SCC-IRG Track 1 - Behavior-driven Building Safety and Emergency Management for Campus Communities

校园社区容易受到各种突发事件的影响，比如活跃的枪手事件和火灾，使学生、教师和其他成员面临重大风险。这个智能和互联社区综合研究资助（SCC-IRG）项目旨在探索将人类行为系统和稳健地纳入建筑设计和应急协议的新方法。具体来说，该项目的重点是了解不同角色的人如何通过从人体实验、行为理论和领域专家的见解中收集的经验数据，个人和集体地应对建筑紧急情况。智能人群模拟被开发出来，以代表不同群体的建筑居住者的目标、偏好和行动，以及他们与他人和环境的互动。研究团队与一系列校园社区利益相关者协同工作，包括第一响应者，执法和应急管理人员，以及建筑设计师，以参数化，验证和测试人群模拟。确定适当的应用程序以减轻安全和健康风险。研究活动招募了代表性不足和贫困的学生。此外，该团队还探索了方法、数据集和研究成果在非建筑突发事件和非校园社区中的普遍性。目前用于检查建筑物紧急情况的人群模拟通常依赖于对人类行为的过度简化的假设，缺乏对不同紧急情况的交叉检查，并且忽略了利益相关者的各种需求。为了解决这些差距，本项目考虑了校园社区中常见的两种常见但独特的建筑紧急情况（即活跃枪手事件和火灾），并对个人、社会和环境因素对建筑物居住者人类行为的影响进行了整体建模。该团队实现了基于智能体的模型，并开发了基于高度可重构和可适应的决策理论框架的多智能体人群模拟，以捕获个人和集体行为。该团队还让校园利益相关者和其他社区成员参与一系列研讨会，共同创造可能的和有代表性的紧急情况。通过探索这种“假设”场景的结果，人群模拟软件有可能成为建筑设计师、设施操作员和应急管理人员的宝贵工具。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。

Campus communities are vulnerable to a wide range of emergencies such as active shooter incidents and fires, exposing students, teachers, and other members to significant risks. This Smart and Connected Communities Integrative Research Grant (SCC-IRG) project aims to explore new ways in which human behaviors are systemically and robustly incorporated into building design and emergency protocols. Specifically, the focus of the project is to understand how people in different roles respond to building emergencies, both individually and collectively, with empirical data collected from human-subject experiments, behavioral theories, and insights from domain experts. Intelligent crowd simulations are developed to represent the goals, preferences, and actions of diverse groups of building occupants as well as their interactions with others and the environment. The research team works synergistically with a range of campus community stakeholders, including first responders, law enforcement and emergency management personnel, and building designers to parameterize, validate, and test the crowd simulation. Appropriate applications are identified to mitigate safety and health risks. Underrepresented and underprivileged students are recruited into the research activities. Furthermore, the team explores the generalizability of the methodology, datasets, and research findings to non-building emergencies and non-campus communities.Current crowd simulations for examining building emergencies often rely on oversimplified assumptions about human behaviors, lack cross-examination in different emergency contexts, and overlook the varied needs of stakeholders. To address these gaps, this project considers two common yet distinctive building emergencies (i.e., active shooter incidents, and fires) that are common in campus communities, and holistically models the impact of personal, social, and environmental factors on human behaviors of building occupants. The team implements agent-based models and develops multi-agent crowd simulations that capture individual and collective behaviors based on a highly reconfigurable and adaptable decision-theoretic framework. The team also engages campus stakeholders as well as other community members in a series of workshops to co-create likely and representative emergency scenarios. By allowing for exploring the outcomes of such “what-if” scenarios, the crowd simulation software has the potential to serve as a valuable tool for building designers, facility operators, and emergency managers.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.