A ZUPT Aided Initialization Procedure for Tightly-coupled Lidar Inertial Odometry based SLAM System

Simultaneous localization and mapping (SLAM) is an important research topic in unmanned platforms. The SLAM method, coupled with the Inertial Measurement Unit (IMU), has recently received much attention due to its excellent performance. However, the tightly coupled LiDAR and IMU odometry (LIO) will be interfered with by the unknown initial state during system startup, which will lead to system startup failure in severe cases, especially in the case of low-cost sensors. Therefore, this paper presents a Zero Velocity Update (ZUPT) aided initialization procedure to estimate unknown initial states assuming that the vehicle always starts from stationary. The proposed initialization procedure includes two phases: static phase and dynamic phase, and this paper also proposes a Zero Velocity Detector to distinguish the two phases precisely. The proposed system is evaluated on the datasets gathered from the campus zone with low-cost sensors to evaluate the whole system. The results show that our approach can effectively improve robustness and partly improve precision when the system is bootstrapped.