Collaborative Research: The Role of Stress in Human Crowd Dynamics during Emergency Situations

这笔赠款旨在了解压力是如何在从太空疏散的人中传播的。在紧急情况下快速、安全地疏散空间会带来拯救生命的后果，但众所周知，压力会影响一个人快速离开大楼的能力。人们不太清楚的是，一个人自己的压力状态如何会传染给团队中的其他人，从而潜在地加剧团队压力水平和高效的团队活动。该奖项支持解决这一复杂问题的实验和数学研究。小团体和大团体的实验将在受控空间中进行，在那里将仔细测量人们的运动、压力生理(例如心率)和情绪，目的是确定压力对团体的影响。然后，实验数据将通过基于计算机的方法进行检查，以找出一个人的生理和情绪状态如何影响他们自己的疏散行为，以及群体中其他人的压力状态和疏散行为。然后，这些结果将被用于创建和测试一个模型，该模型预测在高压力情景下疏散人群的行为。这个模型将提供对人类在压力下如何一起行动的复杂因素的洞察，这可以为人群管理策略提供信息，以确保人们与同龄人和他们的环境安全有效地互动。重要的是，这个数学模型可以用来测试人们在实验室实验中不可行、但在真实紧急情况下可能发生的情况下的行为。该项目团队包括来自工程学、数学和心理学的不同研究人员，以应对这一挑战，同时作为跨学科合作的一部分，培训学生和博士后。研究成果将通过同行评议的出版物、会议演讲和当地科学博览会上的非正式公共科学经验来分享。现有的数学模型虽然能够定性地匹配疏散人群的整体运动，但它们只能部分复制个人的运动，这表明它们缺乏关键变量。该项目将使用最先进的设备来跟踪多个人的位置，并使用多模式测量来索引生理和情绪状态，以高保真地重建个人的压力状态及其对他人压力和运动动力学的影响。这个项目跨越了数学、工程学和心理学的洞察力，将阐明在紧急疏散过程中生理和情感因素--特别是压力状态和传染--在群体动力学中的作用。首先，将收集行人实验的多模式数据。对实验数据的信息论分析将量化疏散人群中跨越多个动态特征的应力信息流。这将允许剖析压力状态(生理和情感)、压力传染和运动之间的因果影响。这些关系将形成一个新的基于代理的模型，该模型包括个人的压力特征，并根据实验进行校准。模型预测将在进一步的大人群实验中得到验证。这一经过验证的模型将为疏散实验提供伦理代理，而不会导致不必要的高压力。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。

This grant seeks to understand how stress spreads among people who are evacuating from a space. Evacuating a space quickly and safely during an emergency has life-saving consequences, but it is known that feeling stressed impacts a person’s ability to exit a building quickly. Less understood is how one’s own stress state can contagiously spread to others in a group, potentially exacerbating group stress levels and efficient group movement. This award supports experimental and mathematical research that addresses this complex problem. Experiments with small and large groups will be performed in a controlled space where people’s motion, stress physiology, e.g., heart rate, and emotions will be carefully measured with the aim of determining stress contagion effects on the group. Experimental data will then be examined with computer-based methods to find patterns in how a person’s physiology and emotional state influence their own evacuation behaviors, as well as the stress states and evacuation behaviors of others in the group. These results will then be used to create and test a model that predicts the behavior of an evacuating crowd in high-stress scenarios. This model will offer insight into the complex factors that govern how humans move together under stress, which can inform crowd management strategies to ensure that people interact safely and efficiently with both peers and their environments. Importantly, the mathematical model can be used to test how people behave in scenarios that are unfeasible in laboratory experiments, but which can occur during real emergencies. The project team includes a diverse group of researchers from engineering, mathematics, and psychology to address this challenge while training students and a post-doc as a part of an interdisciplinary collaboration. Research results will be shared via peer-reviewed publications, conference presentations, and informal public science experiences at local science fairs. Existing mathematical models, while able to qualitatively match the overall motion of an evacuating crowd, they can only partially replicate an individual’s movement, suggesting that they lack critical variables. This project will use state-of-the-art facilities that track multiple people’s positions, with multi-modal measures which index physiological and emotional states, to reconstruct an individual’s stress state and its influence on others’ stress and movement dynamics with high fidelity. Bridging insights across mathematics, engineering, and psychology, this project will clarify the role of physiological and emotional factors–specifically, stress states and contagion–in group dynamics during emergency evacuation. First, multi-modal data from pedestrian experiments will be collected. Information-theoretic analyses of experimental data will quantify the flow of stress information within the evacuating crowd across multiple dynamic features. This will allow dissection of causal influences between stress state, both physiological and emotional, stress contagion, and movement. These relationships will inform a novel agent-based model that includes individuals’ stress features, calibrated to experiments. Model predictions will be tested in further experiments with large crowds. This validated model will provide an ethical proxy for evacuation experiments without inducing undue high stress.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.