Proceedings of the Standing Senate Committee on
Transport and Communications
Issue No. 16 - Evidence - May 10, 2017
OTTAWA, Wednesday, May 10, 2017
The Standing Senate Committee on Transport and Communications met this day at 6:46 p.m. to study the regulatory and technical issues related to the deployment of connected and automated vehicles.
Senator Michael L. MacDonald (Deputy Chair) in the chair.
[English]
The Deputy Chair: This evening, this committee continues its study on the regulatory and technical issues related to the deployment of connected and automated vehicles.
I am pleased to introduce our panel of witnesses. From the Innovative Vehicle Institute, we have Mr. François Adam, General Manager; and Mr. Frederick Prigge, Research and Development Director. We also have Professor Ata Khan, Member of ONE-ITS Board of Directors, Civil and Environmental Engineering, Carleton University.
Welcome, witnesses. I invite you to begin your five minute presentations. Afterward, senators will have questions. We will start with Mr. Khan.
Ata Khan, Professor, Member of ONE-ITS Board of Directors, Civil and Environmental Engineering, Carleton University, as an individual: Thank you, Mr. Chair and senators. I am here representing ONE-ITS Research Society. ITS is an acronym for "intelligent transportation systems.'' I have written and presented on connected and autonomous vehicles. From here on, I will call it CAV. From a socio-technical perspective, in 2013, I was invited by the RAND Corporation of the U.S. to provide a prepublication review of their report titled Autonomous Vehicle Technology: A Guide for Policymakers.
I'm glad to be here. My written evidence has three parts, and I'll summarize these quickly here, within five minutes.
Part 1: It appears that during the past year and a half, there has been a change in the direction followed by the CAV industry. They don't really advertise it that much, but because of some accidents that happened due to risky and experimental autonomous vehicles, the industry is backing off a little bit. Now they are focusing on level 4 automation, and then they will move on to level 5 after level 4 is successful. Level 4 has a driver in the loop. Therefore, they have ADAS, advanced driving assistance system.
Another change is that there appears to be a push in the direction of shared autonomous vehicles for the sake of providing mobility as a service.
Car companies right now in North America, Europe and elsewhere have added to their portfolios and ownership a company that will look into how to provide mobility as a service. Car manufacturing companies are no longer just car builders; they will provide what we call mobility as a service.
Another observation regarding change of direction is that they are now realizing that human factors are of prime concern. We have to understand how people relate to automation. Also, when they are designing their vehicles, they would like to build, as much as possible, artificial intelligence and cognitive features. They are also very much concerned with reliability, making sure that the reliability is as good as it can get. Maybe it's not good enough yet. We will discuss that a bit later.
One other thing, for the sake of giving people more than just driving from place A to place B, they are providing an ITS service platform. Within the car, you can have every feature that drivers usually require; for example, route guidance, reserving a parking spot, et cetera.
We are talking about change in the scene in terms of what transportation will be. It will be a lot more than transport.
Part 2 of my presentation is on R&D and demonstration for achieving societal goals, including products with market potential. Let's keep in mind that all the countries that are advanced and have the capability to develop vehicles are all pushing for market share, so why not Canada? We should definitely get our share. Figures 1 and 2 in my written presentation, and table 1, describe the R&D and demonstration challenges.
Part 3: Policy challenges and the role of government. We are not talking about a uniform market; we are talking about two different kinds of markets. One market is going to be highly competitive: Volkswagen against Mercedes against Ford, and so on. This would be a high-competition market. This will be a mass market for vehicles and major parts.
Then we have a low-competition market where governments have to build intelligent infrastructure so that these vehicles can operate and function. The role of government in that market is covered in table 3.
Then what about impacts? We want to have deployment, therefore we are going to change the scenery in terms of how cities will work, how people will live and so on. Those are the impacts I'm covering in table 4.
I'm going to offer a number of observations as a summary. I'm going to read them to you one by one. There are 11 of these, and I will go quickly.
One, CAVs have more to offer than a travel mode.
Two, we need multifaceted policies. It's no longer linear: If there is a problem, there's one solution. It's not the case. You have to look around. If we fix this one here, will it do something else somewhere else? We have to have multifaceted policies.
Three, policies logically should depend upon the level of automation and the market context. A vehicle that drives itself is definitely very different when compared to a vehicle that has the driver in the loop.
Four, increasing automation will require a systematic and comprehensive policy framework, meaning that one has to define exactly how various policies should work together. Supporting methodology is important because we have to test and ultimately implement policies. USDOT has defined 15-point policies at this point as applicable to automated vehicles, and they hope to add more later on.
Five, the public sector, meaning governments, should prepare for deployment. One cannot deploy CAVs in the present traffic environment.
Six, can existing regulations handle level 4 automation? Level 4 is where we have the driver in the loop. Can existing regulations handle level 4 automation? The answer is that we don't know.
Seven, it's highly likely that we will require additional legal capability so as to handle level 5, which is full automation. Definitely there will be a need to consult. There will be a need to have stakeholders talking to each other and so on.
Eight, drivers may not be comfortable with fully automated vehicles. A study done by Deloitte in the U.S. found that about three quarters of drivers are afraid to ride in a fully autonomous vehicle. They will not ride in it. On the other hand, a study by the American Automobile Association found that drivers who have some automation in their vehicles now are more comfortable. Therefore, we need to carry out extensive testing and demonstrations.
Nine, most impacts of deploying automation levels 4 and 5 are positive, and if there are any potentially negative impacts, these can be avoided with policy instruments. I would like to say that there is a lot of speculation. People sometimes go out of their way to come up with statements that have no bearing whatsoever. Therefore, we have to be careful in terms of reading literature as it comes out.
Ten, government R&D assistance should be prioritized on the basis of cost effectiveness. Effectiveness should be related to policy objectives, including developing products that have market value. So cost effectiveness is obtaining the best value for the money.
