RAPID: Tracking Urban Mobility and Occupancy under Social Distancing Policy

该项目通过利用信息技术收集有关人们如何在城市中移动以及在应该保持社交距离时如何使用公共空间的信息，促进科学和公共卫生的进步。保持社会距离政策是为了减缓新冠肺炎等传染病的传播。保持社交距离在预防疾病方面的有效性取决于人们是否以及如何遵守这些命令。通过收集城市人行道和公共空间的视频，我们可以了解人们如何解释和回应社会距离命令，然后，我们可以展示这些行为如何影响一个又一个社区的健康结果。随着冠状病毒传播到其他地区，有关特定行为与疾病传播之间关系的证据将鼓励公众遵守规定。例如，人们相信在户外运动是安全的，只要人们不要跑得太近。允许跑步和徒步旅行；然而，在公园里打篮球就不是了。本项目收集的数据将有助于建立更清晰、有证据支持的指导，说明哪些活动可能是安全和危险的。这些数据在未来的人机交互研究中也很有用。它将帮助自动驾驶汽车、送货机器人和紧急服务车辆等自动系统自动识别人们在做什么，以便他们做出适当的反应。这可以应用于多种目的：它可以帮助应急响应团队快速定位需要帮助的人。它可以帮助汽车和机器人更好地理解不同类型的人类活动，这样它们就可以等待短暂的情况过去，或者避开持续时间较长的活动。该项目将收集纽约户外行人和轻型车辆（自行车、踏板车、滑板车）活动的数据，数据来源包括：a)行驶在城市中的车辆的行车记录仪录像；b)公共网络摄像头收集的视频流；c)当地公民自愿提供的手机地理定位数据，以形成社会距离政策下的城市交通和空间占用情况地图。这些数据将使研究人员能够推断公共空间中人们的活动、背景、来源和目的地。这些信息可以揭示在哪里，以及为什么呆在家里的命令被执行和没有被执行。它还可以帮助确定哪些领域的基本服务或活动没有得到最佳分配或设计，以减少不必要的互动。这些数据还可用于政策指令和感染模式的事后分析，从而更好地为政策和社会距离设计提供信息。它促进了我们对公共政策如何转化为实地行为和活动，以及个人行为和社会互动的出现如何影响社会健康结果的理解。这项工作展示了当今计算机识别和移动技术在捕捉人类行为方面的应用，将改进和验证传染病预防模型，也将为公共政策提供信息。此外，它将增强机器人社区识别城市空间中人类活动的能力。通过建模社会距离行为，我们可以更好地为未来的公共城市空间设计适合社会的人机交互。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 promotes the progress of science and public health by using information technology to collect information about how people move through cities and use public spaces when they are supposed to be social distancing. Social distancing policy is intended to slow the spread of infectious diseases such as COVID-19. The effectiveness of social distancing in preventing disease depends on whether and how people follow these orders. By collecting video of city sidewalks and public spaces, we can understand how people are interpreting and responding to social distance orders, and, later, we can show how these behaviors affect health outcomes neighborhood by neighborhood. As the coronavirus spreads to other locales, evidence about how specific behaviors correlate with disease spread will encourage public compliance. For example, it is believed that it is safe to exercise outdoors, as long as people are not running too close with one another. Running and hiking are permitted; however, using parks to play pickup games of basketball is not. The data collected in this project will aid in establishing clearer, evidence-backed guidance for what safe and dangerous activities might be. The data will be also useful in future human robot interaction research. It will help autonomous systems, such as autonomous cars, delivery robots, and emergency service vehicles, to automatically recognize what people are doing, so that they can respond appropriately. This can be applied to a variety of purposes: It can help emergency response teams to quickly locate people that need help. It can help cars and robots to better understand different kinds of human activities, so that they might wait for momentary situations to pass, or steer around activities that will be longer in duration. This project will gather data on outdoor pedestrian and light vehicle (bicycle, scooter, skateboard) activity in New York from a) dashcam footage from vehicles driving through the city, b) video streams gathered from public web cameras, and c) mobile phone geo-location data volunteered by local citizens, to form a map of urban mobility and space occupancy under social distancing policy. This data will enable researchers to infer the activities, contexts, origins and destinations of the people in public spaces. This information can reveal where and, in turn, why stay at home orders are and are not being followed. It can also help to identify areas where essential services or activities are not distributed or designed optimally to decrease unnecessary interaction. This data can also be used in post-hoc analysis of policy directives and infection patterns, so as to better inform policy and social distancing design. It advances our understanding of how public policy translates to behaviors and activities on the ground, and how the emergence of individual behaviors and social interactions influence social health outcomes. This work demonstrates the application of current-day computer recognition and mobile technology to capture human behavior, will improve and validate models for infectious disease prevention, and also will inform public policy. In addition, it will augment the robotic community's ability to recognize human activities in urban spaces. By modelling social distancing behaviors, we can better design socially appropriate human robot interaction for public urban spaces in the future.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.