A sliding mode semi-active control for suspension based on neural network

On basis of introducing the essential characteristics of electro-rheological fluid, this paper presents semi-active control for suspension based on ER damper and establishes a quarter car dynamic model. Its state space equation is derived. According to sliding mode control theory, the parameters of the switching surface are determined by the method of pole assignment. A proportion switching method and an equal near rate is used to amend the dynamic quality of sliding mode motion in the proposed controller. Meanwhile, the RBF neural network arithmetic is introduced to optimize the sliding mode control result. Then use Matlab/Simulink software to set up the model including passive and semi-active control and simulate. The results show that the semi-active control is more effective than the passive control, the body acceleration has declined by 20% and ride quality has improved observably.