Australia and New Zealand Driverless Vehicle Initiative Summit
16 November 2017
Adelaide Convention Centre
I am very pleased to have the opportunity to speak at this International Driverless Vehicles Summit, convened by the Australia and New Zealand Driverless Vehicles Initiative.
There are few issues of greater importance in transport policy than the transition to automated or driverless vehicles.
Today I want to speak about the appropriate public policy response in Australia to this coming transition.
No two countries will approach this issue the same way—and in the first part of my remarks I will look at the Australian context for the arrival of driverless vehicles.
Next, I want to discuss the process by which consumers, in Australia and globally, are likely to take up driverless vehicles—remembering that the take up of a technology is quite a different issue to the development and introduction of that technology.
In the last part of my remarks I will turn to what I see as the public policy priorities—for the Australian Government as well as state and territory governments and indeed local government—in preparing for and responding to this transition.
The Australian Context for the Arrival of Driverless Vehicles
Let me turn firstly then to the Australian context for the arrival of driverless vehicles.
To start with, Australia is one fairly small market among many in the world—with about 1.2 million new cars sold each year, of about 95 million sold globally. That means we are going to be a standards taker not a standards maker.
That is quite normal. You can look at mobile phones, or digital televisions, or a whole host of other technologies, including of course today's motor vehicles, and we see the same principle—Australia is an adopter of global standards rather than a country which seeks to impose its own standards. This approach makes good sense for a host of reasons, including the fact that Australian consumers get the benefit of pricing based on global economies of scale.
Yet while we are a small market, we are nevertheless quite an attractive one. This is shown by the very wide range of makes and models of motor car on sale in Australia today, with 67 brands and 350 models available.
We are heavy users of cars and transport compared to many other countries: with 731 vehicles per thousand people, we rank seventh in the world, and the number of kilometres driven per adult ranks us above major economies such as Germany, France and the UK.
With more than 2 million Australians commuting for more than 45 minutes each day, the opportunity to convert time you presently use sitting behind a wheel to other uses is going to look pretty good to a lot of us.
And of course there is the proven propensity of Australians to be early adopters of technology.
All of these factors suggest that we could well be a market where there is quite strong take up once driverless vehicles are available to the consumer.
The Likely Take Up of Driverless Vehicles
Let me turn then from the Australian context to a fundamental issue which is very much a global one—how quickly are we going to see a take up of driverless vehicles. The answer to this question will be fundamental in determining our policy responses—but unfortunately nobody knows what the answer is!!
At one end of the spectrum, recent modelling commissioned by the Queensland Department of Transport and Main Roads, suggests that saturation of highly automated vehicles in the Australian fleet could occur roughly between 2050 and 2060.1 At the other end of the spectrum, last year Telstra Chief Technology Officer caused a stir with his prediction that by 2030 all cars could be driverless.2
If we start to unpack this issue a little more, the first question is, how quickly does the fundamental technology get to the required level so that a vehicle can safely operate in Level 5 automation—on any kind of road or weather, and with lots of other vehicles on those roads still being driven by pesky, unreliable humans governed by infuriatingly unpredictable algorithms.
Every week we see claims from vendors that we are only a few short years from that point.
Jim McBride of Ford said last year that the underlying technologies were four to five years away.3
Market researcher David Galland wrote in Forbes earlier this year that he predicted ten million driverless cars on the road by 2020.4
BMW, Honda, Ford, General Motors, Toyota, Volvo, Renault-Nissan,Daimler and Tesla are all predicting they will have driverless vehicles on the market in five years.
I know ADVI's view is that that Level 5 full automation will happen in the period 2026–2030.
Once the technology is ready to take the market, the next question is, what is the business model used to deploy the technology? A lot of the popular discussion of driverless cars assumes that the business model will be exactly the same as with today's cars: the consumer buys a car and owns it for a number of years; the cost is sufficiently high that it is typically the second most valuable asset the consumer owns; and it is used for one to two hours a day, sitting parked on the street or in a garage for the rest of the time.
I think it is instructive to look at the way that mobile phones were taken up in Australia and around the world. In June 1996 there were 2.6 million mobile services.5 Today there are well over 30 million mobile services in operation6—a ten-fold increase in twenty years.
