Association of Critical, Infrequent Data for Network-Centric Multiple Frame Association for Distributed Multiple Target Tracking

The ability to detect, assess and resolve ambiguity in the measurement- or feature/attribute-to-track association process is fundamental to the success of correct fusion of information and to the robustness of combat identification algorithms that depend on the correctness of this association process. To address these issues, Numerica proposes a program based on the detection of short term ambiguity and the use of a Bayesian network tracking data base (BNTD) that makes use of infrequent or very late data to resolve the association uncertainty in support of combat identification. Bayesian network methods are also used to store, retrieve, query, and modify the database to support long-term ambiguity assessment and resolution. The overall objective of this proposed Phase I effort is to demonstrate a proof of concept for using critical but infrequent or very late data to resolve track and combat identification ambiguity in decentralized, distributed MHT/MFA tracking environment. Implied in this objective is the subordinate issue of the construction of proper frames of data on which the association ambiguity rests.