THE STANDING SENATE COMMITTEE ON TRANSPORT AND COMMUNICATIONS
EVIDENCE
OTTAWA, Tuesday, October 17, 2017
The Standing Senate Committee on Transport and Communications met this day at 9:30 a.m. to study on the regulatory and technical issues related to the deployment of connected and automated vehicles.
Senator Michael L. MacDonald (Deputy Chair) in the chair.
The Deputy Chair: This morning, the committee will continue its study on connected and automated vehicles.
For the first panel, I would like to welcome Mr. Kirk Steudle, Director of the Michigan Department of Transportation He is also a member of the Intelligent Transportation System’s Program Advisory Committee to the U.S. Department of Transportation.
Director Steudle, thank you for attending our meeting and agreeing to speak with us today. I invite you to begin your presentation and, afterwards, the senators will have questions.
Kirk Steudle, Director, Michigan Department of Transportation: Good morning and thank you for allowing me to speak to you on the regulatory and technical issues related to connected and automated vehicles and the experience that we’ve had here in the state of Michigan.
I can summarize this into four phases: be flexible, be broad, be modest and be learners.
I believe you have my written testimony, so I’m going to skip through some of that and just give you the highlights.
First of all, be flexible. Michigan first passed laws regarding automated vehicles in 2013 with effective dates in 2014. Those were well-intended and honest efforts. They were also short-lived. Just two years later, manufacturers and technical leaders were telling us that the laws were barriers to the most forward-looking developers and they were starting to look elsewhere for their activities.
The problem was that the laws were “test only.” They allowed testing under tight controls but no more. They attempted to keep development at a slow and managed pace, but the technology was speeding up. Developers were already at the doorstep of wanting to test on public roads and move closer to routine deployment. Just two years after enacting the “test only” restrictions, Michigan lawmakers amended that to allow open testing on public roads and full deployment. The key is that the law now allows operations rather than limits just to testing.
Had we stuck to the 2013 laws, we would have closed off important research in Michigan and encouraged its development elsewhere. The point is that we need to be flexible in regulation as the technology is refined and new opportunities arise. The 2016 laws that were passed made important changes and removed the “test only” restriction. However, I anticipate we will be back into this discussion in the next year or two as we begin to look at more specific details like references to the “driver” and to “operate.”
The second point is to be broad. At first glance, we thought a good revision to the 2013 laws would be to just strike the “test only” restriction in our laws. Quickly, we learned that there were more developments and more issues that really needed a comprehensive response. The trucking industry was interested in allowing truck platooning, which is a series of trucks that are electronically tethered together where the lead vehicle sets the speed and consequently the following distance of the following trailer vehicles.
Developers and manufacturers needed new and innovative test sites. The laws formally established the American Center for Mobility at the site of the old Willow Run bomber plant from World War II.
Manufacturers and tech companies strongly desired the ability to run automated vehicle networks. Think of Lyft and Uber without a driver. The first draft of the legislation did not even include this issue, but it became far and away the leading and most contentious point in the debate.
Each of these issues could have been considered at their own pace and separately, and it was important to address them at the same time to be responsive to very distinct requests from actual developers. Second, addressing these issues collectively helps reinforce the message I think that we desire: that the traditional centre for automotive research and production will remain at the forefront of this new phase of mobility.
The third piece: be modest. This might seem like a contradiction to the last point about being broad, but I want to emphasize that government regulators need to be mindful of their role and respectful to its limitations. Michigan state government cannot and should not attempt to tell automakers and technology start-ups where to invest and how to develop their products. We have an important responsibility to safeguard the public interest, but this does not mean micromanaging or steering the direction of highly technical and innovative research.
At the same time, states in the U.S. need to recognize their limitations in vehicle regulation. In the U.S., safety considerations of a vehicle’s design, construction and performance have traditionally been the responsibility of the federal government, or NHTSA. States have regulated driver licences, liability, operations of vehicles and traffic management. States need to resist the temptation to stray into “just a few” equipment standards because of the novel nature of these vehicles. In the short term, such efforts will push developers and manufacturers to other locations where they can do the work more freely, but in the long run, multiple state regulations will only confuse and confound the industry and allow overseas competitors to gain a great advantage.
