AI, AI, Captain

AIAI_Captain_BodyThe SmartShip is the autonomous ship by another name, ironically though it won’t mean independence, but convergence, forging the new blue logistics chain as part of the intelligent transport systems of the future. 

When I first started working in offices in the late 1980’s desks were already groaning with chunky word processors, big switchboards, fax machines and huge printers. And they were always going wrong.

At one ad agency in London there was a sign in the photocopying room which read, “Remember—if the photocopier won’t work it’s probably because you upset the fax machine—they’re all in the same union.” I’ve made that joke ever since, but I’m going to have to stop soon. Because shortly it won’t be so funny.

I don’t know how many of you recognise this issue’s cover image. For those that don’t it’s a still from what was voted in 2010 by the Moving Arts Film Journal as the greatest film of all time. 2001: A Space Odyssey, directed by Stanley Kubrick features a couple of astronauts en route to Jupiter on a spacecraft which is controlled by a sentient computer called HAL.

Problem is that HAL goes haywire and kills the crew, aside from Dave, the astronaut whom he traps outside the spacecraft pod-bay and refuses to let back in. Dave eventually does get back in and deactivates HAL which gradually loses its mind as its various circuits are removed.

There was some discussion here about using the image of HAL’s glowing red light on the cover, but I won. My argument being that 2001 is widely regarded as one of the most influential films ever made. So whether you’ve actually seen it or not you will most likely have seen something that was made because of it, and all those thousands of depictions and repetitions have shaped and influenced the public—and therefore probably your—view of artificial intelligence.

That influence has extended far beyond movies. Back in 2011 it was revealed that users who ordered the iPhone 4S’s intelligent assistant Siri to ‘Open the pod-bay doors’ would be met with the chilling response, “I’m afraid I can’t do that, Dave.” Just the same as HAL does. On one hand it’s a great engineer-planted joke, but it’s also a recognition that Siri isn’t just a cool gadget on your phone. It is artificial intelligence. And it demonstrates how pervasive and deeply embedded into society that fear of being outwitted by something that is more intelligent than you, but isn’t human, has become.

It’s activated when we hear the term driverless car, drone or unmanned ship, but is that reasonable? When people talk about the future of shipping and its SmartShips is that what we’re heading for? Is some Master somewhere sometime soon going to get maritime’s first ‘I’m-afraid-I-can’t-do-that-Dave’ moment?

I don’t think so. Not because we won’t have the kind of artificial intelligence which is capable of outwitting us, but because of the reasons why, and the global context in which, SmartShips are developing, and how they’ll operate.

WEF_01

The recent World Economic Forum in Davos took as its theme the ‘Fourth Industrial Revolution’, and you can read a very concise description of what that means from this issue’s Futurenaut, Christopher Rex.

The recent World Economic Forum in Davos took as its theme the ‘Fourth Industrial Revolution’, and you can read a very concise description of what that means from this issue’s Futurenaut, Christopher Rex. If you want to understand what Industry 4.0 means for shipping, well, we covered that back in April 2014, pointing out that as the world digitised, manufacturers—shipping’s customers—were profoundly changing the way they operated, that highly integrated supply chains were going to be pivotal to sustaining competitive advantage and that shipping was the weakest link.

I argued back then that shipping had to recognise and close the digital competence gap between the way it did business and the way its customers and the rest of the logistics channel did. Failing to do so would mean it couldn’t integrate with the cyber-physical systems that were going to be the engine of this new industrial revolution.

The concept of the cyber-physical system was somewhat challenging for shipping when I used to talk about it in 2014. But now the ramifications of the Internet of Everything is beginning to hit home for people, the idea of a component which knows what it is, where it needs to be, what it needs to do and can communicate all of those things to other components around it, isn’t so mind-boggling.

That’s because the roadmap which gets us to the cyber-physical systems of Industry 4.0 can sort of be broken down into about 4 steps. Shipping is currently wrestling with the first two steps: Connectivity, and Sensors and Data.

Unlike our shore-based counterparts, the availability of cheap ubiquitous connectivity hasn’t extended to the deep sea, and it probably won’t ever be as cheap and ubiquitous as on land. But it is getting far, far better. The broadband offerings now are solid and affordable and the advent of the high throughput satellite services over the coming few years will provide previously unthinkable bandwidth at sea. So that’s one part.

The other is the Sensors and Big Data wave that shipping is surfing. Newbuilds are being delivered with virtually all the sensors you could want, and retrofitting them isn’t expensive. As an industry we’re beginning to appreciate how the application of sensors—and in the case of GE’s Direct Write, sensors you can ink onto components—changes the game. Because sensors generate data. Big Data. Class Society DNV GL admits it’s been taken by surprise by the growth of Big Data, but it’s catching up fast, as are other Class societies like ClassNK, featured this issue launching its Ship Data Centre.


AIAI-01

These are the first couple of building blocks along the path to Industry 4.0 and shipping has woken up to them. Data is now being generated in massive quantities and a good proportion of it is being transmitted ashore. But that’s where we’ve got a bit bogged down, to the extent that a lot of the industry is still asking what Big Data is for and querying how much data it really needs.

