Edge AI: Real-Time Decision-Making

A Technical seminar report submitted in partial fulfillment of the requirement for the award
of the degree of

BACHELOR OF TECHNOLOGY

IN
COMPUTER SCIENCE AND ENGINEERING

Under the esteemed guidance of

Dr Radha Seelaboyina
Associate Professor

By

Alle Sai Shivanandha (21R11A0557)

Department of Computer Science and Engineering

Geethanjali College of Engineering and Technology(Autonomous)

Accredited by NAAC with A+ Grade: B.Tech. CSE, EEE, ECE accredited by NBA Sy.
No: 33 & 34, Cheeryal (V), Keesara (M), Medchal District, Telangana – 501301

i
 Geethanjali College of Engineering and Technology (Autonomous)
Accredited by NAAC with A+ Grade : B.Tech. CSE, EEE, ECE accredited by NBA Sy. No: 33 & 34,
Cheeryal (V), Keesara (M), Medchal District, Telangana – 501301

Department of Computer Science and Engineering

CERTIFICATE

This is to certify that the B.Tech Technical Seminar report entitled “ Edge AI: Real-
Time Decision-Making” is a bonafide work done by Alle Sai Shivanandha
(21R11A0557), in partial fulfillment of the requirement of the award for the degree of
Bachelor of Technology in “Computer Science and Engineering” from Jawaharlal
Nehru Technological University, Hyderabad during the year 2024-2025.

Dr. Radha Seelaboyina HoD – CSE

Associate Professor Dr A Sree Lakshmi
Professor

Coordinator

Examiner(s)

ii
 ACKNOWLEDGEMENT

It is with profound gratitude and sincere appreciation that I extend my heartfelt thanks to all
those who have played a significant role in the successful completion of this technical
seminar.

First and foremost, I express my deep sense of respect and gratitude to our Honourable
Chairman, Mr. G. R. Ravinder Reddy, for his constant encouragement and for nurturing
a culture of academic excellence and innovation within the institution.

I would also like to express my sincere thanks to Dr. S. Udaya Kumar, Director, for his
visionary leadership and continued support, which have provided the ideal environment and
motivation for academic enrichment through activities such as this seminar.

My heartfelt appreciation goes to Dr. K. Sagar, Principal, for his steadfast guidance,
infrastructural support, and motivation that have facilitated the successful presentation of this
seminar.

I am deeply grateful to Dr. A. Sree Lakshmi, Head of the Department of Computer

Science and Engineering, for her academic leadership, encouragement, and continuous
support throughout the seminar preparation.

I extend my sincere thanks to the Technical Seminar Coordinators, Dr. S. Vishwanath

Reddy, Ms. M. Swaroopa, for their organized planning, timely guidance, and constructive
feedback which contributed immensely to the smooth and effective completion of the
seminar.

A special note of gratitude is reserved for my project guide, Dr. S. Radha, Associate
Professor, whose expert supervision, insightful suggestions, and dedicated mentorship have
been instrumental in shaping the direction and outcome of this work.

Lastly, I am ever grateful to my parents and family for their unconditional love, support,
and motivation. Their belief in me has been a source of strength and inspiration throughout
this journey.

With genuine appreciation, I acknowledge every individual who, in one way or another,
contributed to the successful completion of this technical seminar.

With warm regards,

Alle Sai Shivanandha
21R11A0557
Department of Computer Science and Engineering
Geethanjali College of Engineering and Technology
iii
 ABSTRACT

Edge AI enables real-time decision-making by processing data locally on IoT devices,

reducing latency and bandwidth usage while enhancing privacy. It powers applications like
autonomous vehicles, industrial automation, and healthcare, where speed and efficiency are
critical. This seminar explores lightweight AI models, federated learning, and hardware
acceleration, addressing challenges like energy constraints, data heterogeneity, and security.
Case studies, such as Tesla’s Autopilot and smart factory predictive maintenance, highlight
its impact. Emerging trends like TinyML and 6G enabled edge computing showcase the
future of decentralized intelligence, making Edge AI a cornerstone of next-gen autonomous
systems.

