Neuromorphic Sensor for Missile Defense

Proposal seeks to build and test a Persistent wide field-of-view (FOV) infrared (IR) camera array for detecting and tracking dim objects. The optical design is based on the successful Tau PANDORA (visible) array systems, now in use for tracking low earth orbit (LEO) and geostationary earth orbit (GEO) satellites. In Phase I, propose to collect daytime shortwave IR (SWIR) imagery on satellites with a single camera. Analyzed this data to demonstrate quantitatively the successful use of Convolutional Neural Networks (CNNs) to identify satellites distinct from the clutter. Tau proposes to construct a prototype sensor array to demonstrate the feasibility of utilizing Neuromorphic Machine Learning (NML). The primary research in Phase II will be directed towards developing an effective training algorithm of Spiking Neural Networks (SNN), an advanced Neuromorphic network, while we continue on present path by improving the Generation II Neural Networks previously investigated. Approved for Public Release | 22-MDA-11215 (27 Jul 22)