两自由度并联机器人的RBF神经网络辨识滑模控制策略研究

RBF Neural Network Identifying & Sliding Mode Control of a 2-DOF Parallel Robot

针对两自由度并联机器人的轨迹跟踪问题,提出一种基于RBF神经网络辨识上界的滑模控制策略。该方案利用RBF神经网络对被控对象的不确定上界进行辨识,将所得的上界值适时送到滑模控制器,既发挥了RBF神经网络具有逼近任意函数的优点,又保留了滑模变结构控制的快速性和鲁棒性,达到了理想的控制效果。

A control law that RBF neural network identifying upper bound of uncertain value and sliding mode control strategy for 2-DOF parallel robot trajectory tracking was proposed.RBF neural network was used to identified the upper bound of uncertain value.In sliding mode controller,the ranges of parameters were changed to fit for the servo mechanism.The control strategy not only keeps the identifying function of RBF neural network,but also retains the high-speed and robustness of sliding mode control.Ideal control effect can be achieved.