关于SISO系统参数辨识在时域中的最优信号设计问题

Problems about the Design of Optimal Signals in Time Domain for SISO System Parameter Identification

本文探讨了时不变线性随机SISO系统的参数辨识的最优信号问题。首先将Fisher信息矩阵的概念推广到随机信号的情况,然后建立了一个输出空间的G——最优准则,并且证明该最优准则与参数空间的D——最优准则是等价的,由此导出了一个最优信号算法。

This paper mainly surveys the field of optimal input design for linear single input/output parameter estimation. Firstly, the Concept of Fisher information matraix is extended to the conditions of stochastic signals, and then a G—optimal criteria belonging to output space is set up. And it is proved that D—optimal criteria and G—optimal criteria are equivalent, thus deducing an algorithm of the optimal stochastic signal.