journal

Point Wise or Feature Wise? A Benchmark Comparison of Publicly Available Lidar Odometry Algorithms in Urban Canyons

Feng Huang, Weisong Wen, Jiachen Zhang, Li-Ta Hsu

IEEE Intelligent Transportation Systems Magazine · Nov 2022

Abstract

Robust and precise localization is es-sential for an autonomous system with navigation requirements. Lidar odometry (LO) has been extensively studied in the past decades to realize this goal. Satisfactory accuracy can be achieved in scenarios with abundant environmental features using exist-ing LO algorithms. Unfortunately, the performance of the LO is significantly degraded in urban canyons with numerous dynamic objects and complex environmental structures. Meanwhile, it is still not clear from the existing literature which LO algorithms perform well in such challenging environments. To fill this gap, this article evaluates an array of popular and extensively studied LO pipelines using the data sets collected in urban canyons of Hong Kong. We present the results in terms of their positioning accuracy and computational efficiency. The three major factors dominating the performance of LO in urban canyons are concluded, including the ego-vehicle dynamic, moving objects, and the degree of urbanization. According to our experiment results, point wise accomplishes better accuracy in urban canyons while feature-wise achieves cost-efficiency and satisfactory positioning accuracy. Authorized licensed use limited to: Hong Kong Polytechnic University. Downloaded on December 09,2024 at 10:46:50 UTC from IEEE Xplore. Restrictions apply.

Problem Addressed

LiDAR odometry performance degrades significantly in urban canyons due to numerous dynamic objects and complex environmental structures, yet existing literature does not make clear which publicly available LO algorithms perform well in such challenging environments.

Key Contributions

Comprehensive benchmark of publicly available LiDAR odometry pipelines on urban-canyon datasets from Hong Kong
Comparison in terms of positioning accuracy and computational efficiency
Identification of three dominant degradation factors: ego-vehicle dynamics, moving objects, and degree of urbanization

Conclusions

The benchmark concludes that point-wise methods achieve better accuracy in urban canyons, while feature-wise methods offer computational efficiency with satisfactory positioning accuracy. The performance of LO in urban canyons is dominated by the ego-vehicle dynamics, moving objects, and the degree of urbanization.