Future Ships – Smart For Sure, Unmanned Maybe?: Volker Bertram (RINA)

June 6th, 2017 by JCB
Posted in Articles, Technical and Training | No Comments »

“The idea (of an unmanned ship)is not new, it has been around for decades but the difference is the technology now exists,” said Oskar Levander (Rolls-Royce) in 2013. Indeed, at least for 30 years now there have been recurrent proposals of unmanned shipping. Several technological trends,and the advance of autonomous technologies for cars and trucks, explain why the discussion of smart ships has gained so much momentum recently.

Most visions of unmanned shipping involve a mix of autonomous systems and remote control by humans. An autonomous ship has software that may advise human operators or even replace human decision making and action. Automatic braking systems in cars are an example of such high-level autonomy. As with cars,
we will see increasingly autonomous systems for ships. These will make ships safer and easier to operate. Autonomous ships may still have small crews onboard, e.g. for service tasks or inspections or for occasional manual control. In all likelihood, future smart ships will employ a mix of local and remote control with redundancy for vital systems in case the communication link breaks down or local systems fail.

Technology changes support higher autonomy. For unmanned (or even autonomous) shipping, there are technical, legal, economic and emotional barriers to overcome. Let us look first at
the technical issues.

Collision (and grounding) avoidance is a key nautical task. Autonomous collision avoidance may use the same information sources as used by crews today (radar, transponders, electronic charts), but would need to compensate for the human vision and hearing of crews onboard manned ships. One option may be using remote sight and hearing, i.e. using cameras and microphones and remote operators to supply required information for sufficient situation awareness. But alternative technologies are evolving rapidly, such as machine vision (automatic ship identification by cameras), laser-based systems (Lidar), and enhanced ECDIS (where e.g. drifting icebergs are integrated into satellite tracking services).

Automatic avoidance path planning is mature already. At present, execution lies in the responsibility of the helmsman, even if he does nothing more than confirm a suggested manoeuvre. Autonomous collision avoidance systems are expected to relieve ship crews of dull watch tasks, but require overcoming several barriers, most notably regulatory permission, to use higher levels of autonomy (execution rather than advice).

While collision and grounding avoidance receives the most attention in the discussion about autonomous ship operation, there are other navigational tasks that need to be considered. Berthing is a difficult and complicated operation, even for experienced ship masters. Berthing of large unmanned cargo ships with tug assistance has been tested successfully in simulations. For

smaller ships with special manoeuvring equipment, Japanese and American researchers have demonstrated in field tests that automatic berthing is feasible.

Differential GPS, suction or electromagnet systems are available for final berthing and quay-side attachment. Tug assistance in port and in emergency situations (e.g. engine failure) requires establishing communication and physical links. All the required communication infrastructure is already in place.

Tug operators should obtain log-in permissions to the towed ship’s commands if the ship is unmanned. Physical links may be established using cooperative robotics as demonstrated in field tests. Emergency response requires quick decisions. Autonomous damage control systems are likely to be accepted readily as there are no emotional barriers to rapid emergency response – smarter sprinkler systems are more readily accepted than smarter navigational systems. Communication for ships will increase in any case beyond the now usual ship-to-ship/ship-to-shore communication. Satellite bandwidth and coverage, while still a limiting factor for some applications and in some parts of the world, shows similar growth rates as general computer technology. The required communication infrastructure, including cyber-security technology, evolves rapidly and will most likely not be a critically limiting issue for smart shipping.

Tasks related to cargo are more problematic, although less frequently discussed. Cargo needs onboard supervision for reasons of safety (cargo shifting, fires), security (theft, tampering, stowaways) and preserving cargo quality (livestock, refrigerated goods). In many cases, the diligence of automated systems may outweigh their disadvantages for cargo related tasks.

Depending on ship and cargo type,requirements and ease of automation vary. For bulk carriers and tankers, no special considerations appear to be necessary. For container vessels, cargo supervision may be required for certain container cargo types, e.g. refrigerated containers under deck. For multi-purpose vessels, transport of refrigerated cargo below deck is not usual, but the trade of these ships makes them unlikely candidates for low-crew or no-crew shipping anyhow. For LNG carriers, Japanese researchers have developed and tested systems for automatic fault detection, diagnosis and corrective actions since the 1990s. Such systems have already been installed on a number of LNG carriers. All cargo documents could be digital with (automated) electronic transmission to ports and other stakeholders. The “Internet of Things” foresees that containers and other unit cargo will have their own chips and near-field communication.

