一类仿射非线性系统的自适应神经网络输出反馈变结构控制

Adaptive neuro-network-based output-feedback variable structure control for a class of affine nonlinear systems

本文研究了一类仿射非线性系统的输出反馈控制问题.在介绍文献[4～6]的基础上,提出一种基于神经网络参数化技术的自适应变结构输出反馈控制方案,该方案能够避免使用严格正实(SPR)条件,它不仅能够保证收缩条件的可行性,而且还可以分析闭环系统的稳态和暂态的一致有界性,并能够对观测增益和控制参数的选取进行清楚地分析.

Based on the results of prior work, we employ the neuro-network parameterization and adaptive variable structure control to propose an output-feedback control for a class of nonlinear affine systems, avoiding the requirement of the strictly positive realness（SPR） condition. This scheme ensures the desired contraction in control, the boundedness of steady-states and the stability of the closed-loop system. It also provides an easy and clear way for selecting the parameters of the observer and the controller.