As the self-driving car revolution chugs along, it appears that the British public isn’t quite ready to take things into second gear.
A survey released today by global technology company Thales revealed 57% of UK citizens would not feel safe in a self-driving car.
Timeline for Automation
- February 28, 2020
- October 24, 2019
With the UK government’s goal to ensure there are “genuine driverless vehicles” on Britain’s roads by 2021, this latest research suggests that government and industry have much to do to ensure public perception keeps pace with the rapid advancement in driverless car technology.
Why are people afraid of driverless cars?
According to the World Health Organisation, there are around 1.35 million deaths globally as a result of road traffic crashes.
The number of these deaths caused by human error range from 90% to 95%, suggesting that the biggest danger to humans on the road is other humans.
The central goal of self-driving cars is to drastically reduce these fatalities: an algorithm doesn’t drink, text, or get tired.
But this idea is yet to permeate into public consciousness. 49% of those surveyed believe that self-driving cars will cause a rise in potentially fatal accidents.
Where does this fear come from?
At a roundtable discussion on self-driving car safety hosted by Thales at the Royal Institute, London, two of the key reasons for the lack of public trust in autonomous vehicles were the fear of the unknown and the much-publicised self-driving Uber crash last year, in which a car in autonomous mode fatally struck a pedestrian.
But what can be done to make the public feel safer in a self-driving car?
Simulations: “The worst driving test you’ve ever been on”
One of the ways in which autonomous cars can be safer than a human driver is the number of hours of testing it can go through in a simulation.
The idea is to perform millions of hours of testing, a much greater scale than would be possible to do in the real world, in an environment that is completely safe.
The State of Technology This Week
“If I can create a virtual environment that has all of the richness of the real world and the different grouping entities within it – different road layouts, different materials and weather conditions – that an autonomous vehicle will be expected to encounter, then I can start to create different scenarios for that autonomous vehicle in order to find out how it’s going to behave in those conditions,” said Timothy Coley, product specialist at XPI and Thales UK.
But how can this level of testing be proven to the public?
One idea is to create a government-recognised certification that proves a self-driving car has been put through a sufficient number of hours of simulated and real-world testing and under appropriate conditions.
“It’s a bit like a driving test,” explained Kirsty Lloyd-Jukes, CEO of Latent Logic, a spin-out company from the Computer Sciences department at Oxford that is combining simulations with real-world tests on the roads of Oxfordshire.
“It’s like the worst driving test you’ve ever, ever been on. Imagine a mean instructor, all those horrible situations that you hope you’ll never have, we’re going to put the self-driving car against all of those and see if it passes.”
Latent Logic is also involved with OmniCAV, a DfT funded initiative to create a proof-of-concept for certifying self-driving cars.
Lloyd-Jukes believes that it’s important to have a global dialogue around what the certification process should be for a self-driving car and as part of their project, Latent Logic is involving the general public in the process.
“With the general public, we’re asking them ‘what should the driving test be? What would you need to know, to know that a self-driving car is safe?’” she said.
Trials and demonstrations
Another way to shift public perception and is to make the unknown known. Professor Paul Jennings, a physicist with Warwick University’s Manufacturing Group, said that it is “encouraging” to see more members of the public being taken through trials and demonstrations of self-driving cars.
“If you speak to people who have been lucky enough to be in a self-driving car, you often get the same response, that for the first couple of junctions people are on edge, then suddenly, you learn to trust it through experience and after a few miles it becomes normal,” he said.
As well as improving trust, he added, you can start to measure some of the societal benefits from these trials, such as easing congestion and increased mobility.
Improving self-driving car security
With self-driving cars connected to a large network in which a stream of data is shared with other cars on the road, the public also fear that self-driving cars could be subject to cyberattacks.
Nearly a third (29%) of the public is concerned about this, with slightly more than a third (35%) worried about connectivity failures.
Gareth Williams, VP of cyber security at Thales UK, says that the public is right to have these concerns.
“All of those things create a new cyberattack surface that we haven’t necessarily seen in the field of automotive so far,” although he added that such threats are already present in other industries.
To solve this problem, he said, manufacturers will need to ensure that cybersecurity is built into self-driving cars from the outset.
“But it takes more than that,” he explained. “You need to create groups of trust, you need to have the right trust componentry and management infrastructure around how you are going to look after that data, deal with how, when you’re moving huge amounts of software, […] how you protect those things and how you make sure that if something happens you’re able to control it and deal with those problems.
“It is about embedding those cyber principles right from the start rather than coming up as a secondary concept and idea”.
Self-driving car safety: Time will tell
The panel added that the automotive industry can learn from other industries, such as aviation, which shares lots of its data among companies to improve safety.
In fact, 65% of those surveyed said they feel safe when flying on-board an aeroplane. And, where pilots must regularly practice in a simulator, a self-driving car could also have simulator check-ups to ensure everything is operating smoothly – a sort of virtual MOT.
Simulations, certification, demonstrations and security are all areas that can improve public perception, but history has taught us that once a technology becomes widely adopted, safety fears eventually dissipate.
During the opening of the first major UK steam train network in 1830, for example, the local MP was killed after being struck by a passing carriage.
“So it got off to a fairly inauspicious beginning,” said Coley, “but now rail is one of the most reliable and safe modes of transport. So whatever fear or misunderstandings there are at the very outsets of this technology, it’s clear that there is a long way to go before really establish self-driving cars as being safe.”
Dr Alvin Wilby, VP of research, innovation and technology at Thales UK, added that “people do get used to it [new technologies]”, pointing out that London’s “DLR has been quite happily operating without drivers since it came into service 25 years ago”.
The full Thales white paper Making driving as safe as flying in an autonomous world can be found here.