Performance of source number estimation using Gerschgorin disks

Abstract

The performance of classic MUSIC algorithm depends on the signal source number estimation directly. However, underestimate has negative influence on the DOA estimation. The signal source number estimation is one of the most important parts in array signal processing and other relevant field. In this dissertation, based on ULA, the influence of signal number on DOA, the influence of overestimate and underestimate on MUSIC algorithm are discussed and analyzed. Afterwards, AIC and MDL which are based on information theoretic criteria and method based on Gerschgorin disk are introduced in details. Finally, through computer simulations, the evaluation of each algorithm is given respectively and under comparison to highlight the performance of Gerschgorin estimator in signal source number estimation.