Subspace Detection of Surprise Events for the BMC2 Architecture

We propose to develop subspace classifiers for detecting and monitoring dynamic and surprise events (e.g. missile launches warhead and decoy deployment, artillery fire, etc.) from a remote platform. The algorithms will be tailored to the Advanced Spectral Sensor developed by SSSC, which is capable of adaptively measuring the spectra of surprise events anywhere within a large field of view at greater than video rates. This sensor is also capable of accurately locating the event with little impact on data bandwidth. The classifiers will be based both on the existing library of surprise event data and on data collection opportunities that arise during the course of the contract. The classifiers will exploit target specific spectral-temporal features and also be tailored to support the unique aspects of the Advanced Spectral Sensor. Information about the target subspace will be incorporated into the hardware implementation to create an advanced sensor system that can be integrated into the next-generation MDA BMC2 architecture.