The future is almost here. Back in the days when science fiction was optimistic rather than dystopian, we were promised all sorts of mobility boons: moving pavements to spare us the trouble of walking, jetpacks to propel us from A to B at hernia-inducing speed and, of course, driverless cars, which would never crash, never lose their way and never make their passengers worry about breath-testing, directions or competence.
Needless to say, jetpacks are mentioned in nobody’s dispatches and the high-speed moving walkway installed in 2002 at the Paris Metro’s Montparnasse station was quietly abandoned a few years later after a series of passenger tumbles. Driverless cars, however, are a more realistic proposition entirely.
When it comes to development, the United Kingdom inevitably lags behind the technical behemoth that is Silicon Valley, where the market leader, Google-affiliated Waymo, has already tested autonomous vehicles on ten million miles of public roads in 25 cities since 2009 and run billions of miles of computer simulations.
There has been serious US investment from Google ($1.1bn from 2009 to 2015), Ford ($4bn until 2023) and General Motors/Honda ($19bn through GM’s Cruise Automation offshoot), while the world’s biggest car company, Japan’s Toyota, has piled $1.3bn into Uber’s CAV (connected autonomous vehicle) Unit alongside Denso and SoftBank.
The UK is probably behind China, too, where Audi, BMW and Daimler have been allowed to test vehicles in Beijing and Shanghai and driverless cars have already used the 5G network to share data in Fangshan.Yet the sheer size and varying road/climate conditions in the US and China will cause difficulties, unlike in Singapore, where its small size, homogeneous roads and technical sophistication combine to make the city state a world leader. Even so, Britain remains in the fast lane, according to the Society of Motor Manufacturers and Traders (SMMT).
The goal is revolution but the process has to be evolutionary. There are six levels of AVs (autonomous vehicles), which will be phased in gradually. L0 is a vehicle with no automated features whatsoever. Already prevalent, L1 supports rather than replaces a driver with extras such as cruise control. Automation steadily increases through L2, L3 and L4 to the CAVs of L5, which will be connected to the 4G/5G infrastructure via sensors on street lights in urban areas, suggests Justin Benson, UK head of automotive at KPMG.
At L5 level, the need for any driver input will be eliminated. "That’s when you’re at the point where occupants are sitting in the car reading The Daily Telegraph," says Steve Cropley, editor in chief of Autocar. L5 and L4 cars will be connected to each other but the relationship with lesser-connected cars remains unclear.
The speculative nature of a sea-change that will affect all our lives makes the future fiendishly difficult to predict. Each major player has different ideas about when, where and how the revolution will take place.
BMW predicts fully driverless CAVs will be operational between 2025 and 2030. The SMMT, the UK motor industry’s umbrella association, estimates that 100 per cent of new vehicles will be partly connected by 2026, although only 30 per cent will be L2 by 2030 and that L5 vehicles will not be introduced until 2035.
Some things are clear: when the revolution happens, Britain will be in the vanguard, partly because its unitary legal system will make adoption much simpler than in the more devolved, state-based US system.
The SMMT’s report, Connected and Autonomous Vehicles: Winning the Global Market, argues that while Britain is behind the US, Japan and South Korea in terms of enabling infrastructure, it is ahead of the game in the potential for automating roads and it leads the world in enabling regulations and market attractiveness. That, the SMMT believes, means the UK is the "world’s number one location for mass-market potential of connected and autonomous vehicles".
Google-affiliated Waymo has tested AVs in 25 US cities
Whether the UK ultimately wins the global race to market will depend on a number of factors, not least sustained support from government and an orderly exit from the European Union. "How the motor industry will react is a question we’ve put to them," adds Cropley.
"They say nothing changes overnight and that demand and machinery change at different rates. What is certain is that engineers who’ve been designing basically the same car for 40 years are totally thrilled. It’s an enormous change but in slow-motion: if you banned the production of diesel cars tomorrow, they’d still be on the roads in 20 years’ time."
