Tsukada Lab. Computer Network Group
← Publications Int'l conference

Neural Error Covariance Estimation for Precise LIDAR Localization

Minoo Dolatabadi, Fardin Ayar, Ehsan Javanmardi, Manabu Tsukada, Mahdi Javanmardi

International Conference on Intelligent Computing and its Emerging Applications (ICEA 2024), Tokyo, Japan, pp. 21–26, Nov 28, 2024

Summary in three lines

  • LiDAR map matching needs reliable error covariance estimates for sensor fusion.
  • A neural network predicts covariance using a dataset generated for that purpose.
  • In a Kalman-filter evaluation, the localization error improved by 2 cm.

Figures from the paper

Figure 1: Neural error covariance estimation pipeline
Figure 1. Neural error covariance estimation pipelineSource: Figure 1 from Dolatabadi et al., International Conference on Intelligent Computing and its Emerging Applications (ICEA 2024), 2024 (arXiv:2501.02558, authors’ version)The upper pipeline builds training covariance targets from point clouds and matching results. The lower pipeline uses learned features to predict covariance.

Abstract

Autonomous vehicles have gained significant attention due to technological advancements and their potential to transform transportation. A critical challenge in this domain is precise localization, particularly in LiDAR-based map matching, which is prone to errors due to degeneracy in the data. Most sensor fusion techniques, such as the Kalman filter rely on accurate error covariance estimates for each sensor to improve localization accuracy. However, obtaining reliable covariance values for map matching remains a complex task. To address this challenge, we propose a neural network-based framework for predicting localization error covariance in LiDAR map matching. To achieve this, we introduce a novel dataset generation method specifically designed for error covariance estimation. In our evaluation using a Kalman filter, we achieved a 2 cm improvement in localization accuracy, a significant enhancement in this domain.

How to cite

M. Dolatabadi, F. Ayar, E. Javanmardi, M. Tsukada, and M. Javanmardi, “Neural Error Covariance Estimation for Precise LIDAR Localization,” in International Conference on Intelligent Computing and its Emerging Applications (ICEA 2024), Tokyo, Japan, 2024, doi: 10.1145/3732437.3732738.

BibTeX

@inproceedings{Dolatabadi2024,
  title         = {Neural Error Covariance Estimation for Precise LIDAR Localization},
  author        = {Minoo Dolatabadi and Fardin Ayar and Ehsan Javanmardi and Manabu Tsukada and Mahdi Javanmardi},
  url           = {https://arxiv.org/abs/2501.02558},
  doi           = {10.1145/3732437.3732738},
  eprint        = {2501.02558},
  archiveprefix = {arXiv},
  primaryclass  = {cs.RO},
  year          = {2024},
  date          = {2024-11-28},
  urldate       = {2024-11-28},
  booktitle     = {International Conference on Intelligent Computing and its Emerging Applications (ICEA 2024)},
  pages         = {21--26},
  address       = {Tokyo, Japan},
  abstract      = {Autonomous vehicles have gained significant attention due to technological advancements and their potential to transform transportation. A critical challenge in this domain is precise localization, particularly in LiDAR-based map matching, which is prone to errors due to degeneracy in the data. Most sensor fusion techniques, such as the Kalman filter rely on accurate error covariance estimates for each sensor to improve localization accuracy. However, obtaining reliable covariance values for map matching remains a complex task. To address this challenge, we propose a neural network-based framework for predicting localization error covariance in LiDAR map matching. To achieve this, we introduce a novel dataset generation method specifically designed for error covariance estimation. In our evaluation using a Kalman filter, we achieved a 2 cm improvement in localization accuracy, a significant enhancement in this domain.},
  keywords      = {ITS},
  pubstate      = {published},
  tppubtype     = {inproceedings}
}