Claim Missing Document
Check
Articles

Found 7 Documents
Search

Novel Battery Management with Fuzzy Tuned Low Voltage Chopper and Machine Learning Controlled Drive for Electric Vehicle Battery Management: A Pathway Towards SDG P, Vinoth Kumar; S, Priya; D, Gunapriya; Batumalay, M
Journal of Applied Data Sciences Vol 5, No 3: SEPTEMBER 2024
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v5i3.236

Abstract

Electric vehicles have a significant impact on the SDGs, specifically climate action, affordable and clean energy, and responsible consumption and production patterns. The present work focuses on a battery management system to effectively utilize the power from the battery to drive the brushless DC motor (BLDC) by tuning the low-voltage buck boost converter as a chopper circuit with fuzzy. The photovoltaic system acts as an additional source to charge the battery when the battery is not connected to the load, and at running conditions, fuzzy logic control enhances efficiency and provides smooth, adaptive control under varying load conditions. Also, the machine learning technique is used for drive control and automation operations. The energy in the BLDC is regulated by managing the voltage and current in a photovoltaic-powered low-voltage chopper by tuning the proportional integral derivative (PID) controller for an ideal balance between reliability and a quicker reaction. The K- Nearest Neighbour (KNN) machine learning algorithm, due to its simplicity and effectiveness in classification, ensures the enhanced reliability and efficiency of the BLDC motor system with commutation and speed control. When fuzzy and the KNN machine learning algorithm are used, the development of systems for control and automation is expedited. The work also shows the results of a study that compared the interoperability of proportionate machine learning and fuzzy controlling algorithms developed with MATLAB. In order to do a critical analysis of the data, the results are compared with the graphs. The integration of the Internet of Things (IoT) and cloud technology with the use of KNN for BLDC motor control can enhance system proficiency with monitoring and display of the observed voltage, current values of the motor, sensorless control, fault diagnosis, and predictive maintenance. The work is also connected with the SDG and impacts due to the efficient operation of electric vehicles.
Lower Limb Prosthetics using Optimized Deep Learning Model – a Pathway Towards SDG Good Health and Well Being S, Prakash; M, Jeyasudha; S, Priya; Batumalay, M
Journal of Applied Data Sciences Vol 5, No 3: SEPTEMBER 2024
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v5i3.239

Abstract

This research article aimed at revolutionizing prosthetic leg technologies to enhance accessibility, affordability, and environmental sustainability. With a focus on addressing the diverse needs of amputees globally, the program integrates principles of eco-design, community engagement, technological innovation, and policy advocacy to foster inclusive and resilient societies which leads to the attainment of Sustainable Development Goal (SDG) Good Health and Well Being. Lower Limb Prosthetics of Activity Recognition is an innovative field combining prosthetic technology and activity recognition systems. The challenge of activity recognition in lower limb prosthetics to optimize the performance and responsiveness of mock limbs. In this work, the problem is overcome by using the Optimized deep learning technique, which improves activity recognition in lower limb prosthetics. The proposed methodology consists of (1) Pre-processing (2) Feature extraction (3) Feature classification. The collected images are pre-processed via improved wavelet demonizing and Empirical mode decomposition. From pre-processed data, the features are extracted using an improved sliding window method. The obtained extracted features are moved on to the Feature classification process. The classification process is done by the Optimized Long short- term memory. They are designed to better capture dependencies and patterns in sequential data, which makes them highly effective for tasks involving time series, natural language processing and other sequential data problems. Optimization can be done by proper data preprocessing and tuning the data from data extraction. The weight of the LSTM model is optimized to improve the performance of this model by the improved Black Window Optimization Algorithm. The main contributions of the paper are to obtain the best classification accuracy, an optimized LSTM model is introduced in this paper, and the weight of the LSTM model is enhanced by the improved Black Window Optimization algorithm. It improves the performance of the proposed system.
Convolutional Neural Network for Battery System Monitoring and SOC Estimation for Ev Applications to Achieve Sustainability S, Priya; P, Vinoth Kumar; V, Sridevi; Batumalay, M; D, Gunapriya
Journal of Applied Data Sciences Vol 5, No 4: DECEMBER 2024
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v5i4.237

