用T-S型模糊神经网络的机械手轨迹跟踪自适应控制

对于常见的将 CMAC神经网络前馈控制器和常规反馈控制器相结合的机械手轨迹跟踪控制方案 ,它的控制性能同时受神经网络前馈控制器学习能力和反馈控制器控制精度的制约。该文提出的采用 T- S型模糊神经网络的机械手轨迹跟踪自适应控制方案充分利用了 T- S模糊模型的特点和优点 ,以一种基于简化的 T- S型的模糊神经网络作为前馈控制器 ,同时反馈控制器也采用 T- S型模糊神经网络实现。针对三自由度机械手轨迹跟踪问题的仿真实验表明 ,采用 T- S型模糊神经网络的机械手轨迹跟踪自适应控制方案是可行的和有效的

Adaptive control of an industrial manipulator trajectory tracking system using a T-S model based fuzzy neural network

Performance of the familiar control scheme for robot manipulator trajectory tracking which uses a CMAC (cerebellar model articulation controller) neural network feed forward controller and a conventional feed backward controller is restricted by both the learning capability of the neural network feed forward controller and the control precision of the feed backward controller. The control scheme presented in this paper uses a simplified T S model based on fuzzy neural network as feed forward controller and a T S model based on fuzzy neural network as feed backward controller; hence, it can take advantage of the characteristics of the T S fuzzy model. Simulation results of a 3 DOF industrial manipulator trajectory tracking system are presented to show the proposed control scheme is practical and effective.