GMM-LIO: Adaptive and Robust LiDAR-Inertial Odometry with Gaussian Mixture Model Voxel Map
Zishun Deng, Can Li, Wanbiao Lin, Lei Sun
AI summary
Problem
Traditional LiDAR-inertial odometry systems degrade in complex or sparse environments due to fixed parameters and map representations that cannot accurately model intersecting surfaces or geometric discontinuities.
Approach
The framework dynamically adjusts voxel resolution and surface covariance based on state uncertainty and local point density, while employing a Gaussian Mixture Model voxel map to capture multi-modal geometries and a robust MAP estimator to suppress outliers.
Key results
- Information-theoretic adaptive voxel resolution and covariance estimation
- Dynamic Gaussian Mixture Model voxel map for complex geometry
- Robust MAP estimator with anisotropic information matrix for outlier suppression
- 36% relative accuracy improvement over leading LIO baselines
Why it matters
Enables reliable autonomous navigation in challenging real-world environments where conventional odometry systems fail due to geometric ambiguity or sensor degradation.
Abstract
Tightly coupled LiDAR–inertial odometry (LIO) systems are critical for autonomous navigation, yet their perfor- mance often degrades due to insufficient adaptability to diverse environments and limitations in map representation. To address these limitations, this paper presents GMM-LIO, a robust and adaptive LIO framework that integrates a novel information- theoretic scan processing module and a high-fidelity Gaussian Mixture Model (GMM) voxel map structure. At its core, GMM- LIO features a two-level adaptive front-end that dynamically modulates voxel resolution based on state uncertainty and ad- justs surface covariance estimation according to local point den- sity on a standard voxel grid. Furthermore, GMM-LIO employs a dynamic Gaussian Mixture Model voxel map to accurately model intersecting surfaces. The entire system is formulated as a robust Maximum a Posteriori (MAP)-based estimator, which employs an Iteratively Reweighted Least Squares (IRLS) solver together with a principled anisotropic information matrix to handle measurement outliers. Extensive evaluations on diverse public and self-collected datasets demonstrate that GMM-LIO achieves state-of-the-art accuracy and robustness, with a 36% relative improvement over leading LIO baselines.