|Acquisition Program: ||PMA-265 F/A-18 E/F|
| ||RESTRICTION ON PERFORMANCE BY FOREIGN NATIONALS: This topic is “ITAR Restricted”. The information and materials provided pursuant to or resulting from this topic are restricted under the International Traffic in Arms Regulations (ITAR), 22 CFR Parts 120-130, which control the export of defense-related material and services, including the export of sensitive technical data. Foreign nationals may perform work under an award resulting from this topic only if they hold the “Permanent Resident Card”, or are designated as “Protected Individuals” as defined by 8 U.S.C. 1324b(a)(3). If a proposal for this topic contains participation by a foreign national who is not in one of the above two categories, the proposal may be rejected.|| Objective: ||Develop a hierarchical aided target recognition (AiTR) algorithm to provide robust real-time automatic recognition of High Range Resolution (HRR) profiles of moving ground targets over tactical depression and aspect angles.
|| Description: ||Stationary relocatable ground targets can employ a variety of camouflage, hide, and signature altering processes to defeat fire control systems. A moving target, while sacrificing many of its deception options, is more difficult to detect and recognize. Radar HRR profiles—one-dimensional (range) representation of target scatterers—provides a standoff all-weather capability for recognition of moving targets. The major drawback of HRR recognition—high pilot workload required for manual recognition—can be reduced by a factor of 100 using AiTR. Based on guidance from Navy pilots, AiTR, to be most effective, must provide a high-confidence decision at some level of recognition (combatant, tank, T-72) within a few seconds. However, current HRR AiTR approaches with heavy reliance on statistical algorithms, measured HRR profiles, and identification level (T-72) decisions have not shown the consistent high confidence decision necessary to realize this reduction in pilot workload for small tank-size ground targets. The hierarchical algorithm sought in this SBIR must provide a methodical search of the target recognition space as well as utilization of the HRR time series for “find x” and general recognition problems for target sets of 20 to 30 ground combatants. Although a top-down divisive approach from detection to classification to recognition to identification is sought, agglomerative techniques that roll up the decision from identification to more general recognition levels will also be considered. Critical to success will be a measure of confidence/uncertainty that will allow the AiTR to decide not only at what recognition level to output the target decision to the pilot, but at what point in the HRR time series it has a high confidence decision. The algorithm must utilize synthetic HRR profiles for template generation or feature sets for statistical classifiers and allow for rapid insertion of new target classes in operational applications. Because of the limited pilot time available, the algorithm must have a high correct decision rate, low false alarm rate, and an “other class” capability that rejects private and commercial vehicles. It must operate against forward-looking tactical radars over as great an aspect and depression angle range as possible. Finally, real-time processing in tactical aircraft processors is required.
|| ||PHASE I: Develop and demonstrate the feasibility of a hierarchical approach to recognize measured HRR signatures of moving ground mobile target signatures. Show performance for different encounter geometries. Include estimates of the CPU and memory requirements for the approach.
|| ||PHASE II: Demonstrate automatic recognition of measured moving and ground target HRR signatures, and provide for unambiguous extrapolation to operational data sets. Present final CPU and memory requirements. Demonstrate:
· Robust correct recognition/ID (80%) for
o Combatant similar confuser false alarm rate of 20% or less
o Other class leakage rate of less than 1%
o All encounter geometries
o Full signatures of 20 range cells on a tank-size target at aspects of 45 degrees on either side of front or back aspects of the target.
|| ||PHASE III: Develop and complete automatic aided target recognition (AiTR) software system or a set of software modules/tools and incorporate into existing/legacy systems and platforms.
|| ||PRIVATE SECTOR COMMERCIAL POTENTIAL: HRR recognition can provide a reduction by a factor of 100 and greater in the manual tasking required for monitoring of security facilities and areas of interest for detection of moving targets that might be used for terrorist purposes. This will provide benefits to the commercial security sector and Homeland Security as well as force protection in hostile areas.
|| References: ||1. HRR profile and general radar technology
2. Donald R. Wehner, High Resolution Radar. Norwood, MA: Artech House., 1987.
3. Jerry L Eaves, Edward K. Reedy, Principles of Modern Radar. New York, NY: Van Nostrand Reinhold Company Inc., 1987.
4. R. O. Duda, Peter E. Hart, David G. Stork, Pattern Classification. New York, NY: John Wiley & Sons, Inc., 2001.
5. Keinosuke Fukunaga, Introduction to Statistical Pattern Recognition Second Edition. San Diego, CA: Academic Press Inc., 1990
Confidence and uncertainty measures
6. Glen Shafer, A Mathematical Theory of Evidence. Princeton, NJ: Princeton University Press, 1976.
7. Judea Pearl, Probabilistic Reasoning in Intelligent System: Networks of Plausible Inference. San Mateo, CA: Morgan Stanley Publishers, Inc., 1988
8. George J. Flir, Tina A. Folger, Fuzzy Sets, Uncertainty, and Information. Englewood Cliffs, NJ: Prentice Hall, 1988|
|Keywords: ||Real-Time Image Processing; Object/Target Recognition and Identification; Combat Identification (CID); High-Range Resolution (HRR) Automatic Target Recognition (ATR); Template Based Target Identification; Profile Signatures|