequences. Furthermore, techniques like Repeat fusion PCR-
The chassis relevant genes selected from the literature were based gene disruption and 5-fluoroorotic acid mediated ura4-
analyzed and an enrichment analysis was performed using negative selection were also identified. Bottom-up
Gene Ontology (GO) (V.2021-08-18) and KEGG (V.99.1) approaches included sequencing of the whole genome, its
databases. Gene_Locus tags of D. radiodurans were subsequent combining through Golden Gate Assembly,
uploaded in the database for detailed cellular activities of the Gibson Assembly and Transformation-associated
genes. Cellular components, biological processes and Recombination Cloning (TAR). However, in silico tools
molecular functions of the differentially expressed genes including Bio Studio revealed the virtual assembly of the
were extracted and used to construct a combined bar plot desired genome but other than those, less computational
signifying the vital functions of the genes (see Figure. 5). The methods were mentioned.
bar plot was constructed by utilizing ggplot2 package of R
(V.4.1.0). B. 134 genes found to be associated with ideal chassis
features
E. Literature Review DEGs
A total number of ten final categories were created and a total
A brief literature search was done, a total of 10 papers were number of 134 genes were associated (see Figure. 3) (some
selected from PubMed to learn individual functions of the categories used the same genes). Categories were classified
differentially expressed genes, consequently, these detailed as biofilm formation, constitution of protein compounds, high
descriptions were analyzed further to create correlations with transformability, production of a novel gene product, survival
the chosen chassis features to understand the predicted robustness and oxidative stress resistance, low percentage of
response of D. radiodurans in space. cryptic genes, strong cell envelope, general stress resistance
genes were annotated.
III. RESULTS
Out of the 134, 10 genes were involved in biofilm formation
A. Fifteen features were identified for D. radiodurans as a in D. radiodurans, mostly they consisted of genes encoding
chassis membrane autotransporters, quorum sensing, attachment,
Eighteen chassis features were identified for D. radiodurans pathway and interaction proteins. 1 gene was pinpointed that
as a strong chassis which can be used in Synthetic Biology. contributes in the production of a cofactor involved in stress
These features consisted of properties like biofilm formation, circumvention.
constitution of protein compounds, a simplified genome, a Seven genes contributed to the transformability in D.
well-studied metabolome and genome, absence or low radiodurans, these are involved in genetic competence,
percentage of cryptic genetic segments, presence of model translocation of DNA, DNA processing and recombination of
organism attributes, an ease of handling, availability of less the environmental DNA. Additionally, 8 genes are involved
non-essential genes, low redundancy of secondary gene in the production of a novel carotenoid (deinoxanthin), most
metabolites, efficient transformability, ease of handling, of these genes encode monooxygenases, prenyltransferases,
production of a distinct gene product, survival robustness and oxidoreductases, phytoene synthases and carotenoid
oxidative stress resistance, strong cell envelope (increased desaturases. 41 genes selected regulate and promote oxidative
toughness) and low vacancy for evolutionary processes. stress resistance in D. radiodurans [22] and which abundantly
include dismutases, catalases, peroxidases, kinases and
thioredoxins. Five genes studied showed some high
evolutionary potential and as expected were involved in DNA
repair, replication and homologous recombination.
Interestingly, 3 genes were cryptic, classified as
uncharacterized genes with no annotated functions. Robust
cell envelope category encapsulated 7 known and
uncharacterized genes encoding proteins which operate
membrane maintenance, structure and protein transport.
Furthermore, 55 genes were added in the general stress
resistance category, DNA replication machinery, UV-
resistance, desiccation tolerance, DNA repair machinery and
manganese complexes genes are the major runners during the
incidence of individual stressors.
Figure 1. A list portraying all the identified chassis features that
characterize an optimal chassis. Features more efficient in D. radiodurans
(green and light green), features weak in D. radiodurans (cyan, blue and
light blue)

Seventh International Conference on Aerospace Science & Engineering 234

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

C. Differential Expression highlighted genes involved in

stress resistance
A
Comparative analysis between the 134 chassis genes and the
3164 entries in the transcript dataset yielded 134 common
entries (see Figure. 4). Out of these 134 genes, 17 were
statistically significant (q_value < 0.05) log2 fold change
(log2FC) values calculated were -2 < x < 1. Most of these
genes were involved in DNA repair and recombination, some
of them are involved in cellular response to desiccation and
radiation. Regarding functions, most of these are associated
with DNA binding, ion binding, a minor fraction show
peroxidase, helicase and endonuclease activity.
Differential Gene Expression analysis revealed a total of 25
genes, out of the 134, which were all statistically significant
i-e. q_value < 0.05. 14 genes were differentially expressed
(q_value < 0.05, -0.5 > log2 FC > 0.5) (see Table. 1).
DR_0070, DR_1998 and DR_2263 protect cells by cleaving
of peroxides, sequestration of oxidative damage and
circumvention of ionizing radiation damage by genome
reconstitution. Interestingly, analysis of the data showed the
presence of some hypothetical genes with uncharacterized B
functions. DR_0972 and DR_1021, DR_1916, DR_B0135,
DR_1709, DR_1538, DR_1913, DR_0715, DR_A0259,
DR_1172 and DR_B0092 were involved in catalyzation of
the DNA repair process, unwinding DNA and RNA by
nucleic acid binding, stabilization of the protein-RNA
complex, membrane-mediated manganese ion transfer,
maintain DNA supercoiled structure and glycosylase activity.
Although some genes passed the p_value threshold,
differential expression was not seen in the plots when log2 FC
cutoff was set to 1 and 2. Hence, the above-mentioned results
were obtained when it was set to 0.5 (see Figure. 2).
Enrichment analysis revealed the above-mentioned functions
for the 14 Differentially Expressed Genes (see Figure. 5), the
remaining 11 genes that were only significant in relevance to
the log2 FC values function by binding to heme groups and
metal ion groups, others act by binding to sulfur-iron clusters,
regulating gene expression by sensing various stimuli in the
environment and by predicted synthesis of Figure 2. A Log2FC cutoff of 0.5, 1 and p-value cutoff of 0.05 was set
to identify differentially expressed genes using a volcano plot. Log2FC
deoxyribonucleotides due to the breakdown of when set to (A) 1 did not emanate any significant results, (B) 0.5 value
ribonucleotides (ribonucleotide reductases). KEGG showed showed differential expression of 14 genes.
results for 4 genes (katA, DR_A0259, recG and G/U
mismatch-specific DNA glycosylase) that were involved in
different pathways, Gene Ontology showed results for all of
the 14 genes with their molecular functions, biological
processes and cellular components, however, significant
results were not found in GO regarding the cellular
component category.
Table 1. A table showing the Differentially Expressed Genes (DEGs) with their corresponding
q_values and log2FC values.

Seventh International Conference on Aerospace Science & Engineering 235

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

41%
Percentage of genes

30%

7% 6%
5% 4% 5%
2%

General stress Oxidative stress Cryptic genes Efficient Biofilm Production of a Less Strong cell
resistance resistance transformability formation novel gene evolutionary envelope
product processes
Chassis categories

Figure 3. A bar chart showing the percentage of genes associated with each chassis feature category

Figure 4. A heatmap showing the comparative expression of the selected 134 genes on the ground (ctrl) and exposed to LEO conditions aboard the ISS
(from the dataset)

Seventh International Conference on Aerospace Science & Engineering 236

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Figure 5. Enrichment analysis bar plot of the 14 Differentially Expressed Genes. X-axis shows the p_value (—log10 to amplify the small value), Y-axis
shows the Gene Ontology (GO) Terms and their assigned IDs.

IV. DISCUSSION
The reported findings are reported on the basis of differential
gene expression, comparative analysis and literature search
that the identified specific genes associated with each chassis
feature were differentially expressed in the LEO

Seventh International Conference on Aerospace Science & Engineering 237

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

environment, which modifies our understanding of the including base-excision repair, recombination and radio-
relationship existing between the genes associated with the resistance, additionally, this also might portray D.
features of an efficient chassis and the apparent differences radiodurans ability to allocate majority of its genes/proteins
induced on their expression under various stress conditions. in the prevention of stress and oxidation rather than growth
Genes associated with repair mechanisms and ion transport during LEO exposure.
were upregulated, interestingly, genes with the similar In different microbes, reactive oxygen species (ROS) are
functions were also downregulated. Most of the significant produced which cause DNA damage, ROS are generated in
genes belonged to the high transformability, survival consequence to desiccation, starvation, radiation etc. These
robustness and oxidative stress resistance (Figure. 3). factors, including H2O2 can also engage in reduction by Fe2+
Our findings indicate that there is a cross-talk between to form hydroxyl radicals which can cause further damage to
individual chassis features of D. radiodurans and the cellular chemistry [38]. Two of the proteins that prevent
corresponding genes associated with it, this data can provide and combat against the ROS are the gene/proteins known as
a better understanding about the individualized functions of Dps-1 and Dps-2 (DNA protection during starvation), these
these genes and how they make D. radiodurans an optimal act by binding to DNA followed by the stimulation of Fe +2
chassis for space-related applications. Although most of oxidation to Fe+3 to make it non-reactive through Fenton
DEGs were involved in DNA repair, recombination and reactions [38]. Our data (Table. 1) shows that both of these
alleviation of oxidative stress, a gene (DR_1709) related to genes/proteins (dps-1 and dps-2) were differentially
Mn (II) transport (mntH) was upregulated in the LEO expressed, of which dps-1 is upregulated and dps-2 is
conditions as compared to the ground control (ctrl) which downregulated, contradictory to the results of [38] which
suggests protein protection by circumventing the production showed the superior ability of dps-2 to combat ROS by low
of oxidative protein products as demonstrated [14]. The affinity binding to DNA and the inability of dps-1 to protect
manganese transporter gene alleviates oxidative stress which against ROS damage because of the formation of less stable
might give D. radiodurans a substantial advantage as a iron complexes. This upregulation of dps-1 as compared to
chassis over other less resistant microbes in extreme space the ground ctrl may suggest that ROS production in D.
environments or future missions to Mars. Moreover, two of radiodurans aboard the ISS was mild which caused the
the catalases (DR_1998, DR_A0259) that were significantly effective genes/proteins (dps-2) against oxidative damage to
expressed under LEO conditions, can be the repercussion of downregulate and instead managed the situation by utilizing
the negative effects of ionizing radiation on the water content a high affinity DNA binding gene/protein (dps-1), because of
of D. radiodurans causing radiolysis, consequently, aiding the cell’s low utility. Apart from this, the ground ctrl cells
the production of ROS (H2O2 and O2-), which causes these showed increased transcription of replication proteins,
catalases to catalyze the disproportionation of H2O2 as transposases and ribosomal proteins (RlmN, RpmC and
peroxidases [33]. Superoxide dismutases (sodA and sodC) RtcB) were also common in all the control replicates, this
genes in D. radiodurans are also involved in preventing observation was consistent with our data (Table. 1) which
oxidation [22], however they were absent from the list of showed the downregulation of the RNA helicase gene
DEGs, this might suggest that the LEO conditions were not (DR_B0135) which might be needed to proliferate efficiently
challenging enough for D. radiodurans survival due to which under normal conditions. One of the most important methods
a few of its chassis specific genes were upregulated. of DNA repair in D. radiodurans is base excision repair,
Although, many genes in D. radiodurans are not well Uracil-DNA glycosylases perform mismatch repair by
characterized our results identified three of them (DR_0972, excising uracil from DNA as part of the pathway, the
DR_1021, DR_1172) that were also upregulated in the LEO incorrect addition of uracil to the DNA molecule is the
environment, which may have various roles in repair consequence of error during replication or by cytosine
mechanisms, cell morphology, biofilm formation and deamination, causing G: U and A: U mismatches [39, 40].
oxidative stress resistance. These cryptic genes might provide Reference (39), described the DNA glycosylase from D.
some genetic characteristics that could contribute to D. radiodurans and how it is specific to G: U and A: U
radiodurans extended survival; however, the presence of mismatches unlike E. coli, in accordance with this study the
such genes proposes a correlation with the cryptic gene results show the upregulation of G/U and A/U mismatch
segment feature mentioned above (Figure. 1), to study this DNA glycosylase which suggests the activation of base
relationship requires additional molecular studies to better excision repair in the LEO exposed cells.
understand their tasks and how they can enable us to utilize Among other antioxidant systems in D. radiodurans like
D. radiodurans as a synthetic device. thioredoxins, glutaredoxins and peroxiredoxins, there are also
Growth and proliferation of D. radiodurans are affected in a whole family of oxidative stress resistance genes/proteins
the LEO and simulated space environment or under known as the OsmC which consists of three subtypes: Ohr,
prolonged exposures to radiation (Ionizing and UV-radiation) OsmC and YhfA [41]. Three proteins: DR_1857, DR_1538
[34], the primary genes/proteins involved in growth during of and DR_1177 encoded by D. radiodurans, are from the
recovery of D. radiodurans from ionizing radiation are also OsmC family, OsmC in D. radiodurans is conserved in other
involved in oxidation prevention, Reference [35], described Deinococcus species, these proteins are peroxidases and
DR_1172 as a desiccation tolerance gene/protein and is catalyze the disproportionation of H2O2 into its unreactive
shown as uncharacterized gene/protein in our data, is also form. Moreover, ddrB gene (DR_0070) in D. radiodurans is
shown to be contributing to growth after its exposure. DNA sensitive to IR [42] which is also validated by our results by
gyrase subunit A and ATP Dependent DNA helicase (recG) its significant upregulation, under IR exposure, DR_1538 in
have a reported role in efficient growth of D. radiodurans [36, our data (Table. 1), shows differential expression and is
37], downregulation of these two in our calculated data upregulated under LEO conditions, which demonstrates that
indicates (Table. 1) their involvement in repair mechanisms

Seventh International Conference on Aerospace Science & Engineering 238

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

this peroxidase has a direct activity in alleviating D. Deinococcus Species. International journal of microbiology, 2017,
5231424. https://doi.org/10.1155/2017/5231424
radiodurans oxidant load.
[10] Minton K. W. (1994). DNA repair in the extremely radioresistant
Although, results were statistically significant (when log2FC bacterium Deinococcus radiodurans. Molecular microbiology, 13(1),
was set to 0.5), the log2FC values of all genes were less than 9–15. https://doi.org/10.1111/j.1365-2958.1994.tb00397.x
1 which suggests that there was no substantial difference in [11] Zahradka et al. (2006). Reassembly of shattered chromosomes in
the expression of the genes when cells were exposed to LEO Deinococcus radiodurans. Nature, 443(7111), 569–573.
conditions outside the ISS (see Table. 1) [17], this could have https://doi.org/10.1038/nature05160
been due to the protective effects of ozone layer that filters [12] Krisko, A., & Radman, M. (2013). Biology of extreme radiation
resistance: the way of Deinococcus radiodurans. Cold Spring Harbor
out the harmful radiation, regardless this low intensity of perspectives in biology, 5(7), a012765.
radiation combined with other stress factors of space were https://doi.org/10.1101/cshperspect.a012765
able to initiate a low yield defensive response in the D. [13] Battista J. R. (1997). Against all odds: the survival strategies of
radiodurans. Deinococcus radiodurans. Annual review of microbiology, 51, 203–
Additionally, the diversity of our results was limited by the 224. https://doi.org/10.1146/annurev.micro.51.1.203
smaller number of genes that were associated with their [14] Daly, M. J et al. (2010). Small-molecule antioxidant proteome-shields
in Deinococcus radiodurans. PloS one, 5(9), e12570.
respective chassis features, there may have been other genes https://doi.org/10.1371/journal.pone.0012570
that contribute to the properties of D. radiodurans as a [15] Lim, S., Jung, J. H., Blanchard, L., & de Groot, A. (2019). Conservation
chassis. Although, computational data shows recognizable and diversity of radiation and oxidative stress resistance mechanisms in
results, some of these DEGs have multiple roles which might Deinococcus species. FEMS microbiology reviews, 43(1), 19–52.
https://doi.org/10.1093/femsre/fuy037
explain why the magnitude of their differential expression
[16] Tian, B., Xu, Z., Sun, Z., Lin, J., & Hua, Y. (2007). Evaluation of the
was low, to understand this phenomenon, a study needs to be antioxidant effects of carotenoids from Deinococcus radiodurans
designed that explores specialized functions of individual through targeted mutagenesis, chemiluminescence, and DNA damage
genes and how these genes are connected with other elements analyses. Biochimica et biophysica acta, 1770(6), 902–911.
of the genome. Furthermore, the information gathered in this https://doi.org/10.1016/j.bbagen.2007.01.016
research can be used to explore D. radiodurans in space [17] Ott, E et al. (2020). Molecular repertoire of Deinococcus radiodurans
after 1 year of exposure outside the International Space Station within
conditions while considering each of its features as a chassis. the Tanpopo mission. Microbiome, 8(1), 150.
https://doi.org/10.1186/s40168-020-00927-5
ACKNOWLEDGMENTS [18] Panitz, C., Frösler, J., Wingender, J., Flemming, H. C., & Rettberg, P.
We thank Arsalan Riaz (Precision Medicine Lab) and Abdus (2019). Tolerances of Deinococcus geothermalis Biofilms and
Planktonic Cells Exposed to Space and Simulated Martian Conditions
Salam (Precision Medicine Lab) for providing their expertise in Low Earth Orbit for Almost Two Years. Astrobiology, 19(8), 979–
regarding the R program, Laiba Shah (Precision Medicine 994. https://doi.org/10.1089/ast.2018.1913
Lab) for providing helpful insight throughout our research [19] Tsai, C. H., Liao, R., Chou, B., & Contreras, L. M. (2015).
and Hina Zamir (Precision Medicine Lab) for proofreading Transcriptional analysis of Deinococcus radiodurans reveals novel
the manuscript. small RNAs that are differentially expressed under ionizing radiation.
Applied and environmental microbiology, 81(5), 1754–1764.
https://doi.org/10.1128/AEM.03709-14
REFERENCES
[20] Kuldell, N. (2015). Biobuilder. Sebastopol, CA: O'Reilly.
[21] Chi, H et al. (2018). Engineering and modification of microbial chassis
[1] Afshinnekoo et al. 2020. Fundamental Biological Features of for systems and synthetic biology. Synthetic and systems
Spaceflight: Advancing the Field to Enable Deep-Space Exploration. biotechnology, 4(1), 25–33.
Cell, 183(5), 1162–1184. https://doi.org/10.1016/j.cell.2020.10.050 https://doi.org/10.1016/j.synbio.2018.12.001
[2] Frösler, J., Panitz, C., Wingender, J., Flemming, H. C., & Rettberg, P. [22] Liu, Y., Liu, L., Li, J., Du, G., & Chen, J. (2019). Synthetic Biology
(2017). Survival of Deinococcus geothermalis in Biofilms under Toolbox and Chassis Development in Bacillus subtilis. Trends in
Desiccation and Simulated Space and Martian Conditions. biotechnology, 37(5), 548–562.
Astrobiology, 17(5), 431–447. https://doi.org/10.1089/ast.2015.1431 https://doi.org/10.1016/j.tibtech.2018.10.005
[3] Mancinelli, Rocco & Klovstad, Melisa. (2000). Martian soil and UV [23] Kim, J et al. (2016). Properties of alternative microbial hosts used in
radiation: Microbial viability assessment on spacecraft surfaces. synthetic biology: towards the design of a modular chassis. Essays in
Planetary and Space Science. 48. 1093-1097. 10.1016/S0032- biochemistry, 60(4), 303–313. https://doi.org/10.1042/EBC20160015
0633(00)00083-0.
[24] Wang, Y., Xu, G., Wang, L., & Hua, Y. (2019). Distinct roles of
[4] Horneck et al. 2003. Critical issues in connection with human missions Deinococcus radiodurans RecFOR and RecA in DNA transformation.
to Mars: protection of and from the Martian environment. Advances in Biochemical and biophysical research communications, 513(3), 740–
space research : the official journal of the Committee on Space 745. https://doi.org/10.1016/j.bbrc.2019.04.042
Research (COSPAR), 31(1), 87–95. https://doi.org/10.1016/s0273-
[25] Chen, Z., Tang, Y., Hua, Y., & Zhao, Y. (2020). Structural features and
1177(02)00662-2
functional implications of proteins enabling the robustness of
[5] Horneck, G., Klaus, D. M., & Mancinelli, R. L. (2010). Space Deinococcus radiodurans. Computational and structural biotechnology
microbiology. Microbiology and molecular biology reviews : MMBR, journal, 18, 2810–2817. https://doi.org/10.1016/j.csbj.2020.09.036
74(1), 121–156. https://doi.org/10.1128/MMBR.00016-09
[26] Ithurbide et al. (2020). Natural Transformation in Deinococcus
[6] Anderson et al. (1956). Studies on a radio-resistant micrococcus. I. radiodurans: A Genetic Analysis Reveals the Major Roles of DprA,
Isolation, morphology, cultural characteristics, and resistance to DdrB, RecA, RecF, and RecO Proteins. Frontiers in microbiology, 11,
gamma radiation. Food Technology, 10, 575-578. 1253. https://doi.org/10.3389/fmicb.2020.01253
[7] Gerber, E et al. (2015). Deinococcus as new chassis for industrial [27] Ott, E et al. (2017). Proteometabolomic response of Deinococcus
biotechnology: biology, physiology and tools. Journal of applied radiodurans exposed to UVC and vacuum conditions: Initial studies
microbiology, 119(1), 1–10. https://doi.org/10.1111/jam.12808 prior to the Tanpopo space mission. PloS one, 12(12), e0189381.
[8] Laurent Lemée, Evelyne Peuchant, Michel Clerc, Martin Brunner, https://doi.org/10.1371/journal.pone.0189381
Hanspeter Pfander. Deinoxanthin: A New Carotenoid Isolated from [28] Rothfuss, H., Lara, J. C., Schmid, A. K., & Lidstrom, M. E. (2006).
Deinococcus radiodurans. Tetrahedron / Tetrahedron [London]; Involvement of the S-layer proteins Hpi and SlpA in the maintenance
Tetrahedron Suppl; Tetrahedron (London), 1997, 53 (3), pp.919-926. of cell envelope integrity in Deinococcus radiodurans R1.
⟨10.1016/S0040-4020(96)01036-8⟩. ⟨hal-00155200⟩ Microbiology (Reading, England), 152(Pt 9), 2779–2787.
[9] LaGier M. J. (2017). Predicted Cold Shock Proteins from the https://doi.org/10.1099/mic.0.28971-0
Extremophilic Bacterium Deinococcus maricopensis and Related

