Predicting Human Behavior in Autonomous Systems: A Collaborative Machine Teaching Approach for Reducing Transfer of Control Events

2025-05-15

AI Summary

Frequent false triggers of Transfer-of-Control (ToC) in non-critical scenarios undermine the robustness and operational efficiency of autonomous systems. Method: This paper proposes a collaborative ToC prediction and assistance framework grounded in human–machine interaction data. Departing from conventional sensor-only approaches, it innovatively leverages non-expert user operational behavior—collected via an interactive simulation platform (an industrial vacuum cleaner environment)—to train an LSTM-based temporal behavioral model. Furthermore, it introduces a “collaborative machine teaching” paradigm, enabling AI to directly acquire domain priors from human problem-solving behaviors, thereby overcoming limitations of perception-only methods. Results: Experiments demonstrate substantial reduction in unnecessary ToC events; the model achieves high generalizability using only routine interaction data. Real-world evaluation confirms significant improvements in system robustness and human–autonomy collaboration efficiency.

Technology Category

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

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

As autonomous systems become integral to various industries, effective strategies for fault handling are essential to ensure reliability and efficiency. Transfer of Control (ToC), a traditional approach for interrupting automated processes during faults, is often triggered unnecessarily in non-critical situations. To address this, we propose a data-driven method that uses human interaction data to train AI models capable of preemptively identifying and addressing issues or assisting users in resolution. Using an interactive tool simulating an industrial vacuum cleaner, we collected data and developed an LSTM-based model to predict user behavior. Our findings reveal that even data from non-experts can effectively train models to reduce unnecessary ToC events, enhancing the system's robustness. This approach highlights the potential of AI to learn directly from human problem-solving behaviors, complementing sensor data to improve industrial automation and human-AI collaboration.

Problem

• Reducing unnecessary Transfer of Control events in autonomous systems
• Predicting human behavior using collaborative machine teaching
• Enhancing human-AI collaboration through data-driven fault handling

Innovation

• Data-driven method using human interaction for AI training
• LSTM-based model predicts user behavior effectively
• Reduces unnecessary Transfer of Control events robustly