Multi-Frequency Radar Discrimination

Toyon Research Corporation and Lockheed-Martin propose to develop discrimination algorithms that process any number of measurements from X-band and/or S-band radars to discriminate countermeasures and other objects from reentry vehicles (RVs). Our approach makes use of statistical models that have been developed for the effective classification of ground targets and aircraft using high-range-resolution radar profiles of the targets. Using this framework, we will develop new models that are trained using X-band and S-band signatures of RVs and other objects. We will also develop fusion algorithms that can combine evidence from multiple measurements from either radar to arrive at effective discrimination decisions. To develop prototype algorithms in Phase I, we will use data that is available from the government, at Lockheed-Martin, or at Toyon. We will process a simple dataset and train our discrimination algorithms. We will evaluate the algorithms using a sequestered set of the data to demonstrate the feasibility of our approach. The evaluation will include a quantitative estimate of our performance given varying numbers of measurements from each radar type. If the algorithms show promise, we will plan for a Phase II effort.