RBF神经网络滑模变结构控制在并联机器人中的应用

并联机器人系统结构复杂,具有强耦合、非线性等特点。滑模变结构控制对参数不确定性和外部扰动具有强鲁棒性,不需要被控对象精确数学模型且基于该方法的控制器设计过程是自然解耦过程,适用于并联机器人控制,但是滑模控制普遍存在抖振问题。鉴于此,该文提出RBF神经网络与滑模控制相结合的控制方法,利用RBF神经网络对滑模控制器切换项的增益进行调节,可以有效地降低滑模控制的抖振,获得较好的控制效果。仿真结果表明,该控制方法跟踪性能好,系统误差小,具有较强的鲁棒性,可以满足并联机器人的控制要求。

The application of RBF neural network sliding mode control on parallel robot

The parallel robot has a complex system,the strong coupling and nonlinear characteristics.The sliding mode control is not sensitive to uncertainly parameter and external disturbance,which dose not need accurate mathematical mode of controlled object and the process of sliding mode controller is a natural decoupling process.This method is applicable to parallel robot control,but which has the shortcoming of chattering.In view of that,this paper proposes a control method that combines RBF neural network with sliding mode control.By the use of RBF neural network to adjust the sliding mode control＇s gain of the switching,which effectively weaken the chattering and obtain the good control effect.The simulation results show that the control method has good tracking performance,small system error and strong robustness,and can satisfy the requirements of the parallel robot control.