Two-stage active sonar network track-before-detect processing in a high clutter harbor environment

Abstract

Reliable active acoustic detection and tracking of small targets while minimizing the number of false alerts is challenging in a shallow, multipath-inducing, high-clutter harbor environment. These targets often exhibit low target strength and the acoustic clutter fields can be dense and highly dynamic. One approach to detect and track targets in this environment involves a two-stage tracker due to the processing gain required to continually track a weak target in such a significant clutter field. The first stage uses more than a single frame of observations to not only initiate tracks but also to contribute to track maintenance. The second tracking stage is initiated by the tentative target tracks extracted from the first stage. Track segment association algorithms are employed to combat periods or locations track intermittency that inhibit detection and cause tracks to fail and restart. This paper describes the results of this two-stage approach to track-before-detect processing applied to a networked harbor surveillance active acoustic detection and tracking system that consists of a large number of relatively simple nodes, employs a windowed Hough-Transform tracker at the beam level, and a Kalman-based, multi-object tracker at the single- and multi-node levels.