Mitsubishi Electric Develops Radar-Based Artificial Intelligence To Predict Tsunamis

mitsubishi develops radar-based artificial intelligence to predict electric tsunamis
mitsubishi develops radar-based artificial intelligence to predict electric tsunamis

Mitsubishi Electric, in collaboration with the Japan Foundation for the Support of Civil Engineering, developed an artificial intelligence technology that uses radar-detected tsunami velocity data to predict flood depths in coastal areas.

Using Mitsubishi Electric's MAISART technology, this artificial intelligence technology will produce accurate predictions for just seconds after detecting a tsunami, and will support the rapid implementation of evacuation plans to prevent possible disasters in coastal areas.

MAISART estimates the flood depth with a high hit rate immediately after a tsunami is detected.

Artificial intelligence technology; He learns the relationship between tsunami velocity and flood depths using simulations on data such as the epicenter, the degree and direction of fault fluctuations. Artificial intelligence accurately estimates flood depths with a margin of error of approximately 1 meter.3 The estimation is made within seconds after the radar detects the tsunami speed and direction. Thanks to fast estimation, it will be possible to prevent or reduce disasters as a result of the faster planning and implementation of evacuations.

Results of simulation evaluations using various test environments simulating potential earthquakes in Nankai Trench

Evaluations to date have focused on hypothetical earthquakes at the Nankai Trench, a major fault line running approximately north-east / south-west along the coast of Japan. With the technology developed, another step will be taken and hypothetical earthquakes in other parts of Japan will be examined, and studies on how tsunamis can affect various ports, as well as other coastal structures and municipalities. In addition to the displacement of faults, the study will also examine tsunamis caused by submarine landslides, which are very difficult to predict with traditional methods.

As an earthquake country, Japan is concerned about the possible damage tsunamis could cause to coastal areas. In order to implement effective evacuation measures, flood depths need to be estimated quickly and accurately before the tsunami reaches land. With traditional methods, the depth of the flood can be estimated in a matter of minutes with a margin of three meters, but the new technology developed by Mitsubishi Electric produces estimates in just a few seconds and supports the rapid implementation of evacuation plans.

Accurate estimation of flood depths requires knowledge of ocean surface currents over a wide area. Mitsubishi Electric developed the technology after learning that this information could be collected in a 50 km area using a special radar device.4 By combining the new radar technology with Mitsubishi Electric's MAISART artificial intelligence technology, it became possible to make accurate predictions of the flood in just a few seconds. .

When the new technology was first developed, it needed simulation for possible tsunami conditions (earthquake centers, fault displacement degree and direction, etc.), but artificial intelligence learned these results over time and started to predict flood depths at a high speed when a real tsunami was detected.


Rail Industry Show 2020


Be the first to comment