Finally, eleven, we should have high-quality objective research. Be aware of technology push. If we go the route of technology push, we will have vehicles with no steering wheel. They will have no pedals. If something goes wrong, you have to bring a crane to take them away somewhere. So policymakers have to be very careful, meaning that we must be aware of technology push.
Now we have to have objective research, and definitely policies have to be developed before we start to implement technology.
Thank you.
The Deputy Chair: Thank you, Mr. Khan.
Mr. Adam?
[Translation]
François Adam, General Manager, Innovative Vehicle Institute: Honourable senators, I first want to give you some context to help you determine the type of questions we can answer.
The Innovative Vehicle Institute is a college centre for technology transfer. There is a network in Canada called Tech-Access Canada, for technology access centres, that is made up of about 30 centres. We belong to that network, whose main mandate is to help businesses, especially SMEs, accelerate their innovation. We work with the businesses. We help them to develop new products and to get new technologies to market faster.
Our technological specialty is advanced transport. Since 1996, we have been working on the electrification of transport, and for two years now, thanks to a subsidy from the Natural Sciences and Engineering Research Council, NSERC, we have developing off-road autonomous vehicles, with applications in agriculture, industrial handling and the transportation of people on private land.
Regarding autonomous vehicles, it is important to understand that today's vehicles are already collecting data. The manufacturers that are interested in autonomous vehicles collect an enormous amount of data, because it will be used to create neural networks that will be put into the vehicles that will become autonomous. This data is essential to create the algorithms that will drive tomorrow's vehicles. This data can be very interesting to analyze accidents, or to learn to understand what happens in traffic flows. It could be interesting for the authorities to have access to these data, and perhaps a law that would allow access to certain categories of data could be considered.
We are technological specialists. We use different technologies, such as sensor fusion. We will have more interesting things to say to you on technological issues. It's important to know that as vehicles become autonomous, the percentage of autonomous vehicles on roads will one day reach 100 per cent, and this will be a watershed moment in terms of road safety regulation.
Currently, the vehicles that are on our roads are built to reduce the risk of injury in case of impact, which is very expensive. Vehicles are heavy. There is a lot of security equipment on board, such as airbags. These devices will no longer be necessary when all vehicles are autonomous. Reducing the need for automobile safety regulation will have an important impact on the cost of vehicles. It will make mobility more financially accessible, and it will also impact energy consumption. We think that today's vehicles are heavy for what they do, which is transport people. Lighter vehicles will need much less energy to move.
This concludes the main points I wanted to raise this evening.
Frederick Prigge, Research and Development Director, Innovative Vehicle Institute: Mr. Khan spoke of levels of automation four and five. I imagine that other people have discussed this with you at previous meetings. I will nevertheless summarize them so that we understand each other.
Our current automobiles are at level 1, which means that they provide warnings. The vehicle emits a signal when you leave your lane or approach an obstacle. At level 2, you no longer hold the steering wheel, but you look at the road. This is what Tesla calls "auto pilot''. This isn't an auto pilot at all, but that is what Tesla calls it. The Mercedes S-Class has this feature. With level 3, you no longer have to look at the road. You can text or read a book. However, within 10 or 15 seconds, the driver must be able to take control of the car. At level 4, you can sleep. The vehicle can leave the road and park if it sees a situation it does not understand. At level 5 there is no steering wheel, no pedals and no driver on board.
I was thinking about a few points while on my way to this meeting. Our country is very vast. Half of the population lives in rural areas. We focus on the cities a lot, but autonomous vehicles will be useful in rural areas first. The challenges of autonomous navigation are fewer in the country. There are fewer things that move, the roads are clearer and often the quality of lines and pavement is better. A lot of factors create advantages. This may allow smaller municipalities to offer public transit with autonomous minivans, for instance.
It will also reduce the cost of transporting merchandise to remote cities. The cost of transporting merchandise is enormous in Canada. Trucks will be able to travel on highways for several hours while the driver is sleeping, and this will help to reduce costs. At this time in Canada we have a shortage truck drivers. This is a problem in the forestry sector in particular. There is wood that is not coming out of our forests because we can't find drivers to drive the trucks. And yet these are private roads where there is little traffic, where autonomous vehicles could potentially be used. I don't want you to think only about vehicles that transport passengers, but about all vehicles, be it farm tractors or others.
The last point I would like to make is the following. Canada is not an island. It is part of a continent. Whether we like it or not, autonomous and connected vehicles are coming, or are already here to some degree. Consequently, we must have an intelligent legislative framework that will allow us to integrate them safely. At this point, it is too late to say that we don't want connected vehicles. All of the cars purchased in the past three years have a cellular modem and connected service on board. We can't turn back the clock where that is concerned. Our job now is to frame it to suit ourselves.
There are two aspects we need to think about. A vehicle does not have to be connected to be autonomous. We talk a lot about infrastructure. However, if you want a vehicle to travel as it does now, without knowing what is happening four kilometres down the road, it has to depend exclusively on its perception of its immediate environment. Make no mistake — given the size of our country, we will not be able to have telecommunications on all Canadian roads. Our vehicles will have to be able to function well everywhere, even when they don't have access to a network. And so the vehicle has to be fully autonomous.
We also have to think about what this will mean for Canada's aging population. Being able to have access to vehicles you don't need to drive when it is dark or the weather is bad, or to go to the hospital or in other such situations, presents an advantage. These vehicles will help to improve the quality of life of people who can no longer drive, at least in certain circumstances.
[English]
The Deputy Chair: Thank you, gentlemen, for your presentations. We'll now start our questions.
[Translation]
Senator Boisvenu: Mr. Khan, you referred briefly to government financial support for research. We know that Budget 2017 mentioned a sum of approximately $950 million over five years, which means $200 million a year to create an industrial cluster in this area. If we compare Canada to other countries, we are losing ground to the United States, Germany or the Netherlands, who are ahead of us in developing this technology. Does the money we will invest over the next five years make Canada the poor relative? Should we make a bigger effort to catch up, or at least to maintain our current level of knowledge and development?
