Reducing DOA estimation error in extended ES-Root-MUSIC for Uniform Rectangular Array

In this paper, we aim at applying ES-Root-MUSIC method for Uniform Rectangular Array (URA) to be able to use Spatial Smoothing Preprocessing (SSP) in Direction-Of-Arrival (DOA) estimation. We propose a novel virtual array manifold and interpolation matrix to reduce D OA estimation error in Extended ES-Root-MUSIC method. Array interpolation process like ES-Root-MUSIC achieves high estimation accuracy with small computational cost, but ES-Root-MUSIC can be applied only when the number of virtual array elements is odd. The extended ES-Root-MUSIC, introducing SSP-like transform into virtual array, can estimate DOAs for both odd and even virtual array numbers. However, the accuracy of the Extended ES-Root-MUSIC becomes worse than ES-Root-MUSIC when the number of virtual array elements is odd. This paper proposes a novel definition of the virtual array manifold and interpolation matrix toward higher resolution DOA estimation. Also we consider the axes of coordinates and relaxing interelement distance in URA.