Finally, be learners. We need to return to an absolutely fundamental component of success. We have to cultivate an environment of learning about the technology and thinking ahead to how to visualize and solve new issues. Several months ago, a hot question was how would an automated vehicle recognize the hand signals of a traffic officer when signals were not functioning or following a major sporting event? First, we had to analyze the situation to realize it’s a two-step process. The vehicle’s automated driving system must learn to disregard the traditional signal and then respond to the officer. One promising solution has already emerged. Officers would have reflectorized vests with signal equipment sewn or embedded in to communicate with the vehicle.
More fundamentally, we have to recognize this field is going to need a steady flow of new engineers, software engineers and thinkers, perhaps from unconventional backgrounds. The technology is new and amazing, but the fuel is still ingenuity and creativity. This gets to the broader question of renewing our education system, but that’s truly essential.
In conclusion, thank you for the opportunity to share these brief comments. I genuinely appreciate your interest and hope these examples from Michigan will be helpful to you. I’d be happy to answer any questions that you might have.
The Deputy Chair: Thank you, Mr. Steudle. We’ll begin with questions from the senators.
Senator Griffin: I really enjoyed your presentation. Thank you for all of your points. I’m interested in all of them, of course, but in your last one in particular, namely, be learners. I don’t know if it’s any different in the U.S. than in Canada, but I suspect not a whole lot different. In the adult population, quite a large proportion of people are functionally illiterate. In terms of your fourth point, renewing our education system, I can see that perhaps that would work for the school system, but what about the great mass of us who already graduated supposedly from the school system with a certain amount of knowledge, and yet many of us are not at all good at functioning in our daily lives because we can’t read well enough or we don’t know our numeracy well enough? Can you give me some further comments on how you would see not just working with the school system — I think I can visualize that — but with the adult population? These are the people who are buying vehicles now.
Mr. Steudle: I would characterize it two ways. First of all, part of the learners, in particular to the school district — and I want to clarify one piece — really has two components. That has developing the future developers of the technology and teaching the general population about what this technology does. I think we would all probably agree that the younger they are, the more easily they adapt to the technology and changes that we all see every day.
With the second part to the broader citizenry, our experience has been when somebody actually rides in one of these vehicles, a truly automated vehicle, it doesn’t take very long and they become very comfortable with the technology. Having said that, the technology has to be intuitive. It can’t have a whole list of requirements, do step A, step B, step C. It really has to be much like an elevator in a building. In the early days, they called it a magic box. You get in the magic box and push a number in reference to a floor. In automotive terms, that box now becomes a pod that you sit in and you touch a destination. Maybe it’s one that says “home” or “school” or “grocery store” or something like that. You push the button, and it interacts on its own and manoeuvres to get to your destination.
We’re in the middle of a very messy transition. I think everybody can realize, when you get to that utopia, that makes a lot of sense, but the transition, in the meantime, is going to be very difficult.
Senator Griffin: Yes, that’s my feeling about it too; it is going to be difficult.
I have another question on a different topic. A lot of experts think automated vehicles will lead to less congestion on our highways; others are saying the opposite. What has been your feeling on this?
Mr. Steudle: I’ve heard both of those. I’ve had a lot of places ask me, “What should we do with our big parking lots? When can we turn them into parks?” I said, “Not yet. That’s one version of what the future looks like.” As you referenced, the vehicle miles travelled very likely could decrease. It could also just as likely increase. In fact, one study that I read showed it going up four times what it is today.
One of the things that’s different is if a whole platoon of vehicles is all being controlled kind of like a swarm of bees, they all move together, then you don’t have the congestion — the stop/start. What happens with congestion today is somebody hits the brakes and then it backs up, so people aren’t necessarily always paying attention. You do eliminate a lot of that congestion if the flow of traffic is all consistent, and it’s consistent because the vehicles are talking to each other and they’re sensing what’s happening around them.
While I agree with your first piece, the future has a lot of different variables. I do see when the traffic flows, particularly if you think of a motorway, expressway or freeway where everybody is going in the same direction and there are not a lot of cross streets, I do see that evening out because you can decrease the amount of space between the vehicles because you’re not depending upon the driver to look and actually see a taillight. The vehicle knows that the whole platoon is slowing down or speeding up.
