A problem in the development of intelligent robots is that efficient computational and representational methodologies have not been available for emulating knowledge and expectation driven behavior so basic to human cognition and problem solving. Even when techniques such as geometric modeling are used for representing objects in the robot world, methods for linking such representations with sensory feedback do not exist. The authors propose the use of intermediate representations called sensor-tuned representations for linking CSG-based solid modeling with sensory information. It is pointed out how object recognition can be done with sensor-tuned representations. Results of manipulation experiments produced by the current implementation of the system are reported.
|Title of host publication||Unknown Host Publication Title|
|Number of pages||10|
|Publication status||Published - 1987 Jan 1|