An Autonomous, not an Ordinary World
by Ceri Done
Autonomous vehicles are no longer the stuff of science fiction. Rather, they are a technological reality now limited, at least temporarily, by a lag between the technological development of the autonomous systems and the vehicles they control, and the regulators and other relevant market participants, such as, in the maritime sector, the IMO, Flag States, Class and insurers. However, that lag or gap is closing.
Here we present an introduction to the subject of vehicular autonomy in a maritime context and the broad legal issues arising, which we will develop in a series of focussed articles over the coming months.
What is an autonomous vehicle?
General press coverage has largely focussed on the development of autonomous cars, with the development and deployment of Tesla’s autopilot system, and the VAG’s competitor systems, being just two such examples. However, the complexity of the environment within which motor vehicles operate, their proximity to and interaction with the public as pedestrians and other road users, and the risks that entails, means that the first truly autonomous vehicle in general commercial operation might not be a car or truck, but (on a small scale at least), an aerial system or ship.
What do we mean by autonomous vehicles? “Autonomous vehicles are vehicles which can drive themselves without human supervision or input. Unmanned vehicles are vehicles which are either controlled remotely, or perhaps operate autonomously. Vehicles can also operate semi-autonomously: taking some control of aspects of their driving, whilst a human retains control of others”1. In order for a vehicle of any sort to operate autonomously, it must be able to perceive its environment, make decisions about where it is safe and desirable to move, do so and at the same time continuously monitor its own ‘health’.
There are degrees or levels of autonomy. In the maritime context, the Norwegian Forum for Autonomous Ships anticipates eight levels of autonomy:
Levels of Autonomy
|1||Direct control||Minimal automation and decision support.|
|2||Decision support||Decision support and advice to crew on bridge. Crew decides.|
|3||Automatic bridge||Automated operation, but under continuous supervision by crew.|
|4||Periodically unmanned||Supervised by shore. Muster crew if necessary.|
|5||Remote control||Unmanned, continuously monitored and direct control from shore.|
|6||Automatic||Unmanned under automatic control, supervised by shore.|
|7||Constrained autonomous||Unmanned, partly autonomous, supervised by shore.|
|8||Fully autonomous||Unmanned and without supervision.|
This is not the only scale in operation – Lloyd’s Register, for instance, has proposed a similar 7 level scale in its “Cyber Enabled Ships – Draft ShipRight Procedure”. Taking the Norwegian Forum’s scale, at present level 8 autonomy seems an unlikely eventuality, with at least some shore supervision always being required. Indeed, according to Rolls Royce’s model, autonomous ships will still need human input from land, making connectivity between the ship and the land-based “crew” essential.
The motivations for the development of autonomous technology and vehicles are many, including safety, efficiency in the use of space, fuel and crew and improvements to quality of life and work. Research and analysis by Allianz Global Corporate and Specialty (AGCS) indicates that 75% to 96% of marine accidents and that approximately 75% of the value of almost 15,000 marine liability insurance claims analysed over a five year period, equivalent to more than US$1.6 billion, can be attributed to human error. The scope for improvements in safety by reducing the human factor in shipping, and the consequential cost savings, is therefore considerable. In terms of efficiencies, by one consultant’s estimate, the cost of carrying crew can account for up to 44% of a ship’s costs. Whilst that might be an outlier at the upper end of the scale, there is no doubt that the changes in ship design that autonomy enables, by the removal or reduction of the accommodation block and ballast systems, to name but two design areas, will yield operational savings.
The Maritime Revolution
In the maritime context, autonomous and unmanned technology has been in use in underwater and surface settings for some time, but not in a broad commercial cargo and passenger carrying context.
In a commercial context, to quote Mikael Mäkinen, President – Marine, of Rolls Royce, “Autonomous shipping is the future of the maritime industry. As disruptive as the smartphone, the smart ship will revolutionise the landscape of ship design and operations”. Rolls Royce, working with Deltamarin, DNV GL and Inmarsat on the Advanced Autonomous Waterborne Applications project (AAWA), aim to have a proof of concept demonstrator by the end of this year, a remote controlled ship in commercial operation by 2020 and an autonomous unmanned ocean going ship in operation by 2035.
Rolls Royce is far from being the only player in this market. The European Commission is co-funding a research project, MUNIN, which “aims to develop and verify a concept for an autonomous ship, which is defined as a vessel primarily guided by automated on-board decision systems but controlled by a remote operator in a shore side control station”2 and in 2016 the Norwegian Government designated the Trondheim Fjord as an official test site to aid in the development of this technology. Other exciting examples include the “YARA BIRKELAND”, the world’s first zero emission autonomous container feeder ship, developed by the fertiliser producer Yara International ASA, in partnership with Kongsberg Gruppen ASA. The “YARA BIRKELAND” is slated for launch in 2018 and will sail first under remote control (2019) and then autonomously (2020) between three ports in southern Norway.
NYK has also recently revealed plans to test a remote controlled vessel across the Pacific Ocean in 2019 and BHP Billiton Ltd. is also studying the development of autonomous vessels. “Safe and efficient autonomous vessels carrying BHP cargo, powered by BHP gas, is our vision for the future of dry bulk shipping” wrote BHP’s vice president of freight Rashpal Bhatti recently.
The widespread adoption of autonomous technology depends, in large part, upon the development of well-defined regulatory frameworks, and the development of trust in the safety and security of the systems. Incorporating autonomous or unmanned vehicles into the existing legislative and regulatory frameworks presents a significant legal challenge but it is not an insurmountable challenge by any means.
Work is already underway at the IMO and at the June meeting of its Maritime Safety Committee, the MSC agreed to include the issue of autonomous surface ships on its agenda, by way of a scoping exercise to determine how the safe, secure and environmentally sound operation of Maritime Autonomous Surface Ships (MASS) may be introduced in IMO instruments. According to the IMO “The scoping exercise could include identifying: IMO regulations which, as currently drafted, preclude autonomous / unmanned operations; IMO regulations that would have no application to autonomous / unmanned operations (as they relate purely to a human presence on board); and IMO regulations which do not preclude unmanned operations but may need to be amended in order to ensure that the construction and operation of MASS are carried out safely, securely and in an environmentally sound manner”.
Autonomous systems and the algorithms at their core will have to meet strict criteria as to safety and collision avoidance, criteria that are in development. To that end, in 2015 the MAXCMAS project (MAchine eXecutable Collision regulations for Marine Autonomous Systems) was launched, with the aim of developing a COLREGs compliant path planner for autonomous vessel guidance and control.
Maritime laws and conventions were not drafted with crewless ships in mind and changes will be necessary. The MSC has stated that “proper consideration should be given to legal aspects, including where the responsibility would lie in case of an accident involving a MASS, its consequences to the cargo, and also the implications to the shoreside”. This reference by the MSC to an ‘accident’ raises the obvious question of whether or not changes to the Collision Regulations will be required and how a salvage will be conducted, among others. Further, as the technology is adopted, there will inevitably be a period during which ships with differing levels of autonomy operate, so creating the potential for further complication as those ships interact. There may also be other incidental legal changes, such as establishing standards of data management and the need to store data on the ship in a blackbox, so that data is preserved in the event of a casualty, to name but two examples.
Class is another link in the regulatory chain. As we have mentioned, for a ship to function autonomously, it must be able to perceive its environment and the other users of that environment so that it may navigate safely and avoid collisions. “The decision algorithms behind this need perfecting, as it requires an interpretation of maritime rules and regulations. This leads to interpretation challenges for the programmer. The development of control algorithms for autonomous vessels will be a gradual and iterative process and subject to extensive testing and simulation”3.
In a subsequent paper, we will explore the MSC’s work, the regulations which require amendment or adjustment, and the work of Class and Flag States in this field. For now, we leave you with this conundrum – should an autonomous system be infallible (an unlikely scenario when it has been developed by humans) or should it be permitted to make mistakes and if the latter, who bears the risk and liability for those mistakes – as the technology develops and its adoption increases, will we see a liability regime with product liability for the operating system at its core?
The widespread adoption of autonomous technology is also dependent upon the development of insurance to provide cover for the potential risks and liabilities that will evolve over time. “The insurance industry’s expertise in risk management will be a factor in the adoption of autonomous and unmanned technology. In an area where regulation and safety standards are yet to be developed, insurers can encourage prudent progress by making their own risk assessments and providing policies for responsible operators”4.
Liability will be a key consideration in the development of insurance products, with liability having significant bearing on the pricing and structure of risk transfer. “What needs to be explored, is to what extent other liability rules, such as product liability, would affect traditional rules of maritime liability and insurance in the field of autonomous shipping”5.
Any insurance cover will also need to address the cyber risk. From the system manufacturer’s perspective, high standards of system resilience, encryption and redundancy will need to be engineered from the start to counter that risk and, as autonomous vessels become more prevalent, changes to product liability cover may also be required.
In another article in this series, we will consider the risks arising, and how the insurance market is adapting and responding to those risks.
In conclusion, the disruptive revolution of vessel autonomy is well resourced, well supported and well underway. It’s a brave new world.
If you have any questions about this article, please contact Ceri Done, whose details appear below.
1 Autonomous Vehicles, Handing Over Control: Opportunities and Risks for Insurance. Lloyd’s.
3 Autonomous Ships: The Next Steps. Rolls Royce.
4 Autonomous Vehicles, Handing Over Control: Opportunities and Risks for Insurance. Lloyd’s.
5 Autonomous Ships: The Next Steps. Rolls Royce.