Project Titles for BTech EC and EI, Spring 2024-25, ECE Dept.

, NIT Rourkela

Sl. No. Title of the Project -1 Brief Description of the Project-1 Title of the Project-2 Title of the Project-3 Brief Description of the Project -3 Title of the Project-4 Brief Description of the Project -4 Title of the Project-5 Brief Description of the Project -5
Brief Description of the Project-2

This project is essentially the design of signal

conditioning circuit (capacitance to digital
converter) for capacitive micro-sensor interfacing. Design and implementation of power
Different circuit blocks like amplifiers, filters, management circuit for Piezo/electromagnetic
Measurement of high pressure through piezo- Development of ECG/EEG acquisition
modulator etc. need to be designed in Cadence Development of a multi channel embedded energy harvesters to get regulated dc output
Capacitive Sensor Interface, Capacitive High Pressure Instrumentation electric sensor and interface a micro-controller Data Acquisition System for Multi IoT for Healthcare for acquisition system either with discrete components Self-powered Instrumentation
1 analog design environment and simulations to be data acquisition system using STM32 or voltage for driving other system to make it self-
to Digital Converter: (Possible DRDO Collaboration) module STM32 to get the pressure profile inside a sensor Instrumentation System of ECG/EEG or ASIC and development of a mobile System Design
carried out to meet the desired specifications. This other microcontrollers. powered system. It will be based on discrete
gun chamber. App.
project will provide a full chip IC design experience components and circuit design basic
with basic focus on sensor interfacing, may lead to knowledge is required.
good publication. Need to learn Cadence circuit
design tool if not familiar.

Design the hardware for neural On-chip detection of sudden cardiac arrests
Develop a complete resilient FPGA structure for re- Design of defect tolerant controller for Develop a self-reconfigurable payload architecture Low power radiation hardened memory Design low power and robust volatile and Design of low-power ECG
2 Design of fault tolerant FPGA Board: Design of AI accelerator processing unit to speed-up AI and heart attacks (knowledge of signal
programmable payloads. payload in nano-satellites used in low-orbit nano-satellites. design non-volatile memory array Processor
applications processing and digital VLSI needed).

Brain tumour detection and

3 Face Recognition : Computer Vision Computer Vision . . . . . .
classification

User interactive software development Implementation of parallel computing

Development of algorithm for beamforming
Exploring data driven control techniques for Investigation and implementation of physics for basic matrix operations like inverse, techniques for basic matrix operations like
and sensing applications and hardware
wireless communication or electromagnetic informed machine learning. Prerequisite: Basic SVD, eigen decomposition. Prerequisite: matrix inversion, eigen decomposition, SVD,
verification using software defined radio.
analysis. Prerequisite: Basic knowledge on control knowledge of neural networks and machine Basic knowledge of matrix operations, etc. Prerequisite: Basic knowledge of matrix
4 Data Driven Control : Machine Learning Beamforming and Sensing Prerequisite: Basic knowledge of matrix Software Development Parallel Computing Techniques
theory. Target work area: Either wireless learning. Target work area: Either wireless integration and differentiation. Target operations with basic programming knowledge.
algebra and digital communication. Target
communication or electromagnetics. Target communication or electromagnetics. Target work area: Algorithm implementation. Target work area: Parallel computing algorithm
work area: Wireless communication. Target
programming skill: MATLAB/Python/ C/C++. programming skill: MATLAB/Python/C/C++. Target programming skill: implementation. Target programming skill:
programming skill: MATLAB/Python/C/C++.
MATLAB/Python/C/C++. MATLAB/Python/C/C++.

Color information serves as an important

cue for many computer vision
applications. Color names are required in
many real-world tasks such as Color
Blind assistance, image retrieval and
This project will deal with studying as well as
This project will deal with studying as well as This project will deal with studying as well as scene understanding. Color naming in
understanding Human Gait which is unique
understanding whether an given audio file is understanding whether an given video file is images is a challenging problem due to
This project will deal with studying FPGA signature while a person is walking. Nowadays
authentic or tampered. Deep learning algorithms authentic or tampered. Deep learning algorithms shadows, view angles, illumination
AI based models for Deepfake Audio AI based models for Deepfake Video Lightweight Cipher design for IoT architectures for design of Lightweight Deep Learning Based Color Deep Learning methods for Human it is a preferred biometric in comparison to
5 have to be developed in order to firstly detect have to be developed in order to firstly detect conditions and surface reflections.
Detection: Detection applications Ciphers which can be part of wearables and Boundary Recognition Gait Analysis Face and Fingerprint recognition as there are a
whether an audio file has undergone tampering. If whether an video file has undergone tampering. If Manual labelling of color names for real-
other IoT applications lot of challenges in them. This project will help
yes then the next task is to localize which time yes then the next task is to localize which time world images in applications like search
the student get a good industry exposure as
instant of the audio file was tampered. instant of the video file was tampered. engines, fashion parsing and tracking is
well as make them industry ready for a job.
a tedious task and time consuming. Here
the student will investigate Deep learning
methods for color naming task. This
project will give hands on deep learning
which is the latest industry standard.

