A Robust Approach for LiDAR-Inertial Odometry Without Sensor-Specific Modeling

Figure 1: Our robust LiDAR-inertial odometry system is directly operational in different environments, sensor configurations, and robotic platforms with distinct motion behaviours, all without any change in configuration or modeling approach. We depict the local map result of our odometry system in four distinct scenarios, shown clockwise from the top left: urban city with Ouster OS1-128 and built-in InvenSense IMU mounted on a car; mixed indoor-outdoor university buildings with Hesai QT64 and Alphasense IMU on a backpack (data from Tao et al. [31]); forest with Hesai XT32 and Xsens MTi-100 IMU mounted on the SAHA tree-harvesting machine (see Jelavic et al. [14]); and parking lot with V elodyne VLP-16 and onboard IMU on a Unitree Go1 quadruped (data from Ou et al. [25]). Abstract-- Accurate odometry is a critical component in a robotic navigation stack, and subsequent modules such as planning and control often rely on an estimate of the robot's motion. Sensor-based odometry approaches should be robust across sensor types and deployable in different target domains, from solid-state LiDARs mounted on cars in urban-driving scenarios to spinning LiDARs on handheld packages used in unstructured natural environments. In this paper, we propose a robust LiDAR-inertial odometry system that does not rely on sensor-specific modeling. Sensor fusion techniques for LiDAR and inertial measurement unit (IMU) data typically integrate IMU data iteratively in a Kalman filter or use pre-integration in a factor graph framework, combined with LiDAR scan matching often exploiting some form of feature extraction. We propose an alternative strategy that only requires a simplified motion model for IMU integration and directly registers LiDAR scans in a scan-to-map approach. Our approach allows us to impose a novel regularization on the LiDAR registration, improving the overall odometry performance. We detail extensive experiments on a number of datasets covering a wide array of commonly used robotic sensors and platforms. We show that our approach works with the exact same configuration in all these scenarios, demonstrating its robustness.