Training Models to Detect Successive Robot Errors from Human Reactions

2025-10-10

AI Summary

This study addresses the challenge in human–robot interaction (HRI) of timely detecting and responding to repeated robot failures. We propose a behavior-driven approach for robot error detection and failure-stage identification, leveraging nonverbal human responses—such as posture, facial expressions, and gaze—captured via video from 26 participants exposed to recurrent robotic errors. Temporal behavioral features are extracted to construct personalized user models, enabling individualized error detection and fine-grained classification of failure stages: onset, accumulation, and collapse. To our knowledge, this is the first work to systematically model the progressive human response to consecutive errors, moving beyond conventional single-error detection paradigms. Experimental results demonstrate an error detection accuracy of 93.5% and a failure-stage classification accuracy of 84.1%, significantly enhancing HRI systems’ adaptive responsiveness to persistent faults.

Technology Category

• Intelligent Robots: Human-Robot Interaction
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

As robots become more integrated into society, detecting robot errors is essential for effective human-robot interaction (HRI). When a robot fails repeatedly, how can it know when to change its behavior? Humans naturally respond to robot errors through verbal and nonverbal cues that intensify over successive failures-from confusion and subtle speech changes to visible frustration and impatience. While prior work shows that human reactions can indicate robot failures, few studies examine how these evolving responses reveal successive failures. This research uses machine learning to recognize stages of robot failure from human reactions. In a study with 26 participants interacting with a robot that made repeated conversational errors, behavioral features were extracted from video data to train models for individual users. The best model achieved 93.5% accuracy for detecting errors and 84.1% for classifying successive failures. Modeling the progression of human reactions enhances error detection and understanding of repeated interaction breakdowns in HRI.

Problem

• Detecting successive robot errors through human reactions
• Modeling progression of human responses to repeated failures
• Using machine learning to classify stages of robot failure

Innovation

• Training models to detect robot errors from human reactions
• Using machine learning to recognize stages of robot failure
• Modeling progression of human reactions enhances error detection