International Journal of Engineering Research and Sustainable Technologies (IJERST)

Editorial Message

2025-04-02T05:01:24+00:00

International Journal of Engineering Research and Sustainable Technologies (IJERST)

Volume 3, No.1 Online ISSN: 2584-1394

Message from Editorial Desk 25th March 2025

Dear Readers, Researchers, and Contributors,

It is with immense pride that we present the latest edition of the International Journal of Engineering Research and Sustainable Technologies (IJERST). This issue marks a significant milestone in our journey, as IJERST has officially received DOI approval from CrossRef and has become a registered member of this esteemed organization. This achievement will greatly enhance the indexing, visibility, and accessibility of our published research, ensuring that scholarly contributions reach a broader global audience.

Additionally, we are pleased to announce that IJERST has been assigned an Archival Resource Key (NAAN), further solidifying our commitment to long-term digital preservation and accessibility. This development will facilitate seamless citation, retrieval, and archival of our articles, reinforcing the credibility and impact of the research we publish.

In this edition, we continue to uphold our mission of promoting pioneering research and technological advancements in engineering and science. The featured articles showcase innovative approaches, sustainable solutions, and groundbreaking methodologies that address contemporary challenges in various engineering disciplines.

Our dedication to academic excellence and research integrity remains unwavering. As we move forward, IJERST will persist in expanding its reach by forging strategic partnerships with academic databases, research institutions, and indexing platforms. These efforts aim to further elevate the journal’s influence and ensure that our authors' contributions gain the recognition they deserve.

We extend our sincere appreciation to our esteemed authors, diligent reviewers, and dedicated editorial board members. Your support and contributions have been instrumental in making IJERST a trusted platform for high-quality research dissemination.

We encourage scholars and researchers to submit their innovative work to IJERST and become part of our growing academic community. With the advancements in indexing and digital archiving, we are committed to providing a robust platform for knowledge exchange and impactful research.

Warm regards,

On behalf of Editorial Team Dr.V.RameshBabu Managing Editor

Dean-University Journal , Dr.M.G.R. Educational and Research Institute Chennai,Tamilnadu,India

dean-univ.journals@drmgrdu.ac.in / ijerst@drmgrjournals.org

COMPARATIVE ANALYSIS OF TRANSFORMER MODELS FOR SENTIMENT CLASSIFICATION IN CODE- MIXED INDIC LANGUAGES

2025-04-02T05:24:31+00:00

Multiple language usage in a single message, or code-mixed text, has increased dramatically as a result of increased social media engagement. Because of this, activities involving Natural Language Processing (NLP), such as sentiment analysis and cyberbullying identification. Models that can effectively manage linguistic variability while retaining high accuracy are needed to address these issues. We investigate transformer-based designs that improve classification performance by utilizing knowledge transfer strategies. RoBERTa, GPT-2, XLM-RoBERTa, and IndicBERT are used in our method, which enhances classification accuracy by the transfer of sharing-private information across code-mixed and monolingual tasks. Results from experiments show that our multi-task framework surpasses single-task models with high accuracy on all datasets with:IndicBERT achieved 96.86% for Hinglish, XLM-RoBERTa achieved 96.95% for Punglish, and IndicBERT obtained 97.55% for Tanglish. In order to advance reliable NLP applications in multilingual environments, this project highlights the transformers' multi-task learning capabilities in enhancing performance on low-resource and code-mixed languages.

