| S.No |
Name of the project |
Abstract |
| 1 |
DESIGN AND DEVELOPMENT OF MACHINE LEARNING ENABLED SENSING FRAMEWORK FOR URBAN LIGHTING SYSTEM MANAGEMENT |
The exponential growth of urbanization coupled with the increasing demand for energy-efficient infrastructure necessitates innovative solutions for managing street lighting systems effectively. In response to this need, this research presented the design and implementation of a state-of-the-art smart street light monitoring system with integrated fault detection capabilities. The proposed system leverages advanced sensor technologies, including current, voltage, Light Dependent Resistor (LDR), and Global Positioning System (GPS), to collect real-time data on street light performance and environmental conditions. This data is processed using microcontroller technology and relay modules to enable precise control and monitoring of individual street lights. Key features of the system include its ability to detect faults in street lights through comprehensive analysis of current, voltage, and ambient light levels. Utilizing GPS functionality, the system facilitates accurate location tracking of faulty street lights, thereby expediting maintenance and reducing downtime. Furthermore, the system offers both local and remote monitoring capabilities. Local status indication is provided via an I2C display, allowing for on-site assessment of street light performance. Meanwhile, remote monitoring via an Internet of Things (IoT) platform enhances operational efficiency by enabling centralized control and real-time data analytics. |
| 2 |
AI BASED STRENGTH PREDICTION OF GEOPOLYMER CONCRETE USING WASTE TYRE RUBBER |
Geopolymer concrete is an environmentally friendly and low-cost alternative to conventional Portland cement concrete. Incorporating waste tyre rubber into geopolymer concrete improves its strength and durability. The model predicts the strength of geopolymer concrete reinforced with waste tyre rubber using historical data and experimental data. It factors in the specific gravity of the geopolymer mix, the concentration of waste tyre rubber, and the curing temperature. The AI-based strength prediction model has applications in the construction, engineering, and environmental sectors. The methodology involves casting of M30 grade geopolymer concrete of 3 beams of size 700×150×150 mm for casting the beams which contains fly ash, alkaline activator, coarse aggregates, M sand and waste rubber tyres as our raw materials. The width of the crack is 0.02 mm. By acquisition of high-resolution images of beams during testing of beams, which are pre-processed to enhance relevant features. Graphs have been extracted based on AI model. The compressive strength is obtained to the compressive strength by applying loads, flexural strength and abrasion resistance of the beams are obtained by testing. The purpose of this paper is to assess the current state of the research being done on geopolymer concrete for the use of waste tyres. The maximum and minimum strength of the beam is 19.52 N/mm2 and 457.14 N/mm2. The purpose of this study is to determine the hardened properties like flexural and compressive strength using AI model which is compared with experimental and predicted values. It also aims to account for additional factors that may influence the strength of geopolymer concrete reinforced with waste tyre rubber. |
| 3 |
EXPERIMENTAL INVESTIGATION OF PADDY STRAW ASH BASED GEOPOLYMER CONCRETE |
Environmental issues resulted from cement production have become a major concern today. To develop a sustainable future it is encouraged to limit the use of this construction material that can affect the environment. Cement replacement material was proposed to partially replace cement portion in concrete. Geopolymer is the best solution to reduce the use of cement in concrete. Geopolymer is a hardened cementitious paste made from fly ash, alkaline solution and geological source material. The development of fly ash and paddy straw ash (PSA) as the source material for geo polymer concrete was studied through the observation of the hardened specimen strength and durability properties. Paddy straw ash is a byproduct from the burning of paddy straw at a temperature lower than 6000C.this means that it is in a form that is soft and easy to grind. Rice husk ash is rich in silica about 90%, 5% carbon and 2% K2O. The specific surface of PSA is between 40- 100m2 /g. It is extremely prevalent in East and South-East Asia because of the Paddy production in this area. The addition of PSA as a silica source also had an effect on the strength of Geopolymer. The strength and durability increased with an increase in silica content. |
| 4 |
EXPERIMENTAL INVESTIGATION ON GEOPOLYMER CONCRETE SLAB USING WASTE TYRE RUBBER |
