Mobile robot navigation in enclosed large-scale space

In a large-scale s ace, navigation may occur among very dispersed landmarks, further apart than the range of sensing of an autonomous vehicle. In this work we investigate the feasibility of construction of a landmark-based cognitive map, whose elements are the obstacles perceived by a robotic vehicle during exploration of an unknown, large-scale environment. This cognitive map can then be used as an aid for goal-oriented navigation in such a challenging environment. A map construction algorithm is described suitable for a mobile robot with the ability of temporarily marking a single location in an enclosed environment containing polygonal objects. The algorithm is being verifies with a LEGO-Technic-based autonomous vehicle, equipped with a 2 d.0.f. arm and relying on inaccurate odometric and short-range proximity sensing . In spite of its limited and inaccurate internal and external sensing abilities basic skills experimentally demonstrated by the robot include pick-and-place of a portable marker, obstacle detection, as well as characterization and recognition of polygonal objects. These skills, in conjunction with the approximate odometric measurements collected by the vehicle, also represent the repertoire of behaviors exploited in map-assisted navigation.