未知非线性系统自适应改进RBF滑模控制

Adaptive improved RBF neural network sliding mode control for unknown nonlinear systems

针对一类未知的非线性系统,利用输入/输出线性化将其变换为部分线性可控系统,通过RBF神经网络对未知非线性函数进行逼近,提出了一种基于RBF神经网络的自适应滑模控制,并设计了自适应滑模控制器;提出了一种连续函数,很好地减少了抖振现象,使得闭环系统状态一致稳定最终有界。实验结果验证了方法的有效性。

A class of unknown nonlinear system is transformed to part-linear controllable system by input-output linearization and unknown nonlinear function is approximated with RBF neural network.An adaptive sliding mode control based on RBF neural network is proposed and an adaptive sliding mode controller is designed.A continuous function is proposed which can greatly reduce chattering phenomena and make closed loop system have uniform stability and ultimate bound.The results show that the proposed method is effective.