Into the future: the technology of tomorrow
David Patraiko, Director of Projects for The Nautical Institute, charts the future of maritime technology – and points out some areas of likely risk
There are a number of new technologies which are beginning to come into use on board ship – and while you might not encounter them for a few years, they will become increasingly common as the fleet is updated and renewed. Some of these technologies provide more data to help make decisions – for example, LIDAR (like radar but uses laser light) improves target detection and identification. Likewise, optical technologies are being evolved to augment human sight.
Other new developments aim to make the most of this improved input. Decision support systems, for example, help amalgamate and analyse data from various sources for activities such as collision avoidance. They might serve as a useful check for human decisions – or may in the future even replace them, although this day is likely a long way off (see page nine for more information). While this research might one day lead to a role for navigators ashore, for the moment it is important to learn to make the most of them while at sea. There are many more examples of technologies that are likely to affect the way navigators work in future:
Decision Support Systems (DSS)
DSS utilise input from a number of sources, including computer systems, sensors and human operators, combining and processing this information to give an overview that helps humans to make decisions. They can help manage situations where there is so much data that there is a risk of the operator being overwhelmed – or not being able to process the raw data in the first place.
Common examples in shipping are weather routing, where a shore-based team with extensive meteorological data and knowledge might be able to offer specific and bespoke advice to mariners on what optimum routes will result in better fuel consumption, reduced weather damage and optimal arrival times. Another example might be a collision avoidance system that uses computer algorithms to identify radar, AIS and possibly optical inputs and compares them with the Colregs to ‘suggest’ possible decision to be made to avoid a collision.
While DSS can be very useful, mariners must understand the basis on which decisions are made. This includes the type of inputs being used, and the strengths and weaknesses of each of those inputs. In terms of weather routing, for example, does the advice recognise all the parameters including traffic and the best practice of seamanship? In terms of collision avoidance, does the system have a full view of the situation, including small craft that may not have AIS? Does it recognise underkeel clearance (UKC)? As we progress into the future all these systems will get smarter – but mariners will need to be aware of the background against which the advice is given, and the impact of either following or overriding it.
Artificial Intelligence (AI)
Artificial Intelligence, often referred to as machine learning, is where computers learn for themselves, initially with significant help by humans. Popular examples involve audio and visual recognition. Many of you will have ‘smart speakers’ where you can ask a computer to answer a question or carry out a simple command. You may have noticed that these systems become better and more accurate the more you use them. That’s because they are designed to learn how to interpret your voice and your preferences. Any improvements are not made by humans; this is the software itself ‘learning’ and improving from previous input.
In the context of shipping, a common AI application is image recognition. There are a number of watercraft that use optical tools to ‘view’ their surroundings, but how do they know what they are looking at? In the first instance they view a buoy, boat, ship etc… and have no idea what it is or what reactions are required. However, human operators spend thousands of hours ‘teaching’ the system to understand what it is looking at. As a result of all this input, eventually the system ‘learns’ what is what and how to identify objects. This type of technology has been very successfully proven with autonomous cars and even with medical diagnosis. In shipping terms, such ‘intelligent systems’ will particularly support mariners in poor visibility – though they will also be useful in good visibility to alert seafarers to things they should have noticed and perhaps didn’t.
As AI systems develop, mariners will need to understand their benefits and limitations. There will be opportunities for skilled mariners to be involved in their development and regulation. The Nautical Institute has, however, been clear that teaching an AI system should never be allowed to distract officers on watch!
Maritime Autonomous Surface Ships (MASS) is the IMO term for ships that operate with reduced or no human involvement on board. The way in which MASS should be introduced, regulated and controlled is currently being debated at the IMO (see page nine for more information). While this may take years to conclude, the reality is that there are already thousands of autonomous craft navigating every day. These are mostly small vessels working in controlled ‘trial’ environments and in niche operations, such as military or research. This is an evolving field, though, and one which will continue to grow, offering both disruption and opportunities.
There are various levels of autonomy, from direct remote control of individual vessels to fleet swarm operations. All of these present their own opportunities and risks. Given the current investment in ships designed to be operated by humans, it is unlikely that the world fleet will be replaced or converted to MASS in the short or even medium term. However, the level of automation on manned vessels will increase as more and more tasks have computer input.
As increased levels of automation are adopted it will be important to clarify what role the associated technology plays in the process. Will it replace human activity – for example, automatic plotting on ECDIS? Will it augment human abilities, as advanced target detection technology does now, or will it work in collaboration with humans, like decision support systems?
Needs of the user
Wherever technology takes us, it is important that it is developed with the needs of the user in mind. To enable this, mariners will need to continuously assess their relationship with automated systems – and give feedback to their companies and The Nautical Institute (you can do this by emailing us at email@example.com) about what works and what doesn’t.