Last Tuesday, an earthquake with a magnitude of 5.1 on the Richter scale occurred in California, near San Francisco Bay. If this type of earthquake is quite common there, what is a little less so is that more than a million Android users have received notifications on their devices to warn them that an earthquake earth “is knocking”.
When an AI anticipates earthquakes
How does this system work? Concretely, the latter does not really make it possible to predict an earthquake but it detects the first signs of this shock before its effects are felt by the public. Warning citizens can therefore allow them to take shelter.
In detail, the United States Geological Survey has developed ShakeAlert, a system of sensors that makes it possible to feel seismic waves called P waves through the ground. Depending on the criteria retained by the device, the latter determines whether S waves, of stronger intensity, are present, because the latter can cause greater damage.
The alert is then sent to the government tools and that of Google, which can then come into action. The efficiency is all the greater as the users do not have to install any particular application but they receive a notification which informs them of the danger.
And the initiative of the Mountain View firm does not stop there. According to our colleagues from Wired, the Tech giant has indeed decided to transform Android smartphones into miniature seismic sensors. Each telephone is thus equipped with an accelerometer which can pick up the signals of an earthquake. It then sends a message to a detection server along with its location data. What allow a more precise reconstruction of the place where the earthquake occurs.
Note that this is of course not the first time that new technologies have been used to prevent this type of natural disaster. We were talking to you, for example, in 2020 about this AI developed by the University of Science and Technology of China and the China Earthquake Administration. This system is able to predict earthquakes shortly before the effects are felt.
Based on very large volumes of seismic data, this device seems very promising. According to the first results dating back to the time, he was able to predict earthquakes more than effectively than humans out of a total of 446 earthquakes.