Among high-resolution direction-of-arrival (DOA) estimators, it is known that the maximum likelihood estimation (MLE) has the most desirable performance in resolution. However, when the SNR falls below a certain threshold, its performance degrades considerably. The prime concern of this research is to try to provide a new estimator which can maintain an acceptable performance for the hard conditions under which the MLE fails. The way to achieve this estimator is to introduce a robust hierarchical statistical model to the sampled data. Simulation results have verified that the resultant robust estimator provides improvement in resolution and possesses certain robustness with respect to the misspecification on the signal model.<<ETX>>
[1]
I. Ziskind,et al.
Maximum likelihood estimation via the alternating projection maximization algorithm
,
1987,
ICASSP '87. IEEE International Conference on Acoustics, Speech, and Signal Processing.
[2]
Ilan Ziskind,et al.
Maximum likelihood localization of multiple sources by alternating projection
,
1988,
IEEE Trans. Acoust. Speech Signal Process..
[3]
J. Berger.
A Robust Generalized Bayes Estimator and Confidence Region for a Multivariate Normal Mean
,
1980
.
[4]
I. J. Good,et al.
The Robustness of a Hierarchical Model for Multinomials and Contingency Tables
,
1983
.