The next big one: Detecting earthquakes and other rare events from community-based sensors

Can one use cell phones for earthquake early warning? Detecting rare, disruptive events using community-held sensors is a promising opportunity, but also presents difficult challenges. Rare events are often difficult or impossible to model and characterize a priori, yet we wish to maximize detection performance. Further, heterogeneous, community-operated sensors may differ widely in quality and communication constraints. In this paper, we present a principled approach towards detecting rare events that learns sensor-specific decision thresholds online, in a distributed way. It maximizes anomaly detection performance at a fusion center, under constraints on the false alarm rate and number of messages per sensor. We then present an implementation of our approach in the Community Seismic Network (CSN), a community sensing system with the goal of rapidly detecting earthquakes using cell phone accelerometers, consumer USB devices and cloud-computing based sensor fusion. We experimentally evaluate our approach based on a pilot deployment of the CSN system. Our results, including data from shake table experiments, indicate the effectiveness of our approach in distinguishing seismic motion from accelerations due to normal daily manipulation. They also provide evidence of the feasibility of earthquake early warning using a dense network of cell phones.