Not everyone is asking that of course. It’s always dangerous to talk about ‘the industry’ in a homogenous sense, because it contains a very diverse set of organisations, and, particularly on the supply side, many of those operate not just in maritime and shipping but across a variety of other verticals.

Rolls-Royce is one, and it’s transferred its years of data expertise in aerospace to maritime. You can read more about what they’re doing with it in ‘Buying the Cow’ elsewhere in this issue, but the bottom line is that what Big Data needs—the next building block of the cyber-physical system, the next step on the roadmap—is analytics, and the algorithms behind them. In order to generate value data needs to become information and that’s where analytics and algorithms are essential.

Sensors attached to big oily bits of machinery, by and large, shipping gets. Algorithms not so much. Data analytics and algorithms require sophisticated and specialist expertise that shipping and maritime has not routinely employed or encountered, but it does exist, and it is turning data into information.

However, whether it’s via ABB’s newly unveiled Integrated Operations Centre, Maersk’s in-house monitoring centres or Wärtsilä’s condition based maintenance, in order to bring those analytic algorithms to bear on the data, it has to be brought back to a central processing centre ashore.

The data has to be crunched by algos and—perhaps surprisingly—still considered by human engineers before it becomes actionable information. And importantly, once that information is actionable the action is being taken by the human rather than the system.

AIAI-02

Andree Underthun, operations manager (left) and Stian Braastad, global technical support manager at ABB’s Integrated Operations Centre.

It’s up to the guy in the control room at Maersk, Wärtsilä, ABB etc., to pick up the phone or send an email to the Master or the Chief via that comparatively expensive satellite link, and tell him to go and turn something up, down or off. And of course, it also requires someone on the vessel to immediately carry out the instruction.

That might differ where remote support is enabled and the ship operator has either implemented its own, or signed up to a manufacturer’s remote support package. In that case the action will be taken directly from the shore-based control centre. But it still requires a human to intervene.

Despite all the technology, we’re still operating with humans very much in the loop. Which is why even though it might not feel like it right now, the journey from step one to three has been fairly straightforward. The next part is likely to feel like less of a step and more of a leap. Not because of the current limitations of technology, but because many of us haven’t yet understood its power and the changes that intelligent networks are going to drive globally.

The entire Big Data and SmartShip debate has revolved around technical, engineering data delivering optimisation and efficiency, which is very important, but not the whole story. I always talk about Big Data being mistaken as an IT transformation wave when in fact it’s a business transformation wave and that goes to the heart of the SmartShip. For several years now the buzz word has been Ship Efficiency, which is cost control by another name. Leaving aside the fact that operating an expensive asset efficiently should be a basic requirement of any shipping company, it focusses totally on cost. Big Data is about cost containment in real time, but it’s about other things too. It’s about tracking of physical items, real-time forecasting and reinventing business processes.

Step four on the path—the SmartShip as an autonomous, cyber-physical system—is underpinned by distributed networks powered by edge computing where instead of bringing data to the centre to be processed, it is analysed and turned into actionable information by the individual component or system, which then takes action itself without the human intervening.

Finding value in Big Data when you restrict its use to operational efficiency is challenging. As Inmarsat Maritime COO Trond Leira said at the Futurenautics roundtable on the subject, “It is very difficult to get good business models just around engineering data, I know companies have tried for a long time to make some monetary value out of it.”

Despite all the manufacturers and Class societies talking about efficiency, that is not the part of the equation that adds value.  That’s going to come from step four, when the ship becomes a cyber-physical system—unmanned or not—and able to operate as part of the much wider intelligent transport system developing now.

Step four on the path—the SmartShip as an autonomous, cyber-physical system—is underpinned by the falling cost of microprocessors and the development of distributed networks powered by edge or fog computing, where instead of bringing data to the centre to be processed, it is analysed and turned into actionable information by the individual component or system. And instead of waiting for a human to intervene, that component or system will take action itself via actuators, based on the decision its algorithm makes.

That cyber-physical system is the essence of an autonomous ship, one which could sail unmanned. Now shipping has got itself very bogged down about unmanned ships arguing that they wouldn’t offer any benefit over manned ships. But that’s not surprising when you only consider the unmanned ship in the context of cost saving.

Class societies and engineers are myopically focussed on the efficient operation of the asset itself and how much everything will cost, but no one yet has understood the value the autonomous ship brings as part of a much bigger intelligent network.

That value isn’t just monetary, it extends to safety and welfare too. Fears about drone ships floating about the ocean having to take impossible decisions in drastic situations are the same as those being raised about autonomous cars. And the answer to both is that these autonomous vehicles aren’t developing in isolation, and are designed to operate as part of this far wider, intelligent system.

The Internet of Everything will see hundreds of billions of intelligent devices thrown online in the coming years of which autonomous cars, trucks and ships will be just be a part. What is being created is a vast, globally distributed cyber-physical system of intelligent, connected objects that will eventually develop into an external brain, or perhaps more accurately, an external nervous system.

At the moment shipping has a lot of less than satisfied customers, and what an awful lot of them are frustrated with is shipping’s inability to integrate its operations with theirs. A survey of INTTRA members identified digital interaction over cloud platforms as number one on their wish list.

