Maximum Likelhood Methods for Direction-of- Arrival Estimation
Petre Stoica, K.C. Sharman · 1990
This paper considers five methods of direction-of-arrival (DOA) estimation, all of which can be derived from the maximum like- lihood (ML) principle. The ML method (MLM) results from the ap- plication of the ML principle to the statistics of the observed raw data. The standard multiple signal classification (MUSIC) procedure, called MUSIC-I in this paper, can be obtained as a brute force approximation of the MLM. An improved MUSIC procedure, named MUSIC-2, is ob- tained by applying the ML principle to the statistics of certain linear combinations of the sample noise space eigenvectors. The MUSIC pro- cedures are more computationally attractive than the MLM. However, from a statistical viewpoint they may be very inefficient. A procedure which compromises between the good performance of the MLM and the computational simplicity of MUSIC is a method of direction esti- mation (MODE-I) which is derived as a large sample realization of the MLM. Finally, the fifth method, called MODE-2, is obtained by using the ML principle on the statistics of certain linear combinations of the sample eigenvectors. MODE-2 is computationally less demanding than the MLM (it is of the same complexity as MODE-1) and statistically more efficient. The computationally and statistically efficient MODE-2 introduced in this paper is considered to be a significant new result in this field. The paper also contains a numerical comparison of these five DOA estimation methods, which confirms our analytic results on their theoretical performance levels.