Obstacle Perception And Obstacle-Aviding Strategy Research Of Mobile Robot Based On Laser Range Finder

Environment perception and autonomous obstacle-avoiding are two key technologies in the field of mobile robot and are the basis of other advanced task. They still need upgrading although they are based on extensive research on various sensors. This paper relies on LMS2000 laser rangefinder made by SICK Corporation of German and TJ-1 mobile robot to explore the environment perception and autonomous obstacle-avoiding. profound research is done in theory and practice.Firstly, this paper establishes the coordination system module, laser-range finding module and analyzes the error of laser-rangefinder systematically on the TJ-1 mobile robot platform. The paper also deduces the relationship among the three coordination systems in mobile robot navigation, which realizes the mutual data transformation in robot obstacle-avoiding and regulating. In order to decrease data storage and increase the system's dynamic performance, environment characteristic data are utilized to substitute original laser data. The paper stores the environment data in the form of trajectories'vector direction other than in the form of grid, which is an improvement of grid method. We put forward a simplified environment modeling method: polar coordinate vector method. This paper extracts environment obstacles by identifying obstacle groups which is proper to robot navigation. Under the theory of rolling window we can get obstacle groups by identifying obstacles in real-time window and merging obstacle, the distance between whom are less than the robot width.Secondly, obstacle-avoiding strategies are studied. Control polices comprising emergency avoiding, precise avoiding, fuzzy avoiding and dynamic avoiding are designed and it can percept and adjust regulating policies autonomously to adapt vicarious situation with the environment changing. Every policy is based on polar spaces, establishes rolling window and detecting local environment information via sensors. Sub-goal point, which is figured out in the local regulating window, will be used as temporary destination to avoid obstacles when reaching actual goal.The obstacle-perception and obstacle-avoiding strategy are described in VC++ language. Finally we have experiments on the TJ-1 platform including single obstacle, multi-obstacles and dynamic obstacle avoiding. The experiment results prove that the robot can pass around obstacles smoothly and certifies the validity of obstacle perception and avoiding.