A mid-1990s report of the University of Hamburg identified frequent maintenance of diesel engines (e.g. changing of filters) as a major show-stopper for unmanned shipping. The advent of LNG as a fuel is a major game changer in this context.

Machinery maintenance is likely to reduce drastically with the adoption of LNG as a fuel. Maintenance requirements will further decrease when switching from diesel engines to fuel cells. Fuel cells have no moving parts and can run on LNG. They may become commercially competitive with generator sets by 2020. At least another decade may pass before they are used for main propulsion in larger ships, but eventually machinery spaces can be envisioned that require no maintenance for months to years.

Unmanned shipping is cheaper – really? Little attention has been paid so far to the economy of unmanned ships. The enthusiasts of unmanned shipping frequently overestimate savings and underestimate costs. “A ship without a crew is certainly technically feasible, but is it also economically profitable? (…) A reduction in the crew cost by investment in sophisticated (technology) both onboard the ship and ashore might add more costs than what is saved (by) having no crew onboard. It is today more realistic to concentrate on a minimum manning of four-six crew onboard and shore based maintenance and support. With advanced design, this is a realistic goal (…)”, summarised Kai Levander in 1994. Around the same time, a study at University of Hamburg similarly found that an unmanned Panamax containership would be economically unattractive.

While decreasing costs of automation and the prospect of low-maintenance ships have shifted the balance in favour of more automated shipping, several economic aspects remain  in the discussion about unmanned shipping.Suppliers of autonomous technology for commercial shipping declare their systems generally as “advisory” to avoid liability issues. With time and experience, more confidence in autonomous systems may overcome liability concerns, but residual risks will remain and suppliers will require additional financial incentives for taking such risks.

Although autonomous systems are expected to increase safety, there may be an initial phase with higher insurance fees than for conventional ships. With proven performance, eventually the insurance fees should be lower, reflecting the  higher level of safety.

Currently, pirates take ship, cargo and crew, demanding and receiving ransom for all. Unmanned ships would be safer than manned ships in this respect. Unmanned ships may have shut-down mechanisms for main propulsion and manoeuvring systems. In this case, pirates would require ocean-going tugs to get
ship and cargo into friendly ports.

A new concern is cyber-jacking where criminals obtain electronic control of unmanned ships. This is also a serious concern for sea  traffic management and largely autonomous,but manned shipping. Awareness of the cyber-security issue starts to spread in the maritime industry and at least partial solutions are on the horizon.

Several factors increase the initial costs of autonomous ships, including additional equipment for automation (sensors, communication, computers,software), high-quality equipment for reduced maintenance work or higher reliability, additional equipment to permit redundancy in key systems, and additional security systems (against terrorism, cyber-jacking or piracy). On the other hand, there will be lower resale value (initially) for such ships as the first autonomous ships are likely to operate in territorial waters only. In this case, a national authority may grant permission to operate even though regulations for international ship operation are not fulfilled. With flag state and port state approval limited to one country only, the resale value of such a ship  would be low. With wider adoption  of autonomous ship technology worldwide, this barrier will vanish
in time.

Disputably higher costs per crew member may have to be taken  into account – some argue that automation will require more highly qualified crews; some that automated systems will be easier to operate and require less training. The actual costs for training and the availability of trained crews will depend on the degree of automation and the degree of user-friendliness of future designs for autonomous systems. The additional costs make net cost savings for autonomous shipping debatable, especially since future requirements for systems (“you would have these anyhow onboard a ship in the future”) and future costs of automation equipment (“prices of sensors, computers and hardware are decreasing rapidly”) are by nature speculative. But they are not allowed. The lack of a legal framework for unmanned shipping is a fundamental hurdle. Several issues need to be addressed within the maritime legal framework, foremost IMO regulations, before we may see commercial unmanned shipping. For example, many international maritime conventions are stated to apply to ‘ships’. Manned ships with some autonomous systems fall clearly under this definition, while small unmanned surface vessels do not. Whether unmanned cargo vessels are treated as ‘ships’, and have to follow all international requirements for manned ships is  yet to be formalised. Similarly, it is
not clear whether a remote operator in a shore centre is the ‘captain’ addressed in current maritime regulations. Flag states and port states will need to agree on the interpretation of the applicability
of existing regulations.

