Evolutionary Algorithms in Target Acquisition and Sensor Fusion

Summary

In this paper, we present evolutionary algorithm techniques for signal processing. A proprietary evolutionary algorithm (EA) is applied to the task of discriminating stationary “targets” from “clutter” in both a radar and imaging sensor. Three applications are made of the EA. First, the EA is used to derive an algorithm for radar target detection including features and decision logics; second, the EA is used to derive an algorithm for an imaging sensor target detection; and third, the EA is used to derive a fused detector, which combines the information from both sensors to make a better detection decision. After detection, the EA may also be used to develop a fused classifier. The EA used can be categorized as a program induction algorithm. That is, the result of the evolutionary process is one or more computer programs that implements a detection or classification algorithm. A technique for real time adaptation of target acquisition algorithms is also discussed.