LiDAR SLAM Utilizing Normal Distribution Transform and Measurement Consensus

Ashwin Vivek Kanhere and Grace Xingxin Gao

Abstract:	Parametric point cloud approximations, such as the Normal Distribution Transform (NDT), are used for LiDAR navigation that requires compact maps. However, the state-of-the-art in these algorithms does not provide a reliability assessment. In addition, the computational efficiency needs to be improved for real-time applications. We propose a consensus-NDT Simultaneous Localization and Mapping (SLAM) framework, which utilizes measurement consensus, to increase computational efficiency and estimate localization reliability. In our approach, measurement consensus is quantified using a two-tiered method. In the first tier, a voxel consensus metric evaluates feature-level measurement consensus. In the second tier, voxel consensus metrics for all features are combined to yield the localization consensus metric, which estimates localization reliability. By selecting high measurement consensus voxels, computational efficiency is improved during the LiDAR odometry and map update steps. We perform experiments with real-world data which show that the proposed framework improves localization computation efficiency without compromising accuracy as compared to na?ve NDT SLAM. Additional experiments show that the localization consensus metric is a valid metric for estimating localization reliability.
Published in:	Proceedings of the 32nd International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2019) September 16 - 20, 2019 Hyatt Regency Miami Miami, Florida
Pages:	2228 - 2240
Cite this article:	Kanhere, Ashwin Vivek, Gao, Grace Xingxin, "LiDAR SLAM Utilizing Normal Distribution Transform and Measurement Consensus," Proceedings of the 32nd International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2019), Miami, Florida, September 2019, pp. 2228-2240. https://doi.org/10.33012/2019.16954
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In