Driverless vehicles represent a once-in-a-century opportunity that the automotive industry needs to handle deftly if companies are not to end up like electronics firm Foxconn, as a supplier (albeit an extraordinarily profitable one) rather than innovator, or DeLorean, where fiscal incompetence and hubris derailed pioneering technology.
"A car model has a product cycle of 10 years: three years in design and seven in the showrooms," says Benson. "Car manufacturers see this as an opportunity, but they need to work out how to respond to that opportunity and not get caught out."
There are templates for success, be it Mercedes founder Emil Jellinek grasping the potential of horseless carriages at the turn of the previous century or Ford introducing the assembly line to car manufacture in 1913. More recently, Tesla recognised the value of rebooting the electric car market, which had lain dormant since the rise of the internal combustion engine.
Regulation and policing
As noted, L1 is already here: advanced driver assistance systems, such as autonomous emergency braking, cruise control, self-parking and speed limiters, are increasingly standard. "The technology is here but implementing it is going to be very disruptive to the supply chain, local authorities, banks, the car industry and finance companies," says Benson.
If L5 cars don’t need to be driven by driving-licence holders, theoretically they can be used by everyone: four-year-olds, 94-year-olds, the infirm, the drunk, the drugged and the criminal. In practice, though, nobody believes the existence of L5 cars will mean the end of less technologically sophisticated vehicles. New worlds need to be policed and regulated.
The UK has already looked at self-driving licence requirements and, in 2018, introduced the world’s first AV insurance legislation, although Britain currently differs from most European countries in insuring the driver rather than the vehicle. (One possible solution is that fleet operators or manufacturers may bear the risk, rather than non-drivers of AVs.)
The insurance industry believes the driver must be able to resume full control in L1, L2, L3 and some L4 cars. In its 2017 paper on the topic, the Automated Driving Insurance Group noted that while "vehicle automation offers society and individuals huge benefits in increased mobility and safety", it is concerned by "the gradual increase in levels of automation and the potential confusion and greater risks these intermediate systems may offer". It warns specifically that L1, L2 and L3 cars should not be described as "automated".
Uptake of autonomous cars will surely result in fewer accidents as more than 90 per cent are caused by driver error. Vehicles will probably need to be overseen by fleet operators using a system that may resemble those used in air-traffic control or on the railways, which enable the respective industries to learn from each (rare) accident.
But there will always be the predictably unpredictable: careless pedestrians, unexpected bumps in the road, blown tyres and malfunctioning engines. And when some vehicles are CAV and some are not, who gets the claim and blame when the inevitable happens?
Basic infrastructure will need a radical overhaul. Traffic laws will require updating and universal 24-hour 4G connectivity and smart street lighting will become necessities rather than aspirations. South Korea and the US are already upgrading to 5G. Where does public transport fit in and will universal speed limits become as obsolete as the figure walking in front of an early car with a red flag or lantern?
The SMMT suggests that once the government has invested £10bn in infrastructure and road maintenance, driverless vehicles will ignite £62bn of economic growth a year. Some £46bn of this would be down to enhanced consumer productivity – in its optimistic scenario, every driving commuter will save 42 travelling hours a year because fewer cars will be on the roads.
But new revenue streams are also expected to emerge, as well as 20,000 jobs in areas such as software and hardware development, and vehicle fleet and network management. The wider impact on the UK job sector could be even more marked, reports the SMMT, with more than 400,000 jobs created in adjacent industries, including telecommunications and logistics.
The social benefits will be legion, too, the organisation claims, with 3,900 lives saved and 47,000 serious accidents prevented until 2030, when, it predicts, a world-leading one-in-every-five UK miles travelled will be automated, and there will be 18 million driverless cars on the world’s roads.
Where will the AV revolution take its first steps? Rural areas with minimal traffic are as unrepresentative as central London, where traffic shuffles along at walking pace, although for driverless cars to fully integrate into the transport system, both extremes will have to be accommodated. Expect a gradual phasing in. Financed by government and industry, tests have taken place in Milton Keynes, Coventry, Bristol and Greenwich.
