Uncertainty as to when a ship will turn up at the quay has been a never-ending irritant for ports and harbours the world over. But the problem will soon be history thanks to artificial intelligence. As a first mover, Oslo Port Authority in Norway is this spring set to implement an innovative machine learning algorithm that reliably calculates the estimated time of arrival (ETA) of vessels sailing from near and far, in a move that promises significant savings for stakeholders involved in port calls as well as other efficiencies beyond the port gates.
A problem looking for a solution
Ships have been transporting passengers and merchandise for centuries, and today 90% of world trade travels by sea. It is by far our most economical means of transport. But the uncertainty linked to the actual (versus theoretical) arrival time of vessels in port endures as a major bugbear for port authorities and quayside service providers. Late arrivals especially represent the biggest source of disruption in logistics chains that involve a sea leg. Avoiding such bottlenecks is a problem itching to be solved. In Europe, estimates of arrival time are today keyed in manually by a shipping agent or the ship’s crew via public reporting portal SafeSeaNet (SSN), but the platform is rarely updated and data is often missing.
On a mission to make the day-to-day business of customers like the Port of Oslo much more predictable, technology experts Grieg Connect have plugged the gap with a state-of-the-art solution leveraging the power of AI. Various machine learning methods are applied to analyze historical AIS vessel traffic data together with the position, course and speed of vessels in real time, as well as environmental and weather data. The result is a tool that calculates ETA automatically and with remarkable reliability.
Competitive traction for shipping
«This technology not only adds commercial value for us as a port, but also increases efficiency and predictability for our customers, and everybody else who supplies goods and services at the port,» said Jens Petter Christensen, Port Captain Harbour Master at the Port of Oslo.
“As a supplier of enterprise system solutions for ports and terminals, it’s vital that we always stay ahead of the game with the latest technology. This innovation provides an accurate source of ETA information that can boost the competitiveness of the maritime transport industry while at the same time helping the environment. The risk of shipping lagging behind is losing market share to logistics operators offering freight by road and rail, which are not nearly as energy efficient,” said Grieg Connect CEO Vidar Fagerheim.
Sights set on ‘Just in Time’
As frontline transport hubs, ports need to address both the challenges and opportunities presented by digitalization. Machine learning especially can be applied in ways that strengthen maritime transport in terms of increased efficiency, safety and, not least, the environment.
Work on creating an algorithm that could accurately predict the ETA of ships began in 2018 as part of a collaboration involving the Port of Oslo as well as the ports of Stavanger, Ålesund, Kristiansund and Nordmøre, and Bodø. The development project, entitled ‘Just in Time’, was partly funded by the Norwegian Coastal Administration while Grieg Connect as technology partner and solution owner invested considerable funds of its own.
The term ‘Just in Time’ is borrowed from the automotive industry and describes the organizational principle that ensures all necessary parts are in the right place at the right time, at every stage of production. By analogy, if you look at a port call as a point on the maritime assembly line, the host port has a key role both in providing its own services, as well as facilitating those of its partners and third parties, in a similar targeted manner. The goal was and remains efficient port management and turnarounds with prompt quayside service delivery based on accurate ETA prediction.
A multitude of applications
The algorithm has further strong commercial benefits. For individual vessels, using precise ETA to optimize sailing time can reduce operating costs through fuel savings, while also reducing carbon footprint with lower emissions. Shipping agents, vessel managers and shipping companies will be able to leverage the tool in multi-party communications to enhance central planning. It could also facilitate new interactive business models for price differentiation based on traffic distribution – similar to energy flow optimization prevalent today in the power management industry. Moreover, customers could potentially pay less if ships were to call at a port at a time the port authority deems most convenient. Ports can win competitive edge by incentivizing ships to call at the best time, perhaps slowing down if necessary, creating a feedback loop where customers are also happy.
A new level of safety
From a safety perspective, the ETA algorithm is especially useful in triggering real-time warnings over vessels behaving abnormally based on historical data (anomaly detection). This significantly improves the ability to prevent potential powered groundings, collisions and other accidents. This added functionality will be offered to public authorities, Vessel Traffic Services (VTS) and other security operators who monitor shipping traffic.