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

Regents Of The University Of Michigan - Ann Arbor, Ann Arbor MI

Investigators

Abstract

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.