Maximum likelihood methods for direction-of-arrival estimation
Authors
Discussion
Loading...
Member takes
Implication
Methodological study demonstrates performance tradeoffs among five direction-of-arrival estimators in sensor arrays, highlighting MODE-2 as an efficient and optimal alternative.
Key Points
- To derive, analyze, and compare five direction-of-arrival (DOA) estimation methods based on the maximum-likelihood principle to balance computational complexity and statistical accuracy.
- Formulated five DOA estimators: exact maximum likelihood (MLM), standard MUSIC (MUSIC-1), an improved MUSIC variant using noise-space eigenvector statistics (MUSIC-2), and two method of direction estimation variants (MODE-1 and MODE-2).
- Conducted numerical simulations to validate theoretical performance levels and computational demands across all five estimators.
- Standard MUSIC (MUSIC-1) serves as a brute-force approximation of MLM, whereas MUSIC-2 enhances accuracy by exploiting the statistics of linear combinations of sample noise-space eigenvectors.
- MODE-2 matched the low computational complexity of MODE-1 while demonstrating superior statistical efficiency and closely matching theoretical performance limits.
Cite This Study
Stoica et al. (1990) studied this question.