非确定因素下汽车半主动悬架的智能控制

以含磁流变阻尼器的1/4 车辆非线性半主动悬架模型为研究对象,在充分考虑该非线性系统未建模动态的基础上提出了具体的神经网络与滑模变结构控制相结合的智能控制策略,有效抑制了悬架系统的振动,使车辆行驶的平顺性与舒适性得以提高.应用神经网络的在线学习能力对非线性动力学模型的不确定部分及外界未知扰动进行了神经网络估计,确定了未知函数的上确界,构造了控制系统的滑模变量并且合理设计神经网络的自适应规律使状态变量快速接近原点.通过稳定性分析证明了此种控制方法是全局渐近收敛的,并且对未建模动态具有强鲁棒性.数值仿真结果验证了该种控制方法的有效性,得到了阻尼器两端控制电压的变化规律.

INTEllIGENT CONTROL OF AUTOMOTIVE SEMI-ACTIVE SUSPENSION WITH UNCERTAIN FACTORS

Taking a nonlinear semi-active magnetal-rehological suspension model of a quarter car as a research objective, a concrete intelligent control strategy that associates neural network technique with a sliding mode-variable-structure-control method is presented on the basis of a nonlinear system in an unknown dynamic status. It suppresses the vibration of a suspension system effectively and can improve the ride performance and ride comfort. The unknown part of the nonlinear dynamic model and outside disturbance are estimated using neural network theory which has on-line learning ability and the upper boundary of the unknown part is determained.. The state variables approach the original point along the constructive sliding mode by designing the reasonable self adaptive law of neural network. The stability analysis conforms that this control method is convergent and robust to the unknown dynamic part in global. Simulation results make sure the validity of the control method, and the control voltage of the damper is obtained.