Performance analysis of ML estimate for DOA in an abnormal case
Hongyi Yu, Zheng Bao · 2002
The estimation of DOA, extensively used in radar, and communication, has drawn much attention. In this paper, an analysis of the performance of the ML method of direction-of-arrival estimation is made, under the condition that the estimated signal number is different from the real signal number. Based on a discussion of the estimation of an infinite snapshot, an asymptotic distribution of the estimation is derived, from which the analytic expressions of asymptotic estimation variance are given. By the results, a research of the relationship of estimation variance to separation in incident angles is made, and some interesting results are attained. The results expected by the theory are in good agreement with the computer simulation.