1

INDUSTRIAL TRAINING REPORT

On
“INTERNET OF THINGS(IOT)”
Done at
“UniCoverge Technologies Pvt Ltd”
by
MAYANK SONI (2100320310074)

Submitted to
Department of Electronics & Communication Engineering
in partial fulfilment of the requirements for the Degree of

Bachelor of Technology
in
Electronics & Communication Engineering

ABES Engineering College, Ghaziabad

Dr. A.P.J. Abdul Kalam Technical University, Lucknow
November, 2024
2

CERTIFICATE RECEIVED
3

ACKNOWLEDGEMENT
We, express our sincere thanks to our supervisor, Dr. Ritu Agarwal of ECE Department, ABES
Engineering College, AK TU University for guiding us right from the inception till successful
completion of the project. We would also like to thank our HOD Mr. Deepak Garg for her
valuable guidance and cooperation to decide the topic and its content.

Signature:

Name : Mayank Soni

Roll No: 2100320310074

Date:
4

TABLE OF CONTENTS
1) Abstract 6
2) Company Profile 7-12
a. About UniCoverge Technologies Pvt Ltd
b. About upskill Campus (USC)
c. The IoT Academy
d. Objectives of this Internship program
e. Glossary
3) Problem Statement 12
4) Existing and Proposed solution 13
5) Proposed Design/ Model 14
a. High Level Diagram
6) Performance Test/Result 15
a. Test Plan/ Test Cases
b. Test Procedure
c. Performance Outcome
7) Work scope 16
8) Conclusion 17
9) Reference 18
5

Figures Description
1. Figure 1 : About Company 7
2. Figure 2 : Applications 8
3. Figure 3 : More Applications 10
4. Figure 4 : Workflow 11
5. Figure 5 : High Level Diagram 14
6

1. ABSTRACT

This report provides details of the Industrial Internship provided by upskill Campus and The IoT
Academy in collaboration with Industrial Partner UniConverge Technologies Pvt Ltd (UCT).

This internship was focused on a project/problem statement provided by UCT. We had to finish
the project including the report in 6 weeks’ time.

My project was about embedded systems and IoT, with a strong focus on object detection. The
project's goal was to create a system capable of identifying things in real time utilizing
embedded sensors and microcontrollers, then communicating the observed data to a cloud
platform via IoT communication protocols for further analysis.

This internship gave me a very good opportunity to get exposure to Industrial problems and
design/implement solution for that. It was an overall great experience to have this internship.
7

2. COMPANY PROFILE

I. About UniConverge Technologies Pvt Ltd

A company established in 2013 and working in Digital Transformation domain and providing Industrial solutions with
prime focus on sustainability and RoI.

For developing its products and solutions it is leveraging various Cutting Edge Technologies e.g. Internet of Things (IoT),
Cyber Security, Cloud computing (AWS, Azure), Machine Learning, Communication Technologies (4G/5G/LoRaWAN),
Java Full Stack, Python, Front end etc.

Figure 1 about company

i. UCT IoT Platform ( )

UCT Insight is an IOT platform designed for quick deployment of IOT applications on the same time providing valuable
“insight” for your process/business. It has been built in Java for backend and ReactJS for Front end. It has support for
MySQL and various NoSql Databases.

• It enables device connectivity via industry standard IoT protocols - MQTT, CoAP, HTTP, Modbus TCP, OPC UA

• It supports both cloud and on-premises deployments.

It has features to
• Build Your own dashboard
• Analytics and Reporting
8

• Alert and Notification

• Integration with third party application(Power BI, SAP, ERP)
• Rule Engine

Figure 2 Application

ii. Smart Factory Platform ( )

Factory watch is a platform for smart factory needs.
9

It provides Users/ Factory

• with a scalable solution for their Production and asset monitoring

• OEE and predictive maintenance solution scaling up to digital twin for your assets.

• to unleased the true potential of the data that their machines are generating and helps to identify the KPIs and
also improve them.

• A modular architecture that allows users to choose the service that they what to start and then can scale to
more complex solutions as per their demands.

Its unique SaaS model helps users to save time, cost and money.
10

Figure 3 more applications

iii. based Solution

UCT is one of the early adopters of LoRAWAN teschnology and providing solution in Agritech, Smart cities, Industrial
Monitoring, Smart Street Light, Smart Water/ Gas/ Electricity metering solutions etc.

iv. Predictive Maintenance

11

UCT is providing Industrial Machine health monitoring and Predictive maintenance solution leveraging Embedded
system, Industrial IoT and Machine Learning Technologies by finding Remaining useful life time of various Machines used
in production process.

