CAREER: Psychology-aware Human-in-the-Loop Cyber-Physical-System (HCPS): Methodologies, Algorithms, and Deployment

人类和智能技术之间的无缝互动代表着一个重要的机遇，但它也带来了一系列令人生畏的挑战。了解CPS和人类之间的复杂相互作用对于塑造未来CPS技术的最终社会成果至关重要，特别是在人-网络-物理系统(HCPS)领域。在这种情况下，CPS所做的决策可能会对智能城市、交通系统和人群管理等环境中的个人产生不同的影响，构成我们所说的“多人-CPS(MHCPS)”。特别是，MHCPS概括了在同一环境中共存的具有相互冲突的偏好和不同观念的大量个体，受社会心理动态和CPS的决策过程的影响。MHCPS由两个相互交织和耦合的动力组成，即生理心理和社会心理，后者概括了不断变化的社会互动模式，这些模式塑造了个体沿着利他、亲社会、中立或反社会行为的连续体的思想、感觉和行为。研究强调了社会心理状态对整体系统性能的影响，其中亲社会行为的增加与系统遵从性和接受度的提高相关。这项研究的目标是设计CPS决策算法，既能理解物理系统，也能理解人类与这些系统交互的社会心理本质。这种心理感知的决策算法将实现和谐的互动，促进公平，并优化整体表现。这个NSF职业项目旨在通过将社会-心理动态和社会紧张融入工程解决方案来弥合当前CPS工程实践中的差距，以应对社会规模CPS当前和未来的挑战。这项研究引入了新的算法、模型和工具，将这些动态集成到工程解决方案中，极大地扩展了研究人员和从业者可用的研究工具箱。该研究包括：(I)开发新的人类建模技术以捕获多人CPS(MHCPS)应用中的社会心理动态，(Ii)将CPS决策的目标形式化以增强MHCPS中的社会动态，(Iii)设计神经符号顺序决策算法，以考虑MHCPS中的物理和社会动态以及约束，(Iv)创建因果意识偏好学习算法以捕获MHCPS中的权衡和社交紧张，(V)在现实生活应用中部署决策算法，考虑部署/资源约束及其对结果系统的影响，以及(6)实施系统化的软件框架，以促进实际执行管理和预防保健方案。这项研究可以显著影响工程研究的方向，以增强的社会考量促进对CPS技术的理解和发展。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The seamless interaction between humans and smart technologies represents a significant opportunity, but it also presents a daunting set of challenges. Understanding the intricate interactions between CPS and humans is essential for shaping the ultimate societal outcomes of future CPS technologies, particularly within the realm of Human-Cyber-Physical Systems (HCPS). In this context, decisions made by CPS can yield disparate impacts on individuals within settings like smart cities, traffic systems, and crowd management, constituting what we term as "Multi-Human-CPS (MHCPS)." In particular, MHCPS encapsulates a multitude of individuals with conflicting preferences and diverse perceptions coexisting within the same environment, influenced by both social-psychological dynamics and the decision-making process of CPS. MHCPS consists of two intertwined and coupled dynamics, physical and social psychology, the latter encapsulating the ever-changing patterns of social interactions that mold individuals' thoughts, feelings, and behaviors along the continuum of altruistic, prosocial, neutral, or antisocial behavior. Research underscores the impact of social-psychological states on overall system performance, where increased prosocial behavior correlates with heightened system compliance and acceptance. The objective of this research is to design CPS decision-making algorithms that understand both the physical system and the social-psychological nature of human interactions with these systems. Such psychology-aware decision-making algorithms will enable harmonious interactions, foster equity, and optimize overall performance.This NSF CAREER project aims to bridge a gap in current CPS engineering practices by incorporating social-psychological dynamics and societal tensions into engineering solutions to address current and future challenges of societal scale CPS. This research introduces new algorithms, models, and tools that integrate these dynamics into engineering solutions, significantly expanding the research toolbox available to both researchers and practitioners. The research entails: (i) developing novel human modeling techniques to capture social psychological dynamics in Multi-Human CPS (MHCPS) applications, (ii) formalizing objectives for CPS decision-making to enhance social dynamics in MHCPS, (iii) designing neurosymbolic sequential decision-making algorithms that account for both physical and social dynamics and constraints in MHCPS, (iv) creating causality-aware preference learning algorithms to capture tradeoffs and social tensions in MHCPS, (v) deploying decision-making algorithms in real-life applications, considering deployment/resource constraints and their impact on the resulting system, and (vi) implementing systematic software frameworks to facilitate the real-life implementation of MHCPS. This research can influence the direction of engineering research significantly, advancing the understanding and development of CPS technologies with enhanced social considerations.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.