Jeongsoo Park

Robotics Research (United States), Korea Advanced Institute of Science and Technology

Papers

Total Citations

161

H-Index

About

Jeongsoo Park is at the forefront of legged robotics, specializing in reinforcement learning, high-speed control, and autonomous navigation for quadrupedal robots. His major contributions address critical challenges in deploying robots on complex, deformable, and discrete terrains—environments where traditional controllers often fail. Park’s most influential work, "Learning quadrupedal locomotion on deformable terrain" (2023, 147 citations), pioneered simulation-based reinforcement learning approaches that enable stable, high-speed locomotion on soft ground, bridging a key gap between simulation and real-world application. He further advanced autonomous navigation with "Learning Semantic Traversability With Egocentric Video and Automated Annotation Strategy" (2024), enhancing robots’ ability to interpret urban scenes for safe path planning. His hierarchical navigation pipeline for high-speed legged robots on discrete terrain (2025) tackles the complex optimization of long-horizon dynamics. Notably, Park led the development of RAIBO2, a highly efficient quadruped that completed a full marathon on a single battery charge—a landmark achievement in energy-efficient robotics. With a growing citation impact, Park’s work is shaping the next generation of agile, intelligent legged robots for real-world deployment.