Japan – Fujitsu Introduces AI technology Enabling Highly Accurate Prediction of Vessel Collision Risks on Complex Maritime Routes

0
64

Fujitsu Limited today announced the introduction of a highly-accurate AI technology to predict vessel collision risks on complex sea routes including bay areas. Field trials were conducted between November 2020 and September 2021 under an outsourcing contract with the Japan Coast Guard to demonstrate the usefulness of the new technology.
Fujitsu previously leveraged AI technology to calculate and predict vessel collision risks based on the current position, speed and direction of the ship. Now, an additional algorithm has been added to this technology in order to calculate the degree to which the vessel is following its course, enabling more accurate risk prevention.

This new feature makes it possible to only provide alerts in situations with a high risk of collision – changes of the course along a route will thus no longer be falsely detected as dangerous steering.

Field trials of the technology were conducted at the Tokyo Wan Vessel Traffic Service Center under an outsourcing contract with the Japan Coast Guard, which manages maritime traffic control services. Results of the trial demonstrated that this technology can reduce unnecessary alerts by about 90% on all sea routes, especially on curved sections where alerts frequently occurred.

This improved technology enables the early recognition of vessels prone to risks and a speedy initial response in order to prevent collisions, while also contributing to improving maritime traffic safety by reducing excessive traffic control operations as well as human errors.

Based on the results of the field trials, Fujitsu plans to provide safe navigation support services beginning in March 2022 for players in the maritime industry globally. Through this service, Fujitsu aims to ensure both the safety of maritime traffic control as well as vessel operation and will support the construction of resilient maritime traffic systems.

Background

More than 99% of Japan’s imports and exports are conducted through seaborne trade. However, the recent impact of COVID-19 has greatly restricted people’s movement, making logistics more important than ever. At the same time, recent serious maritime accidents around the world have drawn increased attention to the importance of ensuring the safety of maritime traffic. This not only includes direct damage to the ship hull and cargo, but also indirect damage caused by lost opportunities due to inoperable ships, as well as serious damage to human life and the environment.

Many maritime accidents are said to be caused by human error. Particularly in busy sea lanes near ports and in bays, there is a need for technology that helps operators to understand the movements of vessels and provide them with information to avoid risks.

However, most of the methods to predict vessel collision risks currently in practical use take into account that vessels navigate along straight lines starting from their current positions. As a result, unnecessary alerts frequently occur when a vessel navigates on curved sections of sea routes, officially defined by laws and regulations like the Maritime Traffic Safety Act. Therefore, the correct timing of risk notifications to vessels depends on the experience and skill of the operation controller.

In fiscal 2019, under an outsourcing contract with the Japan Coast Guard, which operates a navigation support system that combines radar and the Automatic Identification System (AIS)(1) to analyze movements of vessels in order to provide them with relevant information to improve the safety of navigation, Fujitsu leveraged AI technology to predict vessel collision risks in the Tokyo Bay area and to detect other areas where collision risks are highly concentrated. Fujitsu has confirmed the effectiveness of this method for early detection of collision risks.

Fujitsu additionally conducted joint field trials with the Japan Coast Guard in order to improve maritime traffic safety and to establish a technology that can support the operations of the Japan Coast Guard.

About the newly developed technology

Until now, predictions of vessel collision risks had been calculated based on the current position, speed and direction of the ship. Fujitsu has now developed a new, additional algorithm to calculate if a ship follows a specified route (patent pending).

With conventional technology, excessive alerts had occurred in cases where two vessels were judged to travelling straight along their current course in the vicinity of curved sections of sea routes – by utilizing the new algorithm described above, collision risks will now be categorized as low when the two vessels follow a specified route in a curvilinear manner.

This new algorithm thus reduces unnecessary alerts and makes it possible to more accurately determine the risks of collision between vessels (patent pending).

Joint Field Trials with the Japan Coast Guard
1. Date and Time
November 17, 2020 (Tuesday) to September 2, 2021 (Thursday)
2. Location
Tokyo Wan Vessel Traffic Service Center (Place: Naka-ku, Yokohama City, Kanagawa Prefecture)
3. Contents and results of the field trials

In cooperation with the Japan Coast Guard, Fujitsu conducted statistical and quantitative evaluations in the Tokyo Bay area of the effectiveness of its collision risk detection technology by comparing cases where the new technology has been applied with cases where it had not been applied.

After a fixed evaluation period, Fujitsu confirmed that unnecessary alerts could be reduced by approximately 90% in comparison to conventional technology for all sea routes including sea routes with curved sections.

In addition, Fujitsu conducted a collation analysis of the operational records for the actual maritime traffic control operations of the Tokyo Wan Vessel Traffic Service Center which manages maritime traffic control services. This analysis was based on records of the operations performed by the operation controller, including the provision of information to vessels, as well as information drawn from interviews with the operation controller, and the alerts detected by the newly established technology.

As a result, about 95% of the high risk events (2) for which the Operator issued warnings or recommendations were correctly judged as high risk, and it was confirmed that this technology is close to the Operator’s judgment and useful for operational support.

Future Plans

By upgrading the conventional collision risk prediction technology with this new algorithm, Fujitsu aims to provide a safe navigation support service that can detect collision risks with high accuracy even in curved sections of sea routes by March 2022 to players in the maritime industry globally.

Fujitsu also aims to provide a service equipped with an algorithm currently under development that quantitatively evaluates whether a vessel is following a route by analyzing the characteristics of the vessel, such as its size and type, as well as data from past navigation performance by September 2023. In this way Fujitsu will be able to ensure both the safety of maritime traffic control and vessel navigation and to contribute to the construction of a resilient maritime traffic system.

(1) Automatic Identification System (AIS):
system that collects and provides information of the position, speed and course of vessels using radio waves
(2) Correct judgement of about 95% of high risk events:
53 out of 64 cases were correctly evaluated as high risk events exceeding the threshold limit value, 8 further cases were correctly judged as high risks events close to exceeding the threshold limit value

About Fujitsu

Fujitsu is the leading Japanese information and communication technology (ICT) company offering a full range of technology products, solutions and services. Approximately 126,000 Fujitsu people support customers in more than 100 countries. We use our experience and the power of ICT to shape the future of society with our customers. Fujitsu Limited (TSE:6702) reported consolidated revenues of 3.6 trillion yen (US$34 billion) for the fiscal year ended March 31, 2021. For more information, please see www.fujitsu.com.