iv
 LIST OF ABBREVIATIONS

S. No. Abbreviation Full Form

1 IOT Internet of Things

2 AI Artificial Intelligence

3 ML Machine Learning

4 DL Deep Learning

5 VR Virtual Reality

6 MR Mixed Reality

7 BIM Building Information Modeling

8 5G Fifth Generation

9 6G Sixth Generation

10 NLP Natural Language Processing

11 SDK Software Development Kit

12 MQTT Message Queuing Telemetry Transport

13 GPU Graphics Processing Unit

14 ONNX Open Neural Network Exchange

15 NCS Neural Compute Stick

v
 TABLE OF CONTENTS

S. No. Contents Page No.

1 Introduction 1

2 Literature Survey 2

3 Objectives 4

4 Methodology 5

5 System Architecture 6

6 Implementation 8

7 Usecase 10

8 Advantages and Limitations 14

9 Future Scope 16

10 Conclusion 17

11 Bibliography 18

vi
 1. INTRODUCTION

Edge Artificial Intelligence (Edge AI) is transforming the way intelligent systems operate
by enabling AI processing directly on edge devices, such as sensors, smartphones,
microcontrollers, and embedded systems. Unlike traditional AI models that rely heavily on
centralized cloud computing, Edge AI processes data locally at the source. This decentralized
approach reduces the reliance on cloud infrastructure and significantly minimizes latency in
decision-making processes. As a result, systems powered by Edge AI can respond to real-world
stimuli in real time, which is essential for time-sensitive applications across various industries.

One of the key advantages of Edge AI is its ability to enhance bandwidth efficiency. In
conventional cloud-based models, massive amounts of raw data are transmitted from edge devices
to remote servers for processing, which consumes a significant portion of available network
bandwidth. Edge AI mitigates this issue by analyzing and filtering data locally, sending only
relevant or summarized information to the cloud if necessary. This not only conserves bandwidth
but also reduces overall operational costs. Additionally, Edge AI contributes to data privacy and
security, as sensitive data can be processed on-site without ever leaving the device, reducing
exposure to potential breaches or unauthorized access.

Edge AI plays a critical role in the success of Internet of Things (IoT) and autonomous
systems, which often require immediate decision-making capabilities. In the context of self-driving
vehicles, Edge AI enables the real-time analysis of sensor data—such as images, radar, and lidar—
allowing the vehicle to make split-second decisions that ensure passenger safety. Similarly, in
industrial automation, edge-enabled machines can detect anomalies, trigger alerts, and take
corrective action without waiting for instructions from a centralized system. In smart healthcare,
wearable devices equipped with Edge AI can monitor vital signs and issue real-time alerts in
response to abnormal conditions, providing immediate care to patients in critical situations.

1
 2. LITERATURE SURVEY

1. Edge Intelligence: The Confluence of Edge Computing and Artificial Intelligence

Authors: Y. Zhou, F. Liu, Z. Xu, et al.

Published in: ACM Computing Surveys (CSUR), 2019

Summary: This survey explores the integration of AI and edge computing, discussing
architectures, frameworks, and challenges. It emphasizes how processing at the edge
significantly reduces latency and supports real-time inference, especially in IoT and cyber-
physical systems.

2. Edge Computing for the Internet of Things: A Survey

Authors: F. Bonomi, R. Milito, J. Zhu, et al.

Published in: IEEE Internet of Things Journal, 2018

Summary: This paper discusses edge computing as a key enabler for the IoT ecosystem.
It highlights how localized data processing improves responsiveness, reduces cloud
reliance, and supports latency-sensitive applications like autonomous vehicles and smart
cities.

3. Edge AI: On-Demand Accelerated Inference at the Edge

Authors: H. Li, K. Ota, M. Dong

Published in: IEEE Network, 2020

Summary: This survey investigates how AI models can be optimized and deployed on
resource-constrained edge devices. It focuses on techniques for real-time decision-making
in autonomous systems and edge-based deep learning, including model compression and
hardware acceleration.

2
4. A Survey on Edge Computing Systems and Tools

Authors: S. Yi, Z. Hao, Z. Qin, Q. Li

Published in: Proceedings of the IEEE, 2018

Summary: This work explores platforms and tools supporting edge computing. It provides
an overview of real-world applications where low-latency decision-making is essential,
including smart manufacturing and autonomous control systems.

5. Edge Intelligence for the Internet of Things: A Survey

Authors: J. Chen, X. Ran

Published in: IEEE Internet of Things Journal, 2019

Summary: This paper outlines AI challenges and opportunities at the edge, especially for
real-time analytics in IoT environments. It details use cases in healthcare, transportation,
and industrial IoT, where AI-driven edge decisions are crucial.

