A New Approach to Environment Representation with Its Applications in Robot Navigation

High-precision mapping and localization contribute to mobile robot navigation. It also provides convenience, robustness and safety for navigation. But for the single feature and large-scale scene, to build high-precision mapping and localization has great challenges and difficulties. Meanwhile, high precise is unnecessary for simple task. Therefore, this paper proposes a new methods of environment representation to achieve the presentation of the long corridor environment. Based on this environment representation, the dynamic path planning and state transformation approach are introduced. Finally we construct the robot navigation system and use Gazebo which is a simulation platform in ROS system to implementation. Extensive experiments tested on Gazebo demonstrate that our method reaches the state-of-the-art.