In semantic communication, the semantic

features are extracted from different
multimedia data for different applications by
the semantic encoder and then channel Reconfigurable intelligent surfaces have
encoding is performed to carry out the data the capability of improving coverage and Underwater navigation and positioning
Different multimedia data like speech, text, and For semantic communication, the semantic features
transmission. To practically realize these error rates in wireless communication. algorithms will be explored and validated
video will be analyzed for different applications and are extracted according to the application. After the
encoders, reconfigurable intelligent surfaces However, a prototype design along with experimentally. In case of any obstructions, the
Semantic Encoder Design using Deep the semantic features will be extracted. For this Channel Encoder Design for Semantic semantic encoder, the data needs to be suitable for Underwater Navigation and
6 RIS for Semantic Communication (RIS) can be used to model the DL layers RIS/STAR-RIS Prototype Design the control of the phase shifts need to be drones and UAV deployed in an underwater
Learning: purpose, the design of semantic encoders will be Communication communication, hence channel encoder design is Positioning
present in the semantic and channel practically realized at cheap cost. Further environment must implement positioning and
proposed to implement the suitable DL essential to restructure the semantic data. This
encoders. The different phase shifts of the RIS in optical domain and simultaneously navigation algorithms to calculate position of
architecture. requires the use of DL architecture.
RIS can be adjusted to model the weights of transmitting and reflecting RIS prototypes the obstruction and create its navigation path.
the layers. This control requires the design of will be designed as part of this project.
a RIS controller on a FPGA board. The
project will require knowledge of Python and
Verilog.

In this project MATLAB will be used for initial In this project MATLAB will be used for initial Development of Smart Assistive Development of AI and GSM based smart Design of compact energy
Modeling and Design of 12 bits SAR Modeling and Design of 12 bits Sigma Design of MEMS sensors and Design of MEMS sensors and energy Design of compact energy harvesters for
7 modeling and subsequent design will be through modeling and subsequent design will be through Technologies for hard of hearing and technologies for hard of hearing and visually harvesters for wireless sensor
ADC: Delta ADC energy harvester harvester wireless sensor nodes or wearable devices
Cadence Cadence visually challenged people challenged people. nodes

Safety and security are major issues where

technology plays a significant role. Monitoring and
surveillance of crucial areas for safety, security
threat detection, and response facilitate information Computer vision has a potential
and intelligence gathering. However, adverse/ application for the automatic detection of
Inadequate exposure in captured images during
complex environmental conditions like haze hinder diseases to assist physicians in
wireless capsule endoscopy (WCE) brings Satellite /Aerial Image suffers from low visibility
tasks by degrading the captured image/video Underwater Image suffers from low visibility diagnosing and treating diseases.
challenges for both manual and automated and limited resolution. Scene content
Deep Learning based Model for Scene scene. Given the environmental condition of India, Deep Learning based Image and limited contrast. Scene content However, medical images of various
diagnosis, as early lesions and abnormalities may Deep Learning based model for Deep Learning based model for Deep Learning based model for improvement is required for effective object
8 Surveillance during Complex an academic initiative focused on building cutting- Enhancement and Color Correction of improvement is required for object detection modalities impose different challenges in
be difficult to detect under low-light conditions. underwater image restoration Medical Image Analysis Satellite /Aerial Image Analysis detection and recognition for different
Environmental Condition : edge technology in monitoring and surveillance for Low-Light Medical Images and recognition for different computer vision image segmentation, anomaly detection,
Therefore, there is a growing need for computer vision tasks aerial scene
safety and security during such environmental tasks. and disease classification. Employing
advancements in low-light WCE image surveillance.
conditions is the need of the hour. For real-time deep neural networks, this research aims
enhancement.
scene surveillance and monitoring for a wide range to detect different diseases from different
of users, a lightweight framework be developed imaging modalities automatically.
which can be run on resource-constrained
platforms to restore the scene for different
computer vision task.

