Projects / Delivering Earthquake Warnings Using Smartphones

Delivering Earthquake Warnings Using Smartphones

Intelligent Infrastructure

Creating intelligent infrastructures to manage scarce resources and to help realize the social, cultural and economic potential of communities.

Delivering Earthquake Warnings Using Smartphones

Synopsis
Researchers are teaming up to deliver early earthquake warnings with smartphones in California. The project leverages the recent development of warning systems using traditional seismic networks (ElarmS by Prof. Richard Allen) and using smartphones to capture and collect acceleration timeseries (iShake by Prof. Alex Bayen). By linking these two efforts to both enhance the performance of the warning system and start delivery of warnings to individuals’ smartphones, the researchers aim to reduce the overall impact of earthquakes, both on industrial machinery and on people’s safety.

The project focuses on 1) analyzing the acceleration data recorded by iShake to determine if it can be used as part of the ElarmS earthquake detection and shaking prediction system. 2) Developing the necessary real-time processing thread to extract the relevant information from the smartphone data and combine it with existing analysis that uses the seismic networks. And 3) continuing to develop the iShake app to receive shaking warnings from the integrated systems and then alert the phone users.

Details
We are exploring the use of accelerometers in smartphones to record earthquakes. We have developed an application for Android phones based on previous work with iPhones to record the acceleration in real time. These records can be saved on the local phone or transmitted back to a server in real time. A series of shake table tests were conducted (and more tests will be conducted soon) to evaluate the performance of the accelerometers in these smartphones by comparing them with high quality accelerometers. We also recorded different human activities using these smartphones. Different features were extracted from the recordings and were used to distinguish earthquakes from daily activities. We implemented a classifier algorithm based on an artificial neural network, which shows a 99.7% successful rate for distinguishing earthquakes from certain typical human activities.

Two kinds of smartphones have been used in this research: iPhone and Android phones. The applications on these phones are iShake and droidShake. Data was collected mainly in three ways: (1) Continuous recording of different human activities, e.g. walking, running, sitting, taking the bus, etc. (2) Trigger-based data from various users sent to a server. This method requires that the phone stay steady for certain amount of time. Then, if the acceleration exceeds the pre-determined threshold, it triggers the algorithm to send data before and after the trigger to the server. (3) Data recorded during the shake table tests with earthquake input signals. These three types of data were used to distinguish earthquake signals from non-earthquake signals.

This initial study shows the potential of using smartphones to detect earthquakes. By using multiple phones in the future, we can achieve higher accuracy. A network consisting of these smartphones may work as a supplemental network to the current traditional network for scientific research and real-time applications.

Published Material

Qingkai Kong, Richard Allen, Stephen Thompson, Jonathan D. Bray, Ana Luz Acevedo-Cabrera. “Using Smartphones to Detect Earthquakes,”

Available at http://seismo.berkeley.edu/annual_report/ar11_12/2012rs24.pdf

Richard Allen. “Transforming Earthquake Detection?” Science Vol 335, 20 January 2012. Available at http://seismo.berkeley.edu/~rallen/pub/2012allen/AllenScience2012.pdf