In our project “Experimental investigation of geopolymer concrete using waste rubber tyre”. We have casted two geopolymer concrete slab using waste rubber tyre. The materials used in this casting process for the slab are “concrete, sand, aggregate, ground granulated blast furnace (GGBS), flyash. The mix design for the slab 1 is M20 and the mix design for the slab 2 is M25. The mix ratio for M20 grade concrete is (1: 1.75: 3) and the mix ratio for M25 grade concrete (1: 1.3: 2.3). The dimension of the geopolymer concrete slab is (500x500x100mm). Later the materials are mixed and it is poured into the slab mould, before this process. The reinforcement has been added inside the geopolymer concrete slab. 12mm rod has been used in the M20 and M25 grade concrete slab. The spacing for the reinforcement rod is about 10cm. After this process, the cover blocks are kept inside the slab to provide additional strength. Then, the slab has been completely casted and then the geopolymer concrete slab has been cured for a period of 28 days. and then, the two slabs have kept in “Universal testing machine” and compression test have been conducted in these two slabs to find the compressive strength of the geopolymer concrete slab. |
| 5 |
FLEXURAL BEHAVIOUR OF THE REINFORCEMENT CONCRETE INCORPORATE WITH CABLE WIRE AND CULLET GLASS |
This project explores the synergistic effects of incorporating cable wires and cullet glass into reinforced concrete structures to improve flexural behavior. The combination of high-tensile strength cable wires and the unique properties of cullet glass aims to enhance the structural integrity, durability, and sustainability of traditional reinforced concrete. In this project we are casting 4 beams of size 70 x 15 x 15 cm and testing it for finding its flexural behaviour. The research involves a comprehensive investigation into the mechanical properties of cable-reinforced concrete, assessing its ability to withstand bending loads. Additionally, the incorporation of cullet glass particles into the concrete mix aims to contribute to both the mechanical and aesthetic aspects of the structures. The methodology includes material characterization, laboratory testing, and structural analysis. Mechanical tests will assess the flexural strength, ductility, and crack resistance of the novel composite materials. Structural analysis using advanced modelling techniques will provide insights into the overall performance and behavior of the proposed system. The expected outcomes of this project include a deeper understanding of the synergies between cable wires and cullet glass in reinforced concrete, as well as the development of guidelines for practical applications in construction projects. |
| 6 |
EXPERIMENTAL INVESTIGATION OF RICE HUSK ASH BASED GEOPLOYMER CONCRETE |
This study investigates the utilization of rice husk ash (RHA) as a promising pozzolanic alternative to partially replace class-C fly ash in geopolymer concrete, with the dual objectives of enhancing concrete strength and mitigating the corrosive impact of harsh environmental conditions. The experimental approach involves incorporating RHA at varying replacement ratios (30% and 40%) in combination with different geopolymer molarities (10M and 12M), followed by dry curing in a solar dryer for different durations (7, 14, and 28 days). The primary focus is on assessing the impact of these variables on the strength properties of the concrete. Comprehensive testing, including compressive strength, flexural strength, and tensile strength tests, along with SEM analysis, has been carried out to evaluate the suitability of partially substituting class-C fly ash with RHA in geopolymer concrete. The findings of this study, which represent a novel contribution to the field, highlight that the addition of fine RHA particles in conjunction with geopolymer leads to the attainment of the requisite strength for flexural, compressive, and tensile properties. Furthermore, SEM analysis has been instrumental in elucidating the bonding strength of the concrete, providing valuable insights into the performance of RHA-incorporated geopolymer concrete. By shedding light on the efficacy of RHA as a pozzolanic alternative and its potential to enhance the performance of geopolymer concrete, this research contributes to the evolving body of knowledge in sustainable construction materials and paves the way for further advancements in eco-friendly concrete production. |
| 7 |
LANELINK – CARPOOLING APPLICATION |
LaneLink is a cutting-edge web application that aims to transform dailycommutes by promotingeffective carpooling among users. LaneLink uses a user-friendly interface and powerful matching algorithms to link commuters with appropriate travel routes and timetables, promoting shared trips for a greener,more cost-effective, and less crowded transportation system. LaneLink’s key features include customizable profiles, which allow users to select preferences such as preferred departure times, route flexibility, and passenger criteria. The application’s sophisticated matching mechanism uses these preferences and real-time traffic data to recommend the best carpooling arrangements, assuring convenience and dependability for all users. Furthermore, LaneLink promotes safety and security by adopting strong verification methods and offering user ratings and feedback. The platform promotes a sense of community and trust among members, encouraging more people to embrace carpooling as a sustainable transportation solution. Through these capabilities, the app enables users to make educated decisions that help to reduce carbon emissions and traffic congestionin their areas. Overall, LaneLink represents a forward-thinking response to currenttransportation concerns, leveraging technology to encourage shared mobility and improve urban quality of life. LaneLink, with its emphasis on convenience,safety, and sustainability, is on track to become the go-to platform for commuters looking for smarter, greener ways to travel |
