An approach for automated passive sonar contact localization

Abstract

This work seeks to advance the state-of-the-art of fully automated detection, tracking, and localization of passive sonar contacts for large diameter unmanned underwater vehicles (LDUUV). A particle filtering track-before-detect-like approach is taken to develop three core algorithms to perform the above functionality. Each of the core algorithms are examined in pertinent respects using a real data set representative of newer high-resolution LDUUV sonar systems. The data set is highly cluttered, has varying noise levels, and consists of moving contacts that fade, cross, and merge in the bearing space. For comparison, a mainstay point-measurement-based particle filter approach using probabilistic data association is employed and shown to have greater difficulty addressing the scenario challenges than the developed approach herein.

Reference

T. Northardt, “An approach for automated passive sonar contact localization”, IEEE Underwater Signal Processing Workgroup, Rhode Island, 2015.