CAREER: Enhancing Trust-Driven Human-Autonomy Interaction: Modeling Trust Dynamics and Supporting Trust Calibration

该项目旨在解决人类与技术合作关系的核心问题：设计师如何促进对技术的适当信任的建立？自动驾驶汽车和协作机器人等先进技术正在进入经济的各个领域，并将从根本上改变人们的生活和工作方式。然而，只有人们对这些技术建立起适当的信任，才能充分发挥这些技术的经济、安全和健康潜力。本项目旨在理解和模拟信任的形成和演变，并发展适应性自治，促进适当信任的建立。该项目通过培养下一代科学家进行人类自主互动，并为K-12学生开展外展活动，重点是增加女性和代表性不足的少数民族的参与，以及为在职专业人员开展外展活动，以帮助他们适应未来的工作场所，其中人类和自主代理将越来越多地作为一个团队工作，从而促进STEM教育和劳动力发展。研究工作主要有三个方面：(1)建立信任形成和演化的时间动力学模型；(2)从行为和生理信息中评估个体对自主性的信任程度；(3)开发方法，使自主代理能够调整其行为，并引导人达到成功操作所需的更理想的信任水平。该项目将在一个平台上进行多项“人在环”研究，在多种情况下，人类与自主无人机互动，包括需要一系列任务和决策的搜索和救援。在此基础上，建立信任动态模型、信任推理算法和自适应方法，并对其进行测试和验证。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This project aims to tackle a fundamental question at the heart of human-technology partnership: How can designers facilitate the establishment of appropriate trust in technology? Advanced technologies such as autonomous vehicles and collaborative robots are entering every sector of the economy and will fundamentally alter the way people live and work. However, realizing the full economic, safety, and health potential of these technologies is only possible if people establish appropriate trust in them. This project aims to understand and model the formation and evolution of trust, and to develop adaptive autonomy that facilitates the establishment of appropriate trust. The project advances STEM education and workforce development by nurturing the next generation of scientists in human-autonomy interaction, and by developing outreach activities for K-12 students with an emphasis on increasing participation of women and underrepresented minorities, and for working professionals aimed at helping them adapt to the future workplace wherein humans and autonomous agents will increasingly work as a team. The research work has three main thrusts: (1) modeling temporal dynamics of trust formation and evolution; (2) estimating a person's trust in autonomy from behavioral and physiological information; and (3) developing methods that enable the autonomous agent to adapt its behavior and guide a person toward a more desired level of trust required for successful operation. The project will conduct multiple human-in-the-loop studies on a platform with humans interacting with autonomous drones in several situations, including search and rescue where a sequence of tasks and decisions are required. Based on the data, trust dynamics models, trust inference algorithms and adaptive methods will be built, tested and validated.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.