New high resolution direction of arrival estimation using compressive sensing

In military communication and several direction finding applications, precise estimation of the DOA, incorporated with smart beam-forming techniques, can provide useful solutions for interference mitigation and jammer identification. Compressive Sensing (CS) based DOA approaches offer a competitive advantage over the traditional DOA estimation techniques due to their ability to obtain an estimate with measurements collected at a single time instant (i.e. snapshot). However, in reality, CS based methods suffer from grid quantization errors as the source may not exactly align on the discrete grid points. This creates a mismatch between the observation and measurement matrices resulting in an inaccurate estimate. This paper proposes a fundamentally new, recursive algorithm to estimate incoming DOA of a source with zero quantization error. The recursive algorithm is strongly convergent and demonstrates superior performance relative to previous methods. The performance of the algorithm is validated by comparison to the theoretical Cramér-Rao lower bound (CRLB). An array geometry optimization technique is also proposed that enables the recursive algorithm to achieve the CRLB with just two iterations.