SCARA型机械臂的智能神经网络控制研究

在传统机械臂动态控制基础上,将RBF神经网络与模糊逻辑控制相结合,提出一种新的智能控制方法——RBF模糊神经网络控制方法。该方法使用两个子网络分别对关节1和关节2实施控制,最后通过协调级网络来消除和减小两个关节之间的耦合作用。以仿真实验结果说明该网络结构简单明确,计算更加有效,并通过适时地修正网络参数,在线调整了模糊隶属函数的中心值和宽度,所设计的控制器能快速、稳定地跟踪到期望轨迹。

Research on the Control of Intelligent Neural Network of SCARA- robot Manipulator

Based on the traditional dynamic control by robot manipulator,and connecting the fuzzy logical control with RBF neural network,this paper introduces a new intelligent control method,RBF fuzzy neural network control,which provides two parallel sub-networks to control joint 1 and joint 2 of the robot manipulator separately,and eliminates or reduces the couplings between two joints through the protocol-level network,and by using the simulation results,illustrates that the network has simple and explicit structure,higher effective calculation,self-adjusting of the network parameters,and can easily make on-line adjustment of the central value and width of the network membership function,and the controller designed can achieve good tracking performance rapidly and stably.