Building geometric feature based maps for indoor service robots

This paper presents an efficient geometric approach to the Simultaneous Localization and Mapping problem based on an Extended Kalman Filter. The map representation and building process is formulated, fully implemented and successfully experimented in different indoor environments with different robots. The use of orthogonal shape constraints is proposed to deal with the inconsistency of the estimation. Built maps are successfully used for the navigation of two different service robots: an interactive tour guide robot, and an assistive walking aid for the frail elderly.