Singapore and IBM Test Artificial Intelligence Systems
The Maritime and Port Authority of Singapore (MPA) and IBM have jointly completed a three-month trial of artificial intelligence systems designed to improve safety and efficiency in the port.
The trial involved three modules of the MPA-IBM SAFER project, which will be rolled out progressively beginning in September 2017.
Project SAFER, derived from “Sense-making Analytics For maritime Event Recognition,” is a collaboration between MPA and IBM to develop and test new analytics-based technologies aimed at supporting increasing Singapore’s growth in vessel traffic.
There are seven modules under project which offer new capabilities for automating and increasing the accuracy of critical tasks that previously relied on human observation, reporting, Very High Frequency (VHF) communication and data entry. The trial involved three modules: Automated Vessel Movement Detection, Infringement Analytics and Pilot Boarding Detection.
Automated Vessel Movement Detection
MPA’s Port Operations Control Centre handles more than 1,000 vessel movements daily in Singapore’s port waters. One of the many roles of the Vessel Traffic Management officers is to enter the start and end time and location of vessels into the Port Traffic Management System whenever a ship master reports their movement over the radio system.
Using cognitive and analytics technologies to detect and predict vessel movements, the new module reduces radio communication between MPA and ship masters and eliminates the need to enter ship movement details by automatically detecting the start/end time and location of vessel in real time. In addition, SAFER improves the accuracy of the information in movement time and location by up to 34 percent and frees up officers to carry out their other roles.
MPA’s port inspectors enforce regulations on marine safety and environmental protection and coordinate and respond to any marine incidents in the port. Common infringements include operating in port waters without a valid permit, switching off transponders and speeding.
Previously, inspectors were guided by their intuition to look for suspicious activities rather than quantitative information when patrolling the waters. With the machine learning based analytics and vessel prediction models developed for the SAFER system, they are now able to detect suspicious or abnormal vessel behavior through alerts that are be sent them. This enables them to take a more targeted approach and improve efficiency.
Pilot Boarding Detection
MPA works closely with PSA Marine (PSAM) to ensure that 95 percent of vessels requiring pilotage service are served within 15 minutes. Currently, MPA conducts audit checks when there is an appeal or dispute. The SAFER module enables MPA to automatically detect the pilot boarding time, thus validating PSAM’s pilotage service level.
The other modules, to be rolled out in January 2018, will provide advanced information on traffic density within Singapore port waters; detect illegal bunkering activities; detect vessels moving into prohibited areas and predict vessel arrival time.