Vertical vibration control to the in-wheel-motor electric vehicles with static eccentricity based on a reinforcement learning method

Dawei Zhang,Chen Zhong,Shuizhou Liu,Peijuan Xu,Yiyang Tian

DOI: https://doi.org/10.1177/10775463241264047

2024-07-26

JVC/Journal of Vibration and Control

Abstract:Journal of Vibration and Control, Ahead of Print. The in-wheel-motor electric vehicle (IWM-EV) is hailed as the epitome of driving ingenuity within the realm of electric vehicles. Nonetheless, the intricate nature of its components, compounded by the intricate interplay of multiple force fields, poses a significant detriment to ride comfort. In the present study, an IWM-EV driven by a permanent magnet synchronous motor was employed as a representative case study. Initially, the calculations were conducted to determine the unbalanced magnetic force (UMF) in the presence of static eccentricity of the stator. Subsequently, the characteristics of UMF across different ratios of static eccentricity as well as different velocities in the time domains were analyzed. Furthermore, the road-electromagnetic-mechanical model was developed to investigate the influence of UMF on the vertical vibration of IWM-EV under static eccentricity, comparing it against the scenario devoid of UMF. Finally, a reinforcement learning control approach was adopted to regulate the active suspension system, comparing its efficacy with that of passive suspension and semi-active suspension (specifically, skyhook control). Through extensive simulations, the results demonstrated that the reinforcement learning control strategy derived from the road-electromagnetic-mechanical model outperforms the other two control strategies, exhibiting commendable resilience and adaptability across diverse road surfaces and velocities. This study unveiled the potential of RL methods in enhancing riding comfort through active suspension control.