Gunmin Lee
Papers
3
Total Citations
15
H-Index
2
About
Gunmin Lee is a robotics researcher whose work focuses on socially-aware navigation, multi-agent reinforcement learning, and robot control in crowded, human-centric environments. His most impactful contribution, "SCAN: Socially-Aware Navigation Using Monte Carlo Tree Search" (2023, 10 citations), addresses the critical challenge of enabling robots to navigate densely populated spaces without causing discomfort to pedestrians. By integrating Monte Carlo Tree Search with social norms, Lee’s global planner ensures both efficiency and human comfort—a foundational step toward seamless human-robot coexistence. In "Stage-Wise Reward Shaping for Acrobatic Robots" (2025, 4 citations), he introduces a constrained multi-objective reinforcement learning approach that simplifies complex reward design, enabling robots to learn acrobatic maneuvers through intuitive, stage-based strategies. His work "MAC-ID" (2024) further advances multi-agent reinforcement learning by balancing local coordination with individual diversity, improving navigation in dynamic crowds. With a growing citation footprint, Lee’s research is shaping the future of socially-aware robotics, offering practical frameworks for robots that move safely and naturally among people.
Research Focus
Key Achievements
Top Papers
- 1SCAN: Socially-Aware Navigation Using Monte Carlo Tree Search10 citations · 2023
- 2
- 3