Smart machines and AI processes are proving their worth in the ferry and container shipping sectors
In December 2018, Rolls-Royce and Finferries demonstrated the capabilities of autonomous technology on board Falco, inviting 80 guests and journalists along for the ride.
Using Rolls-Royce Ship Intelligence technologies, Falco sailed into the busy waters of the Turku archipelago and detected and manoeuvred autonomously around several obstacles.
On the return leg of the Parainen-Nauvo route, the ferry was operated entirely under remote control by a captain in an office building in the centre of Turku.
Falco is a 53.8-m long double-ended car ferry equipped with twin azimuthing Rolls-Royce thrusters, and has been in service with Finnish state-owned Finferries since 1993.
Finferries is a partner in developing autonomous shipping technologies produced by Rolls-Royce Ship Intelligence along with Svitzer, Intel, ESA, Inmarsat, AXA, Google and Bosch. These technologies have come together in a range of sensors attached to Falco, which gave the vessel a 3D picture of its surroundings. Data from these sensors is relayed to Finferries’ remote operating centre, from which a captain can take control of the vessel if necessary.
Rolls-Royce has racked up 400 hours of sea trials of its technology in the Turku archipelago, including of the autodocking system, which allows the vessel to automatically alter course and speed when approaching the quay and dock without human intervention.
During the trial, the vessel undocked itself using Rolls-Royce’s autodocking system. Using a suite of technologies under the research project Safer Vessel with Autonomous Navigation, Falco used these systems to avoid a series of vessels in its path.
Finferries’ chief executive Mats Rosin added “As a modern shipowner our main goal in this co-operation has been on increasing safety in marine traffic as this is beneficial for both the environment and our passengers. But we are equally excited about how this demonstration opens the door to the new possibilities of autonomous shipping and safety.”
On the return trip, Falco was remotely operated by the Finferries Remote Operation Centre in Turku by Capt Tuumas Mikkola. He successfully steered Falco by remote control to the docking point, where the vessel used the automatic docking system to berth.
Meanwhile, Tallink Grupp is to take part in a European Space Agency-funded project to develop autonomous vessel navigation techniques.
Aalto University, the Finnish Geospatial Research Institute, Fleetrange and Tallink aim to use sensors, machine learning and artificial intelligence (AI) to develop the new techniques, with an emphasis on safety.
The Finnish Geospatial Research Institute’s research manager Dr Sarang Thombre said “A single sensor is never sufficient for providing complete safety-critical information to the crew.
“They always refer to multiple devices providing overlapping information so that defects in any one device can be easily identified and excluded.
“An autonomous navigation system should work on a similar principle.”
The research project will see Tallink Grupp’s newest vessel Megastar used for practical field tests on the Helsinki-Tallinn route on the Baltic Sea.
Data from a range of sources including satellite navigation, vessel transponders and sensors installed on Tallink’s Megastar ferry – including visual, audio, radar and LiDAR – will be processed using AI and machine-learning software, with the goal of automatically identifying objects such as navigation aids and other vessels to improve situational awareness.
“When such information is combined with established vessel navigation rules and regulations, it can potentially enable the vessel to navigate with minimal human guidance even in dense traffic conditions of the Baltic Sea,” said Dr Thombre.
As vessel sizes increase, logistics processes around container shipping become ever more complicated.
Several innovations are coming to the fore aimed at harnessing the power of machine learning and artificial intelligence to optimise logistics processes, and as a global transhipment hub, it is not surprising that Singapore’s maritime cluster is at the fore in developing these.
Global trade balances result in larger numbers of containers being shipped to Europe and the US from Asia than returning. This means empty containers are taking up space in western ports, while Asian terminals and factories suffer from a lack of available container stock.
An attempt to solve this issue is being developed by the Centre of Excellence in Modelling for Next Generation Ports (C4NGP), a collaboration between the National University of Singapore, the Singapore Maritime Institute and industry partners including Softship Data Processing, supported by Singapore’s MPA.
“This is not a new problem, but against the current backdrop of squeezed container rates, tight profitability and the relentless growth in containerised trade, it is beginning to become a significant headache,” says Softship Data Processing’s Asia-Pacific managing director Lars Fischer. The sheer numbers involved in terms of levels of container stock, distance and shipping rates make it extremely difficult to establish a cost-effective solution using manual methods such as spreadsheeting, he adds.
C4NGP’s solution is a mathematical algorithm that can configure supply and demand scenarios, allowing software to assess all repositioning options within set parameters and arrive at the best option.
“It will present a helicopter view of box positioning allowing operators to visualise exactly where their containers are and how best – or if – to reallocate them,” said Mr Fischer.
“The aim is to reduce the time carriers spend transporting 'fresh air' in empty containers thereby reducing costs and increasing container usage and turnaround.”
Elsewhere in Singapore, Cyberlogitec Global is developing methods of applying machine learning to optimise vessel stowage and terminal operations.
The company’s OPUS Stowage system incorporates an optimisation engine, which plans out stowage on board container vessels. The engine applies heuristic algorithms to automate the stowage process, factoring in variables such as container destination, cargo weight, ship stability, ballast optimisation and more, to optimise stability, stowage capacity and efficiency, reducing wastage and improving planning productivity. Cyberlogitec Global business service team manager Wai Mung Low notes that for a vessel capable of carrying 5,000-8,000 containers, a planner might need three hours or more to produce a stowage plan, while the OPUS Stowage system can produce an optimised plan in as little as 20 minutes.
The OPUS Terminal system can bring similar benefits for terminals. It incorporates a configurable, pattern-based automated planning engine that uses algorithms to calculate optimal and safe loading and discharging activities of vessels, among other features.
The result is shorter port calls for vessels, and ports and terminals having a quicker turnaround of berths. Such solutions are seeing demand around the world: Safiport Derince opted to use the OPUS Terminal system at its newest container terminal in the Marmara region of Turkey in November 2018, while Busan New Container Terminal adopted it in March 2019.
Cyberlogitec is also developing machine learning and robotic process automation in its heuristic engines. This simplifies booking requests, allowing the system to execute unstructured document analysis and automate simple and repeatable data-entry tasks, automatically picking relevant information to register a booking out of emails from clients without filling in a form. “The shipping business is very data driven, yet carriers are finding less and less people willing to do the data-entry work,” says Mr Low, explaining that the timing is ideal for bringing this kind of technology in.