This chapter discusses the role of governments in Australia in preparing for the social issues surrounding driverless vehicles, aside from those discussed in previous chapters. The focus of this chapter is on the importance of consistency Australia-wide: evidence received by the Committee broadly agreed that the successful introduction of autonomous vehicles into Australia relies on a consistent approach.
As noted throughout this report, one of the key points made by a wide range of submitters and witnesses to this inquiry was that Australia needs to adopt a nationally consistent approach to driverless vehicles. This applies to infrastructure standards and to regulatory approaches.
Toyota Australia, for instance, made this point, noting that Australia’s relatively small population size makes it more important that manufacturers see the entire country as one market with nationally consistent infrastructure, rules and guidelines:
… in order to get the best outcomes from driverless vehicles and the related technologies, we see the need for a consistent approach across Australia, across all the states and territories. Again, it is very positive. The publication of the national trial guidelines, which I think are expected out in May this year, will create a good foundation. This single national guideline, we hope, will then create, as organisations wish to undertake trials throughout Australia, some consistency in the application of the approach.
In November 2015, the National Transport Commission (NTC) was given a central role in coordinating Australia’s regulatory approach to autonomous vehicles, as the NTC’s Chief Operating Officer Dr Geoff Allan explained to the Committee:
When the Transport and Infrastructure Council, which is a COAG committee of transport ministers from around Australia, met in late 2015, they stated that they wanted Australia to work towards harmonised standards and regulations in relation to automated vehicles to ensure that Australia was well positioned to adopt new technologies. That is a nationally consistent approach, so we position ourselves well in the international market. To achieve that, ministers asked the NTC to prepare for more autonomous vehicles by identifying regulatory and operational barriers… When we reported back to ministers in November 2016 ministers endorsed a plan to remove some of those barriers to allow for the uptake of automated vehicles.
The NTC’s Mr Marcus Burke further emphasised that the Australian approach to the regulation of autonomous vehicles is a national one, and is based on concerns raised by industry:
We are aiming to develop an approach which is national. We do not want to have a situation where vehicles cannot cross state borders due to different regulations in different parts of the country, and that is something we have heard very clearly from industry. We want to support innovation and support the safe deployment of this new technology in order to in particular reduce the current death toll and rate of serious injuries on our roads. We heard very clearly from industry on the need for consistency both nationally and with emerging international standards, given that Australia is a relatively small market for vehicles.
Infrastructure needs and readiness
This section discusses the infrastructure needs of autonomous vehicles and the different levels of infrastructure these vehicles require to existing driver-controlled vehicles.
The Department of Infrastructure and Regional Development (DIRD) highlighted that, currently, the infrastructure needs of autonomous vehicles are unclear:
Automated vehicles may also require the deployment of new kinds of digital infrastructure (e.g. communications infrastructure, accurate satellite positioning) or may require aspects of physical infrastructure to be designed and maintained to a particular standard (e.g. road signage, line markings, road geometry). At the current time, given the early stage of automated vehicle development and trials, there is significant uncertainty about what the exact future requirements might be.
An additional challenge is that not all developers of automated vehicles will use the same enabling technologies, meaning that it is possible that different vehicles could have different technical requirements. There are also different approaches to the design and maintenance of physical infrastructure across Australia, which in the past has presented a barrier to the deployment of technologies such as automatic speed zone recognition.
The Royal Automobile Club of Western Australia (RAC WA) made a similar point, noting that planning to date has not sufficiently recognised that traffic patterns are likely to change with the introduction of autonomous vehicles:
… much of the longer-term planning for road infrastructure requirements is informed by transport models which do not take account of the implications of AVs on travel demand and behaviours. This situation is unlikely to be unique to WA, and further research is needed to better understand these implications.
