Adapting Robot's Explanation for Failures Based on Observed Human Behavior in Human-Robot Collaboration

TL;DR

The paper addresses how robots can anticipate and mitigate user confusion during failure explanations in human-robot collaboration. It combines multimodal data from 55 participants, a data-driven predictor of confusion induction, and an adaptive mechanism that adjusts explanation detail in real time, validated by leave-one-participant-out cross-validation. Key contributions include an automated confusion assessment via Algorithm 1, a Random Forest predictor achieving 89.54% accuracy and 0.878 precision, and a statistically validated mechanism that reduces confusion by decreasing explanation depth when appropriate. The results highlight the potential of real-time behavioral interpretation to tailor explanations, improve trust, and enhance efficiency in collaborative robotics.

Abstract

This work aims to interpret human behavior to anticipate potential user confusion when a robot provides explanations for failure, allowing the robot to adapt its explanations for more natural and efficient collaboration. Using a dataset that included facial emotion detection, eye gaze estimation, and gestures from 55 participants in a user study, we analyzed how human behavior changed in response to different types of failures and varying explanation levels. Our goal is to assess whether human collaborators are ready to accept less detailed explanations without inducing confusion. We formulate a data-driven predictor to predict human confusion during robot failure explanations. We also propose and evaluate a mechanism, based on the predictor, to adapt the explanation level according to observed human behavior. The promising results from this evaluation indicate the potential of this research in adapting a robot's explanations for failures to enhance the collaborative experience.

Figures (8)

• Figure 1: Human Confusion: This figure illustrates cases of confusion during robotic explanations for a robotic failure.
• Figure 2: Human Reaction Modelling for Failure Explanation
• Figure 4: Mechanism Evaluation: Confusion induction across different tasks based on whether suggested explanation adjustments were followed or not.
• Figure : (a)
• Figure : (a)