Episode 6 - The Future of Motoring
What will motoring be in 25 years?
How to listen
In this podcast we ask the experts what motoring will be in 25 years with the introduction of autonomous and electric vehicles. They share the current projects happening across the world to make self-driving vehicles a reality.
- Dr Rebecca Michael, Head of Public Policy at RACQ
- Professor Paul Salmon from the Centre for Human Factors and Sociotechnical Systems at the University of the Sunshine Coast
- Professor Andry Rakotonirainy from CARRS-Q at QUT
You're listening to RACQ Living.
Anthony Frangi: Hello and welcome to the RACQ Living podcast. I'm Anthony Frangi. When the Ford Motor Company launched the Model T Ford back in 1908, did they ever foresee that one day cars would operate without a driver? Well, more than 100 years later and a vehicle capable of sensing its own surroundings and navigating without a human is now a reality. And makers of these autonomous vehicles are predicting that they will improve traffic conditions, reduce crime, assist the elderly and disabled, and even eliminate driver fatigue forever.
But in the wake of the first pedestrian death involving a self-driving test vehicle, could these cars of the future be short lived? That's the topic of today's podcast. Joining our panel is Professor Paul Salmon from the Centre for Human Factors and Socio-technical Systems at the University of the Sunshine Coast, Doctor Rebecca Michael, Head of Public Policy at RACQ and Professor Andry Rakotonirainy from the Centre of Accident Research and Road Safety, Queensland, better known as CARRS. Welcome everyone.
Rebecca Michael: Thank you.
Andry Rakotonirainy: Thank you.
Paul Salmon: Thank you.
Anthony Frangi: Over to you Rebecca, as the Head of Public Policy at the RACQ, how excited are you about the future of motoring in this country and I guess what you're seeing around the world?
Rebecca Michael: I guess what we're seeing in transport is probably the biggest disruption that we had seen since cars, was the last one basically. And it's interesting that RACQ came into effect of the back of the last major transport disruption. What we see around the world is all sorts of global mobility trends and lot of potential, but there is still a lot of instate uncertainty. We're not sure exactly what's going to take off and when it's going to take off. I mean a disruption only becomes a disruption when consumers and motorists value it and they actually take it up and it has critical mass and it disrupts their transport sector.
Rebecca Michael: So I think one of the big challenges as well as keeping track of all these trends is actually working out and mapping that instate uncertainty and knowing which one is going to take off and where do we put our investment.
Anthony Frangi: Andrey, for you, how excited are you about the future?
Andry Rakotonirainy: The technology is very exciting, but the main challenge is, there is a difference between creating robotic car and putting it in the public roads. And from my central point of view, the Centre for Accident Research and Road Safety, our main concern is trying to establish whether it will be safer or not. On the paper, it will be safer because if we look at the fatal five: fatigue, seat belt, etc., etc., it might help, but the main issues, we will face a transition period in which the public roads will be, will have different type of cars: semi-autonomous to autonomous and normal car, and how those cars will interact with other road users and between them. This is big unknown.
Anthony Frangi: And Paul, for you?
Paul Salmon: Yes, similar to Andry. I guess we're really quite excited about the technologies that are coming in and the capabilities that they will have. But from, I guess, a research point of view, we're slightly sceptical, slash concerned, about how these things are being designed and how they're going to be integrated into the system. What you're going to have really, is a very complex mix of fully automated vehicles, part automated vehicles, vehicles with no automation at all, and I guess the concern from our work is, are designs taken into account? It will all probably work very well when everything is fully automated, but when there's a mix of different levels of automation, and if you also add to that things like vulnerably road users like cyclists, which will be on the increase, there's a worry that the design isn't taking this into account and it's going to kind of create some significant difficulties I guess.
Rebecca Michael: I think that there is a recreational element to driving that people enjoy, some people enjoy, not me in particular, but some people, and when we move down this automated route, obviously, that recreational element will be minimised. Often, I know that Paul Turner at RACQ talks about, that he'll have to go a special track to drive his Mustang, you know, because he won't be able to drive it on the road anymore if it's not automated. There is a lack of probably confidence in automated driving systems at the moment and we've seen a number of crashes overseas using automated driving system, so I think there is a long way to go to building the public's confidence in these technologies, separate to even issues that might have around recreational driving.
Anthony Frangi: Yeah, it is, it is early days. Paul, for you, do we want automated cars you think?
Paul Salmon: Yeah, it's an interesting question and there certainly is some research coming out that shows people are not particularly positive about the projected benefits of automated vehicles. Some of our research is showing this quite interesting increasing desire or need to remain connected to things like social media and work pressures which I think will tilt it back in favour of automated vehicles. So, I think, if they get the design right and they get the integration right, I think people, a lot of people will be wishing to use fully autonomous vehicles and use the extra time that you get from that to do things like emailing and do work and things like that.
