Adaptive predictor of surface reverberation in a bistatic sonar

Abstract

A bistatic sonar employing LMS adaptive spatial prediction is used to process against surface reverberation from the convergence zone (CZ). Hydrophones (120, 130) spatially separated from the primary array (100) are used as references. The critical behavior exploited by the invention is that since the reverberation may be viewed as an extended source for the reverberation, the correlation drops off with separation between the reference and primary sensors, while it stays a constant for the plane wave target return. The reverberation is nonstationary functionally dependent on the signal (even though the back scatter is statistically uncorrelated with the signal), and spatially extended over the sector of the CZ annulus cut out by the transmitter azimuth beamwidth. The detection of the target is based on the sudden appearance of one strong target point source (50) within a densely packed region of weak point sources that have been constantly present and whose sum is much larger than the target. The use of more than a single reference leads to a minimal improvement in detection performance and may actually degrade performance due to increased algorithm noise. Detection performance tends to increase with increasing separation distance between the primary and the reference hydrophone.