Dr. Rajkumar S C is a highly accomplished individual with a strong background in Computer Science and Engineering. Currently, I hold the position of a Teaching Fellow at Anna University Regional Campus Madurai. My academic journey includes a B.E., M.E., and Ph.D. in Computer Science and Engineering, showcasing my dedication to knowledge acquisition and expertise in the field.
With close to 9 years of experience in teaching, I am deeply committed to academia and the development of future professionals. My areas of specialization revolve around Intelligent Transportation Systems, Deep Learning, Cloud Technologies, and Machine Learning & IoT Technologies, where I continuously strive to expand my knowledge and contribute to advancements in these domains.
My research contributions have been significant, and I take pride in the publication of multiple papers in esteemed international journals such as Elsevier and Wiley.
RESEARCH, TEACHING, or OTHER INTERESTS
Computer Engineering, Artificial Intelligence, Information Systems, Human-Computer Interaction
12
Scopus Publications
Scopus Publications
A zero-shot LLM framework for multimodal grievance classification, urgency scoring, and abuse detection in civic feedback systems S. C. Rajkumar, D. Yuvasini, Shitharth Selvarajan, Nithya Rekha Sivakumar Scientific Reports, 2026 A unified model is presented for civic grievance redressal, integrating multimodal complaint intake, zero-shot semantic routing, sentiment-derived urgency estimation, and behavior-sensitive abuse detection within a scalable microservice architecture. The framework consolidates components that are typically handled independently by combining transformer-based text processing, CTC-enabled speech transcription, affective-intensity modeling, and longitudinal user-behavior analysis into a coherent decision pipeline. Typed and spoken complaints are projected into a shared semantic representation using a MobileBERT zero-shot classifier, while a recurrent neural network trained with Connectionist Temporal Classification (CTC) provides robust transcription of multilingual and dialect-rich voice submissions. Urgency indicators obtained from lexicon-based sentiment analysis are incorporated into time-aware escalation logic, and abuse mitigation integrates toxicity scores with a repetition-weighted behavioral model to identify and regulate systematic misuse. The platform operates as a containerized microservice ecosystem with WebSocket-enabled real-time updates and AES-encrypted data storage. Experiments conducted on a 1000-sample multimodal dataset show consistent performance, including 92.4% routing accuracy, 0.041 MAE in urgency estimation, 96.2% toxicity precision, 96.8% SLA compliance, and sub-150 ms end-to-end latency. These outcomes indicate suitability for deployment in linguistically diverse and resource-constrained civic environments. Planned extensions include enhanced multilingual ASR, adversarially robust toxicity modeling, and incorporation of image-based grievance modalities.
A hybrid approach combining images and questionnaires for early detection and severity assessment of Autism Spectrum Disorder Rajkumar S.C., Stefano Cirillo, Yuvasini D., Luisa Solimando Image and Vision Computing, 2025 In this research, we propose a novel integrated system for the early diagnosis and cognitive enhancement of infants with Autism Spectrum Disorder (ASD). The system combines two core modules: the Behavioral Analytic Module and the Cognitive Skill Enhancement Module. The Behavioral Analytic Module includes a Questionnaire Analysis Sub-module, which utilizes Random Forest classifiers to analyze questionnaire data, and an Image Analysis Sub-module, which employs a fine-tuned VGG16 Convolutional Neural Network to process facial images. These sub-modules independently assess ASD likelihood and combine their outputs to generate a comprehensive diagnosis using a weighted averaging technique. The Cognitive Skill Enhancement Module integrates interactive games and web-based animations designed to improve cognitive abilities and daily living skills in toddlers with ASD. Additionally, a reward system is incorporated to reinforcement learning outcomes, adaptively calculating rewards based on the infants’ progress. The proposed system aims to provide a holistic approach to ASD diagnosis and intervention, offering an effective tool for early detection and tailored cognitive development. The system’s efficacy is demonstrated through comparative analysis, showing a 93% improvement in diagnostic accuracy and a 92% enhancement in cognitive skill development among toddlers with ASD.
Automatic and Enhanced Reflexive System for Concise Answer Evaluation Using BERT Model and Bi-LSTM Jegatha Deborah L., Rajkumar S. C., K. Suganya Navigating Usability and User Experience in A Multi Platform World, 2024 In intelligent tutoring systems (ITS), the automatic and enhanced reflexive system for concise answer evaluation (AERSCAE) components that evaluate students' responses to questions are crucial. However, applying deep learning to AERSCAE encounters difficulties due to the challenges of accurately scoring short answers and the scarcity of training data. Using bidirectional encoder representations from transformers (BERT)-based deep neural network, this system improves short-answer texts' understanding via capturing contextual nuances. A specialized semantic refinement layer, bidirectional long short-term memory (Bi-LSTM), is integrated into the AERSCAE system to enhance its performance. Bi-LSTM's temporal dependencies and bidirectional processing enable the generation of high-precision scores. Integrating BERT and Bi-LSTM deep learning architectures into AERSCAE systems enhances their short-answer evaluation capability.