Anthony Frangi: Are we going to be doing things like emailing and watching videos and maybe talking to someone on the phone? Are these those sort of activities, do you think, that we'll be doing while we're moving around a city or around the country, Andry? What do you think?
Andry Rakotonirainy: It depends on the type of vehicle that you are going to have. If it's semi-autonomous then you will have to intervene at some stage to take control of the car, because at this stage, an autonomous car that can go from A to B, A and B being anywhere, anytime, any weather, it doesn't exist. The fully autonomous cars now work on a particular route within a particular condition, so we are still far away from the science fiction view of sitting in the car and doing whatever you want.
Anthony Frangi: So that's still many years.
Andry Rakotonirainy: I don't think that it will happen in the next two or three years, so I think it’s still long way.
Rebecca Michael: Automated vehicles, their deployment will be limited and it will be limited by the landscape and road conditions and a raft of other factors. I think that these bullish predictions about, you know, the next couple of years, are designed to drive up share price of companies. I think some of the issues that we need to deal with at the moment is that we've had OEMs that have been driving, making cars and standardising cars for a hundred years yet, we have tech entrants that are coming in now that aren't used to the same compliance and standardization principles that we operate under. There are raft of other issues that we need to get over first.
Anthony Frangi: And what are some of those other issues?
Rebecca Michael: Case in point, Tesla's level 3 conditional automated vehicle, you actually have to have you hands on the wheel at all times, ready to respond and take over if you need to, whereas Cadillac, GM's Cadillac model says: please don't put your hands on the wheel at all. Are we going to get a situation where Mum and Dad both have a level 3 activated conditional automated vehicle and you hop in the wrong one and forget, am I meant to put the hands on the wheel or am I meant to not put my hands on the wheel? So I think those kinds of standardisation issues will be a problem but we also are moving into an area of liability when you start having automated driving systems in place that are produced by particular entity, which may be a bought component by the manufacturer. So if there's a crash on the automated driving conditions, who's at fault? Is it the person sitting in the car who didn't have their hands on the wheel, or is it the system? A whole range of liability issues which will have far reaching ramifications.
Anthony Frangi: Paul I can see you nodding there about liability. Is this a part of the research? It's one of those sort of issues?
Paul Salmon: Yeah, absolutely. So the kind of liability litigation angle, for example: we're doing some research which is looking at people's perceptions of blame when a crash happens with an autonomous vehicle versus not an autonomous vehicle. And what we're expecting really is when a fully or highly autonomous vehicle crashes with say, a cyclist and kills them, people's perception of blame is then going shift, obviously from what it is now, on the driver, to car manufacturers or even to government levels. So you'll see the shift in who people see is responsible for road deaths, which is really quite interesting.
Andry Rakotonirainy: If you look at the current technical approach to create AV, lots of what we call machine learning, meaning that you have data about traffic or about the movement of the car and you use the data to learn about what is a safe behaviour, so it's huge amount of data, that's what Google is trying to do now. We've all heard of Google car, they record those data and those are data to be used to learn what is the right behaviour.
Anthony Frangi: So, the car is learning all of the time?
Andry Rakotonirainy: Yes.
Anthony Frangi: Every time it goes on the road, it's getting smarter. Would that be true?
Andry Rakotonirainy: Yeah, yeah. So, there's that data about the movement of the car and the algorithm which is the machine learning, which is basically the smart. There are two different things: data and algorithm. Someone can just blame the data, how the data were collected, and someone can blame the algorithm, that the algorithm were well designed. On top of that, you have the vehicle manufacturer. So, who is responsible for a crash can be very complicated things to achieve.
Rebecca Michael: It would definitely shift our liability model and you know we might move to a product liability model like they are in the US, where it's not a determination of blame necessarily within that landscape. It'll have ramifications for insurance premiums but I think at the end of the day, the premise of autonomous vehicles is that they will improve safety. And I think one of the issues that we're grappling with ethically, not only is, as Andrey alluded to, the algorithms around the millions of permutations that will happen out on the road every day. Who's making those ethical decisions around how that vehicle will behave, but similarly, when we look around the globe we can say that there's over million road deaths each year and so how safe is safe enough for autonomous vehicles? So what will we tolerate?
Rebecca Michael: And Paul soft of alluded to before that, when an autonomous vehicle's involved in a crash, it is extremely topical and we're quick to say that technology won't work. How safe is safe enough and how do we decide that and what's the risk that we're prepared to tolerate on our roads? Are we happy for families to be doing research and development in level 3 autonomous vehicles out in the public? So, there’s some interesting ethical questions there.
Anthony Frangi: Have any of you been in an automated vehicle?