Seventh International Conference on Aerospace Science & Engineering 239

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
[29] Jiang, S et al. (2017). DrwH, a novel WHy domain-containing [47] Demontis, G. C., et al. (2017). Human Pathophysiological Adaptations
hydrophobic LEA5C protein from Deinococcus radiodurans, protects to the Space Environment. Frontiers in physiology, 8, 547.
enzymatic activity under oxidative stress. Scientific reports, 7(1), 9281. https://doi.org/10.3389/fphys.2017.00547
https://doi.org/10.1038/s41598-017-09541-2 [48] Jouhten, P et al. (2016). Yeast metabolic chassis designs for diverse
[30] Farci, D et al. (2014). New features of the cell wall of the radio-resistant biotechnological products. Scientific reports, 6, 29694.
bacterium Deinococcus radiodurans. Biochimica et biophysica acta, https://doi.org/10.1038/srep29694
1838(7), 1978–1984. https://doi.org/10.1016/j.bbamem.2014.02.014
[31] Zhang, L., Yang, Q., Luo, X., Fang, C., Zhang, Q., & Tang, Y. (2007).
Knockout of crtB or crtI gene blocks the carotenoid biosynthetic
pathway in Deinococcus radiodurans R1 and influences its resistance
to oxidative DNA-damaging agents due to change of free radicals
scavenging ability. Archives of microbiology, 188(4), 411–419.
https://doi.org/10.1007/s00203-007-0262-5
[32] Ueki, T et al. (2020). An Escherichia coli Chassis for Production of
Electrically Conductive Protein Nanowires. ACS synthetic biology,
9(3), 647–654. https://doi.org/10.1021/acssynbio.9b00506
[33] Kobayashi, I et al. (2006). Characterization of monofunctional catalase
KatA from radioresistant bacterium Deinococcus radiodurans. Journal
of bioscience and bioengineering, 101(4), 315–321.
https://doi.org/10.1263/jbb.101.315
[34] Kawaguchi, Y et al. (2020). DNA Damage and Survival Time Course
of Deinococcal Cell Pellets During 3 Years of Exposure to Outer Space.
Frontiers in microbiology, 11, 2050.
https://doi.org/10.3389/fmicb.2020.02050
[35] Liu, Y et al. (2003). Transcriptome dynamics of Deinococcus
radiodurans recovering from ionizing radiation. Proceedings of the
National Academy of Sciences of the United States of America, 100(7),
4191–4196. https://doi.org/10.1073/pnas.0630387100
[36] Wu, Y., Chen, W., Zhao, Y., Xu, H., & Hua, Y. (2009). Involvement of
RecG in H2O2-induced damage repair in Deinococcus radiodurans.
Canadian journal of microbiology, 55(7), 841–848.
https://doi.org/10.1139/w09-028
[37] Kota, S., Charaka, V. K., & Misra, H. S. (2014). PprA, a pleiotropic
protein for radioresistance, works through DNA gyrase and shows
cellular dynamics during postirradiation recovery in Deinococcus
radiodurans. Journal of genetics, 93(2), 349–354.
https://doi.org/10.1007/s12041-014-0382-z
[38] Reon, B. J., Nguyen, K. H., Bhattacharyya, G., & Grove, A. (2012).
Functional comparison of Deinococcus radiodurans Dps proteins
suggests distinct in vivo roles. The Biochemical journal, 447(3), 381–
391. https://doi.org/10.1042/BJ20120902
[39] Moe, E., Leiros, I., Smalås, A. O., & McSweeney, S. (2006). The
crystal structure of mismatch-specific uracil-DNA glycosylase (MUG)
from Deinococcus radiodurans reveals a novel catalytic residue and
broad substrate specificity. The Journal of biological chemistry, 281(1),
569–577. https://doi.org/10.1074/jbc.M508032200
[40] Lindahl T. (1974). An N-glycosidase from Escherichia coli that
releases free uracil from DNA containing deaminated cytosine
residues. Proceedings of the National Academy of Sciences of the
United States of America, 71(9), 3649–3653.
https://doi.org/10.1073/pnas.71.9.3649
[41] Shin, D. H et al. (2004). Structure of OsmC from Escherichia coli: a
salt-shock-induced protein. Acta crystallographica. Section D,
Biological crystallography, 60(Pt 5), 903–911.
https://doi.org/10.1107/S0907444904005013
[42] Norais, C. A., Chitteni-Pattu, S., Wood, E. A., Inman, R. B., & Cox, M.
M. (2009). DdrB protein, an alternative Deinococcus radiodurans SSB
induced by ionizing radiation. The Journal of biological chemistry,
284(32), 21402–21411. https://doi.org/10.1074/jbc.M109.010454
[43] Taylor P. W. (2015). Impact of space flight on bacterial virulence and
antibiotic susceptibility. Infection and drug resistance, 8, 249–262.
https://doi.org/10.2147/IDR.S67275
[44] McLean, R. J., Cassanto, J. M., Barnes, M. B., & Koo, J. H. (2001).
Bacterial biofilm formation under microgravity conditions. FEMS
microbiology letters, 195(2), 115–119. https://doi.org/10.1111/j.1574-
6968.2001.tb10507.x
[45] Novikova N. D. (2004). Review of the knowledge of microbial
contamination of the Russian manned spacecraft. Microbial ecology,
47(2), 127–132. https://doi.org/10.1007/s00248-003-1055-2
[46] Schiwon, K et al. (2013). Comparison of antibiotic resistance, biofilm
formation and conjugative transfer of Staphylococcus and
Enterococcus isolates from International Space Station and Antarctic
Research Station Concordia. Microbial ecology, 65(3), 638–651.
https://doi.org/10.1007/s00248-013-0193-4

Seventh International Conference on Aerospace Science & Engineering 240

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Event Rate of Extreme Mass Ratio Inspiral Systems

in M32 Galaxy
Asad Ali1,2 Rafia Sarwar1,2
1Department 1 Department
Muhammad Waqas 1,2 of Applied of Space Science
2Space and Astrophysics Research Lab
1Department of Space Science Mathematics & Statistics
2Space and Astrophysics Research
2Space and Astrophysics Research Lab (SARL), National Center of GIS and
(SARL), National Center of GIS and Lab (SARL), National Center of GIS Space Applications (NCGSA),
Space Applications (NCGSA), and Space Applications (NCGSA), Institute of Space Technology, Islamabad
Institute of Space Technology, Islamabad Institute of Space [email protected]
[email protected] Technology, Islamabad
[email protected]

Abstract—After growing consensus of scientific fraternity velocities are dependent on mass of central MBH, and their
that, there exist a black hole (MBH) in the center of every mutual scaling relationship is represented by M• – σ [1].
galaxy. The surrounding environment of MBH in the galactic
center contains a very dynamic presence of stellar cluster Interaction between stars & MBH in galactic center occur either by
giving rise to systems comprising of a massive black hole gas produced by stars which then accreted onto the MBH or stars
(MBH) and a population of stellar-mass compact-object (CO), can directly plunge into MBH following radial orbit or inspiral
typically stellar-mass black hole (BH), neutron stars (NS) and gradually loosing angular momentum in form of GWs & giving
white dwarfs(WD). As the COs become gravitationally bound rise to EMRI systems. Both interactions contribute to overall mass
with the central MBH forming extreme mass ratios binary and spin of MBH. LISA, space-based mission in collaboration with
systems, having prolonged 104 to 105 cycles, which are European Space Agency (ESA) will be realized in 2034, which will
anticipated candidates for future space-born Laser have a sensitivity range to detect these GWs producing events in
Interferometer Space Antenna (LISA) to detect gravitational milli-Hertz [10-4 to 10-1 Hz ] window [2], produced by MBH of
waves from these inspiralling binaries. These CO-MBH billion solar masses, orbited by few solar mass COs.
systems are producing complex gravitational waves (GW)
during their evolution with weak strain amplitudes, leading to On the available observed data, velocity dispersion M• – σ and
the last stable orbit around the MBH. We inspect the luminosity relation, 0.01 Mpc−3 is the MBH density in the local
relationship between signal–to–noise ratios (SNRs) and universe. In Milky Way, M31 and M32 galaxies, observationally it
informative parameters of these gravitational wave generating is of scientific knowledge that they possess MBH and a compact
binary sources, and developed a generalized linear model stellar nucleus in the galactic center. Since the mass of the COs are
inferring their properties. Intrinsic parameters of Analytical 100 - 102 M⊙ , their mass ratio is typically q = m / M ∼10−4 - 10−5 .
Kludge (AK) waveform model were restrained, by employing That’s why they are termed as extreme-mass-ratio-inspirals
the well-established understanding of the MBH to extrapolate (EMRIs), these binary systems are likely to radiate GWs with in
the scaling relation, which envisages a fiducial fit for back–of– most sensitive frequency range of LISA. This places them in a
envelop computations of SNRs. We compute the averaged most probable GWs radiating sources detectable by the LISA
probability of 9.39 EMRI events detectable in LISA mission mission. EMRIs systems evolve right in the heart of galaxy, near to
lifetime, to occur in M32 galaxy, considering the sensitivity of MBH, where sufficient number of COs and stellar remnants are
the LISA detector. available for capture by the black hole and cusp of stars has been
formed around them, even if they have gone through any recent
Keywords—Compact Objects (COs), Analytical Kludge merger. Strain sensitivity of LISA facility lies at 3 mHz, most of
Waveform(AKW), Laser Interferometer Space Antenna (LISA), the potential sources for this frequency spot should have a MBH of
Extreme Mass Ratio Inspirals(EMRIs), Massive Black Hole the order of 106 M⊙. In EMRI systems, COs mass range is
(MBH) normally 4 orders of magnitude smaller than the mass of the
SMBH. This can be considered as the perturbation, to background
I. INTRODUCTION space-time of the central object. Every black holes is unique, and
can be characterized by three parameters mass, spin and charge.
Over the last few decades, astrophysical understanding of Since electromagnetic repulsion is much stronger than gravitational
galactic centers in galaxies has increased dramatically. attraction (about 40 orders of magnitude), assuming that
Giving consideration to space based telescope like Hubble astrophysical black holes have negligible net charge, we consider
Space Telescope (HST) and ground based telescopes the Kerr metric, where black holes are described only by their mass
which improved their observation results by using adaptive and spin. Orbital inclination by the spin axis leads to considerable
optics, they have studied kinematics of stars or gas in the effect as Kerr MBHs involves axisymmetric geometry contrary to
exotic sub parsec region for distant and milli parsec region of symmetric geometry of Schwarzschild MBHs. Stellar dynamics
Milky Way galaxy. Remarkable conclusion is drawn from play most important role in EMRI system evolution, as capture of
these results that a dark compact object sitting in the center COs principally depends on stellar dynamics in vicinity of MBH
of every galaxy, most probable a MBH of 106 to 109 M⊙. whilst orbiting it with high velocities, Strong Mass Segregation
(SMS) and scrambled eccentricities become very important too.
The nucleus of galaxy contain star cluster, ranging from 107
COs follow eccentric orbits, with relativistic dynamics, till the last
to 108 stars with in a few parsec around MBH. In nucleus of stable orbit(LSO), but these parameters depend upon the initial
galaxy, star densities are around 106pc−3 and their relative moment of EMRI system when CO is initially captured by MBH.
velocities vary from 100 kms−1 to 1000 kms−1. These As EMRI evolve and number of orbits go high, strength of GW
signal increases and it can be observed effectively till Last Stable

Seventh International Conference on Aerospace Science & Engineering 241

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Orbit (LSO). COs orbit mostly in the strong gravitational field cusp-formation around MBH (Babak et al., 2017) in galactic
before falling into MBH. In this process of orbiting and inspiralling centers, if a galaxy has gone through a recent merger such
CO dissipates orbital energy and angular momentum, emitting that the cusp has disrupted around its central MBH, those
continuous GWs of the order of milli-Hertz, detectable by LISA systems are out as a candidate of EMRI formation. We are
band.
We will use foundation work of [4] inspiral orbit using waveform
studying M32 galaxy, which observably has found to have a
simulation by Barack and Cutler, 2004 [5] at the later orbital stages cusp formed around its central MBH (Amaro-Seoane et al.,
of EMRIs till before CO falls into MBH. EMRIs are formed mostly 2007; T.L, 1998), and mass of M32 MBH is 2.5 × 10 6 M⊙
from, eccentric trajectories emerging from scattering processes and (Karachentsev et al., 2004). The z - independent scaling
N-body gravitational effect of neighboring stellar species. Yet, few relation of mass function is (Babak et al., 2017).
of the COs are ejected to higher orbits by these dynamical
processes that may never arrive at the formation stage of an EMRI.
The expeditious waveform amplitude of GW of EMRI waveform 𝐵
will be orders of magnitude less than that of detectors noise, 𝑑𝑛⁄𝑑 (𝑙𝑛𝑀) = 𝑛𝑜 (𝑀⁄3 × 106 𝑀⊙ ) (1)
nonetheless, detection of few such events the mass and spin
parameters of the MBH. It will demystify MBH, and it can be here M is the mass of MBH, for LISA detectable MBH
wholly described in terms of no-hair theorem. range no is .0002 Mpc−3 , and B is 0.3, considering lower
Gravitational waveforms from such inspiraling sources have limit of EMRI event rate.
significant prospective for gravitational physics.
B. Intrinsic Rates of Stellar Remnants around MBHs
Primary goal of this project is to compute event rates of EMRIs in Stellar remnants (WDs, NSs or stellar mass BHs) orbiting
M32 (NGC221) galaxy, anchoring a MBH in its galactic nucleus. around MBH in GC, their properties contribute to formation
In introduction (1) by primitive deliberation of gravitational wave
of EMRI processes. MS stars nearing SMBH faces tidal
source extreme mass-ratio inspirals (EMRIs) genesis, knowing that
disruption and do not resist the tidal forces by SMBH. NSs
such exotic source is the most probable candidate detectable, by
space-based laser interferometer space antenna (LISA). LISA
which have evolved through different evolutionary paths
mission with the help of European Space Agency (ESA) will be reflect distinctive signatures through dissimilar distribution
realized in 2034 and it will have detection window of the entire sky peak and mass cutoff values. Binary NSs systems and
probing GWs emitting sources up to the red-shift of the order z = neutron star-white dwarf systems show consistent respective
20. New Gravitational wave Observatory (NGO) will revealing a peaks at 1.33⊙ accretion of 0.22⊙ and 1.55⊙ suggesting
hidden Universe where electromagnetic spectrum principally significant mass which has occurred during the spin-up
cannot probe.We will establish an understanding of the phase. where as the simulated BH mass range was
environment around MBH, surrounded by COs of different mass [.5−100]M⊙ (Abbott et al., 2020). Cusp isformed around
range and the impact of various astrophysical processes on the dynamically relaxed clusters around MBH, heavier species
event rates of EMRIs, and computing formalism for event rate in sink closer to MBH, resulting in SMS (Amaro-Seoane and
section (2). Further in (3), we will discuss our approach towards Preto, 2011) and become the most probablysource of EMRI
EMRI waveform construction by playing the reduced parameter formation. WDs mass range is fraction of solar mass,
space for our own good in context with computation ease. Signal
thereforewe set the range of COs to be [.5−100]M⊙ around
processing, Sensitivity of detector and spectral density assessment
to obtain Signal to noise ratio will be discussed in (4). Section (5) MBH to be very wide, to understandtheir effect on resulting
will include the results and future implications of this work. GWs sources EMRIs. Intrinsic rate of COs around MBH is
given by Power law:
II. EVENT RATE IN M32 GALAXY 𝑎
𝑅 = 𝑅𝑜 (𝑀⁄106 𝑀⊙ ) (2)
7
Observations suggest the presence of 10 M⊙ of stars in
galactic center of M32 galaxy by closely studying its
luminous density profile. These stellar density profiles only where a = {−.15, −.25, −.25} with event rates Ro = {400, 7,
represent the brighter species present in the galactic center, 20}Gyr−1 for BHs, NSs and WDs respectively [9]. Given
where as COs like Black Holes (BH) which are heavier of estimated intrinsic rate R are results of MCMC simulations and
does not include the direct plunges. Direct plunges refer to the
all COs, and orbit very close to MBH are not included captures in which CO scatters to the radial orbits and plunges
because of obvious reasons. But they are potential directly without undergoing the inspiral stages.
candidates of EMRI forming systems.
EMRI formation depends upon the astrophysical processes C. Formalizing Event Rate
like Strong Mass Segregation, resonant relaxation, tidal To formalize event rate we will be needing number density
disruption, triaxiality and density profiles of galactic centers. of comoving MBHs, EMRIs formation intrinsic rate, and
Event rates of EMRI counts in any galaxy is very uncertain, mission life time. Whilst running our wave form models we
as we are studying M32 galaxy exclusively, we will use very chose mission life time to be 2 years. M32 galaxy exist at
optimistic astrophysical presumptions. .77MPc from Milky way galaxy, having redshift of z =
0.00017 in distance, we ignore this red shifted component of
A. MBH Population in Universe comoving volume for simplicity and negligible effect. Spins
MBH masses and spins are crucial to understand in the quest of MBH is still highly uncertain, hence, the ∫p(a)da is
of Event rate estimation. MBH possessing higher spin normalized to 1 which is the probability density function pdf of
distributions will produce more EMRIs. In LISA detectable spin parameter, ranging from 0 to 1. So,
range of MBH. Distribution of these MBH evolve as a
function of spin and masses co-moving at particular red-shift 𝑀𝑚𝑎𝑥
𝑁𝐸𝑀𝑅𝐼 = 𝑡𝑙𝑖𝑓𝑒 × ∫𝑀𝑚𝑖𝑛 𝑅 × 𝑑𝑛⁄𝑑 (𝑙𝑛𝑀)(𝑑(𝑙𝑛𝑀)) (3)
z. Estimation of such galaxies in local universe which host
MBH in their centers in co-moving volume of local universe
is a challenging problem. Formation of EMRIs depend upon where Mmin = 104 M⊙ and Mmax = 107 M⊙ [10]. Dirac Delta
function is applied to the above expression so that limit of

