• Introduction

• News

• Dataset

• Getting Started

• Results

• Citations

• License

PALoc presents a novel approach for generating high-fidelity, dense 6-DoF ground truth (GT) trajectories, enhancing the evaluation of Simultaneous Localization and Mapping (SLAM) under diverse environmental conditions. This framework leverages prior maps to improve the accuracy of indoor and outdoor SLAM datasets. Key features include:

• Robustness in Degenerate Conditions: Exceptionally handles scenarios frequently encountered in SLAM datasets.
• Advanced Uncertainty Analysis: Detailed covariance derivation within factor graphs, enabling precise uncertainty propagation and pose analysis.
• Open-Source Toolbox: An open-source toolbox is provided for map evaluation, indirectly assessing trajectory precision. Experimental results demonstrate at least a 30% improvement in map accuracy and a 20% increase in direct trajectory accuracy over the ICP algorithm, across various campus environments.

• 2024/03/26: Early access by IEEE/ASME TMECH.
• 2024/02/01: Preparing codes for release.
• 2024/01/29: Accepted by 2024 IEEE/ASME TMECH.
• 2023/12/08: Resubmitted.
• 2023/08/22: Reject and resubmitted.
• 2023/05/13: Submitted to IEEE/ASME TRANSACTIONS ON MECHATRONICS (TMECH).
• 2023/05/08: Accepted by ICRA 2023 Workshop on Future of Construction.

Our algorithms were rigorously tested on the Fusion Portable Dataset.

Below is our sensor kit setup.

The complete code will be released post-journal acceptance. For a sneak peek:

• Watch on Bilibili
• Watch on YouTube

X Degenerate (Translation)	Yaw Degenerate (Rotation)

To plot the results, you can follow this scripts.

We follow the assumption of pose Independence as barfoot does(TRO 2014) as equation (13), there is no correlated cov between this two poses. Left-hand convention means a small increments on world frame.

Since GTSAM follow the right-hand convention on SE(3) , we need to use the adjoint representation as equation (14).

For referencing our work in PALoc, please use:

@ARTICLE{hu2024paloc,
  author={Hu, Xiangcheng and Zheng, Linwei and Wu, Jin and Geng, Ruoyu and Yu, Yang and Wei, Hexiang and Tang, Xiaoyu and Wang, Lujia and Jiao, Jianhao and Liu, Ming},
  journal={IEEE/ASME Transactions on Mechatronics}, 
  title={PALoc: Advancing SLAM Benchmarking With Prior-Assisted 6-DoF Trajectory Generation and Uncertainty Estimation}, 
  year={2024},
  volume={},
  number={},
  pages={1-12},
  doi={10.1109/TMECH.2024.3362902}
  }

The map evaluation metrics of this work follow Cloud_Map_Evaluation. Please cite:

@article{jiao2024fp,
  author    = {Jianhao Jiao and Hexiang Wei and Tianshuai Hu and Xiangcheng Hu and Yilong Zhu and Zhijian He and Jin Wu and Jingwen Yu and Xupeng Xie and Huaiyang Huang and Ruoyu Geng and Lujia Wang and Ming Liu},
  title     = {FusionPortable: A Multi-Sensor Campus-Scene Dataset for Evaluation of Localization and Mapping Accuracy on Diverse Platforms},
  booktitle = {2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  year      = {2022}
}

This project's code is available under the MIT LICENSE.