Sparse Reconstruction-Based Underwater Source Localization Using Co-Prime Arrays

Traditional matched-field processing (MFP) techniques for underwater source localization such as Bartlett and Minimum variance Distortionless Response (MVDR) suffer from low resolution and sensitivity to environment/system mismatch respectively. This paper considers the problem of underwater source location estimation using a combination of sparsity imposed reconstruction method with newly proposed structure of non-uniform linear sensor arrays, regarded as co-prime arrays. This combination overcomes the above mentioned drawbacks of the traditional methods. Co-prime arrays consist of two subarrays with M and N sensors respectively. By using co-prim arrays the degree of freedom can be increased from O(N) to O(MN). It is a general observation that sparse reconstruction algorithms exhibit high computational workload. To reduce this burden, we use Bartlett for initial estimation as Bartlett is known as robust but low resolution method. The sparsity based algorithms discretize the range of concern onto a grid and off-grid sources can lead to mismatches in the model and weaken the performance notably. In this work we apply the joint sparse recovery method to reduce the off-grid mismatches. It is observed with the help of numerical examples that the proposed method has a number of advantages over other existing methods, such as high resolution, robust estimation and less off-grid mismatches.