基于小波分解的系统辨识最优实验设计

Optimal Experiment Design for Wavelet-Based System Identification

提出了一种基于小波分解的系统辨识最优输入信号的设计方法。将小波的多分辨率分析理论应用于控制系统的最小二乘辨识中,并且与闭环系统的控制性能指标关联起来,从而得出一种最优辨识输入信号的设计方法。这种方法是与控制相关的一种辨识方法,能获得比普通的系统辨识方法更好的结果。仿真结果证明了这种方法的有效性。

This paper proposes a method for optimal input signal design in system identification. We firstly apply the wavelet multiresolution theory to the least-squares system identification, and associate it with the control performance index of the closed-loop system, and then present a method for optimal identification input signal design. This method is a control-relevant identification method, and it could get better results as compared with the existing system identification methods. The simulation result proves the effectiveness of the proposed method.