基于遗传算法权值优化的汽车悬架最优模糊控制

Controlling vehicle suspension via hybrid scheme of fuzzy controller and optimal controller with adjusted weightings based on genetic algorithm

结合车辆非线性主动悬架系统数学模型,首先设计出悬架系统的线性二次型整定(LQR)最优控制器,其中控制器的加权系数利用遗传算法优化搜索获得,然后设计悬架系统的模糊控制器,进而开展LQR最优控制、模糊控制以及两者并联结合的复合控制方法对比研究.利用遗传算法对加权系数进行搜索优化,可有效解决传统LQR控制器加权系数不易确定的问题.该复合控制方法可获得相比单一控制方法更优的悬架系统控制效果,在车辆不同行驶工况条件下,能进一步降低车身垂直振动加速度、悬架动挠度和轮胎形变,明显提高了汽车的行驶平顺性和操纵稳定性.

From a nonlinear automobile model of active suspension system, a linear quadratic regulator （LQR） based on optimal control theory is designed firstly, in which the weighting coefficients are obtained by genetic algorithm. Secondly, a fuzzy controller with two inputs and scheme including optimal controller and fuzzy controller is one output is designed. Consequently, a hybrid control studied through the comparison for the single methods. In our method, the genetic algorithm can overcome the difficulties in selecting the weighting coefficients subjectively and improve the efficiency of design LQR controller. Meanwhile, the proposed hybrid method can further decrease the vertical vibration acceleration of body, the dynamical deflection of suspension and tire deformation, and improve the riding comfort and handling stability.