Andry Rakotonirainy: Yes.
Anthony Frangi: Oh, okay. Andry, tell us about your experience.
Andry Rakotonirainy: The first five minutes, or ten minutes, it's very interesting and you can, you could be little bit, I would say, concerned or afraid, but after that, it's just great. If you are talking about the shuttle bus, for example, which doesn't have a driver, it doesn't have a steering wheel, you go there with 5 or 6 people and the shuttle bus move. And to be honest, you don't really see much difference, you just thinking, there's no driver, it's very slow moving car, but there is, from my perspective, there is no safety concern that I would raise about the shuttle bus.
Rebecca Michael: I've been, I think, in that bus.
Rebecca Michael: I think, again, it depends on the network that you're operating in. So the shuttle bus that I was in was in a closed network, there was no other vehicles on the road and I think as a passenger you have a high degree of confidence. I think that confidence is increased by the fact that you could walk faster than the bus, so you sort of roll the dice that the impact is going to be pretty minimal. And I think that's probably a bit of the problem of where we're at, at the moment in Australia, there's plenty of those trials around the place, I don't think, to speak to Andrey's point, that you're getting a true representation of people's perception of automated driving under those kinds of conditions. Having said that, most modern cars have some form of automation, if you look at lane keep assist and those kinds of things. So it is incremental and I guess people will become or adjust to that increasing technology. The big learning curve will be when the steering wheel goes.
Paul Salmon: I've been in the similar kind of...
Anthony Frangi: So, you've all been in the shuttle bus?
Paul Salmon: Been in the shuttle bus, but I do have a colleague who, we do a lot of work within the UK, who's got a couple of autonomous vehicles for research purposes. He's shown me footage of him in one, and I think Rebecca's point is very good, it depends on the environment that you're in in the vehicle, because, I've seen footage of him on the freeway and it's all very relaxing and things like that, but when the freeway becomes quite busy and things start to get quite complex in the traffic environment, it becomes really quite scary, because the autonomous vehicle is trying to do things that it perhaps can't cope with, and then so the driver has to intervene and take control.
Anthony Frangi: Things like?
Paul Salmon: Could be maintaining a set distance from a vehicle in front, the vehicle in front might start to accelerate and to overtake a truck for example. And so once that vehicle's gone out of the area in front of the autonomous vehicle, this vehicle is speeding up to catch up the vehicle it’s then gone and there's a truck appearing very close to you, and as a human driver, who's not really experienced that before, that can be quite scary. You feel like you have to take control, otherwise the autonomous vehicle's going to head straight into the back of the truck. And I think the other scenario that my colleague was telling me about was, they seem to work quite well on the freeways or highways, but for example, in urban environments, they don't currently recognise roundabouts. So they'll go straight over them if you let them. So if you were really disengaged from the driving task and it was in autonomous vehicle mode, and you didn't monitor the environment and see the roundabout, the car would go straight over it.
Anthony Frangi: Oh dear. It's not good for a Australia, is it? I mean, how many roundabouts do we have in the city alone or in the state alone? But I'm sure they're working on it.
Paul Salmon: But I think Rebecca's point earlier on, is quite important to me in this area. I think we're kind of racing ahead too quickly with it. I think there're issues in the design of these systems that we really should be ironing out and thinking about now before we go to these full trials, and you've seen a couple of significant incidences in trialling of the autonomous vehicles and I think that's why the research we're doing at USC were really starting to look at predictive modelling approaches. So if we take autonomous vehicles and how they're being designed now, we're basically trying to model what kind of crashes and collisions will occur in the road environment so we can then design those problems out now before it actually happens. And I think that's kind of where we're at now in that space.
Andry Rakotonirainy: Within CARRS, we do different type of research in the area of AV. We do simulations of AV and we are particularly interested in how AVs or the other road users share the road. For example, I'm a cyclist and I've been involved in few crashes. When I engage into intersection, I always look at trying to have eye contact with other road users before proceeding. With AV, you wouldn't have that because you don't have a driver. So how do you know that you have confidence that AV saw you? That car, the AV, should have some way to convey their intentions somehow. That is very complicated to do. We do research with driving simulator and we're about to embark in doing research with real car on a closed road, of course.
Rebecca Michael: One of the points of difference is that Australia in not adopting a vendor-led approach to regulation of automated vehicles. We're doing our on research and then the government is proactively designing a regulation to support the safe deployment. I'm not convinced it's the same all the way around the world, America does seem to have a very strong OEM tech company-led approach to trialling of autonomous vehicles and the government is largely hands off at the moment with some of these approaches. So I think that we're well placed here in terms of our approach to research and the rigor with which we're doing it. My concern is that we are a global economy these days and Australia will get the vehicles that they want to send us. And in interest of standardisation, harmonisation across the globe, around standards and those kinds of things, I'm mindful that we just don't want to accept what we are told is safe and that we actually do those tests ourselves and actually determine the safety and have a compliance regime that ensures that we can identify problems in market and address them before they take lives.
