Research Analyzer
← Back ICRA 2026

TeachingBot: Robot Teacher for Human Handwriting

Zhimin Hou, Cunjun Yu, David Hsu, Haoyong Yu

PDF

AI summary

Key figure (auto-extracted from paper)
TeachingBot significantly improves handwriting quality and learner engagement by dynamically adapting physical guidance to individual writing styles.
Robot-assisted teaching Physical human-robot interaction Handwriting learning Variable impedance control Probabilistic trajectory generation

Problem

Scaling up one-on-one physical skill instruction is limited by a shortage of human teachers, while existing robotic teaching systems struggle to adapt to diverse individual writing styles and maintain optimal guidance levels to prevent dependency or disengagement.

Approach

The system captures each learner's unique handwriting style using a probabilistic GMM-GP model to generate personalized intermediate trajectories, then applies a variable impedance controller that dynamically adjusts physical guidance strength based on real-time performance.

Key results

  • Significantly improved handwriting quality over three baselines
  • Increased accuracy in replicating character structure and stroke details
  • Higher interaction forces indicating improved active engagement
  • Adaptive trajectory generation successfully bridges individual styles to standard forms

Why it matters

It provides a scalable, adaptive framework for robot-assisted motor skill teaching that balances personalization with effective physical guidance, benefiting education, rehabilitation, and human-robot interaction.

Abstract

Teaching and learning physical skills often require one-on-one interaction, making it difficult to scale up, as there are not enough human teachers. Robots offer an attractive alternative. This paper presents TeachingBot, an adaptive robotic system that teaches handwriting to human learners through physical interaction. Robot teaching poses two major challenges: (i) adapting to the individual handwriting style of the human learner and (ii) maintaining an engaging learning experience. For the first challenge, TeachingBot uses a probabilistic model to capture the human learner’s writing style from their writing samples. Drawing on the insight that effective teaching bal- ances standardization with individuality, the system generates a personalized teaching trajectory that aligns with the human learner’s natural writing. For the second challenge, TeachingBot employs variable impedance control to guide the human learner, dynamically adjusting the strength of physical guidance based on the human learner’s writing performance. Human-subject ex- periments demonstrate the effectiveness of TeachingBot, showing clear improvement in learners’ handwriting and engagement over baseline methods.

Index terms

Human-Centered Robotics Physical Human-Robot Interaction Human Performance Augmentation

Related papers