[English]
Mr. Khan: Canada is lucky that we benefit from the presence of our neighbour, regardless of what is happening over there right now. We have right here in Ottawa space allocated for testing autonomous vehicles, and the City of Ottawa's economic development department is making sure that they set it up.
Who will test what vehicles? We have QNX belonging to BlackBerry. They are number one in the world in terms of software that cannot be hacked. That is the number-one ingredient. We do not want cars that can be hacked.
Second, Apple has rented space, as has Ford. They are all coming to Ottawa, to Canada, to do the R&D and high- level testing.
Now, what about the government? Government has some limitations, and government maybe should not really give too much money to companies, because there is enough market potential for companies to be on their own. Some assistance, yes. The U.S. is doing it. The U.K. is doing. They are all doing it.
How do we compare? I think the U.S. is ahead. They have deeper pockets, much more resources and so on. Canada is okay. We are very generous to this new technology, from what I understand.
The U.K. is taking a different approach. There was a study done for the House of Lords, something similar to your study. One comment they made was, "Why are we focusing only on automated vehicles for transport? What about automation in other sectors?'' Meaning, look at it broadly. The same technology could work for automation in agriculture, automation in aviation and other things. They are taking a different view, but, yes, they are investing money as well.
Are we doing what we should be doing? The answer is yes.
Are we as far ahead as others? In some respects, we are ahead. In other respects, we have to catch up.
[Translation]
Senator Boisvenu: Thank you very much. Your answers are really very clear.
Mr. Prigge and Mr. Adam, I am trying to understand the role of your centre, and it still seems quite complex and vague. You say that currently you are collecting a lot of data. What methods do you use to collect this data?
Mr. Adam: The vehicles themselves are equipped with cameras, sensors that record all of the data and upload it to enormous servers. The data is then used to train neural networks that will learn to drive like human beings. What you have to understand with this type of technology is that we are using human experience to train an algorithm that will be able to drive like a human being. This type of deep learning technology will also lead to inferences. Millions if not billions of cases will have been submitted to it, and so it will be able to make inferences and to arrive at circumstances that were never actually submitted to the algorithm, and it will be able to make decisions about new experiences.
Senator Boisvenu: If I understand correctly, you are collecting data on human behaviour to eventually arrive at robotic or smart applications.
Mr. Adam: All of the stimuli the car receives and the commands the human provides to respond to a situation are taken into account. What the program is trying to do is reproduce that type of response.
Senator Boisvenu: And do I as a consumer provide this data?
Mr. Adam: One of the large manufacturers who are doing this at this time is Tesla. They collect huge amounts of data from their users because their automobiles are already equipped with many sensors to collect that type of information. The new generations of cars are equipped with most of the sensors that allow them to collect this data, like the most recent Chevrolet Volts, for example.
Senator Boisvenu: Is the driver aware of the fact that he is taking part in this study? Is there some way of waiving our rights? Is this being done without the driver's knowledge, or on a voluntary basis?
Mr. Prigge: Our expertise is with heavy vehicles and off-road vehicles; we do not work in the automobile sector. In the case of Tesla, I know that this is very clearly explained in the sales contract. Does anyone read the 28-page sales contract? That is another issue. I don't know how the other manufacturers work.
Senator Boisvenu: I will look into this tomorrow.
[English]
Senator Galvez: Thank you very much. This is so interesting. It's like living in the future.
Because we are talking in the future and we have to be prepared for the future, I would like to know the time frame between going from level 1 to level 5.
To touch on the problem with the data, I am very happy to give my data away if that data will benefit me and will improve my level 0 because I'm still driving. But at the point when my data is going to serve other people, I think I would like to have a rebate because of the data I'm producing.
But most important is the liability. If I am involved in an accident and if this data is going to come back to me in the form of someone saying, "Oh, you were guilty because you were not driving the right way,'' that is one of my concerns.
Mr. Adam: This is a good question.
First of all, right now we're already at level 2, I would say. You can find cars that have level 2 features, so the driver is highly helped, I would say, to keep the lane and there's automatic braking. The Super Cruise, that's level 2 automation.
Level 3 is not there yet because they don't have the experience to let the car run by itself and wake you up when it's time to take back the steering wheel. You can say that right now we're at level 2.
When will we see all the other levels and how quickly it will attract the customer? I really don't know. But, frankly, some people are afraid of the technology and there are some things we need to know. Some people are thinking about the car of the future being like a place where you will have a sofa and look at the TV shows. You have to think about travel sickness. Some people cannot read while in a car because they will get that sickness. It will be worse in a car where you don't look in front of you and you try to look at something else.
There's not a lot of knowledge about how it will affect humans to be in that type of vehicle for long rides. It's not being studied that much at the moment. We don't have a lot of results on that. There's a lot to be learned about that.
[Translation]
Senator Galvez: The second part of my question is about the infrastructure needed for these automated vehicles. I expect that this will involve electricity and not oil or gas; am I correct?
Mr. Prigge: We can build a fuel-powered autonomous vehicle; that isn't a problem. Mercedes and other manufacturers are already doing this. Looking at it from the opposite angle is what is interesting, because the autonomous vehicle will make it possible for more cars to be electric. Issues involving autonomy, recharging, and route planning are easier with an autonomous vehicle because the vehicle deals with it. You can drive more slowly, it's not a problem, and you have time to do other things. We think that the autonomous vehicle will allow for greater penetration of the electric vehicle. Any vehicle, whatever its means of propulsion, can be automated. At this time, the military have trucks that drive themselves. In areas that are too dangerous to have a pilot on board, the truck drives itself and does what it has to do. And yet they are large diesel trucks from the 1960s.