Senator Bovey: Thank you for your presentation. Having had the opportunity to ride in one of these vehicles, I can say you do get very comfortable as a passenger very quickly.
I’m interested in the July 2017 agreement between Ontario and Michigan to continue the collaboration on testing and development and marketing of the AV and CV technology. Can you tell us a bit about the details of that agreement?
Mr. Steudle: Actually, that was one of the things that I did glance over in my presentation. In July, we had an autonomous drive from Detroit into Windsor, up to Sarnia, across the Blue Water Bridge up to Charter City and culminated with the signing of this agreement. The agreement was basically they were going to work together.
One of the very near-term things is in two weeks the Intelligent Transportation Systems World Congress will be in Montreal. In Michigan, we are very big supporters of that. In that document, it says that Ontario would get very involved with that conference as well. So that’s one short-term thing that here in a couple of weeks we’ll be able to check off and say yes, we’ve agreed to do that and yes, we’re there.
More broadly, it’s about sharing demonstrations and experience across the border, such as the technology. A week and a half, ago I was at the Blue Water Bridge with the U.S. Army’s research and development arm, called TARDEC. The vehicle convoy drove across the Blue Water Bridge to the plaza and then turned around. Many Ontario officials were there and in discussions with the research centre about how do they take that into consideration further into Ontario.
It really is a collaborative piece. There isn’t any kind of an agreement that says, “You do this and we’ll pay you,” either one way or the other. It really is about collaboration between the two countries, recognizing that what we do in this particular space really impacts our economy and we’re much more closely tied and aligned than, say, the United States and some southwest state.
Senator Bovey: What about regulations? I’m confused or concerned, or maybe both, that in Canada we have, obviously like the States, the provinces have responsibilities that the federal government doesn’t have jurisdiction over, and I’m intrigued by this cross-border agreement. What does that mean for the international regulations and safety of these vehicles? If somebody is driving down the highway and you’ve got these platooning connected vehicles taking up a lot of space, how do people get off a highway?
Mr. Steudle: That’s a great question. That’s one that came up during the discussion about truck platooning. The Michigan law specifically says that a truck platoon must break to allow access for exit and entrance ramps. If a vehicle is coming down an entrance ramp and there happens to be a platoon, you can easily see getting by with two trucks platooning, but what if there are 10 trucks platooning? The law says you must break.
Importantly, in truck platooning there is always a driver that has a commercial driver’s licence behind the wheel and he is still steering. The only thing different is they don’t have their feet on the pedals and they’re following at a close distance. It will mean that the driver still has to pay attention. As the vehicle comes down, he has to tap his brakes, disengage the platoon and let the vehicle in or, conversely, if someone is trying to exit.
Senator Bovey: Are these covered in the international regulations between Canada and the U.S.?
Mr. Steudle: They’re in Michigan law. I don’t know right now that there is an Ontario province law. They’re looking at Michigan’s laws to determine what they should adopt. I don’t believe they have the ability to do truck platooning, again, from an operations perspective. They may have the ability to do some testing, but I don’t think they have the operations component that Michigan does. I would be very hesitant to suggest they do or they don’t because I don’t know the Ontario law.
Senator Galvez: Thank you very much for your presentation. Thank you very much for being here today with us.
I’m interested in the more technical and specific details of your experience, your experiment and your pilot. You said that you wanted to be more modern, productive and mutually beneficial. Can you please put that in parameters and tell me if your objectives were attained?
Mr. Steudle: That’s our broader perspective. Where our laws were at with the 2013 testing restriction, we saw companies that were saying, “We’re going to leave. We can’t do what we need to do here, so we’re going to leave the state of Michigan and go to another state that has more progressive laws. Maybe it doesn’t specifically say test only, and we’re going to go there to do the testing.”
That piece is the economic component. If we didn’t do that, we could see the major employment area of our economy moving away. One thing we’ve said is that it’s one thing to lose a plant to another area, but to lose the technical R&D would be a staggering event for Michigan. Seventy-five per cent of North America’s research and development companies are within 100 miles of southeast Michigan, including in Ontario. That represents 376 companies. So we’ve said we really have to protect this market or it will go away.
Senator Galvez: Thank you very much, because that makes the reasons for your pilot clear. I understand that you want to protect your R&D and that you want to protect intellectual property. I understand that very well.