PLL based dielectric resonator feedback oscillator PLL based dielectric resonator feedback oscillator
design to improve the phase noise of the CW design to improve the phase noise of the CW
signal. The PLL based VCO will be implemented signal. The PLL based VCO will be implemented on
on a PCB substrate. This is an ISRO STIC project. a PCB substrate. This is an ISRO STIC project.
This project has a huge potential to establish RF This project has a huge potential to establish RF
RF transistor device modeling is a
Design and development of the and microwave based start-up business. Although Design and development of the and microwave based start-up business. Although
challenging part for any RF designer. To SIW based MIMO transceiver system SIW based MIMO transceiver system design is
prototypes of Phase Locked Loop this project is totally hardware-based, theoretical prototypes of Phase Locked Loop (PLL) this project is totally hardware-based, theoretical SIW based MIMO transceiver
implement various RF active circuit, it is design is an emerging technology. SIW based MIMO antenna design for an emerging technology. Students get
9 (PLL) based Dielectric Resonator knowledge is also required to accomplish the based Dielectric Resonator Oscillators knowledge is also required to accomplish the RF Transistor nonlinear device modeling system design for sub-six GHz 5G
mandatory to have perfect model of the Students get benefited through this 5G Communication benefited through this knowledge of designing
Oscillators (DROs) at 13.8 GHz for project. Since this project has covered frequency (DROs) at 15.26 GHz for Satellite project. Since this project has covered frequency communication
device. This project deals with hardware and knowledge. both software and hardware.
Satellite Communication: range of Ku-band, this will be tremendously helpful Communication range of Ku-band, this will be tremendously helpful
software based circuit design.
to gain the technical knowledge and improve the to gain the technical knowledge and improve the
analytical and troubleshooting skills. Most analytical and troubleshooting skills. Most
importantly, measurement exposure would also be importantly, measurement exposure would also be
beneficial for the overall growth as a start-up RF beneficial for the overall growth as a start-up RF
engineer in the R&D domain. engineer in the R&D domain.
 Accurate extraction of retinal vascular
components is vital in diagnosing and
treating retinal diseases. Achieving precise
The ambient assisted living (AAL) technology
segmentation of retinal blood vessels is
aims to provide more safety and self-
challenging due to their complex structure
sufficiency, while permitting older persons to
and overlapping vessels with other
This project aims to development of Machine The main objective of this project is to identify This project is mainly focused towards live self-dependently in their homes.
anatomical features. Existing deep neural
Learning algorithm for cardiovascular disease different mental disorders or mental conditions like the driver behavior analysis from Monitoring of activities of daily living (ADL) is
ON-BOARD CARDIAC HEALTH DEVELOPMENT OF LIGHTWEIGHT networks often suffer from false positives at PHYSIOLOGICAL SIGNAL RECOGNITION OF HUMAN
detection from ECG signals. The algorithm will be MENTAL HEALTH MONITORING FROM anxiety, stress, emotions from physiological signals multimodal physiological signals like one of the key ideas of AAL. This becomes an
10 MONITORING USING MACHINE MODEL FOR RETINAL VESSEL vessel branches or missing fragile vessel MONITORING AND DRIVER BEHAVIOR FOR AMBIENT
implemented in a Raspberry Pi based system for PHYSIOLOGICAL SIGNALS like EEG, ECG, Respiration signals using Machine EEG,EMG,GSR, ECG,Eye tracker to interesting research idea in modern world,
LEARNING: SEGMENTATION patterns. Also, deployment of the existing BEHAVIOR ANALYSIS ASSISTED LIVING
on-board cardiac health monitoring system Learning based techniques. Finally, this model will ensure proper traffic rule maintainance where condition monitoring of various ADLs
models in resource-constrained
development. be implemented in an on-board system. and avoiding accidents. and their automatic classification is a big
environments is challenging due to their
challenge. This project proposes a new
computational complexity. An attention-
approach for activity recognition using ML
based and computationally efficient
techniques.
architecture needs to be developed to bridge
this gap while enabling improved
segmentation of retinal vascular structures.

This project includes design, development This project includes design and analysis This project includes design and analysis of
This project includes development and testing of
Feature detection of edible using RFID This project work includes implementation of RFID Wearable RF Sensors for Smart Investigations on metasurface-based and testing of metasurface-based antenna Investigations on Phased Array of phased array antenna for defense Investigations on THz antenna and antenna and metasurface as
11 wearable RF sensor/antenna for smart healthcare
and ML: system and detection of food contamination Healthcare unidirectional broadband antenna for IRNSS. It also need balanced power Antenna applications which also need ML for metasurfaces reflector/absorber. The analysis of antenna
system.
divider. solving challenging problems. performance in scattering environment.
In this project the objective is to develop deep The objective is to develop infrared image The objective is to develop distributed signal The objective is to develop different
Domain adaptation based In this project the objective is to track an object
12 Deep learning-based object detection. : learning architecture for different object detection Multimodal image superesolution. superesolution using side information from the Distributed signal processing. processing architecture for estimation and classification algorithms under domain UAV based object tracking.
classification. using a camera mounted on an UAV.
problem. optical image. detection problem. shift.
Next-generation memory like MRAM and
On-chip memory computation is a hardware design ReRAM offer fast read and write speeds, MBIST is a self-testing and repair mechanism
It is an Attack-Resilient Memory Architecture
IoT security is the process of protecting these technique that performs computations within or next low power consumption, and non- which tests the memories through an effective
for the Internet of Things. Complete design
13 Hardware and Security: devices from unauthorized access and threats. (IoT On chip memory computation to the processing elements of a chip. Complete Attack resilient memory architecture Non-volatile memory design volatility, making them suitable for use in Memory Built-in Self-test (MBIST) set of algorithms to detect possibly all the
will be implemented using cadence tool.
and Python programming) design will be implemented using cadence tool. embedded systems. Complete design faults. Complete design will be implemented
(VLSI Circuits)
(VLSI Circuits) will be implemented using CAD tool. using CAD tool. (VLSI Circuits)
(VLSI Circuits)

