Japanese researchers develop machine-learning technique to detect natural disasters
Image credit: Pixabay
In light of recent natural disasters, such as those seen in Indonesia last month, researchers from Tokyo Metropolitan University have applied machine-learning techniques to achieve fast, accurate estimates of local geomagnetic fields using data taken at multiple observation points, potentially allowing detection of changes caused by earthquakes and tsunamis.
The researchers, Yuta Katori and Professor Kan Okubo, have developed a deep neural network (DNN) model, a method used to estimate magnetic fields to detect natural disasters early. This model could be used to develop effective warning systems that may help reduce casualties and widespread damage caused by such disasters.
There are systems for warning people just before the arrival of seismic waves already put in place. However, it is often the case that the secondary wave, that is, the later part of the quake, has already arrived when the warning is given. This therefore means a more accurate indicator is required to give residents time to seek safety and to reduce casualties.
According to researchers, earthquakes and tsunamis are accompanied by localised changes in the geomagnetic field, both of which can be detected by a network of observation points at various locations.
“This report describes geomagnetic signal changes generated by earthquakes and tsunami waves,” the study says, “Results show that detection of their occurrence using geomagnetic field measurement is effective for providing an early alarm system for disaster mitigation.”
With the machine-learning technique, the university researchers were able to take measurements at multiple locations around Japan and created an estimate of the geomagnetic field at different, specific observation points.
Using the DNN model, the Tokyo based researchers fed the algorithm a vast amount of input taken from historical measurements, resulting in the algorithm creating and optimising a complex, multi-layered set of operations that effectively maps the data to what was measured.
Katori and Okubo were able to create a network that can estimate the magnetic field at the observation point with unprecedented accuracy by using half a million data points taken over 2015.
According to the researchers: “Given the low computational costs of DNNs, the system may potentially be paired with a network of high sensitivity detectors to achieve lightning-fast detection of earthquakes and tsunamis, delivering an effective warning system that can minimise damage and save lives.”
The study has been published online in the journal IEICE Communications Express.
In August, over a dozen Caribbean countries teamed up with Virgin founder Richard Branson to launch a multi-million-dollar initiative designed to bring green tech to a region frequently blighted by hurricanes.
Also, this June saw the development of a monitoring system that detects underwater earthquakes using pre-existing cross-continent fibre cables that form the basis of the global communications infrastructure.
Sign up to the E&T News e-mail to get great stories like this delivered to your inbox every day.