Mr Alex Foulds of DIRD also noted that infrastructure is a major investment for governments (at all levels) – not just in terms of money but of timeframes of infrastructure projects:
I think it is important that the investment also is guided by real-world experience. For example, if you build a road, it takes four years to build—it is with you for 60 years, maybe more. If you build a car and design a car from scratch, it takes two to three years—and it is with you for maybe eight to 10, on the whole. Software—overnight. They are the clock speeds that sit around the way things are developed. So you do not want to commit to a 60-year long piece of—potentially stranded—infrastructure until you know that that is actually what you need, and that it is fit for growth, and that it can accept change.
One of the strongest points made by witnesses and submitters to this inquiry was the importance of consistent road infrastructure across Australia. For autonomous vehicles to operate successfully across the country, infrastructure – in particular, road markings and signs – will need to be consistent nationwide.
Mr Ashley Wells of the Federal Chamber of Automotive Industries reflected on this, noting that Australia’s federal system may pose some problems for automated vehicle sensors:
There are apparently global standards for road signs. Australia has taken a slightly different path, and we do have differences across the states and territories relating to those as well. The federal highway system is somewhat different, but when you get into individual states there can be some differences. So the in-car cameras, and the technology that is going into those, are obviously not sophisticated enough to pick up the nuances that come with the various state and territory changes.
Austroads, the association of road transport agencies across Australia (and New Zealand) emphasised that consistency of approach is central to the work of road agencies:
As has been highlighted in our submission and in the submissions from other stakeholders at these hearings, it is extremely important that the regulatory and operational frameworks to support the deployment of automated vehicles are nationally consistent. In an effort to achieve this Austroads works very closely with the Commonwealth Department of Infrastructure and Regional Development, the NTC and its road agency members in each of the jurisdictions.
Austroads also has an industry reference group whose role is to ensure that industry is engaged in the process of preparing road infrastructure for the arrival of autonomous vehicles.
The Committee heard that a significant issue with Australia’s readiness for autonomous vehicles was that the quality of roads in Australia is variable. This may cause substantial problems – particularly in regional areas – for the functioning of driverless vehicles.
As DIRD noted, infrastructure in regional Australia is not consistently at the level required to support autonomous vehicles:
A significant challenge to the deployment of automated vehicles in regional areas is the provision of supporting infrastructure. Depending on how technology develops, this could include requirements for both physical infrastructure (e.g. sealed roads, signage, road markings) and digital infrastructure (e.g. mapping data or communications infrastructure). Improving infrastructure in regional Australia would require a concerted effort by all levels of government.
Austroads highlighted the scale of the issue – Australia has approximately 900, 000 kilometres of roads, approximately 85 per cent are the responsibility of local councils. Over half of Australia’s road kilometres are unsealed dirt roads, which will present significant issues for automated vehicles in their current form.
Austroads was also able to point to the specific areas of concern raised by the vehicle manufacturers:
With regard to physical infrastructure, with the consultations that we did, particularly with car companies and tech companies, the barriers that they are seeing at the moment—some of them are having issues with lines, where they might be non-existent, inconsistent or deteriorated. They are certainly having issues with inconsistency with road signs, and speed signs in particular. One issue that has been highlighted during our consultations is dynamic electronics signs, LED signs, in that they do not necessarily work at the same frequency rate as the signs in Europe. We have car companies that have actually withheld safety applications, because of the infrastructure that we have. The other issue that has come through, particularly from our international colleagues doing trials, is the condition of the pavement itself. The technology is just at a point where it may not be able to handle where there is significant deterioration of the pavement—potholes et cetera. So, we cannot forget the basics. It is not only those attributes that you think that the sensors are reading but also the condition of the road that is affecting how these vehicles operate.