The primary objective is to develop a

Multistatic Active Drone Surveillance System
(DSS) GPR uses EM wave to detect and
This project aims to develop a classify buried objects inside the shallow
multimodal detection system with subsurface. In this project novel signal
improved real time video surveillance processing software needs to be developed for
capabilities by fusing the camera, buried object detection and classifications.
In this project, focus is to develop a low cost
Avalanche accidents always remain a challenge Infrared, and Radar inputs. There are many steps of signal processing
Developing signal processing software for radar low power SFCW/FMCW Radar platform for
whenever, the troops move around for post Developing conceptual hardware for starting from data acquisition till object
object detection, imaging, classification with user multi-purpose applications like SAR imaging, Development of classical and
RF harvesting based harmonic maintenance, any other operational or Intrusion detection and integrating sensors and subsequent detection and classification. In this process
Radar Simulation Based on friendly interactive visual user interface. Development of low cost RADAR using Surveillance, Object detection and machine learning techniques for
14 active/passive transponder for administrative tasks in avalanche prone areas. classifications based on fusion of laboratory testing of object detection. machine learning can be effectively used in
Matlab/Python Software: Applications: Smart city infrastructure monitoring, SDR module and discrete components classifications. Applications: High resolution Imaging and Detection of Buried
Avalanche Rescue Operation Objective is to develop RF harvesting based multi sensors data Developing deep learning techniques various stages of DSP process for successful
agriculture mapping, intrusion detection, Airport imaging by radar mounted in drone/UAV Object
harmonic active/passive transponder to help for multimodal data fusion and real time object detection and classifications.
Security check in, Automotive Radar, etc. platform, Foliage penetrating radar, GPR,
effective radar detection of buried victim. object detection. Applications: Smart city infrastructure
etc.
Initially FMCW radar-based classifier has inspection and maintenance, Buried Utility
to be integrated with optical camera detection, Landmine and UXO detection,
based detection system. Archaeological inspection, Environment
monitoring, Planetary exploration, Mining
security, etc.

Radio frequency (RF) transmission and reception is The low cost processor as Raspberry Pi will The low cost processor as Raspberry Pi
Doppler weather RADAR is a highly reliable to
RADAR (Radio Detection and Ranging) is possible using SDR which is a firmware to be be utilized to device full algorithm will be utilized to device the product by
provide data to a AI&ML driven forecasting
popularly used for locating and also measures controlled by software. Therefore, attractive features implementation and application in tracking implementation of algorithm on it.
Raspberry Pi (Computer) network system. The Doppler frequency measurements
RADAR system design for tracking velocity (speed with direction) of flying RF transmitter and receiver design are there with SDR to design RF transceivers at any Raspberry Pi based drone tracking drone by means GPS (Global Positioning Application will be to define path to fly for App design for weather radar and
15 design for defining flying path of capture the effect of relation motion. Therefore,
drone/UAV.: objects/drones in three dimensional (3-D) spaces. using software defined radio (SDR). frequency. The developed product will be used for system design. System) and RADAR data fusion. It will be a a drone as will be required for a pre- wind profiling.
drone. such data available for weather forecasting
The product as RADAR hardware design and measuring antenna radiation patterns, transmit new product to be developed and scope of decided mission. It can adaptively
needs to be used in this APP development for
application as tracking of drone it aimed. receive coverage measurement for a given power research is there with it for improving spatial change the route or decide the path with
local weather forecasting.
budget etc. resolution in geographical map generation. AI.

This project will focus on the strategic

This project will focus on leveraging
This project will focus on developing adaptive placement of trusted repeaters, optimization
machine learning for fault management
This project will focus on leveraging advanced algorithms leveraging existing infrastructure in of multipath key propagation, and leveraging This project will explore the hardware
Adaptive Resource Allocation for Intelligent Fault Management in in optical networks, aiming to predict,
ML applications for network traffic machine learning techniques, such as federated optical networks during high-demand scenarios Quantum Key Distribution based Secure node-disjoint paths for secure key and data Hardware implementation of ML implementation of ML algorithms, with a focus
16 Optical Networks in High-Demand Optical Networks Using Machine detect, and localize faults efficiently. By
prediction: learning, for network traffic prediction in optical such as disasters, or major events like Cricket, Optical Networks transmission. The approach will emphasize algorithms on federated learning (FL) for decentralized
Scenarios Using Existing Infrastructure Learning analyzing real-time and historical network
networks. football matches, without deploying additional scalable and efficient methods to enhance data processing.
data, ML models can enable proactive
hardware. security without compromising network
maintenance.
performance.

A portable, real-time ECG measurement system A low-cost function generator has to be designed Taking advantage of ML and available TCAD
A MATLAB front-end has to be designed
has to be implemented by taking market available and implemented using the market-available Different solar cell architectures have to be Design of a MATLAB front-end for Machine learning augmented data, new compact models need to be
Design of portable, low-cost, real-time Design of a low-cost function generator Design and numerical modeling of for the study of semiconductor
17 components to display accurate real-time ECG components for academic laboratory experiments. designed and numerically modeled for the the study of semiconductor compact device modeling for SPICE developed for semiconductor devices
ECG measurement system: for academic laboratory experiments heterojunction Solar Cells devices/nanostructures for academic
data of a person on a computer screen. The The system should generate accurate voltage and performance improvement of the solar cells. devices/nanostructures simulation applicable for faster and more accurate SPICE
hands-on experiments.
system developed should be of low-cost. frequency signals and should be of low-cost. simulation of digital and analog circuits.
There are many instances, where the image have Image Super-Resolution is a machine learning task
very less resolution. Hence, it is difficult to find the where the goal is to increase the resolution of an The aim of this project is to detect the
Image segmentation and classification will be Deep Learning based vehicle detection is
Image Super Resolution for thermal details of the image. The aim of this project is to Image Super Resolution for RGB image, often by a factor of 4x or more, while Satellite/ Aerial Image Segmentation Vehicle Detection in traffic using Ground Penetrating radar signal subsurface object from the received GPR
18 done using deep learning to classify urban, used to improve traffic safety and reduce
images using machine learning. : use machine learning algorithms to increase the images using machine learning. maintaining its content and details as much as and Classification deep learning Processing using Machine learning Signal. It is always a challenge to detect
agricultural, and forested areas. accidents.
resolution.Different activities will be detected using possible. The end result is a high-resolution version objects in the presence of noise and clutter.
deep learning techniques. of the original image.
OTFS and NOMA combined together can
This project focuses on the performance analysis This project aims for average bit/symbol error rate support high data rate for high speed
Performance of IRS-assisted Error rate analysis of RSMA system Performance analysis of OTFS-NOMA
19 of IRS-assisted communication systems in terms of analysis of rate splitting multiple access (RSMA) communication devices. Performance NA NA NA
communication systems: under generalized fading system
outage probability, error rate, etc. under generalized fading environments. analysis of such system models is the prime
focus of this project.

