Weighted Incoherent Signal Subspace Method for DOA Estimation on Wideband Colored Signals

Abstract

Wideband direction-of-arrival (DOA) estimation is a key part in array signal processing. The existing algorithms for the wideband DOA estimation are often studied in the situation of uniformly distributed energy. In addition, all the frequency bins are weighted equally in these algorithms. However, these algorithms perform unsatisfactorily when encountering wideband colored signals with nonuniform energy spectrum. To improve the performance of DOA estimation for wideband colored signals, we proposed two weighting methods, which are based on the perturbed subspace theory and random matrix theory, respectively. The two methods weight the space spectrum from all the frequency bins according to the mean square error of DOA estimation in each frequency bin. The numerical results show that the random matrix theory based method performs well, due to the inference premise that the dimensions of matrices increase at the same rate. The perturbed subspace-based method, which is concise in calculating the weights, shows high accuracy only at high-signal-to-noise ratio and with adequate snapshots. The effectiveness of the two algorithms are also demonstrated by comparing them to various existing algorithms and the Cramér–Rao bound.

Keywords

• Direction-of-arrival estimation
• random matrix theory
• signal subspace method
• wideband signal