← Back ICRA 2026

Enabling Embodied Human-Robot Co-Learning: Requirements, Method, and Test with Handover Task

Emma M. van Zoelen, Hugo Veldman-Loopik, Karel van den Bosch, Mark Neerincx, David A. Abbink, Luka Peternel

PDF

AI summary

Key figure (auto-extracted from paper)

A structured task design and modified reinforcement learning algorithm can successfully enable reciprocal human-robot co-learning in physically embodied collaboration.

Human-robot collaboration Co-learning Physical human-robot interaction Reinforcement learning Handover task Robot adaptability

Problem

Despite growing interest in collaborative robots, there is limited understanding of how humans and robots can reciprocally learn together in physically embodied tasks, and how to actively facilitate this continuous co-learning process.

Approach

The authors derived five core requirements for co-learning and designed a physical handover task paired with a decomposed Q-learning algorithm that enables synchronous, adaptable robot behavior alongside human partners.

Key results

Five design requirements for human-robot co-learning identified
A collaborative handover task and modified Q-learning algorithm developed
Qualitative evaluation with six participants confirmed co-learning emerged in three dyads
The setup successfully met all five co-learning requirements for successful dyads

Why it matters

Offers a validated framework and practical implementation for researchers and engineers building physically embodied robots that continuously adapt alongside human collaborators.

Abstract

Despite a large body of research on robot learning, it has not yet been thoroughly studied how collaborating humans and robots learn reciprocally. In such situations, both humans and robots continuously learn about each other and the task through interaction. This paper addresses the research question: “How can human-robot co-learning be facilitated in physically em- bodied collaborative tasks?”. First, we derived five requirements for successful human-robot co-learning from literature: shared goal, synchrony, interdependence, adaptability, and transparency. Based on these requirements, we designed a collaborative human- robot handover task and a robot Q-learning method. In an evaluation with six human participants co-learning was indeed found to emerge in the hand-over task. Particularly, for three of the human-robot dyads, our designed setup proved to facilitate co-learning in a way that met all five requirements. The task and robot learning method presented in this paper demonstrate how human-robot co-learning can be enabled in physically embodied tasks.

Index terms

Human-Robot Collaboration Physical Human-Robot Interaction Reinforcement Learning