IdleFL: A Hybrid Peer-to-Peer Distributed Computing Platform for Privacy-Preserving Machine Learning Using Idle Resource Harvesting Durvish Khurana, Chandramohan Dhasarathan Proceedings of the 4th IEEE International Conference on Interdisciplinary Approaches in Technology and Management for Social Innovation Iatmsi 2026, 2026 IdleFL is a hybrid peer-to-peer distributed computing framework that leverages idle computational resources from heterogeneous personal devices to accelerate machine learning (ML) training while preserving data privacy through strict data locality. In contrast to centralized cloud-based solutions, IdleFL enables decentralized collaborative training through federated learning and secure on-device execution. The system adopts a three-layer architecture inspired by operating system design principles, integrating scheduling, sandboxing, and checkpoint-based fault tolerance to support robust and adaptive distributed computation. A real-world deployment across a heterogeneous five-device mesh demonstrates a training speedup of up to <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$3.3 \times$</tex> while maintaining 94.2% model accuracy relative to single-device baselines, with complete task recovery under tested failure scenarios. These results demonstrate the practicality of a free, privacy-preserving distributed ML platform based on idle-resource harvesting.
Integrating AutoML and Explainability: A Unified Approach for Decision-Making in Engineering and Social Sciences Responsible AI Principles and Practices, 2026
A modified social spider algorithm for an efficient data dissemination in VANET Achyut Shankar, Rajaguru Dayalan, Chinmay Chakraborty, Chandramohan Dhasarathan, Manish Kumar Environment Development and Sustainability, 2025 Technical growth in the field of communication and information is an important aspect in the development and innovation of industrial automation and in the recent advances in the field of communications. The recent development of mobile communications has led to worldwide ubiquitous information sharing and has rehabilitated human lifestyles. This communication revolution is now introducing effective information sharing into the automotive industry. The current technology is extending this field of applications for vehicle safety, improving the efficiency in traffic management, offering reliable assistance for drivers and supporting the modern field of vehicle design. With these advances, the vehicular network concept has grabbed worldwide attention. In this article, a novel sampling-based estimation scheme (SES), to initiate the involvements and increase the probabilistic contacts of vehicle communication. The scheme is divided into a few segments, for ease of operations with a perfect sample. The contact duration between two vehicles moving in opposite directions on their overlapped road is lower, but their contact probability is higher. By contrast, the duration of the contact between two vehicles moving in the same direction on their overlapped road is higher, but their contact probability is lower. SES can easily obtain efficient routing by considering the above-mentioned stochastic contacts. Furthermore, we investigate the content transmission among the probabilistic contacts, by using the flow model with probabilistic capacities. The performance of the proposed SES is experimentally validated with the probabilistic contacts in VANETs.
Tensor RT optimized driver drowsiness detection system using edge device Chandramohan Dhasarathan, Sambasivam Gnanasekaran, Arnab Pattanayak, Gourav Kumar, Kartik Vig, Vaibhav Narain, K.M. Deva Narayan, Sunidhi Garg Ain Shams Engineering Journal, 2025 Driver drowsiness has emerged as a major issue in terms of road safety, causing dangerous and sometimes even fatal accidents. Fatalities, serious injuries, monetary losses, and property damages highlight the direct consequences of driver drowsiness. Major causes include driver sleepiness, fatigue, long journeys, distraction (physical and digital), and medical reasons. A reliable, non-invasive, and accurate technology that can detect and alert the driver in case of drowsiness, fatigue or distraction can help avert this rising problem. In this paper, alongwith the proposed project, we have analysed and reviewed the latest research, studies, and experiments in the field of driver drowsiness detection and prediction analysing systems. The proposed approach utilizes multiple models to improve and accurately detecting driver drowsiness, with the models being InceptionV3, ResNet50, VGG-16, and MobileNetV1. The system uses transfer learning techniques for implementing CNN model algorithms to analyze live video from the camera module, allowing for real-time detection of driver behavior such as fatigue or distraction. Utilizing the computing power of the edge device, Jetson Nano and the optimization capabilities of TensorRT, the system achieves rapid inference of input data, enabling rapid decision making based on analyzed input data. An audio warning system set up with a DIY piezoelectric buzzer will be attached to the Jetson Nano to provide real-time feedback to the driver, improving overall road safety. With the proposed project, we were able to achieve highest accuracy of 98.27% and throughput of 141.785 frames/second with MobileNetV1 model. Through TensorRT optimization we were able to achieve maximum throughput of 683.923 frames/second (significant increase of 79.27%) and an accuracy of 98.19% (negligible decrease). Overall, this report presents a significant contribution towards driver drowsiness detection systems by proposing a real time, low-cost, and accurate system that can be installed in vehicles to ensure safe transportation and prevent accidents.