Design of Capacitive Gap To design a dc-dc power To design a dc-dc power converter as Design of current sensors using hall
20 To measure small gap for non contact type gauge. Design of low power amplifier circuit To design linear amplifier circuit To design low pass filter circuit To design filters Do design a ac current sensor
measurement circuit : converter regulator application effect transducer
 Space-Time Adaptive Reconfigurable
Initially, student will select suitable indoor Intelligent Surfaces (STAR-RISs) can
First, design a representative NTN system
THz channel models Incorporating Material enhance signal quality and improve
architecture, considering different types of
Students will initially investigate existing wireless properties (walls, furniture), Human underwater communication performance.
Main goal will be to develop and evaluate a deep nodes (satellites, ground stations, user
communication technologies (e.g., Wi-Fi, blockage, Multipath propagation, Students will initially, study existing
learning-based model for accurate Bit Error Rate terminals). Then, it is required to define the
Bluetooth, Zigbee, LTE, etc) and their suitability for Atmospheric attenuation. Then, they will research on underwater acoustic
(BER) prediction in satellite-terrestrial resource allocation framework, including
healthcare applications. Then they will mainly focus validate channel models with experimental communication challenges, channel
communication networks. Students are required to resources to be allocated (bandwidth, power,
on analysing security vulnerabilities and threats in data. Next, it is required to choose models, and STAR-RIS technology.
collect and pre-process relevant datasets for time), constraints, and optimization objectives.
healthcare wireless systems (e.g., data appropriate simulation tools (e.g., MATLAB). Then, they will design a prototype of
1.Design and Analysis of Secure Deep Learning for BER Prediction in training and testing. Then, they will be asked to 3.Design and Analyze Secure 6G AI-Based Resource Allocation Next, students will perform the data collection
interception, unauthorized access and attacks). Further, main goal will be to investigate 4.Performance Analysis of STAR- underwater communication system
21 Health Care Wireless System and Satellite-Terrestrial Communication design and implement deep learning architectures Communication Systems for Indoor THz Techniques for NTN and model development. Further goal will be to
Further, using physical layer security they will security threats specific to indoor THz RIS Underwater Communications incorporating a STAR-RIS. Considering
Investigation of Security Measures: Networks (e.g., RNN, LSTM, CNN) for BER prediction. Communication. Communications utilize existing simulation tools to model the
evaluate the performance of the proposed secure communication (eavesdropping, jamming, factors like, propagation effects
Further, train and evaluate, and analyze the impact NTN system and integrate the developed
system in terms of security, reliability, latency, and spoofing, data theft) and to explore existing (attenuation, scattering, multipath), water
of different network parameters and environmental AI/ML models and to implement realistic
energy efficiency. Finally, we are expecting to security methods (encryption, authentication, properties (salinity, temperature,
factors on BER. Compare the existing mathematical channel models, user traffic models, and
develop a prototype or simulation model to physical layer security), finally, design a turbidity), STAR-RIS characteristics
complexity with the proposed Deep-leaning-based interference models. Finally, student will
demonstrate the feasibility and effectiveness of the secure communication system architecture (number of elements, reflection
model. highlight the benefits of AI-based resource
proposed system. via deep diving into 6G technologies (AI/ML, coefficients), our goal will be to define
allocation in terms of throughput, latency,
RISs, massive MIMO) and their security and investigate the performance such as,
energy efficiency, and other KPIs.
implications. data rate, latency, reliability (BER/PER),
coverage, energy efficiency, and security.