The GATEway Project tested automated valet parking, driverless grocery deliveries and a shuttle service in Greenwich with £5.3m of government funding. The technology company FiveAI – part of a consortium including Direct Line, Britain’s leading personal motor insurer – was awarded a £12m government grant to develop CAV software for an autonomous car system. It will test cars in the London boroughs of Croydon and Bromley next year.
Benson sees the immediate future as an automated trial at the 2022 Commonwealth Games in Birmingham, followed by a more extensive test in a "geo-fenced" area of a Midlands city, that is, one from which cars with drivers are banned. "It might be different in China but no politician in Western Europe would permit a major trial in a non-geo-fenced area," Benson says.
Technicians analyse data while testing an AV in a pedestrianised zone in Milton Keynes
Naturally, where there are trials there will be errors and humans will remain all too human. "Watch a school bus drop-off," suggests Cropley. "The kids will go different ways but instinctively you know that most – but not all – of the time, one kid will go around the back of the bus and you’ll drive accordingly. I’m not sure machines could anticipate that."
An accident in February 2016 in Mountain View, California, illustrates the kind of subtle fine-tuning required if CAVs are to fulfil their promise. The algorithms in one of Google’s self-driving Lexus SUVs assumed that a bus driver in an oncoming lane at traffic lights would slow down. Assuming the other vehicle would stay put, the bus driver drove on. The resulting crash – in which no one was seriously hurt – prompted Google to introduce 3,500 tests to make sure it didn’t happen again.
There are myriad ethical issues. The headlines when – as will surely happen – a driverless car is involved in a fatal accident for the first time on British roads will write themselves as surely as cars will drive themselves. Indeed, it’s already happened in Tempe, Arizona, when 49-year-old Elaine Herzberg was fatally injured, struck by an Uber car in driverless mode as she tried to cross a major highway with her bicycle.
In the inchoate stages of the revolution, each incident will increase fear and suspicion until it can be proved that there are fewer accidents and the public accepts that road hazards can never be eliminated. Once people have stopped fearing the new technology, human nature being what it is, they will become blasé about it. "People will just step out into the road assuming cars will stop," says Cropley. "That level of complacency needs to be factored in."
Noting that driving-licence applications in the UK have dropped by 6 per cent over the past few years, Benson envisages a generation growing up less attached to owning cars. "Why own a car? People will need a family car for a long time yet, but the number of cars each family owns has been dropping for some time." He foresees an alliance between local authorities, finance companies and fleet operators, whereby customers will use driverless cars for individual journeys.
"What the industry can’t answer is what happens if there are fewer cars on the roads," says Cropley. "There will be reduced ownership but, since cars will be electric, the government will lose swathes of fuel tax income. There will be increased road pricing and people will become more discriminating about doing a mile. This will mean a shift from car ownership to hire and share. But a car is still an indicator of prestige. The last people to keep cars in their garages will be at the high end: owners of Morgans, Ferraris, McLarens."
Driverless vehicles will offer less obvious benefits, not least to the hitherto marginalised. "Self-driving vehicles have the potential to tackle loneliness and isolation and create economic opportunity," Jesse Norman, the minister of state for transport, has argued.
A 2017 SMMT survey pointed out that those aged between 17 and 24, 65 or over and the physically disabled (71 per cent of whom currently say they find it difficult to travel when and how they want) will benefit from becoming more mobile and not having the stress of driving. "It’ll be a big deal when a driverless car can pull up outside granny’s and take her to hospital," says Cropley.
"The cars will also be reconfigured. They’ll have bigger doors, which will make it easier for her to get in."
Elsewhere, there are vested interests who will not look as benignly upon CAVs. The smiling robot cabbie who drove Arnold Schwarzenegger in Total Recall appears less outlandish today but cab drivers, Uber drivers and chauffeurs may not embrace the new dawn. Nor will lorry drivers, bus drivers or anyone else paid to drive. Yet these jobs will not disappear.
