Text Size

Current Size: 100%

Autonomous vehicles: the threadbare emperor

Former campaigner's picture
Photo: car lights

You can't go anywhere near any transport coverage these days without hearing about autonomous and connected vehicles. Spurred by Google and Tesla, huge expectations are being put into this technology as the future for mobility. The Government has set up an office for connected and autonomous vehicles, and trials are being promoted in Milton Keynes, Coventry, Greenwich and Bristol. With freight, lorry platoons are to be trialled, with the vision of long road trains of trucks driven by a single driver adding to the efficiency and safety of road freight. Now a "modern transport bill" is promised, to allow the UK to seize the leadership in autonomous vehicles technology.

You can see why Governments, and politicians generally, would like all this. They are perennially attracted to shiny new things and as (mostly) non-scientists they tend to be susceptible to sales pitches from technology companies. The rail industry has suffered enough from this in the past – for example, maglev technology has come and gone as the obvious replacement for old 19th century steel wheels on steel rails, yet that technology has shown itself to be adaptable enough to outdo maglev technology in cost and delivery – just ask the politicians around Munich and Shanghai that bought the exaggerated maglev claims and are having to live with the consequences. In the same way, the proponents of autonomous vehicles are telling Governments that this is the next big thing and if they invest and legislate now they can be in at the ground floor and secure first mover advantage. When the proponents and promoters are big names like Google and the big automotive companies, the attraction for politicians is enormous.

These technologies are attractive to politicians in other ways. Autonomous vehicles appear to offer huge safety benefits by removing driver error altogether. They can give older people continued independence, collecting them on demand from their own places and taking them to shops, doctors, friends etc. And AVs can be seen as a real solution to transport in rural areas, avoiding worrying about loss-making bus and rail services. 

But for politicians, this technology has another attraction – it avoids them having to do anything difficult on traffic. Autonomous vehicles, so the argument runs, can give people mobility where they like, when they like. They can be used more efficiently so cutting congestion. They allow people to work on the move, in comfortable cocooned environments. How much better than having to manage, fund and regulate public transport, or tell people that they might not be able to drive everywhere at nil cost. AV appears to offer a way out of politically contentious issues like bus lanes, parking controls and charges or any kind of management of demand.  

The rail industry has drunk some of this Kool-Aid. Rail freight operators are worried about lorry platoons reducing heavily subsidised road freight costs still further (a view allegedly also taken by Lord Adonis, who is said to have told the Government not to worry about rail freight any more because it will soon be superseded). Passenger operators and rail planners are concerned that autonomous vehicles will allow commuters and business travellers to work on the move, removing one of rail's main advantages. Opponents of HS2 have seized on the forthcoming universality of AVs as another stick to beat the project with. 

I would say the rail industry has other things to worry about – costs, delivery, structures, user experience and all the other traditional rail issues. AV competition should come a long way down the list. This is because, having talked to people in the technology world and the motor industry, it is clear that, for the UK at least, the AV emperor is looking distinctly threadbare, and is subject to a huge level of hype, a lot of emanating from California.  

Let's go through the AV claims and the practicalities. Platooning lorries might work in the wide open spaces of North America and Australia – on the British motorway network with lots of junctions close together they have severe practical issues. If they are on the inside lane, how do other vehicles join and leave the motorway? If on the outside lane, there are safety issues of other vehicles "undertaking" them on the inside, and of vehicles joining and leaving the platoon crossing all the other lanes of traffic. These kinds of issues torpedoed a proposal for a car share lane on a newly widened stretch of the M1 ten years ago, and there are no signs that anyone has solved them since. So the DfT talk about smaller platoons of maybe two or three lorries – but even those lengths raise these practical issues, and at that level may be less cost effective anyway. It's no surprise that the platooning trial will be on the M6 through Cumbria at night – it's about the only stretch of UK motorway where these issues are likely to raise fewer problems. 

Safety is another huge issue. Put simply, either AVs are going to have to be programmed to avoid all risks of any sort – so that they will be able to sense possibilities of, say, children running into streets and immediately stop or slow down, or they will have to accept a level of risk. If the former, AVs are going to be a very slow form of travel. Even in rural areas, country lanes with overhanging trees and blind corners offer risky environments (the casualty rate on rural roads is higher than on any other), very different from Google's California campus or even Milton Keynes grid of streets. If a level of risk is decided to be part of the package – and Google appear to be expecting this by trying to patent sticky car bonnets so pedestrians, once hit, stick to the front rather than roll off – who will be liable when a crash happens and people are injured or killed? The "driver" isn't in fact a driver so it won't be them – will it be the car programmers, or the manufacturers of the vehicle or the software? This raises plenty of issues for insurance companies and lawyers to argue over – the Government's Modern Transport Bill promises insurance provisions to solve this, but one technology company executive I spoke to thinks that it will require the Government to put up a trillion pound insurance bond to make AVs at all usable. 

And that's before you get into the potential for hacking into these systems and making the vehicles crash. Anyone who has seen the original version of the film "the Italian Job" may remember the Benny Hill character reprogramming Rome's traffic lights to divert the gold bullion vans towards the criminals – AVs offer much more potential for criminals or terrorists to create havoc and death on our streets. 

But perhaps the biggest problem facing AVs is that the prospects they hold out, at least in urban areas, are for more rather than less chaos. Imagine lots of these vehicles cruising city streets – increased congestion is inevitable, making it harder for people to cross roads unless there is a signalised crossing point, leading to a surge in demand for more crossings.  And if the wilder predictions comes true and AVs replace some public transport, especially buses – perhaps in alliance with Uber (another technology company with a lot of hype surrounding it) – then cities will just become gridlocked. 

Even on inter-urban roads, the AV vision only really has a chance of getting anywhere when every vehicle is autonomous. In a world of part autonomous, part conventional driving, the issues I've raised about lorry platoons and safety will still be huge. Some proponents suggest that vehicles might be autonomous for part of their journey, say on motorways and dual carriageways, and then return to driver control on local roads, but such transitions are inherently unsafe (even now, drivers coming off motorways take time to adjust to local roads). 

I'm also unsure about the user reaction. Part of the AV vision is shared vehicles replacing personally owned cars. There are some moves in that direction with car clubs, but I'm not sure users have really had their voice heard on whether they trust this technology sufficiently to abandon their own cars, or even to move to autonomous personally owned ones. We've yet to see what real users, faced with real choices, will do in practice.

Let me be clear – I am not saying that we should ignore AVs or the technology underpinning them. Clearly all this technology offers opportunities for improving safety and efficiency of our road and rail networks – as one London Underground executive pointed out, making trains start from depots automatically and travel to their starting station to collect their drivers would have potentially large benefits. The "digital railway" plans can make use of much of this technology, and the rail industry can and should take the best of it and apply it.  The "connected" vehicle technology that is, probably wrongly, bracketed with AVs, has lots to offer in better communication between road and rail vehicles and the infrastructure at line- or road-side. 

So I'm not rejecting the technology at all. What I am trying to do is to hose down the hype. The British and European context is different from the US one, and with our cities growing and densifying, the ability of trains (and even buses) to move large numbers of people around in limited space isn't likely to be undermined by lots of autonomous vehicles any time soon. I'm also saying that the hype around AVs masks challenges that policy-makers and legislators will have to face. Until they've done so, I think the rail industry should relax.

This is a version of an article originally written for RailReview.