Towards High-Performance Solid-State-LiDAR-Inertial Odometry and Mapping

We present a novel tightly-coupled LiDAR-inertial odometry and mapping scheme for both solid-state mechanical LiDARs. As frontend, feature-based lightweight LiDAR provides fast motion estimates adaptive keyframe selection. backend, hierarchical keyframe-based sliding window optimization is performed through marginalization directly fusing IMU measurements. For the Livox Horizon, newly released LiDAR, feature extraction method proposed to handle its irregular scan pattern during preprocessing. LiLi-OM (Livox mapping) real-time capable achieves superior accuracy over state-of-the-art systems types on public data sets of LiDARs in experiments using Horizon. Source code recorded experimental are available at https://github.com/KIT-ISAS/lili-om.