Model Predictive Control and Deep Reinforcement Learning for Energy Optimization in Battery Electric Vehicles: A Review
Keywords:
Battery Electric Vehicles, Eco-Driving, Energy Efficiency, Model Predictive Control, Deep Reinforcement Learning, Regenerative Braking, AI-Based PredictionAbstract
The rising deployment of Battery Electric Vehicles (BEVs) has accentuated the need for energy-efficient driving strategies to ensure maximum vehicle range and battery longevity. Eco-driving, which intends to optimise energy consumption without compromising vehicle performance, has become an important research focus. This review investigates the use of Model Predictive Control (MPC) and DRL techniques for energy maintenance-orientated eco-driving in BEVs. MPC provides a structured approach to predicting and controlling the dynamic behaviour of the vehicle, considering constraints on battery parameters, regenerative braking, and energy recovery. DRL, conversely, employs a data-driven learning approach to create adaptive driving policies that optimise energy usage in real scenarios. The work investigates alternatives for forecasting maximum energy consumption variations using AI techniques and improvement of vehicle and battery operation. Also, it discusses performance comparison based on predicted results against measured operational data to verify how accurate and feasible these algorithms are. Their latest developments in integrating predictive control with learning-based methods for energy efficiency optimisation are reviewed, and some challenges related to computational burden, real-time implementation, and generalisation across different driving conditions are outlined. In line with this discussion, future research would be orientated toward hybrid methods integrating MPC and DRL for adaptive energy management, as well as the exploitation of promising AI techniques for enhanced prediction and control. This overall study will give good insight into designing intelligent, energy-efficient BEV systems and provide a research direction of eco-driving optimisation.
References
N. Yang, L. Han, C. Xiang, H. Liu, and X. Li, “Model predictive control and deep reinforcement learning for energy optimisation in battery electric vehicles,” Applied Energy, vol. 335, p. 1202, 2023. [Online]. Available: https://doi.org/10.1016/j.apenergy.2023.1202.
N. Yang, L. Han, C. Xiang, H. Liu, and X. Li, “Reinforcement learning-based real-time intelligent energy management for hybrid electric vehicles,” Applied Energy, vol. 335, p. 1202, 2023. [Online]. Available: https://doi. org/10.1016/j.apenergy.2023.1202.
L. Tresca, L. Pulvirenti, and L. Rolando, “A cuttingedge energy management system for a hybrid electric vehicle relying on Soft Actor–Critic deep reinforcement learning,” Transportation Engineering, vol. 19, p. 100308, 2025. [Online]. Available: https://doi. org/10.1016/j.treng.2025.100308.
