Time Domain Adaptive Beamformer for Sonar Signal Processing

Abstract

Passive sonars try to detect, locate and classify underwater acoustic sources by listening to sound waves impinging on sensor arrays and estimating the direction of arrival by beam forming methods. Strong noise sources from own platform and other interfering sources make detection and classification a difficult job to achieve in practical scenarios. Adaptive beam formers aim to minimize the interference from other sources while looking in a desired direction. Often it is necessary to have a wideband frequency response along the look direction to allow estimation of signal parameters in that beam. Wideband time domain beam formers combine desired frequency response along the look direction with interference power minimization thus offering much better estimation of the target spectral parameters along that direction. In this paper, the Frost's beam former was adapted for sonar processing with emphasis on using the beam former output for spectral processing and audio processing and the performance was evaluated under various input scenarios.