| 8 |
Real estate price prediction |
Real estate transactions are pivotal financial decisions, and accurate price prediction a crucial role in facilitating informed choices and preventing potential financial losses. In thisstudy, we propose a machine learning-based approach to forecast real estate prices, with the primary goal of enhancing predictive accuracy and aiding stakeholders in making well-informed decision By comparing the efficacy of various machine learning algorithms, including Linear Regression, Random Forest. The training and validation process, along with model interpretation techniques, are discussed to ensure robust performance. We aim toid entity the most reliable method for predicting real estate prices. Through comprehensive data analysis and model evaluation, our objective is to provide stakeholders with valuable insights and tools to navigate the dynamic landscape of the real estate market effectively. |
| 9 |
Mr. SCRAPPER |
In the age of ubiquitous internet access, the proliferation of spam links poses significant risks to users’ privacy, security, and overall online experience. To address this challenge, we present “Mr. Scrapper,” a novel mobile application designed to identify and classify links as either spam or legitimate with high accuracy. Leveraging a combination of frontend Java interface, Python backendprocessing, and MySQL database management, Mr. Scrapper integrates seamlessly into users’ daily browsing activities. In the age of ubiquitous internet access, the proliferation of spam links poses significant risks to users’ privacy, security, and overall online experience. To address this challenge, we present “Mr. Scrapper,” a novel mobile application designed to identify and classify links as either spam or legitimate with high accuracy. Leveraging a combination of frontend Java interface, Python backend processing, and MySQL database management, Mr. Scrapper integrates seamlessly into users’ daily browsing activities. The significance of Mr. Scrapper extends beyond individual user experiences. By proactively identifying and flagging spam links, it contributes to the collective effort of combating online threats and fostering a safer digital environment for all. Moreover, its modular architecture and reliance on cutting-edge technologiesdemonstrate the potential for innovation at the intersection of deep learning, mobile development, and cyber security. |
| 10 |
PHOTOGRAPHY COMMUNITY |
Lensrivals is an innovative web application designed to provide photographers worldwide with a dedicated platform for participating in photography contests and showcasing their talent. Leveraging modern web technologies and a user friendly interface reminiscent of popular social media platforms, Lensrivals aims to revolutionize the way photographers engage with each other and compete for recognition and cash prizes. Key features of Lensrivals include contest participation, where photographers can upload their best shots according to contest instructions, add captions, locations, and hashtags, and engage with other participants’ submissions. The platform offers a feed page where users can explore recent photo submissions, a personalized profile for managing posted clicks and tracking contest participation, and a comprehensive list of live and upcoming contests to keep users engaged. Lensrivals utilizes Firebase for backend and database management, ensuring real-time data storage and synchronization, while Clerk Authentication adds an extra layer of security to user accounts. The use of NEXT.js for frontend development and Tailwind CSS for styling ensures a sleek and responsive design, enhancing the overall user experience. Shadcn UI Components further augment the platform with visually appealing design elements. With its unique combination of social media features and contest functionalities, Lensrivals aims to foster a vibrant photography community where photographers of all skill levels can thrive. Join Lensrivals today and discover a new way to showcase your creativity, compete for prizes, and connect with like-minded individuals passionate about photography. With its focus on innovation and community building, Lensrivals provides a unique space for photographers to showcase their work, connect with peers, and elevate their skills. Whether you’re an amateur enthusiast or a seasoned professional, Lensrivals offers a platform where your creativity can shine. Join us today and become part of a thriving community of passionate photographers. |
| 11 |
Adaptive CRISIS RESPONSE NETWORK |
In emergency situations, rapid response and efficient communication are critical for saving lives and mitigating damages. This paper presents the design and implementation of a Crisis Response System (CRS) aimed at minimizing emergency response time and enabling victims to promptly access nearby hospitals, emergency services, and notify trusted contacts. The system utilizes modern technologies including geolocation services, mobile applications, and real-time communication channels to streamline the response process. Through rigorous testing and evaluation, the CRS demonstrates its effectiveness in reducing response time and enhancing the overall emergency management process. |
| 12 |
ON ROAD VEHICLE BREAKDOWN ASSISTANCE FINDER SYSTEM |