Figure 4 Workflow

I. About upskill Campus (USC)

Upskill Campus along with The IoT Academy and in association with Uniconverge technologies has facilitated
the smooth execution of the complete internship process.

USC is a career development platform that delivers personalized executive coaching in a more affordable,
scalable and measurable way.

II. The IoT Academy

The IoT academy is EdTech Division of UCT that is running long executive certification programs in collaboration with
EICT Academy, IITK, IITR and IITG in multiple domains.

III. Objectives of this Internship program

The objective for this internship program was to

☛ get practical experience of working in the industry.

☛ to solve real world problems

☛ to have improved job prospects.

☛ to have Improved understanding of our field and its applications.

☛ to have Personal growth like better communication and problem solving.

12

IV. Glossary
Terms Acronym
Embedded System ES
Internet of Things IoT
Microcontroller MCU
Ultrasonic Sensor US
Cloud Platform CP

3 Problem Statement
This project's goal is to use Internet of Things (IoT) and embedded systems technologies to create and implement
an object detecting system. Automated systems that can precisely identify an object's presence, movement, or
absence inside a specific space are becoming more and more necessary in a wide range of industrial and consumer
applications. Conventional object detection systems are generally restricted by fixed infrastructure or depend on
manual monitoring, which results in inefficiencies and increased costs.
Constructing a dependable, real-time system that can identify objects and transmit the resulting detection data to
a centralized platform for additional processing and decision-making is the difficult part. The system needs to be
scalable, low in false positives and negatives, and energy-efficient in order to function in a variety of situations.
To enable remote monitoring and data analysis, the system should also smoothly interact with current IoT
frameworks.
In order to overcome these obstacles, our project is creating a system that transmits data via Internet of Things
communication protocols while detecting objects using embedded sensors and microcontrollers. The ultimate
objective is to develop an object detection system that is affordable, dependable, and simple to implement. This
system can be applied to a range of industrial applications, including automated manufacturing processes,
inventory management, and security systems.
13

4 Existing and Proposed solution

Typical object detection systems today use simple sensor arrays, like infrared or ultrasonic sensors, or more
conventional techniques, like camera-based monitoring. These techniques have a number of drawbacks even
though they might work well in some situations. More sophisticated camera-based systems are typically more
expensive to install and maintain, which limits their applicability for mass usage. Conversely, basic sensor arrays
sometimes have a restricted range and may have trouble correctly detecting things in settings that are complicated
or varied. Furthermore, there is a chance of inaccuracy because a lot of these devices are quite sensitive to external
variables like temperature and illumination.
Since many traditional systems do not interact with IoT frameworks, real-time data transfer is another major
limitation that makes remote monitoring and data analysis difficult.

On the other hand, my suggested remedy makes use of IoT technologies in conjunction with embedded sensors
and microcontrollers to overcome these constraints. This technology combines infrared and ultrasonic sensors to
identify objects more accurately, even in difficult-to-reach places. IoT protocols like MQTT are used to transfer
the discovered data in real-time to a cloud platform, allowing for remote monitoring and immediate data analysis.
This method not only increases detection accuracy and scalability but also provides a reasonably priced substitute
for pricey camera-based systems.
14

5 Proposed Design/ Model

The suggested object detection system, which makes use of embedded systems and the Internet of Things, is
designed in a methodical manner to guarantee dependable performance and effective development. The first
design phase, the middle development stages, and the final deployment and result comprise the three key stages
of the process.

1. The first stage of design

Choosing the right components and designing the system architecture are the main priorities of the first phase.
Selecting the sensors (such as infrared and ultrasonic) and microcontroller (MCU) that will process the sensor
data falls under this category. Planning the communication protocol—like MQTT—to send the data to a cloud
platform is another aspect of the design.

To visualize the system, a block diagram or flowchart is made, showing how data moves from the sensors to the
microcontroller and then to the Internet of Things platform.

