Published March 2026

Special Issue

No description available.

Volume 1

Issue 1

Articles in this Special Issue

Secure online banking REST API

Authors:Mathan Kumar, Nivash, Sathya Sri, Deepan*, Senthil Prakash*

Abstract

Online banking has become an indispensable part of the modern financial ecosystem, necessitating systems that are not onlyhighly functional but also rigorously secure and scalable. This paper presents a robust backend solution designed using JavaFull Stack technologies, specifically the Spring Boot Framework, to replace traditional, monolithic banking software with alightweight, decoupled RESTful architecture. The proposed system addresses critical security vulnerabilities inherent in legacyarchitectures by implementing Stateless Authentication using JSON Web Tokens (JWT) and Spring Security, effectivelymitigating risks associated with Cross-Site Request Forgery (CSRF) and session-based attacks. Data integrity is mathematicallyguaranteed through ACID-compliant transaction management using the @Transactional architecture, ensuring that fundtransfers are processed reliably without data loss during network failures. Furthermore, the application utilizes Spring Data JPAfor efficient interaction with a MySQL database, providing built-in protection against SQL injection. This research demonstratesa production-ready approach to building secure financial applications, prioritizing modularity, maintainability, and advancedcryptographic data protection.

Keywords

REST API, spring boot, JSON Web Token (JWT), stateless authentication, acid transactions, and cybersecurity.

Download

Educational rag system for specialized textbook

Authors:Muthupandi, Praveen, Praveen, Mahalakshmi*, Senthil Prakash*

Abstract

Large Language Models (LLMs) have shown strong abilities in understanding and generating natural language, but they oftenproduce inaccurate or misleading responses when applied to specialized educational domains. This paper introduces anEducational Retrieval-Augmented Generation (RAG) system tailored for textbook-based learning, where accuracy is critical.Instead of depending completely on the model's internal knowledge, the method recovers appropriate knowledge from acurated collection of domain-specific textbooks. It utilizes a fusion recovery methodology that links keyword matching withsemantic similarity to detect the most appropriate content, which acts as context for a locally deployed LLM to make preciseanswers. Developed as a web-based platform using frameworks like Django, the system minimizes dependency on externalAPIs and includes a preprocessing mechanism to convert visual elements like diagrams into descriptive text. Overall, thesystem provides accurate, explainable, and domain-focused educational support.

Keywords

Retrieval-augmented generation, large language models, hybrid retrieval semantic search, hallucination reduction.

Download

Reconnectify alumini association platform

Authors:Hethenthira, Sriram, Vakish, Mahalakshmi*, Senthil Prakash*

Abstract

Reconnectify-Alumini Association Platform is a web-based application designed to establish and maintain a strong connectionbetween alumini, students, and educational institutions. In many institutions, alumini data and communication are managedusing traditional methods such as spreadsheets, social media groups, and manual records, which often lead to inefficiency,lack of organization, and limited interaction. This project proposes a centralized and structured digital platform for aluminiengagement. The system enables alumini to create professional profiles, share job opportunities, and participate in events,while students can connect with alumini for mentorship, career guidance, and networking. Administrators can manage users,monitor activities, and control system operations efficiently. Key features include user authentication, profile management,alumini networking, event management, job posting, messaging, and notification services. The application is developed usingthe MERN, and MongoDB for database management. Secure authentication is implemented using JSON Web Tokens (JWT),and user passwords are protected through bcrypt encryption. The system follows a modular architecture ensuring scalability,maintainability, and flexibility for future enhancements.

