Modeling the signal processing of a multi-sonar system to determine interference solutions

The quality of a multibeam echo-sounder sonar (ES) signal used in seafloor mapping is degraded by an interfering sidescan sonar (SS) signal. The ES and SS signals occur at approximately the same time, are in the same frequency band, and arrive from approximately the same direction. The ES signal is a 2 ms continuous wave pulse with a first null bandwidth of 1 kHz centered at 187.5 kHz. The interfering SS signal is an 8 ms linear frequency modulated chirp with a bandwidth of 2.5 kHz centered at 183 kHz. The presence and properties of the interfering SS signal are explained using a model of the ES signal processing. Model results as well as spatial and temporal frequency analysis of the recorded ES signal show that most of the interfering SS signal can be removed by zero phase filtering with a 10 tap boxcar filter. This additional filtering reduces the resolution of the multibeam echo-sounder system from 1.318 meters to 1.873 meters. Because this reduction in resolution is not acceptable, the design trade-offs of moving either the ES or SS frequency bands are discussed. Statistical analysis of seafloor detections, before and after the additional frequency filtering shows that there is a difference between filtered and unfiltered detections that is dependent upon the angle of arrival. The unfiltered detections are on average between 0.4% and 1.6% deeper than the towfish altitude of 50 m