E-In Son

About

Hi! I am a Ph.D. student at the SNU Vehicle Intelligence Lab, advised by Prof. Seung-Woo Seo.

My research focuses on learning-based autonomous navigation, with particular interest in visual navigation, traversability estimation, and representation learning. Building on earlier work in unstructured, off-road environments, I am now moving toward more general navigation and robotics, including embodied agents such as world models and vision-language-action models.

Publications

ORION

Ordinal Neural Collapse as a Representation Prior for Visual Navigation ECCV 2026

E-In Son*, Jung-Taak Kim*, Seung-Woo Seo

European Conference on Computer Vision (ECCV), 2026

Paper Project Page

PTS-Map

PTS-Map: Probabilistic Terrain State Map for Uncertainty-Aware Traversability Mapping in Unstructured Environments RA-L 2025

Dong-Wook Kim, E-In Son, Chan Kim, Ji-Hoon Hwang, Seung-Woo Seo

IEEE Robotics and Automation Letters (RA-L), 2025

Paper Code

Follow the Footprints

Follow the Footprints: Self-supervised Traversability Estimation for Off-road Vehicle Navigation based on Geometric and Visual Cues ICRA 2024

Yurim Jeon, E-In Son, Seung-Woo Seo

IEEE International Conference on Robotics and Automation (ICRA), 2024

Paper Code

Adaptive Robot Traversability Estimation

Adaptive Robot Traversability Estimation Based on Self-Supervised Online Continual Learning in Unstructured Environments RA-L 2024

Hyung-Suk Yoon, Ji-Hoon Hwang, Chan Kim, E-In Son, Se-Wook Yoo, Seung-Woo Seo

IEEE Robotics and Automation Letters (RA-L), 2024

Paper

Traversability-aware Adaptive Optimization

Traversability-aware Adaptive Optimization for Path Planning and Control in Mountainous Terrain RA-L 2024

Se-Wook Yoo*, E-In Son*, Seung-Woo Seo

IEEE Robotics and Automation Letters (RA-L), 2024

Paper Project Page

Projects

Mobile Robot for Autonomous Exploration in Unstructured Environments

Agency for Defense Development (ADD)

End-to-end visual navigation with multi-modal localization for long-horizon missions in unknown, hazardous terrain. I lead the planning stack, from learning-based path prediction to real-time safe replanning (2024 – 2026), and previously developed the path tracking module (2022 – 2023).

Awards & Honors

Earth Rovers Challenge (1st Place)
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
Oct 2024
Academic Excellence Scholarship
Konkuk University
2018 – 2020