You were asking us earlier when autonomous vehicles will arrive on the market. I don't have a crystal ball and I never do predictions, because I'm always wrong. However, today if you purchase a very modern farm tractor, you are at level four. You program the route for your tractor and the tractor will plow your field and come back to you afterward. You never have to sit in the tractor it's all done by GPS, by differential radio and all kinds of technologies. There has been a lot of progress in off-road vehicles. Since there are far fewer human beings around, there are fewer legal constraints. The off-road vehicles are already there.
[English]
Senator Bovey: Thank you for your presentations. I'm really interested in the social aspects of all of this.
Mr. Khan, I liked your phrase that we need to be careful of the technological push and make sure we have the testing and studies, and to hear that we don't yet know what it's going to be like to be in one of these vehicles. I'd be interested to know what your sense is and have you expand on the services these vehicles will be able to provide beyond taking us from point A to point B. What are some of the social dimensions and human concerns that we should be looking at?
Mr. Khan: First, I'd like to backtrack a little bit. The history of automation for vehicles, I think, started with Google. They proved that vehicles can drive themselves, as long as you train them, meaning that you take them around a course and from there on you don't have to touch anything.
Yes, then came along another company that is now in the business of producing vehicles, and they hope to have fully or nearly fully autonomous vehicles within a year at the cost of $35,000. But these are not approved by the U.S. government. They have not gone through the detailed approval process yet. People may buy them with the impression that the driver will always be in the driver's seat, with their hands on the steering wheel.
What has happened is that the president of the company made an announcement at the time that they sold their first vehicle, saying that if you switch on the autopilot, then you can watch whatever you want, meaning you can read the paper. There have been accidents and some lives have been lost. Therefore, we are now back to, as my colleagues said, at a level lower than fully autonomous.
We are now beginning to climb the ramp and we are saying let us develop technology, test it properly, get it approved properly and give it to the people. That's one dimension.
The other is the one you mentioned, and that is the social aspect. I think there will be a marvellous revolution in mobility in many ways. From my written evidence, in table 1, I will highlight a few items.
It shows the capabilities of level 4 vehicle automation. The driver is in the loop. The first thing we will notice is an extension of human driver capabilities. There is no way that a human driver can match what the computer and communications can deliver.
The other one is active safety. Not every driver is alert all the time, so one may be distracted, fall asleep or whatever. Therefore, active safety will come forward and will do the job.
There are other things which are very important for commercial purposes. What if people want route guidance, if a certain bridge has a problem or there's an accident somewhere? We would like to know how to find our way through the network, so route guidance is very important. A lot of people pay money for that.
Tolling can be done within the vehicle. One can reserve a parking space in a busy place, like downtown Montreal or Toronto. So many of these are coming up now.
In terms of enriching the lives of drivers, it is going to be a dimension that was not imaginable, but now it will be the case.
There are other capabilities. A vehicle may develop a problem. Therefore, it can send a signal and the driver will reach a destination where someone will be there to change a battery or something else. It is a diagnostic capability.
What if there's an accident and someone has a heart problem and needs attention very quickly? The system will activate the chain of events whereby help could be given to them at the right place.
In other words, from a social perspective, we're going to be very rich in getting services, and hopefully it will be at a low cost level.
Have I answered your question?
Senator Bovey: Yes, you have. Thank you.
Mr. Prigge: Just out of curiosity, I want you to know that the first cross-America trip in a fully autonomous vehicle was in 1986, at an average speed of slightly over 60 miles per hour. This is not new stuff. It was a bit sketchy, but it did work; it didn't crash.
On the social benefits, there are lower health care costs. Accidents are caused by humans, most of them. Of course, there have been fatal accidents with autonomous vehicles, it's true, and it will keep happening but I hope we can get it much lower than with human drivers. Since Canada has a public health system, it's in our best interest to try to keep people safe.
As I said in my earlier statement, I think the aging population of seniors will benefit greatly from that added mobility. That's definitely a social benefit that we can get from those vehicles.
Regarding productivity from mines and farms, we're still a pretty primary-sector economy in Canada. We dig stuff out of the ground and we cut trees. That's a big part of our economy. That will help us greatly to be competitive on the international market with autonomous vehicles dedicated to those tasks.
As stated, six or eight weeks ago there was a huge jam on Autoroute 13 in Montreal during a snowstorm. No authority knew there were 2,000 people stuck on the highway at 3 a.m. This should not happen with connected vehicles. It should flag an emergency response, which could benefit everyone on a social level.
Mr. Adam: If I could add to that, public transit may benefit from these vehicles because these vehicles could be used in bringing people from the last mile to the main public transit system. More people will be interested in public transit if they just have to call a little widget that will come to their home, pick them up there and bring them to the subway station or the main bus line.
In using these types of vehicles in the dense, urban core, you won't have to bother with parking anymore if they are used as a tool like "mobility as a service.'' You get in, you get out; you don't care about parking. That's interesting.
Senator Eggleton: Professor Khan, in the first part of your presentation, you talked about an emphasis on shared services of automated vehicles. You then used the phrase "mobility as a service.''
It sounds like something quite different from what we do nowadays. If we want a car, we buy one from the dealer and take it home. This sounds like a different concept. Can you expand on that?
Mr. Khan: Car companies, original equipment manufacturers, realize that if they do not expand their offerings, they may lose market share in terms of vehicles being produced. So they said that in addition to selling a vehicle to a person or an organization, they want to have companies that are attached to them, and they will provide services that could be "first mile,'' meaning that if you want to go to a GO train station, it will take you there or bring you back, take you to the doctor or wherever you want. They will fill in the gap between private motoring and public transport. They will provide what they call "mobility as a service.''
There's more to it, actually. In Europe, they're very advanced. They can create, for a fee, a complete package whereby they can say, "Take your bike, leave it here and hop on to one of these vehicles. We'll take you over there, and then from there you go somewhere else.'' They're giving them a complete multi-modal package. This is all being done by companies that are in the business of manufacturing cars and so on.
