Robust reverberation reduction for a stationary target in the shallow water

Abstract

Reverberation has an adverse effect in target echo detection, especially in shallow water. The low-rank and sparsity decomposition method can effectively reduce reverberation, based on the characteristic of low-rank for reverberation and sparsity for a moving target, in multiple pings of detection. However, the complexity of the target state challenges the performance of this method. In the case of a stationary target, the distinctions of the characteristics between reverberation and the target become ambiguous. After reverberation reduction is performed, the target echo can appear in the low-rank matrix rather than in the sparse matrix. Consequently, the target cannot be detected. There is a degraded reverberation reduction performance. To guarantee a meaningful decomposition, based on the random orthogonal model and the random sparsity model, the identifiability condition for the decomposition is derived. According to the condition, the method of sparse matrix is increased in dimension size is proposed, to manage detection probability shrinkage caused by the stationary target. Finally, the robust reverberation reduction performance is verified via some number simulations and synthetic data. It is demonstrated that the stationary target could be detected under a large ping number of the matrix.