Multiple Database Evidence Accrual Techniques

"Current Intelligence, Surveillance, and Reconnaissance (ISR) systems have access to data originating from a number of assets each with distinct information content and reporting characteristics. A process for combining the information from these assets toform an accurate and integrated tactical picture is needed. Reports may be delayed or consist of track segments. Correct report association is arguably the key factor driving the performance of target tracking and identification, and ultimately thefidelity of the collective tactical assessment. ISI proposes a novel application of Bayesian Networks (BN) to accomplish multitarget, multisensor report association and identification. Practical system implementation is afforded by trading optimality withcomputational requirements through user-specified and/or automatic control of tracking parameters. A BN architecture structured for efficient implementation of a multiple hypothesis tracking regime, with the potential for optimal Bayesian association andID, has been set forth with the flexibility to intrinsically handle association of delayed reports to fused-tracks. Procedures for performing the track fragment association within the BN framework are established. Military applications include battlefieldsurveillance, target identification, situational awareness analysis, and high-value asset location. Commercial applications include air traffic control, traffic monitoring and flow analysis,