Edge-Intelligent Electric Vehicle Charging Coordination for Grid Load Balancing and Renewable Integration

The increasing penetration of electric vehicles (EVs) into modern power systems introduces new challenges for grid load balancing and renewable energy integration. Traditional EV charging infrastructures rely on centralized cloud-based control, resulting in high communication latency, reduced scalability, and limited responsiveness to dynamic grid conditions. This study proposes an edge-intelligent framework for real-time EV charging coordination, leveraging IoT-enabled sensing and localized AI inference to optimize grid load distribution and enhance renewable energy utilization. The framework deploys lightweight deep learning models (CNN + LSTM, XGBoost, Random Forest) on edge devices such as Jetson Orin Nano and Raspberry Pi 5. The application enables behavior-aware scheduling and dynamic pricing adjustments directly at charging stations. A large-scale U.S. Department of Energy dataset is used to train and validate the models. Experimental results demonstrate that the system improves charging station utilization by 27.6%, reduces peak grid load by 24.5%, and lowers user charging costs by 29.8% while maintaining high prediction accuracy (R² > 0.92, MAE = 0.0182 kWh). Furthermore, the decentralized architecture reduces communication overhead by approximately 80%, supporting faster decision-making and improved grid responsiveness. By intelligently coordinating EV charging based on real-time user behavior, grid conditions, and renewable availability, the proposed approach offers a scalable and practical solution for sustainable smart grid operation.