Subarray processing with coprime and minimum redundancy arrays

Abstract

A coprime sensor array (CSA) refers to a non-uniform linear array consisting of two undersampled uniform linear arrays (ULAs) with coprime numbers of elements, respectively. Its beam pattern can be enhanced by multiplying beampatterns of individual subarrays. This technique, called product processing, provides improved resolution because its sum coarray length is extended from the product of beampatterns. Nevertheless, the extended coarray inevitably includes holes that result in higher peak sidelobes than the full ULA case. In this paper, we propose a technique to reduce peak sidelobes through the combination with minimum redundancy arrays (MRAs) and an additional subarray. Two ULAs comprising the coprime array are thinned by the MRA technique, and the reduced number of sensors is then utilized to construct the third subarray. The third array suppresses peak sidelobes by placing nulls at the peak sidelobe locations. The characteristics of the proposed technique are analyzed in terms of white noise gain (WNG) and maximum sidelobe level (MSL). It is demonstrated that the proposed technique can improve the MSL with similar number of sensors, thereby reducing the estimation error in the presence of multiple sources.