TWO-STAGE DC-DC ISOLATED CONVERTER FOR EV BATTERY CHARGING WITH AI CONTROLLER

2025-04-02T05:44:35+00:00

The Electric vehicles (EVs) require efficient and reliable charging systems to optimize battery performance and lifespan. Traditional converters, however, face limitations in meeting these demands, especially under dynamic charging conditions. This project presents a novel approach to EV battery charging using an AI-controlled two-stage DC-DC isolated converter, enhanced with an artificial neural network (ANN) for real-time optimization. The proposed system utilizes a two-stage converter architecture, where the first stage boosts the input voltage, and the second isolated stage provides galvanic separation, ensuring safe and efficient power transfer to the EV battery. An ANN- based controller dynamically adjusts the converter parameters based on input voltage, battery state-of-charge (SOC), and output current. This intelligent control minimizes energy losses, improves voltage regulation, and reduces thermal stress on the converter components, leading to enhanced system efficiency. Simulation results indicate that the ANN controller adapts effectively to varying input conditions, maintaining optimal charging rates while safeguarding battery health. This AI-driven approach offers a promising solution for fast, efficient, and adaptive EV battery charging, aligning with the increasing demand for smart and sustainable transportation technologies.

SMART IOT ENABLED CLEANER WITH LIDAR NAVIGATION AND AUTO-DOCKING

2025-05-14T11:16:03+00:00

The Smart IoT-Enabled Cleaner with LiDAR Route and Auto-Docking extend points to make a brilliantly, independent cleaning arrangement that leverages cutting-edge advances to upgrade productivity, client comfort and shrewd integration. This cleaner utilizes LiDAR (Light Location and Extending) innovation for exact mapping and route of its environment, empowering it to distinguish deterrents and arrange its cleaning way in real-time, guaranteeing careful coverage. With an IoT network, the cleaner can be remotely observed and controlled through versatile or web applications, permitting clients to plan cleaning sessions, track advance and get status upgrades. The auto-docking includes guarantees that the cleaner naturally returns to its charging station when its battery is moo or after completing a cleaning cycle, ensuring continuous option. By tending to the impediments of conventional cleaning gadgets, this inventive arrangement improves mechanization, optimizes route and presents smart capabilities, making it a profoundly viable and flexible device for both cutting-edge homes and commercial spaces. The integration of LiDAR route, IoT highlights and self-maintenance capabilities underscores the project’s commitment to progressing smart cleaning innovation, advertising both comfort and supportability

DISTRIBUTED INTELLIGENCE FOR CAMPUS PARKING ALLOCATION AND TRAFFIC OPTIMIZATION

2025-05-14T12:03:29+00:00

Efficient campus parking management is needed to minimise congestion, wastage of resources, and delays caused to students, staff, and visitors. The study focused on applying machine learning techniques to improve parking management systems using real-time data analysis. Such ML techniques comprise decision trees, SVMs, and neural networks developed through cameras and sensors for predicting parking availability, optimising resource allocation for multiple locations, and directing vehicles to them. The system dynamically adapts to varying traffic conditions, load levels, and periods of the day to allow further optimisation of space usage and improved mobility. Besides other advantages, predictive analytics endorses strategic infrastructure planning and improves operational efficiency. In essence, the findings of the study prove that the ML solutions for parking will enhance both campus mobility and traffic congestion during peak hours, thus encouraging sustainability.

SECURIG HEALTHCARE DATA IN THE CLOUD AGAINST INFERENCE ATTACKS

2025-05-14T12:03:27+00:00

The E-HealthCare cloud system demonstrates important possibilities for improving healthcare quality and the well-being of people. However, it has not spread, especially due to the most important concerns about security and privacy. Existing methods for storage rely on basic access control models and are susceptible to inference attacks. In this article, we have set up an inference attack for attacks on e-HealthCare cloud systems that use fine-tuned access controls. We propose a new approach using a two-layer encryption scheme that allows for secure and effective data manipulation. The first encryption layer contains marked access guidelines. This can be tailored to any data attribute where fine access control is achieved and encryption efforts are reduced simultaneously. As a rule, role attributes and access guidelines are hidden within a second encryption layer to protect privacy. The cloud is permitted to perform a variety of mathematically intensive tasks on behalf of the user without disclosing private data, in order to use the computing power of the cloud without affecting the confidentiality of the sensitive information. To achieve this, we develop a blind data access protocol using Paillier encryption. The results of comprehensive security analysis and performance assessments confirm that the proposed solution is as efficient and efficient as a secure cloud management system.