6. Artificial Intelligence at the Edge

Authors: M. Chiang and T. Zhang

Published in: Nature Electronics, 2019

Summary: This article reviews the technological evolution and impact of AI at the edge,
particularly for low-latency, energy-efficient inference in mobile and embedded systems.
It showcases applications in robotics and smart healthcare.

3
 3. OBJECTIVES

Edge AI is revolutionizing real-time decision-making by bringing artificial intelligence

directly to edge devices, such as sensors, cameras, and embedded systems. This approach
drastically reduces latency, enabling immediate responses without relying on distant cloud servers.
It also enhances data privacy and bandwidth efficiency by processing information locally. In IoT
and autonomous systems, such as self-driving cars and smart healthcare, Edge AI ensures
continuous and intelligent operation even in low-connectivity environments. As a result, Edge AI
is emerging as a critical technology for the next generation of responsive, efficient, and secure
intelligent systems

• To explore the integration of Artificial Intelligence with Edge Computing for faster
and decentralized data processing in real-time applications.
• To analyze how Edge AI reduces latency by performing computation closer to the data
source, eliminating delays associated with cloud-based systems.
• To examine the role of Edge AI in enabling real-time decision-making for critical
applications such as autonomous vehicles, industrial automation, and smart healthcare.
• To investigate how Edge AI enhances data privacy and security by limiting the need to
transmit sensitive data to centralized cloud servers.
• To evaluate the impact of Edge AI on bandwidth efficiency, by processing and filtering
data locally before sending only essential information to the cloud.
• To identify hardware and software challenges involved in deploying AI models on
resource-constrained edge devices.
• To review current architectures and frameworks that support the implementation of
Edge AI in IoT ecosystems.

4
 4. METHODOLOGY

1. Literature Review
Study existing research on Edge AI, IoT, and real-time systems to understand current
trends, technologies, and challenges.
2. Problem Identification
Define the specific real-time decision-making issue or scenario (e.g., autonomous vehicle
obstacle detection, industrial fault prediction).
3. System Architecture Design
Develop a suitable Edge AI architecture combining sensors, edge devices (e.g.,
Raspberry Pi, NVIDIA Jetson), and AI models.
4. Model Selection and Optimization
Choose lightweight AI models (e.g., MobileNet, TinyML) and optimize them for
deployment on edge devices using techniques like model pruning or quantization.
5. Implementation
Deploy the AI model on the edge device and integrate it with sensors or IoT modules for
real-time data collection and inference.
6. Testing and Evaluation
Measure system performance in terms of latency, accuracy, energy efficiency, and
decision response time under different conditions.
7. Result Analysis
Compare edge-based performance with cloud-based systems to highlight improvements
in real-time decision-making.
8. Conclusion and Future Work
Summarize findings and propose future improvements such as advanced edge hardware,
5G integration, or federated learning for collaborative intelligence.

5
 5. SYSTEM ARCHITECTURE

1. Data Sources
• What it does: Captures real-world data in real time.
• Examples:
o Cameras (for video/image recognition)
o IoT sensors (temperature, humidity, motion)
o Microphones (for audio analysis)
• Purpose: Acts as the input for the AI system.

2. Data Preprocessing
• What it does: Cleans and formats raw data before inference.
• Techniques:
o Noise reduction (e.g., smoothing sensor data)
o Normalization/scaling (bringing values to a standard range)
o Image resizing, frame selection (for vision tasks)
• Purpose: Ensures data is usable by the AI model and reduces errors.

3. Edge AI Inference
• What it does: Performs real-time predictions at the edge.
• Tools & Frameworks:
o TensorFlow Lite, ONNX Runtime, OpenVINO, NVIDIA TensorRT
• Model Types:
o Lightweight ML/DL models (MobileNet, Tiny-YOLO, etc.)
• Purpose: Avoids sending all data to the cloud, reducing latency and bandwidth use.

6
4. Real-Time Decision Logic
• What it does: Analyzes the output of the AI model and decides on actions.
• Examples:
o If "person detected", then "sound alarm"
o If "temperature > threshold", then "shut down motor"
• Purpose: Converts AI output into actionable steps using business logic or embedded
rules.

