iG-LIO: An Incremental GICP-Based Tightly-Coupled LiDAR-Inertial Odometry
Zijie Chen, Yong Xu, Shenghai Yuan, Lihua Xie
Abstract
This work proposes an incremental Generalized Iterative Closest Point (GICP) based tightly-coupled LiDAR-inertial odometry (LIO), iG-LIO, which integrates the GICP constraints and inertial constraints into a unified estimation framework. iG-LIO uses a voxel-based surface covariance estimator to estimate the surface covariances of scans, and utilizes an incremental voxel map to represent the probabilistic models of surrounding environments. These methods successfully reduce the time consumption of the covariance estimation, nearest neighbor search, and map management. Extensive datasets collected from mechanical LiDARs and solid-state LiDARs are employed to evaluate the efficiency and accuracy of the proposed LIO. Even though iG-LIO keeps identical parameters across all datasets, the results show that it is more efficient than Faster-LIO while maintaining comparable accuracy with state-of-the-art LIO systems. The source code for iG-LIO has been open-sourced on GitHub: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/zijiechenrobotics/ig_lio</uri> .