基于多种群遗传算法的主动座椅LQG控制器优化  

Optimization of Active Seat LQG Controllers Based on Multiple Population Genetic Algorithm

作　　者：胡启国[1] 骆艳丽 HU Qiguo;LUO Yanli(Department of Mechanic and Vehicle Engineering,Chongqing Jiaotong University,Chongqing 400074,China)

机构地区：[1]重庆交通大学机电与车辆工程学院,重庆400074

出　　处：《噪声与振动控制》2020年第2期1-6,共6页Noise and Vibration Control

基　　金：国家自然科学基金资助项目(51375519);重庆市基础科学与前沿技术研究专项重点资助项目(cstc2015jcyjBX0133)。

摘　　要：针对LQG控制器权重系数依靠经验选取的不足,提出基于多种群遗传算法的主动座椅悬架LQG控制器权重系数的多目标优化设计方法。在建立6自由度主动座椅悬架模型的基础上,利用最优控制理论设计LQG控制器;以LQG控制器权重系数为变量,以各性能指标为优化目标,建立基于权重系数的多目标优化函数;利用MPGA算法的全局搜索能力,对LQG控制器权重系数进行寻优,实现对主动座椅悬架最优控制器的优化。仿真结果表明:经MPGA算法优化权重系数后,适应度函数值比GA算法优化时降低0.9左右,验证了此优化方法的有效性。Aiming at the deficiency of the traditional LQG controllers that the weighting coefficients need to be determined by the experience of designers,a multi-objective optimization design method for the weighting coefficients of the LQG controller in active seat suspension based on multiple population genetic algorithm(MPGA)is proposed.On the basis of a 6-DOF active seat suspension model,the LQG controller is designed based on the optimal control theory.By using the weighting coefficients of the LQG controller as the variables and taking all the performance indexes as the optimization goals,the multi-objective optimization function based on the weighting coefficients is established.In virtue of the global search ability of the MPGA,the weighting coefficients of LQG controller are optimized and the optimal controller of the active seat suspension is obtained.The simulation results show that compared with the GA algorithm,the fitness function value is reduced by 0.9 when the weighting coefficient is optimized by the MPGA algorithm,which verifies the effectiveness of the proposed optimization method.

关 键 词：振动与波 多种群遗传算法 LQG控制器 权重系数 主动座椅 

分 类 号：O324[理学—一般力学与力学基础] O328[理学—力学]