Senator Eggleton: You think they will be selling a service, depending where you live, your needs and your access to public transit. It will be a combination of things and they will sell you a service as opposed to a vehicle. But will people still buy vehicles? Will it still be a mixture or do you think this service will become predominant?
Mr. Khan: There are various projections. One thing is certain: Car ownership will not be an essential feature. People will not be depending upon the car as they do now. They can live within the city or in the suburbs and call one of these autonomous vehicles to take them wherever they want to go.
Could that vehicle be serving the needs of 10 or 100 other people? That is not certain, because people being what they are, they are efficiency-oriented and they want service within seconds. The car is there in the garage, so they can just get in, start it and go. How close can we come to that? In other words, what kind of tolerance will there be to say that you have to wait for it? There cannot be an infinite number of vehicles in the city to serve everybody's needs, so we need some simulations.
I think what will happen is that the use of cars will be different. They could be attached to public transit. They could be shared as cars are shared now. They could be hailed as a taxi. There are all kinds of possibilities.
Senator Eggleton: What about people who live in rural areas?
Mr. Khan: I have seen projections or, rather, speculations that because of autonomous vehicles or automated vehicles, people will leave the city and live on farms because they can have these vehicles take them wherever they have to go.
I think that is a fiction. That cannot happen for a number of reasons. These vehicles are expensive and if they are called, maybe it will not cost as much as a taxi, but surely it will cost a lot of money. So people will not move to the farms to live there. We have nothing to worry about. We will still live in cities, and municipalities will still have smart growth, higher densities and so on.
People who live in rural areas can share, within a certain boundary, some autonomous vehicles. For example, the City of Calgary is looking into a project where a suburban community just outside of Calgary is going to have something of that type whereby they can share vehicles. So they don't need fixed public transit of the usual kind; they can have another kind.
Senator Eggleton: Let me ask the Innovative Vehicle Institute group something. According to notes from our researchers, you have something called the ARION project. You mentioned off-road vehicles, like tractors. Tell me about the ARION project.
Mr. Adam: The ARION project is the one I mentioned for which we received a grant from NSERC. We have the participation of several industrials, in that some are interested in developing vehicles for agriculture, industrial material handling and others for people movers, in fact, in private settings.
Last fall, we demonstrated our first prototype of, let's say, a tractor. In fact, what we are doing in that program is developing algorithms for automated navigation of the vehicles that we can use in different types of application that are off-road. We did a demonstration last fall with our partner with a tractor that was going by itself in a field.
Senator Eggleton: These don't go onto public highways? They are totally on private property?
Mr. Prigge: Yes.
When we think of a vehicle, we think of a car. That's only half the vehicles out there. What do we actually manufacture in Canada? We make a lot of buses, trucks, fork lifts, tractors and all that stuff. It's a big chunk but we don't think about those.
Navigation problems in a factory environment are not the same as on the road. You won't use the same sensors and you won't use the same algorithms. What we're aiming at doing, and it is starting to be pretty good, is to have a very general navigation system that can be customized for very different environments, as automakers are really focusing on one task: driving on a road. Try to bring that automated vehicle into an aluminum smelting plant and it won't work. It won't be able to go around. That's what we're doing, basically: off-road autonomous stuff.
[Translation]
Senator Saint-Germain: Technologically speaking, I am sure that this smart technological driver who will be surrounded by a passenger compartment with seats or a cargo load area, in the case of commercial transport, will work. That aspect does not concern me.
Where I see less progress — and I'm following up on Senator Galvez's question — is around the issue of adapting infrastructures. I understand that smart cars used for individual transport by those who will have the means to pay for them will share the road with group or public transit, and commercial transport. What technological improvements will we have to make, such as to signalling systems? What stage are you at in preparing our roads for this cohabitation with smarter multivehicles that will be smarter than most of today's drivers? That is what I don't see.
Mr. Adam: Regarding our activities, since we only work on off-road vehicles, we have not looked at that aspect in our internal IVI reflection.
Senator Saint-Germain: That's an easy answer. Come back to earth and to the road.
Mr. Adam: At this time, manufacturers are trying to develop vehicles that will be able to function without special infrastructures. The vehicle learns to recognize its environment; it recognizes the roads, the other vehicles, the pedestrians and cyclists. It recognizes its environment and knows how to move through it. The communication with infrastructure will probably be useful at other levels aside from navigation; it will be useful, for instance, to see what is happening in the city, and to change the route if the vehicle knows that there has been an incident downstream, for instance. This will allow for more fluid traffic flow.
Senator Saint-Germain: Concretely, in the case of a bus that left Lévis to go to Montreal using Highway 20, could Highway 20 at this time, without any changes, allow an autonomous vehicle to make it to Montreal? Would the vehicle be able to read everything itself?
Mr. Prigge: A highway like Highway 20 is the easiest environment. You can get into a Tesla in Quebec and turn on cruise control — which is mistakenly called the "autopilot'' — and you can go to Montreal without touching the steering wheel or the pedals. The most basic infrastructure for vehicles are clear lines painted on the road. This isn't cutting-edge technology.
Senator Saint-Germain: Railway transport — the first means of transport before the airplane — still exists. How do you see the adaptation of autonomous railway cars and the commercial competition between public transit on roads with completely autonomous vehicles — here I am at level 4 and even 5 — and the train?
Mr. Prigge: We have not worked on train cars up till now. I am sure there are people who are much more competent than we are in this field. These are very competitive companies whose aim is to lower the cost per kilometre for cargo. I am sure that they are already doing this at this time, that is to say the auto-switching of train cars.
Senator Saint-Germain: That is not part of your study. Thank you.
Mr. Prigge: No.
[English]
Senator Duffy: Thank you to the witnesses. It's a fascinating subject.
