基于动态神经网络的一类非线性不确定系统的自适应观测器

Adaptive observer for a class of nonlinear uncertain systems based on dynamical neural networks

基于动态神经网络 ,对一类非线性不确定系统提出了相应的观测器设计方法。在观测器设计中 ,充分考虑了不确定性和神经网络逼近误差对观测器性能的影响 ,增加了鲁棒控制项并设计了相应的参数自适应律 ,以保证良好的观测性能。神经网络的权值在线进行调整 。

Based on dynamical neural networks,a design method of adaptive observer for a class of nonlinearuncertainsystemsisdeveloped.Theeffectsofuncertaintermandneuralnetwork approximation error are well considered. A robust control term and weights updating laws are designed toguaranteetheperformanceofadaptiveobserver.Theweights of neural networks are updating on line, and the off line training phase is not required.