The autonomous vehicle sector has made significant strides forward over the last few years, but exactly how far have we come, and what is a realistic timeframe until driverless cars become a commonplace sight on our roads?
John Paddington provides some welcome insight
When I was approached to write this article to give my opinion on where we are with driverless vehicles, it reminded me that when I talk to the public, friends or family about my work with driverless cars, I often get asked the question - “When do you think I’ll be able to use/have one?”
On the face of it this is simple but it can have a much more complicated answer. So, I decided to use my thoughts on the answer to frame this article.
The first thing to consider when answering this question is what do people want? I often think when talking to people outside of the industry about driverless vehicles, that they specifically mean driverless cars like something out of TV and Film, a go-anywhere, go-anytime vehicle, where the vehicle occupants will not need to think about driving at all and be free to do whatever other activities they want in the meantime.
If you’ve done any work, been to an event or read anything about driverless vehicles, you will most likely have heard of the Society of American Engineers Levels of Automation, where Level 5 is the “holy grail” and corresponds to this go-anywhere, anytime vehicle. In my opinion, this is also what is heavily talked about in the media and by people in the industry. Over the last few years, there’s been bold claims about Level 5 vehicles being available soon, particularly by major players. However, I feel very safe in saying that no one has a Level 5 vehicle currently and it will probably take longer to get there than a lot of people suggest.
Just because there is not a Level 5 vehicle currently available, does not mean there is not a lot of research and development taking place in the world of driverless vehicles. It would also be a mistake to assume that research into driverless vehicles is a recent thing. The UK Transport Research Laboratory (TRL) were testing driverless vehicles in the 1960s/70s, although limited to driving along routes with magnetic tracks. There was also plenty of research in Europe in the ‘90s and early ‘00s. What has changed, though, is the now widespread level of investment and commitment into driverless vehicles, brought about by advancements in technology, particularly in machine learning and artificial intelligence (AI). You will struggle to find a major automotive company or technology company not thinking about driverless vehicles or automation in some way, either through in-house efforts or partnering with universities or start-ups.
In the UK, the Government has rightly decided to encourage research and development, which forms a key part of the industrial strategy.
There are two key drivers in doing this:
- Encourage hi-tech jobs and growth within the UK.
- Ensure that any solutions developed are UK-centric and are not just imported from elsewhere.
This dual focus is shown in the funding of the UK Government Centre for Connected and Autonomous Vehicles (CCAV), which is funded both by the UK Department for Transport and the Department for Business, Energy and Industrial Strategy. CCAV has made £150 million in match funding available over the last few years for collaborative research and development projects through the funding body Innovate UK.
Through my previous employers, Conigital, I was actively involved in three CCAV/Innovate UK-funded projects, Capri, Insight and Synergy. These are centred around developing autonomous shuttle services. My experience from these projects and seeing outputs from other projects in the UK and in Europe is that the focus of the research and development in these projects is on getting the technology right to allow the vehicles to move on their own. A lot of work is going on into machine vision, sensing and AI, as well as related aspects such as platooning and traffic signal communications.
What has emerged in the last few years is that it is not practical to conduct the volume of real-life trials to prove the technology works efficiently, safely and reliably, particularly, if you want to be able to remove a safety driver from the vehicle. Therefore, elements of simulation will be required. Simulation allows large volumes of tests to be undertaken and to consider scenarios which are dangerous or rare in real-life. What is interesting is that simulation is starting to happen at all levels of the technology, not just at the overall vehicle but for sub-systems and even inputs into sensors.
“In the UK, the Government has rightly decided to encourage research and development, which forms a key part of the industrial strategy”
This focus on technology is important but I do wonder if there is not enough work being done engaging with the public to determine the real use cases for vehicles. For me, driverless vehicles will only provide significant benefits to cities and regions when they are used in public transport/ride-sharing configurations linked to Mobility as a Service (MaaS) applications. This is not to say that driverless cars can’t offer benefits, particularly around improved safety and reduced congestion, but if they are primarily used as single occupancy vehicles there will be missed opportunities. There could be adverse effects, for instance increasing vehicle miles travelled and carbon emissions.
The journey experience is vital and the human factor element of autonomous vehicles must not be overlooked. Journey touch points such as finding the right vehicle, opening the doors and telling the vehicle to move away could be significant pain points if not well thought out and solutions offered that work for everyone in society.
Saying this there are good examples of public engagement out there. The Capri project undertook face to face co-design work with 60 members of the public to get their views on all aspects of an autonomous vehicle service.