Seventh International Conference on Aerospace Science & Engineering 242

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

integral converges only when M approaches M• of M32 pericenter), Φ0 (initial azimuthal orbital phase angle) and α 0
galaxy and zero elsewhere. is the initial azimuthal angle of orbital angular momentum
[5].
III. EMRIS IN M32 GALAXY It is still a large Parameter space as far as waveform
Essential features of General Relativity (GR) has been generation is concerned by grid based or MCMC methods.
contemplated by using Analytical Kludge Waveform Model Global mode of such multi dimension parameter space is
(AK). This waveform model AK, takes in account of lowest computationally very intensive. Our project is specifically
order quadrupole moment that retains effect of radiation for M32 galaxy, so, we have information from observational
reaction & effect of orbital precession. AK waveform model astronomy, like mass, sky location and distance to MBH, in
improves the rate at which waveform evolves and yet that way parameter space is left with 10 parameters only.
reflecting its all characteristic features. Additionally, the
A. Model Parameters
probability of EMRI event happening strongly depends upon
capture processes of COs, which can only be determined by Our work surrounds Event rate EMRIs of M32 galaxy,
dynamical astrophysical stellar processes going on in the where MBH of M• = 2.5×106 M⊙ sits in the galactic center
closest vicinity of MBH, orbiting around it with high [18]. Because we are specific to a single galaxy, we can
velocities and random eccentricities. As LISA constellation further reduced our EMRI parameter space, by limiting the
itself will be cart-wheeling and orbiting, it will induce parameters of locality, mass of MBH, distance to the source,
additional Doppler modulations to the receiving signal, that this reduces the complexity of waveform model. M32
put together reduced mode frequency approximation galaxy exists at .77MPc, having redshift of z = .00017 in
realized by (Cutler, 1998). We choose this model distance, ignoring z-component for simplicity and its
considering its computational efficiency in comparison to negligible effects in comoving volume. Let the set of
the available range of waveform models. parameters be, λ = {λ0, λ1, λ2,...., λz } where superscript ”z”
is the tell of parameter number. Selection range for
AK waveform model considers COs instantaneous orbital parameters is as follows:
trajectories as Newtonian in MBH vicinity, assuming
background manifold to be flat. In start of EMRI evolution,  Using stellar kinematics, Hubble Space Telescope
trajectories of CO orbits are taken to be Keplerian ellipses, (HST) imaging, and Jeans Anisotropic Models
and in later stages of EMRIs, relativistic effects are mass of MBH is M• of 2.5 ± 0.5 × 10 6 M⊙ and
introduced stimulated through PN approximation equations. astronomical distance D• = .77Mpc [18].
Appended relativistic effects take account of radiation  Spin ”a” is key intrinsic parameters with the
reaction which is (circularizing and inspiralling) affiliated to strongest impact on the inner-most stable orbit.
CO path in phase–space course and precessions of orbital hence, to nullify the expected discrepancy in the
plane through out formation space. magnitude of spin angular momentum ã = S/M2 ,
we take it to be uniformly distributed U [0, 1].
GWs can be represented completely in three coordinate  We use the J2000.0 coordinates: Ecliptic latitude θs
systems, Barycentric ecliptic coordinate system around solar and ecliptic longitude φs are fixed to be -
system, coordinates of MBH at galactic center, and co- 40.86638889 and 10.67541667 respectively.
rotating coordinates of detector which principally uses low-  Mass of COs intrinsically defined by its type, e.g
frequency approximation proposed by (Cutler, 1998) [11] BHs, NSs, WDs or low mass Main Sequence (MS)
respectively. After choosing source coordinate, Waveform stars. We introduced a randomized uniform
signal is constructed in h+,× polarizations that is based upon
distribution of μ within U[.5, 100]M⊙.
Peters–Mathews work, these effects are then translated into
strain amplitude hI,II and with the antenna response functions  Initial eccentricity was taken in uniform
F+,×, sky location and polarization angle of source is taken in distribution between U[0.15, 0.25].
Barycentric coordinate system.
 Initial orbital frequency νo at t = 0 of CO’s capture
Full representation of EMRI waveform can be done by 17 by MBH is taken to be uniform between ∼ 2 − 5
physical parameters including both intrinsic–extrinsic mHz.
parameters. If we neglect the spinning effects of CO, 14
parameters are left to consider and are expressed as λAK=  Spin vectors are taken to be isotropic and
{ν0, M, μ, e0 , γ0 , Φ0 , α0 , θS , φS , i, ã, θK , φK , DL}. They symmetrically distributed. Three angles
contribute to the evolution of waveform. Parameters are corresponding to the initial phase space trajectory
chosen, when capture of CO occur by MBH at t = 0, and naively, taken between U[0, 2π].
subscript zero implies that. All of these parameters are
chosen to be dimensionless. Extrinsic parameters are Initial orbital frequency are set in such a way that EMRI
dependence on the location of source and orientation of system evolve and inspirals up to the Last Stable Orbit
detector. Where Polar angle of MBH spin axis is θK and (LSO) for a fixed duration of 2 years. Bank of sampling is
azimuthal angle of MBH spin axis is φK . θS and φS are the obtained to have statistically significant results. For each
source latitude and longitude in ecliptical coordinate system. sample, We generated randomized set of parameters as
Distance to the source is DL which is in Mpc. Intrinsics discussed above. We used the noise model of (Barack and
parameters are ν0 (Initial orbital frequency), M (Mass of Cutler, 2004) and sampled 105 mission realizations.
MBH), μ (Mass of CO), e0 (Initial eccentricity), i (orbital
inclination angel) , ã (Spin of MBH), γ0 (initial direction of IV. SIGNAL ANALYSIS

Seventh International Conference on Aerospace Science & Engineering 243

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

If strain amplitude of waveform signal is s(t), and it is the In our research to estimate EMRI event rate, we studied
time-dependent linear sum of GW signal denoted by h(t) and LISA detectors and its techniques for detection of GWs,
noise n(t). We are reviewing the signal analysis in the summarized astrophysics of EMRIs and processes effecting
prospect of classic LISA design model by (Barack and its formation, evolution and event detection formalization.
Cutler, 2004) [5], were noise n(t) is taken to be non- We broaden our signal analysis base and reviewed LISA
correlated and normal (Gaussian), having zero mean. sensitivity curve and power spectral density analysis to
Modifications for LISA will require contributions of estimate SNRs. Theoretical waveforms model for EMRIs
updated analytic noise distributions. We are using low visualizations were performed by Analytical Kludge
frequency approximation in which triangular formation of Waveform model by (Barack and Cutler, 2004). Code we
LISA detector is assumed to be indistinguishable 2-armed used to generate bank of waveforms were based on previous
detectors, held perpendicular to each other, and will detect work of (Ali et al., 2012).
strain h+,× signal. This approximation considers LISA
constellation to be statically at equal distance (Cutler,1998) A. Estimates of SNR with EMRI Parameters
[11]. If strain amplitude is hI,II (t) detected at each detector is
I and II, then these signals at each detector h I,II are Initial orbital frequencies has positive correlation with
effectively Michelson interferometers response, and emerge SNRs. We express SNR in logarithmic scale as it is a
out as linear sum of perturbative polarization amplitude positive quantity and by log plots we can better understand
A+,×(t) and geometric antenna response functions F I,II (t) of the behavior of SNR with sensitive parameters of EMRIs.
detector’s channels I and II, respectively, and the response We will be expressing statistical indicators like R2 index
functions FI,II (t) depends upon the sky location of the and p − value, which will tell the goodness of regression
source translated in detector’s coordinates. model. Initial orbital frequency has R 2 index of 0.762 with
Waveform model templates are constructed on the basis of very significant p-value of p = 0. This linear relation of
chosen astrophysical model and its parameter set. Inner initial orbital frequency and SNR marks out that how orbits
product tells how the signal is variant in comparison to a of COs around MBH at time of its capture will affect SNRs.
template and this quantification is called SNR [19]. Inner This strong inter-dependency can also be observed in scaling
product function evaluates the amount of weight by which relation of SNRs where orbital frequencies are translated
the two time-dependent Fourier transformed signals are into semi-major axis a to compliment the scaling relation of
different is called optimal filtering, it solely depends upon SNRs. Next most sensitive parameter is mass of CO,it has a
sensitivity of the detector, which is lower R2 = 0.011 value than initial orbital frequency and has
p − value < 1.81 × 10 −295 . From rest of the parameters,
∞
𝜌2 [ℎ] = 4ℜ(∫0 ℎ∗ (𝑓)ℎ (𝑓)⁄𝑆𝑛 (𝑓) 𝑑𝑓) (4) it is clear that they don’t have any correlation SNR, and are
tabulated in Table 1 with an inadequate level of significance.
where h(f) representing the measure of Fourier Transformed
(FT) waveform signal detected, Sn(f) defined one-sided Statistical density distributions and Kernel Density
Power Spectral Density (PSD) of the detector’s noise and Estimation (KDE) for the set of parameters is given in As
ρ(h) is SNR. If SNR is higher than 20, then we will consider clear from the scatter plots, orbital frequencies relate to the
that signal to be detectable. SNR in a linear fashion. Also, cosine angles of orbital
inclination expresses the expected bi-modal distribution, as
A. LISA Sensitivity the parameter was constrained between 0 and 2π and has
We have used frequency dependent LISA sensitivity curve sinusoidal behavior. Other parametric distributions are non–
approximated by Cutler (2004) which is primarily being Gaussian and non–consequential. The heat map shows no
contributed by single-link optical metrology noise, single multicollinear relation of substantial degree is observed
test mass acceleration noise and galactic confusion noise. amongst the parameters, except orbital frequency and
normalized CO mass has a slight negative correlation.
V. RESULTS AND APPLICATIONS

Seventh International Conference on Aerospace Science & Engineering 244

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig 1: Scatter plots and marginal histograms of independent

parameters with dependent SNR ρ. Red line is used for regression
line establishing a correspondence between parameters and SNR.
(continued from previous page)

Fig 3: Corner Plot of EMRIs parameter and their mutual Correlogram:

Statistical Density Contour Variation, diagonal represent 2-D posterior
Fig 2: Heat-map and Correlation Coefficient Matrix, Negative correlation distribution.
exists between μ/M and νo , Clearly, SNR ρ is highly sensitive to μ/M and νo

Seventh International Conference on Aerospace Science & Engineering 245

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
TABLE I. STATISTICAL P ARAMETRIC CORRELATION our results are mass normalized. It indicates that mass
ONTO THE SNR normalized SNR of an EMRIs still, slightly, varies
with CO’s mass. Strong Mass Segregation (SMS) may or
Parameters R-squared p-values may not be the reason of this effect. eo increasingly
νo .762 0 contributes to SNR with relatively stronger regression
coefficient. Spin remains highly uncertain parameter
ln μ .011 1.81×10−295 proportionally relates with positive coefficient but for higher
eo .0021 7.28×10−57 spin distributions the term will not significantly impact
SNR. Likewise, bi modal distribution of inclination angle
γo 6.89×10-6 .363 confines, typically, at maxima’s will remain trivial to the
φo 7.65×10-7 .7621 best-fit.
05
Cos(i) 2.03×10 .1189 C. EMRI Events In LISA Mission Life Time
αo 8.95×10 07
.743
In upper panel of Figure 6.5 express SNR distribution of
S/M2 = â 4.58×10-06 .458 probabilistic EMRIs for 105 mission realizations in LISA
θk 2.57×10 −07 .8605 sensitivity band. ρ = 20, was set to be the threshold limit of
detection, all EMRIs are louder then this limit and hence
φk .000116 .00019 detectable. AS M32 is very near to earth so it is an expected
If p<<0.05, indicates strongly significance, if p ≤ 0.05 it is just significant. trend to observe D• = .77Mpc. The expectation of high
maximum SNR is observed at ρmax = 4.7 × 106 with median
B. Back-of-Envelop Estimates of SNR value at ρ median = 59829 for 105 realizations, simulated for
set of 10 parameters. SNR remains loud even for a single
At time of capture of CO, semi-major axis is calculated for for each
EMRI simulation. To approximate the fiducial-fit, quantitatively, it
detectable EMRI during mission lifetime.
is better to use a parameter of gravitational radius rg to understand
the the initial orbit dynamics in terms of influence radius of MBH.
EMRIs event rate sweeps between the semi major axis range in
terms of gravitational radius is 8.81rg ≤ a ≤ 16.24rg. Semi-major
axis a, mass of CO , initial orbital eccentricity eo , cosine of orbital
inclination cos and spin of MBH â contributes to our best fit model,
and rest of the parameters are has no or very negligible
contribution.
log(ρ̂) = −2.86 (a / rg ) − 0.026log (μ/M⊙ ) + .75 eo − .031 Cos(i)
+ .005 â + 14.5 (5)
Power law was generated by using multiple linear regression
method with (forward and back-ward) step-wise approach. The
best fit line was drawn using a generalized linear model (glm–
function) based on Gaussian error distribution. EMRI signal is Fig 5: Histogram expressing SNR distribution of probabilistic EMRIs for
sensitive to each parameter of the model. 105 realizations with LISA sensitivity.

We calculated EMRI event rate for 2 years LISA mission

life, and calculated NEMRI = 9.39 per year. Decimal part of
the result is unrealistic, event rate of 9.39 per year is
averaged for 105 mission realizations. To observe the
probability distribution of event rates, We increase the event
rate by factor of 10000 and expressed in 6.5. The
distribution is in agreement with the Poisson distribution.

