Artificial Intelligence techniques in Vehicle-to-Grid (V2G) systems: A review, comparative study, and model evaluation
Mahmoud Elnady, Štěpán Ožana
Abstract
The increasing adoption of electric vehicles (EVs) has positioned Vehicle-to-Grid (V2G) systems as a cornerstone for the integration of renewable energy into the power grid. By enabling bidirectional energy flow, V2G systems contribute to grid stability, load balancing, and peak shaving. However, the complexity of real-time energy management in V2G requires advanced solutions. Artificial Intelligence (AI) techniques have emerged as powerful tools for optimizing various aspects of V2G, including demand prediction, scheduling, battery health monitoring, and grid stabilization. This paper reviews state-of-the-art AI methods applied to V2G systems, categorizing them into machine learning, deep learning, and optimization algorithms. A comprehensive literature review highlights the development trajectory, challenges, and achievements in applying AI to V2G systems. A comparative study evaluates these models based on accuracy, efficiency, scalability, and adaptability. Additionally, a case study on implementing an LSTM-ILP hybrid model for V2G optimization in a residential community demonstrates practical application and performance benefits. Insights into future research directions are also provided.