Technical study of magnetorheological semi-active suspension control strategy for vehicles based on reinforcement learning

The suspension system is an important part of the car, the main function is to alleviate the road surface does not level brought about by the shake, to ensure that the vehicle driving smoothness and comfort, so as to improve the vehicle handling performance and safety, improve the vehicle to reduce the vibration performance more and more become a hot topic in the industry. Among them, semi-active suspension is a popular suspension system. This paper explains the significance and classification of the role of the suspension, and gives a relevant introduction to the magnetorheological fluid, on the basis of which it combines the domestic and international research results, looks forward to the development trend of the semi-active suspension of the vehicle, analyzes the advantages and shortcomings of four important semi-active suspension control strategies, and finally introduces the reinforcement learning and combines the reinforcement learning with the existing control strategies of the semi-active suspension, and designs a reliable reward function to enhance the suspension's damping performance by taking into account the multifaceted driving factors. The reward function to enhance the damping ability of the suspension by considering multiple driving factors, which can provide a reference for further research on the control of magnetorheological semi-active suspensions for vehicles.