Semi-active suspension preview control based on visual perception

A preview control technique based on visual perception is proposed to address the body vibration problem generated by the semi-active suspension system when driving on typical discrete road surfaces. Firstly, by merging YOLOv5 target detection and stereo vision technologies, typical road surface properties and distance information are obtained in real time. Secondly, a semi-active suspension control system incorporating preview fuzzy feed-forward and Active Disturbance Rejection Control feedback is devised based on current vehicle condition information. Finally, Matlab/Simulink simulates tests with realistic discrete road conditions. The experimental findings demonstrate that the visual preview control technique may greatly increase the semi-active suspension system’s damping performance, which in turn improves the vehicle’s handling stability and ride comfort.