Evidence based target motion analysis

Abstract

An evidence-based approach to contact tracking (E-BaCT) is presented. The E-BaCT paradigm is different from classical target motion analysis and offers the potential advantages of intrinsically incorporating data association, heuristic rule sets, and constraints without the need for statistical models. Given a set of admissible contact trajectories, evidence in the form of features is extracted from available data and other information to form objective functions. The multiple objective functions are aggregated into a single objective function from which a tracking solution and a measure of solution quality are obtained. The E‑BaCT paradigm is presented as a design process and a specific realization is shown to use the Radon or Hough transform to extract some of the features and fuzzy inference for feature aggregation. A bearings-only target motion analysis application illustrates the algorithm’s robustness in a complicated scenario involving multiple merging-splitting-crossing and intermittent contacts.

Reference

Nardone, B. Guimond,  “Evidence based target motion analysis”, SIA Technology and Engineering Conference, Adelaide, Australia, 2013.