PATTERN RECOGNITION TECHNIQUES FOR EARLY WARNING RADAR DISCRIMINATION

The project objective is to examine descrimination techniques that can be implemented without modifying EWR hardware or software. Pattern recognition techniques is one approach. With approximately 100 seconds of Radar Cross Section (RCS) vs. time data, a trained observer can generally discriminate between Reentry Vehicle (RV) and tank like signatures when the data is displayed over time because the signature of a tank results in a visual pattern that is very different from that of a RV. We have a large set of actual EWR RCS data on a wide variety of recognition code against real world data. In phaseI, different ways of representing the data will be analyzed and potential pattern recognition algorithms will be designed, coded and tested. In phase II, code will be implemented in an existing EWT site.