Research Analyzer
← Back ICRA 2026

Human-Like Robot Action Policy through Game-Based Empathetic Inference for Human-Robot Collaboration

Yubo Sheng, Yiwei Wang, Haoyuan Cheng, Huan Zhao, Han Ding

PDF

AI summary

Key figure (auto-extracted from paper)
Empathetic intent inference combined with proactive action planning enables robots to exhibit human-like behaviors and collaborate effectively with humans.
Human-robot collaboration intent inference game theory theory of mind proactive control human-like behavior

Problem

Human-robot collaboration faces a double-blind intent inference problem where neither agent knows the other's goals, making it difficult for robots to adapt and behave in ways humans perceive as natural.

Approach

The authors develop an empathetic intent inference framework using game theory and second-order theory of mind to estimate human intent, then apply proactive and reactive action policies to guide robot behavior.

Key results

  • Empathetic proactive policies achieve Turing-like human indistinguishability
  • Proactive nonego policy improves goal tracking and reduces effort in simulated surgery
  • Identification of key behavioral traits driving human-like perception
  • Proactive nonego strategy outperforms nonempathetic baselines in performance and satisfaction

Why it matters

This framework advances the design of collaborative robots that naturally adapt to human partners, benefiting fields like surgical robotics and industrial automation.

Abstract

Harmonious human–robot collaboration requires the robot to behave like a human partner, which raises the critical question of what factors make the robot do so. This article proposes a series of policies based on empathetic and nonempathetic intent inference, proactive and reactive action planning, and ego and nonego action styles to examine, which modules enable robots to exhibit human-like behaviors. Two series of experiments are con- ducted with human subjects to test the performance of the proposed controllers. In Experiment 1, the participant must identify whether the collaborating partner is a human, similar to a turing test. The classification results empirically verify that the designed empa- thetic proactive policies enable the robot to exhibit human-like behaviors. Experiment 2 indicates that the proposed policy can be applied to complex collaborative tasks, and this result is consistent with the findings of Experiment 1. From empirical evidence from the experiments, we believe that empathy and proactive policies are essential elements to enable robots to perform human-like actions.

Index terms

Physical Human-Robot Interaction Cooperating Robots Cognitive Human-Robot Interaction Human Factors and Human-in-the-Loop

Related papers