Fig 4: Mass–normalized SNR as a function of parameters obtained

by GLM a, cosi, μ, eo . Red line corresponds to the best–fit
estimates log(ρ̂) = −2.86 (a / rg ) − 0.026log (μ/M⊙ ) + .75 eo −
.031 Cos(i) + .005 â + 14.5 of SNR. Radial parameter ranges
between 8.81rg ≤ a ≤ 16.24rg
Fig 6: Poisson probability distribution normalized over 105realizations
Most influential parameter is semi-major axis a be the to for number of EMRI events N EM RI = 9.39 per year be detectable in ISA
estimate SNR. It is anti correlated with SNR, closer initial frequency band form GC of M32 over mission duration of 2–years.
orbits will result in louder EMRI event. Negative weak
dependency exist between CO’s mass and SNR even though GR manifests the classical theory of gravity associated with
the curvature of the manifold and provides precise

Seventh International Conference on Aerospace Science & Engineering 246

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

measurements in for testifying strong gravity. There exists LISA. V. Extreme mass-ratio inspirals”, Physical Review D, vol 95,
no turntable parameters in GR besides Newton’s pp. 103012, May 2017.
Gravitational constant that gives definite predictions. Thus [3] Pau Amaro-Seoane, “Detecting Intermediate-Mass Ratio Inspirals
tests to GR have strong implications probing new physics. From The Ground And Space”. Phys. Rev. D, vol. 98, pp. 063018,
Since, GR might not be the ultimate theory of gravity, September, 2018.
confronting such tests of GR in a strong gravity regime is [4] C. P. L. Berry, and J. R. Gair, “Observing the Galaxy’s massive
therefore consequential. These tests correspond to the low blackhole with gravitational wave bursts”, Monthly Notices of the
energies and scales much shorter than Planck length. At Royal Astronomical Society, vol. 429, pp. 589–612, February 2013.
present, there is no finalized theory of quantum gravity. [5] L. Barack, and C. Cutler, “Lisa capture sources: Approximate
Statistical considerations of EMRIs thus quantifies the waveforms, signal-to-noise ratios, and parameter estimation
modifications to general relativity. GWs studies are accuracy”, Phys. Rev. D, vol. 69, pp. 082005, April 2004.
dependent on matched filtering techniques too, in available [6] Pau Amaro-Seoane, “Relativistic dynamics and extreme mass ratio
waveform models. Waveform models are very sensitive to inspirals”. Living Reviews in Relativity, vol. 4, May 2018.
slightest changes in orbital evolution of COs. These effects [7] David Merritt, “Dynamics of galaxy cores and supermassive black
will be apparent by the mismatching of signal with holes”. Reports on Progress in Physics D, vol. 69, pp. 2513–2579,
theoretical models. August 2016.
[8] J. R. Gair, Christopher Tang, and Marta Volonteri, “LISA extreme-
Having frequent occurrences of EMRIs at lower red shifts z, mass-ratio inspiral events as probes of the black hole mass function”,
investigating the waveform templates will allow us to Phys. Rev. D, vol. 81, May, 2010.
measure luminosity distance D L (z) within 10% precision [9] Clovis Hopman, “Extreme mass ratio inspiral rates: dependence on
and few square degrees of the sky localization. The distance the massive black hole mass”, Classical and Quantum Gravity, vol.
of the source can be deduced from the amplitude of the 26, April 2009.
signal, z, in turn, can be measured. The data can be used to [10] J. R. Gair, “Probing black holes at low redshift using LISA EMRI
compute cosmological parameters like Hubble constant Ho . observations”, Classical and Quantum Gravity, vol. 26, April 2009.
Difficulty lies in the absence of EM counterparts of EMRIs. [11] Curt Cutler, “Angular resolution of the LISA gravitational wave
However, the analysis can be done by associating the detector”, Physical Review D, vol. 57, pp. 7089–7102, June 1998.
independent red shifts measured by available EM data from [12] S. Gillessen et al., “An Update on Monitoring Stellar Orbits in the
the galactic catalog that makes use of a cluster of the Galactic Center”, The Astrophysical Journal, vol. 837, February 2017.
galaxies correlating to the cosmic web, at particular z, to [13] F. Yusef-Zadeh, D. Choate, and W. Cotton, “The Position of
spot the MBH hosting EMRI. Sagittarius A ∗ at the Galactic Center”, The Astrophysical Journal,
vol. 518, pp. L33–L35, May 1999.
ACKNOWLEDGMENT [14] Ali, A., Christensen, N., Meyer, R., and Rover, C., “Bayesian
inference on EMRI signals using low frequency approximations”,
This research was conducted under the supervision of Classical and Quantum Gravity, vol. 29, pp. 145014, June 2012.
Dr.Asad Ali. He provided the simulation source code based [15] A. Ali, “Bayesian Inference on EMRI Signals in LISA Data”, PhD
on his previous studies in the context of MLDC round 4. We thesis, The University of Auckland, New Zealand, 2011.
acquired the knowledge of Galactic EMRIs and [16] R. O. Hansen, “Multipole moments of stationary space-times. Journal
methodology to compute SNRs from Rafia Sarwar. of Mathematical Physics”, vol. 15, pp. 46–52, 1994.
Dr.Fazeel Mehmood Khan helped us sorting the queries and [17] Christopher P. L. Berry et al., “Astro2020 Science White Paper: The
checking the consistency of scaling relations used. unique potential of extreme mass-ratio inspirals for gravitational-
wave astronomy”, White paper submitted to Astro2020 (2020
Decadal Survey on Astronomy and Astrophysics), March, 2019.
[18] Igor D. Karachentsev, Dmitry I. Makarov, and Elena I. Kaisina.
Updated nearby galaxycatalog.The Astronomical Journal, 145(4):101,
Mar 2013.10.1088/0004-6256/145/4/101.ISSN 1538-3881.doi:URL
References http://dx.doi.org/10.1088/0004-6256/145/4/101.
[19] Lee S. Finn, “Detection, measurement, and gravitational radiation”
[1] John Kormendy, and Luis C. Ho, “Coevolution (Or Not) of Phys. Rev. D, vol. 46, pp. 5236–5249, December 1992.
Supermassive Black Holes and Host Galaxies”, Annual Review of
Astronomy and Astrophysics, vol. 51, pp. 511–653, August 2013.
[2] Stanislav Babak et al., “Science with the space-based interferometer

Seventh International Conference on Aerospace Science & Engineering 247

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Strategizing Pakistan’s Space Policy for Defensive

Discourse: A Case Study of Pulwama-Balakot Crisis
Noor Ul Huda Atif
Centre for International Peace and Stability
National University of Science and Technology
Islamabad, Pakistan
[email protected]

Abstract— Outer space is a rapidly developing frontier around the formulation of space policy during the Cold War
that offers opportunities in all aspects of commercial and era has had overwhelming influence on South Asian region.
strategic domains. A well-established policy, which adheres South Asia has been a hot seat of geopoliticatl conflicts
to an appropriate strategy, proves to be helpful in realizing historically. Ever since the independence in 1947 following
these opportunities. For a state such as Pakistan, that the separation of subcontinent, the two new states with
relatively has a small-scale space program with handful of grievances, grudges and completely different ideologies were
satellites, which bear the burden of providing both born. This rivalry between two newly created states was due
commercial and strategic support to country, it requires a to a flash point in the form of unresolved territorial dispute of
well-planned policy to exploit outer space technology Jammu and Kashmir. Over the time, the states of Pakistan
emerging in international security domain. Pakistan and India instead of resolving the disputes have only
should advance its space program for the strategic and amplified enmity through the increased military prowess and
defensive discourse as the hostile acts of the adversary in stockpiles of conventional and unconventional arms that has
the region grows as evidently witnessed in the Pulwama- led to increased strategic instability in the region. The near
Balakot Crisis of 2019. The National Space Policy of equal military expertise and built up has sustained deterrence
Pakistan provides a framework for the development and in the past seven decades however, the imbalance of space
capabilities between both countries account for an effective
launch of Lower Earth Orbit (LEO) and Geo-synchronous
space policy directed to enrich the space program of Pakistan
satellites, yet it lacks in various aspects of defensive
with strategic space assets capable of performing
strategy. It requires revision to reach technical Intelligence, Surveillance and Reconnaissance (ISR) for the
advancement taking place in the field of space sciences armed forces.
leading to the development of Anti-Satellite (ASAT)
weapons. While emerging technologies are making way India and Pakistan share belligerent relations, which have
into the international security, incidents like Balakot strike the tendency to aggravate over the slightest event. Evidently,
2019 suggests that Pakistan should direct its space policy a large scale bombing at a military convoy in Pulwama,
for its security and take a defensive discourse to devise a Illegally Indian Occupied Kashmir (IIOJK) led to military
revised space policy. This paper resolves to explore the escalations on both sides of border initiated by the Indian Air
Force, as they released payload at Balakot, an area in the
options for Pakistan to devise a defensive space policy that
Khyber PakhtunKha (KPK) province. The initial scanning of
secures the space assets and the country.
the area was conducted via the PRSS-1, the only strategic
Keywords— Space Policy, Pakistan Space Policy, remote sensing satellite of Pakistan, which enabled Pakistan
Defensive discourse, Discourse Analysis, Pulwama- to effectively deny the false Indian allegation of destroying a
Balakot, National Space Policy Madrassa harboring 350 terrorist [1]. The incident
emphasizes on the significance of strategic space assets and
I. INTRODUCTION is used as a case study for the initiating a defensive discourse
Outer space serves as an unknown, mysterious frontier among the policy-makers in this paper.
for humans, yet to be completely explored, developed and II. PULWAMA-BALAKOT CRISIS-2019
exploited. In the short while of the space ventures, the
exploitation of outer space ousted the exploration and India and Pakistan share confrontational relations since
development mainly due to the militarization, which gained partition with Kashmir being the flash point. The seven
momentum given the space race. The space policies of USSR decades of independence of both states have seen severed
and US revolved around the newly initiated space race rather diplomatic, economic and military ties and 2019 only added
than space security and sustainable use for a better future. to the fuel. As the Indian Prime Minister Narendra Modi
Over the time lack of effective policy and strategies for the prepared for elections in April 2019, Pakistan anticipated
use of space led to placement of spy satellites and Anti- that the reflection of his anti-Pakistan rhetoric would
Satellites (ASAT) for offensive and defensive measures. A manifest into a mishap that will help PM Modi gain the
comprehensive space policy entails the elements of national popular vote for a second regime in India as Prime Minister.
security, defense, science and technology, research and
On 14 February 2019, a suicide bomber hit an Indian Army
development, academics and education and the law and
convoy at Lethpora in the Pulwama district of Illegally
advocacy. Like any other international affair, the astropolitics
Indian Occupied Jammu and Kashmir (IIOJK). Immediately

Seventh International Conference on Aerospace Science & Engineering 248

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
after the attack, Indian media resorted to its usual banter Within a few hours of attack, Pakistan using its only
against Pakistan. The venomous Indian media as always strategic satellite employed with the Remote Sensing
took the opportunity to accuse Pakistan of the Pulwama Capabilities, PRSS-1 was able to detect the region where the
attack where 40 soldiers of Indian armed forces had died. Indian Aircraft released its payload, damaging cultivation.
Pakistan however, denied the Indian allegations and The satellite images from PRSS-1 confirmed that there was
condemned the attack that was intentionally being blamed to no infrastructure damage on ground and area, which was
aggravate the fire of belligerence between both states. repeatedly being claimed by the Indian media. Later the
On 19 February, the Prime Minister of Pakistan Mr. Imran high-resolution images taken by the European Space
Khan attributed the Pulwama attack to Indian government’s Imaging on the next day of strikes also established that there
desire of conducting a false flag operation against Pakistan were no signs of casualties and no infrastructural damage on
for the upcoming elections, as had been anticipated. Pakistan the site. The Australian Strategic Policy Institute, while
was prepared for a false flag operation by India given the analyzing the satellite data from WorldView-2 satellite,
deteriorating political and military morale on other side of rejected the Indian claim of the strike that killed 350 militant
border [2]. under training. Micheal Sheldon, a digital forensics analyst
The Indian Air Force (IAF) in the early of 26 February 2019 from Atlantic Council conducted an independent
intruded into the international borders of Pakistan and investigation on the issue and stressed that the target site
conducted a failed attempt at neutralizing an alleged terrorist showed no signs of any infrastructure damage and that the
training camp in Balakot, Khyber PakhtunKha, Pakistan. botch was mysterious in the light of the autonomous nature
According to Pakistan, the Air force intercepted three Indian of the supposedly used missiles. A major comparison was
Air Force teams attempting to enter the Pakistan at various drawn by the Reuters, which conducted an investigation
sectors. Pakistan Air Force apprehended two of the three based on high-resolution satellite images by “Planet Labs”.
teams at the border by the combat air patrol aircraft, where It showed images of the said target site from April 2018,
the third one was able to cross the Line of Control (LoC) compared them to the images of February 2019 after the
from the Kiran Valley near Muzzafarabad. However, the strike, and concluded that the landscape was unchanged. The
other two intercepted teams propelled the third team to remarks made by Reuters were that “there were no
retreat, flee the scene releasing their payload, and dump discernible holes in the roofs of buildings, no signs of
their fuel in an open area missing their said target. The scorching, blown-out walls, displaced trees around the
Indian Air Force, government and media claimed to damage madrasa or other signs of an aerial attack”. Satellite imaging
a militant training camp in Balakot killing 350 militants and remote sensing played an essential role in the
under training. However, their claims were rendered manifesting Indian propaganda against Pakistan.
catatonic by the international media, defence analysts and 2) High-Tech Combat Aircrafts
organizations. In this particular case, the satellite technology Anticipating that IAF would return with more confidently,
proved to be a valuable asset to silence the false prerogatives Pakistan had already set up a trap of communication
of India [3]. jammed environment for it. When the IAF aircrafts entered
The 27 February 2019, Indian Air Force again attempted the the region, they lost contact with one another and were
same stunt of entering the borders of Pakistan challenging unable to communicate back at the ground station. The
the country’s sovereignty. This time they met with the JF-17 multi-role JF-17 and F-16 aircrafts employed with high-tech
fighter aircraft formations rather than the patrolling aircrafts combat equipment were able to bring down the Indian Mig
of Pakistan Air Force (PAF). The confidence from their and Sukhoi Fighter jets. The communication jammers
previous attempt compelled them to conduct airstrikes in the blocked the signals to the aircraft radars which practically
broad daylight; however, they were apprehended and chased blinding the aircrafts. PAF’s JF-17 fighter aircrafts though
away by the PAF aircrafts. The aircrafts engaged in an air used under the conventional capacity during the Operation
combat where the Pakistan shot down a Mig-21 and a Su- Swift Retort in 2019, are being upgraded to the electronic
30MKI aircrafts of Indian Air Force and captured an Indian Warfare suit, which will enable it to be a smart machine and
Air Force pilot. assess its situation and react to it [4].
A. Technical Component of Air-Combat III. OUTER SPACE POLICY FOR PAKISTAN
The Pulwama-Balakot crisis and the air combat that Outer space is an essential entity in the international security
followed it were dependent on two main elements of the and the advance technological world and requires
technology: the satellite imagery and remote sensing legitimization for its utility. The global importance of the
technology and secondly, the high-tech combat aircrafts. domain has become evident as defining the future. The
The air combat that was considered a conventional air frontier that requires the immediate humanly attention is
combat hinted towards an electronic war where modern exploited more than its development and exploration with
technology had a significant role to play. The high imagery the change in global security architecture. As a global
from satellites essentially established that IAF had common, the outer space is an entity that holds significant
wrongfully entered Pakistan challenging its sovereignty and value in the international arena in modern times mainly due
damaging cultivation in the early hours of 26 February 2019. to its ability to provide various services both for commercial
The high-resolution satellite images from various satellites and military purposes. However, the utilization of outer
around the world confirmed that no infrastructural damage space requires a comprehensive set of guidelines, which
was adhered during the Indian intrusion. On 27 February adhere to the international laws and treaties. In a fragile
2019, the high-tech combat equipment employed in the region of South Asia, a comprehensive space policy will
aircrafts played a significant role in Operation Swift Retort. have to entail the main elements of national security and
1) Role of Satellite Technology in Crisis strategic stability. The hostile relations of the two states of

Seventh International Conference on Aerospace Science & Engineering 249

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
India and Pakistan negatively affect the regional peace and engage in large number of space assets and satellites that can
stability. The similar scenario in future can lead to severe match the advance high-tech satellites in space by the
consequences as India now has the Anti-Satellite weapon. adversary. The emerging space technology as an essential
India has made clear of its adversarial ambitions over time component of the international security manifests threats
and Pulwama-Balakot crisis evidently suggests that it will be emerging from space for Pakistan from the adversary in the
willing to make such an attempt again in the future. region.
Space acts as a force multiplier for the armed forces for the B. A Defensive and Strategic approach
intelligence and reconnaissance missions. Defensive
measures embedded in the space policy provide the For Pakistan, a strategic and defensive approach to the space
modernization that is essential for the forces fighting wars in policy entails the security to space assets, ground systems,
a technically advanced era. Pakistan’s policy-makers and control and command centre, strategic assets and their
academia is in dire need of opening a discourse for a movements and the terrestrial humans. Spatial defense
defensive space policy. Although a comprehensive policy measures are enhanced through the thrust areas of
comprises of various components, however, in this case it is intelligence, surveillance, reconnaissance and navigation.
evidently essential for Pakistan to follow the defensive Exploitation of space for real-time military utilization has
discourse and open to measures for securing strategic and significant advantages. It gives an upper edge over the
spatial assets from the threats emerging from space activities adversary in the communication aspect and the
of adversary. Pakistan’s missile capabilities evidently reconnaissance missions. Real-time spatial information used
manifest that the country is proficient enough to enrich its efficiently can be significantly employed for early-warning
space program with strategic satellites employed with ISR missile attacks, deliverance of precision guided munitions
capabilities. through satellite signals, or jamming enemy networks.
Space technology is specifically essential for the Air Force
A. Pakistan’s Space Vision and acts as a force multiplier by allowing battlefield
Pakistan as a nascent space state is determined to broaden situational awareness through space communication and
the horizon of its capabilities in the field of space space sensors. The link established between the fighter
technology. The increasing significance of outer space in aircrafts and electronic radar satellites allows
technically advanced era specifies the growth of modern communication with network centric systems across the
technology with the influence of command, control and country. The assistance in Ballistic Missile Defence system
communication. The dual-purpose satellites, including the is a significant role of space technology. The imperative of
commercial and strategic spatial assets and satellites space situation awareness is essential for gathering real-time
evidently strengthen the national security. intelligence on the adversary aircraft, missiles and space
Pakistan’s Space Vision was developed in 2011 with the aim borne threats from the Anti-Satellites weapons. As in the
of utilizing space for socio-economic benefits. Pakistan has case of Operation Swift Retort, PAF resorted to jamming the
been geo-strategically a significant state in the international communication and radar systems of adversary aircrafts,
security events that accounts for high-tech military gains. similarly, the space assets can also be rendered inoperable
The recent event of Balakot strike and the following by resorting to jamming that gives an upper edge.
Operation Swift Retort, where the adversary has used C. Current R&D to modify the Defensive Measures
coercive measures to entice through its belligerent actions is
compelling enough to incorporate the strategic and The direction of the space policy currently can be estimated
defensive measures into the vision plan. The ‘space vision’ through the budget of 2020-2021 and budget 2021-2022 for
functions by an operational framework of Pakistan’s Space SUPARCO. Pakistan is determined to enhance its research
of Program -2040. Under this framework, Pakistan has set and development in the field of space sciences and space
its ambitions of developing and launch of certain quantity of technology. However, the lack of dedicated research centers
satellites. This framework will provide the institution with and satellite development institutes are a hurdle in
the roadmap for development of five Geo-stationary realization of the Pakistan Space Vision. In the year 2021,
satellites that may include the multi-purpose and navigation SUPARCO was authorized for the establishment of Space
satellites and six Low Earth Orbit (LEO) satellites which Application Research Center in Gilgit-Baltistan and Pakistan
will include the remote sensing and optical remote sensing Space Center. These centers will be responsible to undertake
satellites and the communication satellites. the feasibility study and research for various satellite
programs to be launched in future. These future projects
The space vision that was initially named the space vision under the research and development include the Pakistan
2040, was to ensure the research and development of said
Remote Sensing Synthetic Aperture Radar Satellite and
eleven satellites by the year 2040, which is a very minimal
Pakistan Satellite Navigation Program, Pakistan Multi-
quantity. As a nascent space faring state, Pakistan is to
Mission Satellite (PAKsat-MM1), Pakistan Optical Remote
engage in large number of space assets and satellites that can
Sensing Satellite (PRSS-02), advance system for PAKSat-
match the advance high-tech satellites in space by the 1R Ground control segment and the development of Online
adversary. The emerging space technology as an essential Satellite Image Service. The Radar Satellite will develop
component of the international security manifests threats
Pakistan’s reconnaissance and surveillance capabilities
emerging from space for Pakistan from the adversary in the
assisting the military in detection of friendly and adversary
region [5].
objects approaching. PRSS-1 is already employed as
The space vision that was initially named the space vision country’s only strategic satellite and an Optical Remote
2040, was to ensure the research and development of said Sensing Satellite PRSS-2 will enhance will strategic
eleven satellites by the year 2040, which is a very minimal capabilities further [6].
quantity. As a nascent space faring state, Pakistan is to

