Whole-Arm Sensing
Proprioceptive sensor-based contact point localization and force identification.
This project focuses on Contact Estimation for robotic manipulators. By leveraging joint encoders, joint torque sensors, and base force-torque sensors, our methods enable robot arms to estimate multi-contact locations and identify contact forces simultaneously without requiring external cameras or artificial tactile skin.
Key Research Topics
- Contact Diffusion Models (CDM): Applying denoising diffusion probabilistic models to proprioceptive sensing to robustly localize multiple simultaneous contact points.
- Multi-Contact Particle filter with Exploration Particles (MCP-EP): Analytically and optimizationally resolving contact locations and external force vectors on the robot’s links.
- Real-time Proprioceptive Sensing: Implementing efficient estimators capable of running at high frequencies suitable for feedback control loops.
Related Publications
- CDM: Contact Diffusion Model for Multi-Contact Point Localization (ICRA 2025)
- Seo Wook Han, Min Jun Kim
- [arXiv] [IEEE Xplore]
- Learning Multi-Contact Localization via Diffusion Models with Proprioceptive Sensing (RSS 2025 Ws)
- Seo Wook Han, Min Jun Kim
- Contact Estimation Diffusion Model for Collaborative Robots (ICCAS 2024)
- Seo Wook Han, Min Jun Kim
- Proprioceptive Sensor-Based Simultaneous Multi-Contact Point Localization and Force Identification for Robotic Arms (ICRA 2023)
- Seo Wook Han, Min Jun Kim
- [arXiv] [IEEE Xplore]
- Multi-Contact Point Localization and Force Identification for Collaborative Robot Arm using Proprioceptive Sensor (KRoC 2023)
- Seo Wook Han, Min Jun Kim