A PERCEPTION SYSTEM FOR OBJECT RECOGNITION, ACQUISITION AND TRACKING IN CLUTTERED ENVIRONMENTS

HUMAN RESOURCE EFFICIENCY IN SPACE OPERATIONS CAN BE SIGNIFICANTLY IMPROVED WITH INTELLIGENT SYSTEMS WHICH CAN PERFORM INTRICATE OR MUNDANE TASKS WITH MINIMAL HUMAN INTERACTION WHILE MAINTAINING A FLEXIBLE MAN-HACHINE INTERFACE. THIS PROJECT PROPOSES THE DEVELOPMENT OF A SOPHISTICATED PERCEPTION SYSTEM THAT CAN BE INTEGRATED WITH A VARIETY OF IMAGING SENSOR SYSTEMS AND IS CAPABLE OF AUTONOMOUSLY RECOGNIZING OBJECTS FROM ARBITRARY (3D) PERSPECTIVES. IN ADDITION, THE PROPOSED SYSTEM WILL BE CAPABLE OF AUTONOMOUS (OR OPERATOR DESIGNATED) OBJECT ACQUISITION AND TRACKING IN CLUTTERED ENVIRONMENTS AGAINST STATIONARY OR MOVING OBJECTS. THE PHASE I EFFORT WILL ESTABLISH THE PRELIMINARY REQUIREMENTS FOR THE SYSTEM AND WILL CONCENTRATE ON THE DESIGN OF THE CRITICAL ELEMENTS OF THE PERCEPTION SYSTEM WHICH INCLUDES OBJECT RECOGNITION, ACQUISITION, TRACK AND OPERATOR INTERFACE. THE MAJORITY OF THE EFFORT WILL BE DEVOTED TO DEVELOPING A SIMULATION WHICH WILL DEMONSTRATE THE KEY ASPECTS OF THE SYSTEM AGAINST ARTIFICALLY GENERATED IMAGE DATA. IT IS ANTICIPATED THAT THE SIMULATION WILL VALIDATE THE BASIC BUILDING BLOCKS OF THE PROPOSED SYSTEM AND BOUND THE SYSTEM'S OVERALL PERFORMANCE. THE PROPOSED SYSTEM HAS APPLICATION TO AREAS SUCH AS AUTOMATED MANUFACTURING (E.G. ASSEMBLY AND QUALITY CONTROL) AND SECURITY SYSTEMS.