最小距离谱分析

Minimum Distance Spectral Analysis

本文提出了一种新的谱估计方法——最小距离谱分析(MDSA),它是所有满足相同观测自相关函数的谱中与光验谱的距离最小者.该方法具有结果的唯一性、幂等性等性质,是一种凸集投影方法.本文给出的几个具体例子表明了 MDSA的优良性能.

A new spectral analysis method named minimum distance spectral ana-lysis(MDSA)is developed in this paper.Minimum distance spectrum has a mini-mum distance from a priori spectrum as compared with other spectrums that con-sist with the same measured autocorrelation functions.The result of MDSA has suchmathematical properties as uniqueness and idempotence.MDSA is a convex pro-jection method.Simulations showing the performance of MDSA are also illustra-ted.