A Neyman-Pearson approach to the development of low cost earthquake detection and damage mitigation system using sensor fusion

Damage mitigation of critical infrastructures onset of an earthquake is a very important matter. In our work, we developed an earthquake warning and protection system through P-wave sensing that detects earliest onset of an earthquake before damaging ground shaking occurs. Our approach captures P-wave using redundant channels (dual sensors) and minimizes the possibilities of false triggers from harmless, momentary ground shaking. The system comprises a piezoelectric detector and a digital three-axis silicon accelerometer. The core of this study focuses on the detection algorithm using the collected data to decide if an earthquake is occurring. The algorithm is adaptive and is based on a probabilistic approach over a sliding window to maximize the detection of events, keeping the probability of false alarms under a fixed rate. Finally experimental results are provided showing the system meets the expected performance with both artificial and real seismic stimuli. One possible extension of our approach could be implementing geological wireless sensor networks using contemporary smartphones for data acquisition.