Date
June 2023
Review issue
- 33
Author
Neville Stanton, Professor Emeritus of Human Factors in the Transportation Research Group at the University of Southampton
Share
A brain as good as ours
The automotive industry has witnessed huge change over the last decade, with the effort to reduce reliance on fossil fuels seeing exponential growth in the sale of hybrid cars and electric vehicles (EVs). But what about the much-touted yet still unseen fully autonomous, mass-produced driverless car? One expert suggests there is a long wait ahead for that change to come.
Neville Stanton has spent his life studying why people behave in certain ways. As Professor Emeritus of Human Factors in the Transportation Research Group at the University of Southampton, he has been teaching and researching human factors, ergonomics and, originally, psychology since the mid-1980s.
His expertise and insight have also been used in designing high-tech systems: submarine control rooms for the Royal Navy, flight decks for civil aviation and the development of driverless cars.
For instance, in one project for Jaguar Land Rover, Stanton was asked to assess how quickly people respond when they need to take control of an automated car in an emergency.
His team’s research revealed huge variability, from a few seconds to half a minute. “It was a shock for the vehicle industry,” recalls Stanton, “because they’d been working on the idea of 10 seconds.
“The beautiful thing about human beings is that we’re all different. But that makes design quite difficult.”
Stanton’s response was to make the handover protocol as customisable as possible, and the times improved. “Not dramatically – but people were happier. When we had a one-size-fits-all, most people were pretty unhappy: there was either too much information or not enough. Finding that Goldilocks point was difficult, so we allowed drivers to find it themselves, and it made a huge difference.”
Stanton has seen vehicle automation take enormous strides since a car “with a boot full of computers” first drove him, mildly terrified, around a test track. But he remains sceptical that fully driverless cars will be with us anytime soon.
For a start, he believes being driven in an automated car can be more stressful than manual driving. Look at driving instructors, he says. They are required to watch the learner driver, the road, and other road users and make constant informed decisions about if and when to intervene.
“That’s extremely stressful and demanding, much more than driving yourself,” he argues. “Yet that’s what we’ve designed for vehicle automation.
“People can’t keep it up. 20 minutes, that’s roughly the average, and you’re drained. People aren’t designed for constant vigilance – but you’re expected to constantly maintain vigilance over the automated system.” As he says with a wry smile, no one sits in front of a full washing machine in case it breaks down.
“We go off and do something else. Now, if we could do that with vehicle automation, that would be fantastic – the point of automation should be to allow you to get on with something else. But unfortunately, at the moment, it’s not good enough.”
Stanton believes cars will need artificial general intelligence – “a brain at least as good as ours,” in other words – before they can be fully automated.
“Arguably, on the motorway, you could use artificial narrow intelligence; we’re creeping towards that. But it would have to be a lot more intelligent to manage the complexity of town driving, where there are pedestrians, cyclists, balls bouncing out and little boys chasing after them. It’s complex and demanding.”
After all, he adds with a hint of exasperation: “Even driving manually we still manage to run people over. So we’re a long way off. People make guesses of 50 to 100 years, but who knows when automated general intelligence will be with us. For motorway driving, well, we keep on saying 10 or 20 years – but we’ve been saying that for 30 years.”
Stanton’s slight exasperation at road users’ remarkable ability to keep killing and maiming each other is perhaps understandable. It is an area he has long worked in, and most recently was involved in a ground-breaking project to overhaul how the UK assesses road traffic accidents.
The UK has long had independent national investigation units for air, maritime and rail, but never for roads. It is an extraordinary omission: around 1,700 people are killed on Britain’s roads yearly, with tens of thousands more suffering life-changing injuries.
By contrast, in aviation, around 50 people die each year; in rail, around 300 (tragically, most are not accidents); and in maritime it is around 15. “The numbers on the road are phenomenal,” concedes Stanton. So, while aviation, maritime and rail accidents will be reported to the relevant national investigation unit, on roads it falls to the local police force to investigate. This means findings are localised which, Stanton argues, is not conducive to broader understandings or learnings.
“A traffic cop, who may or may not have had any training in a road safety investigation, writes the report and tries to identify the causes. It’s normally things like careless driving, drink and drug driving, not wearing a seatbelt, excessive speed and distracted driving (including mobile phone use). They call them the fatal five.
“But what we need to do is get underneath that: we consider those things to be the symptoms, not the causes of road traffic collisions.”
Over the last half a century or so, the number of people dying on British roads has come down dramatically, but there has been little movement over the last decade. A new approach was clearly required. “There are no more new learnings to be had using the traditional methods of education, enforcement and engineering,” argues Stanton. “We need to dig deeper – we’ve done all we can that way.”
Working on a project funded by the Department for Transport and National Highways, and managed by the RAC Foundation, Stanton helped build the case for a national Road Safety Investigation Branch. He defined what methods the project should use and helped train the investigators, before carrying out a meta-analysis – which revealed, clearly, that many factors that create preconditions for collisions were going undetected, and so unchanged.
The research, together with other reports from the RAC Foundation, convinced government, and a Road Safety Investigation Branch was announced by the Department for Transport in June 2022 (a formal launch date has not yet been confirmed).
Stanton says the annual cost of running the unit will likely be between £7m and £10m – small change when you consider road traffic collisions cost a staggering £30bn annually.
Plus, he says with a rather resigned air, “Government has a role in protecting people against themselves. People worry about flying – but the most dangerous part of your journey is driving to the airport.”
Stanton’s fascination with why people behave and react as they do remains undiminished. Although he has retired from day-to-day teaching, he maintains an office at the university and continues to carry out consultancy work. “Even though I’m retired now, I still want to learn more; I’ve never stopped asking why,” he says.
“It gets you a long way and gets people to start thinking there may actually be a better way of doing things – of looking to the future and saying, we could do this better. More efficiently, more effectively and more enjoyable. There’s an awful lot of frustrations people have in their work – we need to think about doing things differently.”