Complementary Observation, Long Duration Track Fusion (COLD Track Fusion)
Small Business Information
6 New England Executive Park, Burlington, MA, 01803
Director of Cognitive Fus
Director of Cognitive Fus
AbstractMaintaining continuous track of high-value ground targets in complex environments using only a single sensing modality such as radar, presents a number of challenges for which the use of complementary sensors offers a potential solution. Even withsimultaneous radar modalities such as ground moving target indicator (GMTI) and high-resolution range profiles (HRR), there are a number of common scenarios in which the track of a specific target can be lost and subsequent reacquisition of track isdifficult. A significant problem is targets moving in and out of hide sites, where track is lost and identification of targets emerging from hide is uncertain. We propose that the fusion of complementary sensing by GMTI/HRR-radar and MSI/HSI spectralsensors have the potential to overcome these limitations and support long-duration tracking and reacquisition of targets emerging from hide. The proposed effort will explore the many synergies between GMTI tracking and HSI learning & recognition of targetspectral signatures. Tracking can be aided by HSI & HRR target features, and context extracted using spectral data mining. Target learning can be assisted by tracking (in conjunction with spectral anomaly detection) to localize the target of interest in asmall moving area. We will demonstrate the feasibility of such strategies using an existing multi-modality (EO, HSI) 3D site model of Mobile, Alabama, and embedding moving targets (possessing HSI and HRR signatures, and GMTI detections from a simulatedradar sensor). We will utilize our existing software for multi-hypothesis tracking together with our neural assisted target learning & recognition (i.e., data mining) system, to demonstrate the potential of fused GMTI/HRR/HSI tracking of targets in and outof hide. We anticipate that this approach to multi-modality fused tracking will enable long duration target tracking and reacquisition of lost tracks as targets emerge from hide. We expect this study will also reveal significant issues in distributedsensor resource management. Success in Phase I with demonstration of feasibility, will lead to a Phase II effort in which a prototype integrated tracking & target learning/search system is developed and applied to data sets of interest. This integratedGMTI/HSI tracker will find use in military battlefield systems, homeland defense systems for monitoring critical transportation infrastructure, traffic monitoring systems, drug interdiction, and vehicle pollution monitoring (ground and maritime).
* information listed above is at the time of submission.