Machine intelligence applied to radar image understanding

Concepts from artificial machine intelligence are applied to range-Doppler images of simple space objects. The relevant features to be extracted from the radar data include the position and type of all scattering centers in body coordinates as well as the motion parameters for the object. The goal is to produce a representation of the imaged 3-D object that is appropriate for recognition. The proposed approach combines the power of classical methods with the flexible representations and control structures afforded by the field of machine intelligence. The authors have built a recognition system using signal processing primitives to extract features, semantic model building and matching, and a control structure that is based on a blackboard architecture. Performance evaluation results for the object modeling system are presented