Hyeonjae Gil
Papers
5
Total Citations
58
H-Index
2
About
Hyeonjae Gil is a leading researcher in robotics perception, specializing in LiDAR-based place recognition and lifelong mapping for autonomous systems. His work addresses critical challenges in robot localization and SLAM (Simultaneous Localization and Mapping), particularly under dynamic and spatiotemporally varying conditions. Gil’s most impactful contribution is the **HeLiPR** dataset (47+ citations), the first heterogeneous LiDAR dataset designed for inter-LiDAR place recognition, enabling robust matching across different sensor types and environmental conditions. He also developed **ELite**, an ephemerality-aided lifelong mapping framework that seamlessly aligns multi-session data, removes dynamic objects, and updates maps end-to-end—a breakthrough for long-term robot deployment. His work on **Helios** introduces overlap-based learning and local spherical transformers for heterogeneous LiDAR place recognition, addressing the limitations of traditional spinning LiDAR approaches. Additionally, Gil has explored cross-modal perception through night-to-day thermal image translation for deep thermal place recognition. With over 50 total citations and multiple publications in top venues, Hyeonjae Gil is shaping the future of robust, long-term robotic navigation in real-world environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2Ephemerality Meets Lidar-Based Lifelong Mapping5 citations · 2025
- 3
- 4
- 5