In the event that an individual’s vehicle breaks down, the On-Street Vehicle- Breakdown Application (ORVBA) is a suitable solution for them to seek help in remote areas. Clients of ORVBAFA will be the selected public, and they will establish contact with the specific professional via a dependable application procedure. The framework known as On Street Vehicle Breakdown Application (ORVBA) exclusively targets mechanics who are legally supported and maintained. It is extremely irrelevant that some consumers in a continuous framework have their own master educational file. Additionally, they have no idea whether their cars will break down or experience mechanical issues in isolated locations or in any other far locations from their reliable repair firms. Users can search for a list of professionals in any place (or) surrounding areas who can help them in confusing situations caused by their vehicles’ mechanical difficulties under the proposed-On Street Vehicle Break-down Assistance Finder System (ORVBAFS) development. |
| 13 |
LoRa-Wan: A RESILIANT FRAMEWORK FOR EMERGENCY COMMUNICATION SYSTEM |
In times of crises like natural disasters, ef ective communication is crucial for coordinating responses and saving lives. Traditional infrastructure often fails during such events, leaving communities vulnerable. LoRaWAN technology of ers a resilient solution, enabling long-range communication between remote devices and gateways. Its low-power protocol makes it suitable for emergency scenarios, ensuring reliability and scalability. LoRaWAN’s capabilities facilitate communication even in harsh environments, strengthening collaboration with stakeholders. Additionally, its unlicensed radio frequency bands reduce traf ic, while bidirectional communication supports real-time data transmission and remote device management. |
| 14 |
Lane Detection and Alerts for Autonomous Driving |
During the driving operation, humans use their optical vision for vehicle maneuvering. The Road lane marking acts as a constant reference for vehicle navigation. One of the prerequisites to have in a self-driving car is the development of an Automatic Lane Detection system using an algorithm.Computer vision is a technology that can enable cars to make sense of their surroundings. It is a branch of artificial intelligence that enables software to understand the content of images and video. Modern computer vision has come a long way due to the advances in deep learning, which enables it to recognize different objects in images by examining and comparing millions of examples and cleaning the visual patterns that define each object. While especially efficient for classification tasks, deep learning suffers from serious limitations and can fail in unpredictable ways.This means that a driverless car might crash into a truck in broad daylight, or worse, accidentally hit a pedestrian. The current computer vision technology used in autonomous vehicles is also vulnerable to adversarial attacks, by manipulating the AI’s input channels to force it to make mistakes. For instance, researchers have shown they can trick a self-driving car to avoid recognizing
stop signs by sticking black and white labels on them. |
| 15 |
Traffic automation System using Reinforcement learning |
Traffic Automation System (TAS) leveraging machine learning techniques to optimize traffic flow in urban areas. The system aims to alleviate congestion, reduce travel time, and enhance overall traffic efficiency. Central to the system’s functionality is the utilization of advanced machine learning algorithms for real-time data analysis and decision-making. The TAS operates on a comprehensive model that incorporates various theories and methodologies. Primarily, the system relies on reinforcement learning (RL) algorithms to develop adaptive traffic control strategies. RL enables the TAS to continuously learn and adapt its control policies based on feedback received from the environment, such as traffic volume, congestion levels, and historical traffic patterns. Furthermore, the project integrates predictive modelling techniques, including recurrent neural networks (RNNs) and Long Short-Term Memory (LSTM) networks, to forecast future traffic conditions. By analyzing historical data and current traffic trends, the system can anticipate congestion hotspots and proactively adjust traffic signals to mitigate potential bottlenecks. TAS incorporates graph theory principles to optimize traffic flow across interconnected road networks. Graph-based algorithms are employed to determine the most efficient routes for vehicles, considering factors such as distance, road capacity, and traffic density. Traffic Automation System presented in this paper represents a holistic approach to traffic management, leveraging machine learning, predictive modelling, and graph theory to optimize urban mobility and enhance the overall transportation experience to the mass. |
| 16 |
SAFEROUTE AI SENTRY – FOR POTHOLE AWARENESA IN AUTONOMOUS DRIVING VIA YOLOV8 |
