Maximizing mutual information in horizontal linear arrays.

Abstract

Estimating the bearing of a sound source using a towed horizontal array is a common array processing problem. In a passive sonar system, it is common to null sources from the forward endfire direction to eliminate the noise of the towing vessel. This research designs linear nonuniform symmetric arrays of a fixed aperture for estimating the bearing of sound source. Specifically, the hydrophone spacings for a symmetric linear array are chosen to maximize the upper bound on mutual information between the true bearing and the bearing estimate in white noise. The arrays’ maximizing the mutual information while nulling forward endfire look significantly different from the uniform arrays commonly used in towed systems, and also differ from arrays designed to minimize spacing redundancy. Arrays maximizing mutual information are helpful when bearing estimation is considered as a quantization problem to assign the source to the correct partition. In this approach, increasing the mutual information should reduce the probability of error (Pe) in choosing the partition containing an unknown source. Simulation results demonstrate that the optimum arrays proposed here have much lower Pes than the uniform array. [Work funded by ONR code 321US.]