Dr. Khan, you mentioned fear of early adoption. I don't think you used that phrase, but that people would be fearful of getting in a vehicle where they did not have a sense of control. People have read about the U.S. space program, and the original spaceship wasn't even going to have a window because it was all automated. John Glenn insisted they put in a window so it looked like they were actually flying the thing when it was being flown by a computer.
How do you see that fear of the public to adopt, and secondly, the interaction between platoons? We have had other witnesses who talked about platoons of trucks. You have humans who are terrified on the highway as it is, and now we will have semi-autonomous vehicles, platoons of trucks and these terrified drivers all intermixing. How does that resolve itself?
Mr. Khan: Automation is going to take place gradually, meaning that first we will have vehicles that can sense a lane, and if there is a departure, they can give an alert. If there is a problem, the brakes are applied automatically. People will get used to it.
There have been complaints: "I do not see any problem; why have you applied my brakes? I almost went into the windshield.'' That's a design problem. They can tweak it a little bit and therefore it will be more sensitive to how the human being feels.
It is inevitable that we will gradually see automation taking hold. Now, what will happen is that maybe 1 per cent of the fleet will be automated and 99 per cent will not. What will happen when we have 50 per cent? What will happen when we have 80 per cent? No one has really looked into it in the same level of detail as the House of Lords' study. They have actually spent a lot of money in doing that. Their conclusion was that we cannot really have automation in a meaningful way without changing the intelligence in the infrastructure.
Going to the platooning aspects on the highway, Swedish R&D has proven that with communications technology you can link a number of trucks. Because they are closer and there is less aerodynamic drag, they save fuel. So, technologically, it is possible for form platoons of freight vehicles.
What will happen when there is a very long platoon of trucks? What will happen to people who are driving? Will they be afraid to travel near them? That's a human factor that no one has really addressed yet. It is still to be addressed. We are going to see a step function. Bit by bit, people will get used to it.
Let me pose one important challenge to my colleagues who are in the technology business. If we have a good driver — alert, not aggressive — can technology replace that driver, the same level of intelligence, given that there could be millions of options? On a routine basis, how many decisions do we make? Do I merge now or not? Do I look that driver in the eye and say, "Will you let me or not?'' What about the road? How slippery is it? How much distance is there? Can technology replace human intelligence? Is it possible?
Mr. Adam: Yes. In fact, technology has so much power. The processing power of the computers we put in these vehicles are — let's say, for example, a board that is available to put in the cars right now to do that power processing. It has the power of 150 MacBooks in a single board. It will check all the surroundings many times per second. You have no idea. No human can do that. It's much more aware of its environment than any human. It will have information about the traction of all wheels in real time, so we cannot, as humans, be as efficient as that machine.
But that doesn't mean that machine never fails. They will fail. Any computer or hardware can fail because of workmanship problems. There are going to be problems.
Right now 93 per cent of accidents on the roads are caused by human error; if we reduce that by half, that's already a great improvement in the death toll.
Senator Duffy: Do you see a parallel between what you're proposing for the highways and what we saw when Airbus developed the fly-by-wire aircraft 25 years ago? There was no longer a steering wheel for the pilots; they had a joystick. There were no longer hydraulic cables inside the plane pulling the rudders; it was all done by servos and electric motors. Nobody thinks about it now when you go get on board an aircraft to realize it's essentially — and it can take off and land on autopilot. We don't like to tell the passengers that because we don't want them to be nervous. Is that essentially what we're adapting to the highways?
Mr. Adam: Kind of.
Mr. Prigge: Those technology breaks are always scary. That was one that people were not aware of.
But that fright about a platoon of trucks, people had the same fears when they started to have cars on the road. They were riding their horse, and there was this car coming by. They were afraid of it, but we adapted and everything was okay.
I'm not saying it's going to be easy and perfect, but yes, if systems are well implemented, and when they have a purpose, when we know why we're doing that, we should manage it properly. It shouldn't be an issue.
You are referring to that Airbus case. Right now, I think Infiniti that is making two car models where there is no steering column. Your steering is a joystick. There is no mechanical link to the wheels. People are not afraid that's going to break and they are going to lose their steering, because it has been tested and retested.
Thankfully our neighbours in the south are quick to take you to court, so that helps us. The manufacturers of those vehicles don't want to be sued. They usually do their homework. Not for emissions, though.
Senator Runciman: Thank you for being here. I wanted to talk with Professor Khan about policy development and the role of government going forward and the recommendations the committee might make. I know you have discussed that in your paper. We have heard testimony with respect to the legal framework and liability and litigation. You are also talking about cost effectiveness as a policy. That would be a refreshing initiative, no doubt about that.
I wanted to spend time with you on the regulatory role. I hadn't heard about the new car assessment program, but perhaps there was earlier testimony which I wasn't present for or missed.
You talked about Tesla putting cars on the road. Marketing them as fully autonomous without having approval of the government, that boggles my mind. But we have had a history, recent cases, where governments seemed to be operating on assumptions and trust. We have seen it with Volkswagen, which was an environmental issue essentially, but also with General Motors, which cost lives, where these things are not brought to the public's attention and addressed.
I'm concerned about that. I'm not big on regulation. But initially, as we move into this area, with so many unknowns, and we're talking about these systems used by people who might wish to do us harm, for example, plus the privacy concerns. We had a witness who did polling on it. They said 75 per cent of the manufacturers have no strategies with respect to how to combat hacking.
I would like to get your best advice on how involved government should be, at least in the initial stages, to ensure that we're protecting consumers and the public at large.
Mr. Khan: Those who work for the government, it is a challenge they have never imagined. Here comes a manufacturer who classifies a vehicle as fully autonomous, level 5, meaning you take it out and it will take care of you, no problem. How do they know? Have they tested these vehicles under all possible conditions, especially in Canada, where lane marks cannot be seen for three or four months? There is a bad rain, or someone stole a sign and it is no longer there, or there is sun shining from the other side and the cameras/LiDARs cannot see.