Secure session key pairing and a lightweight key authentication scheme for liable drone services Rajkumar ．S．C, Jegatha Deborah ．L, Vijayakumar ．P, Karthick ．KR Cyber Security and Applications, 2023 Recent advancements in drone technology have created new application opportunities, particularly for small drones. However, these advancements raise concerns about security, adaptability, and consistency. Data security is jeopardized by flying intelligent devices. The distributed nature of drones, their accessibility, mobility, adaptability, and autonomy will all have an effect on how security vulnerabilities and threats are identified and controlled. However, attackers and cybercriminals have begun to employ drones for malevolent reasons in recent years. These attacks are frequent and can be fatal. There is also the matter of prevention to consider. The communication entities of the drone network can communicate securely via authentication procedures. Such solutions, however, must strike a balance between security and portability. However, the proposed technique is implemented to improve security to avoid attacks and provides a secure, lightweight, and proven solution to a key agreement for drone communication. A novel certificate-less Drone integration approach that depends on trusted authorities centres to help communication entities establish their key pairs while keeping those same trusted authorities centers from knowing about them has been devised. The proposed scheme results achieved higher security of 94 percent than existing schemes.
Optimized traffic flow prediction based on cluster formation and reinforcement learning S.C. Rajkumar, Jegatha Deborah L., Vijayakumar P. International Journal of Communication Systems, 2023 SummaryIn recent days, the traffic flow information is collected using the global positioning system through the Internet, which is yet to become ubiquitous. A novel technique is proposed for the intelligent transportation system, which leads to reduce the traffic congestion that will become an unavoidable phenomenon in the near future. This system uses a magnetic sensor to identify the type of the vehicle and the exact vehicle count in the traffic environment based on variation in the magnetic flux. This information is transmitted to the cloud server with the help of cluster by utilizing the nearby proximity services. An intelligent agent that uses reinforcement learning is implemented in the cloud server to learn the real‐time traffic flow from multiple sources for the prediction of a valid and optimized route suggestion for the registered users. This work is implemented, and implementation results show that the proposed work achieves an accuracy of 98.36%. Hence, this intelligence method for VANETs will certainly account for improved traffic prediction to the vehicle transportation. It can reduce the vehicles waiting time in traffic and that would minimize the fuel consumption. It will make an eco‐friendly environment of reduced carbon dioxide emissions in urban cities.
A novel secure e-learning model for accurate recommendations of learning objects R. Karthika, S.C. Rajkumar, L. Jegatha Deborah, S. Geetha Secure Data Management for Online Learning Applications, 2023 Since the COVID-19 pandemic situation, secure e-learning recommendation systems are the preferred mode of learning. They enable time and location-independent learning. Even though both institutions and students accept the use of technology in their educational processes, personalized course learning materials based on each student’s learning preferences are not without limits and security problems. No-blockchain recommendation engines, including collaborative and content-based filtering, may now better focus their recommendations by incorporating information about the users and the content they’re looking at. There is no consideration for the preferences of students, such as their preferred methods of learning, which will help to ensure the accuracy and security of course learning materials. Thus, in order to address these issues and enhance the consistency, security, and quality of recommendation systems, we propose an innovative secure decision support algorithm that incorporates ratings and learning styles in order to securely deliver courses as learning objects. Additionally, the proposed work focuses on enhancing security while conducting online courses and assessments through the use of a novel secure learning algorithm. This contributes to the enhancement of the secure and predictive rating of learning objects and recommends to the e-learner the most highly rated learning objects. Furthermore, the experiment showed that integrating collaborative filtering and content-based filtering techniques to develop our suggested secure hybrid algorithm resulted in a more accurate predictions for e-learning purposes.
Passive-Awake Energy Conscious Power Consumption in Smart Electric Vehicles Using Cluster Type Cloud Communication Pandi Vijayakumar, S. C. Rajkumar, L. Jegatha Deborah International Journal of Cloud Applications and Computing, 2022 Nowadays, electric vehicles (e-vehicles) have a significant impact on the current intelligent transportation system, with the goal of establishing a smart environment in the near future. Furthermore, when an intelligent system is integrated with IoT technologies, it produces more efficient results to the society. This research work examines the impact of energy degradation on the wireless transmission to optimize power consumption using a passive-awake cloud-cluster communication system, thereby extending the lifetime of an energy-constrained electric vehicle. Wireless communication means that electromagnetic waves draining a steady amount of energy from the condenser, even if the device is not connected to the internet, which constitutes the main constraint for a long-distance electric vehicle. In this paper, a passive-awake assistant is proposed, which significantly reduces power consumption.
An improved public transportation system for effective usage of vehicles in intelligent transportation system S.C. Rajkumar, L. Jegatha Deborah International Journal of Communication Systems, 2021 SummaryProcuring usage of the public transportation system enhances the promising effect of limiting the number of own vehicles usage in the contemporary world. The present research advocates a new paradigm of the Intelligent Transportation System (ITS) in the near future, to rescue fossil fuel and to maintain a healthy environment for the current generation. To provide this facility, Long Short Term Memory (LSTM) based intelligent learner has been proposed. This intelligent learner is mainly used to predict high vehicle demand requests in order to utilize a public transport system effectively. In this way, excess usages of vehicles are reduced from low vehicle demand request locations to the locations where high vehicles demand requests are generated. Moreover, a new enhanced approach has also been designed to establish communication between the onboard vehicles and the passengers for instant reservation of their seats based on real‐time sensors. To achieve the effective usage of the public transportation system, an effective dynamic scheduling algorithm that dedicates more convenient travel in the complex transportation system, has been proposed. The proposed system results are evaluated using real‐time transport data, which are collected from major cities and they are implemented to predict the exact vehicles demand. The performance results are compared with various existing methods and the proposed system has proved its efficiency than the existing methods. When the proposed system is implemented, it improves 87% usage of public transportation as well as the usage of taxis and own vehicles would be reduced drastically in the city.
