Publications
Detection of Shallow Water Low Frequency Sources Using Modal Peak Energy Detection
Abstract
Detection and bearing estimation with horizontal arrays in shallow water is difficult due to multipath. This paper develops a passive broadband technique for combining the the multipath energy to improve detectability and clutter rejection in shallow water environments. The method uses a normal mode model to predict the location of peak arrivals in the frequency-azimuth plot so that they can be combined. The new technique, called Modal Peak Energy Detection (MoPED), was inspired by the Sub-band Peak Energy Detection (SPED) approach that has been successfully implemented in operational sonar systems. The SPED processor sums peaks in a single beam across frequency to produce a bearing time record. The MoPED processor sums peaks across both frequency and beams, using a ``mode map'' derived from an environmental model. SPED suffers in complicated shallow water multipath environments when the peaks are significantly spread in azimuth. In these environments, MoPED can significantly improve detection probability and clutter rejection with little additional computational complexity.
The paper appeared in the Proceedings of the 13th Adaptive Sensor Array Processing Workshop at MIT Lincoln Laboratory, June 2005.