Hyperspectral and Persistent Sensor Signal Processing Platform and Algorithms

CG2’s goal on this SBIR will be to provide greater capacity for on-board, localized sensor data processing on remote platforms such as UAVs. For this SBIR, CG2 will develop an “embedded avionic HPC” (EA-HPC) system that will include on-board GPU accelerated sensor processing tracking algorithms on an embedded hardware platform derived from COTS HPC components. The EA-HPC will be developed using a multi-stage approach from Phase I proof of concept (TRL 3), through in-lab validation of the prototype (TRL 4 & 5), ground-based in-field utiliza-tion (TRL 6) and finally on-board in-flight utilization (TRL 7). This will allow the Air Force to establish and utilize the benefits of this system at multiple points in the development cycle; first to the range, and ultimately to the battlefield. For this SBIR, CG2 will team with Numerica Corp, (Juan Vasquez, Ph.D) to integrate multi-target tracking algorithms on the EA-HPC. The embedded hardware platform will be derived from COTS HPC components provided by CG2’s parent company, Quantum3D. BENEFIT: The performance boost gained by the EA-HPC will deliver substantial improvements to solu-tions of hyperspectral and persistent surveillance tracking problems. The increase in perfor-mance provided by the EA-HPC on-board system will result in superior utility and translation of the collected sensor input. This will provide many benefits by improving the warfighters’ ability to track and maintain “chain of custody” of multiple moving targets across sensors, spec-tra, weather conditions, time, and from operator to analyst, to decision maker, to intelligence analyst or post-action analyst, as required in the field.