Hundreds of laws may have to change as the use of driverless vehicles increases in Australia. Liability is just one of the thorny areas to be reviewed, writes DENISE CULLEN.
It was a clear, sunny morning on Adelaide’s Southern Expressway. Perfect driving conditions, you could say, except that the man sitting in the driver’s seat of the car heading down the freeway at 70 kilometres an hour neither had his eyes on the road, nor his hands on the wheel.
Instead, the car’s on-board sensors, cameras and software were controlling functions such as steering, braking and accelerating. The 3km journey, in a Volvo XC90 in November 2015, was one of Australia’s first on-road demonstrations of an automated car, and the auto industry is convinced it is a harbinger of things to come.
Automated or “driverless” vehicles are already used in industry and incorporated into public transport planning. For example, mining giants such as Rio Tinto operate fleets of remote-controlled driverless trucks on outback sites, while the Sydney Metro rail network currently under construction proposes to use driverless trains to shuttle an additional 100,000 commuters across the city every hour. Driverless vehicles are also poised to revolutionise our public roads, boosting motorists’ safety, mobility, productivity and sanity. The most optimistic predictions say all vehicles on Australian roads will be automated by 2030. But are we ready for them?
Automated vehicles still have a whiff of science fiction about them, but peek inside any suburban garage and you’ll see they’re nothing new. From cruise control and reversing cameras through to automatic transmission and braking systems, the most unremarkable of vehicles already possess a range of active driver assist technologies.
According to the Australian Driverless Vehicle Initiative (ADVI), we’re now at the “partial automation” stage of the driverless journey where, despite such whiz-bang wonders, drivers’ hands must remain on the wheel. Next comes “conditional automation”, where all aspects of driving are automated, but drivers must be prepared to resume control when prompted. Following
this is “high automation” and, finally, “full automation”, where steering wheels and driver’s seats can be tossed on the scrap heap.
By curbing the potential for human error, road safety benefits are expected to flow from the introduction of automated vehicles. Anyone who’s ever been rear-ended at traffic lights knows that human drivers are far from perfect – prone to driving while fatigued, intoxicated, emotional or distracted. Last year’s Driving Into the Future report published by Clayton Utz notes that more than 90 per cent of road accidents are caused by human error.
These results are borne out in our annual road toll. Bureau of Infrastructure, Transport and Regional Economics (BITRE) figures revealed that there were 1,300 road deaths in 2016, a 7.9 per cent increase on the previous year. Serious injuries caused by road crashes also appear to be increasing, Australian Institute of Health and Welfare statistics show.
Commuters routinely stuck in peak hour gridlock will also appreciate driverless vehicles’ promise of reduced congestion. The cost of congestion is estimated to increase from $6.1 billion in 2015 to more than $12.6 billion in 2030 in Sydney, according to NRMA’s Accelerating Our Smart Transport Future report. NRMA spokesman Peter Khoury says advanced technologies that will allow autonomous vehicles to communicate with elements of the road infrastructure, and with each other, will help smooth current traffic woes.
Fully autonomous vehicles are flagged to reduce car park snafus too. CISCO’s 2014 Smart+Connected City Parking report estimated that 30 per cent of all traffic congestion in urban areas is caused by drivers looking for available car parking spaces. Khoury points out that autonomous vehicles, which drop off passengers and then proceed to remote parking stations outside the CBD until required to make a pick up, will free up publicly available car parks. This will also bring increased mobility to people unable to drive, such as the elderly, children, or those with disabilities.
Frustrated motorists seem open to these changes. When the University of Michigan’s Transportation Research Institute asked, “What is your general opinion regarding autonomous and self-driving vehicles?”, 61.9 per cent of respondents held positive views. Just over one quarter (26.7 per cent) were neutral, while only 11.3 per cent responded in the negative.
Despite such optimism, Associate Professor Jeremy Woolley, Director of the Centre for Automotive Safety Research at the University of Adelaide, advises us to keep cool heads regarding a “utopian” driverless future. “Taking it all back a notch to reality, we’re a long way off total automation,” he explains. “We’re likely to see automated vehicles in very controlled environments, such as airports and car parks, but when they move out onto public roads, all becomes a lot more complex.”