Of the many issues and details to  be addressed, let us look at three frequently raised concerns:

Minimum manning regulations – Every Flag Administration must ensure its ships are manned to the level and competence required by international rules and standards.  For unmanned ships, future regulations would be guided most likely by an “equivalent safety” approach adopted in many recent additions to maritime regulations, following a goal-based rather than prescriptive legal philosophy. One possibility may be a separate
registry for unmanned ships.

Seafarers in distress – Various international frameworks address the legal and moral obligation to help persons in distress at sea. In maritime Search & Rescue efforts, unmanned ships could deviate from planned voyages to assist search efforts with their (superior) sensors. Rescue operations are more problematic. We can envision robotic devices to assist seafarers in distress, as proposed by the University of Rostock. However, retrieval of seafarers and subsequent (medical and other) care are unsolved problems. One possibility would be to exempt unmanned ships (the same as unmanned oceanographic platforms) from the obligation to assist seafarers in distress.

Evaluation of the safety of autonomous systems – Autonomous systems will reduce the number of accidents due to human error. However, decision-making of ‘intelligent’ (machine learning)systems may introduce different risks. New generations of increasingly autonomous systems will require a new approach to safety evaluation. Processes for evaluating autonomous systems may resemble the processes for certification of competence applied to human crews: a combination of loosely monitored regular performance combined with simulator testing of dangerous or critical situations.

Homo Maritimus versus Homo Informaticus: The marine industry views the issue of unmanned ships predominantly with much scepticism. Michael Barnett of the Southampton Institute  summarised the findings of a formal study on alternative manning structures on ships: “One major  conclusion of the participants was that, although technically feasible, unmanned ships were unlikely to appear in the foreseeable future for commercial and political reasons.” Considerable consensus was found that unmanned ships might be technically feasible by 2030, but “the opinion of the respondents was split on both the desirability and the likelihood of adoption.”

Shipping is a conservative industry. “There are many reasons for resistance to IT systems in the shipping industry; some not very noble, some due to ignorance, some to arrogance, some to laziness, some to personal convenience,” says Giampiero Soncini, CEO of SpecTec. There are also good reasons for the conservatism in the industry, notably that ships are high-value, long-life assets subject to rather rigid international regulations.

Then there is society’s intolerance of system or machine error. Being human, we can forgive human error much more easily than technical failure. Tolerance of accidents due to human error is much larger than tolerance of accidents due to failures of autonomous systems. This illogical imbalance impedes adoption of new technology.

We also frequently encounter emotional barriers. These are not to be dismissed lightly. Seafarers and trade unions look with concern at ideas which (at least in their perception) implicitly devalue their profession, and may threaten employment and degrade working conditions. IMO, several national authorities and non-governmental organisations (NGOs) have voiced concern regarding low-crew ships with increased work load for the remaining crew. Whether this is actually an argument for or against autonomous systems depends on the implementation of these systems. Poorly designed (partial) automation may result in boredom,  fatigue and stress, ultimately making the seafarer profession unattractive and shipping less safe. This is  detrimental to our goals and values in the maritime industry. Human factors and ergonomics should play a larger role in the design of autonomous systems to accelerate the process of acceptance in the marine community.

The issue for now is the Smart Ship, not the unmanned ship Technical developments and general acceptance (and subsequently legal frameworks) will evolve gradually. At each stage we will review and learn before moving on to the next level of autonomy. For now, the issue is the arrival of the ‘smart’, user-friendly and much safer ship, which will familiarise a wider maritime community with possibilities of autonomous technology. The unmanned ship may come at some point in time, but it will not come overnight. Instead, it will come as part of an evolutionary process that will take time – and give us the time to find wide acceptance of this type of future shipping.

This article was reproduced with the kind permission of the Chief Executive of the Royal Institution
of Naval Architects.