KIST (Korea Institute of Science and Technology) · Center for Humanoid Research, Artificial Intelligence and Robot Institute
Physical Intelligence & Embodied Robotics Lab
We study how robots can physically interact with, perceive, and reason about the real world—bridging the gap between AI and embodied intelligence by integrating physical embodiment through force-aware sensing, visuomotor policies driven by imitation learning and VLA, and coordinated whole-body control for agile, contact-rich interaction.
Welcome new members to PIER Lab! M.S. students Seungseop Lee (Yonsei Univ.) and Namyoon Kim (Korea Univ.), intern researchers Seungwon Jang (Kwangwoon Univ.) and Suhyeon Pyo, and undergraduate student researchers Byeol Nam (Ewha Womans Univ.) and Taehyun Lee (Univ. of Seoul) join us from March 2026.
Paper “CSC-MPPI: A Novel Constrained MPPI Framework with DBSCAN for Reliable Obstacle Avoidance” accepted at IROS 2025!
Jun 23, 2025
Welcome undergraduate student researchers Dogeun Kim (Univ. of Seoul) and Junsoo Park (Sungkyunkwan Univ.) to PIER Lab!
Jan 01, 2025
Paper “Intuitive and Interactive Robotic Avatar System for Tele-Existence: TEAM SNU in the ANA Avatar XPRIZE Finals” published in International Journal of Social Robotics!
We propose CSC-MPPI, a novel constrained Model Predictive Path Integral framework that integrates DBSCAN clustering for reliable obstacle avoidance in robot navigation tasks.
@inproceedings{park2025cscmppi,title={CSC-MPPI: A Novel Constrained MPPI Framework with DBSCAN for Reliable Obstacle Avoidance},author={Park, Leesai and Jang, Keunwoo and Kim, Sanghyun},booktitle={IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},year={2025},pages={19645--19652},doi={10.1109/IROS60139.2025.11246839}}
We present a motion planning framework for mobile manipulators that enables navigation including door traversal. The system integrates whole-body control with task planning to handle the complex constraints of door interaction.
@article{jang2023door,title={Motion Planning of Mobile Manipulator for Navigation Including Door Traversal},author={Jang, Keunwoo and Kim, Sanghyun and Park, Jaeheung},journal={IEEE Robotics and Automation Letters},year={2023},volume={8},number={7},pages={4147--4154},doi={10.1109/LRA.2023.3279612}}
We present a motion planning approach for closed-chain constraints using a probabilistic roadmap with improved connectivity, enabling reliable planning for robots with kinematic loops such as parallel manipulators.
@article{jang2022closedchain,title={Motion Planning for Closed-Chain Constraints Based on Probabilistic Roadmap With Improved Connectivity},author={Jang, Keunwoo and Baek, Jiyeong and Park, Suhan and Park, Jaeheung},journal={IEEE/ASME Transactions on Mechatronics},year={2022},volume={27},number={4},pages={2035--2043},doi={10.1109/TMECH.2022.3175260}}
We present a contact state estimation method for peg-in-hole assembly tasks using a Gaussian Mixture Model, enabling robust estimation of contact states during precision assembly operations.
@article{lee2022contact,title={Contact State Estimation for Peg-in-Hole Assembly Using Gaussian Mixture Model},author={Lee, Haeseong and Park, Suhan and Jang, Keunwoo and Kim, Seungyeon and Park, Jaeheung},journal={IEEE Robotics and Automation Letters},year={2022},volume={7},number={2},pages={3349--3356},doi={10.1109/LRA.2022.3146949}}
We propose a regularized hierarchical quadratic programming (R-HQP) approach for real-time whole-body motion generation, ensuring robust and stable solutions in the presence of singularities and conflicting task constraints.
@article{hong2021regularized,title={Regularized Hierarchical Quadratic Program for Real-Time Whole-Body Motion Generation},author={Hong, Seongil and Jang, Keunwoo and Kim, Sanghyun and Park, Jaeheung},journal={IEEE/ASME Transactions on Mechatronics},year={2021},volume={26},number={4},pages={2115--2126},doi={10.1109/TMECH.2020.3032522}}