It is an immense responsibility, and it will be well beyond 2025, 2035, maybe 2040, before any administrator will certify that this vehicle is safe.
Let's backtrack. The industry has realized that the day of fully autonomous vehicles being approved by governments is not coming soon. They want to make their money by developing level 4.
Mercedes Benz says that all you have to do is put your hand on the steering wheel: "You make the first move and we will take care of you from there on. We will make a turn safely; therefore, you will never go off the road. If there is a problem in terms of distraction, we will take care of you. If there is a problem in terms of traction of the road,and the tire is not providing friction, we will take care of you. If the cars are connected, in that case they will communicate with each other; therefore, there will be extra safety built in.''
In terms of responsibility that the governments have from a safety perspective, it is unprecedented. There has been nothing like this before.
Going to the social side, we cannot stop technology. The technology is always there. It will come. Someone is going to build it and someone will buy it. How come approximately a quarter million people made a down payment to buy a Tesla, fully autonomous, certified or not? They may have signed a paper saying, "Yes, I will watch; I will be in the driver's seat.'' But when they realize that automation is there, they can switch it on and forget about everything else.
Senator Runciman: What is the process today? You mentioned NCAP. Is that a governmental agency? It's not? It's an industry-wide agency?
It seems to me that if these mistakes, if you will, by major manufacturers, which are covered up, are not caught in the approval stage, obviously we're going into much more complex areas. I would think that there has to be a recommendation that the government play a more active and involved role in approval of these vehicles before they go on the road.
Mr. Khan: Yes. Six months ago, the United States Department of Transport, in association with National Highway Traffic Safety Administration, developed a 15-point safety assessment checklist for manufacturers, meaning that you say you have a product, you go through this checklist. Where is it? Level 2, level 3, level 4 or level 5? What does it have? "Show us your papers. Have you tested that?'' So they are getting around to implementing the process.
What I'm afraid of is that sometimes technology will be ahead and regulations will try to catch up. That should not happen. We cannot really take a chance. People think that if they're buying it, the government is allowing it. It may not necessarily be safe.
Senator Bovey: Because of the work you're doing with large vehicles, does that include emergency vehicles? You talked about the crisis in Montreal with the blizzard, which gives rise to this question: Are you testing ambulances and fire trucks and emergency vehicles with this technology?
Mr. Prigge: Since we work for private companies who hire us, we don't have enough grants to live on that money. We get hired for private contracts. We have not had an ambulance or a fire truck company come to us to work for them. The answer, sadly, is no.
Senator Bovey: That could be interesting.
Mr. Prigge: Absolutely.
Mr. Adam: More buses and trucks.
Mr. Prigge: Especially in dangerous situations like what the military did, it's a good thing to have an automated vehicle.
Senator Galvez: I am interested in that technology and the intellectual property and how the technologies you are developing will move the Canadian economy. You mentioned NSERC, and my colleague wanted to know if we are investing enough. What are we getting as Canadians?
Mr. Adam: That was exactly the purpose of that research program. In fact, as we were developing algorithms for automated navigation for Canadian industrials, the idea was to make sure that Canada could position itself in that market, and that in the long run we won't become users of technology from Europe. There was nobody in that field in Quebec. That's why we are working with them to make sure that they will have products that will be automated in the coming years. We decided to keep the IP at IVI, and we licensed it to our industrial partners.
Senator Galvez: These industrial partners are Canadians?
Mr. Adam: Yes, all Canadians.
[Translation]
Senator Boisvenu: I have a question, out of curiosity, Mr. Adam: how does your centre function? You are a private enterprise; do you have contracts?
Mr. Adam: We are a non-profit organization affiliated to a college. We work closely with private businesses. We mostly do applied research on product development. In fact, we have three specialties: applied research, technical assistance to businesses, and information.
Senator Boisvenu: Do you have any human resources?
Mr. Adam: We have a 25-person team, mostly engineers who work on development.
Senator Boisvenu: The research in this area is evolving.
Mr. Adam: Yes, constantly.
Senator Boisvenu: If one goes to your offices, can we physically see what you do?
Mr. Adam: Certainly.
Senator Boisvenu: It's very interesting.
[English]
Senator Duffy: On the question of software, Professor Khan, you mentioned QNX and BlackBerry. You mentioned that Apple has come here. I gather QNX and BlackBerry are writing software in Kanata. BlackBerry we understand because it is a Canadian company; they did have operations here as well as Waterloo.
What attracts a company like Apple from Cupertino, California, to come to Ottawa to work on autonomous vehicles? Or is that what they are doing?
Mr. Khan: I'm not sure about why QNX, when it was bought by BlackBerry, was moved from California to Ottawa. Why not locate it right next door to the owner, BlackBerry, in Waterloo, Ontario? It could be because Ottawa is the high-technology centre — the Centre for Advanced Technologies.
Senator Duffy: These people who work in high-tech like to talk to each other and have a community where they can trade ideas, which forms a nucleus.
Mr. Khan: Yes. Let me give some background as to how these things came together.
I don't know if you have access to my written document. I would like to draw your attention to figure number 2. That will place the Ottawa scene in terms of what is happening. Figure number 2 is the high-level architecture for an automated vehicle. It has basically four boxes. One is called "Advanced driver information systems.'' The other one is called "Advanced driving features,'' such as platooning and cooperative driving. The third one is "ITS services platform.'' The last one is called "Infotainment systems.''
QNX made their money by developing infotainment systems. Their product is in millions of cars around the world. They take pride in saying that their software can never be hacked.
Then they got the idea, why stop at infotainment? Why not go to the advanced driving assistance system and use the same technology and develop that. This is what they are doing now?
