High spatial resolution radar using thinned arrays

The structure and size of a phased array antenna defines the beam pattern and the resultant spatial resolution of the radar. Often an exceedingly large number of radiating elements are required to synthesize a given array aperture in order to enhance the ability of the radar to unambiguously distinguish closely spaced targets. Classical random array theory provides tools for thinning of huge arrays without spatial aliasing, but at the cost of high sidelobe levels and unstable performance. Traditional beamforming and classical numerical methods (such as MUSIC and ESPRIT) cannot accurately estimate direction-of-arrival (DoA) for thinned arrays. In this paper, we employ a spatial compressed sensing (SCS) framework for accurate DoA estimation by a randomly thinned phased array. For the first time, we also sparsify a phased array-MIMO hybrid and suggest preferred thinning methods for the hybrid that uses non-overlapping subarrays. We estimate DoA for these arrays using the recently proposed multi-branch matching pursuit (MBMP) algorithm. Numerical experiments show MBMP outperforms classical beamforming and orthogonal matching pursuit by several orders of DoA recovery error.