Seventh International Conference on Aerospace Science & Engineering 250

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
IV. CONCLUSION https://indianexpress.com/article/what-is/mission-shakti-asat-missile-
5645166/
Pakistan’s ISR capabilities lag in the space technology,
which weakens the defence of state and the handful of
spatial assets owned by the country. The incident of
Pulwama-Balakot crisis evidently directs to the emergence
of electronic warfare, where the victory will bselong to the
technical sound side. In the incident, Pakistan mainly relied
on the international media and analysts to decipher the
satellite data and images in order to nullify the false Indian
claims. If Pakistan had its own ISR capable satellite, the jet
movements could have been intercepted in time and the
incident could have been prevented from occurring. It is
evident from the incident that ISR capable satellites play a
vital role; they would have assisted in protection of state and
land sovereignty. The policy-makers must incorporate the
enhanced spatial defensive measures so to avoid such
mishap in future as the adversary has progressed further
since the incident with the development of Anti-Satellite
weapon and various intelligence, surveillance and
navigation satellite [7]. Along with that the Basic Exchange
Cooperation Agreement (BECA) that will ensure the real-
time geospatial information sharing between United States
and India that will enable both the states to eye the naval and
ground movements of Pakistan and China at Gwadar Port
and the China-Pakistan Economic Corridor (CPEC) project.
Pakistan Air Force is determined to upgrade the fighter
aircrafts at par with the suit of electronic warfare where it
will be able to configure its surroundings automatically, the
additional linkage of space communication and information
will enhance the interception and networking properties with
the ground system. This combination will provide the
intelligent combat property to the fighting aircraft. The
satellites operating as the eyes in the sky will increase the
functionality of the combat machines and provide amplified
battlefield situational awareness.
Pakistan’s strategic approach to the space assets requires a
defensive path rather than the offensive one. India’s
offensive account as a space power threatens the strategic
stability of the region and security of Pakistan. The current
research and development requires a speedy process, which
will enable the country to defend against offensive
adversarial actions. The changing International security
situation has once again increased Pakistan’s geostrategic
significance and hence the need of defensive spatial
measures for the country.
V. REFERENCES
[1] SPD Official 1 (Militarization of Outer Space: threats to security of
Pakistan and strategic stability of South Asia), interviewed by Noor Ul
Huda Atif, SPD Head Quarter, Rawalpindi, 31 May 2019.
[2] “After Balakot, Pakistan Threatened to Respond to India's Six Missiles
'Three Times Over': Report”, News 18, accessed April 4, 2019,
https://www.news18.com/news/world/pakistan-threatened-to-respond-to-
indias-six-missiles-three-times-over-report-2071289.html.
[3] “Indian aircraft violates LOC, scramble back after PAF’s timely
response: ISPR”, Dawn News, Accessed 4 August 2021,
https://www.dawn.com/news/1466038
[4] Pakistan Air Force Officer (Strategizing Pakistan’s Space Policy for
Defensive Discourse), interviewed by Noor Ul Huda Atif,, 21,August 2021
[5] Emerging Issues, 2012. “Regulation of Space Activities,” no. September
[6]Khan, Javed Iqbal. "Federal Budget Publication." Federal Budget 2021.
Finance Division , 2021.
[7] Express Web desk, “What is Mission Shakti-ASAT,” The Indian
Express, March 27,2019, accessed June 22,2019,

Seventh International Conference on Aerospace Science & Engineering 251

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Multi-GNSS Navigation Performance Assessment

and Improvement Quantification
Arif Hussain Syed Hadi Hussain Shah Hina Magsi
GNSS and Space Weather Lab Department of Electrical and Computer GNSS and Space Weather Lab
Sukkur IBA University, Engineering Sukkur IBA University,
Sukkur, Pakistan Muhammad Ali Jinnah University, Sukkur, Pakistan
[email protected] Karachi, Pakistan [email protected]
[email protected]

Arlsan Ahmed Faheem Akhtar Chachar

GNSS and Space Weather Lab, Department of Electrical Engineering
Sukkur IBA University, Sukkur IBA University,
Sukkur, Pakistan Sukkur, Pakistan
[email protected] [email protected]

Abstract— The Global Navigation Satellite Systems (GNSS) I. INTRODUCTION

offer an easy, efficient, and cost-effective way to determine the
location, time, and velocity anywhere around the globe. The positioning and navigation technologies have become
However, the existing systems have only been successful in more prevalent and pervasive in our modern and technology-
outdoor or open-sky environments and often lack consistency in driven society. The significance of accurate, reliable, and
maintaining the required navigation performance (RNP) uninterrupted localization and navigation services is obvious
thresholds in urban contexts where availability and accuracy of for certain applications such as road, rail, aerial and marine
GNSS are degraded due to signal blockage, multipath (MP) transport, precision agriculture, surveying and multi-modal
effects, and NLOS reception. In recent years, GNSS technology logistics [1]–[4] , but certain high-tech and safety-critical
has undergone remarkable developments in terms of satellite applications such as autonomous driving architectures (i.e.,
constellations, signal quality and transmission, receiver Unmanned Ariel vehicles, auto-bot), intelligent transportation
technology and positioning estimation techniques. The systems, e-Health, Geo-fencing and other safety-critical
completion of emerging navigation systems (i.e., BeiDou and applications majorly rely on accurate positioning and precise
Galileo) along with the modernization of existing constellations navigation to carry out essential operations [5]–[8]. Among
(i.e., GPS and GLONASS) promises significant improvements the numerous technical solutions to precise positioning and
in satellite visibility, geometry, quality of navigation signals, and
navigation [9], [10],the Global Navigation Satellite System
thus overall localization accuracy. The aim of this paper is to
(GNSS) has established itself as an efficient and cost-effective
record, quantify and compare the performance of satellite-
based navigation systems in single and multi-GNSS mode under
technology for outdoor navigation.
different operating environments. For this purpose, a series of The GNSS is long-established, prevalent and all-weather
field experiments are carried out in multi-GNSS and GPS-only satellite-based navigation system that determines the location,
mode with total observation period of 24 hours and 864,000 data time and velocity anywhere around the globe. The first
samples at standard open-sky and degraded site (i.e., satellite based navigation (Global Positioning System (GPS))
surrounded by high-rise buildings). In the multi-GNSS mode, all was declared fully operational system in 1995 with a
four global constellations (i.e., GPS, GLONASS, Galileo and
constellation of 24 satellites in Medium Earth Orbit (MEO)
BeiDou) are utilized simultaneously for the first time over
[11], [12]. However, poor positioning performance i.e.,
Pakistan. The performance assessment in this paper is carried
out using fundamental quality indicators, e.g., satellite visibility limited availability and degraded accuracy was a major hurdle
or service availability, satellite geometry in terms of PDOP, and to the practical realization of GPS in several critical
statistical accuracy measures (i.e., Circular Error Probable applications [13], [14]. In recent years, the landscape of
(CEP) and Distance Root Mean Square (DRMS)). The results of satellite-based positioning has experienced dramatic changes
performance assessment indicate that combining multiple and advancements in terms of satellite constellations, signal
satellite constellations in multi-GNSS mode results in increased transmission, quality of signals, receiver technology and
satellite visibility/availability, better geometry and thus mitigation models [15]. As of 2020, the GNSS has grown from
significantly improved localization accuracy. It is then further the U.S. GPS - only system to four independent global
analyzed that increased satellite availability in multi-GNSS constellations including the Russian GLONASS, European
mode may not result in consequential improvements in Galileo, and Chinese BeiDou, leading to remarkable boost in
positioning accuracy at standard open-sky environment, satellites availability. Besides four global constellations, there
however, significant improvements were observed at the are further regional i.e., QZSS (Japan), IRNSS (India) and
degraded site. The statistical accuracy of reported positions was Satellite Based Augmentation Systems (SBAS) substantially
found to be improved by five times and outages were reduced by aiding the availability and accuracy. Compared to the early
almost 50% of the observation period in the multi-GNSS mode 2000s, when a single constellation receiver (GPS-only) relied
at the degraded site. on less than 10 satellites to estimate a positioning solution
[16], [17], now a multi-constellation GNSS receiver
Keywords— GNSS, Performance Assessment, Positioning
Accuracy, Environment Contexts.
(GPS+GLONASS+Galileo+BeiDou) can acquire more than
30 satellites on a given geo-location [18]. Furthermore, GNSS
modernization has enabled signal transmission at multiple

Seventh International Conference on Aerospace Science & Engineering 252

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

frequencies, increasing capability in terms of signal single and quad-constellation mode respectively to estimate
acquisition, atmospheric correction, multi-path detection, the navigation solution [22]–[24]. The availability can be
resilience to interference and precise positioning. The defined as percentage of time that the services of the system
modernized radio navigation signals are stronger, accurate and are available to user, which is an indication of the ability of
more resilient to natural or man-made interference [19]. These the receiver to provide reliable positioning information. In
notable advancements in satellite-based navigation systems in this paper, the service availability is characterized and
terms of satellite constellations (i.e., Satellite visibility and quantified by two events i.e., Tracking and outage.
geometry), signal transmission, quality of signals and receiver
technology promise significant improvements in GNSS
performance.  Tracking : 𝑁 ≥ 𝑁𝑇𝐻
 Outage : 𝑁 < 𝑁𝑇𝐻
Keeping in view above fact, this experimental case study
is intended to assess, quantify and compare the performance Where, 𝑁 represents the number of tracked satellites and 𝑁𝑇𝐻
of a GNSS receiver in a single (GPS) and multi-GNSS mode shows the satellite threshold or minimum number of satellites
in different signal reception conditions. For this purpose, a required for successful operation. For single-constellation
series of field experiments are carried out at pre-surveyed and
𝑁𝑇𝐻 = 4 and quad-constellation 𝑁𝑇𝐻 = 7.
carefully selected experimental sites i.e., standard open-sky
and degraded environment. At first, the GNSS performance in
both configuration modes is evaluated at nominal open sky 3) Position Dilution of Precision (PDOP): The
and degraded environment (i.e., surrounded by high rise satellite geometry, which indicates the spatial distribution or
buildings) using fundamental quality indicators, e.g., satellite actual geometric locations of the satellites in the sky as
visibility or service availability, satellite geometry in terms of perceived by the GNSS receiver, contributes significantly to
PDOP, and statistical accuracy measures (i.e., Circular Error the overall accuracy of the navigation solution. For given
Probable (CEP) and Distance Root Mean Square (DRMS)). number of satellites, the poor geometric distribution (i.e.,
Then, in order to compare the resilience of multi-constellation satellite clustered together in the one part of sky) results in
GNSS with single-constellation under degraded degraded positioning accuracy and the severity of this effect
environmental conditions, the improvement in navigation is usually estimated by Position Dilution of Precision
performance due to combination of all four satellite (PDOP). The PDOP is geometrically determined factor that
constellations is computed and compared with reference to refers to positioning inaccuracies caused by degraded spatial
GPS-only mode. distribution of satellites [25]. The better geometric
distribution (i.e., satellites spread over sky) yields a lower
II. GNSS PERFORMANCE ASSESSMENT
PDOP value while higher PDOP values are the indication of
METHODOLOGY AND FIELD EXPERIMENTATION
degraded satellite geometry. From estimated/reported
This section presents a detailed overview of the coordinates (i.e, x, y, z), the PDOP can be calculated as.
performance assessment framework adopted in this paper,
fundamental quality indicators of GNSS and field 𝑃𝐷𝑂𝑃 = √𝜎𝑥 2 + 𝜎𝑦 2 + 𝜎𝑧 2
experimentation which includes the details of the technical
equipment and candidate sites on which experiments are where, 𝜎 2 represents the variance.
performed.
4) Statistical Accuracy: The accuracy is the key
A. Performance Assessment Framework parameter of the navigation performance that measures the
The availability and accuracy of GNSS-based localization deviation/dispersion in the reported position with reference to
and navigation is highly influenced by several systematic the true/mean positions. It is the statistical measure of
errors (e.g., clock biases, orbital errors and receiver noise), performance that is associated with the probabilistic
atmospheric effects (e.g., ionospheric scintillation, solar and distribution of positioning error. In positioning and
geomagnetic storm) and environmental factors (blockage, navigation systems, accuracy usually refers to the radius of
fading, multi-path and/or NLOS reception), all of which can circle around a mean/true position also referred as confidence
lead to serious inaccuracies or even outages in navigation region. The smaller the radius of region or circle, the higher
services [20], [21]. The scope of this research is limited to the accuracy [26]. In this study, two most commonly used
investigation of environmental factors by analyzing accuracy parameters i.e., Circular Error Probable (CEP) and
fundamental quality indicators of GNSS i.e., satellite Distance Root Mean Square (DRMS) are used for
visibility, service availability, PDOP and statistical accuracy performance assessment.
measures. a) Circular Error Probable (CEP): The CEP is a
1) Satellite Visibility (N): Satellite visibility represents fundamental statistical accuracy measure and is defined as the
the number of satellites that can be acquired or tracked at a radius of a circle, centered on the true position and expected
given geo-location and time of the day. It has direct and to include 50% reported positions. For example, if a CEP of
determinable effects on accurate positioning and precise 3 m is mentioned and 2000 solutions are computed then it is
navigation. With more satellites on hand, better positioning highly expected that 1000 values will in a circle having radius
accuracy can be achieved by selecting a group of potentially of 3 m [18], [27]. Using the standard deviation (σ) of
strong satellites [18]. estimated coordinates, the Circular Error Probable can be
estimated as
2) Service Availability: For proper operation and
accurate positioning and navigation solution estimation, a 𝐶𝐸𝑃 = 0.62𝜎𝑥 + 0.56𝜎𝑦
GNSS receiver requires a minimum number of satellites and
below these thresholds, the receiver experiences outages. b) Distance Root Mean Square (DRMS): The DRMS is
Specifically, a minimum of 4 and 7 satellites are required for

Seventh International Conference on Aerospace Science & Engineering 253

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

a two-dimensional accuracy parameter that includes the 65%

of reported positions within the confidence region [18], [27]
It is estimated by taking the square root of the average of III. PERFORMANCE EVALUATION RESULTS
squared positioning errors. In this section, the navigation performance achieved by
GPS-only and multi-GNSS receivers at two candidate sites is
presented and discussed in detail.
𝐷𝑅𝑀𝑆 = √𝜎𝑥 2 + 𝜎𝑦 2
A. Performance Assessment in GPS-only mode
The performance evaluation results of a GNSS receiver in
where, 𝜎 2 represents the variances of reported coordinates. GPS-only configuration and in terms of fundamental quality
An example of statistical accuracy of reported positions can indicators such as satellite visibility or service availability,
be seen in Fig. 1. The positioning solution estimated at every PDOP and standard accuracy measures (CEP and DRMS)
epoch is plotted as blue point while the red and green circle over entire experimental period of 6 hours on both candidate
show the CEP and DRMS respectively also known as sites are shown in Fig. 4. Fig. 4(a), (b) and (c) in the top half
confidence region. of figure shows the performance assessment results in
standard open-sky environment which is characterized by the
absence of reflective surfaces, and with optimal conditions for
satellite tracking or acquisition (e.g., negligible obstructions)
as shown in Fig 2. While, Fig. 4(d), (e) and (f) together
represents the overall performance achieved by a GPS
receiver on a degraded site having reduced satellite visibility
and high probability of signal blockage and multipath effects
due to presence of obstructions.
The satellite visibility along with service availability
which is characterized by Tracking and Outage events are
shown in Fig. 4(a) and Fig. 4(d) for Open-Sky and Degraded
environments respectively. As expected, no outage at all was
experienced by the receiver in an open sky environment
during the entire observation period of 6 hours (00:00 – 06:00)
due to optimal signal reception conditions. The average
satellite visibility, N, was found to be 09 which is quite a good
number as compared to required threshold for single-
constellation receiver (i.e., N¿= 4). With adequate satellite
visibility, there is a high probability of accurate estimation of
positioning and navigation solution. On the other hand, the
Fig. 1 Statistical Accuracy of reported position estimated in terms of CEP. receiver experienced significant outages at the degraded site
which can lead to disastrous consequences in safety-critical
B. Field Experimentation applications. As it can be clearly seen in Fig. 4(d), the receiver
In this study, a total of four field experiments are was only able to provide the positioning solution for less than
conducted on two carefully selected experimental sites i.e., 3 hours with the outage of almost 52% due to the presence of
standard opensky and degraded environment. The first site signal blocking scenarios as shown in Fig. 2. The average and
i.e., open sky has excellent signal reception conditions with no maximum number of satellites dropped to 4 and 6 respectively
blockage and mostly LOS reception as shown in Fig. 2(a). On and this limited satellite visibility may also result in poor
the other hand, the second site (Fig. 2(b)) which is
characterized as degraded environment has significant
chances of signal obstruction and multipath/NLOS reception
that can result in degraded navigation accuracy. These two
experimental sites, namely Standard Open-Sky and Degraded
Environment represent the two main operating environments
for GNSS receiver. The details of each experiment performed
are given in Table I.
The experimental setup used in this study is sown in Fig.
3, which consists of PolaRx5s GNSS receiver, high precision
antenna and computing machines. In each experiment GNSS
raw data was logged at the rate of 10 Hz for an observation
period of six hours. The raw data is then processed and
fundamental quality indicators are extracted for performance
assessment.