The Flourish project looked at mobility requirements for older people. Recently, the Coventry-based autonomous pod manufacturer Aurrigo started a six-month trial with the charity Blind Veterans UK in Brighton to evaluate how autonomous vehicles can help the blind and partially sighted. The latest CAV3 projects funded by Innovate UK also have a stronger public engagement aspect and I’m looking forward to seeing the results of these projects.
“Driverless vehicles will only provide significant benefits to cities and regions when they are used in public transport/ride-sharing configurations linked to Mobility as a Service applications”
Safety is another key point for public acceptance of autonomous vehicles. Often in the media, this is associated with ethics and is often talked about in conjunction with the trolley problem. Simply put, the trolley problem is if you can’t avoid a fatal accident but have a choice about what actions you take, what do you do? This, in part, is due to the excellent research carried out by MIT on the Moral Machine (http://moralmachine.mit.edu/).
If you’ve not done this test, it’s a useful self-reflection exercise. However, I think the trolley problem is often overstated and I don’t believe driverless cars will have the time, intelligence or sensing capability to really make these kinds of decisions anytime soon.
There’s a whole raft of softer ethical issues around driverless vehicles that are more interesting, particularly for rural areas. For instance, in what circumstances should autonomous vehicles let vehicles move in from side roads when there’s queues of traffic? This also links to legal and regulatory considerations. For instance, it’s common for drivers in the UK to mount the pavement to make space for an emergency vehicle, although this is strictly illegal. These kinds of issues will need to be resolved and to the credit of the Law Society in the UK, they have been considering issues already, as shown in their recent consultation on Autonomous Vehicles, which is well worth a read. (https://www.lawcom.gov.uk/project/automated-vehicles/).
As well as legislation regarding the use of autonomous vehicles, there will be a whole raft of regulations on a range of topics to be considered, for instance what is the best class of vehicle to represent autonomous shuttles, should regulation be like that of existing taxis or buses or something else. Again, these are issues being explored in current projects but need resolution. There are also decisions to be made about the role of government and whether it takes a strong-role in shaping how vehicles/services operate or whether it lets industry take the lead. I expect this to vary between countries and potentially between cities/regions.
I would say most autonomous vehicles on the roads today either fit into the category of research and development or early trials with clients. The move to commercialisation will be a challenge for all. This is a challenge for start-ups in the industry who need to show consistent growth in revenues over time to keep attracting funding required at each stage to grow their businesses.
Cost will play a huge part in commercialisation, both for consumer and commercial vehicles. There will be early adopters willing to pay a premium to trial the latest technologies, but both up-front and operating costs will need to fall if self-driving vehicles are to become widespread. This is particularly important in the public transport industry where operating margins are low. I also expect that as autonomous vehicle technology is proven, we will see larger capacity vehicles, particularly shuttles/buses becoming available as they can be more cost effective.
Linked to this there is also a question of whether we will see true level 5 vehicles, as there could be significant costs in covering all the edge cases for Level 5 vehicles whereas a Level 4 vehicle may well be enough if used only in geofenced areas as a public-transport/on-demand type service. Level 5 vehicles are more likely to be needed for passenger cars where there is more expectation of using the vehicles all the time.
A challenge for the UK and other governments will be to consider how to support business(es) to commercialise the good work undertaken in the research and development projects. Funding will likely be needed but also creative approaches to help provide support to end-clients to get initial commercial projects established.
A similar question will apply to local governments, at the town, city and regional levels, who will need to determine the best way to encourage and support commercial rollout of vehicles. Particularly, around decisions on infrastructure which may have long lifespans (often 20-30 years). This is something that I am particularly interested in my new role. Here, in the West Midlands, testbeds have been established to help work with private enterprise to determine the most effective infrastructure.
“A challenge for governments will be to consider how to support business(es) to commercialise the good work undertaken in the research and development projects”
As you can see, it is difficult to answer the question “when can we expect to see a driverless car like in the films” as there are many factors in play but I think it is a matter of time until we see them, not if we will see them. What I do think is that the first truly driverless vehicles will not be quite as imagined in the films but more limited to driverless buses or shuttles operating in controlled environments, such as airports and hospitals. Later, we’ll see fully driverless cars and after that, who knows? It wouldn’t surprise me to see all sorts of different use cases, such as driverless hotel rooms, meetings rooms and shops.
FYI .... .......
John Paddington is Innovation Integration Lead (Public Sector) at Transport for West Midlands, UK and a vice chair of the ITS (UK) Mobility as a Service Interest Group.