We now come to why we need connected vehicles and what problem they will solve. I was reading in the news about your Katy Freeway. It has 26 lanes. It was constructed in 2008. It is the biggest highway in your country, and it is already blocked. Can you please tell me how these automated vehicles will solve your problem of these highways that are losing their purpose?
Mr. Steudle: First, I don’t know which highway you were referring to. I know that in Michigan we don’t have one that’s 26 lanes wide. Somewhere in the country, we probably do, in one of the big metropolitan areas, Los Angeles or Atlanta.
However, from a Department of Transportation perspective, we’re interested in safety. Last year 35,000 motorists died on American roadways. In Michigan, it was 1,081. What has been proven is that automated, connected technology will reduce those crash numbers by about 80 per cent. If you look at crashes in total, 94 per cent of crashes are caused by human error. Even if you can only remove 50 per cent of those crashes and 50 per cent of the fatalities, that means 17,000 people are still alive; specifically in Michigan, 500 people are alive this year that died last year in highway crashes. One component is safety. From a Department of Transportation perspective, that’s what we’re most interested in, the safety component.
The second one is the congestion component. You get to big, urban areas where there is lots of congestion. To do nothing, you will have the same congestion in the future as you do today, if not worse. The only other method is to try to make that more efficient. If the headways between the vehicles are consistent — that is, if they’re consistently 20 feet, like it is mostly in rush hour in urban areas — but they’re going a consistent speed, you end up with eliminating a lot of the starting and stopping that causes congestion.
If you carry this further down the path, right now in the United States, the traffic lanes are 12 feet wide, but the maximum width of a vehicle is 8 feet. There’s two feet on each side of the vehicle, so there’s four feet between vehicles. If the vehicles know exactly where they are and they know where the other one is, it doesn’t need to be four feet anymore. The space where three or four lanes could easily be covered with closer traffic lanes that allow for more to move through there in a platoon together.
Senator Saint-Germain: Thank you very much, Mr. Steudle, for a very interesting presentation. In your conclusion, you referred to the importance of working on potential interoperability issues across borders. As you know, at the end of our hearings, we have to make recommendations to the Canadian government. Could you develop more for us some of these interoperability issues that you are preoccupied with and on which the Canadian government should perhaps work harder in order to tackle these issues?
Mr. Steudle: Let me give you probably the best example I can think of. Right now in the U.S. Congress, there are two versions of bills that deal with automated vehicle deployments. One has passed the House of Representatives; another one is in the Senate right now. Those bills are trying to quantify what the role of the federal government is, meaning vehicle regulations, and what the role of state government is. As I referenced earlier, that’s drivers, operators, insurance and liability issues and then traffic management.
The current bills that are in Washington right now are wrestling with exactly the issue that you’re asking. Right now, we have 50 states. We’re attempting to kind of come together on some of our state regulations, but they’re different. The California law is very different from the Michigan law. Then there are the states in between. Florida’s regulations are closer to Michigan’s. In order to have harmonization, even within the United States, we have to get all of those in alignment so that when you drive from Michigan to California, you don’t have to stop and turn off the system in Kansas.
I think the same thing will hold true for the Government of Canada across all of your provinces. I do know that there are people all across your country that are looking at how should we do this. I think you’re going to have the same approach. While your provinces are bigger, you’re still going to have a border where Quebec and Ontario meet that where there will be a potential problem. The federal government here is attempting to try to clarify what interstate travel is like and what the interoperability has to be.
From a technology standpoint, many technology companies have said, “We don’t need any other interaction. Our vehicle is smart enough to go down the road and know what’s around it and can interface with the current environment.” That’s true, but it doesn’t lead to a very efficient system where the vehicle talks to the other vehicle that knows it’s about to stop, or it’s moving forward, or there’s an accident up ahead. That’s the connected component to it.
That’s where it becomes more difficult because there are different companies that have different ideas of how they should communicate and what should be communicated back and forth. It’s a bit difficult because you don’t want to pick a winner and loser. However, from a technology perspective, it’s operating on a communications platform that’s regulated by the federal communications commission. It gets intertwined in federal government policy very quickly.