2. Stages of Intermediate Development

Parts of the system are constructed and tested in the development stages. In order to ensure precise object
detection, the microcontroller must first be programmed to interpret sensor data in real-time. Subsequently, the
integration of the IoT communication module facilitates the transmission of the discovered data to the cloud.
This phase involves extensive testing in many settings to improve the accuracy and dependability of the system.
Additionally, power optimization techniques are used to guarantee energy efficiency, which is vital for Internet
of Things devices.

3.Final Results and Implementation

The system is built in its whole and put through one more round of testing and validation in the final stage.
Deployed in a real-world setting, the object detection system continuously identifies items and sends data to the
cloud. The end result is a fully functional system that can be remotely observed using an Internet of Things
platform, offering real-time data analysis and insights. To make sure all goals are achieved, the system's
performance is assessed in comparison to the original problem description.

5.1 High Level Diagram

Figure 5: HIGH LEVEL DIAGRAM OF THE SYSTEM

15

6 Performance Test / Result

A focus on accuracy, power consumption, processing speed, and durability was part of the test plan to make sure
the system would work well in industrial environments.

I. Test Plan/ Test Cases

Testing for Accuracy:

Test Case 1: Confirm that the system recognizes items of different sizes and forms.

Test Case 2: Evaluate false positives and negatives in various environmental settings.

Examination of Power Consumption:

Test Case 3: To verify energy efficiency, measure power consumption while in operation and standby.

Speed of Processing:

Test Case 4: To guarantee real-time performance, measure the interval between object identification and data
transfer.

Examination of Durability:

Test Case 5: Evaluate the system's dependability in high humidity and harsh temperature environments.

II. Test Procedure

Record detection performance under various conditions while placing items within sensor range.

Testing for Power Consumption: Calculate and examine how much energy is used under various operating conditions.

Processing Speed: The amount of time it takes to detect an object and send data.

Durability Testing: Put the system in harsh environments and keep an eye on how it works.

III. Performance Outcome

There were few false positives and negatives and a high accuracy rate from the system.

Power Consumption: Within reasonable power constraints, the system performed well.

Processing Speed: With very little delay, the system satisfied real-time processing requirements.

Durability: In high humidity and extremely cold temperatures, the system operated with dependability.
16

8 Future work scope

Considering the project's time limits, a number of improvements were noted for further investigation. Improving
object detection methods to get higher accuracy, especially in congested or complicated situations, is one
possible area of focus. This might be accomplished by utilizing sophisticated machine learning techniques. To
increase the system's detecting capabilities, adding other sensor types—like image-recognition cameras—is
another way to make improvements. Future development might concentrate on scalability as well, which would
enable the system to oversee numerous sensors or devices inside a bigger network for all-encompassing
industrial monitoring. Improving power efficiency with cutting-edge methods may help decrease energy usage
and prolong battery life. Furthermore, adding advanced data analytics to the cloud platform might help with
improved decision-making and offer deeper insights.
Enhancing the IoT platform's user interface would also make it more accessible and useful. Finally, a thorough
robustness test carried out in harsh circumstances would guarantee the system's dependability in a range of
operational settings. Future improvements like these might greatly increase the system's usefulness and
functionality.
17

9 Conclusion
The ESP32-based object detection project successfully demonstrated that a low-cost, power-
efficient microcontroller can handle basic object detection tasks. Using the ESP32-CAM module,
we achieved real-time image capture and detection with reasonable performance in simple
scenarios. Despite processing limitations, the ESP32 proved effective for lightweight, remote
monitoring applications where affordability and power efficiency are priorities.

While it lacks the power for complex detections, this project shows the ESP32’s potential in IoT
applications like surveillance and automation. Future improvements could include cloud
processing and optimized models to enhance detection accuracy within its constraints.
18

10 REFERENCE
[1] R. M. Xian Wee, T. Connie and M. K. Ong Goh, "Integrating Object Detection and Optical Flow Analysis for Real-time
Road Accident Detection," 2024 International Symposium on Intelligent Robotics and Systems (ISoIRS), Changsha, China,
2024

[2] Z. Li et al., "Aerial Image Object Detection Method Based on Adaptive ClusDet Network," 2021 IEEE 21st International
Conference on Communication Technology (ICCT), Tianjin, China, 2021

[3] L. Lin, C. Shi, N. Wan, W. Lu and K. Gao, "Research and Implementation of Small Object Detection Algorithm for Power
Embedded Devices," 2021 IEEE 2nd International Conference on Information Technology, Big Data and Artificial
Intelligence (ICIBA), Chongqing, China, 2021