The self-driving car is arriving right now – not all at once, but model by model and step by step.
By Toby Hagon
Only a few years ago, the car that drives itself was Jetsons-style science fiction – nice to dream about but unlikely to appear in suburban driveways.
In 2016, the picture is very different. With the flick of a digital switch, fledgling electric car brand Tesla recently added software to its autopilot system that allows its cars to be parked remotely via a smartphone app. Mercedes-Benz is about to offer its new mid-sized E-Class with an Active Lane Change Assist function that can recognise other vehicles and lane driver.
BMW, VW Audi, Ford, Nissan and other car companies are in hot pursuit, collectively investing billions of dollars building vehicles that are starting to take the steering wheel and pedals out of the equation.
Step by step, these technologies are already making car travel not just more relaxing but also safer and less congested. This is shaping up as the biggest change to car travel since Henry Ford built his first assembly line more than 100 years ago.
Setting the schedule
Among the experts, there’s no longer any argument against the idea that car control is starting to change radically. The debate is now over how far that change will go, and how fast. Will the completely self-driving car be on the market by 2020 or 2030?
Technology giants Google and Apple, neither of which has ever sold a single car, are working on hugely ambitious projects to develop completely self-driving cars. Autonomous car expert Brad Templeton, who has advised Google on its vehicular ambitions, believes mass production of fully self-driving cars will happen rapidly once the technologies prove themselves.
The car companies are trying to evolve by “taking a car and putting a computer in it and trying to make it drive better”, he says.
“This technology is almost certainly going to come out incrementally.” Chris Urmson, Google
Templeton thinks the more revolutionary approach of Google and Apple will do better but acknowledges that the traditional car makers have been moving ahead, too. For the average consumer buying a car today, these advances are what matter most.
Templeton nominates Mercedes-Benz as the most progressed with autonomy. Its driverless program began in 1995 when the company set up its Silicon Valley research and development centre. Mercedes’ US/German combination is now creating driver assistance features – such as auto-braking, active lane-keeping, adaptive cruise control and blind spot warning systems – that reduce or eliminate crashes and let drivers do less work in mundane driving situations such as freeways and thick traffic.
This approach allows people to enjoy driving when they want, and it has already spread beyond the firm’s big luxury S-Class models.
BMW is also well advanced with driverless technology, as is rival Audi (part of the Volkswagen Group), electric car specialist Tesla, Nissan’s Infiniti arm and the ever safety-oriented Volvo. BMW may have an extra edge: it has partnered with Chinese search engine giant Baidu to test autonomous vehicles in China, the world’s biggest car market.
Bumps in the road
The cars now on the market don’t drive themselves full-time, but they can remove much of the work of driving in certain situations, including long freeway trips, traffic jams and parking.
Self-driving capabilities are now starting to outstrip legislation. For example, the latest BMW 7 Series can be remotely parked using a key fob, though that feature wasn’t offered at its Australian launch, because it didn’t yet have regulatory approval – something BMW and others are working on.
This approval issue highlights a challenge also noted by Boston Consulting Group’s Xanthi Doubara: car travel is regulated by cities, states and nations, and all have to be brought on board.
“There is still a lot of work to be done on various sides – not just on the car manufacturer and technology company sides, but also on the legislator and regulator side,” says Doubara, who expects fully self-driving cars only in 2025 or later, after much real-world testing.
The established car companies mostly agree that further progress towards full autonomy – where people take no part in driving – is an extremely complex technical task. Mercedes describes “very high hurdles” for driverless technology; even its most optimistic experts see full autonomy as arriving in 2025 or 2030 at the earliest.
“All of the research we’re doing at the moment is showing people are increasingly interested in not having to drive.” Gerard Waldron, Australian Road Research Board
Experts agree that freeway driving is the focus for early driverless technology, as it’s relatively predictable, with little or no side traffic, cars moving in the same direction, good road markings and, generally, a lack of vulnerable road users, such as pedestrians and cyclists. Suburban driving is more challenging and city driving can vary vastly across the world.
At Ford, vice president of research and advanced engineering Ken Washington argues that vehicles will need “a high-resolution, detailed annotated map” before full autonomy becomes a realistic prospect.
Boston Consulting Group says car sensors, too, need extensive development. Washington agrees, saying fully autonomous vehicles “require sensors that really allow you to identify high-resolution objects in the environment”.
Ford recently incorporated third-generation laser sensors in its development vehicles. They can see about 200m ahead, allowing speeds above 100kmh with the ability to react to obstacles ahead. Even with that foresight, many self-driving vehicles will initially be confined to certain areas, roads and conditions, according to Washington. For instance, bad weather might make autonomy temporarily impossible.
