Collaborative Research: Project Smart-Recon: Smart Device-Enabled Reconnaissance after Earthquakes

每当发生灾难性事件时，都会在受影响地区部署侦察队，对建筑物进行目视检查。小组给建筑物贴上红色、黄色或绿色的标签，以表明它们可能的状况和允许的用途。这一过程往往需要数周时间。这项工作的核心前提是，可以利用公民对智能手机和设备的广泛拥有来自动化并显着加速这一首次侦察工作。在该项目下，将研究如何整合由传感器执行的测量，这些传感器通常是大多数现代智能设备的一部分（例如加速度计，陀螺仪等）。来推断它们在地震中的运动信息。通过提高自然灾害后评估建筑物损坏的速度和准确性，拟议的侦察技术将减少对公民的潜在危险和困难，并将节省大量费用。了解这些信息还将使第一响应者能够优化其响应，并使物理检查团队能够优先考虑其工作，从而最大限度地减少灾难后的混乱。在教育方面，该项目将对人力资源开发产生重大影响。通过将土木工程和电气工程连接起来，参与该项目的学生将获得两个学科交叉的多学科教育。为了实现该项目的目标，将开发新的算法，使智能设备能够感知（或学习）它们所处的表面类型，并利用这些知识推断地震事件期间它们的运动信息。由于每个设备的运动可能会受到二次运动（例如在表面上滑动）的污染，因此将采用信号处理技术来研究如何使用经历相关运动的多个传感器的总体观察来产生高度准确的地板运动估计。此外，亦会进行研究，探讨装置在建筑物内的位置及装置量度的参数所需的准确程度，以确保对抗震结构需求作出有意义的评估。自动化的第一次侦察工作将通过计算层间漂移率并将这些比率与已知的损坏极限进行比较来实现。因此，可以在地震事件发生后的几分钟内对建筑物的损坏程度进行电子标记。

Whenever a catastrophic event occurs, reconnaissance teams are deployed within the affected areas to conduct visual inspections of buildings. The teams tag the buildings red, yellow or green to indicate their probable condition and permitted use. This function often takes weeks. The central premise of this work is that widespread citizen ownership of smartphones and devices can be leveraged to automate and significantly accelerate this first reconnaissance effort. Under this project, research will be conducted on how to integrate measurements performed by sensors that are typically part of most modern smart devices (e.g. accelerometers, gyroscopes, etc.) to infer information about their motion during a seismic event. By increasing the speed and accuracy with which building damage may be assessed in the aftermath of natural disasters, the proposed reconnaissance technology will reduce potential hazards and hardships to citizens and will provide enormous cost savings. Knowing this information will also enable first responders to optimize their response and physical inspection teams to prioritize their efforts, thereby minimizing confusion in the aftermath of a disaster. On the educational front, this project will have a substantial impact on the development of human resources. By bridging civil and electrical engineering, the students who will work on this project will attain a multi-disciplinary education at the intersection of both disciplines.To attain the project's objectives, new algorithms will be developed to permit smart devices to sense (or learn) the type of surface they are on and use that knowledge to infer information about their motion during a seismic event. Since the motion of each device may be contaminated by secondary motion, e.g. sliding on a surface, signal processing techniques will be employed to investigate how ensemble observations across multiple sensors that experience correlated motion can be used to yield highly accurate estimates of floor motion. Studies will also be conducted to explore the necessary level of accuracy required for device location within a building and device-measured parameters to ensure a meaningful assessment of seismic structural demands. The automated first reconnaissance effort will be enabled through computation of interstory drift ratios and comparing those ratios to known damage limits. As such, it is possible to electronically tag buildings for their level of damage within minutes of a seismic event.