Abstract
An effective memory structure for an intelligent robot and a flexible route-finding algorithm based on the memory structure are discussed in this paper. The contents are as follows.
(1) A stratified memory structure composed of several strata, one overlapping slightly with another, has good flexibility.
(2) An English-like language restricted in a reasonable manner is adopted as the input language, which can be translated into the inner form suitable for a robot.
(3) A route-finding algorithm that makes use of the memory structure mentioned above can make fairly effective search of the route even in the case that a robot has ambiguous information on its environments.
These memory structures, the interpretation procedure and the route-finding algorithm are skillfully combined in a robot system, and the inferential ability to take off ambiguities in input and stored information is brought forth as the result.