Contribution

A Trajectory Consistency Metric for GNSS Anomaly Detection with LiDAR Odometry

---

Abstract

While Global Navigation Satellite System (GNSS)-based robot localization is successful in open scenarios, it quickly becomes unreliable in GNSS-degraded environments such as forests. With the increasing interest in using autonomous robots in forestry, it becomes more important to have reliable localization in forest environments, which are among the most challenging areas for GNSS-based localization. Having an estimate for the quality of the localization can help achieve this. While GNSS receivers provide uncertainty estimates based on signal characteristics and the satellites’ constellation, practical experience shows that these values are less meaningful in forests. This paper presents an error metric that exploits the properties of commonly used robot localization setups to assess the quality of the localization. This assessment is based on a comparison between a LiDAR odometry-based local trajectory estimate and a GNSS-based global trajectory estimate in their respective coordinate systems. A qualitative Analysis shows that the metric enables meaningful Statements about the quality of position estimates derived from GNSS measurements in the global coordinate system

Keywords: anomaly detection, GNSS, LiDAR odometry

How to Cite:

Wipfler, H. & Steinbauer-Wagner, G., (2025) “A Trajectory Consistency Metric for GNSS Anomaly Detection with LiDAR Odometry”, ARW Proceedings 25(1), 127-128. doi: https://doi.org/10.34749/3061-0710.2025.25