Development of a Robust Active Suspension Preview Control Method Based on Visual Perception
Xiang Zhang, Zhijun Fu, Minghui Cui, Dengfeng Zhao · 2024
Although robust active suspension control problem has been well studied, solving robust control via unknown road surface recognition (such as speed bump detection) has not been fully solved. This paper develops a$\mathrm{H}\infty$preview control method to complete the robust active control design for suspension systems, which uses the ideas of visual perception proposed for preview control. First, a novel detection approach of typical working conditions of urban road (speed bump pavement) based on YOLOV5 is designed using camera, along with image fitting technology. Then, an augmented system is constructed using the recognized speed bump information, so as to reformulate the active suspension control into a modified robust regulation problem. On this basis, the linear matrix inequalities (LMI) technique is introduced to develop the$\mathrm{H}\infty$preview control of the suspension system. Finally, simulation results are given to verify the effectiveness of the proposed method.