Google and other researchers, including autonomous machines expert David Mindell, have also highlighted the “handoff problem” – the task of ensuring drivers are ready to take back control of the car when the semi-autonomous system can’t cope. Google has named this as a reason their work is centred on full autonomy.
Generating creative and innovative ideas: enhancing our creativity
Google joins the fleet
Google has been working on a fully autonomous car for several years. Templeton says it has a unique advantage: almost 2.5 million km of “real testing on real roads mixed in with other cars”. Put all other car companies together and they don’t have that much data.
Google has long said it is aiming for a fully autonomous car by the end of this decade. Now it, too, seems to be acknowledging autonomous cars are arriving step by step. Google’s self-driving car project director Chris Urmson said in a March speech that the technology was “almost certainly going to come out incrementally”.
He also conceded that the first fully self-driving cars might be restricted to clear weather and good roads and that self-driving cars must improve on what we already expect on the road.
“When it does come out, we’re going to want it to be better than human drivers.”
Yet the benefits of full autonomy are enormous, with everything from driverless taxis and trucks to vehicles transformed into an office space or entertainment room. Mercedes-Benz gave a tantalising vision of what could be possible in a self-driving car with the radical Vision F 015 concept. Seats that spin around to create a lounge-like space as well as giant screens were key to the car’s inviting interior.
Going all the way
Gerard Waldron is the managing director of the Australian Road Research Board and one of Vehicle Initiative, a consortium of about 60 companies keen to establish a framework around which autonomous technology is introduced and deployed. It includes insurers, road agencies, governments, car makers, tech providers and banks, and Waldron sees a different post-autonomy world.
“Once you do get rid of the steering wheel, there’s a paradigm shift in the way people buy cars, use cars or get access to cars,” he says.
“All of the research we’re doing at the moment is showing people are increasingly interested in not having to drive.”
“Many people are going to switch from buying cars to buying rides.” Brad Templeton, consultant
Along with electric propulsion, Waldron believes driverless tech could up-end the traditional car production and sales model.
“Why would you tie up 50 grand [for a car] to sit 23 hours a day in your driveway or in a car park?” he asks.
“I think most of us will end up with a mobility plan that is pretty much like your mobile phone plan. There will be mobility companies with fleets of these vehicles … and you will come to a deal with them.”
Waldron goes as far as to suggest that some car makers may not adapt: “Everything that has made a car company strong in the past is potentially now a weakness,” he says.
Templeton agrees there will be enormous changes. “For many people, they’re going to switch from buying cars to buying rides,” he says.
“It’s going to be an entirely different way of thinking about buying cars, what type of cars we want, what kind of cars people make and what companies make them.”
Doubara, on the other hand, thinks traditional car makers are well positioned and attuned to the importance of advancing autonomy. Her extensive research shows an industry ready to react, if anything to protect its patch.
Driving social change
Society, too, is set to change. With autonomous cars dominant, parking stations would no longer be required in built-up areas, and there would be no queuing for a space; you could leave the car at the entrance to the car park – or simply alight out the front of your destination. Doubara believes driverless ride sharing could ultimately compete with public transport, reducing the need for costly rail networks or other infrastructure.
But crash avoidance is the most obvious – and appealing – part of the autonomous driving puzzle. All things working well, self-driving cars will be involved in far fewer crashes than the human-controlled vehicles of today.
“The benefits of highly connected autonomous driving cars are clear, with all the promise of making driving far safer, less stressful, more efficient and more rewarding,” said Carlos Ghosn, CEO of the Nissan-Renault Alliance, at the recent New York International Auto Show.
“There is still a lot of work to be done on various sides.” Xanthi Doubara, Boston Consulting Group
Templeton says there is the potential to save more than 1.2 million lives globally each year.
“The sooner we stop killing all those people the better,” he says, adding that the annual road toll is far greater than the number of war-related deaths. “We have the chance to stop it.”
When there are crashes, Templeton says apportioning blame will be done in a matter of seconds.
“There’s going to be a complete 3D recording of everything and logging of everything,” he says. “There won’t be any arguing about who’s at fault; it’ll be more going into details of, ‘Was somebody negligent?’”
Experts broadly agree the arrival of driverless cars is not a question of if, but when. Whenever that day comes, we look to be heading towards a very different driving world.
Not quite like the Jetsons, but an upheaval the car industry hasn’t experienced before.
Home drone delivery lifts off