Australia is likely to be an early adopter of driverless cars, and they could be on its roads in as little as 10 years’ time, according to some of the nation’s leading researchers in autonomous vehicle technology and policy.
Speaking at an online media briefing ahead of Australia’s first on-road trial of driverless cars in South Australia this weekend, researchers involved in the initiative said that, based on a combination of manufacturers’ predictions and Moore’s Law, we might expect to see fully autonomous vehicles driving on Australian roads somewhere between 2025 and 2030.
And they even foresee a time in the not-too distant future when it will be illegal for humans to drive cars themselves in most on-road situations.
But, of course, there are a multitude of factors that will need to fall into place – laws changed, public confidence gained, technology improved, bugs ironed out, etc – before this can happen.
To this end, South Australia last month became one of the few states in the world – the US states of California and Nevada are others – to pass legislation allowing the trial of autonomous vehicles on public roads.
And though the rule change was not necessary for the November 7 and 8 trials – these will take place on a closed section of the Southern Expressway – it positions South Australia as a key player in this potentially huge emerging industry.
The trials – which will coincide with the Southern Hemisphere’s first international conference on driverless cars, to be held in Adelaide – will test manoeuvres including overtaking, lane changing, emergency braking and using the highway’s on and off ramps.
It’s all part of the Australian Driverless Vehicle Initiative – a huge project involving many different stakeholders, including SA’s Flinders University, the state government, the Australian Road Research Board, Bosch, Volvo and representatives from the Swedish transport administration.
So why are driverless cars such an important development for society – and such an investment priority for tech giants like Google and emerging companies like Uber? And what will it take to get them on our roads?
As Professor Rocco Zito – Head of Civil Engineering at Flinders University and developer of two autonomous vehicles – has noted, we actually already have autonomous vehicles on our roads.
In fact, as he told the media briefing on Monday, most of the world’s major auto makers currently have models with some level of autonomy, ranging from collision avoidance, to lane departure warnings, and automatic parallel parking.
As for the kind of autonomous vehicle most people picture – with the driver’s seat passenger sipping on a coffee and reading the paper as the car ferries the kids to school – that is some years, and much policy development, away.
“We have the technology,” Zito told the media briefing on Monday. “It’s not really a technology question any more, it’s really an integration question.
“We need vehicle to vehicle communication and vehicle to infrastructure communication – that is, vehicles talking to each other and talking to the infrastructure.
And we need law changes, Zito adds. “Driverless cars is not the only technology today that is at least one step ahead of policy,” he said.
Then there is community acceptance. What do we need to do to convince drivers to take their hands off the wheel – which, by the way, is still currently illegal under the Vienna convention?
And then there’s the coordination of it all, which, says Zito, is what the stakeholders in the Australian Driverless Vehicle Initiative are trying to get happening.
According to Professor Michael Regan – the chief science officer for the ARRB group who has been leading the Australian Driverless Vehicle iInitiative for about a year now – we are currently in phase one of the journey to fully driverless cars.
“This is the demonstration phase, where we work out what needs to be done to prepare Australia for driverless vehicles; including in the department of public acceptance – to gauge how people feel about these vehicles.
To Regan, the human factor will be one of the big challenges in the uptake of autonomous vehicles: after all, it will be the humans who still need to take over control of the cars if they fail or reach the limits of their control.
So there are issues of driver inattention, skill degradation and even increased incidence of motion sickness that need to be addressed.
The bottom line is, fully autonomous vehicles don’t quite work 100 per cent yet. Sure, Google’s driverless car has clocked up 1 million miles of testing “virtually incident free”, but as UniSA research professor Anthony Finn wonders, is 1 million miles enough? Should 5 million miles, or 10 million miles of incident-free testing be required here?
“The cars are better at operating in some areas than others,” Finn told the briefing. “If a new stop light appears overnight, they may not know to obey it.”
Their mapping systems alone, he says, would require a sort of “google street view on steroids.”
“They don’t really discriminate between dangerous and harmless situations. They don’t interpret the intent of people – and humans are rather good at that.
“And they’re not very good at making ethical decisions; obviously there is a sliding scale of what to do in an accident. …These are all deficiencies (in the technology).
“We have to work out what it will take to convince average people to get into one of these cars and allow it to drive you from here to Sydney – or would you put your kids in there?” Finn said.
But Professor Toby Walsh, a UNSW expert in artificial intelligence, thinks driverless vehicles will be very much welcomed by society, for the reasons of safety and sustainability, and will be transformational in these areas.
“I think we will one day look back and wonder why we let people drive such dangerous cars in the past,” Walsh predicts. “I suspect people will only be allowed to drive cars themselves on race tracks,” or other specially designated areas.
Another intelligent transport expert, this time from Victoria’s Swinburne University of Technology, Hussein Dia believes the mass adoption of driverless cars could be disruptive on many levels and to many sectors.
He quotes a study conducted in the Portuguese capital, Lisbon, the results of which suggested that driverless vehicles used in peak hours could deliver the same level of mobility – that is, the same number of car trips – using only 35 per cent of the number of vehicles.
In a 24-hour scenario, the study found you would need only 10 per cent of the existing number of vehicles to achieve the same amount of mobility – but only if you had a high penetration of efficient public transport.
“Why is Uber investing in driverless vehicles? Apple … is rumoured to be moving into this space as well. Why is Silicon Valley interested in this space? Why is Google investing in the shared mobility spaces?
“This will transform our society in a major way,” Dia said on Monday.
Not least of all, the insurance industry. Driverless vehicles could reduce accidents by 90 per cent, says Professor Dia, and thus reduce car insurance premiums by as much as 75 per cent.
As to when all this will happen in Australia, the consensus seems to be somewhere between 2025 and 2030.
“Manufacturers are saying around 2030, maybe,” said Zito. “it will depend on when we get vehicle-to-vehicle and vehicle-to-infrastructure communication happening.”
“Australia will be an early adopter,” said Walsh. “We have an ageing workforce of truck drivers, high unemployment rates. We’re likely to be adopting these technologies very early on.
To Regan, public acceptance will be a key barrier. But in the end, “it will happen fairly quickly, where we will have more and more situations where vehicles can drive fully autonomously.”
And Professor Finn agrees: “When it happens, it will happen quite quickly. Using Moore’s Law as a guide, perhaps 2025-2030.”