层叠CMAC补偿的并联机器人变结构控制研究

针对复杂机器人系统的不确定性,提出一种层叠小脑模型关节控制器(CMAC)神经网络同滑模变结构控制相结合的控制策略。首先利用改进CMAC学习机器人系统的不确定信息,并作为前馈补偿来确保跟踪误差的快速收敛,再通过滑模变结构控制器消除CMAC网络的逼近误差和不可重复随机干扰的影响。采用Lyapunov直接法进行控制律选取,分析表明系统可实现全局渐近稳定。在6-6并联机器人的轨线跟踪仿真试验中显示了良好的鲁棒性和精确性.

Variable Structure Control for 6-6 Parallel Manipulators Based on Cascaded CMAC

This paper proposes a novel control scheme with respect to the uncertainty of the complex robot system, which combine cascaded Cerebellar Model Articulation Controller (CMAC) with Variable Structure Control (VSC). Firstly, a improved CMAC is used to learn the uncertainty of robot system, and it is used as a feed-forward compensator, fast tracking error convergence and better learning stability are obtained through using cascaded CMAC. Then, a VSC term is used to reduce the effect of CMAC estimate error and unrepeatable disturbances. The control law is chosen based on Lyapunov direct method. Simulation of 6-6 parallel manipulator trajectory tracking shows that good stability and accuracy are obtained.