不确定离散系统的神经网络自适应控制方法研究

摘要

In this paper, adaptive neural network controller of the uncertain discrete-time systems is designed. Because it does not assume that the system slate is measurable, it uses an observer to estimate unmeasured states. The results compared with existing discrete systems, the controller has fewer direct adaptive parameters. Therefore it can easily achieve engineering algorithm. By using Lyapunov analysis, all of the closed-loop system signals are guaranteed to be the uniformly ultimate boundness ( UUB ). It can achieve the system output tracking the reference signal to bounded compact set. A simulation example is taken to verify the effectiveness of the method.%本文研究了不确定离散系统的神经网络自适应控制器的设计.因为它不需要假设系统状态是可测的,一个观测器用来估计不可测状态.与现有离散系统的结果相比,该控制器具有较少的直接自适应参数.因此,可以很方便地实现工程算法.利用Lyapunov分析方法,所有的闭环系统的信号是保证最终有界(UUB),并且能够实现系统输出跟踪参考信号到有界紧集.一个仿真例子,验证了该方法的有效性.