Obin Kwon

Seoul National University, University of Illinois Urbana-Champaign

Papers

Total Citations

H-Index

About

Obin Kwon is an emerging robotics and computer vision researcher whose work spans embodied AI navigation, indoor localization, and humanoid robot control. His research addresses some of the most challenging problems in autonomous systems: enabling robots to understand and navigate complex environments with human-like intelligence. Kwon's most recognized contribution, "Topological Semantic Graph Memory for Image-Goal Navigation" (2022, 7 citations), introduces a sophisticated framework allowing embodied robots to incrementally build landmark-based semantic memory graphs, enabling intelligent target-seeking in unknown environments — a significant step toward truly autonomous robot navigation. His work on "WayIL" (2024) tackles the practical challenge of robot localization using human-readable wayfinding maps, bridging the gap between abstract human spatial representations and precise robotic positioning systems. Most recently, Kwon has ventured into humanoid robotics with CHILD (2025), a whole-body teleoperation system that advances joint-level control for humanoid robots, expanding the frontier of complex manipulation tasks. Across his growing publication record, Kwon demonstrates a consistent focus on making robots more capable of operating intelligently in human-centered environments, positioning him as a promising contributor to next-generation autonomous systems research.