Steerable sparse linear array design based on compressive sensing with multiple measurement vectors.

Abstract

This paper presents the synthesis of a non-uniform linear array for two-dimensional Range-Bearing underwater acoustic imaging based on the multiple beam, measurement vectors (MMV) Compressive Sensing (CS) method. This method considers multiple steering directions in the mathematical model and exploits the sparsity of the oversampled linear aperture to obtain a thinned array structure. In this paper, a simulation study to synthesise the sparse linear array (SLA) for a Forward Looking Sonar system considering multiple beams in the field of view (FOV) was conducted. The comparison of the array synthesised using single measurement vector compressive sensing (SMV CS), which considers only the broadside beam in the mathematical model, is also presented. From the imaging point of view, when the SMV CS synthesised SLA is steered in a direction other than broadside, spurious/unwanted lobes appear in the beam pattern (BP). To overcome this problem, the MMV method and demonstrate its performance efficacy with a simulation study are proposed. The design of SLA for 37.5, 75, 150, 300, and 600 kHz with angular resolution 6.4°-0.4° with 11-149 beams in the FOV keeping the aperture fixed are presented, as is a design keeping angular resolution fixed at 1.5° to generalise a design of SLA for the above mentioned interested frequencies.