Back in 2010, there was a safety pilot model deployment that looked at these vehicles on the connected side. This was before automated came along. On the connected side, the vehicles were talking to each other and to traffic signals. The automakers have worked out that communications protocol so that if they’re sending information, they’re sending the same type of information so it can be received on the other end.
From a technology perspective, the technology has to figure that out. But from the broader government perspective, I think the ability to smooth out some of the edges between the interoperability, meaning the traffic laws are similar, is an important step as well. It’s a bit dicey. I can’t speak specifically about the Canadian experience. In the United States, however, strong proponents of the states have the right to set the rules exactly as they want them and they can prohibit something if they don’t want it. The federal government is saying, “No, you can’t do that. We have to have it all the same.” There are some interesting discussions going on with that debate.
Senator Saint-Germain: Do you believe a pilot project between, let’s say, one Canadian province and one or two American states would be helpful in the perspective of better harmonization?
Mr. Steudle: Absolutely. In fact, I would welcome an opportunity to continue a collaboration between Michigan and Ontario. We do a lot of things with Ontario. As I said before, the state of Michigan has more in common with Ontario, frankly, than it does with many of our other states.I would welcome an opportunity to continue that collaboration between Michigan and Ontario. We’ve got three international crossings that could be very helpful.
The Deputy Chair: Before moving to the second round, Mr. Steudle, I want to pick up on this question of pilot projects that the senator just mentioned.In January of this year, the U.S. Department of Transportation announced 10 designated pilot sites to encourage testing and information sharing of AV technology. One of those sites was designated in Michigan. Can you give us an update on what progress, if any, has been made on this?
Mr. Steudle: That’s the American Center for Mobility at Willow Run, and it is a 500-acre site that, as we speak, is about probably four weeks away from having the high-speed loop, which is a 2.5-mile high-speed loop that has three level interchanges that the vehicles drive through. It has entrance and exit ramps and a tunnel that curves so that as the vehicle approaches the dark tunnel all it sees is a dark hole. It can’t see the light at the end of the tunnel because it’s around a curve. It has concrete and asphalt pavement and different types of guardrails. That high-speed loop will be open for testing at the beginning of December.
At the same time, they’re working at other multiple environments that include a multi-lane boulevard that goes under that high-speed loop. There is a multiple-lane bridge that goes over the top. There is what they call the NHTSA intersection, which is a very specific design that the federal government has that vehicles have to be able to manoeuvre through. That’s in the process of being built as well.
As it develops further, there will be off-road testing, flat surface testing that allows for different manoeuvres to be done on a very flat surface, and then there is a campus that is being built as we speak. There are buildings being put in place that will allow companies that are there on a temporary basis to rent out a stall that is approximately 20 feet by 80 feet. Then in the spring much larger buildings, which will be for the foundation members, will be starting as well.
From a funding perspective, the state of Michigan is funding the big part of the construction. Since it was a challenge grant from the Michigan Strategic Fund to the American Center for Mobility, it came up with the private funds to match and they’ve secured founders.
There are seven founder positions. AT&T signed on early, as did Toyota and Ford Motor Company. There are three others that are about to be announced. I can’t tip my hand here until they do their press releases about joining as well. And they’re in final negotiations with the seventh one. Those seven founding members have each put up the equivalent of about $5 million apiece to help this for the first three years. They’re starting to work now on the next level, which is the sponsorship level, which I think is in the $2-million range, and there are approximately 20 of those coming as well.
Now, another component to ACM is the education component, and just this morning the governor will be signing an agreement with the 17 universities, colleges and community colleges down at ACM to collaborate on their curriculums and develop the stem education back at the middle school, high school, college and the community college levels to develop the technical training that’s required for, say, a mechanic who ultimately has to work on these vehicles. And then, of course, at the university level, there will be the education component for the engineers and the computer programmers as well.
That consortium is getting signed today where all 17 of these educational institutions are coming together to collaborate instead of compete, because so many times we see different colleges and universities that compete with each other for the same students, and this is an attempt to use government money and funding more efficiently.
The Deputy Chair: Thank you very much. That’s a very comprehensive answer.
Senator Galvez: Extending on my first question about the goals of the use of connected vehicles, I understand that the goals are related to efficiency and safety. I can see that if we do mass transport and if all these vehicles are all standard and very similar, because you’re going to be more efficient with electrification, mass transport, security, and trucks for transport, but are you saying that these companies — Toyota, General Motors, Chevrolet — will start building the same type of vehicle and not a variety of models? You’re laughing, so please tell me why.
