Hyeonjae Gil* · Dongjae Lee* · Giseop Kim · Ayoung Kim
(* Equal Contribution)
ELite is a LiDAR-based lifelong mapping framework that integrates ephemerality—the likelihood of a point being transient or persistent—into the entire pipeline. Unlike binary classifications of map elements (static vs. dynamic), ELite introduces a two-stage ephemerality concept to represent changes across different time scales:
- Local ephemerality: Captures short-term variations (e.g., moving cars vs. parked cars).
- Global ephemerality: Represents long-term changes (e.g., parked cars vs. new buildings)
By leveraging ephemerality, ELite seamlessly aligns multiple sessions, removes dynamic objects, and matains accurate maps.
Code will be available soon!
If you use ELite for any academic work, please cite our paper.
@INPROCEEDINGS { hjgil-2025-icra,
author={Hyeonjae Gil and Dongjae Lee and Giseop Kim and Ayoung Kim},
title={Ephemerality meets LiDAR-based Lifelong Mapping},
booktitle={Proceedings of the IEEE International Conference on Robotics and Automation (ICRA)},
year={2025},
month={May.},
address={Atlanta},
}If you have any questions, please contact:
- Hyeonjae Gil (now9728@gmail.com)
- Dongjae Lee (dongjae0107@gmail.com)
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT)(No. RS-2024-00461409), and in part by the Robotics and AI (RAI) Institute.