Modeling Synergistic Coordination in Multi-Agen Human and Human-Machine Systems

社会、行为和经济科学理事会提供博士后研究金，为最近的博士毕业生提供获得额外培训的机会，在知名科学家的赞助下获得研究经验，并在本科和研究生培训之外拓宽他们的科学视野。博士后奖学金还旨在帮助新的科学家指导他们的研究工作，跨越传统的学科领域，并利用独特的研究资源，地点和设施，包括在国外的地点。这个博士后奖学金支持一个新兴的跨学科学者在人机交互领域。许多日常活动的成功取决于合作个体的有效行为协调。这些活动包括通过装配线移动材料，与家庭成员一起装载洗碗机，与同事握手，以及操纵遥控车辆修理卫星。最近在人类运动科学、心理学和复杂动力系统领域的研究揭示了对这些相互作用进行建模的新方法。这个项目的目的是研究这些人类启发的模型是否可以在机器人和人工代理中实现，以开发高度鲁棒和灵活的人机系统。因此，该项目将通过将这些系统的行为建立在自然人与人交互的动态中，从而在基于交互式人类-机器人和人类-人工智能体的系统中取得重要的新进展。更广泛地说，拟议的项目对工作场所机械，自动驾驶汽车和医疗技术的设计和开发具有重要意义，这些技术需要高水平的精确人机合作。理解如何在机器人和人工智能体中实现人类启发的行为动态模型也将有助于辅助技术的发展，包括假肢，机器人辅助和治疗技术，这些技术更好地适应特定最终用户的需求。合作的个体表现为一个单一的协同单位，他们的行动和行为通常以自组织的方式协调，需要很少或不需要明确的指导或先验规划。该项目的目标是开发和测试动态（微分方程）模型，捕捉人类多代理协调的协同自组织，并将这些模型部署在人工代理（虚拟和机器人代理）中，以创建高度鲁棒和相互响应的耦合人机系统。该项目旨在推进协调人机系统开发的新框架，该框架来自物理，信息和生物力学过程的经验证据，这些过程塑造和限制了人类成功和适应性联合行动行为的动态。使用一组物体移动和传递任务，该项目的具体目标是证明：（1）在执行物理联合行动任务的成对人类代理之间发生的协同协调的动态模型如何能够作为人工代理性能的基础;以及（2）可以在人类和人工代理的交互系统中实现，以产生稳定和自适应的鲁棒行为协调模式，在人与人之间的互动中观察到的。拟议的研究整合了行为，认知和社会心理学，人类运动科学，计算机科学，复杂性科学，工程和哲学的当代方法，并将推进我们对人类联合行动的行为动力学的基本理解。此外，拟议的项目将对基于辅助，诊断和治疗感觉运动技术的交互式机器人和人工代理的开发和设计产生广泛和变革性的影响。

The Directorate of Social, Behavioral and Economic Sciences offers postdoctoral research fellowships to provide opportunities for recent doctoral graduates to obtain additional training, to gain research experience under the sponsorship of established scientists, and to broaden their scientific horizons beyond their undergraduate and graduate training. Postdoctoral fellowships are further designed to assist new scientists to direct their research efforts across traditional disciplinary lines and to avail themselves of unique research resources, sites, and facilities, including at foreign locations. This postdoctoral fellowship award supports a rising interdisciplinary scholar in the area of human computer interaction. The success of many everyday activities depends on the effective behavioral coordination of cooperating individuals. These activities include moving materials through an assembly line, loading a dishwasher with a family member, shaking hands with a co-worker, and maneuvering a remotely operated vehicle to repair a satellite. Recent research within the fields of human movement science, psychology and complex dynamical systems has revealed new methods for modelling these interactions. The objective of this project is to examine whether these human inspired models can be implemented in robotic and artificial agents in order to develop highly robust and flexible human-machine systems. Accordingly, the project will result in important new advances in interactive human-robotic and human-artificial-agent based systems by grounding the behavior of these systems in the dynamics of natural human-human interaction. More generally, the proposed project has significant implications for the design and development of workplace machinery, autonomous vehicles, and medical technologies where high levels of precise human-machine cooperation are necessary. Understanding how to implement human inspired behavioral dynamic models in robotic and artificial agents will also help with the development of assistive technologies, including prosthetics, robotic aids, and therapeutic technologies, that are better tailored to the needs of the specific end user.Cooperating individuals behave as a single, synergistic unit, with their actions and behaviors often coordinated in a self-organized manner, requiring little or no explicit direction or a priori planning. The objective of this project is to develop and test dynamical (differential equation) models that capture the synergistic self-organization of human multi-agent coordination and deploy these models in artificial agents (virtual and robotic agents) to create highly robust and mutually responsive coupled human-machine systems. The project is designed to advance a new framework for the development of coordinated human-machine systems, one that is derived from empirical evidence of the physical, informational and biomechanical processes that shape and constrain the dynamics of successful and adaptive joint-action behavior in humans. Using a set of object moving and passing tasks, the specific aims of the project are to demonstrate (1) how dynamical models of the synergistic coordination that occurs between pairs of human agents performing a physical joint-action task can serve as the basis for artificial agent performance and (2) can be implemented in systems of interacting human and artificial agents to produce stable and adaptive patterns of robust behavioral coordination equivalent to that observed during human-human interaction. The proposed research integrates contemporary methods from behavioral, cognitive and social psychology, human movement science, computer-science, complexity science, engineering, and philosophy and will advance our fundamental understanding of the behavioral dynamics of human joint action. Moreover, the proposed project will have broad and transformative implications for the development and design of interactive robotic and artificial agent based assistive, diagnostic, and therapeutic sensorimotor technologies.