Wide-area Motion Imagery and RF Compressive Sensing Applications Navy STTR FY2012.A Sol No.: Navy STTR FY2012.A Topic No.: N12A-T026 Topic Title: Wide-area Motion Imagery and RF Compressive Sensing Applications Proposal No.: N12A-026-0095 Firm: Invertix Corporation 8201 Greensboro Drive Suite 800 McLean, Virginia 22102 Contact: Brecken Uhl Phone: (575) 646-9316 Web Site: www.invertix.com Abstract: The Global Information Grid (GIG) has access to massive amounts of data from a vast number of sensors and sources. Extracting information from this quantity of sensor data is a significant challenge. Sensor data is often large, which taxes limited-bandwidth GIG interconnections. Current approaches rely almost entirely on end-user analytics, which is inefficient in that it requires raw data to reach the analysts from distant nodes of the GIG before information is extracted. The proposed effort overcomes this issue, both saving limited bandwidth and improving analytic efficiency by extracting information close to the source. We apply concepts of both compressive sensing (CS) and compressive processing (CP) to the information-generating chain that flows from the sensor through the GIG to analytical resources, and then back to the GIG as a product. Objective, quantitative means are used to optimize state-of-the-art CS and CP services deployed at locations throughout the distributed sensor-GIG-analyst network. The result is a Distributed Compressive Sensing and Processing (DCSP) framework with a comprehensive joint application of CS and CP to a large, deployed network. Benefits: The Distributed Compressive Sensing and Processing (DCSP) framework enables maximum use of sensor data and the capabilities of both the GIG and the analysts that interact with it. The use of optimally-compressed data files allows for better use of limited global information grid (GIG) resources (network bandwidth) and enables the deployment of automated detection and classification services throughout the network architecture. With DCSP, analysts are presented with more raw data equivalents (in the form of compressed sensor results) per unit time, while still having the ability to request the original raw data as needed. Furthermore, analytic effectiveness is improved through classification, detection, and other meta-data results from distributed, automated analytical engines. The result is a greater volume of equivalent data processed per unit time, producing more actionable knowledge. The DCSP framework has immediate application to DCGS-N, for surface navy, airborne, and Marines, where bandwidth is limited and out of DCGS control. In addition, local storage, pre-processing, and fusion of sensor data in emerging reference cloud based architectures will increase commercial application beyond Navy to DCGS-A, DCGS-AF, RT-RG, Border security, Homeland Security and many other potential users with similar large data, constrained-network challenges. Return