5. Actuation / Control Systems

• What it does: Triggers physical or digital responses based on decisions.
• Examples:
o Turn motors on/off
o Activate cooling systems
o Send alerts/notifications
• Purpose: Executes control in the physical world instantly.

6. Edge Data Logging & Temporary Cache

• What it does: Stores recent data and results temporarily at the edge.
• Benefits:
o Works offline in case of network issues
o Supports rollback or troubleshooting
• Purpose: Enables quick data recovery and ensures system resilience.

7
 6. IMPLEMENTATION

1. Define the Use Case

• Identify the real-time decision problem (e.g., object detection, anomaly detection,
predictive maintenance).
• Clarify expected outcomes and latency requirements.
2. Choose Appropriate Edge Hardware
• Select devices with computing capability (e.g., NVIDIA Jetson, Raspberry Pi,
Google Coral, Intel NCS).
• Ensure compatibility with sensors and support for AI inference.
3. Collect and Prepare Data
• Gather real-world data from edge sensors or historical datasets.
• Label data if necessary (especially for supervised learning).
• Preprocess: clean, normalize, or transform data for model training.
4. Train and Optimize AI Models
• Train ML/DL models on cloud or local machine.
• Optimize for edge using quantization, pruning, or converting to formats like:
o TensorFlow Lite (.tflite)
o ONNX (.onnx)
5. Deploy Model to Edge Device
• Use SDKs and runtime environments like:
o TensorFlow Lite Runtime
o ONNX Runtime
o OpenVINO Toolkit
• Test inference speed and accuracy on real-time input.
6. Integrate Decision Logic
• Write application logic to interpret AI output (e.g., if-confidence-thresholds).
• Add conditional flows to trigger responses.
8
7. Connect Actuators or Output Systems
• Interface with physical systems (motors, relays, LEDs) or software (dashboards,
alerts).
• Use GPIO, MQTT, or APIs based on the device and use case.
8. Enable Edge Logging and Offline Caching
• Log critical decisions locally in case of cloud disconnection.
• Store recent input/output for debugging or future training.

9. Set Up Cloud Sync (Optional)

• Use cloud to:
o Push software updates or improved models
o Aggregate and analyze long-term data
o Send notifications or integrate with enterprise systems
10.Test and Monitor

• Perform stress testing and edge-case handling.

• Use real-time monitoring dashboards or cloud-based health checks.
• Ensure the system meets real-time performance and accuracy targets.

9
 7. USECASE

Edge AI is transforming industries by enabling real-time decision making directly on devices

without relying on constant cloud connectivity. By processing data locally, Edge AI reduces
latency, enhances privacy, and ensures quicker responses—making it ideal for applications that
demand immediate action. From smart surveillance and autonomous vehicles to healthcare
monitoring, industrial automation, agriculture, and more, Edge AI empowers systems to act
instantly on sensor or camera data, improving efficiency, safety, and user experience across a wide
range of real-world scenarios.
1. Smart Surveillance
• Use Case: Real-time threat detection and anomaly identification using AI-powered
cameras.
• How Edge AI Helps: Processes video locally to detect suspicious activity instantly,
reducing the need to stream large video files to the cloud and improving response time
while enhancing privacy.

2.Autonomous Vehicles
• Use Case: Real-time object detection, navigation, and decision-making.
• How Edge AI Helps: Processes sensor data (e.g., camera, radar, lidar) on-board to make
split-second driving decisions, crucial for safety and performance, as seen in Tesla’s
Autopilot system.

3.Industrial Automation & Predictive Maintenance

• Use Case: Early detection of equipment faults and automated control of machinery.
• How Edge AI Helps: Analyzes vibration, temperature, and operational data at the
machine level to predict failures before they happen, avoiding downtime and reducing
maintenance costs.

10
4.Smart Agriculture
• Use Case: Crop monitoring, soil health analysis, and pest detection using sensors and
drones.
• How Edge AI Helps: Enables on-site data processing to provide farmers with real-time
insights, optimizing irrigation, pesticide use, and harvest timing without relying on
constant internet access.

5.Smart Cities
• Use Case: Traffic management, energy optimization, and public safety monitoring.
• How Edge AI Helps: Local edge devices can analyze traffic flows, monitor air quality,
or detect accidents, helping city infrastructure respond dynamically and efficiently.
6.Healthcare Monitoring Devices
• Use Case: Real-time patient monitoring through wearables and medical sensors.
• How Edge AI Helps: Processes biometric data locally (heart rate, oxygen levels, etc.) to
detect abnormalities immediately and alert healthcare providers or emergency services
without latency.

