Multi-Resolution Tracking for Fire-Control

Data Fusion Corporation (DFC), Wright State University, and Unisys Systems Corporation will jointly develop a computationally scalable multiresolution approach for air-to-air tracking and data association. This approach features a wavelet-based multiresolution approach to track estimation. This multiresolution approach permits dynamic regulation of computational burden through control of both spatial and temporal resolution. It includes an interacting multiple-mode approach for tracking maneuvering targets. This tracking algorithm will be combined with a multiresolution version of a data association algorithm based on mean-field, event-averaged, maximum-likelihood estimation. The mean-field theory approach to data association avoids the combinatoric costs of most permutation-based association algorithms such as MHT and JPDA. This work also considers how these methods can be controlled by DFC's sensor manager. The goal of the work is computationally tractable algorithms for air-to-air tracking of low-observable targets in clutter.