This project focuses on classifying

Diabetic Retinopathy (DR) is a severe eye This project utilizes convolutional neural This project uses graph neural networks
hyperspectral images captured by
condition caused by diabetes, leading to vision This project focuses on analyzing human gait networks (CNNs) to classify and recognize (GNNs) to recognize dynamic hand or body
remote sensors to analyze land cover
impairment or even blindness if not detected early. patterns using machine learning and deep learning human activities (e.g., walking, running, gestures from temporal data. Gestures are
and urban areas. By leveraging deep
Based on retinal fundus images, the project aims to techniques to estimate psychological states, such sitting, or jumping) from sensor data or video represented as graphs, where nodes
learning techniques like convolutional
classify DR into various severity levels (e.g., No as stress, anxiety, or depression. By extracting inputs. The system processes time-series correspond to key joints and edges capture
neural networks (CNNs), it processes
DR, Mild, Moderate, Severe, or Proliferative DR). Human Gait-based Analysis for features like stride length, speed, and posture from data from wearable sensors or frames from Hyperspectral Image spatial and temporal relationships. GNNs
Diabetic Retinopathy Grading using Human Activity Recognition using high-dimensional spectral data to identify Dynamic Gesture Recognition using
22 This project focuses on using deep learning Psychological State Estimation in video or sensor data, the system identifies subtle video sequences, extracting spatial and Classification for Remote Sensing effectively model these interactions to classify
Deep Learning Techniques: Convolutional Neural Networks and classify distinct regions. The project Graph Neural Network
methods to grade DR from retinal images Behavioral Monitoring behavioral changes linked to mental health. Models temporal features for activity detection. Application gestures in real time. The project processes
involves data preprocessing and training
automatically. It involves preprocessing to enhance like LSTMs or CNNs process the temporal and CNNs are trained on datasets to ensure high datasets, extracting robust spatiotemporal
models to classify areas. This
the image’s contrast, training models, and spatial data, ensuring accurate state estimation. accuracy. This solution has applications in features. Applications include human-
classification aids in environmental
evaluating performance. The automated system The project aims to provide early psychological healthcare, fitness tracking, and smart computer interaction, sign language
monitoring, agriculture, and disaster
aids in accurate early detection and improving intervention and improve mental health care. surveillance systems, providing real-time and recognition, virtual reality, and efficient gesture-
management with enhanced accuracy
patient outcomes. robust activity recognition. based systems.
and efficiency.

Design algorithms for improving vehicular Design optimization algorithms for

Integrated Sensing and Design algorithms to improve accuracy of landslide Design Lightweight federated learning Design optimization algorithms for delivering
Integrated Sensing and Communication communication by integrating the sensing Federated Learning for Intelligent 4.Resource Allocation for delivering improved QoS to vehicular 5.Resource Allocation for
23 Communication for Landslide detection using data captured with mmWave models for intelligent vehicular string stability to vehicular platoons over the
for Vehicular Communication capabilities of the vehicles with their communication Vehicles Vehicular Communication over 5G users over the high frequency 5G Vehicular Platoons over 5G
Detection: sensors. communication over 5G high frequency 5G channels
transceivers channels

Design and Development of FSO/LASER based High-Speed FSO System for DATA Design and development of High-Speed Object detection and data
FSO based Autoalligned Energy Design and development of IP Camera based Object detection and data transmission AI Based Crop selection and Development of AI Based Crop selection and
24 Autoalligned Energy Harvesting for UAV Criminal Activity recognition system Transmission using quantum FSO System for DATA Transmission using transmission using UAV to
Harvesting for UAV applications : Criminal Activity recognition system using UAV to ground station recommendation recommendation
applications communication quantum communication ground station

Automated visual surveillance in dynamic

scenes, especially for humans, is currently
one of the most active research topics in
computer vision. It has a wide spectrum of
In this project, various deep-learning methods will
In this project, various deep-learning methods will promising applications, including the
be applied to classify diseases affecting the lungs
be used to identify the MR images of tumor and non- detection of anomalous behaviors, interactive
using chest X-ray and CT scan images. A
tumor classes from various slices of brain MR surveillance using multiple cameras, crowd
supervised learning approach will be employed,
images. Transfer learning has been used to classify flux statistics, and congestion analysis, etc.
leveraging annotated datasets to train deep-
tumor and non-tumor brain MR images. The method The objective of this project or research is to
learning models for accurate disease detection
of transfer learning has been widely used to develop algorithms to detect anomalous
using different techniques. The main aim is to
overcome the problem. The lack of data is a major events such as fighting, riots, road accidents,
improve classification performance while Deep Learning-based Anomalous Events
Deep-learning-based Disease Detection Identification of Brain Tumor using disadvantage in the field of medical imaging. There etc., using Deep Learning-based weakly
25 addressing challenges such as overfitting and class Detection in Untrimmed Surveillance NIL NIL NIL NIL
using Medical Images.: Deep-learning Techniques are many frameworks in the area of classification, supervised video anomaly detection models.
imbalance. Although many frameworks exist for Videos
but still, there is a need to design better and more As anomalies are context-dependent and
disease detection, there is still a need for models
accurate frameworks for the identification of brain can have infinite possible configurations, the
with higher accuracy and lower computational
tumors and classify them based on their features. model-based approach can be used to
complexity. The primary objective of this project is
So, the main aim of the project is to design and classify any events. The proposed models
to design and propose a supervised learning-based
propose a highly accurate and comparatively less will be based on motion and appearance
deep-learning framework that efficiently and
computationally complex model for fulfilling the patterns and focus primarily on sparsely and
accurately detects and classifies fatal diseases
task. modestly crowded scenes. The developed
using biomedical images.
models will be optimized for achieving
desired online performances such as high
fps, low computational resources, and good
accuracy.

