9319412 Canny The research addresses continuing exploration of sensing and manipulation strategies for simple robot hardware, in the RISC (Reduced Intricacy in Sensing and Control) paradigm. RISC robotics, advocated by the author and Prof. Ken Goldberg of USC, is an attempt to fuse automation and robotic technologies. It uses traditional automation hardware such as parallel-jaw grippers and optical beam sensors, together with geometric planning and sensing algorithms, to accomplish more cost-effective and reliable operations that are also easy to set up and reconfigure, and flexible enough to support small batch sizes and rapid changes in part design needed in future flexible/agile manufacturing systems. The research will address model-building algorithms for optical beam and other sensors, accurate model acquisition, high-rate recognition, grasping and fixturing with minimal numbers of elements, and automatic design of RISC workcell elements. The effectiveness of the RISC approach will be tested with some challenging case studies, starting with an assembly program for a model-aircraft engine.
