SBIR-STTR Award

This proposal addresses the challenge of incorporating non-kinematic features into tactical active sonar tracking algorithms used in anti-submarine warfare. The central focus is the development of three general methods that use feature information (i.e., their statistical characterization) to improve both data association and filtering. The methods are general in that they apply to a wide variety of features and trackers. They will be implemented and tested in STRs Multi-Stage Multi-Hypothesis Tracker (MS-MHT). Based on gains seen in other domains, these feature-aided tracking techniques are expected to improve multi-target tracking performance in the presence of heavy clutter and high ambient noise. In addition, STR proposes to develop an information theoretic approach to assess the impact non-kinematic features have on the data association problem. The approach is applicable to all three methods and could potentially be used to dynamically make decisions about whether or not to integrate a feature into a tracking solution.

Benefit:
Improved multi-target active sonar tracking capabilities in anti-submarine warfare in challenging environments (heavy clutter, high ambient noise). Could impact operational systems such as the AN/SQQ-89. The development of three methods to exploit non-kinematic features in multi-target trackers that are applicable to a wide variety of features. Moreover, the technology is applicable to other sensing modalities (e.g. radar and infrared imaging) and commercial applications (e.g. automotive radar). The development of an information theoretic approach to quantify the impact non-kinematic features have on data association. This could potentially be used to dynamically control whether or not a feature should be used by a tracker.

Keywords:
Feature-aided tracking, Feature-aided tracking, Multi-Stage Multiple Hypothesis Tracking (MS-MHT), coupled recursive filtering, Active Sonar in Anti-submarine Warfare, state augmentation, Kullback-Leibler Divergence, single-ping scoring

This proposal addresses the challenge of incorporating non-kinematic features into tactical active sonar tracking algorithms applied to surface anti-submarine warfare. The central focus is the development of three general methods that use feature information (i.e., their statistical characterization) to improve both data association and filtering. The methods are general in that they apply to a wide variety of features and trackers. They will be implemented and tested in STRs Multiple Hypothesis Tracker (MHT). Two avenues to expand the core capability of the feature-aided tracking techniques are also proposed. The first explores multi-stage MHT architectures that fuse kinematic and feature information from multiple sensors. The second avenue explores algorithms to auto-tune the tracker to avoid feature model mismatch. Based on gains seen in other domains, these feature-aided tracking techniques and architectures are expected to improve multi-target tracking performance in the presence of heavy clutter and high ambient noise.

Benefit:
It is anticipated that this work will lead to improved multi-target active sonar tracking capabilities for surface anti-submarine warfare in challenging environments (heavy clutter, high ambient noise). In particular, this work is aimed at improving the performance of the AN/SQQ-89 Surface Ship ASW Combat System. The feature-aided tracking methods being developed are applicable to a wide variety of features and trackers. The technology is also applicable to other sensing modalities (e.g., radar and infrared imaging) and to commercial applications (e.g., automotive radar). It is anticipated that the multi-stage architectures will allow measurements from several sensors to be fused to together, multiplying the impact of the feature-aided techniques. Lastly, the auto-tuning capability should allow the techniques to be used even when the feature statistics are non-stationary.

Keywords:
Multi-Stage Multiple Hypothesis Tracking (MS-MHT), Multiple Hypothesis Tracking (MHT), single-ping scoring, composite confirmation logic, state augmentation, Anti-Submarine Warfare, Feature-aided tracking, Active sonar