Keywords

Alumini management, MERN stack, web application, role-based access control, JWT authentication, mongodb, node.js, react.js, student networking, event management

Download

Adaptive traffic signal control using reinforcement learning

Authors:Dhanush, Ragunath, Visweshwaran, Senthil Prakash*

Abstract

This paper presents the design, development, and experimental evaluation of a Reinforcement Learning (RL) based AdaptiveTraffic Signal Control (ATSC) system for intelligent urban traffic management. The proposed approach formulates traffic signaloptimisation as a Markov Decision Process (MDP) and applies Deep Reinforcement Learning (DRL) algorithms specificallyDeep Q-Networks (DQN), Proximal Policy Optimization (PPO), and Advantage Actor-Critic (A2C) to enable real-time adaptivedecision-making under dynamic traffic conditions. The system is organized into four integrated modules covering userinteraction and visualization, RL algorithm execution, reward-based feedback, and traffic simulation. Experimental outcomesshow that RL agents, particularly PPO, consistently surpass conventional fixed-time and actuated control methods across keymetrics including average waiting time, queue length, throughput, and estimated emissions. Network-wide efficiency is furtherimproved through multi-agent coordination using a Centralised Training and Decentralised Execution (CTDE) strategy. Theframework provides a scalable, modular, and extensible platform for future research and smart city integration.

Keywords

Reinforcement learning, traffic signal control, markov decision process, deep q-network, ppo, sumo, multi-agent systems, smart city, adaptive control.

Download

Farm management and crop marketing system

Authors:Bhavatharani, Dhanya Sri, Santhoshini, Senthil Prakash*

Abstract

This study describes the creation and implementation of a Farm Management and Crop Marketing System designed to enhancefarm productivity and improve farmers’ access to markets. The system offers an online platform to handle various agriculturalactivities, including crop monitoring, expense tracking, and income tracking. It enables direct communication between farmersand customers, reducing dependency on intermediaries. The proposed system integrates farm data management with an onlinecrop marketing platform, enabling farmers to maintain accurate records and make better decision-making. Features such asreal- time data updates, financial analysis, and direct marketing improve efficiency and transparency. The system alsoaddresses challenges such as lack of digital awareness, poor record-keeping and limited access to markets. Experimentalanalysis shows that the system enhances productivity, minimizes human work, and improves financial returns. Upcomingimprovements involve adding weather prediction features and AI-driven crop suggestions.

Keywords

Farm management, crop marketing, agriculture system, financial tracking, digital farming.

Download

Personalized adaptive tutoring system using knowledge tracing

Authors:Asin, Aslee, Sowndarya, Senthil Prakash*

Abstract

This paper evaluates the effectiveness of a personalized adaptive tutoring system using knowledge tracing as an intelligentstyle toward increase beginner studying. Outcomes. Conventional learning methods regularly implement a repaired uniformteaching methodology, which fails to address individual learning differences. To overcome this limitation, the proposed systemleverages knowledge tracing techniques to model and monitor a student’s learning progress over time. Knowledge Tracingenables the system to estimate a learner’s mastery level by analysing their responses to questions and interactions withlearning content. Based on this continuous assessment, the system dynamically adapts the difficulty level, content sequence,and type of instructional materials to suit each student’s needs. The study explores various knowledge tracing models,including probabilistic and machine learning-based approaches, for accurate prediction of student performance.

Keywords

Knowledge tracing, adaptive tutoring system, personalized learning, student modeling, machine learning, elearning.

Download

Secure token based authentication with JWT and rest API using rest framework

Authors:Prasanth, Sivamani, Tharun, Balamurali*, Senthil Prakash*

Abstract

This research paper investigates the architectural potential of JSON Web Tokens (JWT) and Django REST Framework (DRF)as a robust, secure, and stateless infrastructure for modern web-based authentication systems. In the evolving landscape ofcyber threats, traditional session-based authentication mechanisms often suffer from high server overhead and scalabilitybottlenecks due to persistent state management. This study explores the implementation of a decentralized authenticationmodel that leverages cryptographer signing to ensure data integrity and user privacy. By utilizing a stateless approach, wedemonstrate how JWT facilitates seamless cross-domain communication and reduces latency in distributed environments. Thepaper further discusses token-based security protocols, including access and refresh token cycles, while comparing theefficiency of this model against conventional von Neumann-inspired session architectures in terms of speed, energyconsumption, and on-chip learning for API security.

