Development of Novel Knowledge-Based Space-Time Adaptive Processing (STAP) Techniques Using Support Vector Machines (SVMs) Navy STTR FY2005 Sol No.: Navy STTR FY2005 Topic No.: N05-T013 Topic Title: Development of Novel Knowledge-Based Space-Time Adaptive Processing (STAP) Techniques Using Support Vector Machines (SVMs) Proposal No.: N054-013-0449 Firm: Virtual EM Inc. 2019 Georgetown Blvd Ann Arbor, Michigan 48105-1532 Contact: Tayfun Ozdemir Phone: (734) 222-4558 Web Site: www.virtualem.biz Abstract: Virtual EM Inc. proposes to develop STAP techniques using machine learning algorithms for eventual implementation in critical radar systems. These algorithms require less time to train than neural networks. Virtual EM's simulation packages will be used to train the algorithms for array element failure, multi-path and jamming mitigation as well as interference from ground and maritime clutter. The performance of the algorithms will be evaluated in comparison to conventional reduced order STAP techniques. Benefits: On the military side, Navy's E-2 Advanced Hawkeye program is in critical need for advanced STAP algorithms that can mitigate jamming, multi-path and clutter interference. The technology that will be developed from this SBIR effort will fill this critical need. On the civilian side, MIMO systems are the target. Such systems will soon be common place in WLAN and cellular communication systems. MIMO systems increase range and bandwidth by making use of the multi-path, and the technology from this SBIR project will have direct application in this area. Return