基于神经网络非线性系统辨识和控制的研究

Identification and Control of Nonlinear Systems Using Neural Networks

本文提出了由静态的前馈网络和稳定的滤波器构成的非线性系统的辨识模型．在神经网络固有的逼近误差存在的情况下，从理论上计论了神经网从应用于辨识和控制过程中系统的稳定性问题．最后研究了在非线性系的轨迹跟踪过程中增加滑动控制来补偿神经网络的逼近误差，从而提高系统的跟踪性能．

In this paper we present an identification model constructed by static feedforward neural networks and stable filters for nonlinear dynamical systems.Adaptive identification and control schemes based on neural networks are shown to guarantee stability of the system, even in the presence of neural network approximation errors. Finally, sliding control is used to compensate for inherent network approximation errors in order to improve the tracking performance.