Unified Bayesian Feature-Aided Association, Identification, and Tracking
Small Business Information
SCIENTIFIC SYSTEMS CO., INC. (Currently Scientific Systems Company, Inc.)
500 West Cummings Park - Ste 3000, Woburn, MA, 01801
AbstractCurrent techniques for detecting, tracking, and identifying ground targets frequently fail because geokinematic data alone is insufficient to sort out real-world complexities. Such difficulties can be ameliorated through the use of feature-aided association, identification, and tracking, (FAAIT) i.e., the use of target identity information to aid the association of reports to tracks and thereby to enhance the tracking and identification of targets. However, many difficulties limit the effectiveness of fusion of geokinematic data with feature data from multiple sensor or source types. Scientific Systems Company, Inc. (SSCI) and its subcontractor Lockheed Martin Tactical Systems (LMTS) propose the use of generalized Bayesian techniques to: (1) optimally fuse multisource geokinematic and feature data using likelihood functions; (2) robustly deal with unknowable real-world uncertainties using generalized likelihood functions; and (3) deal with multiple targets in complex scenarios using generalized multi-hypothesis correlation techniques or, more generally, approximate multitarget recursive Bayes nonlinear filtering techniques. The major objectives of Phase I will be to: (1) identify features from several disparate sensor types which could be combined to aid in target classification or report association; (2) develop the necessary theoretical extensions; (3) develop the necessary computational approximations; and (4) implement a concept-feasibility prototype algorithm; (5) test this algorithm in simulated, reduced-complexity scenarios; and (6) develop a detailed plan for further analysis and implementation in a Phase II effort. Phase II will emphasize development, testing, and evaluation of a prototype feature-aided tracking, association, and identification (FAAIT) algorithm to be tested using simulated or real data from an expanded set of feature or sensor types. Commercial applications of the approach will also be investigated by SSCI and LMTS during Phase I and fully developed during Phase II. The technology proposed in this proposal has specific applications to military and commercial applications. Military applications include surveillance of ground stationary and moving targets. Commercial applications include airborne and spaceborne mapping and charting. The technology developed under this effort can be directly extended to commercial distributed traffic monitoring for Intelligent Highway Systems and for law enforcement applications such as border surveillance.
* information listed above is at the time of submission.