Title–DIVING SAFETY MONITORING SYSTEM: A DEEP LEARNING BASED DRIVER DROWSINESS DETECTION USING YOLOV8

Author– Sudharani Bangarimath, Dr. Sangram Patil and Jambukeshwar Pujari

Abstract

Traffic accidents kill and injure most people. Traffic accident injuries kill a million people annually, according to the WHO. Lack of sleep, rest, or fatigue can cause drivers to fall asleep, endangering themselves and others. This research focuses on driver sleepiness detection and response. The artificial intelligence represents advanced technology, offering solutions to create a majority of vision-based applications. In this research, a driver drowsiness detection system has been developed utilizing the deep learning Yolo v8 algorithm. An image dataset comprising two classes, namely drowsiness and non-drowsiness, is utilized for training with Yolov8. Image classification is conducted based on the condition of the driver’s eyes, specifically whether they are open or closed. The validation of the trained model indicates an accuracy of 96.85% and an F1-Score of 97.5%.

Index Term–Deep Learning, Driver Drowsiness, AI. WHO, Vision-based Applications.

DOI- 10.30696/IJEEA.XIII.II.2025.01-12

Reference to this paper should be made as follows:  Sudharani Bangarimath, Dr. Sangram Patil and Jambukeshwar Pujari, (2025), “Diving Safety Monitoring System: A Deep Learning Based Driver Drowsiness Detection Using Yolov8” Int. J. Electronics Engineering and Applications, Vol. 13, No. 2, pp. 01-12.

Title–CLIMATE PATTERN PREDICTION USING HYBRID SPATIOTEMPORAL MODELS

Author– I Wayan budi sentana ,Ram Kumar Solanki, and Chetan Chauhan

Abstract

Accurate climate pattern prediction is crucial for disaster preparedness, resource management, and policy planning. Hybrid spatiotemporal models, combining spatial dependencies and temporal dynamics, offer improved predictive capabilities. This paper explores hybrid approaches integrating convolutional neural networks (CNNs) for spatial feature extraction with recurrent neural networks (RNNs) or transformers for temporal modeling. Experiments on historical climate datasets demonstrate enhanced accuracy in predicting temperature, precipitation, and extreme weather events. Challenges include data sparsity, model interpretability, and computational efficiency. Future research aims to integrate multi-source data, improve model generalization, and deploy scalable predictive frameworks for real-time climate monitoring.

Index Term–climate prediction, hybrid spatiotemporal models, convolutional neural networks (CNNs), recurrent neural networks (RNNs), transformers, temperature forecasting, precipitation modeling, extreme weather prediction, environmental data analytics, real-time climate monitoring.

DOI- 10.30696/IJEEA.XIII.II.2025.01-12

Reference to this paper should be made as follows: I Wayan Budi Sentana, Ram Kumar Solanki, and Chetan Chauhan (2025), “Climate Pattern Prediction Using Hybrid Spatiotemporal Models” Int. J. Electronics Engineering and Applications, Vol. 7, No. 2, pp. 13-28.

Title–AUGMENTING AR/VR EXPERIENCES USING MULTIMODAL TRANSFORMER MODELS

Author– Ram Kumar Solanki, Abhishek M Dhore, and Amit R Gadekar

Abstract

Augmented Reality (AR) and Virtual Reality (VR) applications benefit from multimodal data integration to create immersive user experiences. Transformer models capable of processing and fusing visual, auditory, and textual information offer promising avenues for enhancing AR/VR systems. This paper explores multimodal transformer architectures for real-time AR/VR augmentation, leveraging cross-modal attention, contextual embedding, and synchronized data streams. Experiments demonstrate improvements in environment understanding, interactive responsiveness, and user engagement. Challenges related to computational efficiency, latency, and multimodal alignment are discussed. Future work focuses on scalable architectures, adaptive fusion strategies, and personalized AR/VR experiences.

Index Term–augmented reality (AR), virtual reality (VR), multimodal transformers, cross-modal attention, data fusion, real-time interaction, contextual embedding, user engagement, adaptive architectures, immersive computing.

DOI- 10.30696/IJEEA.XIII.II.2025.01-12

Reference to this paper should be made as follows: Ram Kumar Solanki, Abhishek M Dhore, and Amit R Gadekar (2025), “Augmenting AR/VR Experiences Using Multimodal Transformer Models” Int. J. Electronics Engineering and Applications, Vol. 7, No. 2, pp. 29-44.

