Tightly-coupled Lidar-inertial Odometry and Mapping in Real Time

We proposed a tightly-coupled lidar-inertial odometry and mapping method for real-time 6DoF state estimation in complex environments e.g. unexplored, full of obstacles or GPS denied. The proposed lidar-inertial odometry is a tightly coupled, nonlinear optimization-based method, fusing pre-integrated IMU measurements, feature points and ground plane extracted from lidar data. High-frequency state estimation is obtained by the proposed odometry. Further, a low-frequency lidar mapping method with scan-to-map refinement and global pose-graph optimization is proposed for more accurate pose estimation and globally consistent map. Real-world experiments are performed in indoor and outdoor scenarios to validate accuacy and generalizability of our method and compare against state-of-art real time lidar odometry and mapping methods.