Legislative shifts ahead
Technology, and people’s willingness to embrace it, is advancing faster than lawmakers can respond. Panellists on an Automated Autos session at this year’s World Science Festival pondered myriad questions. What does “a person in control of driving a vehicle” mean in the context of a vehicle that has no driver? Who is liable in the event of an accident – the manufacturer, the owner, or the software designer? Does Australian Consumer Law apply to a defective driverless vehicle in the event of an accident?
The National Transport Commission (NTC) last year identified 716 laws, rules, and regulations ripe for overhaul.
These were drawn from international law (e.g. the 1968 Vienna Convention on Road Traffic), national law (e.g. Australian Road Rules) and state and territory legislation (e.g. passenger transport, tolling, and criminal codes).
In its related Regulatory Reforms for Automated Vehicles policy paper, the NTC said Australia risked lagging behind the rest of the world if it failed to make recommended regulatory changes. Falling out of step with international standards and conventions, the authors warned, could be a “significant barrier” to introducing automated vehicles in Australia.
According to Owen Hayford, a partner in Clayton Utz, and co-author of the firm’s Driving Into the Future report, a good starting point would be to streamline the process for applying for permits to test automated vehicles on public roads without a human driver.
“The current process involves applying for ad-hoc exemptions from state and territory road agencies,” he explains. “Each road agency presently responds to these applications differently and imposes different conditions. Unless the processes in each state and territory are aligned, it is likely that inconsistent approaches will deter international manufacturers from testing automated vehicles in Australia.”
Concepts underpinning Australian road and vehicle regulation must also change. “Most of our laws are based on the assumption that a human being is at the helm of the vehicle,” says Michael Tooma, Asia Pacific lead partner of Clyde & Co’s Regulation and Investigation practice. Progression to highly and fully automated vehicles will disrupt this definition of “driver”. Other concepts such as “control”, currently interpreted as having one hand on the steering wheel, also require revision.
The NTC targeted 143 potential changes to vehicle compliance requirements. For example, Australian Design Rules specifying the inclusion of features such as steering wheel, speedometers, rear view mirrors and brake pedals, may become redundant as fully automated vehicles roll out. More than 50 rules governing licences and permits, 80 governing interactions with enforcement officers such as police, and 25 governing interaction with other road users were also flagged for revision.
“We’re likely to see automated vehicles in very controlled environments, such as airports and carparks, but when they move out onto public roads, all becomes a lot more complex.”
ASSOCIATE PROFESSOR JEREMY WOOLLEY, University of Adelaide
Playing the blame game
Liability for crashes will be one of the biggest issues. Despite anticipated safety benefits, Hayford notes, “Accidents will continue to occur, and those who suffer loss will need to be compensated.” Should liability for automated vehicle accidents be borne by the owner? The operator at the time? Or the corporation that manufactures, supplies, installs or maintains the vehicle? Hayford suggests the answers to these questions might change as each new stage of automation evolves.
In the meantime, some are asking whether automated vehicles should be segregated from traditionally-driven ones. Tooma refers to the first fatal crash, in Florida in May last year, involving a Tesla Model S operating in autopilot mode. The car was T-boned by an errant tractor trailer which neither the Tesla’s cameras, nor its driver, noticed. Tesla subsequently issued a statement saying that while autopilot was improving, it “still requires the driver to remain alert”.
Such scenarios spark further questions. In an article for Precedent, Hayford asks how much warning automated vehicles will give their human drivers. Must the driver resume control instantly? Will the warning period be sufficient to allow human drivers to safely resume control after engaging in non-driving tasks? (It was claimed that the Tesla Model S driver was watching a Harry Potter movie at the time of the crash.)
Would last-minute human intervention in such emergencies even be enough to escape trouble? “A human driving that vehicle might have swerved to avoid that outcome, or might not have done anything differently,” Tooma says.
Philosophical debates are also stirring. Will automated vehicles be programmed to protect their occupants at all costs? Faced with an unavoidable collision in one of two directions, will algorithmic calculations see them hit the elderly pedestrian, or the mother pushing a pram?
Khoury says the brave new world of automated vehicles will reshape whole industries, particularly transport, to an extent not seen since the days of horse and cart. Smarter technologies will raise fresh privacy and data ownership concerns. “We need to get our heads around it, and we should not be afraid of having the social and moral discussions now,” Khoury says. “We might not have all the answers, but we should at least be starting to ask the questions.”