The advent of self-driving vehicles has marked a transformative milestone in contemporary transportation, offering unprecedented levels of safety, efficiency, and convenience. However, the persistent challenge of navigating unpredictable road conditions, particularly in the presence of potholes, poses potential safety risks to autonomous driving. This research introduces an innovative cloud-powered next-generation self-driving safety system that leverages the capabilities of artificial intelligence, specifically the You Only Look Once version 8 (YOLOv8) model, in tandem with the wand (Weights & Biases) deep learning platform. This integration facilitates pothole detection and advanced navigation, significantly enhancing the safety standards of autonomous driving. The selection of YOLOv8 is based on its exceptional accuracy and speed in object detection. YOLOv8 utilizes a singular neural network to predict object bounding boxes and class probabilities directly, enabling rapid and precise object detection. The cloud-based architecture of this system supports continuous model updates and refinements, ensuring adaptability to evolving road conditions and pothole variations. Beyond its application in pothole detection, this system holds the potential to redefine the landscape of autonomous transportation, ushering in a new era of safety and reliability in self-driving technology. The proposed approach not only addresses the challenges posed by potholes but also establishes a foundation for safer and more efficient autonomous travel, showcasing the transformative impact of cutting-edge technology in the field of self-driving vehicles. |
| 17 |
THERMAL VISION AND OBSTACLE DETECTION FOR LOW VISIBILITY ENVIRONMENTS |
Navigating through low visibility environments, vital for exploration and rescue operations, faces limitations with traditional visual-based systems. This project develops a robust navigation system for such environments using thermal vision and LiDAR within ROS. Integrating a thermal camera and YD LiDAR sensor, the system aims to enhance situational awareness and obstacle detection. The methodology includes sensor integration, software development, and testing. Expected outcomes include improved navigation accuracy and adaptability, contributing to safer and more efficient missions in challenging conditions. Through this project, advancements in navigation technologies for low-visibility environments are pursued, with implications for various applications requiring exploration and safety in adverse conditions. |
| 18 |
EXPERIMENTAL INVESTIGATION ON CEMENT MORTAR WITH PARTIAL REPLACEMENT OF CEMENT AND FINE AGGREGATE AS GLASS POWDER AND E-WASTE |
This experimental investigation explores the effects of partially replacing cement with glass powder (10%, 20%, 30%) and sand with E-waste powder (constant 10%) in mortar. The study aims to assess the potential of these alternative materials in enhancing mortar properties while reducing the environmental impact of traditional concrete production. The experimental procedure involves preparing concrete mixtures with varying replacement levels of cement with glass powder and sand with E-waste powder. Test specimens are then prepared and subjected to various tests to evaluate the fresh and hardened properties of the mortar, including compressive strength, split tensile strength, flexural strength, workability, water absorption, and durability. By comparing the test results of the mixtures with different replacement levels, the study seeks to assess the impact of the partial replacement of cement with glass powder and sand with E-waste powder on the properties of mortar. 30% of glass powder and 10% of E waste powder test sample gives the maximum compressive strength of 24.33 N/mm sq, durability, and the environmental benefits of using waste materials in mortar production. This investigation contributes to the understanding of sustainable and eco-friendly concrete mix designs and may offer valuable implications for the construction industry in adopting more environmentally conscious practices. |
| 19 |
YOLO LANDSLIDE SENTRY: A FRAMEWORK FOR RAPID DETECTION AND RESPONSE |
This project introduces an innovative approach to landslide detection in hairpin bend regions using the You Only Look Once (YOLO) object detection framework. Hairpin bends, with their unique topography, pose specific challenges for landslide detection, necessitating a tailored solution that integrates advanced computer vision techniques. The proposed methodology combines high-resolution satellite imagery data to create detailed terrain models of hairpin bend areas. YOLO, known for its real-time object detection capabilities, is adapted to identify potential landslide triggers, including slope instability and changes in vegetation cover, within these complex landscapes. Real-time monitoring systems, including ground- based sensors and weather stations, are strategically placed to continuously capture environmental conditions. Integration with the YOLO-based detective model enables the early identification of potential landslide threats, facilitating the implementation of targeted early warning systems. Community engagement remains a crucial aspect of this approach, involving local residents in the development of evacuation plans and preparedness strategies. The synergy between YOLO-based technology and community involvement creates a comprehensive solution for proactively managing landslide risks in hairpin bends |
| 20 |
CONVOLUTIONAL NEURAL NETWORK FOR HUMAN DETECTION |
