时滞离散非线性系统基于NN预测的准滑模控制

NN-prediction-based quasi-sliding mode control for discrete nonlinear system with time delay

研究一类具有输入滞后的离散非线性系统的准滑模控制问题 .根据滑模控制原理和神经网络的逼近能力 ,提出了一种基于神经网络预测的准滑模控制器设计方法 ,给出了神经网络预测器的自适应算法 .通过理论分析和仿真结果 ,证明了神经网络预测器的自适应算法是收敛的 ,闭环准滑模控制系统是稳定的 。

The problem of quasi sliding mode control for a class of discrete nonlinear systems with input lag is studied in this paper. Discusses the theory behind the sliding mode control and the approximation capability of neural networks, and the synthetical design scheme for NN prediction based quasi sliding mode controller and the adaptive algorithm for NN predictor given to reduce prediction errors, and concludes from theoretical analysis and simulation results, that the adaptive algorithm is convergent, the closed loop quasi sliding mode control system is stable, and the tracking errors can converge to a neighborhood of zero.