Mr. Steudle: That’s a great question. I live just outside Detroit, so I have a long family history in the auto business. I chuckled when you talk about all the different models and that’s what differentiates them all. I think those models will be here for a long time. We have to go much further down the road where we end up with one utilitarian pod that is just going down the road. Maybe we as a society will ultimately choose to spend our money doing something else, but that’s yet to be decided.
I remember reading an article from Bob Lutz, who is a long-time auto executive at Ford, Chrysler and General Motors, in which he talked about his vision. He’s in his upper 70s, I believe. I hope I didn’t say something wrong about his age. He had a vision that looked similar to what you’re talking about, that 20 to, say, 40 years away the transport system will look different and you will have some kind of a pod. The question becomes: Is it a customized pod that’s just yours or are you using one that everyone else uses? That’s a question we as a society will have to figure out. However, I do think there will be cars for a long time.
As you referenced, we also think about General Motors and Toyota or Ford. Just last week there was a new company from Ann Arbor, Michigan, that was testing an autonomous shuttle that has six to eight seats and was driving on the streets of the city of Detroit. It was taking employees from an employee parking garage to their office headquarters, which was about a mile away. It ran from seven o’clock at night to ten o’clock at night, taking employees back and forth. It went on a fixed route, so it wasn’t what we call “out in the wild.” It knew the streets and where it was going.
I believe that shuttle was built by Polaris in Minnesota. It meets all the federal standards, and it was in live traffic. It can go a maximum of 24 miles an hour, but big-city traffic goes maybe 15 or 20 in the core urban area. It moved passengers from one area to another. It had no steering wheel and no brakes. It had a person, a safety ambassador, in the front seat. If they needed to interact with the vehicle, they could, but to my knowledge they didn’t have to at all. They moved passengers back and forth.
The comments that I read from some of the media reports were that people were very comfortable with that vehicle. They didn’t feel that it should have stopped when it didn’t or that it should have accelerated. They felt comfortable with the way the vehicle operated.
I offer that as a suggestion. Here is a small, start-up company that may be bought by somebody bigger, but it’s the technology piece that’s fitting on top of a shuttle. This is the “be learners” piece that I talked about. It’s about now that you have that, how does it solve a situation in the future?
One of the biggest issues with transit systems is the first mile and the last mile. You get off at the station and you’re still a mile and a half away. If these types of shuttles can move people in those areas, it helps solve that problem. It also helps, I think, with rural mobility. I think we can take that technology and expand it into the rural parts of our state, or in a country, or frankly, the entire world, once it’s mapped. You have to know where everything is located. It’s on kind of a fixed route, but I think the technology can get us to where we provide much more mobility for seniors and for people who are disabled and are maybe stuck in their homes.
Senator Bovey: I want to go back to the testing. I’m intrigued with the facility that you say you’ve set up with the off-ramps, on ramps, and so on. I had the opportunity to ride in one of these vehicles. We were riding in one on a road where there was a line only down the middle and not on the side. You talked about lines on both sides and that we won’t need four feet between cars. What about the many roads in this country and yours that just have a centre line? That’s part of the testing question.
Coming from Winnipeg, we have a season that you could call pothole season. You talked about different surfaces. I wonder about potholes, because there are many in both our countries, and weather. Again, coming from Manitoba, a lot of winter transport up north is done on ice roads. Going forward, I wonder if these usual but perhaps less usual scenarios are being looked at.
Mr. Steudle: They absolutely are. There is one other facility I didn’t talk about that is called Mcity. It is at the University of Michigan in Ann Arbor. It was created and opened in 2015. It’s a 35-acre site that has 30 different attributes that a vehicle will encounter on public roads. It encounters a roundabout, dirt roads, a railroad crossing and urban settings in very small segments.
To your question about winter testing, a year and a half ago, Ford Motor Company did the first winter testing in Mcity with an autonomous vehicle. It was one of their first prototype vehicles that had LiDAR sensors on it. They had it out in a snowstorm at night travelling those roadways. The learning was exponential. They really learned about LiDAR, how the snow packs on it, how it messes up the signal and how it couldn’t see a pavement marking.
