Learning Contact Tasks Skills Based on DMP and Affordance Templates
Hunjo Lee, Gi-Hun Yang
AI summary
Problem
Learning from demonstration struggles to generalize contact-rich tasks to unknown or varying poses, limiting real-world robotic applicability.
Approach
A teleoperation-based framework records expert position and wrench data, then uses dynamic movement primitives for trajectories and affordance templates to learn adaptive contact strategies.
Key results
- 100% success rate on valve-turning across three novel poses
- 60-70% reduction in contact force and torque RMS vs. baseline
- 60-80% success on precision peg-in-hole insertion (40 µm clearance)
- Validated robust pose generalization for contact-rich assembly
Why it matters
Enables robots to reliably execute complex contact manipulation in unstructured environments, advancing autonomous manufacturing and assembly.
Abstract
Learning from demonstration (LfD) enables robots to learn experts’ skills by human demonstration. Re- cently, LfD has been developed for learning and performing skills in contact-rich tasks. However, task performance has not been generalized to unknown poses in contact-rich tasks. In this paper, we propose a teleoperation-based learning from demonstration (LfD) framework for performing contact-rich tasks in unknown poses. Expert demonstrations are collected via a bilateral teleoperation system, with an orientation syn- chronization algorithm aiding intuitive manipulation. From demonstrations, position and wrench profiles are recorded. Task trajectories are learned using dynamic movement primitives (DMP), while strategy learning allocates input and compliance spaces based on affordance templates to adapt motion during contact. By combining trajectory and strategy learning, the framework successfully reproduces manipulation behaviors in novel configurations. Experiments on turning-valve and peg-in- hole insertion validate the method, showing improved success rates and robustness to pose variations.