Artificial Intelligence Techniques for Optimizing Electric Vehicle Charging with Parallel Convolutional Neural Network: Coordinating Smart Grids and Intelligent Transportation Systems: Trends and Challenges

Abstract

The rapid growth of electric vehicles has created new challenges for modern power grids, including dynamic charging demand, grid stability, and efficient energy distribution. This review examines recent advances in artificial intelligence techniques for intelligent EV charging within smart grids and intelligent transportation systems. Deep learning models, particularly Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, and hybrid CNN–LSTM architectures, have demonstrated excellent performance in forecasting charging demand and optimizing charging schedules. These approaches effectively capture spatial and temporal patterns, reducing peak load and improving overall grid efficiency. Advanced architectures such as parallel CNNs and attention-based models further enhance prediction accuracy by integrating traffic conditions, weather information, and grid load data. Hybrid models, including ROCNN-BiLSTM, improve feature extraction and reduce uncertainty in charging behaviour, supporting more reliable decision-making. Optimization techniques such as reinforcement learning and swarm intelligence enable adaptive, real-time charging strategies that minimize operational costs and balance energy demand. Despite these advancements, challenges related to scalability, computational complexity, data heterogeneity, and cybersecurity remain. Future research should focus on lightweight, explainable, and energy-aware AI frameworks integrated with edge computing to support secure, efficient, and sustainable EV charging in next-generation smart grid ecosystems.