Business model innovation played a critical role in this rapid take up. Selling phones on a two or three year contract allowed the consumer to purchase a phone worth a thousand dollars or more with no money upfront. Offering prepaid services was critical to getting lower spending consumers to take up mobile phones.
So when I see commentators argue that we do not yet know what the winning business model will be for driverless vehicles—and it will be a key determinant of take-up—that makes a lot of sense to me.
Let me venture cautiously into this space with a few observations. Firstly, closed or restricted locations like campuses seem to offer obvious opportunities for the early introduction of driverless vehicles.
Secondly, we are already seeing, and will see much more of, driverless vehicles in such closed settings, moving goods not people. Some examples include Rio Tinto using driverless trucks on remote mine sites; and the ‘autostrad’ carriers used at the Patricks terminal in Port Botany to move containers around.
When I visited Heathrow Airport last year, I was told that they plan to have all airside vehicles driverless by 2020.
Thirdly, it looks likely that driverless vehicles will break down today's rigid demarcations between private and public transport, and rail and road transport. In a future world we could have driverless vehicles operating over dedicated high speed corridors for part of a journey, probably in convoy like a train without rails, and then separating out to complete individual journeys on suburban roads.
That links to a fourth prediction which seems to be pretty widely accepted: in the future we will not necessarily own a car. Instead more and more of us will choose to dial up a vehicle on a smartphone whenever we need it; it will take us where we want to go; and then it will go on to serve the next customer.
Of course this model has been around a very long time—in the form of taxis, hire cars and more recently ride sharing services.
3.8 million Australians use taxi services; and of them 54 per cent now use ride sharing services such as Uber, GoGet and GoCatch; and 28 per cent prefer ride sharing services almost exclusively.7
Pulling all this together I think we can have a high level of confidence that automated vehicles will become increasingly important in specific settings such as campuses, or in specific applications such as mining and industrial settings.
Whether they become a mass market consumer phenomenon rapidly— once the technology is robust—remains a more complex question, and a lot is going to depend on what consumers decide to do.
A big factor here will be cost. Consumers might love the idea of a driverless car—but if costs them two or three times as much as a Toyota Corolla or a Kia Rio then we are not likely to see mass market take up.
A second, related, factor is that because cars are such an expensive purchase, people typically hold them for several years.
So even if you love the idea of a driverless car, you are unlikely to go out and buy it while your current car—which you paid a lot of money for only a few years ago—still has a fair bit of life left in it.
The third potential barrier is safety. There is every reason to believe driverless cars will be substantially safer than human driven ones—but we humans don't always think rationally about risks, and it is not hard to see how a couple of well-publicised crashes involving driverless cars could undermine public confidence.
Public Policy Priorities
I have spoken at some length about factors that will determine the take up of automated vehicles—and my key point is that most of those factors are out of the hands of government. Let me turn then to the final area I want to discuss: what should be the policy priorities for Australian governments, federal and state and territory, when it comes to autonomous vehicles.
The first of four priority areas I want to highlight is ensuring that, as a nation, we are well positioned to take up autonomous vehicles once they are in the market and there is consumer demand for them.
From the government perspective, this means encouraging tests and trials, having appropriate safety protections in place for commercial deployments, and resolving insurance and other legal issues.
Australia is making solid progess in these areas.
Earlier this year all ministers around the country adopted national guidelines for trials. This means that if a trial requires vehicle needs exemptions from current laws, we can provide this in a safe and nationally consistent fashion.
Just last week at a meeting of federal, state and territory transport Ministers we agreed that Australia will develop a flexible safety assurance system for automated vehicles. This is designed to allow industry to demonstrate that vehicles are safe while international standards are still being developed. We aim to have this in place by 2020.
We also have extensive work underway on insurance and other legal issues. Ministers last week also agreed to change the way that the road rules are enforced, so that the human drivers won't have to have one hand on the wheel while using specific automated systems, such as self-parking. More work is planned or underway on issues including access to data, the legal definition of ‘driver’ and compulsory third party insurance.
The second area in my view is actions to deliver the required communications infrastructure.
In January the Australian Government announced a $12 million funding commitment to test the next generation of satellite positioning technology, known as a Satellite-Based Augmentation System or SBAS.
Current satellite positioning technology has an accuracy of around 5 to 10 metres, which can tell us which road a vehicle is on, but not which lane it is in. SBAS can provide accuracy of better than 10 centimetres, which would provide the lane level accuracy that will be needed by a range of connected and automated vehicle applications. More accurate positioning data would also benefit other sectors such as mining, agriculture and aviation.