Abstract

The necessity to develop an alternative energy source to handle the looming energy crisis has arisen due to the recent rise in energy consumption. This is most likely to happen with grid-synchronized electric vehicles (EVs), since vehicle-to-grid (V2G) technology is one of the smart grid's technological advancements that permits energy exchange between EVs and the grid. The merging of EVs with the grid influences the whole electricity system and is susceptible to imbalances in supply and demand, frequency, and voltage. The proposed work focuses on effective and smart control of the Single Ended Primary Inductance Converter (SEPIC) converter with efficient control techniques employed for battery management systems for electric vehicle charging. PI oversees controlling the converter. The battery's calculated state of charge (SOC) is used to make a paradigm-shifting sequence for the converter with workable optimization strategies to lower imbalance issues when EVs are connected to the grid. This leads to the achievement of sustainable development goals (SDGs). Purpose. When the SEPIC converter is connected to a photovoltaic source, it needs to be analyzed in terms of how it switches operations. The source also needs to be used efficiently when it is connected to a battery. Monitoring and SOC estimation of the battery need to be efficiently performed with a quicker response for EV applications. Methods. Convolutional neural networks (CNN) were used to solve the issue; these networks considerably enhance response times and boost system reliability overall. Results. The system operates on the principle that when the battery level is less than 60%, the battery is charged through buck operation, and it is discharged through the boost mode when the SOC exceeds 60%. When linked to the grid, the PI controller regulates both power and practical value. The proposed system demonstrates how battery management-based CNN and SEPIC can switch at high speeds. The system's research directions were established for the results' later application to experimental samples for energy efficiency and process innovation.
Quantum-Enabled Secure and Energy-Efficient Protocols for Smart Grid Communication Systems Al-Qaraghuli, Sara; Jameel, Sarah Haitham; Majid, Mohammed Nouri; Jawad, Aqeel Mahmood; Saeed, Matai Nagi; Batumalay, M
International Journal of Engineering, Science and Information Technology Vol 5, No 4 (2025)
Publisher : Malikussaleh University, Aceh, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52088/ijesty.v5i4.1193

Abstract

The development and evolution of the smart grid into complex, cyber-physical energy systems make it essential to secure?communication among the distributed components. The rise of quantum computing has made it even more pressing to develop protocols that?are secure outside the limitations of classical cryptosystems. In this paper, it proposes a quantum-assisting secure communication scheme (QASCP) to?boost the security and energy for smart grid communication systems. The proposed protocol combines quantum key distribution with lightweight entropy-based mutual authentication and dynamic session management. It is designed to defend grid assets such as control centers, smart meters, and distribute energy?resources from sophisticated adversarial models, including quantum-capable threats. The approach consists of system level simulation utilizing a?co-simulation framework customized for quantum smart grid communication. The performance of this scheme was compared?against classical and PQ lattice-based schemes in terms of the authentication latency, energy consumption, entropy preservation, and scalability to handle the load and delay effects, under the assumptions of different loading and delay scenarios. Simulation results show that QASCP is able to reduce the energy consumption and authenticity latency, simultaneously it keeps the high throughput and leaves strong entropy?under attack scenarios. The protocol is also shown to?remain robust with varying quantum bit error rates as well as having a smaller memory footprint on popular network topologies. The results provide evidence for the practical integration of?quantum-secure communication in smart grid architectures. By addressing security and performance simultaneously, the protocol?provides a path to future-proof energy networks which can support dependable operations in a quantum-enhanced environment. This could be future enhance for energy efficiency.
Data-Driven Evaluation of a Gamified Breath-Holding Training Application to Improve CT Scan Quality and Reduce Patient Anxiety P, Vinoth Kumar; M, Ganga; K, Vijayakumar; K, Umamaheswari; Devarajan, Gunapriya; Batumalay, M
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.804

Abstract

This study presents the development and evaluation of Breathe Well, an innovative three-tiered Graphical User Interface (GUI) application designed to address motion-induced step artifacts and patient anxiety during Computed Tomography (CT) scans. The core idea of the application is to combine relaxation techniques, guided breathing exercises, and gamified training modules within a single interactive platform that allows patients to practice breath-holding and anxiety control prior to scanning. The objective is to enhance patient cooperation, reduce involuntary movement, and improve overall image quality while minimizing the time healthcare staff spend on manual breath-hold instruction. The study involved a comparative analysis between a control group and an intervention group trained using the Breathe Well system. Quantitative results demonstrated a significant improvement in imaging outcomes, with the mean artifact score decreasing from 3.1 ± 0.8 in the control group to 2.1 ± 0.7 in the intervention group (p 0.01). Psychological assessment using the State-Trait Anxiety Inventory (STAI) revealed a marked reduction in patient anxiety, with mean scores declining from 48.6 ± 6.4 before training to 38.2 ± 5.8 after using the application (p 0.01). Qualitative feedback further confirmed increased patient confidence, comfort, and comprehension of CT procedures. The findings indicate that integrating gamified digital interventions into pre-scan preparation significantly improves both patient experience and diagnostic precision. The novelty of this research lies in the creation of a self-guided, multi-level digital platform that bridges behavioral training and imaging technology, offering a scalable, patient-centered solution for modern radiology workflows.
TF-EffBiGRU-AttNet: A Novel Deep Learning Framework for Spatio-Temporal Energy Demand Forecasting in Electric Vehicle Charging Networks Prakash, S; Aruna Mary, S; Sudhagar, G; Batumalay, M
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.805