TABLE I. FIELD EXPERIMENTATION

Observation
S. No. Experimental site Mode
Period
1 Standard Open-Sky GPS-Only 06 (Hours)
2 Degraded Environment Multi-GNSS 06 (Hours)
3 Standard Open Sky GPS-Only 06 (Hours)
4 Degraded Environment Multi-GNSS 06 (Hours)

Seventh International Conference on Aerospace Science & Engineering 254

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 2 The candidate sites used in the field experimentation; (a) Standard open-sky environment having excellent signal reception conditions (b) Degraded

Fig. 3 Experimental setup consisting of Septentrio’s PolaRx5s Multi-constellation GNSS receiver used in the field experimentation to acquire GNSS data.

satellite geometry which ultimately contribute to more reported in both experiments in the single-constellation mode,
inaccuracies in positioning estimation. it can be concluded that satellite geometry is significantly
influenced by environmental conditions and it is more likely
The PDOP is a key index used for analyzing the effect of to be affected in urban or dense urban areas.
satellite geometry on the accuracy of positioning solutions.
The obstructions present in between the satellite and receiver Once the satellite visibility and geometry are checked, the
not only result in reduced satellite visibility but also affect the accuracy of the reported position is analyzed in open-sky and
spatial distribution of satellites and yielding the poor satellite degraded environment cases. As expected, the adequate
geometry which is indicated by a higher PDOP value. Fig. satellite visibility along with excellent geometry in open-sky
4(b) and Fig. 4(e) show the reported PDOP value in open sky case results in good accuracy of positioning solution. The
and degraded environment respectively for the observation accuracy of reported position in terms of CEP and DRMS is
period of 6 hours. In open-sky views, the PDOP value was shown in Fig. 4(c) and Fig. 4(f). For the case of the open-sky
found to be less 2 for more than 90% of observation time environment, the values of CEP and DRMS were observed to
which is the indication of excellent geometric distribution of be 1.15 m and 1.36 m respectively. In most navigation
satellites and this was expected due to optimal environmental applications, the CEP of less than 1.5 m is usually considered
conditions with negligible blockage. However, at the as an acceptable level of accuracy. However, significant
degraded site, the extremely poor PDOP values were observed inaccuracies in positioning solution were reported at the
with average PDOP value greater than 10 which is normally degraded site due to severe signal obstruction, poor geometry
considered as imperfect satellite geometry. The geometry of and multipath signal reception. The value of CEP was around
satellites in the sky and accuracy of positioning solution are
highly inter-linked because the PDOP values determines the
amplification factor of positioning errors. For the same
number of satellites, poor geometric distribution can result in
decreased positioning accuracy. Based on the PDOP values

Seventh International Conference on Aerospace Science & Engineering 255

 Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 2 The performance assessment of GNSS receiver in quad-constellation mode at standard open sky and degraded environment using fundamental quality.

16 m and DRMS was found to be 22.24 m at the degraded site The accuracy levels estimated for the reported position in
which shows a quite large inaccuracy. multi-GNSS mode at open-sky and degraded site are shown in
Fig. 5(c) and Fig. 5(f) respectively. As expected, the increased
B. Performance Assessment in Multi-GNSS mode satellite visibility along with better geometric distribution
The performance evaluation results of a GNSS receiver in resulted in significantly improved positioning accuracy. For
multi-GNSS configuration and in terms of fundamental the open-sky case, the accuracy parameters were found to be
quality indicators such as satellite visibility or service less than 1 during entire observation period of the experiment.
availability, PDOP and statistical accuracy parameters (CEP In fact, almost 90% values of the estimated positioning
and DRMS) over entire experimental period of 6 hours on solution were observed to be in confidence region having a
both experimental sites are shown in Fig. 5. Fig. 5(a), (b) and radius of 0.4 m which is usually considered as highly accurate
(c) in the top half of figure shows the performance assessment level of positioning accuracy in most of the navigation
results in standard open-sky environment while Fig. 5 (d), (e). applications.
and (f) represent the overall performance achieved by a GNSS
receiver on a degraded site. The integration of all four satellite Compared to GPS-only mode at open-sky, the statistical
constellations such as GPS, GALILEO, BeiDou Russian accuracy was found to be improved by 1.7 times in multi-
GLONASS resulted in remarkable boost in satellite visibility GNSS mode as CEP reduced to 0.80 m. Although, it is a very
and availability of direct Line-of-Sight (LOS) signals as good improvement but GPS-only receiver also gives a
shown in Fig. 5(a) and Fig. 5(d). As it can be clearly seen in sufficiently accurate positioning solution in open sky views.
Fig. 5, the increased satellite visibility results in reduced Compared to the GPS-only mode at open-sky, statistical
outages, better satellite geometry and thus improved accuracy was found to be improved by 1.7 times in multi-
positioning accuracy. GNSS mode as CEP reduced to 0.80 m. Although, it is a very
good improvement but GPS-only receiver also gives a
In both cases, the receiver was found to be in complete sufficiently accurate positioning solution in open sky views.
tracking mode without any interruption. The average satellite However, very noticeable accuracy improvements were
visibility, N, was observed to be 38 at open-sky and 24 at the observed at the degraded site in multi-GNSS in comparison
degraded site which is a quite large number as compared to with GPS-only receiver configuration. The availability of
single-constellation where average satellite visibility was redundant range measurements due to combination of all four
found to be less than 10 in both cases. Similarly, noticeable satellite constellations resulted in significantly improved
improvements were observed in satellite geometry; the PDOP accuracy of positioning solution. The CEP and DRMS were
value was found to be less than 1 for the whole observation found to be 3.21 m and 4.40 m which are almost 5 times
period at open-sky site and less 2 for 90% of experimentation smaller than their counter parts in singleconstellation. Hence,
time at the degraded site as shown in Fig. 4(b) and Fig. 4(e) it can be established that combination of multiple navigation
respectively. Interestingly, even at the open-sky site with
almost same environmental conditions, the average PDOP
value was reduced by around 54% due to increased number of
tracked satellites. In the same way, at the degraded site, the
average PDOP value was 1.77 which almost 7 times less than
its counterpart in open-sky case. These are the very
noteworthy improvements in satellite visibility and geometry
which promise a significant accuracy enhancements even on
the degraded sites.

Seventh International Conference on Aerospace Science & Engineering 256

 Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 5 The performance assessment of GNSS receiver in quad-constellation mode at standard open sky and degraded environment using fundamental quality indicators.

system can significantly mitigate the environmental effects by [3] C. Stallo et al., “GNSS-based location determination system
leveraging the redundancy of range measurements. architecture for railway performance assessment in presence of
local effects,” 2018 IEEE/ION Position, Locat. Navig. Symp.
IV. CONCLUSION PLANS 2018 - Proc., pp. 374–381, Jun. 2018, doi:
10.1109/PLANS.2018.8373403.
In this paper, a comprehensive study on the GNSS [4] S. Lo, “Navigation for Aviation,” Position, Navig. Timing
performance in single and multi-constellation mode using Technol. 21st Century, pp. 1871–1891, Dec. 2020, doi:
fundamental quality indicators and under different signal 10.1002/9781119458555.CH61.
[5] G. B. Michael Dooley, Paolo Pirjanian, Nikolai Romanov, Lihu
reception conditions (i.e., environments) is presented. The Chiu, Enrico Di Bernardo, Michael Stout, “Application of
navigation performance is evaluated by quantifying and localization, positioning & navigation systems for robotic enabled
analyzing the fundamental quality indicators of GNSS i.e., mobile products.”
satellite visibility along with service availability, PDOP and [6] A. F. G. Ferreira, D. M. A. Fernandes, A. P. Catarino, and J. L.
statistical accuracy of reported position. Based on the detailed Monteiro, “Localization and Positioning Systems for Emergency
Responders: A Survey,” IEEE Commun. Surv. Tutorials, vol. 19,
analysis of performance evaluation results on both candidate no. 4, pp. 2836–2870, Oct. 2017, doi:
sites, it can be concluded that single-constellation (GPS) 10.1109/COMST.2017.2703620.
receiver can only provide a user with accurate, reliable and [7] C. B. Lin, R. W. Hung, C. Y. Hsu, and J. S. Chen, “A GNSS-Based
seamless positioning and navigation services in open-sky Crowd-Sensing Strategy for Specific Geographical Areas,”
views. However, poor signal reception conditions (i.e., Sensors 2020, Vol. 20, Page 4171, vol. 20, no. 15, p. 4171, Jul.
blockage or multipath) in degraded environments make it very 2020, doi: 10.3390/S20154171.
[8] S. Suryaatmadja and N. Maulani, “CONTRIBUTIONS OF
hard for a GPS receiver to function properly without using any SPACE TECHNOLOGY TO GLOBAL HEALTH IN THE
aiding devices. On the other hand, the integration of all four CONTEXT OF COVID-19,” J. Adm. Kesehat. Indones., vol. 8, no.
satellite constellations such as GPS, GALILEO, BeiDou 2, pp. 60–73, Jun. 2020, doi: 10.20473/JAKI.V8I2.2020.60-73.
Russian GLONASS resulted in remarkable boost in satellite [9] P. D. Groves, “Principles of GNSS, Inertial, and Multisensor
visibility and availability of direct Line-ofSight (LOS) that Integrated Navigation Systems, Second Edition | Artech books |
IEEE Xplore.” https://ieeexplore.ieee.org/document/9101092
resulted in reduced outages, better satellite geometry and thus (accessed Aug. 01, 2022).
improved positioning accuracy. In fact, the accuracy of [10] Nasir SaeedHaewoon NamTareq Y. Al-NaffouriTareq Y. Al-
reported positions was found to be improved by almost five NaffouriMohamed-Slim Alouini, “A State-of-the-Art Survey on
times and the outages were reduced by almost 50% in the Multidimensional Scaling Based Localization Techniques.”
multi-GNSS mode at the degraded site. https://www.researchgate.net/publication/333678948_A_State-
of-the-
ACKNOWLEDGEMENT Art_Survey_on_Multidimensional_Scaling_Based_Localization_
Techniques (accessed Aug. 01, 2022).
This work was supported by the Higher Education [11] N. R. Council, “The Global Positioning System: A Shared
Commission (HEC) of Pakistan through the National National Asset,” May 1AD, doi: 10.17226/4920.
Research Program forUniversities (NRPU) under Grant [12] B. W. PARKINSON, T. STANSELL, R. BEARD, and K.
GROMOV, “A History of Satellite Navigation,” Navigation, vol.
6250/Sindh/NRPU/R&D/HEC/2016. 42, no. 1, pp. 109–164, Mar. 1995, doi: 10.1002/J.2161-
4296.1995.TB02333.X.
REFERENCES [13] P. Misra, B. P. Burke, and M. M. Pratt, “GPS performance in
[1] G. N. S. S. European, “GNSS Market Report,” doi: navigation,” Proc. IEEE, vol. 87, no. 1, pp. 65–85, 1999, doi:
10.2878/565013. 10.1109/5.736342.
[2] I. Skog and P. Händel, “In-car positioning and navigation [14] W. Y. Ochieng and K. Sauer, “Urban road transport navigation:
technologiesa survey,” IEEE Trans. Intell. Transp. Syst., vol. 10, performance of the global positioning system after selective
no. 1, pp. 4–21, 2009, doi: 10.1109/TITS.2008.2011712. availability,” Transp. Res. Part C Emerg. Technol., vol. 10, no. 3,

Seventh International Conference on Aerospace Science & Engineering 257

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021
pp. 171–187, Jun. 2002, doi: 10.1016/S0968-090X(02)00008-6.
[15] N. Joubert, T. G. R. Reid, and F. Noble, “Developments in Modern
GNSS and Its Impact on Autonomous Vehicle Architectures,”
IEEE Intell. Veh. Symp. Proc., pp. 2029–2036, 2020, doi:
10.1109/IV47402.2020.9304840.
[16] “Global Positioning System Standard Positioning Service
Performance Standard.” https://rosap.ntl.bts.gov/view/dot/16930
(accessed Aug. 01, 2022).
[17] W. Lechner and S. Baumann, “Global navigation satellite
systems,” Comput. Electron. Agric., vol. 25, no. 1–2, pp. 67–85,
Jan. 2000, doi: 10.1016/S0168-1699(99)00056-3.
[18] A. Hussain, A. Ahmed, H. Magsi, and R. Tiwari, “Adaptive GNSS
receiver design for highly dynamic multipath environments,”
IEEE Access, vol. 8, pp. 172481–172497, 2020, doi:
10.1109/ACCESS.2020.3024890.
[19] yu yasyukevich, A. yasyukevich, E. Astafyeva, Y. V
Yasyukevich, A. Yasyukevich, and E. Astafyeva, “How
modernized and strengthened GPS signals enhance the system
performance during solar radio bursts How modernized and
strengthened GPS signals enhance the system performance
Lermontov str,” doi: 10.1007/s10291-021-01091-5ï.
[20] M. Karaim, M. Elsheikh, and A. Noureldin, “GNSS Error
Sources,” Multifunct. Oper. Appl. GPS, May 2018, doi:
10.5772/INTECHOPEN.75493.
[21] A. Hussain, F. Akhtar, Z. H. Khand, A. Rajput, and Z. Shaukat,
“Complexity and Limitations of GNSS Signal Reception in Highly
Obstructed Enviroments,” Eng. Technol. Appl. Sci. Res., vol. 11,
no. 2, pp. 6864–6868, Apr. 2021, doi: 10.48084/ETASR.3908.
[22] D. N. Aloi, M. Alsliety, and D. M. Akos, “A methodology for the
evaluation of a GPS receiver performance in Telematics
applications,” IEEE Trans. Instrum. Meas., vol. 56, no. 1, pp. 11–
24, Feb. 2007, doi: 10.1109/TIM.2006.887190.
[23] A. Khodabandeh, S. Zaminpardaz, and N. Nadarajah, “A study on
multi-GNSS phase-only positioning,” Meas. Sci. Technol., vol. 32,
no. 9, p. 095005, May 2021, doi: 10.1088/1361-6501/ABECED.
[24] A. Hussain, A. Ahmed, H. Magsi, J. B. Soomro, S. S. H. Bukhari,
and J. S. Ro, “Adaptive data length method for gps signal
acquisition in weak to strong fading conditions,” Electron., vol. 10,
no. 14, Jul. 2021, doi: 10.3390/ELECTRONICS10141735.
[25] R. Langley, “Dilution of Precision,” GPS Solutions.
https://www.semanticscholar.org/paper/Dilution-of-Precision-
Langley/c04cd3dc971734ebe2ba1aa113e417e8b28b248c
(accessed Aug. 01, 2022).
[26] C. Specht, P. S. Dabrowski, J. Pawelski, M. Specht, and T. Szot,
“Comparative analysis of positioning accuracy of GNSS receivers
of Samsung Galaxy smartphones in marine dynamic
measurements,” Adv. Sp. Res., vol. 63, no. 9, pp. 3018–3028, May
2019, doi: 10.1016/J.ASR.2018.05.019.
[27] “Measures of Accuracy in the Navstar/GPS: 2drms Vs. CEP.”
https://www.ion.org/publications/abstract.cfm?articleID=12097
(accessed Aug. 01, 2022).

Seventh International Conference on Aerospace Science & Engineering 258

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

GNSS TEC BASED IONOSPHERIC ANOMALIES

RELATED TO THE 2019 MW 6.2 THAILAND
EARTHQUAKE
Shireen Mushtaq Rimsha Ashraf Dr. Munawar Shah
Institute of Space Technology Institute of Space Technology Institute of Space Technology
Islamabad, Pakistan Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected] [email protected]

Punyawi Jamjareegulgarn
King Mongkut’s Institute of Technology
Ladkrabang,
Prince of Chumphon Campus, Thailand
[email protected]

Abstract— Detecting seismic ionospheric perturbation by the GNSS TEC. [7] To recognize the seismo-ionospheric
Global Navigation Satellite System (GNSS) is an inevitable perturbations, the TEC obtained from GNSS stations made a
approach and has proven to be a new field in the investigation great contribution.[8] In the recent years, the GPS-based
of Earthquake (EQ) precursors. In this paper, the temporal TEC anomalies before EQ have been revealed using TEC
Total Electron Content (TEC) anomalies are investigated from
maps and TEC measurements over distinct stations.[9]
the International GNSS Service (IGS) stations around the
epicenter of the November 20, 2019, Thailand EQ (M w 6.2), Anomalous TEC variations were investigated for the
that occurred on UTC=23:50:43 h. Seismo ionospheric Tohoku EQ 2011 over the epicenter beyond the upper
perturbations are estimated with TEC during thirty-one days confidence bound.
(twenty days prior and ten days followed by EQ). The solar
and geomagnetic activity indices are checked in the context of
They found pre-seismic TEC anomalies in quiet
EQ. Quiet geomagnetic activity is observed on the day of EQ. geomagnetic periods. [10] Ionospheric TEC perturbation
However, Kp is larger than 3 immediately after the EQ. Results was examined between 2000- 2019 EQs of different
depict the TEC perturbation measured at three GNSS stations. magnitudes (M > 6.0). They distinguished that the positive
Furthermore, clear positive anomalies are detected within anomalies are greater than the negative anomalies.
three days prior to EQ Mw 6.2, (-3, -2, and -1 days) during Furthermore, these studies show that the daily variations in
quiet geomagnetic activities. Since, Kp >3 after the EQ. TEC anomalies can fund considerable precursors earlier to
Therefore, the ionospheric perturbation is related to the worldwide EQs (M ≥ 6). [11] TEC perturbation of
geomagnetic storms. twenty-four EQs was investigated in China of different
magnitude between Mw (5-8). The TEC results show the
positive and negative variations before EQ. Moreover, the
Keywords: Ionospheric perturbation; Thailand Earthquake;
Total Electron Content. perturbation was not steady and not corresponding to the EQ
magnitudes.[12] The correlation between ionospheric TEC
I. INTRODUCTION variations and three faults are examined of twenty eight EQ
that have greater than 7 magnitudes.[13] The procedures and
For several years, the seismic events have been studied
their relationship with ionospheric anomalies from various developments of ionospheric seismology for the 2008
satellites and a connection has also set up with the Wenchuan EQ and 2011 Japan EQ are obtained from GPS
lithosphere during the main shock preparation period. [1] TEC.[14] The global ionospheric map (GIM) is used to
The ionospheric perturbations related to EQ inside the observe anomalies in the TECU related to thirty five EQs
breeding zone have been accounted from GNSS TEC as pre- that their magnitudes greater than 6.0 happened in China
and post-seismic precursors. [2] between May 1998 - April 2008.[15]

There are several studies about ionospheric anomalies In this paper, ionospheric disturbances are monitored in the
induced by EQ. For example, the recent studies included the form of variation over the Thailand EQ of Mw 6.2 using
investigation of abnormal ionospheric signatures related to TEC to provide an exacting evidence about seismo
Mw greater than 6.5 EQs in New Zealand from TEC.[3] ionospheric anomalies.
Moreover, the investigation of possible statistical relation II. CASE STUDY
among ionospheric TEC anomalies and the global inland
EQs (Mw ≥ 6.0) during 2003-2017 was also proposed.[4] The EQ of November 20, 2019 in Thailand (Mw 6.2)
Earliest ionospheric perturbation associated with EQ was happened on UTC = 23:50:43h at the geographical location
monitored in 1964 after the Alaska EQ (Mw 9.2, March 28, of Latitude (19.45°N) and Longitude (101.36°E), and
1964).[5] To evaluate the ionospheric disturbances, the pre- shallow hypo central depth 10 km according to United
and post-seismic Taiwan EQ (Mw = 6.4) obtained from States Geological Survey (USGS). According to Global
GNSS TEC during quiet geomagnetic storm conditions.[6] Disaster Alert and Coordination System or GDACS
They reported the possible relationship between the EQ and (https://gdacs.org/report.aspx?eventid=
1194401&episodeid=1279805&eventtype=EQ), there are

Seventh International Conference on Aerospace Science & Engineering 259

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 1 The EQ Location (Mw 6.2) is denoted by red filled star. The black color filled circles denote the GNSS stations, while the red circle is the preparation zone
estimated by Dobrovolsky equation, which is about 463.5 km.

no reports of casualties or damage. Some people near the Ŕ = 100.43*M (1)

epicenter and surrounding areas felt aftershocks throughout
where the radius of seismic region is represented by Ŕ (unit:
the morning period. Residents of Chiang Mai, Thailand
have long felt the shakes but no major damage. Up to 8,000 km) and M is the EQ magnitude. Here, the M is 6.2, hence,
people were exposed to strong shaking. the EQ radius is about 463.5 km as depicted in Fig. 1.