This project proposes the development of a

A wireless ECG monitoring system designed
portable, AI-integrated pupillometry device that This project presents a novel, AI-powered portable
for real-time cardiac care. It will enable
leverages machine learning algorithms and solar cooker that utilizes a Fresnel lens to focus
seamless transmission of ECG signals from
computer vision to accurately measure pupil size sunlight and cook food efficiently. The cooker
a wearable sensor to a mobile device. The
and reactivity. By harnessing the power of artificial features a compact, lightweight design and an AI-
system should provide a user-friendly
intelligence, this aims to revolutionize the field of powered control system that optimizes temperature
AI-Driven Portable Pupillometry for interface for real-time ECG visualization,
26 pupillometry, enabling healthcare professionals to AI-Powered Portable Solar Cooker control, stirring, and heat distribution. This Wireless ECG Monitoring System NA NA NA NA
Next-Generation Healthcare: allowing for prompt detection of cardiac
diagnose and monitor neurological and innovative solution aims to provide a reliable,
anomalies and timely interventions. With its
physiological conditions more efficiently and sustainable, and safe cooking method for rural
compact design, low power consumption,
effectively. With its compact design, user-friendly communities, outdoor enthusiasts, and emergency
and cost-effectiveness, it has the potential to
interface, and AI-driven analytics, this has the relief situations, reducing reliance on fossil fuels
revolutionize cardiac monitoring and improve
potential to transform the future of healthcare and mitigating environmental impacts.
patient outcomes.
diagnostics

Microwave absorbers (2 to 18GHz) will be designed

Radar Cross Section (RCS) of an Antenna will be Antenna will be developed on conical Intelligent Surface will be developed to hide an
on cylindrical and conical surfaces for Defence Design and Development of Textile (or Antenna will be developed on hat/shoe/cloths Development of an Antenna on a Development of an Intelligent
Reduction of Radar Cross Section reduced for stealth applications. It will help our Conformal Microwave Absorber for surfaces for Indian Air Force (IAF) object from the RADAR for Defence
27 Applications. The radar system will not be able to Wearable) Antennas for Defence for army applications. Quality research Conical Surface for Defence Surface to hide an object for
(RCS) an Antenna using Metasurface: defence system undetectable. Quality research Defence Applications applications. Quality research articles will Applications. Quality research articles will be
detect airplanes, missiles, ships, etc. Quality Applications articles will be published in IEEE Applications Defence Applications
articles will be published in IEEE be published in IEEE. published in IEEE.
research articles will be published in IEEE
 Neuromorphic computing is an emerging
technology that supports event-driven
data processing for applications
requiring efficient online inference and/or
control. Recent work
has introduced the concept of
neuromorphic communications,
whereby neuromorphic computing is
integrated with impulse
radio (IR) transmission to implement low-
energy and low-latency
Radiofrequency fingerprint identification (RFFI) is a
remote inference in wireless Internet-of-
promising device authentication technique based
Deep Learning based Radio Frequency Neuromorphic (Spiking Neural Things (IoT) networks. Neuromorphic Deep learning based OTFS/FBMC
on the transmitter hardware impairments. The 5G and AI enabled IoT for societal PLL based Telemetry and telecommand PLL based Telemetry and telecommand OTFS/FBMC based Transceiver
28 Fingerprint Identification for detecting Application of AI and IoT Networks) Integrated Sensing and integrated sensing and communications based Transceiver implementation
device-specific hardware features can be extracted applications. receiver implementation in FPGA. receiver implementation in FPGA. implementation in FPGA
IOT /RF transmitters: Communications (N-ISAC), is a novel solution that enables in FPGA
at the receiver by analyzing the received signal and
efficient
used for authentication.
online data decoding and radar sensing.
N-ISAC leverages a
common IR waveform for the dual
purpose of conveying digital
information and detecting the presence
or absence of a radar
target. A spiking neural network (SNN) is
deployed at the receiver
to decode digital data and to detect the
radar target using
directly the received signal.

A RISC-V processor has to be designed using

Design and development of IoT based an air quality monitoring system is designed using ASCON is a light weight encryption protocol Design of a Trusted platform
Design of a 5-satge RISC_V processor Verilog HDL. It has to validated through functional Design and implementation of ASCON TPM consists of hash algorithm, Design of AXI, APB and peripheral Design of AXI, APB and peripheral IPs for
29 Air quality monitoring system using ESP 32 and IOT communication protocol and later used in IoT.. THis design has to be module for internet of thing
for IoT appliction. simulation and has to be implemented in FPGA for light security protocol in FPGA encryption algorithm and RNG.. IPs for RISCV processor RISCV processor
RTOS: on a complete system has to be developed implemented in FPGA and further in ASIC. application
an IoT application.
The core of semantic communication is to
The project aims to address challenges in IRS-aided wireless communications, its
extract the “meanings” of sent information at a
Channel estimation in intelligent Investigation of channel estimation in uplink and Integrated Sensing and Communication (ISAC) reflection and channel models, hardware
Integrated Sensing and Communication Intelligent Reflecting Surface-Aided Integrated Sensing and Detection and tracking of UAVs is an transmitter, and with the help of a matched
30 reflecting surfaces assisted millimeter downlink of IRS-assisted mmWave multiple-input technology for future wireless systems, such as architecture and practical constraints, as well Semantic Communication
for future wireless systems Wireless Communications Communication for UAVs important application of ISAC. knowledge base (KB) between a transmitter
wave communication: multiple-output (MIMO) systems balancing communication and sensing as various appealing applications in wireless
and a receiver, the semantic information can
requirements and mitigating interference. networks.
be “interpreted” successfully at a receiver.