Roadworks and other changes to existing road environments remains one of the major problems for autonomous vehicles to solve, as Mr Ballingal of Austroads further explained:
… in the consultations we did with car companies and supporting tech services, that was highlighted as a key issue. There are basically two ways that a vehicle could identify that it is entering a roadworks area. One is by receiving data—so there could be a map service coming from a cloud service—that is where the permitting comes in; permitting a site to be a work zone. The second is the sensors on the vehicle detecting that it is entering a work zone. That is a real issue at the moment because we do have inconsistencies with our roadwork sites. That is going to be a key area with Austroads moving forward, to try to raise the consistency in the standards in those sites.
While road quality can affect all forms of vehicles, the Committee heard that satellite and internet infrastructure is necessary for autonomous vehicles, and that particularly in regional areas, may require further improvement to support their operation.
In its evidence to the Committee, Austroads noted the importance of this form of infrastructure, in particular ensuring the vehicles have up-to-date information about changes to roads:
With digital infrastructure, data is going to be critical. For the most part, the market will support that. We have mapping data providers that have better data on our roads than road agencies do, but there will be some attributes for which a road operator will still be the authoritative source. It is still within a jurisdiction where the decision is made to close a road, close a lane, give a permit to do roadworks or change a speed zone, and so, somehow, the road agencies have to feed into that supply chain. That is part of the digital infrastructure.
Austroads also emphasised that the digital infrastructure requirements will be dependent on the type of vehicle and its use:
It is not like every vehicle is going to need a cellular connection 100 per cent of the time everywhere. It will depend on how the manufacturers develop that vehicle to operate. We call it an operational design domain, so it is within the boundaries of what it is designed to do. But certainly, moving forward, it is something that as a country we need to give appropriate consideration to. Compared with other developed countries, our geographic coverage of cellular communications is relatively low. It is high by population; it is not high by geography. It could be that if we want to support more automated-use cases, such as truck platooning on rural highways, we will need to give consideration to communications coverage.
The other digital infrastructure that I would highlight is around positioning services, particularly satellite positioning services. On that one, we are working very closely with not just the Commonwealth Department of Infrastructure and Regional Development but also Geoscience Australia.
One resolution to concerns raised regarding digital infrastructure is currently being tested by Geoscience Australia. The Satellite Based Augmentation System, which is a program to overlay information:
An augmentation system is an overlay that works with GPS to improve the accuracy from the standard five to 10 metres that you will get from GPS down to submetre levels of accuracy.
This is relevant because the North Americans, the Japanese and the Europeans—some of the largest car manufacturers—already have augmentation systems in place. The positioning component of the multisensor units that are in vehicles that are being designed now are coming from SBASs, but in Australia we do not have one. We were funded, through MYEFO last year, to undertake a two-year testing program of a satellite based augmentation system across Australia, and then our New Zealand friends subsequently joined in the testing as well, by contributing another $2 million.
We are coordinating that testing through the Cooperative Research Centre for Spatial Information, and right now we are in the process of assessing—we had something like 60 proposals for different testing examples that could be undertaken with the SBAS. We are trying to now cut them down and get groups together to fit within the allocated research budget. Two of those are relevant to this discussion. One of them is around autonomous cars and the other one is around intelligent transport systems in heavy vehicles, both with state government leadership but also industry participation. We think that is going to take us in the right direction as far as transport goes.
The important part about an SBAS is that it is not reliant on a telecommunications network in any shape or form. The corrections actually come from space. So a person sitting out the back of rural Australia, in the Simpson Desert, gets exactly the same access to signals as a person sitting in one of the larger metro areas—in fact, probably improved access than someone who is sitting in an urban canyon for some reason, because of visibility of the satellite itself.
As discussed in chapter 5, the shift to fully driverless vehicles, particularly as part of the larger move towards automation of many tasks, will cause significant changes in the current employment model. In order to minimise the negative effects of that change, all levels of government, alongside industry, will need to prepare.
