Fujitsu AI is better at managing Tokyo Port than humans

The future of intelligent systems is not in mobile gadgets, but in the reorganization of complex and large-scale processes in the economy. One of the first is to solve the problems of logistics - according to the Japanese authorities, 280 emergencies occur annually in the extremely busy Tokyo Bay. To prevent them, Fujitisu Zinrai, a specialized AI that monitors safety at sea, has been working here since last year.

Port AI entered service in December 2019, and in March 2020, the results of the first stage of its tests were summed up. It turned out that the electronic brain issued twice as many warnings and orders to change course to ships than human controllers usually do. At the same time, AI is not reinsured - on the contrary, the speed of passage of ships through the water area has the same priority in its settings as traffic safety. The AI's task is to prevent critical indicators for each of the port subsystems.

Zinrai constantly monitors the parameters of ships and simulates routes and characteristics of their movement. This makes it possible to calculate in advance the relative position of all objects in the water area, predict the appearance of risk zones and separate ships. On average, AI predicts the creation of difficult situations two minutes faster than a person, which gives tremendous time to maneuver. The accuracy of giving instructions to individual ships is also increased.

The final decisions are still made by people, they are also responsible for giving commands, the task of AI is reduced only to the analysis and forecasting of critical situations. And based on the results of the tests, the port authorities decided to leave the system and expand the area of ​​its operation. In the future, similar systems will appear in all ports of the country.