This initiative launched officially last week and will include specific projects on connected and automated vehicles. The outcomes from the test-bed will inform Government decision making in this area.
We are also working to make sure that Australia is well positioned for the introduction of the next generation of mobile telecommunications technology, 5G.
The features of 5G will include improved connectivity, greater network speeds and bandwidth and, critically, very low latency. Unlike current networks, this means that 5G could be used for safety critical communications in connected and automated vehicles, such as real-time collision warnings.
The Turnbull Government's work in this area includes making spectrum available in a timely manner; actively engaging in the international standardisation process; streamlining arrangements to allow mobile carriers to deploy infrastructure more quickly, and reviewing existing telecommunications regulatory arrangements to ensure they are fit-for-purpose.
The Government is strongly engaged with industry on the best way to capitalise on opportunities and emerging issues on 5G.
The third priority is policies to encourage and support Australian businesses in this space—so we are not just consumers of this technology but we have an involvement in producing it.
There are certainly some impressive companies and organisations at work in this space.
I recently had the opportunity to visit Cohda Wireless, a firm based here in Adelaide that specialises in vehicle communication technology, which is important for autonomous vehicles.
Another impressive business is Baraja, based in West Linfield in my electorate. They recently secured a $1 million Accelerating Commercialisation grant in order to assist in developing an improved LIDAR sensing device. This is a critical component in automated vehicles.
The Turnbull Government, through the Department of Industry, Innovation and Science, has also committed $55 million over 10 years to the iMove Cooperative Research Centre, which will explore digital and evolving vehicle technologies. This to be matched by $178.8 million in cash and in-kind participant contributions from industry and academic stakeholders.
This is exactly the transition from traditional manufacturing to high skilled, high value-adding advanced manufacturing at the centre of the Turnbull Government's industry agenda.
I want to conclude by discussing the impact of these changes on our road infrastructure—including how we fund that infrastructure.
This is an issue which brings together all of the factors I've mentioned today—consumer uptake, deployment models and uncertainty about how technology is evolving. Together, these uncertainties make it difficult for government to plan for the future in any comprehensive way, at least at this early stage.
For example, will automated vehicles use existing parking facilities, or will most vehicles be automated taxis that require more drop-off zones?
How will the convenience of automated transport change travel patterns? Should we be planning to avoid increased urban sprawl?
Do we need to provide roads with high quality line markings, or will vehicles rely on 3D maps instead?
There are however some trends that are becoming clearer.
More and more vehicles are likely to use electric motors rather than internal combustion engines. This trend is already underway, and is likely to be reinforced if driverless vehicle technology is taken up. As I have already argued, driverless vehicle technology could well lead to lower levels of household vehicle ownership.
The combined effect of these trends means that the revenue from fuel excise and car registration is unlikely to keep pace with the amount we spend as a nation on operating, maintaining and investing in roads.
Today that figure stands at about $25 billion a year—and it looks likely to increase.
So part of our policy response to the likely increasing role of electric vehicles and driverless vehicles is to consider how we can continue to fund our road system.
Last year the Turnbull Government said that we intended to establish a study, to be led by an eminent Australian, into the way Australia funds and pays for its road networks. I will have more to say about this study, and who will lead it, in coming months.
Let me conclude then by returning to my claim that Australia is likely to be an attractive market for vendors of driverless or autonomous vehicles because of factors such as our high rate of vehicle ownership, the relatively long distances we drive both in our low population density cities and across the vast scale of our regional and remote areas, and our proven appetite to take up new technologies.
But there is much that remains uncertain—including how quickly the technology will be ready to come to market and how quickly consumers will choose to take it up.
Much of this is outside of governments' hands—but on the other hand there are clearly policy areas where governments—and only governments— can take action to prepare for this tremendous change.
That is why the Turnbull Government, working with state and territory governments, has a clear agenda in this space—so that Australians can benefit from the profound changes that driverless and autonomous vehicles are going to bring.
1 TransPosition. (2016), Conceptual sensitivity modelling and analysis on the introduction of autonomous vehicles, prepared for Department of Transport and Main Roads, unpublished, cited with permission.
5 AMTA, ‘Ten Years of GSM in Australia’