Keywords

JWT, Django REST framework, stateless authentication, API security, token-based auth, scalable architecture.

Download

Transfer learning based-plant disease diagnosis system

Authors:Elavarasi, Sanjitha, Santhiya, Senthil Prakash*

Abstract

This paper presents an optimized system for plant disease detection using transfer learning techniques. Plant diseasessignificantly affect crop yield and quality, making early detection making early detection crucial for sustainable agriculture. Theproposed system employs pre-trained deep learning models like ResNet, VGG, and MobileNet to classify plant leaf diseasesfrom images The methodology includes image collection, preprocessing, data augmentation, feature extraction, andclassification. Transfer learning enables the model to leverage knowledge from large-scale datasets, reducing trainingduration and improving accuracy even with limited data. The organization is capable of real-time disease detection and can bedeployed on mobile or web platforms edge devices to assist farmers. Experimental results show that the proposed systemachieves high accuracy and efficiency compared to traditional approaches.

Keywords

Plant disease classification, transmission learning, deep learning, CNN, image classification, agriculture.

Download

Homomorphic encryption electronic voting system

Authors:Iswarya, Monisha, Praveena, Senthil Prakash*

Abstract

This paper presents a secure and privacy-preserving electronic voting system using homomorphic encryption. In traditionalelectronic voting systems, ensuring data security, voter anonymity, and result integrity remains a major challenge. Theproposed system addresses these issues by encrypting votes at the time of casting and allowing computations to be performeddirectly on encrypted data without decryption. Homomorphic encryption ensures that sensitive information is never exposedduring processing, thereby enhancing security and privacy. The system also prevents vote tampering, unauthorized access,and data manipulation. This approach provides a transparent, efficient, and reliable voting mechanism suitable for moderndigital elections. The proposed model ensures accuracy, confidentiality, and trust in the electoral process.

Keywords

Homomorphic encryption, electronic voting, security, privacy, cryptography.

Download

Student collaboration study buddy portal

Authors:Jayasri, Sruthi Yazhini, Swathi Murugapriya, Senthil Prakash*

Abstract

The increasing need for effective peer-to-peer learning in digital education has necessitated the development of student-centric collaborative platforms. This paper presents the Student Collaboration Study Buddy System, a full-stack web application designed to enable students to connect, communicate, and collaborate through a unified digital environment. Traditional learning methods often lack efficient peer interaction and centralized access to study materials, which can limit student engagement and knowledge exchange. To address these challenges, the proposed system offers a structured and user-friendly digital solution that connects learners and facilitates effective collaboration. The portal is developed using modern full-stack architecture, with react employed for building a dynamic and responsive user interface, Node.js and Express.js for handling server-side logic and API integration and Mongo DB for scalable and flexible data storage. The system incorporates secure user authentication and authorization mechanisms to protect user data and ensure controlled access to resources. Key features of the Study Buddy Portal include user profile management, creation and management of study groups, real-time chat and discussion forums, file and note sharing, and a personalized dashboard that provides quick access to relevant activities and materials. Furthermore, the system is designed to be scalable and adaptable for future enhancements. Future developments of the portal may include the integration of artificial intelligence for personalized study recommendations, advanced analytics to monitor student performance, and mobile application support to ensure accessibility across multiple devices. Overall, the Study Buddy Portal represents an effective solution for modern educational challenges by leveraging technology to create a collaborative and engaging learning environment.

Keywords

Collaborative learning, study portal, web application, react, node.js, express.js, mongo db, peer learning, resource sharing, student collaboration.

Download

Enterprise library book management system

Authors:Kamali, Kaviya, Raveena, Senthil Prakash*

Abstract

This paper investigates the potential of The Enterprise Library Book Management System is a comprehensive software solution developed to automate and streamline the day-to-day operations of a library. In traditional libraries, most of the work such as maintaining records of books, issuing and returning books, and managing member details is done manually, which is time- consuming and prone to errors. This system aims to eliminate these issues by providing a digital platform that ensures accuracy, efficiency, and ease of use. The system allows librarians to manage a large collection of books by storing detailed information such as book title, author name, publication, category, and availability status. It also enables easy addition, deletion, and modification of book records. Users can quickly search for books using different filters, which help in saving time and improving user experience.