Title–AI-DRIVEN DISASTER RESPONSE OPTIMIZATION USING REAL-TIME SATELLITE DATA

Author– I Wayan Budi Sentana and Rajesh Kumar Tiwari

Abstract

Disaster response requires rapid and informed decision-making to minimize human and economic losses. AI-driven approaches leveraging real-time satellite data provide actionable insights for disaster monitoring, resource allocation, and response optimization. This paper explores AI models for processing satellite imagery and geospatial data to predict disaster impact, prioritize critical areas, and optimize rescue operations. Techniques such as convolutional neural networks (CNNs), spatiotemporal modeling, and reinforcement learning are applied to real-time data streams. Experiments demonstrate improved response efficiency, predictive accuracy, and resource deployment. Challenges include data latency, cloud coverage, model interpretability, and integration with emergency management systems. Future directions focus on multi-source data fusion, scalable AI pipelines, and decision support frameworks for timely disaster response.

Index Term– Artificial Intelligence, Disaster Response, Satellite Imagery, Real-Time Data Processing, Convolutional Neural Networks (Cnns), Spatiotemporal Modeling, Reinforcement Learning, Geospatial Analytics, Emergency Management, Resource Optimization.

DOI- 10.30696/IJEEA.XIII.II.2025.01-12

Reference to this paper should be made as follows:  I Wayan Budi Sentana and Rajesh Kumar Tiwari, (2025), “AI-Driven Disaster Response Optimization Using Real-Time Satellite Data” Int. J. Electronics Engineering and Applications, Vol. 7, No. 2, pp. 45-56.

Title–REAL-TIME PREDICTIVE MODELLING FOR DISEASE OUTBREAKS USING EDGE AI

Author– I Wayan budi sentana , and Rajesh Kumar Tiwari

Abstract

The rising frequency of infectious disease outbreaks calls for faster, smarter, and decentralized response systems. This study presents a Real-Time Predictive Modelling Framework using Edge AI for early detection and control of outbreaks. By deploying AI models on edge devices—such as healthcare IoT gateways and local monitoring units—the system enables instant data analysis and anomaly detection near the data source. It integrates spatio-temporal analytics, federated learning for privacy-preserving collaboration, and adaptive deep learning that evolves with emerging infection trends. Experimental results show a 70% reduction in decision latency and a 25% increase in prediction accuracy compared to cloud-based models. The framework offers scalability, rapid response, and secure data handling, positioning Edge AI as a key technology for proactive epidemic management. control.

Index Term–Edge Artificial Intelligence (Edge AI), Real-Time Predictive Modelling, Disease Outbreak Detection, Federated Learning, Internet of Things (IoT), Public Health Surveillance, Anomaly Detection, Temporal-Spatial Data Analytics

DOI- 10.30696/IJEEA.XIII.II.2025.01-12

Reference to this paper should be made as follows: I Wayan Budi Sentana, and Rajesh Kumar Tiwari (2025), “Climate Pattern Prediction Using Hybrid Spatiotemporal Models” Int. J. Electronics Engineering and Applications, Vol. 13, No. 2, pp. 57-71.

Title–QUANTUM CRYPTOGRAPHY FOR SECURING MEDICAL DATA AT THE EDGE

Author– Ganesh R. Pathak and Abhishek  M. Dhore

Abstract

The exponential growth of Internet of Medical Things (IoMT) devices and edge computing in healthcare has revolutionized real-time diagnostics and patient monitoring but simultaneously introduced severe security and privacy challenges. Conventional cryptographic schemes, while effective against classical attacks, are increasingly vulnerable to the computational power of emerging quantum computers, which can break widely used algorithms such as RSA and ECC. To address this imminent threat, this paper proposes a Quantum Cryptography–enabled Edge Security Framework (QCESF) that integrates Quantum Key Distribution (QKD) with Edge AI-based access control mechanisms for end-to-end protection of medical data. The proposed model ensures quantum-resilient encryption, secure data transmission, and dynamic authentication across distributed healthcare nodes without compromising latency or system efficiency. By employing BB84 and E91 quantum protocols within an edge–cloud architecture, the framework enables unbreakable key exchange and detection of eavesdropping attempts in real time. Experimental analysis and simulations demonstrate that QCESF achieves a 40–60% improvement in key exchange security, 30% reduction in latency, and enhanced resistance to quantum-based attacks compared to conventional hybrid cryptographic methods. The integration of quantum communication with edge intelligence not only strengthens data confidentiality and integrity but also establishes a scalable, future-proof foundation for next-generation secure healthcare ecosystems.

Index Term–Quantum Cryptography, Quantum Key Distribution (QKD), Edge Computing, Internet of Medical Things (IoMT), Secure Data Transmission, Quantum-Resilient Encryption, BB84 Protocol, E91 Protocol, Edge AI Security, Privacy-Preserving Healthcare Systems.

DOI- 10.30696/IJEEA.XIII.II.2025.01-12

Reference to this paper should be made as follows: Ganesh R. Pathak and Abhishek M. Dhore (2025), “Climate Pattern Prediction Using Hybrid Spatiotemporal Models” Int. J. Electronics Engineering and Applications, Vol. 13, No. 2, pp. 72-86.