One of the key pieces of an automated vehicle is to think of it as multiple checklists. Does it have a GPS map? The computer is saying, “Does it have a GPS map? Does it know specifically where it’s at? Does it have the LiDAR function? Does it have forward-looking cameras and rear cameras? Does it have the radar systems and all the different systems on the vehicle all the way down to can it see the paint lines?” As it goes through and starts checking things off, it feels comfortable and says, “I know where I’m at and I can move forward.” If half of those boxes aren’t checked, it may not have the comfort of where it’s at and it may not move. Maybe I’m making this number up. Maybe there are ten boxes and it says, “Only two boxes are checked. We’re not going to go anywhere.” I’m not an automotive engineer and I don’t want to pretend that I am, but you get the point. There is a checklist that says do I feel comfortable moving or not.
To your point about traffic paint lines, here is a perfect example. I’ve pushed this nationally, even four years ago with Google, namely that every state should just have a good paint line. In Michigan there are 122,000 miles of public roads, but 50 per cent are gravel. They will never have a paint line.
I talked with a friend at the Ford Motor Company who works in this particular space, and he said, “We’ve got that figured out. We can see the edge of a gravel road without a paint line pretty good.” The technology has advanced in a couple of years so that they can see the terrain that’s around them.
I live in a subdivision that has no centre line. It’s a paved road but there is no centre line down the road. I live off of a county road that has no edge lines on the side. I know exactly what you’re talking about.
That’s what the legislation that passed in Michigan in 2016 did. It opened all of those roads up for testing. So now you’re not depending upon doing it just in a test facility. You can now take your vehicle out on a two-lane road that only has a centre paint line or a gravel road and you can see how it operates.
Senator Bovey: Thank you.
The Deputy Chair: Thank you, Mr. Steudle for your participation today. The Senate of Canada appreciates you taking the time to speak with us.
Honourable senators, I am pleased to introduce our next witnesses from the U.S. Department of Transportation, starting with Mr. Finch Fulton, Deputy Assistant Secretary for Transportation Policy, who also serves as a member of President Trump’s transition team focusing on transportation and infrastructure policy. Mr. Fulton is accompanied by Mr. Nathaniel Beuse, Associate Administrator for Vehicle Safety Research at the National Highway Traffic Administration.
I want to thank both of you being for being us with this morning. Congratulations on the fifty-first anniversary of the U.S. Department of Transportation, which was this past Sunday. I invite Mr. Fulton to begin his presentation, and afterward senators will have questions for you.
Finch Fulton, Deputy Assistant Secretary for Transportation Policy, United States Department of Transportation: My name is Finch Fulton and I am the Deputy Assistant Secretary for Transportation Policy here at the Department of Transportation.
I would also like to introduce Mr. Nathaniel Beuse, Associate Administrator of the National Highway Traffic Safety Administration’s Office of Vehicle Safety Research. Nat has a wealth of experience with this department, including having worked for years on the federal autonomous vehicle guidances we have before us. Nat is going to be able to answer questions about the background and history of how we got where we are today as well as provide any technical assistance we may need.
It is our honour to represent Secretary Elaine Chao and the United States Department of Transportation to help inform the committee’s study on regulatory and technical issues related to the deployment of connected and automated vehicles. We thank the Standing Senate Committee on Transport and Communications and Chair Dawson for the invitation to participate in today’s proceedings.
Before we discuss the department’s current approach and framework for connected and automated vehicles, I would like to highlight how important the relationship between the U.S. Department of Transportation and Transport Canada is to Secretary Chao. As you all know, the U.S. Department of Transportation and Transport Canada signed their first agreement more than 45 years ago, and the Trump administration is committed to continuing this cooperation and information sharing, especially as it relates to safety and emerging innovative areas like automated and connected vehicles as well as with our air traffic control systems. In fact, one of the first foreign officials Secretary Chao spoke and met with was Minister of Transport Marc Garneau. In addition, in March I had the privilege of joining Secretary Chao, Minister Garneau and other senior U.S. and Canadian officials to see first-hand the success the Canadian government has had transferring the air traffic control system to Nav Canada. We are very happy to in some small way return some of the hospitality and collaboration and all the efforts you have already shown us, so thank you for that.
