Network-Centric Multiple Frame Association For Distributed Multiple Target Tracking

"Multiple frame data association methods such as MHT (multiple hypothesistracking) and MFA (multiple frame assignments) have establishedthemselves as the preferred tracking method. The reason for the successof this class of methods is that difficult data association decisionsare held in abeyance until additional information is available. Thiscapability has been developed for either single sensor tracking and fora centralized architecture for multiple sensor tracking; however, itsuffers from single-point-failure, delays, and communication loading.Thus, one must turn to a distributed architecture for bothestimation/fusion and data association.The objective of the proposed program is the development of adecentralized network-centric multiple frame assignment (MFA) class ofarchitecture algorithms that affordably preserves the quality of acentralized architecture across a network of platforms while managingcommunication loading and achieving a consistent air picture on entitiesof interest for each platform. Using an extensive set of trackingmetrics on real AWACS data and high fidelity simulated data, issuesassociated with these architectures will be investigated and solutionsproposed. Numerica, Inc. has a very successful record of developing algorithms(specifically in the target tracking area), implementing the algorithmsusing highly-structured object-oriented software techniques, andlicensing the software to defense contractors and the