MUltiSensor Target Recognition System (MUSTRS)

The paper describes the Advanced Research Projects Agency (ARPA) MUltiSensor Target Recognition System (MUSTRS). A smart sensor manager controls forward-looking infrared (FLIR) and millimeter wave (MMW) radar sensors to obtain multiple looks at targets on the ground. Targets in IR images are recognized using a variation on minimum average correlation energy (MACE) filtering and/or a model-based algorithm called key features. Radar data are processed using quadratic distance composite filtering techniques. Evidence is combined using a Bayesian method. The system has been designed to correctly classify time critical mobile targets with very low false alarm rates.<>