Navigating the electric vehicle landscape: A review of smart charging techniques, growth trends, global adoption strategies, and future directions

The integration of electric vehicles (EVs) into existing power grids presents challenges and opportunities for optimizing EV charging and discharging. This study provides a comprehensive review of conventional methods and machine learning approaches, with a particular focus on the reinforcement learning (RL) technique, which has demonstrated superior adaptability and efficiency compared to alternative methods. The application of RL in EV charging and discharging is explored, highlighting the key states, algorithms, constraints, agents, and objectives considered in optimization problems. While previous research has predominantly emphasized EV charging and discharging strategies, this review addresses 2 critical gaps in the literature. First, it examines the integration of renewable energy sources into power grids to meet the growing demand for EVs, without increasing fossil energy consumption for power generation. Second, it investigates the unique barriers hindering EV adoption in developing countries, contrasting these challenges with the successful strategies implemented in developed nations. Strategies, such as prioritizing smaller EV models and implementing battery-swapping systems, have been proposed to overcome some of these challenges. Finally, potential research areas are discussed, including the need to scale RL models for large-scale deployments, integrate real-world power and transportation data into optimization frameworks, and develop tailored approaches for equitable EV adoption.