Research on eco-driving strategy of connected hybrid electric vehicles for traffic influence factors

Abstract

Abstract ：Eco-driving technology is the core means to improve the energy economy and traffic efficiency of connected hybrid electric vehicles. Existing studies mostly focus on the coupling of speed planning and energy management in car-following scenarios, and there are still deficiencies in relevant research on fuel cell hybrid electric vehicles under complex traffic environments. This study proposes a hierarchical eco-driving strategy that integrates speed planning and energy management to achieve multi-objective optimization. Firstly, a driving environment model is constructed based on real traffic scenarios, and the passing reference speed is solved by comprehensively considering road condition information such as signalized crossroadss. Then, a multi-objective cost function is constructed based on the model predictive control framework to realize vehicle speed optimization. The speed condition is transmitted to the Energy Management System (EMS). At the energy management level, an improved Adaptive Noise TD3 (AN-TD3) algorithm integrating an adaptive noise mechanism is proposed, and a comprehensive reward function considering hydrogen consumption cost and component aging is designed to realize the optimal power distribution between the fuel cell and the lithium battery. The joint simulation verification based on MATLAB/Simulink and SUMO shows that the proposed strategy significantly shortens the travel time, reduces the number of vehicle starts and stops and energy consumption. The AN-TD3 algorithm has a faster convergence speed than the traditional Deep Reinforcement Learning (DRL) algorithm, the hydrogen consumption is close to the global optimal of dynamic programming, which effectively suppresses the aging of fuel cells and lithium batteries, and improves the economy and robustness of the whole vehicle operation.