【摘要】 将滑模控制(SMC)引入模糊小脑模型关节控制器(FCMAC),提出了一种多变量滑模FCMAC(SFCMAC).该控制器采用滑模函数转换输入信号,减小输入空间以降低网络存储空间,分粗调和微调两阶段训练网络参数,以保证控制系统的稳定性.将所提出的SFCMAC用于仿人手臂的轨迹跟踪控制,并与SMC,MLP和FCMAC比较,仿真结果表明该控制器能克服系统非线性和不确定因素的影响,控制效果良好.更多还原

【Abstract】 Sliding mode control(SMC) was used in fuzzy cerebellar model articulation controller(FCMAC).A multivariable sliding mode-based FCMAC(SFCMAC) is proposed in which sliding surface variables were specified as the input values of multivariable SFCMAC to reduce the memory spaces.SFCMAC was successively trained at the rough tuning stage and the fine tuning stage to guarantee the stability of the control scheme.The SFCMAC is applied on trajectory tracking control of the humanoid arm,and is compared with SMC,multilayer perceptron network(MLP) and FCMAC.Simulation results show that the SFCMAC is of excellent control quality.更多还原