Efficient Laser-Based 3D SLAM in Real Time for Coal Mine Rescue Robots

摘要

The accurate description of laneway space with self-localization is a key issue when coal mine rescue robots (CMRRs) perform post-disaster exploration and rescue missions. 3D Simultaneous Localization and Mapping (SLAM) is an effective but time-critical with highly challenging task in semi-structured laneway environment. In this paper, we propose a novel real-time 3D SLAM based on Normal Distributed Transform (NDT) which also take pose graph optimization and loop closure to further improve the consistency of mapping. We innovatively extract the floors and walls as plane nodes to construct constraints, in addition to applying pose nodes from the lidar odometry. Edges of the graph are constructed by the observation constraints between pose nodes and plain nodes. A simple and effective loop detection method is used via odometry-based in conjunction with appearance-based approach to build a global consistent map. The proposed method has been evaluated on the public dataset KITTI and field tests in a simulated underground coal mine. The results indicate that our algorithm can achieve lower computational complexity and lower drift in well-conditioned scenes and degenerate scenes, which can provide pose estimation and environment description for CMRRs to realize remote control assistance and automatic navigation in the coal mine rescue missions.