This work investigates the application of YOLOv8, a cutting-edge deep learning model, for real-time human detection. YOLOv8’s efficient single- stage architecture enables rapid object identification while maintaining accuracy. This report explores the implementation of YOLOv8 for human detection tasks. Pre-trained models are utilized, leveraging their inherent ability to recognize humans within a broader object classification. Frameworks like PyTorch and libraries like Ultralytics streamline the implementation process. Creating some manual annotations using the Roboflow application and creating and training over 1000 images and testing and training to attain more accuracy than previously used methods for detection. This work delves into the pre-processing steps for input images, ensuring compatibility with the chosen YOLOv8 variant. The model’s output, bounding boxes with confidence scores for detected humans, is analyzed. Furthermore, the report discusses the applicability of YOLOv8 in real-time scenarios like video surveillance. Potential applications in pedestrian counting and activity monitoring are highlighted. Integration with multi-object tracking algorithms like Deep SORT is explored to enhance functionality. This report emphasized that while pre-trained models offer a convenient starting point, fine-tuning with human-centric datasets can refine detection accuracy. Finally, this project concluded by underlining YOLOv8’s potential as a powerful andadaptable solution for real-time human detection across diverse scenarios. |
| 21 |
INTELLIGENT SURVEILLANCE SYSTEM FOR VEHICLE EMISSION |
This proposed work of manual inspection processes for verifying vehicle emissions, an innovative surveillance system leveraging image processing and computer vision has been developed. This system focuses on real-time detection of smoke emissions from vehicles, addressing inefficiencies and inaccuracies inherent in manual methods. By utilizing deep learning, an advanced object detection model, the system autonomously analyzes surveillance footage captured by strategically positioned cameras. This enables prompt identification of smoke emissions, triggering immediate alerts to relevant authorities via an integrated alert mechanism. Compared to manual inspection methods, the proposed system offers several advantages. It automates the verification process, eliminating the need for manual intervention and reducing reliance on human personnel, thereby improving efficiency. The system’s real-time monitoring capabilities enable proactive enforcement of regulatory standards, ensuring timely interventions and compliance across diverse traffic conditions. The proposed surveillance system represents a significant advancement over manual methods for detecting vehicle emissions |
| 22 |
AN AUTONOMOUS VEHICLE BASED ON V2V EFFULGENCE WITH DEEP LEARNING STANDARD FOR COMMUNICATION |
This proposed work presented the design and implementation of an advanced autonomous vehicle system integrating Vehicle-to-Vehicle (V2V) communication with Visible Light Communication (VLC) technology, alongside deep learning algorithms for image identification. The system comprised intricate hardware components including an ESP32 microcontroller serving as the central processing unit, ultrasonic sensors for proximity detection, motor drivers for DC motors enabling speed control and obstacle avoidance, and toggle switches for signalling turns. The deep learning software component encompassed convolutional neural networks (CNNs) for real-time image recognition, enabling the vehicle to interpret and respond to complex traffic scenarios accurately. The VLC communication system facilitated high-speed and secure data exchange between vehicles, enhancing coordination and safety measures on the road. Through the synergy of hardware mechanisms and sophisticated software algorithms, this work aims to establish a robust autonomous vehicle platform capable of navigating diverse traffic environments with optimal safety and efficiency |
| 23 |
DEEP LEARNING-EMPOWERED CHANNEL ESTIMATION AND CSI FEEDBACK FOR ENHANCED RELIABILITY IN 6G NETWORKS |
This system delved into the realm of Deep Learning (DL) for channel estimation, focusing on crucial aspects such as DL model selection, training set acquisition, and the design of the RESNET50 architecture. With the increasing integration of automated services, machines, vehicles, and sensors, DL is poised to become a predominant paradigm in the 6G era channel estimation. This system advocated for advanced DL techniques to address diverse challenges, including various frequency bands, wireless resources, and geographical environments. It highlighted transfer learning for training DL models and explored federated learning for collaborative task accomplishment. This comprehensive system aimed to guide MIMO communication researchers in integrating DL into their wireless channel estimation applications, fostering robustness and adaptability in diverse environments. By leveraging advanced DL techniques, such as transfer learning and federated learning, researchers can address the complexities of channel estimation across different frequency bands and wireless resources. The adoption of the RESNET50 architecture offers a promising framework for efficient and accurate estimation, further advancing the capabilities of future 6G communication systems. ResNet 50 had the highest accuracy of 99.75% with a loss rate of 0.33, while the other models achieved 98.16%, 98.47%, and 98.56%, respectively. Furthermore, ResNet 50 achieved a validation accuracy of 99.69%, precision of 99.50%, F1- score of 99.70, and AUC of 99.83%. |
Quick Links
