Publications / 2024 Proceedings of the 41st ISARC, Lille, France
Accurate localization plays a crucial role in the effective operation of autonomous robotics systems, especially in dynamic environments such as construction sites. Simultaneous Localization and Mapping (SLAM) utilizing LiDAR sensors has emerged as a popular solution due to its ability to function without external infrastructure. However, existing algorithms exhibit significant shortcomings. Despite current methods achieving high accuracy over long trajectories, they struggle with precision and reliability in complex indoor environments. This paper introduces a novel feature-based LiDAR SLAM system designed to address these limitations and enhance short-term precision and overall robustness. The proposed system is evaluated using both existing datasets and a physical robotics platform, addressing the limitations of current implementations and showcasing improved performance in challenging real-world scenarios, particularly in construction environments.