Table 1 Slant TEC (STEC) is monitored by the Global Positioning

System (GPS) satellite. It is computed in expressions of
The information for the selected EQ in Thailand. TEC unit. The STEC is evaluated via following equations.
No Date Time Latitude Longitude M Depth [17]
. (UTC) (N) (E) w (km)
( )
1 20/11/ 23:50:43 19.45 101.36 6.2 10
2019 (2)

( ) (3)
III. DATA COLLECTION AND METHODOLOGY
In the present paper, the ionospheric disturbances related where f1 and f2 denotes carrier wave frequencies, λ1 and λ2
to Mw 6.2 EQ are studied from the GNSS TEC obtained from denotes wavelengths, l1 and l2 are the carrier phases, N1 and
IGS stations working in the EQ breeding zone. The various N2 depicts the uncertainties of the signal, b1 and b2
morphological parameters (magnitude, depth, and location) symbolizes the carrier phases of the signal, P 1 and P2
of the EQ are attained from USGS denotes the pseudo-ranges, and is the random residuals of
https://earthquake.usgs.gov/earth /search/. Different the signal. To transform STEC into VTEC, the below
geomagnetic indices (i.e., Dst, Kp, Ap and F10.7) are equation is used. [18]
analyzed to verify the space weather conditions (SWC) and
quiet days earlier and afterward the EQ. For this task, the VTEC = STEC × cos (arcsine (RsinZ/(R+H))), (4)
source data during October 31-November 30, 2019 are
where the Earth radius is denoted by R, H is the IPP
regained from International Services of Geomagnetic Indices
(ISGI) (http://isgi.unistra.fr/data _download.php) and OMNI ionospheric height, and Z represents the zenith angle of the
web NASA https://omniweb.gsfc. nasa.gov/form/dx1.html. satellite.

To examine the temporal anomalies, the daily GNSS TEC To recognize the anomalies, we analyze the TEC and DTEC
data are retrieved from the IGS stations within EQ for twenty days prior and ten days next to the main shock
geographical position. Furthermore, the TEC data are day. To recognize the TEC anomalies, the confidence
obtained via this website bounds are computed by using the below equations.
http://www.ionolab.org/webtec/single.html?locale=en. Upper bound (UB) =m+2σ (5)
These IGS stations are in operation during the resultant EQ
preparation zones. The EQ radius is calculated by Lower bound (UB) =m-2σ (6)
Dobrovolsky equation as shown in Eq. (1). [16]

Seventh International Conference on Aerospace Science & Engineering 260

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 2. The Geomagnetic indices afore and next to the Mw 6.2. Here, the Kp shows active geomagnetic storm after the EQ day.

where the standard deviation and the median of the TEC the EQ day (Fig. 3). Because the storm is active after the EQ
time series are denoted by σ and m, respectively. Moreover, day,
the TEC deviation from the upper bound (DTEC) is
we cannot say confidently about the pre-seismic ionospheric
computed by subtracting the original TEC from the anomalies. The TEC deviations (DTEC) from upper bound
respective UB. are calculated for the selected GNSS stations to indicate
TEC Deviation (DTEC) =TEC original – Upper Bound (7) anomalous ionospheric situations and also filtered out the
anomalous days. The DTEC for three IGS stations before
To differentiate either the ionospheric disturbances related and after the EQ can be computed using Eq. (7) and shown
to EQ or geomagnetic storm, this procedure has been used. in Figs. 4-5. We use the temporal and continuous wavelet
All calculations were executed by MATLAB. All maps and transform method to find out the DTEC. The positive
graphs were also plotted with MATLAB functions. anomalies are observed from the results of DTEC. But these
are storm induced TEC anomalies and no clear evidence of
EQ induced ionospheric perturbations came out.
IV. RESULTS AND DISCUSSION
V. CONCLUSION
In this paper, The GNSS TEC data were used to examine the
We investigated the ionospheric TEC anomalies for the
seismo ionospheric anomalies preceding the Thailand EQ. Thailand EQ by analyzing the possible reasons for these
The ionospheric parameters are affected by solar and anomalies based on solar and geomagnetic storm conditions.
geomagnetic storm conditions. The main purpose of The Dst, F10.7, Kp and Ap indices were examined. The
including solar and geomagnetic storm conditions in this studied indices showed the EQ occurred during the quiet
research work is to differentiate whether ionospheric geomagnetic activities. Ionospheric TEC data were
anomalies are induced by the storm or seismic process. In determined by using GNSS data. The TEC results show that
Fig. 2, the solar conditions and geomagnetic storm show the EQ anomalies appeared three days earlier and on the EQ day
EQ happened in the quiet geomagnetic activities (Kp < 3, during quiet geomagnetic storm conditions. But, the storm is
Dst < −10 nT, Ap < 7 and F10.7 < 100). However, the Kp is active after the EQ day and these ionospheric anomalies are
greater than 3 immediately after the EQ day. associated to the storm. As for the computed DTECs, the
positive anomalies are more than the negative anomalies.
The lower and upper bounds are calculated at the same
The data showed the clear ionospheric anomalies over the
period as the geomagnetic storm data by using daily TEC
three days before and on the EQ day.
variation data of the selected GNSS stations near to the EQ
ACKNOWLEDGEMENTS
epicenter. Ionospheric anomalies are examined in the TEC
values from three IGS stations named as CMUM (Chiang
The Author is obliged to the USGS,ISGI,OMNI web NASA
Mai University, Muang), CUUT (Chulalongkorn University,
and IONOLAB for providing data about the TEC data,
Thailand) and CPNM (Chumphon University, Muang). As
respectively.
seen in Fig. 3, the CMUM station showed the positive TEC
anomalies on days -3, -2, -1 before the EQ day and on the
EQ day. But, the storm became active after the EQ day;
therefore, there are no EQ induced pre-seismic anomalies at
the CMUM station. Likewise, the CUUT station showed the
positive TEC anomalies on days -2, -1 before the EQ day
and on the EQ day. The CPNM station also showed the
positive TEC anomalies on day -1 before the EQ day and on

Seventh International Conference on Aerospace Science & Engineering 261

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 3. The TEC measured at three GNSS stations (CMUM, CUUT, CPNM) along with confidence bounds for indicating the seismic
ionospheric anomalies.

Fig. 4. The DTEC values beyond the upper bounds for all three GNSS stations.

Seventh International Conference on Aerospace Science & Engineering 262

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 5. Wavelet spectrums of DTEC values beyond the upper bounds depicting the TEC disturbances before and on the main shock day for all
three Thailand stations (i) CMUM, (ii) CUUT and (iii) CPNM stations.

DTEC associated with upper bounds from all three station.

Seventh International Conference on Aerospace Science & Engineering 263

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

REFERENCES
[1] A. Kiyani et al., “Seismo ionospheric anomalies possibly
associated with the 2018 Mw 8.2 Fiji earthquake detected with
GNSS TEC,” J. Geodyn., vol. 140, p. 101782, 2020, doi:
10.1016/j.jog.2020.101782.
[2] M. Shah and S. Jin, “Pre-seismic ionospheric anomalies of the
2013 Mw = 7.7 Pakistan earthquake from GPS and COSMIC
observations,” Geod. Geodyn., vol. 9, no. 5, pp. 378–387, 2018,
doi: 10.1016/j.geog.2017.11.008.
[3] M. A. Adil, A. Abbas, M. Ehsan, M. Shah, N. A. Naqvi, and A.
Alie, “Investigation of ionospheric and atmospheric anomalies
associated with three Mw >6.5 EQs in New Zealand,” J. Geodyn.,
vol. 145, no. April, p. 101841, 2021, doi:
10.1016/j.jog.2021.101841.
[4] F. Zhu and Y. Jiang, “Investigation of GIM-TEC disturbances
before M ≥ 6.0 inland earthquakes during 2003–2017,” Sci. Rep.,
vol. 10, no. 1, pp. 1–7, 2020, doi: 10.1038/s41598-020-74995-w.
[5] K. Davies and D. M. Baker, “Ionospheric effects observed around
the time of the Alaskan earthquake of March 28, 1964,” J.
Geophys. Res., vol. 70, no. 9, pp. 2251–2253, 1965, doi:
10.1029/jz070i009p02251.
[6] A. R. Abbasi, M. Shah, A. Ahmed, and N. A. Naqvi, “Possible
ionospheric anomalies associated with the 2009 Mw 6.4 Taiwan
earthquake from DEMETER and GNSS TEC,” Acta Geod.
Geophys., vol. 56, no. 1, pp. 77–91, 2021, doi: 10.1007/s40328-
020-00325-1.
[7] S. Pulinets and D. Davidenko, “Ionospheric precursors of
earthquakes and Global Electric Circuit,” Adv. Sp. Res., vol. 53,
no. 5, pp. 709–723, 2014, doi: 10.1016/j.asr.2013.12.035.
[8] Y. Zou and T. Zhao, “Ionospheric anomalies detected by GPS
TEC measurements during the 15 August 2007 Peru earthquake,”
2010 Int. Conf. Microw. Millim. Wave Technol. ICMMT 2010, no.
August 2007, pp. 1216–1219, 2010, doi:
10.1109/ICMMT.2010.5525006.
[9] Y. Zou, “Ionospheric anomalies detected by GPS TEC
measurements during two strong earthquakes in Indonesia,” 2010
2nd Conf. Environ. Sci. Inf. Appl. Technol. ESIAT 2010, vol. 2,
pp. 199–202, 2010, doi: 10.1109/ESIAT.2010.5567309.
[10] M. Akhoondzadeh, “Anomalous TEC variations associated with
the powerful Tohoku earthquake of 11 March 2011,” Nat.
Hazards Earth Syst. Sci., vol. 12, no. 5, pp. 1453–1462, 2012,
doi: 10.5194/nhess-12-1453-2012.
[11] M. Ulukavak, M. Yalçınkaya, E. T. Kayıkçı, S. Öztürk, R.
Kandemir, and H. Karslı, “Analysis of ionospheric TEC
anomalies for global earthquakes during 2000-2019 with respect
to earthquake magnitude (Mw≥6.0),” J. Geodyn., vol. 135, no.
September 2019, 2020, doi: 10.1016/j.jog.2020.101721.
[12] F. Ke, Y. Wang, X. Wang, H. Qian, and C. Shi, “Statistical
analysis of seismo-ionospheric anomalies related to Ms > 5.0
earthquakes in China by GPS TEC,” J. Seismol., vol. 20, no. 1,
pp. 137–149, 2016, doi: 10.1007/s10950-015-9516-x.
[13] M. Ulukavak and M. Yalçlnkaya, “INVESTIGATION of the
RELATIONSHIP between IONOSPHERIC TEC ANOMALY
VARIATIONS and FAULT TYPES before the
EARTHQUAKES,” ISPRS Ann. Photogramm. Remote Sens.
Spat. Inf. Sci., vol. 4, no. 4W4, pp. 383–388, 2017, doi:
10.5194/isprs-annals-IV-4-W4-383-2017.
[14] S. Jin, G. Occhipinti, and R. Jin, “GNSS ionospheric seismology:
Recent observation evidences and characteristics,” Earth-Science
Rev., vol. 147, pp. 54–64, 2015, doi:
10.1016/j.earscirev.2015.05.003.
[15] J. Y. Liu et al., “Seismoionospheric GPS total electron content
anomalies observed. before the 12 May 2008 Mw7.9 Wenchuan
earthquake,” J. Geophys. Res. Sp. Phys., vol. 114, no. 4, pp. 1–10,
2009, doi: 10.1029/2008JA013698.
[16] I. P. Dobrovolsky, S. I. Zubkov, and V. I. Miachkin, “Estimation
of the size of earthquake preparation zones,” Pure Appl. Geophys.
PAGEOPH, vol. 117, no. 5, pp. 1025–1044, 1979, doi:
10.1007/BF00876083.
[17] S. Jin, R. Jin, and J. H. Li, “Pattern and evolution of seismo-
ionospheric disturbances following the 2011 Tohoku earthquakes
from GPS observations,” J. Geophys. Res. Sp. Phys., vol. 119, no.
9, pp. 7914–7927, 2014, doi: 10.1002/2014JA019825.
[18] J. A. Klobuchar, “Ionospheric Time-Delay Algorithm for Single-
Frequency GPS Users,” IEEE Trans. Aerosp. Electron. Syst., vol.
AES-23, no. 3, pp. 325–331, 1987, doi:
10.1109/TAES.1987.310829.

Seventh International Conference on Aerospace Science & Engineering 264

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

INVESTIGATION OF SEISMO IONOSPHERIC ANOMALIES

ASSOCIATED WITH THE MW 6.7 MONGOLIA
EARTHQUAKE
Rimsha Ashraf Shireen Mushtaq
Institute of Space Technology Institute of Space Technology
Islamabad, Pakistan Islamabad, Pakistan
[email protected] [email protected]

Abstract—The seismo-ionospheric anomalies associated with Table 1

the Mw 6.7 Mongolia Earthquake (EQ) occurred on Jan 11 Detailed information about EQ
,2021, are investigated from the Global Navigation Satellite Mw Date Time LAT LONG Depth EPZ
System (GNSS) based Total Electron Content (TEC) acquired o o
(UT) () () (km) (km)
from the stations in seismic region. To observe the EQ
anomalies, it is mandatory to take into account the variations 6.7 11/01/2021 21:32:59 51.12 100.47 10 760.3
in solar and geomagnetic storm activities, the present study has
the worldwide geomagnetic storm level, particularly the
Disturbance Storm Time (Dst), Kp, Ap, F 10.7 indices prior to Furthermore, the anomalies in ionosphere happened due to
EQ and after the EQ. It is clear that the EQ happened during EQs, and raised upward to the ionosphere in the form of
the quiet geomagnetic conditions Kp < 3. But the storm was vertical electric fields into the ionosphere. Similarly, the
active five days before the EQ. The electromagnetic variations evolution of gases such as Radon from the epicenter
within the ionosphere related to EQ are statistically examined produces some chemical modifications which prevail up to
in TEC from IGS stations. Our results from TEC clearly show the ionosphere [8]. Additionally, some seismic ionospheric
that the ionospheric anomalies are induced several days before
anomalies are also revealed as false indications and that are
the EQ near the epicenter and positive anomalies appear on -
4th Day, -3rd Day, -2nd Day, - 1 Day and on main shock day. not considered as EQ precursors [9]. Moreover, the
Based on the obtained results, anomalies in TEC are more anomalies appearing in the ionosphere are also due to the
likely related to the geomagnetic storm indices due to its strong solar storm. Therefore, it is mandatory to investigate
abnormality during the EQ preparation period. Furthermore, the geomagnetic indices a few days prior and afterward the
enhancement and depletion in TEC prior to EQ supports the EQ in the course of the EQ zone. The basic aim of including
presence of storm anomalies over the seismic zone. geomagnetic indices is to investigate whether the
Keywords—Earthquake; Global Navigation Satellite ionospheric disturbance is due to geomagnetic storm or EQ.
Systems; Total Electron Content; Geomagnetic Storm Many studies shows that geomagnetic storm is completely
quiet during the three days prior to EQ [3],[10-12], In order
to further clarify geomagnetic storm anomalies ,the main
I. INTRODUCTION goal of this paper is to statistically analyzed and examined
GPS TEC to monitor anomalies generated in ionosphere for
The seismic activity that yields ionospheric effects have twenty days before and ten days next to the main shock and
gained attention for many years, the GNSS TEC has been additional idea of this study is to explore the variance among
proved to be an efficient tool to monitor and estimate the seismic and geomagnetic storm anomaly.
ionospheric anomalies which are mainly associated with
large magnitude and shallow hypo central EQs within the II. CASE STUDY
EQ preparation zone(EPZ). [1-3]. The partially ionized zone The Mw 6.7 occurred on January 11, 2021 at 33 Km SSW of
of Earth’s upper atmosphere is the ionosphere, it is the Turt in Mongolia at UT =21:32:59 hours having
portion of the atmosphere where ionized molecules and free geographical location of Latitude (51.21 oN) and Longitude
electron density have an impact on the propagation of (100.47oE), as reported by United States Geological Survey
electromagnetic waves of radio frequency. Researches have (USGS) and it is presented in Table.1. This EQ has a
been conducted during the last decade to monitor changes shallow hypo central depth of 10 km, as shown in Fig. 1.
that took place in the ionosphere before or after the main There are about 19000 people living within 100 km of the
shocks of large magnitude EQs. Therefore, the importance epicenter. USGS PAGER estimates that up to 4000 people
of examining the ionosphere layer as precursor in EQ were exposed to strong shaking
prediction research has been revealed [4]. Geomagnetic https://www.gdacs.org/report.aspx?eventid=1251188
storms, atmospheric circulations and magnetic activities are &episodeid=1352410&eventtype=EQ.
the main source for triggering ionospheric disturbances
[5].Recently, GPS TEC assists in studying the relationship
between EQ and ionospheric precursors [6]. The ionospheric
disturbance was first reported above the EQ epicenter in
1964 after the Alaska EQ (M w 9.2, March 28, 1964) [7].