Mr Harrison of the Council of Capital City Lord Mayors made the point that the transition is not far off and will happen quickly:
When you go back to the early days with settlers, there were whipmakers, groomsmen that would look after horses and things like that. Those jobs all disappeared, but with the new technology came new jobs. But there was a long transition. It was 26-odd years before we started seeing a lot of vehicles on the road. That is going to happen much more quickly this time. From a social perspective, for those people who may lose their jobs due to autonomous vehicles there will be new jobs that will come but it is incumbent upon the three levels of government to work with the universities and our TAFEs to identify those new jobs and make sure that we are reskilling people to take on those new jobs and that people do not fall through the gaps with these changes that are inevitably going to come.
Some witnesses identified that Australia has traditionally had problems retaining trained and skilled young workers in the science, engineering and information technology fields, but that Australian universities have recently had a strong focus on autonomous technologies, including vehicles.
The Department of Industry, Innovation and Science noted the importance of planning for this workforce change, and highlighted the Commonwealth Government’s National Innovation and Science and Agenda to encourage increased levels of STEM literacy amongst Australians:
Part of the government’s role in facilitating a move to new and growing industries is enabling workers to develop the right skills for the jobs of the future.
Australia’s workforce will need to be equipped with the right skills and training – from basic digital literacy to complex science, technology, engineering and mathematics (STEM) expertise – to be able to undertake the jobs created around autonomous vehicle development and use. For example, a very large quantity of data will be generated by connected and autonomous vehicles. Different skillsets will be required to build and maintain the systems, servers and processors to protect this asset, as well as manage, analyse, and make sense of the data.
The Committee recognises the importance of a nationally consistent approach to the preparation for and introduction and regulation of driverless vehicles in Australia and was pleased to hear that a consistent approach is currently being taken, with cooperation between the Commonwealth and state and territory governments via the Transport and Infrastructure Council and with a coordinated approach being led by the Department of Infrastructure and Regional Development, the National Transport Commission and Austroads.
The Committee recommends that the Commonwealth Government, in consultation with state and territory governments, continues to coordinate their approach to automated vehicles, ensuring consistent regulations and policy settings.
Noting that the successful deployment of autonomous vehicles in Australia will require further standardisation of road infrastructure, the Committee is of the view that the Commonwealth Government, through the Department of Infrastructure and Regional Development, should coordinate a project to standardise road infrastructure in Australia.
The Committee also notes that the quality of roads in Australia is variable and the issue of road quality, particularly in regional Australia, must be considered in preparing for increasingly autonomous vehicles.
Similarly, the Committee recognises the important work being conducted by Geoscience Australia on digital infrastructure required for autonomous vehicles and encourages that work’s continuation.
The Committee recommends that the Commonwealth Government coordinates efforts to standardise road infrastructure in Australia, particularly as it relates to signs and road markings, and that the Commonwealth Government considers ways to ensure that the benefits of automated vehicles are available across Australia, including in regional Australia.
Notwithstanding the work being undertaken already to prepare the way for driverless land-based vehicles in Australia, the Committee’s inquiry has highlighted that there will be many and wide-ranging social impacts once driverless vehicles become available in Australia, and it is important that those are adequately prepared for.
The Committee recommends that the Commonwealth Government consider the merits of establishing either a dedicated national body or a cross-agency taskforce, in conjunction with state and territory jurisdictions and working with vehicle and software manufacturers, to coordinate Australia’s preparation for the introduction of land-based automated vehicles. This body would have regard to topics including, but not limited to:
Methods of public engagement to ensure that concerns about automated vehicles are addressed and benefits are explained
The employment ramifications, both direct and indirect, of automated vehicles
How to best ensure that people with disability and older Australians are able to benefit from automated vehicle technology
How to best ensure that people in regional and rural Australia can access the benefits of automated vehicles
The infrastructure needs, both physical and digital, of automated vehicles and the role of governments in ensuring that those standards are met, particularly in regional and rural areas of Australia
The ownership, use and security frameworks applicable to the data generated by automated vehicles
Legal liability and insurance implications of automated vehicles.
Ms Michelle Landry MP
30 August 2017