7.Retail & Smart Stores

• Use Case: Personalized shopping experiences and inventory management.
• How Edge AI Helps:
o Uses local cameras and sensors to track customer movement, analyze behavior,
and suggest personalized promotions in real time.
o Monitors shelf stock levels and alerts staff or triggers automatic restocking
without relying on cloud-based systems.

8.Drones for Real-Time Inspection

• Use Case: Infrastructure inspection, agriculture monitoring, and disaster assessment.
• How Edge AI Helps:
11
 o Processes aerial footage on the drone to detect cracks in buildings, crop health
issues, or damaged areas in real time.
o Enables autonomous flight decisions based on what’s observed, even in remote
areas with limited connectivity.

Conclusion

Edge AI is revolutionizing a wide range of industries by enabling real-time, intelligent decision-

making directly at the source of data. From autonomous vehicles ensuring passenger safety, to
smart surveillance enhancing public security, and healthcare devices delivering instant patient
insights—Edge AI minimizes latency, reduces bandwidth usage, and improves privacy. Its
applications in industrial automation, smart agriculture, retail, smart cities, and drone inspections
showcase its versatility and critical role in next-generation systems. By processing data locally,
Edge AI not only optimizes performance and efficiency but also lays the foundation for a future
driven by decentralized, responsive, and intelligent technologies.

Real-World Impact

Safety:

• Instant Threat Detection: Edge AI in cars, drones, and factories detects dangers (like
obstacles, system failures, or intrusions) in milliseconds, allowing immediate action and
preventing accidents.

• Medical Emergency Response: AI-powered devices at the edge (like wearable monitors)
can instantly spot critical health conditions (like heart failure) and trigger emergency
protocols without delay.

• Public Safety Systems: Smart surveillance cameras can autonomously recognize weapons,
fires, or unsafe crowd behavior in real time — improving emergency responses.
12
Trust:

• Data Privacy: Since data is processed locally on the device instead of being sent to a cloud
server, sensitive personal or corporate information stays private, building user confidence.

• Reliable Performance: Edge AI continues operating even if the internet connection is

slow or disrupted, ensuring dependable service (essential for sectors like healthcare,
military, and transportation).

• Explainable Decisions: Localized AI models can be more transparent and easier to audit,
helping users and regulators trust AI-driven actions.

Sustainability:

• Reduced Energy Use: Processing at the edge minimizes the need for heavy, energy-
intensive cloud computing — saving bandwidth and cutting down on carbon footprint.

• Smarter Resource Management: In industries like agriculture and smart cities, Edge AI
enables real-time optimization of resources (like water, electricity, fuel) based on current
conditions.

• Longer Device Lifespans: On-device AI can monitor and predict equipment maintenance
needs, reducing waste by extending machine life and minimizing unnecessary
replacements.

13
 8. ADVANTAGES AND LIMITATIONS
Advantages:

• Ultra-lowLatency
Decisions are made instantly without relying on distant servers.

• Better Privacy and Security

Data stays on the device, reducing the risk of interception or leaks.

• High Reliability
Works even without a stable internet connection.

• Reduced Bandwidth Usage

Less data needs to be sent to the cloud, saving costs and reducing network congestion.

• Energy Efficiency
Saves energy compared to constantly uploading and processing large datasets in
centralized clouds.

• Personalization
Edge AI can adapt in real time to the user's behavior and preferences locally.

• Scalability
Easier to deploy many devices across remote areas without needing huge backend
infrastructure.

Limitations:

• Limited Processing Power

Edge devices (like sensors or smartphones) often have less computing power compared to
cloud servers.

14
• Model Size Constraints
Complex AI models must be compressed or optimized to fit into small devices, which
can reduce their accuracy.

• Hardware Costs
Specialized edge hardware (like AI chips) can be expensive to design and deploy at scale.

• Maintenance and Updates

Managing and updating thousands of distributed devices is challenging.

• Security Risks
Devices can still be physically tampered with or hacked if not properly secured.

• Data Fragmentation
Data is spread across many devices, making it harder to perform centralized analytics or
training without additional orchestration.

• Development Complexity
Building and optimizing AI models for edge deployment requires specialized skills and
tools.