They came to Ottawa for a number of reasons, I think. There are other high-technology companies which are developing communication G4, G5, et cetera. Also because they are close to the government. The other one could be that maybe the big boss of QNX wanted to come to Ottawa because skiing is prevalent. It's interesting to know why high-technology companies came to Ottawa — because of labour and so on.
Senator Runciman: Weather.
Mr. Khan: QNX is here. Apple decided to come. Now there is a big park in terms of having all these technology groups coming together in one location.
Their centre is going to be with Blackberry in Kitchener/Waterloo. They teleconference and travel and so forth.
One thing is certain: There is a core of technologies being developed that cannot be matched anywhere else. Why is this the case? BlackBerry is very advanced, QNX. Ford is very advanced. Apple is very advanced. With all these companies coming together, they're next door to each other, something good is bound to happen.
One thing people are afraid of is that when you have thousands of these vehicles running on Highway 401 in Toronto and if one of them is hacked, how many cars will there be in terms of plowing into each other, travelling at 100 kilometres per hour? An automated vehicle should never be hacked; otherwise, it's a disaster.
Whom do you go to? You go to QNX, the number one provider of that technology. That's the reason Apple and Ford came, and maybe why others will come.
Are they developing software that is way ahead? As a university, we went to the shop of QNX, and they came to visit us. They don't say much. They are very secretive because it means that otherwise someone may get an idea as to how it is done. They are very serious about developing the number one technology for automation. They will not be any other, because there is no other software company whose product is in millions of cars around the world. QNX has that.
Senator Duffy: It's a little like Windows being on virtually every computer in the world. They are that dominant.
In a way, what you're telling us today is that while production of vehicles, the car plants that we used to see of the old days, has now migrated to lower-wage countries, we are in effect replacing some of those jobs with high-tech, good- paying, non-polluting jobs of the future right here in Canada. It's not a dream; it's really happening.
Mr. Khan: We have to take a longer view. Ultimately, all vehicles will be electrified, meaning that they will have batteries.
In terms of building cars, they are not built in one place, to my knowledge. Parts come from here and there and are assembled wherever the manufacturer thinks is a good place; it could be Mexico, Canada, the U.S. or Europe.
What is the future for Canada in the automotive industry, the high-technology type, the automated type? If governments play their role, they will go after incentives so that manufacturing of autonomous vehicles will take place here, because the brain of automotive technology will be here.
The batteries will come from the U.S. There's a big factory somewhere in the desert where they produce millions of batteries and they're very advanced, but that's okay. One can get them as a part of free trade.
Low-wage countries have their own problems. There's a problem of reliability. Mexico is just getting over it. They're convincing others that their workers are as good as the Japanese workers; maybe, maybe not. Canada stands to gain a lot. The main reason they can make a condition that we will give you the engine, meaning the software that will drive the electric, autonomous vehicle, if you locate your plant here. You can have parts built anywhere in the world, but bring it here to put it together.
I hope that, through your good offices and so on, you can give a gentle push to the government to say, "Go after the economic benefits.'' We are at that level where it can be done. Now it can be done. Maybe in another 10 years, it will not be possible.
Mr. Prigge: I would like to come back to why they are here. To my knowledge, QNX is a company from Ottawa. They did not move here. Your analogy to Windows is perfect. What they make is an operating system. It is in many factories around the world. It started as an industrial operating system. They brought that mindset to vehicles.
Mr. Khan: I'd like to correct that, if I may. QNX was a California company bought by BlackBerry. They were transplanted from California to Ottawa.
Mr. Prigge: In the 1980s they were in Ottawa. I'm not that old.
In terms of why they're coming to Canada, it's easy; it's because of artificial intelligence, AI. The University of Toronto and the University of Montreal are at the top of their game. They are maybe among the five best in the world. Those algorithms are built with artificial intelligence. The initial pull to bring these companies in, Google gave some money to the University of Montreal and the U of T. That's why, because we're good in artificial intelligence.
We also have that pool of people who learned during the telecom boom during the late 1990s. Those people are still here; they know how to write very good software and how to build very robust electronics. Nortel and JDS Uniphase might not be here anymore, but those skilled workers are.
[Translation]
Senator Saint-Germain: When level 5 autonomous vehicles are on the market, do you think they will self-finance based on the user-pay principle?
Mr. Adam: Very good question. I am not an economist, but it is certain that a small-capacity autonomous vehicle that can transport a few people will cost much less than a full-sized bus. So if it is used for public transit in smaller municipalities, you have to think that today there are municipalities that offer a public transit service with 40-foot buses that are empty. That's costly.
Senator Saint-Germain: Are there any studies. . . . I understand that you are not doing this, but perhaps Mr. Khan could answer. Are there any businesses, universities and centres studying the economic issues related to the funding of this commercial and industrial product?
[English]
Mr. Khan: If there is mass production of any technology, the cost goes down, of course. The basic ingredients of automation are coming down in price. Once upon a time, the LiDAR that basically measures the distance and so on, now one can have seven or eight of those in an automated vehicle for the same price as one. Ultimately, there will be an increment of about $10,000 as compared to a vehicle that we buy now.
Electric vehicles are going to be subsidized by government for a while for environmental reasons; ultimately, they will pay for themselves.
I'm optimistic that if we overcome the safety issue, the economics will be favourable, for the simple reason that there will be a sharing of the vehicles and the vehicles will be providing mobility services. The vehicles are going to be renewed frequently; therefore, there will be no repair, basically. If its life is gone, you will get another one.
Economics-wise, the studies I did are in a 2012 paper that I wrote for a journal based in Japan. The cost- effectiveness is favourable for automation, yes.
Senator Saint-Germain: Thank you. We'll have to work on this further.
The Deputy Chair: I'd like to thank Mr. Adam, Mr. Prigge and Mr. Khan for their participation this evening.
In our meeting next week, we'll hear from representatives from the University of Alberta and the University of Sherbrooke.
(The committee adjourned.)