Statistical performance comparison of high-resolution DOA estimation methods

A wide comparison of statistical performance among high-resolution methods of direction estimation (MODE) is presented. The development of MODE is first reviewed. In order to investigate their statistical performance clearly, more than 30 kinds of methods are studied and divided into 5 groups. Ten typical approaches are evaluated in resolution and precision respectively. Based on a united simulation model, their performance are compared by using Monte-Carlo statistical analysis. Furthermore, several criteria for the application of MODE are presented to judge their practicality. The theoretical study shows the performance of weighted subspace fitting (WSF) possesses, and indicates that MUSIC, Mini-Norm, TLS-ESPRIT are the most promising methods in practice.