Youngjin Seo

I'm an Integrated M.S./Ph.D. student in the Robotics and Intelligent System Engineering Lab (RISE Lab) at Sungkyunkwan University (SKKU), advised by Prof. Hyungpil Moon.

I'm interested in robot learning for manipulation, with a focus on imitation learning, vision-language-action (VLA) models, and force-based policies.

Email  /  CV  /  Scholar  /  Github

profile photo

Education

  • Sungkyunkwan University (SKKU), South Korea
    Integrated M.S./Ph.D., Intelligent Robotics
    Mar. 2025 – present
  • Sungkyunkwan University (SKKU), South Korea
    B.S., Mechanical Engineering (summa cum laude, top 5%)
    Mar. 2019 – Feb. 2025

Publications
URF URF: A Unified Robot Control-Policy Framework for Stable Contact Aware Manipulation
Jiyou Shin, Youngjin Seo, Jaeseog Won, Sungwon Seo, Hyunjun Kim, Seokmin Yoon, Tuan Luong, Hyungpil Moon
arXiv, 2026
project page / arXiv

A manipulation policy that predicts not only a compliant action but also the controller behavior used to execute it, jointly outputting a virtual target, a stiffness matrix, and an impedance-admittance switch ratio.

NPU benchmark Benchmarking 6D Pose Estimation and 6-DoF Grasp Detection on a Low-Power NPU: Case Studies with PVN3D and HGGD
Youngjin Seo*, Hyojun Yun*, Hong-ryul Jung, Hyungpil Moon (* equal contribution)
International Conference on Control, Automation and Systems (ICCAS), 2025
paper

A benchmark of PVN3D and HGGD on a low-power NPU, measuring whether 6D pose estimation and 6-DoF grasp detection stay practical under embedded compute budgets.

Projects
Development of a K-Logistics Humanoid Robot Integrated with a High-Sensitivity Robotic Hand Based on a Multimodal AI Foundation Model
Sep. 2025 – present
  • Deployed imitation learning and vision-language-action algorithms, including ACT, π0, π0.5, and TA-VLA
  • Set up an Isaac Sim simulation environment with a dexterous hand for demonstration data augmentation
Development of a SLAM-Based Drone System for Cart Monitoring
Apr. 2026 – Aug. 2026
  • Implemented an integrated ground control interface that visualizes real-time drone and cart positions on a 2D map
  • Enabled drone waypoint setting and cart position correction directly through mouse interaction on the map
Development of Core Technology for Mobile Manipulator for 5G Edge-Based Transportation and Manipulation
Jun. 2025 – Dec. 2025
  • Object recognition using YOLO11, SAM 2, and FoundationPose
  • Grasp point detection with HGGD, including collision checking of the detected grasps
  • System integration for grasping with a three-fingered hand
Development of Low-Cost, High-Performance Visual Sensor Technology Integrated with AI Semiconductors and AI Algorithms Designed for Robots
Dec. 2024 – Nov. 2025
  • Converted the architectures of PVN3D and HGGD to run on an NPU
Development of a Simulation Drone Competition Utilizing Realistic Environments and Drone Dynamics
Mar. 2023 – Dec. 2023
  • Developed a multi-drone racing competition in an Unreal Engine simulation environment using AirSim, PX4, and QGroundControl
  • Built an autonomous drone flight algorithm based on PID control within the simulation environment

Teaching

  • AI & Robotics Practical Training for Vocational High School Students
    Instructor, Sungkyunkwan University
    Jul. 2026
  • Kinematics of Machinery (EME3017)
    Teaching Assistant, Sungkyunkwan University
    Spring 2025
  • Design Lab on Vibration and Dynamic Systems (EME3057)
    Teaching Assistant, Sungkyunkwan University
    Spring 2025

Template from Jon Barron's website.