Multi-Phenomenology Discrimination for Feature Aided Data Fusion
We propose to apply a proprietary discrimination technique rooted in the manifold learning literature to discrimination of object type through radar and through electro-optical/infrared sensors, and to use the features computed by this technique to help correlate tracks between sensors. Our discrimination technique is data-type agnostic, meaning that we can apply the same basic algorithm in both phenomenologies, which, in turn, suggests that future work may allow us to more self-consistently perform cross-sensor data fusion. The proposal leverages prior investment by MDA in radar discrimination techniques and is endorsed by a major MDA prime contractor for sensor technologies, increasing its probability of successfully transitioning to the operation BMDS.
Small Business Information at Submission:
Director, Applied Mathematics
DECISIVE ANALYTICS Corporation
1235 South Clark Street Suite 400 Arlington, VA -
Number of Employees: