A reality check for autonomous shipping in 2018
To the observer it is striking that much of the technology underpinning autonomous vessels has been in use for many years with several vessel designs already developed. Indeed, the “Yara Birkeland”, the world’s first commercial autonomous vessel, shows us that the autonomous vessel is in fact already here.
The economic benefits of autonomous vessels are centered around lower operational costs due to reduced fuel consumption and crew costs. Most of the reduced fuel consumption is due to the removal of the accommodation structure, which according to studies can result in a 6% reduction of fuel due to a reduction in weight and air resistance. At the same time removal of the accommodation structure gives a 5% reduction in construction cost and opens up for more cargo space and thus higher freight income.
From this it follows that the full benefit of the autonomous technology is only likely to be obtained if a vessel is completely unmanned. If an autonomous vessel is manned, but with a reduced crew, much of the benefits are lost as the vessel will still incur certain crew costs and, more importantly, needs a costly, heavy and bulky accommodation structure.
The cost reductions must in turn be balanced against higher capital expenditures when constructing autonomous ships. With no crew onboard there is much greater need for technical redundancy systems, such as twin screw rather than single screw propulsion, which drives up costs. According to one study a conservative estimate is that redundancy systems on board an autonomous bulker can be expected to increase construction cost by 10%.
Moreover, less fuel consumption may not necessarily translate into lower costs for an autonomous vessel. The reason is that the Heavy Fuel Oil (HFO) widely used today is cheap, but also dirty, greasy and maintenance intensive. Without humans to ensure the smooth operation of engines operating on such fuel, machinery breakdown is a likely outcome. Autonomous vessels must instead run on higher-grade fuel such as Marine Diesel Oil (MDO) or Marine Gas Oil (MGO). Such fuel is considerably more costly than HFO and estimations have shown that MDO/MGO would have to decrease about 12% in price to justify an investment in an autonomous bulker in the current market.
Interestingly, this may very well change with the new IMO emission rules scheduled to take effect from 2020. Fuel such as MDO or MGO emitting less Sulphur will then become a much more viable alternative to HFO, which can only be used if expensive and space constraining scrubbers are installed instead.
Adding to an already complicated equation is the considerable amount of uncertainty surrounding investments into new technology. An overriding consideration is that return on an investment into new vessels must be calculated for 25 years, the generally accepted life-span of a vessel. This is a powerful disincentive for investments into a new and untried technology such as autonomous vessels.
One important variable is the cost of insurance. As around 80% of marine casualties are caused by human error, there is good reason to think that marine accidents are likely to be reduced due to autonomous vessels, with lower insurance premiums to follow. However, at the same time mitigating the effect of incidents will be challenging without the presence of humans. This casts some doubt on the insurance consequences of unmanned vessels. Insurers cannot easily model a new concept as there is no available data and are therefore likely to err on the side of caution when underwriting the risk.
Autonomous shipping will not be stopped by the various obstacles, but it is a good idea to take your eyes off the headline grabbing deep-sea vessels and instead focus on the smaller coastal and inland waterway vessels. This is where disruption is likely to hit first. One reason for the feasibility of the “Yara Birkeland”, is that she is small in size, only 100 TEUs. Smaller vessels mean smaller investments and thus less risk. The local trade also means she is able to use an electric propulsion system not available to deep-sea vessels. For similar reasons we are likely to see further smaller vessels being ordered and deployed before larger scale investments are undertaken. Such a gradual approach will provide proof of concept that provides reassurance to various stakeholders before committing capital to other larger projects.
In this regard, the 25-year lifespan of existing vessels will act as a highly efficient impediment to the deployment of autonomous vessels. It is unlikely that new technology will become so advantageous that it justifies scrapping last generation vessels before they approach the end of their lifecycle.
In addition to the commercial obstacles come various legal challenges. Shipping is an extremely international business with vessels trading to various countries and continents. Uniformity of regulatory approach is typically achieved through the International Maritime Organization (IMO) where some of the main conventions practically cover the world’s combined merchant fleet. The problem with such a comprehensive system is that it moves very slowly. With the process of regulating autonomous vessels just barely having started within the IMO, the outcome is years away.
