OR-LIM: Observability-aware robust LiDAR-Inertial-Mapping under High Dynamic Sensor Motion
Abstract LiDAR-based Simultaneous Localization And Mapping (SLAM) has been an impactful way for the reconstruction of environmental map. Yet, consistent mapping in sensing-degenerated and perceptually-limited scenes (e.g. multi-story buildings) or under high dynamic sensor motion (e.g. rotating platform) is still challenging. In this paper, we present a novel LiDAR-inertial-mapping (LIM) system with exceptional observability. At its core, it combines a robust real-time LiDAR-inertial-odometry (LIO) module with an efficient surfel-map-smoothing (SMS) module that seamlessly optimize the sensor poses and scene geometry in the meantime. The planar surfels are hierarchically generated and growed from point cloud map to provide reliable correspondences for the fixed-lag optimization. Moreover, the normals of surfels are analyzed for the observability evaluation of each frame. To maintain the global consistency, factor graph is utilized integrating the information from IMU propagation, LIO as well as the SMS. The robustness is extensively tested on the datasets collected by a low-cost multi-beam LiDAR (MBL) mounted on a rotating platform. The experiments conducted on complex multi-story building and large-scale outdoor scenes with various settings of sensor motion have shown the superior performance of our system over multiple state-of-the-art methods.
Demo videos at https://www.youtube.com/watch?v=R8B6AtgdH8E