A brief look at the number of start-ups piling into the transport and logistics space trying to get a piece of what they correctly see as gross inefficiency in maritime operations is instructive. It puts the discussions about autonomy, and the opportunity it creates, into perspective, and it clarifies the roadmap that much-pilloried innovators like Oskar Levander at Rolls-Royce have laid out.

The SmartShip reflects shipping’s experience of a wholesale transportation disruption. The way they’ll create value is by allowing us to become part of the Industry 4.0 cyber-physical systems and wider intelligent transport systems, and that will move us up the value chain for our customers.

It may seem like a giant leap to get there, but as I indicated before, technologically, it’s not far at all. The Sea Traffic Management systems developed by the MONA LISA project are pretty much ready to be deployed, and Rolls-Royce is already utilising a degree of edge computing in its condition monitoring solution.

AIAI-03

But perhaps the most overt signal as to where the industry is going came from the signing of a strategic agreement between Inmarsat and Ericsson late last year. The two companies intend to jointly develop services, solutions and applications to drive industry standards for satellite connectivity and applications integration in the maritime industry. Their shared intent is to facilitate the sharing of cargo, logistics and vessel operational data to streamline the entire maritime supply chain.

In practice what that means is integration between the Inmarsat network and Ericsson’s Maritime ICT Cloud—an end-to-end managed cloud solution that connects vessels at sea to shore-based operations including maintenance service providers, customer support centres, fleet and transportation partners, port operations and authorities. “Enabled by Inmarsat, the Maritime ICT Cloud will ensure that trucks will spend less idle time at ports, cargo will spend less time in transit, and producers will be better able to plan their shipments,” say the companies. That, in a nutshell, is how you add value.

Maritime connectivity has acted as a brake on technology adoption in the past, but now it’s the bridge. And according to Mike Mitsock, VP of marketing at KVH, that’s being recognised. “Since we launched KVH mini-VSAT Broadband 2.0 at London International Shipping Week last September, some people’s eyes seem to have been opened,” he told me. “We’re now at the point where speeds are increasing and costs are declining, and the next step—HTS—will be a quantum leap in the potential for both. Bandwidth will always be a precious commodity at sea, but the wider availability and falling cost have opened the door to different conversations, around strategic, value-based adoption—satcoms as a value-enabler, instead of just being considered a cost centre.”

The reality is that the vertical markets we’re used to are beginning to converge, driven by the enabling technologies via which we are integrating our operations and creating new collaborative ways of working. We need to talk not about shipping, but about Blue Logistics—a phrase you’re going to hear from me much more during 2016. I believe that’s important because it focusses us not on the insular shipping market, but on the world beyond, and the customers and suppliers—and the consumers—that we have to get to know and serve better if we’re to survive.

“I think Big Data will be applied to shipping in different ways with information coming from across the supply chain rather than just from the ship. It will come from the end users of shipping, the consumers, who may look very different in years to come,” agreed Warwick Norman, CEO of RightShip at our Big Data roundtable.

We’re creating a smart world, a world where artificial intelligence is everywhere. So why don’t I think we’ll end up with an M/V HAL? Well, because the trend is not towards one, clever, all-powerful device, but the intelligent network

That end-consumer’s influence is only going to grow. They are insisting on transparency, accountability and security from the companies and products they interact with and purchase, and those preferences will have a major impact on the supply chains of which shipping is an integral part.

Blockchain technologies based on the principle of distributed ledgers are likely to form the basis of secure transactions in the Internet of Everything, and already start-ups are building applications which enable a consumer to digitally investigate the entire lifecycle and journey of the product they’re going to purchase.

That kind of transparency has been anathema to shipping thus far, but it’s going to have to become second nature in the future. Because it’s likely we’re going to be rated for everything from emissions to cyber-security to underwater noise. And that will take a SmartShip and a smart network to deliver.

We’re creating a smart world, a world where artificial intelligence is everywhere. So why don’t I think we’ll end up with an M/V HAL? Well, because the trend is not towards one, clever, all-powerful device, but the intelligent network. A massive grid of individual smart devices each with its own objective but able to negotiate in real-time with each other to achieve the safest, most efficient delivery. The likelihood is that a few devices will go tonto now and again, but the distributed system will cope.

Kubrick’s terrifying vision of a psychotic artificial intelligence killing humans has a twist. At the end of the movie Dave discovers that he and his fellow astronauts have not been given the whole truth about the mission on which they’ve been sent, and why. HAL, however, has known all along. It is being forced to lie to the astronauts by mission control which has been the cause of HAL’s breakdown and malfunction.

SmartShips are coming, but they aren’t threatening. Because they will be what we make them. The most dangerous and unpredictable thing we have to deal with is the same as it’s always been. And that’s us.

 

 Images courtesy © ABB/Metro-Goldwyn-Mayer/Getty Images/WEF-RemySteinegger

ipad_issue10This article appeared in the January 2016 issue of Futurenautics.

free of charge as a PDF, read online and subcribe

Author: admin

Share This Post On