Title :
Robust estimation of the number of sources using an MMSE-based MDL method
Author :
Teng Long ; Lei Huang ; Erke Mao
Author_Institution :
Dept. of Electron. Eng., Beijing Inst. of Technol., Beijing
Abstract :
In this paper, using the mutual information, we bridge the probability density function and the minimum mean-square error (MMSE) between the observed data and the desired signal, and then employ the MMSE to construct an MMSE-based MDL criterion for accurate source enumeration. The presented numerical results demonstrate that the proposed method is superior to the existing MDL methods in detection performance.
Keywords :
array signal processing; eigenvalues and eigenfunctions; estimation theory; least mean squares methods; MMSE-based MDL method; eigenvalue decomposition; minimum desciption length; minimum mean-square error; probability density function; robust estimation; sensor array processing; source number estimation; Eigenvalue decomposition (EVD); Minimum description length (MDL); Sensor array processing; Source number estimation;
Conference_Titel :
Radar Conference, 2009 IET International
Conference_Location :
Guilin
Print_ISBN :
978-1-84919-010-7