Benson envisages layers of service for hire cars. Someone has to take packages from car to doorstep. One trial in Miami stalled when pizza customers living at the top of apartment blocks refused to come down to collect their autonomously delivered meals. Someone has to load and unload lorries (although a 26-tonne, driverless, 5mph truck is making deliveries in Jonkoping, Sweden, and "platooning" convoys, where driven trucks are ahead of and behind driverless lorries, are already common on US highways), some people will always want to emphasise their social status by employing a chauffeur and what of the emergency services, whose driving cannot correspond to all the rules of the road all the time?
"For the driver, driverless cars mean feet off, hands off and brain off," says Cropley. "The last one is the hardest to deliver. Last year, I was in a supervised driverless car in Tokyo. We came to an elevated road section, tramlines and several car lanes all at once. The car essentially said: ‘Sorry mate; can’t do it.’ There was a seven- or eight-second warning but I had to grab the wheel quickly so we didn’t hit anything."
There are various obstacles that may slow the progression of AVs, not least the swamp of confusion that is regulation. "The question is ‘Whose balance sheet does this go on and if you pay £10 for a ride, who gets what?’," says Benson. Since we’re starting from scratch, it won’t be straightforward but, despite all the vested interests and the trial-and-error nature of the changes, a broad, ever-evolving regulatory consensus will emerge over time, made easier by the fact that while internal combustion engine cars have 2,300 moving parts, AVs will only have around 200.
Although regulatory approval and successful trials should boost consumer confidence, consumer demand is another matter entirely. L2, L3 and most L4 cars will be driverless to some extent. It’s hard to imagine they will be cheap and, until mass production lowers prices, Benson sees organisations such as hedge funds owning fleets of cars, instead of consumers.
Then there is a more mercurial issue. Many people love driving and are attached to their cars in a way they are not to other labour-saving devices, such as their dishwashers. Beyond the ownership issues, many may find the driving experience and the sense of control it brings extremely difficult to relinquish.
"The real enthusiasts won’t want to let it go," notes Cropley. "Nor will the people who enjoy the interiors, the comfortable isolation and who quite like driving. Even so, letting the machine take over in situations such as traffic jams will be a huge incentive."
Adverse weather conditions
Problems caused by weather conditions – currently, automated vehicles struggle in darkness and snow – will be overcome but will congestion reduce dramatically? Accidents and roadworks will not be abolished. Satnav has displaced rather than eradicated queues. A hire network akin to the bike-hire system seen in many European cities will increase availability. The new machines will be electric but driving will not be carbon-neutral.
Looking to the future, the SMMT has advocated four ways forward: traffic law changes; 4G across the entire road network; collaboration between local authorities and the car industry; and harmonising international regulations.
Once those fundamentals are in place, once public misgivings have been assuaged and once L5 CAVs are among us, the future will be the present. Some envision a utopia, others see the introduction of L5 cars as taking decades and then being limited to high-quality roads and geo-fenced areas, one litmus test being whether an L5 car can negotiate an area as packed with pedestrians as Oxford Street.
"Given my experience in driverless cars, I can’t see L5 happening within five years. I’d say it’ll be closer to a decade," says Cropley. "People are so addicted to going to places that I can’t see major reductions in traffic, but there will certainly be some. The motor car has brought huge personal travel advantages over the past 125 years and I welcome the revolution: it would be ridiculous to wish for progress to stop. The population keeps growing and we have to find more sophisticated ways to keep it flowing."
Benson’s future is sunny indeed. "My vision is bright open spaces and clean air, with no traffic jams and no crowding on roads," he says. "Fewer cars will mean fewer car parks, so inner-city space can be used more effectively. People won’t be poisoned by emissions and cities will enjoy a renaissance as places to live and work."
Image credit: Getty images