Anthony Frangi: What impact are automated vehicles likely to have on, I guess, the economy and the motoring industry, Paul?
Paul Salmon: Yeah, interesting question. Look, I think the role of the road transport system in economic growth is really well known, so it will have an effect. Again, I think it's, if they get it right and they integrate things like driverless heavy goods vehicles and if they get them into the system and they work well, it will have a positive effect through that way, because there'll be more capacity to ship freight around the country and things like that.
Andry Rakotonirainy: From my view, this development of disruptive technology such as AV is a part of a broader paradigm shift in society and transport in general. There are three paradigms that I think is very noticeable. The first one is: we're still trying to get cars that are fuel efficient and safe, that we're continuing to do that so, ANCAP for example is trying to find out what is the best car safe, so we'll continue that and we'll go through this semi-autonomous to fully autonomous car.
Andry Rakotonirainy: The second point, the second paradigm is: we are now moving through the share-riding type of mechanism. The young population doesn't want to own a car so the notion of ownership will change completely, and that have some impact on how you design the urban environment. When you create a car you wouldn't need a parking or garage anymore. So that notion of ownership will change fundamentally our society.
Andry Rakotonirainy: The third one is very much technology oriented, is that future car will be connected mobile objects. So, if we are talking about fully autonomous car, for example, and you are travelling from home to work with autonomous car, you can do something within, and that will be something probably related to revenue income. You can do work within your car. And the all consequences of that is: who can access to that kind of technology that will introduce some inequality in our society. Building an AV is not just building an AV. The infrastructure has to follow. For example, the traffic light should talk to that car to talk about, to say that it's red or green, so it's a huge investment. I don't think that all suburb will be able to afford that. So that will introduce some inequality to our society.
Rebecca Michael: What's interesting to note, is that the whole commercial sector is well ahead in the game in the space. The whole industrial revolution was based on automation, in that we know that automation in autonomous vehicles probably offers productivity gains to logistics and commercial operations that would drive the uptake in that space far ahead of passenger vehicles. Predominantly for the cost really, I think an autonomous vehicle has four light hours at a cost of about $75,000 each. I can't see any of us being able to, unless your name is Tesla, stump up for those kinds of dollars anytime soon. And what we are seeing is this uptake, we're seeing UPS vehicles are using, driving autonomous vehicles, then deploying little robots to actually deliver packages and things like that to increase their productivity. So I think it's a journey that we will go on. There'll be massive economic gains from automated vehicles in the commercial space.
Anthony Frangi: It's a fascinating area. Any final comments before we go? Firstly to you Rebecca?
Rebecca Michael: It's a real journey that we're on and the end, we're not sure, but we need to map that out actively, we shouldn't be passive in this sort of journey with automated vehicles. And I think that my colleagues here are actively participating in that in RACQ as an advocate is going to stay on the front foot of this. As I say to our members: we're jumping in the Petri dish, were going to shake the experiment.
Anthony Frangi: Good. Paul, any final comments from you?
Paul Salmon: Yeah, I mean, as everybody's kind of touched on it, it's extremely exciting. To kind of hypothesize how the road transport system is going change is exciting. My feeling is we need to learn to walk before we can run and there's all sorts of difficulties and constraints and things that we need to work out before we get to fully automated vehicles, basically.
Andry Rakotonirainy: I think the main issue is that human will be stealing forth somehow in this and we are looking at the safety of humans and there is big unknown. The other issue that we need to solve very quickly is the policy or regulation issues. If we look at in 19th century, when the first automobile was introduced, the first policy inducted in the UK and US, is that the pedestrian should proceed the car with a red flag to tell the other road users that an automobile is coming and the posted speed limit at that time was 5 kilometres per hour. I do believe that the policymaker of that time really tried to find out some best way to protect the road users, but we are facing this problem now. We don't know how those AV will behave so we just sometimes shoot in the dark.
Anthony Frangi: Well, it’s interesting, isn't it? Further discussion, I'm sure, is ahead and as we drive in our vehicles today we do ponder what it would be like in the next 5, 10, 15, in fact 20 years.
Anthony Frangi: Rebecca, Paul, Andry, thanks so much for joining us in this RACQ living podcast.
Andry Rakotonirainy: Thank you.
Paul Salmon: Thank you.
Anthony Frangi: If you would like more information on any of the stories raised today, email us at: firstname.lastname@example.org.
Anthony Frangi: I'm Anthony Frangi. Join me next time for more RACQ living.