Safeguarding user privacy at the edge of fog computing networks in decentralized distributed computing Chandramohan Dhasarathan, Reddy B. Ramachandra, Diwakar Tripathi Swarm Intelligence Theory and Applications in Fog Computing Beyond 5g Networks and Information Security, 2025 Establishing secure identities and authentication measures is crucial to thwart unauthorized access to fog nodes and edge devices. Fog computing environments offer enhanced user privacy protection while leveraging the advantages of decentralized and distributed computing at the network’s edge. Directly conducting data analytics on edge devices without transmitting raw data to the cloud minimizes the risk of sensational sensitive evidence during the data processing phase. Incorporating rigorous access mechanism contrivances ensures that solitary ratified individuals can enter sensitive data, employing role-based control supervision skills. Additionally, homomorphic encryption is advocated, as it enables computation on encrypted statistics without the prerequisite for decryption, facilitating protected data dispensation while preserving privacy. The escalating pervasiveness of IoT diplomacies and growing demand for low-latency applications, fog-based research focus has arisen as a auspicious paradigm that encompasses the cloud computing capabilities of network operations. Nevertheless, the decentralized and distributed nature of fog computing raises substantial concerns regarding the privacy and security of user data. This research article explores various strategies and techniques to safeguard user privacy within the context of decentralized and distributed computing in fog computing networks. The examination encompasses encryption, anonymization, access control, and other privacy-preserving mechanisms designed to address the distinctive challenges posed by the fog computing environment.
Optimizing Blockchain Integration for Secure and Scalable Internet of Medical Things in Healthcare Applications Chandramohan Dhasarathan, Puviyarasi Thirugnanasammandamoorthi, B. Ramachandra Reddy, Diwakar Tripathi Privacy and Security Inss Fintech Healthcare and Social Applications, 2025 The integration of Blockchain technology with the Internet of Medical Things (IoMT) presents transformative potential for healthcare, enhancing data security, privacy, and transparency. As IoMT devices collect and transmit sensitive health data, ensuring privacy and preventing unauthorized access become critical concerns. Blockchain offers a decentralized, immutable ledger that can address these challenges by providing secure transaction recording and audit trails. However, limitations related to scalability and efficiency remain obstacles to broad adoption. This research explores various optimization strategies such as consensus algorithm improvements (e.g., Proof of Stake over Proof of Work), hybrid Blockchain models, and off-chain storage to enhance performance in IoMT environments. Lightweight cryptographic protocols are also proposed to reduce device overhead. Through simulations and real-world case studies, we evaluate these strategies in terms of latency, energy efficiency, security, and compliance. Results indicate that Blockchain, when optimized, significantly enhances trust, interoperability, and usability in healthcare IoMT applications.
Legendre neural network method for solving nonlinear singular systems Intelligent Technologies for Sensors Applications Design and Optimization for A Smart World, 2023
The future of web crowdfunding: An ethereum blockchain approach Intelligent Technologies for Sensors Applications Design and Optimization for A Smart World, 2023
An agent based model for service composition in pervasive computing: Web service computing approach International Journal of Applied Engineering Research, 2014
Web service suitability assessment for cloud computing M. S. Nanda Kishore, S. K. V. Jayakumar, G. Satya Reddy, P. Dhavachelvan, D. Chandramohan, N. P. Soumya Reddy Communications in Computer and Information Science, 2011
RECENT SCHOLAR PUBLICATIONS
Integrating AutoML and Explainability: A Unified Approach for Decision‐Making in Engineering and Social Sciences A Dalmia, C Dhasarathan Responsible AI: Principles and Practices, 175-198 , 2026 2026
When to Hash and When to Learn: A Systematic Performance Analysis of Deterministic Indexing versus Neural Inference for Text Retrieval B Kumar, Y Verma, C Dhasarathan 2026 IEEE International Conference on AI Engineering and Innovations (AIEI), 1-6 , 2026 2026
