Two-step optimal subarray detection based on matched-field processing

The optimal detector based on matched-field processing with a large-aperture vertical linear array (VLA) suffers from the so called 'snapshot deficient problem', which leads to an inaccurate estimate of the noise covariance matrix and hence a degraded detection performance. In this paper, to improve the detection performance with deficient snapshots, a two-step sub-array detector (SAD) is proposed. Due to the reduction on the hydrophone number in the two-step processing, the proposed SAD improves the estimate of the noise covariance matrix. Thus, SAD has a better detection performance than the optimal full array detector (FAD) in a shallow-water correlated noise environment with short snapshots. Numerical simulation results demonstrate the effectiveness of the proposed SAD.