# Long-Term Urban Vehicle Localization Using Pole Landmarks Extracted from 3-D Lidar Scans

**Authors:** Alexander Schaefer, Daniel Büscher, Johan Vertens, Lukas Luft, Wolfram Burgard  
**arXiv:** [1910.10550](https://arxiv.org/abs/1910.10550)  
**Date:** 2019-10-23  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/1910.10550v1/

## Key points

- Paper presents a mapping and localization system using pole-like objects as landmarks for vehicle localization in urban environments
- Pole-like objects are highlighted for their ubiquity and long-term stability
- System consists of a pole detector, mapping module, and online localization module
- Open-source implementation of the system is available on GitHub
- Extensive experiments conducted over 15 months in an urban area to evaluate performance
- System successfully handled varying routes, weather conditions, seasonal changes, and construction sites
- Proposed method outperformed a recently published method in terms of accuracy
- Robust and effective solution for long-term urban vehicle localization using pole landmarks
- Improved reliability and accuracy compared to existing methods

## Summary

The paper titled "Long-Term Urban Vehicle Localization Using Pole Landmarks Extracted from 3-D Lidar Scans" presents a comprehensive mapping and localization system that utilizes pole-like objects as landmarks for vehicle localization in urban environments. The authors highlight the ubiquity and long-term stability of pole-like objects, making them ideal for this purpose. The proposed approach consists of three main components: a novel pole detector, a mapping module, and an online localization module. Each component is described in detail, and the authors provide an open-source implementation of their system on GitHub. To evaluate the performance of their method, extensive experiments were conducted. First, the system was tasked with localizing a mobile robot over a period of 15 months in an urban area using an initial map. The system successfully handled constantly varying routes, differing weather conditions, seasonal changes, and construction sites, demonstrating its reliability for long-term operation. Furthermore, the authors compared their approach to a recently published method in terms of accuracy. The proposed method clearly outperformed the existing method, showcasing its superiority in accurately localizing vehicles based on pole landmarks extracted from 3-D lidar scans. Overall, this paper presents a robust and effective solution for long-term urban vehicle localization using pole landmarks which has improved reliability and accuracy compared to existing methods. The availability of an open-source implementation further enhances the accessibility and potential impact of this research.