Using a Bayesian Network to Predict User Trust in Teleoperation Robots

Trust plays a crucial role in human-robot interaction (HRI), especially in teleoperation scenarios where users control robots remotely. This paper develops a static trust prediction model using a Bayesian approach to improve system design, user experience, and robotic reliability. By integrating prior knowledge with empirical data, we constructed a Bayesian model that estimates trust levels in teleoperated robotic systems. Our model incorporates physiological measures and task performance data to provide a comprehensive trust prediction framework. We evaluated the model’s performance using precision, recall, and F1 score, achieving high precision (0.92), good recall (0.84), and a balanced F1 score (0.79). These metrics demonstrate the model’s effectiveness in accurately predicting trust levels in teleoperated robotic systems. The results underscore the importance of trustworthiness in maintaining user confidence and improving system interactions. This study highlights the potential of Bayesian models in enhancing the reliability and user experience of teleoperated robots, offering valuable insights for future developments in HRI research.