Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Chefs KISS -- Utilizing semantic information in both ICP and SLAM framework

Sven Ochs, Marc Heinrich, Philip Schörner, Marc René Zofka, J. Marius Zöllner

TL;DR

The paper tackles reliable localization for urban autonomous vehicles in GNSS-denied environments using HD maps. It introduces a semantic extension to KISS-ICP for LiDAR odometry and integrates it with Cartographer to enable semantically labeled submaps and loop closure. Key contributions include semantic downsampling that preserves critical objects, label-weighted residuals in ICP, semantic loop closure, and offline filtering of parked vehicles to improve relocalization. Results on SemanticKITTI and KITTI show competitive odometry and substantial reductions in ATE, highlighting improved map accuracy and re-localization robustness for real-world deployment.

Abstract

For utilizing autonomous vehicle in urban areas a reliable localization is needed. Especially when HD maps are used, a precise and repeatable method has to be chosen. Therefore accurate map generation but also re-localization against these maps is necessary. Due to best 3D reconstruction of the surrounding, LiDAR has become a reliable modality for localization. The latest LiDAR odometry estimation are based on iterative closest point (ICP) approaches, namely KISS-ICP and SAGE-ICP. We extend the capabilities of KISS-ICP by incorporating semantic information into the point alignment process using a generalizable approach with minimal parameter tuning. This enhancement allows us to surpass KISS-ICP in terms of absolute trajectory error (ATE), the primary metric for map accuracy. Additionally, we improve the Cartographer mapping framework to handle semantic information. Cartographer facilitates loop closure detection over larger areas, mitigating odometry drift and further enhancing ATE accuracy. By integrating semantic information into the mapping process, we enable the filtering of specific classes, such as parked vehicles, from the resulting map. This filtering improves relocalization quality by addressing temporal changes, such as vehicles being moved.

A Chefs KISS -- Utilizing semantic information in both ICP and SLAM framework

TL;DR

The paper tackles reliable localization for urban autonomous vehicles in GNSS-denied environments using HD maps. It introduces a semantic extension to KISS-ICP for LiDAR odometry and integrates it with Cartographer to enable semantically labeled submaps and loop closure. Key contributions include semantic downsampling that preserves critical objects, label-weighted residuals in ICP, semantic loop closure, and offline filtering of parked vehicles to improve relocalization. Results on SemanticKITTI and KITTI show competitive odometry and substantial reductions in ATE, highlighting improved map accuracy and re-localization robustness for real-world deployment.

Abstract

For utilizing autonomous vehicle in urban areas a reliable localization is needed. Especially when HD maps are used, a precise and repeatable method has to be chosen. Therefore accurate map generation but also re-localization against these maps is necessary. Due to best 3D reconstruction of the surrounding, LiDAR has become a reliable modality for localization. The latest LiDAR odometry estimation are based on iterative closest point (ICP) approaches, namely KISS-ICP and SAGE-ICP. We extend the capabilities of KISS-ICP by incorporating semantic information into the point alignment process using a generalizable approach with minimal parameter tuning. This enhancement allows us to surpass KISS-ICP in terms of absolute trajectory error (ATE), the primary metric for map accuracy. Additionally, we improve the Cartographer mapping framework to handle semantic information. Cartographer facilitates loop closure detection over larger areas, mitigating odometry drift and further enhancing ATE accuracy. By integrating semantic information into the mapping process, we enable the filtering of specific classes, such as parked vehicles, from the resulting map. This filtering improves relocalization quality by addressing temporal changes, such as vehicles being moved.

Paper Structure

This paper contains 5 sections, 2 equations, 2 figures, 3 tables.

Table of Contents

  1. INTRODUCTION
  2. RELATED WORK
  3. METHOD
  4. EVALUATION
  5. CONCLUSION

Figures (2)

  • Figure 1: The main feature of this paper is the improvement of the Cartographer framework utilizing a semantic adaptation of the KISS-ICP vizzo_kiss-icp_2023 approach. Due to this reason we outperform the state-of-the-art and also add the capabilities of Cartographer. This includes map manipulation as seen in \ref{['fig:wd_zoom']} and \ref{['fig:wd_zoom_no-car']}. The backgrounds are desaturated for better visualization.
  • Figure 2: Exmaple trajectories of the KITTI odometry dataset. The graphics are generated utilizing grupp_evo_2017.