Auto(m)ation

Automated Ship
Artists rendition of a Rolls Royce powered unmanned ship

By Captain George Livingstone

I remember an American cartoon from childhood, “The Jetson’s”, I still remember the intro song (“Meet George Jetson, his boy Elroy, Daughter Judy, Jane his wife”).  The Jetson’s lived in “Orbit City’”, a push button, automated, space age world. It never occurred to me that before my days were over we would see some of that world. Well its 2016 and there are (nearly) fully automated passenger jets and we are fast approaching automated cars (thanks to the likes of Google and Tesla). Given the rapid advances in automation for jets and cars it was only a matter of time before companies like Rolls Royce considered automated ships. And it isn’t just a private enterprise, several Baltic nations have been heavily involved in projects looking at the possibility of fully autonomous vessels linked to shore side control. The stated goal is to have autonomous ships (unmanned) sailing the world’s oceans within the next ten years. The theory is the ships will be linked to shore side operation centers controlling the vessel as it goes about its business on the ocean trades.  It is incredible to contemplate but there is no doubt that the determination and technology exist to make it a possibility.  

There’s the rub though, the technology exists but should that be the primary consideration?  Are there practical and other (i.e. ethical) considerations prior to implementing automation? Especially in the dynamic and dangerous world of marine transportation? Have similar questions been considered or debated in aviation or auto? Of course they have and we in marine transportation should take very seriously any outstanding issues automation has caused in either.

As an example, a colleague recently sent me an article, “Why experts worry about a Telsa crash” by Mathew Dolan of the Detroit Free Press newspaper (July 2, 2016).  This article focused on ‘human factors’ associated with automated vehicles in particular and there are Red flags.  “When drivers no longer actively drive the car but still have to monitor road environment, their performance in the monitoring task rapidly deteriorates.”  The article goes on to state that work still to be needs to be done as the Telsa autopilot is explicitly denoted as a “Beta product” (inferior).  In other words, although the technology is well advanced, experts are sounding cautionary notes and questioning whether automation ready for prime. 

There is a fascinating and detailed analysis of ‘automation’ in the best-selling book “The Glass Cage:  Automation and Us” where author Nicholas Carr sounds multiple alarms.  In regards to Google cars, Mr. Carr states “although Google expects commercial versions of its car to be on sale by the end of the decade, that’s probably wishful thinking.”  The vehicle’s sensor and computer systems remain prohibitively expensive, “with the roof-mounted laser apparatus alone going for $80,000.  Many technical challenges remain to be met, even the most powerful computers still have a hard time distinguishing a bit of harmless road debris from a dangerous obstacle.”   He goes on to question the daunting legal, cultural and ethical hurdles that must be addressed with fully automated vehicles.  Where will culpability and liability reside?  With the car’s owner? The manufacturer? The insurance company? With the programmers who wrote the software?  There are a host of issues that must be addressed and agreed to regardless of the fact that the technology exists.

Mr. Carr and Mr. Dolan bring up valid counter points to the application of full automation in aviation and autos.  I would pose questions for automation in marine transportation.  What is the cost to build a fully autonomous ship? (2X, 10X)? Is there enough profit in the bulk cargo, tank ship and container ship world to even consider the idea? If no then is the idea even practical? What about liability (insurance risk?)  International treaties and law? Wouldn’t it all have to be rewritten and negotiated from scratch through the IMO?  

As a working mariner with nearly four decades of sea service I have concerns regarding automation at sea; namely it’s practical application in the real world of ocean transportation.  Just a few examples; how does a ‘shore-based’ operator determine when (or how) to slow a ship when facing heavy weather?  Experienced mariners will tell you one must be on scene ‘feeling’ the ship and how it is reacting to heavy weather.  What about onboard failures at sea?  Fire?!?  Main Engine?  Generators?  Navigation? Container or cargo problems?  How do we integrate autonomous ships into a world dominated by ‘manned’ vessels? Etc.  

Mr. Thor Hukkelas, Principal Engineer Marine Operations for Kongsberg has said “We are surrounded by more and more automatic, apparently intelligent and autonomous systems.  Despite the fact that these systems become increasingly reliable, they will eventually fail.”  For sake of debate, let’s concede an immediate 98% safety record for autonomous ships upon future introduction.  Assuming about 10,000 ships sailing the world’s oceans on any given day, that would still mean about 200 ships everyday would have issues of one kind or another (all over the big wide world). In only one week 1,400 ships would be in need of some kind of repair or servicing (in the middle of the Atlantic?).  In one month nearly 6,000 would theoretically be in need of some kind of repair or attendance.  And on and on it would continue until what? Some kind of disaster(s)? Would autonomous ships ultimately be safer for the Public Trust?

In a feature article on automation and Bridge Team cohesiveness in the July 2016 Seaways, Capt. Nick Nash (V.P NI, Master Princess Cruises) states “Our goal, with proper training and the pilot’s agreement, is to be able to drive a ship into port using our track control system (sophisticated auto pilot) to a position just off the berth – preferably in what our Track Control System calls “Track Mode”.”  He goes on to say “We are now instrument navigators backed up by visual clues” (I respectfully disagree).  Indeed the issue with the likes of “Track Mode” or any other kind of automation is that it must be 100% reliable, 98% is simply not good enough.  In just the last 12 months, a passenger ship entered Sydney Harbor (Australia) with the intent to use “Track Mode” until just off the berth.  Sydney Harbor has a very difficult ocean approach; ships enter between North Head and South Head fully exposed to following swells and must make a very tight 90 degree left turn before proceeding to the inner harbor.  In this particular case something went awry and “Track Mode” could not handle the 90 degree left turn, compelling the pilot to retake direct control of the ship.  

The point here is not to pick on a well-respected, experienced colleague nor is it to take pot shots at automation, there is far too much on the line regarding Public Safety to be combative or negative.  The essential point is that it’s folly to over-rely on automation until such time that automation is 100% full proof and we are not nearly there.  International Marine Transportation should embrace advances in technology and automation but not at the expense of reason, common sense and safety.