Seventh International Conference on Aerospace Science & Engineering 265

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Where STEC denotes Slant Total Electron Content, f1

(1575.42 MHz) is carrier phase frequency of GPS signals
and f2 (1227.60MHz) is frequency (f2) at second end.
P1 and P2 are the pseudo ranges measured in L1 and L2,
respectively. S represents uncertainty of signal and b
denotes the instrumental biases, differential code bias is
represented by d. Moreover,
transformation from Slant TEC into Vertical TEC is done by
the next equation.
) (3)

Where, R denotes the Earth’s radius and H indicates single

layer height thickness. Furthermore, the elevation angle of
the satellite is represented by Z.
To probe into the ionospheric abnormalities and its
relationship with the EQ, the TEC data is acquired from
permanent three IGS stations functioning in EPZ.The named
as KSTU, ULAB, and IRKT was monitored to compute the
upcoming EQ relation with ionosphere. The TEC from IGS
stations are obtained from
http://www.ionolab.org/indephp?p
age=webtec&language=en.
To recognize any TEC anomalies trends, confidence bounds
of TEC values are calculated [15]. TEC more than the upper
bound (UB) is measured as a positive anomaly and TEC less
than the lower bound (LB) as a negative anomaly. The limits
are computed by the median (m) and standard deviation (σ)
of twenty pre and ten post days after the EQ day by the
Fig.1 Geographical position of the EQ epicenter. Black star shows the following equation. Then bounds are computed by below
coordinates of the epicenter and the colored portion depicts the intensity of
EQ. The location map is taken from USGS.
equations [11].
Upper Bound = m + 2σ, (4)
DATA COLLECTION AND METHODOLOGY Lower Bound = m − 2σ. (5)

In the study, we have investigated the ionosphere TEC Where, m denotes median and standard deviation is depicted
variations from IGS stations related to the shallow hypo by σ.
central Mw 6.7 Mongolia EQ.These IGS stations operating To confirm the anomalous ionospheric situations the
in the EPZ. Various morphological parameters of the EQ Deviation TEC (DTEC) is calculated. Deviation TEC can be
comprise the magnitude, depth, and location and calculated for selected IGS stations by this formula:
information about the EQ are obtained from USGS website
https://EQ.usgs.gov/EQs/search/.Moreover,To distinguish TEC deviation = TEC original –Upper Bound (6)
EQ and storm anomalies in the TEC, the examination of
these indices has equivalent significance, the data of IV. RESULT
geomagnetic and solar indices (Kp, Dst, Ap and F10.7) are In this paper, The ionospheric anomalies from IGS stations
acquired from ISGI https://isgi.unistra.fr/datadow nload.php within EPZ are considered within the twenty days earlier
OMNI web NASA https://omniweb.gsfc. nasa.gov and ten days next the EQ of the Mw 6.7, Mongolia EQ. Fig.
/form/dx1.html 2 show the geographical positions of EQ and IGS stations
The examination of TEC anomalies related to EQ, requires that lie in the EPZ.Our outcomes are to delineate the
that IGS stations work inside the EPZ. EPZ, as estimated by geomagnetic storm (Kp>4) or EQ induced perturbations
Dobrovolsky et al.[14] in the form of ρ = (10 0.43M) km. within five days pre-and post-EQ the Mw 6.7 event over the
Radius of the EPZ is represented by ρ, where M symbolizes epicentral regions. Geomagnetic storms are quiet on the day
the magnitude of EQ and its radius is computed by the of EQ but there is a monotonous increase in geomagnetic
Dobrovolsky equation and it is about 760.3 km, as shown in storm analysis a few days before EQ and right after EQ,
Figure 2. meaning the storm is active beyond 5 days pre-and right
TEC is estimated from IGS stations in the form of Slant after main EQ day (Fig. 3). The confidence bounds are
TEC (STEC). It is computed with the ray path of the calculated by using daily TEC data from selected IGS
recorded signal in the unit of TEC, where 1 TEC is stations within the EPZ. The TEC retrieved from IGS
equivalent to 1016 e/ m2. [13]. stations displays irregularities within five days before the
EQ (e.g., -4, -3, -2, -1).Ionospheric anomalies are on the EQ
( )
day. The KSTU station showed positive TEC anomalies on
(1) days -2, -1 before, on the EQ day and one day after EQ.The
ULAB station showed positive TEC anomalies on days -4,-
( ) (2)
3,-2,-1, on the EQ day and one day after EQ as shown in the

Seventh International Conference on Aerospace Science & Engineering 266

 Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 2 Location of EQ epicenter and GNSS permanent stations functioning in the EPZ. The red star depicts the site of the epicenter where small colored circles
represent IGS stations and the red line denotes the EPZ estimated by the Dobrovolsky equation.

Fig. 4. If an original TEC is more than either of the lower or stations in Mongolia functioning nearby epicenter from GPS
upper boundary then we declare that an abnormal signal is TEC, has been discussed. The geomagnetic storm is active
detected. As per previous results positive ionospheric beyond 5 days before EQ of Mw 6.7 (Kp>4) ; it shows no
anomaly is TEC above upper confidence bound (Fig. 4)[10]. association with the epicenter. Investigation on TEC
To provide further proof in the preparation period of M w 6.7 improves our results which depicts that enhancement and
EQ, TEC variances are examined from three IGS stations depletion in TEC prior to EQ supports the induction of
(i.e. IRKT,KSTU and ULAB stations).Temporal and spatial storm anomalies over the seismogenic zone. Lastly, since
graphs show TEC anomalies on the EQ day and three days present study EQ has been preceded by geomagnetic
afore the EQ day (Fig. 4 and 5).The ionospheric anomalies activity, so looking for genuine ionospheric perturbations
are clearly observable in IGS stations. instigated by seismic activity is the task. Finally, the
Similarly, the deviations in TEC among the observed TEC comprehension to what extent EQs and geomagnetic
and upper bound are discovered pre and post the EQ, as conditions are related to one another is undeniable for the
shown in Fig. 5. Enhancement in TEC prior to EQ supports field of EQ prediction.
that anomaly may be due to the Kp>3 storm over EPZ. Fig.5 ACKNOWLEDGEMENT
depicts DTEC anomalies that are because of geomagnetic
storms and no association of activating ionospheric The authors would like to thank the IGS community for the
disturbance has been observed. Fig. 6 showed the cross availability of TEC data. Additionally, we are also grateful
wavelet spectrum of TEC anomalies 20 days before and 10 to the USGS, ISGI and OMNI WEB for making data
days after main shock. It shows that an increase in TEC available for the EQ, solar indices, Geomagnetic indices,
disturbances appeared on the 4th day before the main event. and TEC data.
After the 4th day, the disturbances were prolonged and
enlarged, and attained its maximum value at EQ day. Then,
the TEC disturbances disappeared after 2 days of main
shock. Our findings indicate that abnormal disturbances of
ionospheric TEC lie between -4, -3, -2, -1 and day of EQ.
[16-17] Further, storms are active 5 days before main shock
and right after it, so anomalies is induced by storm.
V. CONCLUSION
In the present paper, seismo ionospheric perturbations of
shallow hypo central EQ having Mw 6.7 monitored by IGS

Seventh International Conference on Aerospace Science & Engineering 267

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 3 Geomagnetic and solar storm conditions during observational period. (a) Dst values, (b) Kp index, (c) F10. 7, (d) Ap.
Red lines depict EQ day.

Fig. 4 TEC measurement recorded at three IGS stations (KSTU, ULAB, IRKT )for thirty days twenty days prior and ten days next to EQ.

Seventh International Conference on Aerospace Science & Engineering 268

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 5 Spatial TEC (upper) and DTEC (lower) recorded at three IGS stations (KSTU, ULAB, IRKT ) shows variations due to geomagnetic storm twenty days
before and ten days after EQ.

Seventh International Conference on Aerospace Science & Engineering 269

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Fig. 6 Wavelet spectrum of DTEC retrieved from IGS station (a) IRKT (b) KSTU and (c) ULAB stations depicts TEC disturbances before the main event
of EQ.

Seventh International Conference on Aerospace Science & Engineering 270

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Start

Selection of seismic event (Mongolia dated 11/01/2021)

Selection DOY for twenty days prior EQ and ten days after

Downloading and processing GPS data from

IGS stations (KSTU,IRKT,ULAB )

TEC calculations Analysis of geomagnetic data and solar activity

Statistical anomaly detection

Anomaly appear due to geomagnetic storm

Yes

No further analysis needed

Fig. 7 Flowchart depicts methodology adopt in Proposed paper

Seventh International Conference on Aerospace Science & Engineering 271

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

REFERENCES
[1] K. Davies and G. K. Hartmann, “Studying the ionosphere with [16] L. M. He, L. X. Wu, A. De Santis, S. J. Liu, and Y. Yang, “Is
the global positioning system,” Radio Sci., vol. 32, no. 4, pp. there a one-to-one correspondence between ionospheric
1695–1703, 1997, doi: 10.1029/97RS00451. anomalies and large earthquakes along Longmenshan faults?,”
[2] Z. Hong-ping and Z. H. U. Wen-yao, “Real - t i m e Monitoring Ann. Geophys., vol. 32, no. 2, pp. 187–196, 2014, doi:
and Prediction of Ionospheric Electron Content by Means of GPS 10.5194/angeo-32-187-2014.
t .,” vol. 45, no. 2, pp. 331–337, 2004, doi: [17] A. Ali, C. Sheng-Chang, and M. Shah, “Continuous wavelet
10.016/j.chinastron.2004.07.00. transformation of seismic data for feature extraction,” SN Appl.
[3] M. Shah et al., “Total electron content anomalies associated with Sci., vol. 2, no. 11, pp. 1–12, 2020, doi: 10.1007/s42452-020-
earthquakes occurred during 1998–2019,” Acta Astronaut., vol. 03618-w.
175, no. May, pp. 268–276, 2020, doi:
10.1016/j.actaastro.2020.06.005.
[4] MUP lukavac, “Deep learning for ionospheric TEC forecasting at
mid-latitude stations in Turkey,” Acta Geophys., vol. 69, no. 2,
pp. 589–606, 2021, doi: 10.1007/s11600-021-00568-8.
[5] E. Astafyeva, “Ionospheric Detection of Natural Hazards,” Rev.
Geophys., vol. 57, no. 4, pp. 1265–1288, 2019, doi:
10.1029/2019RG000668.
[6] A. Kiyani, M. Shah, A. Ahmed, H. Hussain, and S. Hameed,
“Seismo ionospheric anomalies possibly associated with the 2018
M w 8 . 2 Fiji earthquake detected with GNSS TEC,” J. Geodyn.,
vol. 140, no. July, p. 101782, 2020, doi:
10.1016/j.jog.2020.101782.
[7] K. Davies and D. M. Baker, “Ionospheric effects observed around
the time of the Alaskan earthquake of March 28, 1964,” J.
Geophys. Res., vol. 70, no. 9, pp. 2251–2253, 1965, doi:
10.1029/jz070i009p02251.
[8] S. Pulinets, “Ionospheric precursors of earthquakes: Recent
advances in theory and practical applications,” Terr. Atmos.
Ocean. Sci., vol. 15, no. 3, pp. 413–435, 2004, doi:
10.3319/TAO.2004.15.3.413(EP).
[9] S. Jin, G. Occhipinti, and R. Jin, “GNSS ionospheric seismology:
Recent observation evidences and characteristics,” Earth-Science
Rev., vol. 147, pp. 54–64, 2015, doi:
10.1016/j.earscirev.2015.05.003.
[10] M. Freeshah et al., “Analysis of atmospheric and ionospheric
variations due to impacts of super typhoon mangkhut (1822) in
the northwest Pacific Ocean,” Remote Sens., vol. 13, no. 4, pp. 1–
20, 2021, doi: 10.3390/rs13040661.
[11] A. R. Abbasi, M. Shah, A. Ahmed, and N. A. Naqvi, “Possible
ionospheric anomalies associated with the 2009 Mw 6.4 Taiwan
earthquake from DEMETER and GNSS TEC,” Acta Geod.
Geophys., vol. 56, no. 1, pp. 77–91, 2021, doi: 10.1007/s40328-
020-00325-1.
[12] M. A. Adil, A. Abbas, M. Ehsan, M. Shah, N. A. Naqvi, and A.
Alie, “Investigation of ionospheric and atmospheric anomalies
associated with three Mw >6.5 EQs in New Zealand,” J. Geodyn.,
vol. 145, no. March, p. 101841, 2021, doi:
10.1016/j.jog.2021.101841.
[13] H. Li, Y. Yuan, Z. Li, X. Huo, and W. Yan, “Ionospheric electron
concentration imaging using combination of LEO satellite data
with ground-based GPS observations over China,” IEEE Trans.
Geosci. Remote Sens., vol. 50, no. 5 PART 2, pp. 1728–1735,
2012, doi: 10.1109/TGRS.2011.2168964.
[14] I. P. Dobrovolsky, S. I. Zubkov, and I. Miachkin, “Estimation of
the size of earthquake preparation zones,” Pure Appl. Geophys.
PAGEOPH, vol. 117, no. 5, pp. 1025–1044, 1979, doi:
10.1007/BF00876083.
[15] M. Shah and S. Jin, “Pre-seismic ionospheric anomalies of the
2013 Mw = 7.7 Pakistan earthquake from GPS and COSMIC
observations,” Geod. Geodyn., vol. 9, no. 5, pp. 378–387, 2018,
doi: 10.1016/j.geog.2017.11.008.

Seventh International Conference on Aerospace Science & Engineering 272

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Author Index

Author Page Author Page

Aamer Shahzad 66 Dr. Muhammad Adnan Shahid 163

Abdul Hanan 114 Dr. Muhammad Tahir 86

Abdul Hanan Hamid 156 Dr. Munawar Shah 71, 76, 260

Abdur Rafeh Abbasi 156 Dr. Najam Abbas Naqvi 48

Absar Ahmed Khan 56 Dr. Rehan Mahmood 79, 86, 91

Adnan Maqsood 106 Dr. Rizwan Mughal 91

Ahmad Hammad 197 Dr. Salma Zainab Farooq 151

Ammar Mushtaq 213 Dr. Shahid Mahmood 168

Areeba Azhar 124 Dr. Waqas A. Qazi 177,181,191

Arif Hussain 253 Ejaz Nazir 71

Arooj Shafique 91 Erman Senturk 156

Asifullah Khan 140 Faheem Akhtar Chachar 253

Asim Khan 71 Faisal F. Khan 232

Asim Mehmood 19 Gohar Majeed 19
Asim Shahzad 124 Hadia Batool 228

Asiya Khan 114 Hammad Munawar 146

Asra Imtiaz 52 Hamza Akbar Khan 220

Awais Munawar Qureshi 96 Hamza Bin Babar 71

Ayesha Hakim 114 Hamza Malik 146

Ayesha Malik 173 Hassan Sadiq 48, 187

Azeema Azhar 117 Hina Magsi 253

Beenish Batul 42 Hira Aftab 42

Daboor Sultana 76, 156 Hussna Zainab 52

Dr. Ali Sarosh 6 Ijaz Ali 202

Dr. Amir Rashid Chaudhry 206 Ilyas Mahmood Qureshi 197

Dr. Arslan Ahmed 253 Imaan Shahid Raja 206

Dr. Asad Ali 241 Imtiaz Nabi 151

Dr. Hayat Muhammad Khan 91 Irtiza Ali 15

Dr. Hossein Raza Hamdani 52 Javaid Khurshid 112

Dr. Imran Tahir 206 Javid Rashid Ch 206

Seventh International Conference on Aerospace Science & Engineering 273

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Kashif Mehmood 19 Muneef U. Rehman 232

Khalid Mahmood 114 Muzaffar Habib 26

Khudeja Salim 232 Nabeel Uddin Sheikh 206

Labeeqa Zaheer 181 Naeem Akhter 112

M. Naveed Tahir 124 Nauman Latif 197

Mamoona Waheed 173 Noman Ahsan 202

Maria Siddiqua 112 Noor Ul Huda Atif 248

Mazhar Shehzad 201 Obaid Ur Rehman 2

Moaz Rao 197 Osama Ali Ahmed Awan 6

Mohamed Freeshah 71 Oscar Lionel 104

Mohammad Usman 48, 187 Pawan Kumar 104

Mohsin Javaid 96 Punyawi jamjareegulgarn 260

Moosa Bin Saif 168 Rafia Sarwar 241

Muhammad Abubakar Arshad 86 Raja Muhammad Sohail Khan 146

Muhammad Adeel Khan 19 Rehan Mahmood 83

Muhammad Ahmad 202 Rimsha Ashraf 260, 266

Muhammad Anas 26 Rizwana Kausar 96

Muhammad Arif Arshad 126 Sabih Ahmad Khan 220

Muhammad Arqim Adil 71, 156 Saddam Hussain Khan 140

Muhammad Asfandyaar Arshad 91 Sadia 197

Muhammad Ayaz Ahmad 220 Saif Ullah Khan 163
Muhammad Danish Majeed 163 Salem Mohammed Alohani 26

Muhammad Ibtisam Babar 117 Salman Amir 206

Muhammad Ihtisham Babar 117 Salman Qadri 114
Muhammad Mubeen 168 Saqib Mehdi 156

Muhammad Saifullah 114 Saqlain Abbas 194, 202

Muhammad Saqlain 163 Shahzad Hameed 2

Muhammad Shaarif Younas 197 Sheeza Khalid 177

Muhammad Shoaib Ul Hassan 48, 187 Shireen Mushtaq 260, 266

Muhammad Umair Shahid 202 Shujaat Hussain 96

Muhammad Waleed Khan 140 Subhe Seeme 163

Muhmmad Ahsan khan 132 Suleman Qamar 140

Muhmmad Waqas 241 Sumaira Abbasi 228

Seventh International Conference on Aerospace Science & Engineering 274

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

Sunnyaha Saeed 151 Usama Saleem 15

Syed Hadi Hussain Shah 253 Usman Azeem 79

Syed Irtiza Ali Shah 220 Usman Imran 91

Syed Muhmmad Basit Ali 15 Zahid Mehmood 6

Taimur Ali Shams 220 Zohair Anjum 206

Tanveer Ahmad 114 Zuhaib Hassan 2

Tariq Abbas 191 Zukhruf Liaqat 66

Tayyaba Shafaq 206 Zulkarnain Abbas 202

Usama Rehman 19

Seventh International Conference on Aerospace Science & Engineering 275

Institute of Space Technology Conference Proceedings
Islamabad, Pakistan ICASE 2021

PROCEEDINGS
Seventh International Conference on Aerospace
Science and Engineering
December 14 – 16, 2021
Institute of Space Technology
Islamabad, Pakistan

Organized by
National Center of GIS and Space Applications (NCGSA)

Seventh International Conference on Aerospace Science & Engineering 276

 Pa r t n e r s & S p o n s o r s

C o n f e r e n c e S e c r e ta r i at
Institute of Space Technology, 1 Islamabad Expressway,
Islamabad 44000, Pakistan
+92-51-9075799, +92-333-2662857
[email protected]
www.ncgsa.org.pk | www.ist.edu.pk
/IST.islamabad | /ICASE.IST | /ncgsa.ist