Keywords

Library management system, database management system, automation book management, member management, transaction management, report generation.

Download

Real-time visualization of cyber-physical system via mobile digital twin

Authors:Gokkulakkanan, Guhan, Sivasurya, Senthil Prakash*

Abstract

This paper presents a real-time mobile-based Digital Twin system for Cyber-Physical Systems (CPS) visualization using cloud-based technologies. The proposed system replaces traditional hardware-dependent monitoring solutions with a simulation-driven architecture, enabling real-time data generation and visualization without physical sensors. A scalable backend is implemented using Firebase real-time Database to achieve low-latency data synchronization between the simulation module and the mobile application. The system ensures continuous data consistency through real-time cloud updates and an event-driven architecture, allowing seamless bidirectional communication between the virtual model and the user interface. The mobile application, developed using modern frameworks, provides interactive dashboards, graphical visualization, and system status monitoring for enhanced user understanding. The proposed approach improves accessibility, reduces infrastructure cost, and enhances scalability compared to conventional CPS monitoring systems. Experimental evaluation demonstrates efficient real-time synchronization with minimal delay, making the system suitable for academic, training, and lightweight industrial applications. In addition, the system demonstrates a flexible and extensible framework that can be adapted for various real-world applications such as smart monitoring, industrial automation, and educational simulations. The modular design allows easy integration of future enhancements, including IoT-based data acquisition and AI-driven analytics. By combining real-time visualization with a lightweight mobile interface, the proposed system provides a practical foundation for developing scalable Digital Twin solutions in resource-constrained environments.

Keywords

Digital twin, cyber-physical systems, firebase, real-time synchronization, mobile application, cloud computing.

Download

Corporate leave and HR management system

Authors:Abitha, Madhumathi, Ranjith, Senthil Prakash*

Abstract

This paper explains the development of a Corporate Leave HR management system used to manage employee leave in an organization. In many companies, leave handling is done manually using paper or email, which takes more time and may cause mistakes. By digitizing the leave process, the system cuts down on paperwork, speeds up approvals, and improves accuracy. It also makes communication between employees and HR staff clearer and more effective. It also helps in better communication between employees and HR. Overall, this system makes leave management easy and efficient.

Keywords

Leave management, HR system, employee records, automation, leave tracking.

Download

Cooked with care food delivery marketplace

Authors:Muzammil, Saravanakumar, Hariharan, Senthilkumar*, Senthil Prakash*

Abstract

In today’s fast-paced world, access to healthy and home-cooked food remains a challenge, especially for individuals who rely on convenient yet trustworthy meal options. This paper presents Cooked with Care, a community-driven food delivery marketplace designed to connect home-based chefs with customers seeking hygienic, affordable, and freshly prepared meals. Unlike traditional food delivery platforms that focus on single transactions, the proposed system emphasizes long-term service quality by continuously analysing order history, delivery performance, and customer feedback. The application is developed using the MERN stack, enabling a scalable and modular architecture that supports seamless interaction between users, chefs, and administrators. Security and reliability are ensured through role-based authentication, secure API communication, and structured data management. The system incorporates essential features such as meal browsing, menu management, order processing, payment tracking, and delivery coordination, all integrated into a user-friendly interface. By promoting transparency, trust, and consistent service monitoring, the platform not only enhances user experience but also empowers home chefs by providing them with digital visibility and business opportunities. This research demonstrates a practical and scalable approach to building a modern food delivery system that prioritizes community engagement, operational efficiency, and sustainable local entrepreneurship.

Keywords

Food delivery marketplace, MERN stack, home chefs, web application, REST API, user experience, service monitoring.

