Performance Analysis of the Music Algorithm with Spatial Smoothing in the Presence of Coherent Sources

We study the effect of spatial smoothing on the performance of the MUSIC algorithm for directions-of-arrival estimation of partially and fully coherent sources. We show that spatial smoothing improves the sensitivity of the invariant subspaces as a result of progressive decorrelation of the sources, and that the rate at which this happens depends upon the spacing and directions of the sources. This analysis explains the performance improvement observed with moving arrays in the presence of coherent sources. We also show that smoothing reduces the magnitude of the perturbations through subarray averaging, thereby further improving DOA estimation performance. Results of computer simulations are included to support our analysis.