信源数目估计方法的研究

盲信号处理（Blind Signal ProCessing）是近年来信号处理研究的热点之一,＂盲＂是指在没有任何先验信息或者先验信启、不足环境下的假设,因此,该处理方法对环境及信号的应用更加宽泛。盲信源数目是盲分离的首要前提,信源数目的确定更是直接关系到盲分离效果的正确性。本文主要研究了IAC准则,MDL准则和盖式圆准则的盲信源数目估计算法。在不同信噪比情况下,分别对噪声观测信号进行信源数目估计。通过实验证明,盖式圆在含有噪声的情况下,能够相对准确的估计出信源数目,IAC和MDL在低信噪比情况下,不能准确的估计出信源数目。

Optimizing algorithm in determining the number of signals

Blind Signal Processing is one of the hot topics in the study of Signal Processing recent years.The "Blind" means wo have the information which is without any priori information or lack of prior information. Therefore,this treatment method is more inclusive on the environment and the application of the signal.The estimate of number is first premise in the blind separation of source.The number of source is directly related to the correctness of the blind separation and it is very important.This paper mainly studies the algorithm of estimating the number with the IAC standards,MDL standards and Getschgorin theorem.Estimating the numbers of the observation noise signals,In different SNR.Through the experiment,we can see that the Getschgorin theorem can be able to be relatively accurate estimate of the number of source.And IAC and MDL and can’t accurate estimate of the number of source in low SNR.