Joon-Ha Kim

About

Joon-Ha Kim is a leading robotics researcher whose work centers on advancing legged locomotion through cutting-edge control, state estimation, and mechanical design. His primary contributions lie in developing real-time nonlinear model predictive control (NMPC) frameworks for dynamic locomotion on challenging terrains, particularly for quadruped and bipedal robots. Kim’s most cited paper (48 citations) introduces a constrained NMPC on SO(3) for dynamic legged locomotion, while his subsequent work on contact-implicit MPC (2024) enables diverse quadruped motions without pre-planned contact modes—a significant leap in autonomy. In state estimation, his STEP framework (41 citations) innovates with a preintegrated foot velocity factor, avoiding the usual non-slip assumption to improve observability. Kim also contributed to the design of KAIST HOUND (30 citations), a quadruped platform optimized for speed and efficiency using mixed-integer nonlinear optimization. His research has garnered over 250 total citations, with notable achievements including robust push-recovery strategies and invariant smoothing techniques. Kim’s work bridges theory and practice, making him a key figure in the future of agile, autonomous legged robots.

Research Focus

Key Achievements

Top Papers

1. 1
   Real-Time Constrained Nonlinear Model Predictive Control on SO(3) for Dynamic Legged Locomotion

   48 citations · 2020
2. 2
   Legged Robot State Estimation With Dynamic Contact Event Information

   45 citations · 2021
3. 3
   STEP: State Estimator for Legged Robots Using a Preintegrated Foot Velocity Factor

   41 citations · 2022
4. 4
   Design of KAIST HOUND, a Quadruped Robot Platform for Fast and Efficient Locomotion with Mixed-Integer Nonlinear Optimization of a Gear Train

   30 citations · 2022
5. 5
   Contact-implicit Model Predictive Control: Controlling diverse quadruped motions without pre-planned contact modes or trajectories

   22 citations · 2024
6. 6
   Seamless Reaction Strategy for Bipedal Locomotion Exploiting Real-Time Nonlinear Model Predictive Control

   22 citations · 2023
7. 7
   Invariant Smoother for Legged Robot State Estimation With Dynamic Contact Event Information

   17 citations · 2023
8. 8
   Monte Carlo Tree Search Gait Planner for Non-Gaited Legged System Control

   15 citations · 2022
9. 9
   Avoiding Obstacles during Push Recovery Using Real-Time Vision Feedback

   7 citations · 2019
10. 10
    DRPD, Dual Reduction Ratio Planetary Drive for Articulated Robot Actuators

    5 citations · 2022

Key Collaborators

Contact & Links