Detect the leaks in pipelines using IoT and Detect the leaks in pipelines using IOT and Deep
31 IoT based Leak detector 1: IoT based Leak detector 2 Image Classifier in FPGA Classify images using FPGA Video content Analysis 1 Analyse the video in ARM Video content Analysis 2 Video analysis in ARM 2
Machine Intelligence 1 learning

Underwater navigation is a critical technology for

applications such as marine exploration, underwater
robotics, submarine operations, and environmental
monitoring. Traditional navigation methods, such as Long Range (LoRa) technology has emerged
GPS, are ineffective underwater due to signal as a promising solution for underground
attenuation. Therefore, alternative techniques such communication, offering low-power, long-
as Inertial Navigation Systems (INS), acoustic- range, and cost-effective wireless
Reconfigurable intelligent surfaces (RISs) are an
based methods (Ultra-Short Baseline and Long connectivity in challenging environments.
emerging technology aimed to enhance the energy
Baseline systems), and magnetic and optical Traditional communication methods, such as
and spectrum efficiency of wireless communication
sensors are employed for precise underwater Wi-Fi and cellular networks, face significant
links by reconfiguring the propagation environment
positioning. signal attenuation in underground settings,
of radio-frequency (RF)/optical waves. Planar
making LoRa a viable alternative due to its
phased meta-material surfaces or reflect-arrays for
This project explores advancements in underwater sub-GHz frequency operation and high
RF RIS and multiple smart mirrors have been used
navigation, including the integration of AI, sensor penetration capabilities. This project
for activating optical RIS. In this project, we analyze
fusion, and machine learning for enhanced Underground Communication Using explores the feasibility of using LoRa for
32 Underwater wireless communication : the underwater optical wireless communication Underwater Navigation System NA NA NA
accuracy and autonomy. The development of LoRa underground applications, including mining
(UWOC) system performance in terms of bit error
autonomous underwater vehicles (AUVs) and operations, underground sensor networks,
rate (BER), outage probability (OP) and channel
remotely operated vehicles (ROVs) relies on robust disaster response, and pipeline monitoring.
capacity (CC) under the influence of underwater
navigation systems that combine multiple sensors The study discusses signal propagation
turbulence, beam attenuation, pointing errors and
for real-time data processing. Challenges such as characteristics, challenges such as soil
occlusion or blockage due to obstacles. A RIS-
signal interference, environmental disturbances, composition effects and multipath fading,
assisted UWOC proposed to mitigate the effects
and power consumption are discussed, along with and potential solutions using adaptive data
obstacles, pointing errors and turbulence and
potential solutions using advanced algorithms and rates and relay nodes. Experimental results
improve the system performances.
energy-efficient hardware. indicate that LoRa can provide reliable
communication over extended distances
The project highlights the importance of underwater underground, making it suitable for real-time
navigation in fields like oceanography, defense, monitoring and IoT-based applications.
deep-sea mining, and marine archaeology. Future
research directions include the improvement of
hybrid navigation systems underwater

Brain Tumor Detection and

33 Object Recognition: Computer Vision Handwritten Character Recognition Computer Vision 3-D Face Recognition Computer Vision Facial Expression Recognition Computer Vision Computer Vision
Classification

Project involves cyber security & hardware security Development of CAN Project involves simulation & IC
Project involves cyber security & hardware security Development of framework for Industrial
34 Security for Vehicular IoT: Security for IoMT for Internet of Medical Things, a typical fitbit to be Smart Energy Meter Protection. Protecting smart metres against intrusion. Industrial IoT framework design communication protocol with development/FPGA Implementation.This
for a vehicular IoT framework. IoT.
used. additional security. would facilitate vehicle to be intrusion proof

Investigation of the performance of the two source

Functional connectivity analysis of
Identification of EEG-based functional Functional connectivity analysis in EEG functional connectivity (FC) methods, the inverse-
35 human brain as signatures for NA NA NA NA NA NA
connectomes specific to pathologies source space based source FC (ISFC) and the cortical partial
pathologies:
coherence (CPC) to identify an appropriate method.

Patients suffering from quadriplegia or other

neurological impairments are unable to move their
hands or shoulders to control the wheelchair
manually or through the joystick. These people
require a smart electric-powered wheelchair Robotic manipulators find significant applications in
controlled by their head-neck motions instead of healthcare, industry, logistics, defence, etc. This
hand-and-shoulder movements. Therefore, this work proposes a voice-controlled assistive robotic
Head-neck gesture classification for Voice-controlled assistive robotic
36 work proposes the detection and recognition of manipulator for assisting individuals with disabilities. NA NA NA NA NA NA
assistive wheelchair control: manipulator.
head-neck gestures for generating control The use of speech recognition algorithms to
commands for assistive wheelchairs. The gestures precisely control the robotic manipulator through
can be detected with the help of inertial voice commands.
measurement unit (IMU) sensors, flex sensors,
attitude sensors, etc. which can be further
classified to obtain discrete control commands for
wheelchairs.