不确定系统的不确定项观测器设计

Uncertainty observers for uncertain systems

为了克服常规用不确定项的界估计来设计控制器带来的保守性和动态性能差的弊端,本文就连续的和离散的不确定系统,分别提出了不确定项的观测器设计方案.通过引入分布和离散加权平均的概念,对这两种观测器的观测误差精度进行了分析估算,从数学上证明了所得的观测值在加权平均意义下是无偏的.此外,通过适当选择观测器反馈增益,可以得到满意的误差估计精度.把该方法应用到一个控制系统中,仿真结果表明所提出的观测器能很好的估计不确定项和改进闭环系统动态性能.

The estimation for uncertainties in dynamic systems plays an important role m controller design, ana me traditional bound estimation for uncertainties would induce more conservative controller or worse dynamic performances. Motivated by uncertainties estimation via state feedback, we proposed uncertainty observers design schemes for both con- tinuous and discrete uncertain systems in this paper. By introducing the concept of distributed and discrete weighted averages, observation errors are analyzed and the observed values are mathematically proved to be unbiased in terms of the weighted average. Moreover, satisfactory accuracy of observation errors can be achieved via choosing appropriate observer gain matrix. The scheme is applied to a control system. Simulation results show that the proposed observers can efficiently estimate the uncertainties, and the improved closed-loop dynamic performances are achieved.