Work

Cross-Level Degeneracy Perception for Robust LiDAR Localization

Year: 2026

Type: article

Abstract: This work-in-progress extends LIO-SAM with cross-level degeneracy per-ception (CLDP). Geometric and scan-matching weak directions are jointly verifiedand temporally filtered into three reliability states that regulate directional dampingand LiDAR-factor covariance. Tests on five vehicle and two GEODE sequencesshow reduced drift, including a tunnel2 RMSE reduction from 17.01 to 7.29 magainst its best baseline. Additional independent diagnostics examine the relationbetween scene normals and runtime registration information.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.24053196

Authors: sun, yang, meng, xiaolin, wang, qing

Institution: Southeast University

Language: English

Topic: Robotics and Sensor-Based Localization

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7225632094

FWCI:

Cites:

Cited by:

Related to: