Automated pilots of the Baltic – enhancing or weakening navigational reliability?

In the WISE project, part of the work has consisted in identifying resilience gaps related to the Baltic Sea. In the environmental policy WISE sub-group (EPRG), we have focused particularly on maritime safety. In this blog, I reflect upon the significance of recent trends toward automated pilotage from the point of view of navigational reliability.

In a December 2017 piece of news, the CEO of Finnpilot Pilotage, Kari Kosonen, was asked whether he thinks remote pilotage is a threat or an opportunity. “- Definitely an opportunity”, he answered. He is not alone in thinking this. The digitalization and automatization of marine shipping are experiencing something of a boom nationally and internationally. Hopes are high for “autonomous shipping”. “In addition to innovations, this will bring international buzz and investments to Finland, and with them growth and jobs”, according to Harri Kulmala, CEO of Dimecc Oy. Initially, the idea has been to pilot ships remotely or from the shore, but also unmanned pilot vessels managing feeder traffic to island bases such as those of Örregrund, Harmaja, Utö and Isokari are said to be in sight.

That’s what we know. What we don’t know are the details of this sought-after automation, and like so much in policy and management, the details are everything.

Automated pilotage for what class of vessels operating under what types of conditions? Shipping operations even under “normal” sea conditions are rarely invariant, with all manner of surprises possible and contingencies needed to be prepared for. This says nothing about what happens to automatic pilotage when the vessel’s real-time operations are temporarily disrupted or have failed indefinitely, requiring manual control and direct operator management. The real-time risks to be managed for vary not only across normal operations, but also for disrupted, failed and recovered operations, whether the pilotage is automated, manual or the degree to which it is one or the other. Equally important is the question of what level of autonomy is being sought for the vessels and conditions of operations (for example, a dynamically positioned ship relies on a greater level of autonomy than the autonomously remote operating vehicles used by the vessel for its below-sea operations)? Specific details and scenarios are called for and the stakes are high in getting the answers right before proceeding further with the promises of automatic pilotage.

Here’s why. On the one hand, as our EPRG-WISE research on navigational resilience gaps has shown, one of the main unresolved challenges in navigational reliability is that of the human herself. Most of the time, things go well. Sometimes, however, human helmsmen tire and fall asleep behind the rudder. Some helmsmen don’t know how to navigate icy waters. Not all seafarers are fluent in English, which is the official language in international navigation. Humans seek profit and rationalize away crew in order to cut back on costs with implications for human and environmental safety. On the other hand, recent research has also identified many possible new risks with greater levels of pilot automation. For example, the lack of crew onboard might lead to no bodily feeling of how the ship is rocking, that is to no “ship sense” among the on-shore ship remote operators. The predominance of on-shore remote control could also lead to skill shortage and skill degradation. Also, as has been pointed out, babysitting automated systems at a distance might lead to overwhelming boredom among personnel, leading to loss of vigilance and lowered levels of preparedness for surprises.

From the point of view of navigational reliability, then, we seem to face a well-known paradox between demands of autonomy and demands of skilled expert intervention: 1. The higher the autonomy level, the less the demand for a human operator (with skills shortage and degradation as a possible result); but 2. the demand for an experienced operation is highest during unforeseen contingencies, emergencies or crises (when little time remains to take effective action in the face of a now increased degree of system complexity). How to resolve this dilemma?

The standard answer is and must be: more simulations and operator trainings to ensure that human operators remain on top of things. In simulating critical navigational situations of, say, failure of automated navigation, it is possible to demonstrate how one can learn that what one thought one knew one really didn’t know; or that one knew more than one actually thought; or both. Given the changed task description of maritime crew, however, enthusiasm for such simulation exercises might be hard to muster. Indeed, there are already indications that increased automation in shipping is reducing the motivation among students to engage in maritime education, as some now tell us. One way out of the dilemma could be to increase the level of complexity of the simulations themselves. It might, for instance, be simulated that some sort of societal crisis leads to the current navigational systems being compromised, this in turn giving rise to new questions of the locus of control over vessels. Or the simulation could contain elements of no longer being able to manage the navigational system for keeping surprises at bay, such as a situation in which, say, an autonomous cargo ship collides with a passenger ferry containing migrants of unknown origin and health status. These would be highly challenging situations already under current manned conditions, and it might be worth the effort to consider—well in advance—what kind of additional challenges and possible new risks the called for higher levels of automation might bring with them. And far from overwhelmingly boring.