Consistent Map Building by a mobile robot euipped with stereo sensor and lrf

This paper presents an efficient map building technique for indoor mobile robot navigation based on laser range finder and binocular stereo vision sensors. To effectively incorporate different sensors and deal with measurement uncertainty involved in environment perception, this article presents a local map integration approach in which Bayesian filter based dynamic occupancy grid map modeling techniques are employed. The adopted method is discussed in the context of mobile robot Simultaneous Localization and Map-Building (SLAM). In SLAM routine, the integrated local map is utilized as observation input, and Rao-Blackwellized Particle Filter (RBPF) is used for refining location estimation and generating accurate global map. Advantages of our proposal are validated by real experimental results carried on Pioneer robot.