Advanced Data Fusion Using Bayesian Track-Before-Detect

Dispersed sensors offer more information to tracking systems than a single sensor and this additional information should improve the performance. Typical approaches only use information from a single sensor to initialize tracks and then use measurements from other sensors to improve the track estimate. In a new approach, the PMHT will be enhanced to not only identify new track starts, but also should yield better tracking performance by determining if a new track is realistic. With this enhancement to the PMHT, this emerging tracking algorithm will then be able to utilize its excellent clutter rejection to track targets in dense clutter environments. We also believe we have a solution to the optimistic covariance of the PMHT, and we will explore the feasibility of this solution during this effort. Therefore, we are proposing using the enhanced PMHT as a composite tracker on several platforms utilizing simulated sensor data. We think that the enhanced PMHT will improve the single integrated picture across the composite trackers; reduce or eliminate redundant, spurious, and broken tracks; better maintain tracks throughout the missile flight; and deal with out-of-sequence measurements. In addition, the PMHT will demonstrate its superior processing speed.