Hierarchical locative reasoning for autonomous mobile platforms

Describes a fuzzy-based approach to self localization to support real-time indoor robot navigation. The authors use a hierarchical representation of map and environmental objects that are perceived from sensors. Each perceived object is matched against a map object to obtain an estimate of the robot´s location with respect to the map. The authors use fuzzy techniques to characterize partial matches, and to combine estimates discerned by distinct objects. The authors´ approach has been tested on the SRI mobile robot, Flakey