LHMM: A Tightly-Coupled LiDAR-Inertial Hybrid-Map Matching Approach for Robust and Efficient Global Localization
Submitted to IROS 2025.
⚠️ The code will be released soon. Stay tuned by starring and watching the repository!
| Dataset | Abbreviation | Name | Distance (m) | Prior Map | Scene Type |
|---|---|---|---|---|---|
| FusionPortable Dataset | fp_1 | building_day | 666 | Leica BLK360 | Campus |
| fp_2 | corridor_day | 669 | Leica BLK360 | Degeneracy | |
| fp_3 | escalator_day | 263 | Leica BLK360 | Rapid | |
| fp_4f | MCR_fast_01 | 90 | Leica BLK360 | Rapid | |
| fp_4n | MCR_normal_00 | 48 | Leica BLK360 | Rapid | |
| fp_5d | canteen_day | 250 | Leica BLK360 | Scene change | |
| fp_5n | canteen_night | 270 | Leica BLK360 | Scene change | |
| Geode Dataset | geode_1 | stairs_β | 902 | Leica RTC360 | Degeneracy |
| NCD Dataset | ncd_1 | 01_short_experiment | 1610 | Leica BLK360 | Campus |
| ncd_2 | 02_long_experiment | 3060 | Leica BLK360 | Campus |
The code is currently being prepared. The following will be provided soon:
- System dependencies & environment setup
- Compilation & running instructions
- Supported LiDAR devices & datasets
This demo showcases:
- Skeletonization-based compression of prior maps
- Local map updates under the hole-aware keyframe mechanism
- Localization performance and comparisons in challenging scenarios
Contributions, suggestions, and discussions are welcome! Since the project is still being organized, feel free to submit an issue or follow the repository for updates.
For any questions, collaborations, or inquiries, please contact: 📩 junyl@zju.edu.cn or open an issue on GitHub.