Being a driverless car passenger proves ‘unsettling and extraordinary’

How many people does it take to drive a driverless car? Five: a safety driver behind the wheel, an operator to program the route, and three engineers monitoring it in another car behind.

It is, to be fair, barely even a prototype. The autonomous car unveiled in Milton Keynes last week is bleeding-edge engineering, Britain’s entry in a global race to get the first driverless car on the road.

The converted Range Rover Sport can steer itself, speed up and slow down, stop at red lights and move off when they turn green. It can even cope with roundabouts, a fundamental skill in Milton Keynes. The five operators are there to examine every nuance of the car’s reaction to the ever-changing road conditions – cyclists, pedestrians and other drivers, and the weather, to name a few.

The public demonstration of the car by UK Autodrive, a consortium led by engineering company Arup, supported by Jaguar Land Rover, Ford and Tata Motors, should have been a celebratory milestone for British motor manufacturing.

Yet growing excitement about self-driving cars was shattered by the death of Elaine Herzberg in Tempe, Arizona, last week. She was hit by an autonomous Uber that did not apparently detect the 49-year-old when she wheeled her bicycle across the road at night.

“It’s dreadful,” says Tim Armitage, the project director for UK Autodrive. “It’s dreadful for the person involved – everyone involved. And it shows just how important it is to make sure that what we are demonstrating today is safe and that we don’t oversell the technology.”

UK Autodrive’s engineering efforts are focused on safety, but that is not the only concern. The government has invested £250m into several major research projects, involving at least 1,000 people, Armitage estimates. By 2035 the Department for Transport expects the industry to be worth £50bn to the economy – about a third of all UK manufacturing, although the current motor industry contributes about £58bn. Still, there’s a lot riding on the success of self-driving cars.

The government strategy for getting there is all about research. By dangling substantial grants, it has managed to corral car companies, universities and other interests into working together – a contrast to the US, for example, where Uber, Google’s Waymo and Toyota are entirely competitive.

In the UK there are 15 government-sponsored projects led by four major consortiums like UK Autodrive which are coming at the problem from different angles – predicting the behaviour of other drivers, cyclists and pedestrians, understanding the needs of elderly or disabled drivers, and the challenges of motorways.

All the research returns to the same theme: how to make roads safe while allowing people to travel where they need to.

As with any technology, there are glitches. On the demonstration drive in Milton Keynes, we have a near-miss when the car lurches forward at a junction in the car park. The safety driver, Jim O’Donoghue, saves us from an impending collision with a parked car by grabbing the wheel.

Afterwards he’s not quite sure if he realised the car was going wrong or if it warned him it couldn’t cope. “I’ve been driving it for several weeks, so I’m really tuned in to it,” he says, underlining the unconscious, near-visceral familiarity that people have with technology they use regularly. Driver behaviour and human-machine interaction are all important elements of the research.

Being a passenger in this Range Rover is like being driven by a clumsy taxi driver. Acceleration feels aggressive – from the traffic lights, it’s foot-to-the-floor until we hit 30 miles per hour. Braking starts rather later than I would like.

Once up to speed, it drives extraordinarily close to the kerb. The side sensors can judge the distance much more precisely than a human. It all adds up to an unsettling experience that feels entirely unlike riding any other vehicle. But the car, it must be emphasised, is still being built – sensors and controls are adjusted on a daily basis.

These cars are the most eye-catching part of the research, with the aim to design a vehicle that can cope with any situation more safely than a human driver. The other, more immediate focus is on connected cars – using mobile phone technology to allow every vehicle on the road to talk to each other.

Another demonstration shows how drivers in standard cars can be given information, such as whether cars in traffic ahead are braking sharply, or if an ambulance is approaching.

Possibly the most compelling project is on finding parking spaces. Cars with visual sensors can detect parking spaces and share the information with a network of connected cars, and inform the driver where the closest spot is. “About 30% of traffic is people driving round looking for parking spaces,” Armitage says. “In the future we might need far less car parking. Imagine what you could do with the space.”

The research invites questions that are much more fundamental than simply how to build a safe, self-driving car to replicate the UK’s fleet of driver-owned internal combustion vehicles. Instead of owning or leasing one vehicle for all your needs, people could access different types of vehicle for different journey.

Another fundamental question is how much UK roads might change to enable self-driving cars to work more effectively. Cars could connect with each other, but also with road furniture like traffic lights, prompting drivers to slow down ahead of red lights.

The growing infrastructure around electric cars shows how quickly this could happen. Arup laid electric cables and charging infrastructure in Milton Keynes when the council decided it would try to become the centre of electric cars in the UK. “When they first arrived, there were complaints from drivers of internal combustion engine cars that too many parking spaces were given over to charging points. Now they’re all full.”

One of the projects closest to completion involves self-driving pods. The Gateway consortium in Greenwich peninsula, London, has been operating four driverless pods in pedestrian areas, examining how members of the public react.

There are similar pods in Milton Keynes.“We’re hoping that will be ready before the end of the year,” Armitage says. “Singapore is very interested in what we’re doing., “They want autonomous buses because they can’t recruit enough drivers.” It’s a potential export opportunity that would establish the UK’s self-driving credentials on the world stage. So while this Range Rover may not be fit to drive unsupervised yet, we are likely to see more autonomous vehicles operating very soon.

(THE GUARDIAN)