15
 9. FUTURE SCOPE

The future of Edge AI for real-time decision-making is incredibly promising and transformative
across industries. As devices become smarter and networks faster, Edge AI will enable
autonomous systems like self-driving cars, drones, and robots to operate safely and independently.
In healthcare, wearable devices will offer instant diagnostics and monitoring, improving patient
outcomes. Factories and smart cities will rely on Edge AI for predictive maintenance, traffic
control, and energy optimization, creating more efficient and sustainable environments. With 5G
and 6G networks, real-time collaboration between edge devices and the cloud will become
seamless. Privacy and security will also improve through decentralized learning methods like
federated learning. The growth of hyper-personalized services in retail, entertainment, and fitness
will be fueled by faster on-device intelligence. Additionally, Edge AI will help reduce carbon
footprints by minimizing reliance on massive cloud data centers. Critical fields like defense, space
exploration, and disaster response will benefit from rapid, local decision-making without latency.
Overall, Edge AI is shaping a future that is smarter, faster, greener, and more human-centric.

16
 10. CONCLUSION

In conclusion, Edge AI is redefining the future of technology by enabling instant,

intelligent decision-making closer to where data is generated. It offers unmatched advantages in
safety, trust, and sustainability across industries like healthcare, transportation, manufacturing, and
beyond. Despite certain limitations such as hardware constraints and security challenges, the
potential of Edge AI to create faster, smarter, and more sustainable systems is undeniable. As
networks evolve and AI models become more efficient, Edge AI will play a central role in building
a safer, greener, and more connected world.

Looking ahead, continuous advancements in AI algorithms, edge hardware, and

communication technologies like 5G and beyond will further unlock the true potential of Edge AI.
Innovations such as federated learning, tiny machine learning (TinyML), and improved edge
processors will make real-time decision-making even more efficient, secure, and accessible. As
industries invest heavily in edge computing solutions, the collaboration between AI at the edge
and the cloud will become more dynamic, enabling systems that are not only reactive but also
predictive and adaptive. Edge AI is set to become a key pillar of the digital future, driving smarter
living, stronger economies, and a more resilient society.

17
 11. BIBLIOGRAPHY

1. Shi, W., Cao, J., Zhang, Q., Li, Y., & Xu, L. (2016). "Edge Computing: Vision and
Challenges." IEEE Internet of Things Journal, 3(5), 637-646.

2. Satyanarayanan, M. (2017). "The Emergence of Edge Computing." Computer, 50(1), 30-

39.

3. Zhang, C., Patras, P., & Haddadi, H. (2019). "Deep Learning in Mobile and Wireless
Networking: A Survey." IEEE Communications Surveys & Tutorials, 21(3), 2224-2287.

4. Li, Y., Ota, K., & Dong, M. (2018). "Deep Learning for Smart Industry: Efficient
Manufacture Inspection System with Fog Computing." IEEE Transactions on Industrial
Informatics, 14(10), 4665-4673.

5. Zhou, Z., Chen, X., Li, E., Zeng, L., Luo, K., & Zhang, J. (2019). "Edge Intelligence:
Paving the Last Mile of Artificial Intelligence with Edge Computing." Proceedings of the
IEEE, 107(8), 1738-1762.

6. Lin, J., Yu, W., Zhang, N., Yang, X., Zhang, H., & Zhao, W. (2017). "A Survey on Internet
of Things: Architecture, Enabling Technologies, Security and Privacy, and Applications."
IEEE Internet of Things Journal, 4(5), 1125-1142.

7. Wang, S., Zhang, X., Zhang, Y., Wang, L., Yang, J., & Wang, W. (2017). "A Survey on
Mobile Edge Networks: Convergence of Computing, Caching and Communications."
IEEE Access, 5, 6757-6779.

8. Abbas, N., Zhang, Y., Taherkordi, A., & Skeie, T. (2018). "Mobile Edge Computing: A
Survey." IEEE Internet of Things Journal, 5(1), 450-465.

9. Premsankar, G., Di Francesco, M., & Taleb, T. (2018). "Edge Computing for the Internet
of Things: A Case Study." IEEE Internet of Things Journal, 5(2), 1275-1284.
18
10. Mahmud, R., Kotagiri, R., & Buyya, R. (2018). "Fog Computing: A Taxonomy, Survey
and Future Directions." Internet of Everything, Springer, 103-130.

19