Navy Applications of 4th Generation Deeply Coupled Computing Architectures Navy STTR FY2006 Sol No.: Navy STTR FY2006 Topic No.: N06-T005 Topic Title: Navy Applications of 4th Generation Deeply Coupled Computing Architectures Proposal No.: N064-005-0422 Firm: CPU Technology, Inc. 5731 W. Las Positas Boulevard Pleasanton, California 94588 Contact: Mark Scheitrum Phone: (925) 224-9920 Web Site: www.cputech.com Abstract: Warfare is changing; there is a need for airborne and space-based sensor systems to detect small, highly maneuverable targets against a strong clutter background in the presence of jamming. Space-Time Adaptive Processing (STAP) is a new signal processing technique for advanced radar systems that allows for performance enhancements over conventional approaches. Current power, weight, and size constraints make the real-time implementation of full degrees-of-freedom STAP techniques on airborne/spaceborne platforms impractical. CPU Technology and University of Maryland Institute for Systems Research propose to investigate and demonstrate the advantages of applying 4th generation deeply coupled programmable SOC technology, called the QX4, and other adaptive hardware elements to provide airborne platforms with the computing capacity required to perform real-time STAP techniques. It provides 8.5 GFlops of 64-bit performance at only 8 watts of power. The adaptive hardware will utilize floating gate technology to accelerate STAP radar processing. We will code a STAP algorithm and execute the software on a single x86 microprocessor, a QX4 array, and a QX4 array with adaptive hardware to measure their relative performance. Other metrics will include size, weight, power and accuracy. We will also investigate adding STAP cost-effectively to existing radar processors. Benefits: Space-Time Adaptive Processing (STAP) enhances the ability of radars to detect targets that might otherwise be obscured by clutter or by jamming using a two-dimensional filtering technique. This capability enables a radar to see items under vegetation and provides much improved information in urban scenarios where clutter is extremely high and the targets are small. Our goal is to make a high speed real-time STAP radar processor small enough and with minimal power so that STAP can be used by traditional airborne sensor platforms such as the E-2C as well as smaller aircraft such as fighters, helicopters and UAVs. It could potentially be used for the larger missiles as well. Multidimensional processing such as found with STAP is used in many applications including mobile communications, video processing, medical imaging and radar processing. Sensor fusion and autonomous operation could also potentially benefit from the compact performance throughput of QX4 and adaptive hardware techniques. Return