Abstract

Electric Vehicle Charging Stations (EVCS) are key enablers of sustainable transportation, yet accurate forecasting of their energy demand remains challenging due to complex spatial-temporal variability. This study introduces a novel hybrid deep learning framework, Two-Fold EfficientNetV2 BiGRU with Attention (TF-EffBiGRU-AttNet), optimized using the Self-Adaptive Hippopotamus Optimization Algorithm (SA-HOA), to enhance prediction accuracy and computational efficiency in EVCS energy demand forecasting. The main objective is to integrate multi-scale spatial learning, bidirectional temporal modeling, and adaptive feature prioritization within a single architecture capable of robust and interpretable forecasting. The model’s novelty lies in its dual-fold spatial feature extraction using EfficientNetV2 and dynamic optimization through SA-HOA, which adaptively balances exploration and exploitation during training. Experimental validation on two real-world datasets from Palo Alto and Perth demonstrates that the proposed model consistently outperforms state-of-the-art baselines. For the 7-1 forecasting task, TF-EffBiGRU-AttNet achieved the lowest MAE of 0.012 and RMSE of 0.051 for Palo Alto, and MAE of 0.029 with RMSE of 0.12 for Perth. For the 30-7 task, it achieved MAE of 0.0332, RMSE of 0.1654, and MAPE of 0.20% on Palo Alto, and MAE of 0.0235, RMSE of 0.0824, and MAPE of 0.37% on Perth, outperforming Bi-LSTM and EfficientNet by over 60% in RMSE reduction. Moreover, SA-HOA improved optimization efficiency with a best fitness value of 0.0003 and reduced convergence time to 1.2 seconds, surpassing PSO, GWO, and HOA. These results highlight the framework’s ability to capture spatial-seasonal and nonlinear dependencies while maintaining low computational overhead. The findings confirm the model’s potential as a robust, adaptive, and scalable solution for intelligent EV energy demand forecasting, supporting smart grid planning and sustainable energy management.
Leveraging Generative AI in Vehicles for Enhanced Driver Safety and Advanced Communication Systems P, Vinoth Kumar; T, Sri Anadha Ganesh; Batumalay, M; Kumar, S N; Devarajan, Gunapriya; K, Bhuvaneshwari; T, Kesavan; S, Lakshmi Praba; S, Nandhanaa K
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.809

Abstract

This paper proposes an integrated artificial intelligence–based driver assistance system for electric vehicles (EVs) that combines computer vision–based drowsiness detection with a generative artificial intelligence (GenAI)–driven conversational interaction framework to enhance driver safety and human–vehicle interaction. The primary objective of this work is to reduce fatigue-related driving risks while enabling natural, hands-free, and context-aware communication between the driver and the vehicle. The core idea is to tightly couple real-time driver state monitoring with intelligent conversational feedback, allowing safety alerts and voice interactions to adapt dynamically to the driver’s condition. Driver drowsiness is detected using non-intrusive visual indicators, namely eye closure duration and blink rate, extracted from an in-vehicle camera. A drowsy state is identified when eye closure exceeds 10 s or when the blink rate exceeds 6 blinks within a 6 s interval. Upon detection, the system generates multi-modal alerts consisting of audio warnings and vibration feedback, while a GenAI-based natural language processing module provides real-time, hands-free voice interaction. Experimental evaluation was conducted on an ESP32-based embedded prototype across five predefined driving scenarios representing normal and fatigued conditions. The results show stable face and eye detection under normal driving and achieved 100% correct alert triggering in all drowsiness-related cases (3 out of 5 scenarios), with zero false positives observed during non-drowsy conditions (2 out of 5 scenarios). The system demonstrated consistent real-time response and reliable alert activation under fatigue conditions. The main contribution and novelty of this research lie in the real-time integration of generative AI–driven conversational intelligence with embedded computer vision–based drowsiness detection within a unified, resource-constrained platform, which is rarely addressed jointly in existing systems. Overall, the proposed framework provides a practical, scalable, and human-centered solution for intelligent driver assistance in semi-autonomous and future autonomous EV environments.