MASS — here we come, ready or not?

30 Окт

Статья посвящена вопросам, связанным с эксплуатацией автономных судов — marine autonomous surface ships (MASS). Автор показывает актуальность таких вопросов, в частности и особенности, правил эксплуатации MASS, а также участия Морского института Великобритании в разработке этих правил. Автор делится собственным опытом в сфере эксплуатации автоматизированных судов и соображениями относительно целесообразности эксплуатации полностью автоматизированных судов, автоматических судов — роботов. Особое внимание уделяется вопросам безопасности, в том числе избежанию столкновений судов. Кроме того автор обращает внимание на организационные проблемы, связанные с MASS, в частности, с влиянием внедрения MASS на трудовые отношения моряков. Вообще, в настоящей публикации содержится целый фонтан идей по развитию автоматизированных судов и организации внедрения в жизнь подобных технологий. Особо интересной представляется идея использования в автоматическом режиме подводных судов.

We must move fast if we are to keep ahead of developments in this sector.

Some very interesting presentations and discussions during The Nautical Institute’s Seminar and AGM in Malta and several good publications in Seaways have inspired me to put down some of my own thoughts on the subject, based on my own experience on testing autonomous software and hardware on board.
Of course, there are more questions than answers at the moment, but hopefully it is wortli asking them.
Why now?
The first question is: does this problem really merit our attention right now? It seems as if we have plenty of time before the first totally autonomous ship comes over the horizon. In the meantime there are many more immediate issues to deal with: fatigue, human element, climate change, and so on. Is it absolutely necessary to discuss unmanned ships now?
Experience shows that manufacturers are always ahead of industry, dictating their own rules and procedures (just look at ECDIS, where several of the technologies are anything but not user friendly, but we still have to use them and train our officers in their use, creating a giant opportunity for mistakes). Right now, we have the very rare opportunity to tell manufacturers to create their robots according to certain rules. And we can take part in developing these rules! This is the main reason for the NI to take part in these discussions. For once, we have the opportunity to write the rules, not just to play catch up after the event.
MASS and the charterer
From a sceptical point of view, even if marine autonomous surface ships (MASS, to use the established term) are tried and tried in the near future — will charterers take the risk of using them? Financially, they would certainly find it interesting, given that the minimum average monthly crew cost for an ocean going ship is between $70,000- 100.000. A fully autonomous ship would mean no headaches with crew changes, delayed flights, sickness, or health and safety 011 board. And, potentially, with no humans on board — 110 human errors! It certainly seems likelv that some of the larger companies will take this opportunity. It does present a huge task for P&I clubs, as there are no examples of such insurance in the past.
Ready or not?
If we agree that MASS are worth discussing, we can discuss what exactly we want to see at sea. Here are two questions: Are autonomous ships ready to sail right now? And are we ready for them?
Back in 1991-92, I was involved in one of the first attempts at developing autonomous ships, as an officer 011 the ship where the test system was installed. The task of the system was to keep the ship on route. This had good results: the ship moved from A to B, changing course properly at the waypoints and making course corrections where necessary’ without any interference from us. While correcting the course, the programme was continually deciding whether to plot a new course or return to the pre-plotted line, depending on the situation.
The project was closed clue to lack of financing, but it gave me some new experience, and an opportunity to talk to the engineers, who had previous experience in remote and automatic control for aviation and space craft. According to them, a ship is the most difficult object to automate, as there are so many forces affecting the ship’s movement. The system had to be ‘taught’ for quite a long time in order to gather sufficient data to react properly. The same problem arose when it came to tricky navigation — too little data, or too much volume of data, or poor interpretation led several times to the wrong decisions. The only way to avoid them was to ‘train’ the robot.
Both software and data processing capacity have moved on significantly since then, but even the best software can’t predict everything. Technically, autonomous ships probably are ready to sail, but caution is required regarding particular weather and navigational conditions. And of course, any ship needs maintenance from time to time. That means we need to provide some accommodation for temporarv crew (e.g. mooring crew or maintenance crew). And most important — we need to provide life saving equipment. An autonomous ship is not necessarily an unmanned ship.
Collision avoidance
Are we ready to see an oncoming robot at sea? Here, the answer, as far as I can understand is more negative — we don’t know how the robot will react to oncoming traffic. One solution seems to be to limit the use of full autonomous mode to certain times and areas.
The scheme is quite simple: the ship leaves the port w ith almost full crew (mooring team). After the pilot point, the crew disembark (in some areas where navigation is difficult or traffic heavy, navigators may stay longer). The ship is under control of a Shore Control Centre (SCC) operator. Once on a defined route reserved only for autonomous ships, fully autonomous mode could be turned on.
Human operated ships are to avoid these routes.
If a robot ship and a manned ship do meet, who should give way?
It seems reasonable to create a special signal to identify MASS, but this signal should not give them priority! This is a task for software manufacturers — MASS should at all times manoeuvre as prescribed by Colregs [Editor’s note: this is the position that The Nautical Institute has taken at the IMO].
Moreover, all MASS ships in full autonomous mode should follow similar protocols. That is, given the same navigational situation, similar MASS vessels should behave in the same way. For example, in a give way situation, where course, speed, angle, and CPA are the same, two similar MASS vessels will alter course on the same angle.
What if it all goes wrong?
Based on reports of accidents with autonomous cars, we can expect to see problem situations developing with MASS at an early stage of introduction — like mixing up landmarks or unpredictable behaviour in areas of intense traffic. An out of control ship approaching a major port is a real disaster. What can we do? Send a team of paratroopers on board in mid ocean? What it it is just sensor malfunction or wrong interpretation of data? Any ship’s equipment on MASS should be doubly or even triply redundant.
To prevent accidents — and react to them it necessary — the following scheme might be used, particularly for shore controlled vessels.
Operators could be located at certain points on ship’s route, rather like dispatchers in aviation. Emergency teams who will board the ship in case of necessity will be located at the same points. Needless to say, all shore control centre (SCC operators and emergency teams must be professional seafarers with proper qualifications and specific training in autonomous ships. Even ship must come with a password protected red button’ allowing the ship to be switched to manual mode. Better yet, have dual factor authorisation, with two emergency team members each having separate passcodes.
This leads inevitably to the next question. What about bad guys — pirates, smugglers? How can we prevent them from gaining access or control of the MASS? The ‘red button’ giving manual control is to be encrypted by all possible means. The construction of the ship should allow maximum protection from unauthorised access, with the situation around the ship continually monitored and recorded.
In some ways MASS offer considerable security advantages. The only value on board is the cargo and the ship itselt — there are no hostages, and no cash. All doors could be closed and locked at all times, making the ship totally hardened.
This does seem to contradict search and rescue requirements. How can MASS help people in distress? First of all, by reporting their presence to the shore operator, and then launching life rafts. Further procedures must be developed.
Man vs machine?
Last but by no means least, how will MASS affect the labour market? Will it be a case of human mariners vs robots? I guess not. We still have many vacancies in the industry, and the profession of the seafarer is unfortunately becoming less and less popular. Robots will just fill the gaps. There will be enough new positions for humans who want to enter the industry as shore control centre operators, emergency team members, and so on.
On the legal side, IMO is already working on a scoping exercise to identify the dozens of international conventions and other documents to be updated.
In summary, and acknowledging that the list of matters to be considered is huge, a few points about the development of MASS that are worth bearing in mind:
• MASS vessels are real. Their appearance at sea is just a matter of time;
• An IMO Code for MASS operations must be developed as soon as possible;
• The Nautical Institute has a good opportunity to take part in creating rules for such ships in advance;
• It is The Nautical Institute’s task to reflect the views of human mariners, not robots. The NTs participation in developing rules and regulations must remain focused on ensuring maximum comfort and safety on vessels operated by humans;
• MASS technology will need further development and correction based on results from the first ships;
• In order to avoid accidents, which will definitely happen in the initial stages, MASS operating in full automatic mode should be limited to areas awav from usual routes and, it possible, awav from heavy weather areas. Manned ships are advised to avoid these waters;
• Establish ‘dispatcher’ points for the SCC operators and emergency teams;
• SCC operators and emergency team members to be fully trained professional mariners;
• MASS vessels to be protected from unauthorised access. At the same time, they must enable easy embarkation for emergency teams;
• Search and rescue is a weak point at the moment. Special procedures must be developed;
• MASS vessels should include accommodation, food and water stores and life saving equipment for limited crew (eg maintenance, emergency team, or rescuees);
• Special signals to be created to identify’ MASS operating in shore control and fully autonomous mode. This signal should not give priority;
• MASS to obey all Colreg requirements;
• Navigation protocols to be standardised for similar MASS vessels (at least for standard situations);
• The labour market will not be affected by MASS in the foreseeable future.
Blue sky thinking
The most effective solution for mass ocean transport is not MASS as they are currentlv envisaged, but commercial submarines. After loading, the vessel is brought out of port by a human crew. The crew then disembark, and the vessel goes underwater, and moves through the occan under a totally autonomous regime. No wind, no surface waves, and the environment is much more predictable. As there are no crew, no oxygen is needed, so compartments might be filled with inert gas to prevent fire. Ultra long radio and acoustic waves can be used for obtaining data and monitoring; batteries for moving the ship. Sounds fantastic? Twenty years ago, autonomous cars sounded fantastic. Several companies are already working on a commercial cargo submarine concept.

Автор: Captain Alexander Sagaydak, FNI

Источник: Seaways. — 2018. — September. — Р. 26 — 27.