Azimuth DOA Estimation Using Concentric Coprime Uniform Circular Array With Sparse Bayesian Learning

Estimation of direction-of-arrival (DOA) is a fundamental problem in array signal processing. The uniform circular array (UCA) is a highly useful array geometry for DOA estimation, especially in electronic warfare (EW), as it possesses useful properties such as complete 360° azimuth coverage and DOA estimation accuracy that is almost invariant with azimuth unlike linear array (LA). In addition, it is well suited for physical installation on complex platforms such as aircraft, ship, and unmanned aerial vehicle (UAV). Given the inherent sparsity of the DOA parameter in the azimuth domain, the last decade has seen significant research applying compressed sensing and sparse signal recovery with sparse array geometries. The difference co-array of virtual sensors of the physical array is dense; hence, sparse arrays can provide more accurate DOA estimation. In this paper, we propose a novel sparse array geometry called a concentric coprime uniform circular array (CCUCA) and perform azimuth DOA estimation by employing sparse Bayesian learning (SBL) with the difference co-array. This antenna array geometry is well suited for installation on complex platforms used for surveillance in EW compared to a nested sparse circular array (NSCA). The results of numerical experiments demonstrate the performance of the proposed technique. We also present the design of a novel passive frequency-independent CCUCA simulator, which can be realized using radio-frequency (RF) power dividers and delay lines. This can be used for generating multiple RF signals with appropriate phases for laboratory testing of the DOA estimation system realized using phase-coherent, multichannel RF and digital signal processing (DSP) hardware.