Download

Defi automated market maker (AMM) volatility simulator

Authors:Akash, Vasanthan, Vigneshwaran, Senthil Prakash*

Abstract

The rapid evolution of Decentralized Finance (DeFi) has established Automated Market Makers (AMMs) as the cornerstone of decentralized liquidity, yet these protocols remain highly susceptible to market instability. This project introduces a DeFi AMM Volatility Simulator, a robust computational framework designed to model and stress-test liquidity pool dynamics under diverse market regimes. By leveraging the Constant Product Market Maker (CPMM) invariant, $x \clot y = k$, the simulator quantifies the critical relationship between token price volatility and Impermanent Loss (IL) for liquidity providers. The system utilizes stochastic processes, such as Geometric Brownian Motion (GBM), to replicate realistic price paths, allowing users to analyse how rapid fluctuations influence slippage, pool rebalancing, and overall capital efficiency. Ultimately, this simulator serves as a vital risk assessment tool, bridging the gap between theoretical algorithmic trading and practical liquidity management by providing data-driven insights into the sustainability of decentralized exchanges during periods of extreme market turbulence.

Keywords

Automated Market Maker (AMM), liquidity pool, invariant curve, Geometric Brownian Motion (GBM).

Download

Privacy preserving clinical information extraction pipeline

Authors:Pradeep Kumar, Srimathi, Sanjeevi Vishnu, Senthil prakash*

Abstract

This paper presents a privacy-preserving Clinical Information Extraction Pipeline designed to extract structured medical knowledge from unstructured clinical narratives while ensuring robust patient data confidentiality. The proposed pipeline integrates state-of-the-art Natural Language Processing (NLP) techniques including Named Entity Recognition (NER), relation extraction, and medical concept normalization with privacy-enhancing technologies such as differential privacy, federated learning, and de-identification modules compliant with HIPAA and GDPR standards. Clinical entities such as diagnoses, medications, procedures, and laboratory findings are accurately identified and extracted without exposing personally identifiable information (PII). Experimental evaluations conducted on benchmark clinical datasets demonstrate that the pipeline achieves competitive extraction accuracy while maintaining strong privacy guarantees, with minimal utility loss. The results highlight the feasibility of deploying privacy-aware NLP systems in real-world healthcare environments, paving the way for secure secondary use of clinical data in medical research, pharmacovigilance, and clinical decision support systems.

Keywords

Clinical information extraction, privacy-preserving NLP, de-identification, differential privacy, federated learning, named entity recognition, electronic health records, HIPAA compliance.

Download

Real time chat messenger with state persistence

Authors:Mayura, Rithika, Shanmathi, Senthil Prakash*

Abstract

This work aims to design and implement a real-time messaging application that facilitates immediate interaction among users. The system uses modern web technologies to ensure messages are delivered without delay. It also includes state persistence, which allows chat history and user sessions to be saved even after refreshing or reconnecting. The application provides a user-friendly interface, efficient data handling, and reliable performance, making it suitable for modern communication needs.

Keywords

Human-centered AI, artificial intelligence, explainable AI, automation.

Download

Self-correcting code generating with iterative testing

Authors:Jothi, Maheswari, Swetha, Senthil Prakash*

Abstract

Recent advances in Large Language Models (LLMs) have significantly improved automated code generation capabilities. However, despite producing syntactically correct and logically plausible code, LLMs frequently generate programs that fail during execution, suffer from edge-case errors, or do not satisfy functional requirements. This gap between apparent correctness and actual excitability presents a major limitation for real-world software development applications. This project introduces a Self-correcting Code Generation System that applies iterative feedback loops and automated testing to dramatically improve code reliability. Instead of generating code only once, the system follows an argentic workflow in which the LLM generates both the source code and corresponding unit tests, executes them in a secure sandbox, analyzes failures, and iteratively refines the code until correctness is achieved or a retry limit is reached. Research demonstrates that such iterative refinement can improve code accuracy from approximately 40% to over 90% on benchmarks such as Human Evil.

Keywords

Self-correcting code generation, iterative testing, large language models, automated testing, error detection, software development.

Download