Anchoring is the process of creating and maintaining the correspondence between symbols and percepts that refer to the same physical objects. Anchoring must necessarily be present in any physically embedded system that includes a symbolic component, e.g., an autonomous robot including a symbolic planner. Typically, the solution to this problem was developed on a system by system basis for specific domains, and it was hidden in the code. It is only recently that systematic studies of anchoring as a problem per se have started to be pursued in the field of embodied intelligent systems. In this talk, I will advocate the need for a domain-independent definition of the anchoring problem, and will report our current steps in this direction. I will illustrate these steps by showing a few experiments performed on different mobile robots.
Autonomous robots, Perception, Symbol grounding, Cognitive architectures
S. Coradeschi and A. Saffiotti. (2003)
Special Issue on Perceptual Anchoring,
Robotics and Autonomous Systems, 43(2-3).