IdleFL: A Hybrid Peer-to-Peer Distributed Computing Platform for Privacy-Preserving Machine Learning Using Idle Resource Harvesting D Khurana, C Dhasarathan 2026 IEEE International Conference on Interdisciplinary Approaches in … , 2026 2026
Quantum Computing—Real-World Applications and Its Transformative Potential S Goundar, K Shruthi, A Nuhiu, C Dhasarathan, A Bhardwaj Quantum Computing, 233-252 , 2026 2026
Tensor RT optimized driver drowsiness detection system using edge device C Dhasarathan, S Gnanasekaran, A Pattanayak, G Kumar, K Vig, ... Ain Shams Engineering Journal 16 (10), 103620 , 2025 2025 Citations: 10
A modified social spider algorithm for an efficient data dissemination in VANET A Shankar, R Dayalan, C Chakraborty, C Dhasarathan, M Kumar Environment, Development and Sustainability 27 (10), 24659-24702 , 2025 2025 Citations: 24
3D Voronoi Diagram Division-Based Hybrid Weighted Regression Localization Algorithm MK Hasan, S Khapre, C Dhasarathan, S Islam, FRA Ahmed, TE Ahmed, ... Tsinghua Science and Technology 30 (6), 1-24 , 2025 2025 Citations: 1
A low-cost AI-powered system for early detection of diabetic retinopathy and ocular diseases in resource limited settings H Vohra, MK Hasan, SNHS Abdullah, S Islam, AH Abd Rahman, ... IEEE Access 13, 97322-97336 , 2025 2025 Citations: 5
Proactive Fall and Vehicle Crash Detection in Healthcare: Advancing Patient Safety with Intelligent Computer Vision and Human Activity Recognition Systems H Vohra, C Dhasarathan, Y Mathur, R Samaiya, S Verma, JS Monacha, ... International Conference on Smart Computing and Informatics, 289-306 , 2025 2025
Safeguarding user privacy at the edge of fog computing networks in decentralized distributed computing C Dhasarathan, RB Ramachandra, D Tripathi Swarm Intelligence, 133-155 , 2025 2025 Citations: 8
Blockchain-based intelligent digital credentialing system for participatory governance: Design, implementation, and potential implications C Dhasarathan, D Rajaguru, J Subash Chandra Bose SN Computer Science 5 (8), 1051 , 2024 2024 Citations: 7
Digital Twin for Sustainable Farming: Developing User-Friendly Interfaces for Informed DecisionMaking and Increased Profitability C Dhasarathan, S Gnanasekaran The Future of Agriculture: IoT, AI and Blockchain Technology for Sustainable … , 2024 2024
Nature-inspired optimization algorithms based feature selection: Application in credit scoring D Tripathi, BR Reddy, S Dwivedi, AK Shukla, D Chandramohan, ... Journal of Intelligent & Fuzzy Systems, JIFS-219413 , 2024 2024 Citations: 6
Data privacy model using blockchain reinforcement federated learning approach for scalable internet of medical things C Dhasaratha, MK Hasan, S Islam, S Khapre, S Abdullah, TM Ghazal, ... CAAI Transactions on Intelligence Technology , 2024 2024 Citations: 131
Sentimental analysis and prediction of socioeconomic disasters tweets by ML and regular expression P Thirugnanasammandamoorthi, H Kumar, D Ghosh, C Dhasarathan, ... Journal of Intelligent & Fuzzy Systems, JIFS-219417 , 2024 2024 Citations: 2
A nomadic multi-agent based privacy metrics for e-health care: a deep learning approach C Dhasarathan, M Shanmugam, M Kumar, D Tripathi, S Khapre, ... Multimedia Tools and Applications 83 (3), 7249-7272 , 2024 2024 Citations: 54
Application areas, benefits, and research challenges of converging Blockchain and machine learning techniques A Manimaran, S Goundar, D Chandramohan, N Arulkumar Integrating Blockchain and Artificial Intelligence for Industry 4.0 … , 2023 2023 Citations: 10
Internet of Things and Blockchain in Healthcare: Challenges and Solutions N Arulkumar, A Manimaran, D Chandramohan, S Goundar Integrating Blockchain and Artificial Intelligence for Industry 4.0 … , 2023 2023 Citations: 2
Legendre neural network method for solving nonlinear singular systems M Veerasamy, SCB Jaganathan, C Dhasarathan, A Mubarakali, ... Intelligent Technologies for Sensors, 25-37 , 2023 2023 Citations: 3
The Future of Web Crowdfunding: An Ethereum Blockchain Approach SCB Jaganathan, M Veerasamy, A Mubarakali, C Dhasarathan, ... Intelligent Technologies for Sensors, 337-372 , 2023 2023 Citations: 3
MOST CITED SCHOLAR PUBLICATIONS
Prediction of COVID-19 using genetic deep learning convolutional neural network (GDCNN) RG Babukarthik, VAK Adiga, G Sambasivam, D Chandramohan, ... Ieee Access 8, 177647-177666 , 2020 2020 Citations: 145
