Study On Geometric Map Building And Robot Localization In Unknown Indoor Environments

The Autonomous Mobile Robot has the ability to memorize reason, make decision and take action. As one technology of robotics, navigation is the key point which makes a robot become a truly intelligent and autonomous robot. Navigation is composed of three basic problems and the Simultaneous Localization and Mapping (SLAM) is the precondition of the others. This thesis mainly focuses on the SLAM problem based on uncertain environmental information. The aim is to make the robot have the ability of autonomous map development and location identification in unknown structural indoor environments.Firstly, this thesis introduces the history of robot and the three research stages of SLAM, indicating the practicality and academic value of SLAM.Secondly, the thesis reviews the main solutions of SLAM and points out the difficulties and key points of SLAM. Then, the thesis proposes a two-step decomposition method of building map:①The laser data is divided into several fields using density cluster method and so the map building problem becomes line extraction problem in each field;②Line extraction is realized with least square procedure and lines are matched and renewed according to the double-match rule. Precise localization using Extended Kalman Filter (EKF) is realized.Finally, the experimental results prove that the SLAM system fulfills the design requires completely.