车辆悬挂LQR主动控制权矩阵权重参数优化

Weight parameters optimization for the weight matrices in a vehicle suspension LQR active control

针对LQR主动控制算法权矩阵Q和R的取值困难的问题,为降低权矩阵取值复杂度,分别设计权矩阵Q和R的权重参数λ和ξ。建立优化目标函数f,并分析了λ和ξ对f的影响。利用遗传算法(Genetic Algorithm,GA)的快速搜索能力,通过降维寻优,得到目标函数最优解。对GA-LQR的振动控制性能分析结果表明,GA-LQR对乘坐舒适性和操纵稳定性的改善比LQR明显,目标函数f的值更小。该LQR权矩阵权重参数优化方法简单有效,适用于车辆工程领域。

Aiming at the difficulty in determining the values of weight matrices Q and R in a LQR active control algorithm, two weight parameters λ and ξ for the weight matrices Q and R of LQR were designed respectively. An optimal objective function f , was built and the influences of λ and ξ on the f were analysed. The fast searching ability of the Genetic Algorithm（GA）, was utilized to obtain the optimization solution of the objective function through the dimension reduction optimization. The analysis results show that the vibration control performance of GA-LQR on the ride comfort and manipulation stability are more superior than the traditional LQR, and the value of the objective function is smaller. The proposed optimization method is simple and effective, and it is suitable for the application in the area of automobile engineering.