Minku Kim

Ph.D. Candidate in Robotics

Email CV GitHub LinkedIn

About Me

Hi, I'm Minku Kim, a Ph.D. candidate in Robotics at Oregon State University and a member of the Dynamic Robotics and Artificial Intelligence Laboratory, advised by Professor Alan Fern.

Previously, I earned an M.S. in Mechanical Engineering and Applied Mechanics from the University of Pennsylvania. In the Dynamic Autonomy and Intelligent Robotics Lab, advised by Professor Michael Posa and co-advised by Professor Pratik Chaudhari.

I am always happy to discuss robotics research and potential collaborations. Feel free to get in touch.

Research Interests

Legged & Humanoid Robots: reinforcement learning, whole-body control, skill learning, and sim-to-real transfer
Robot Perception: 6D pose and shape estimation, multi-view vision, and sensor-rich feedback

News

2026Humanoid Hanoi was accepted to CoRL 2026. See you in Austin, Texas!
2025Learning a Vision-Based Footstep Planner for Hierarchical Walking Control was accepted to IEEE-RAS Humanoids as an oral presentation. See you in Seoul, Korea!
2025Received the Oregon State College of Engineering Scholarship.
2025Received the Penn Engineering Outstanding Research Award.

Research

Preprint

SAGE: Semantic And Geometric Estimation of 6D Object Pose from Multi-View Observations

Minku Kim†, Ho Jin Choi†, and Nadia Figueroa

Under Review
Preprint

ASM-6D: Generative Correspondences for Fast Object Pose-Shape Fitting

Ho Jin Choi†, Minku Kim†, and Nadia Figueroa

Under Review
CoRL

Humanoid Hanoi: Investigating Shared Whole-Body Control for Skill-Based Box Rearrangement

Minku Kim†, Kuan-Chia Chen†, Aayam Shrestha, Li Fuxin, Stefan Lee, and Alan Fern

Conference on Robot Learning, 2026.

arXiv Project Page Code
RSSW

Dynamic-ASM6D: Real-Time 6D Object Pose and Shape Estimation via Active Shape Models and ADMM

Ho Jin Choi†, Yi-Hsuan Cheng†, Minku Kim†, and Nadia Figueroa

Equivariant Systems Workshop at RSS and IEEE-RAS TC Virtual Poster Session, 2025.

RSS Workshop TC Poster
Humanoids

Learning a Vision-Based Footstep Planner for Hierarchical Walking Control

Minku Kim, Brian Acosta, Pratik Chaudhari, and Michael Posa

IEEE-RAS International Conference on Humanoid Robots, 2025.

arXiv IEEE Xplore Semantic Scholar Oral Presentation

Projects

Quadruped Control

Obstacle Avoidance Using Control Barrier Functions

Using control barrier functions.

Animation of Comparative Analysis of MPC, LQR, and RL Footstep Planners on Uneven Terrain

Legged Locomotion

Comparative Analysis of MPC, LQR, and RL Footstep Planners on Uneven Terrain

A Drake-based comparison of model-predictive, LQR, and reinforcement-learning footstep planners for Cassie.

Optimization and Control

Optimization-Based Obstacle Estimation from Human Demonstrations

Learning obstacle geometry from expert demonstrations through CLF-CBF constraints and probabilistic regression.

Paper Poster

Robot Manipulation

Vision-Aided Pick-and-Stack with a 7-DOF Manipulator

AprilTag perception, gradient-based inverse kinematics, and collision-aware waypoint planning for a Franka Panda.