Scholarlink

<p>As robotic systems become more autonomous and capable, they are expected to work alongside humans as teammates rather than just tools. Trust is a crucial factor in collaborative human-robot interaction (HRI), and appropriate trust in robotic collaborators can influence the overall performance of the interaction. Building upon previous work in modelling trust in HRI, this paper describes a refined mathematical trust model to imitate a three-layered framework of trust, which can estimate human trust in robots in real-time. We show that the refined mathematical model significantly outperformed the existing model. Further, this model was tested and validated in a user study where participants engaged with the NAO robot in four sequential collaborative sessions. The results showed that the model is valid based on the linear regression analysis, with both the trust perception score (TPS) and interaction session being significant predictors for the trust modelled score (TMS) computed by applying the trust model. We also demonstrated that trust levels differed across the three layers of trust. This trust model highlights the model’s potential in developing adaptive robotic behaviours optimized for user trust, which can enhance the development of robotics systems that can respond to changes in human trust level in real time.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Three-Layered Framework for Estimating Human Trust in Robots During Repeated Interactions

  • 期刊论文

  • 发表日期：

    2026年2月18日

查看全文
  • Abdullah Alzahrani,
  • Simon Robinson,
  • Muneeb Ahmad

摘要

As robotic systems become more autonomous and capable, they are expected to work alongside humans as teammates rather than just tools. Trust is a crucial factor in collaborative human-robot interaction (HRI), and appropriate trust in robotic collaborators can influence the overall performance of the interaction. Building upon previous work in modelling trust in HRI, this paper describes a refined mathematical trust model to imitate a three-layered framework of trust, which can estimate human trust in robots in real-time. We show that the refined mathematical model significantly outperformed the existing model. Further, this model was tested and validated in a user study where participants engaged with the NAO robot in four sequential collaborative sessions. The results showed that the model is valid based on the linear regression analysis, with both the trust perception score (TPS) and interaction session being significant predictors for the trust modelled score (TMS) computed by applying the trust model. We also demonstrated that trust levels differed across the three layers of trust. This trust model highlights the model’s potential in developing adaptive robotic behaviours optimized for user trust, which can enhance the development of robotics systems that can respond to changes in human trust level in real time.