Model-based localization for an autonomous mobile robot equipped with sonar sensors

Abstract

Concerns the initial localization problem for a mobile robot with sonar sensors, using consistent languages based on interval-valued fuzzy sets to represent vague preferences. Interval-valued preference structures and techniques for crisp choice are used. A map of the environment is assumed available to the robot. The search procedures proposed are independent of the way the map is constructed, so a module which creates and maintains a map can easily be hooked up with the methods described. Sonar readings are represented by 2D arcs. The width of the arc captures the errors resulting from distance measurements. The localization problem is solved by performing a search on the map, in 2 stages: orientation estimation and position estimation. The search mechanism utilizes the concepts of scaling and interval-valued preference structures to make a choice between possible orientations and/or positions. The proposed procedures are illustrated on several experimental settings. The limitations and advantages of the procedures are discussed.