基于倒谱的非最小相位FIR系统自适应辨识

Cepstrum Based Adaptive Identification of Nonminimum Phase FIR System

本文介绍了一种在高阶谱一维切片的基础上利用倒谱进行非最小相位FIR系统自适应辨识的算法。这一算法具有计算量小，不需要预先假定系统阶次的特点，而且能够自适应地分别辨识出非最小相位FIR系统的最小相位部分和最大相位部分的冲激响应，同时，系统阶次可以在自适应过程中确定。另外由于本算法利用观测信号的高阶级计量进行系统辨识，所以很好地抑制了高斯有色噪声对辨识结果的影响。数值仿真结果证明了本算法的有效性。

This paper presents adaptive identification algorithm of nonminimum phase FIR system using cepstrum on slices of higher order spectra. This algorithm has small computation load, the system order is not re-quired known as a priory, the minimum phase part and the maximum phase part of the impulse response can be adaptively identified respectively, and the order of the system can also be determined during the process. Since this algorithm utilzes only higher order cumulants of the observed signal, the measurement nooise is allowed to be color Gaussian noise. Simulation examples are presented to prove its effectiveness.