One of the legal issues facing autonomous vessels is the safe manning requirements applicable to merchant vessels. Several conventions (and charterparties) require that vessels shall be properly manned, something which is a challenge for an autonomous vessel designed to operate safely without any crew onboard.
The parties are free to amend any contractual wording according to the technical reality, but the convention rules on manning are implemented and mandatorily imposed by the flag states. It could be argued that a vessel from a safety point of view is sufficiently manned with zero crew when designed for this purpose. However, all maritime conventions have been drafted on the basic assumption that vessels have crew onboard and it seems unlikely that flag states will push the boundaries here.
Seaworthiness remains a crucial requirement in shipping found in a great number of contracts and legislation. In English case law it has been defined as “that degree of fitness which an ordinary careful and prudent owner would require his vessel to have”. It seems likely that technical developments will be factored into such a flexible standard, meaning that an autonomous vessel that complies with applicable rules would be considered seaworthy.
Rules on navigation are on the other hand closely linked to the crew, for instance by imposing certain duties on the Master of a vessel. If there is no Master it is unclear how such duties can be discharged. It is unlikely that a duty simply will fall away under such circumstances and a more likely approach by flag states is that amendments will be required to place the duty on a different entity such as the owner, a designated person ashore or the vessel itself.
We must assume that for many years to come unmanned autonomous vessels will be navigating alongside manned vessels. Collision avoidance regulations (COLREGS) must therefore apply equally to both kinds of vessels as the crew onboard a manned vessel must be able to predict what other vessels will do regardless of manning.
From a technical point of view it is probably feasible to create algorithms that comply with COLREGS. A challenge, however, is that the COLREGS often also contain a human element, such as Rule 2(a) which hasi an overriding obligation to act as is required by “the ordinary practice of seamen.” This typically refers to what a reasonably competent mariner should do in a given situation
It is in this regard noteworthy that many of the COLREGS rules refer to what the vessel, not the seafarer, should do in a given situation. Indeed, the 1910 Collision Convention Article 3 refers to the errors of the ship, not the crew. Although for instance good seamanship is a concept based on human decision making and actions, there is conceptually nothing preventing an unmanned vessel from complying with such requirements through its manoeuvring. Such an objective approach is in line with the spirit of COLREGS and may be the way forward for any necessary amendments.
Another challenge is the great variety of maritime liability regimes. Some impose liability on the owner (or bareboat charterer) through actions of the crew (e.g. COLREGS), some impose strict liability on the registered owner (e.g. various pollution conventions), whereas others hold the ship liable in its own right (e.g. maritime liens).
Liability regimes directed at either the vessel or the registered owner are likely to apply equally to manned and unmanned vessels. However, an autonomous vessel will create complications to any liability regime based on crew negligence. Can for instance a bareboat charterer be held liable for a collision of an autonomous vessel? Moreover, the introduction of new software and hardware enabling autonomous navigation of vessels is bound to cause new legal issues. Will for instance software suppliers risk becoming liable for navigation under COLREGS based on algorithms developed by them? The answers to these questions are far from clear, and much work remains before international solutions are in place given the slow pace of the IMO.
By contrast, national legislators can move much faster as international conventions allow for national exemptions. In the «Yara Birkeland» project the Norwegian Maritime Authorities have publicly stated that they want Norway to be the first maritime nation deploying an autonomous vessel. Other nations are likely to follow suit, paving the way for local deployment of autonomous vessels before they make their way into international waters. As the commercial case for autonomous vessels is also strongest for coastal and inland waters vessels, disruption is likely to start locally and expand from there.
It is here useful to bear in mind the transition from sailing ships to steam ships. Despite the superiority of steam ships, the transition took more than 50 years, counting from the point at which there was a true pick-up in registration of new steam ships. This shows that a significant lag can be expected also for autonomous vessels. In other words, autonomous shipping is coming but we are not there yet.