Data privacy model using blockchain reinforcement federated learning approach for scalable internet of medical things C Dhasaratha, MK Hasan, S Islam, S Khapre, S Abdullah, TM Ghazal, ... CAAI Transactions on Intelligence Technology , 2024 2024 Citations: 131
Minimizing the makespan using Hybrid algorithm for cloud computing R Raju, RG Babukarthik, D Chandramohan, P Dhavachelvan, ... 2013 3rd IEEE international advance computing conference (IACC), 957-962 , 2013 2013 Citations: 68
COVID-19 health data analysis and personal data preserving: A homomorphic privacy enforcement approach C Dhasarathan, MK Hasan, S Islam, S Abdullah, UA Mokhtar, AR Javed, ... Computer communications 199, 87-97 , 2023 2023 Citations: 66
A study on metaheuristics approaches for gene selection in microarray data: algorithms, applications and open challenges AK Shukla, D Tripathi, BR Reddy, D Chandramohan Evolutionary intelligence 13 (3), 309-329 , 2020 2020 Citations: 66
User privacy prevention model using supervised federated learning‐based block chain approach for internet of Medical Things C Dhasarathan, MK Hasan, S Islam, S Abdullah, S Khapre, D Singh, ... CAAI Transactions on Intelligence Technology , 2023 2023 Citations: 58
Credit scoring models using ensemble learning and classification approaches: a comprehensive survey D Tripathi, AK Shukla, BR Reddy, GS Bopche, D Chandramohan Wireless Personal Communications 123 (1), 785-812 , 2022 2022 Citations: 57
A nomadic multi-agent based privacy metrics for e-health care: a deep learning approach C Dhasarathan, M Shanmugam, M Kumar, D Tripathi, S Khapre, ... Multimedia Tools and Applications 83 (3), 7249-7272 , 2024 2024 Citations: 54
A bio-inspired privacy-preserving framework for healthcare systems: C. Dhasarathan et al. C Dhasarathan, M Kumar, AK Srivastava, F Al-Turjman, A Shankar, ... The Journal of Supercomputing 77 (10), 11099-11134 , 2021 2021 Citations: 54
A secure data privacy preservation for on-demand cloud service D Chandramohan, T Vengattaraman, P Dhavachelvan Journal of King Saud University-Engineering Sciences 29 (2), 144-150 , 2017 2017 Citations: 46
Energy-aware multilevel clustering scheme for underwater wireless sensor networks S Chinnasamy, J Naveen, PJA Alphonse, C Dhasarathan, G Sambasivam IEEE Access 10, 55868-55875 , 2022 2022 Citations: 36
Natural language processing based sentimental analysis of Hindi (SAH) script an optimization approach H Shrestha, C Dhasarathan, S Munisamy, A Jayavel International Journal of Speech Technology 23 (4), 757-766 , 2020 2020 Citations: 35
Data privacy breach prevention framework for the cloud service C Dhasarathan, V Thirumal, D Ponnurangam Security and Communication Networks 8 (6), 982-1005 , 2015 2015 Citations: 30
A multi-agent approach: To preserve user information privacy for a pervasive and ubiquitous environment D Chandramohan, D Sathian, D Rajaguru, T Vengattaraman, ... Egyptian Informatics Journal 16 (1), 151-166 , 2015 2015 Citations: 28
A comprehensive novel model for network speech anomaly detection system using deep learning approach A Manimaran, D Chandramohan, SG Shrinivas, N Arulkumar International Journal of Speech Technology 23 (2), 305-313 , 2020 2020 Citations: 27
A new privacy preserving technique for cloud service user endorsement using multi-agents D Chandramohan, T Vengattaraman, D Rajaguru, P Dhavachelvan Journal of King Saud University-Computer and Information Sciences 28 (1), 37-54 , 2016 2016 Citations: 27
A modified social spider algorithm for an efficient data dissemination in VANET A Shankar, R Dayalan, C Chakraborty, C Dhasarathan, M Kumar Environment, Development and Sustainability 27 (10), 24659-24702 , 2025 2025 Citations: 24
A novel approach for multi-constraints knapsack problem using cluster particle swarm optimization RG Babukarthik, C Dhasarathan, M Kumar, A Shankar, S Thakur, ... Computers & Electrical Engineering 96, 107399 , 2021 2021 Citations: 19
2M2C-R2ED: Multi-metric cooperative clustering based routing for energy efficient data dissemination in green-VANETs D Chandramohan, A Dumka, L Jayakumar Technology and Economics of Smart Grids and Sustainable Energy 5 (1), 15 , 2020 2020 Citations: 18
FEWSS—framework to evaluate the service suitability and privacy in a distributed web service environment D Chandramohan, T Vengattaraman, P Dhavachelvan, R Baskaran, ... International Journal of Modeling, Simulation, and Scientific Computing 5 … , 2014 2014 Citations: 18
