Compressed Sensing based Tow Ship Interference Cancellation and Target Detection in Shallow Ocean

Abstract

Passive towed arrays assume great importance in coastal surveillance, providing the capability of long range detection and localization of targets. The performance of the towed array sonar is bounded by the intense tow ship radiated signal picked up by the array. This problem is magnified in shallow ocean, owing to the multipath and multimodal propagation of the acoustic waves. A popular solution for mitigating the own platform interference is based on the Eigen analysis of space time covariance matrix (STCM). As length of the sensor array increases, matrix dimension increases, prohibiting the practical realization of this approach. In this paper, a computationally efficient pre-processing technique based on the Compressed Sensing (CS) framework is investigated. The efficacy of the method is initially evaluated with the monte carlo simulations of shallow ocean data model. The scheme is further validated on the experimental data captured during the field trials in the Arabian Sea. It is demonstrated that employing the proposed scheme, tow ship noise is cancelled substantially, aiding the detection of low signal to noise ratio (SNR) targets, while the computational requirements are diminished to a large extent.