We also recognize that there has been active collaboration between Canada and the United States with connected and automated vehicles. I was thrilled to see these efforts reflected in the White Paper on Automated Vehicles in Canada recently put out by the Canadian Council of Motor Transport Administrators. We have and will continue to work with our Canadian counterparts to ensure that North American stakeholder concerns are represented on the international stage.
At the U.S. Department of Transportation, safety is our highest priority. The market introductions of connectivity, improved camera and LiDAR and radar systems, augmented and virtual reality, machine learning developments and other vehicle technologies have rapidly accelerated in recent years. It is these systems that will form the basis for higher levels of automation, and we expect global market entry very shortly of automated vehicle systems that start to perform more and more of the driving task.
You have heard many times the sobering statistic that human error is the cause of over 90 per cent of vehicle crashes. In 2016, there were 37,461 people killed in crashes on United States roadways. Current crash avoidance and automated vehicle technologies have the potential to help drivers avoid these crashes.
This past September, the U.S. Department of Transportation launched a new federal guidance to support and encourage the growth of automated vehicles, Automated Driving Systems 2.0: A Vision for Safety. The guidelines build on previous policy and incorporate feedback received through public comments, including those from our international partners and Congressional hearings. A Vision for Safety paves the way for the safe deployment of advanced driver assistance technologies while providing voluntary guidance that encourages best practices and prioritizes safety. The guidelines also include technical assistance to states and best practices for policy-makers. The guidelines specifically address the safe deployment of Automated Driving Systems’ automation levels 3 through 5, which correlate to conditional, high and full autonomous system categories. The guidelines also contain 12 priority safety design elements for consideration, including vehicle cybersecurity, human-machine interface, crashworthiness, consumer education and training and post-crash ADS behaviour.
This voluntary guidance also provides a flexible framework for industry to use in choosing how to address a given safety design element, thus keeping the door open to private-sector-driven innovation. In addition, to help support public trust and confidence, A Vision for Safety encourages entities engaged in testing and deployment to publicly disclose Voluntary Safety Self-Assessments of their systems to demonstrate their varied approaches to achieving safety. You may have heard that on October 12Waymo became the first company to publish their voluntary safety assessment.
In addition, A Vision for Safety supports the wealth of new opportunities to transportation that ADSs provide for persons with disabilities by encouraging consideration of the unique needs of this population in the design, development and the policy surrounding autonomous vehicles.
Over the coming months and years, the United States Department of Transportation, the National Highway Traffic Safety Administration and other federal agencies will continue to take a leadership role in encouraging the safe introduction of automated vehicle technologies into the motor vehicle fleet and on public roadways, working in collaboration in the areas of policy, research, safety standards, freight and commercial use, infrastructure and mass transit.
At the direction of Secretary Chao, the moment she released A Vision for Safety, the department began the process of developing a holistic department-wide framework to facilitate the adoption of surface automated driving systems that will include the National Highway and Traffic Safety Administration, the Federal Highway Administration, the Federal Motor Carrier Safety Administration, the Federal Transit Administration and the Maritime Administration.
The United States Department of Transportation is committed to facilitating and furthering automated vehicles while maximizing beneficial safety, mobility and efficiency outcomes. We will also seek opportunities to reform regulations that may hinder these innovations unnecessarily. We look forward to continue working with our Canadian counterparts to advance automated vehicles, including interoperability issues across borders. We are embracing a future where automated vehicles can help avoid crashes; improve mobility options for the elderly and persons with disabilities; and enable new economic opportunities, where the cost of fuel and time spent commuting is dramatically reduced, where millions more people have access to affordable and reliable transportation, and most importantly, where highway fatalities and injuries are dramatically reduced.
Thank you once again for having us here today. We’ll be happy to answer any questions you may have.
Senator Eggleton: Thank you for your presentation today. You talked a bit about accidents and fatalities and how, in future, we hope to avoid a lot more of them with automated and connected vehicles, but let me ask you about this incident that occurred in May 2016 involving a Tesla vehicle and a truck. The investigators found that the truck driver’s failure to yield the right-of-way and the car driver’s inattention due to overreliance on vehicle automation were the causes here. What has been done about this? We’re seeing, in this decision, a concern about overreliance on vehicle automation, and yet we’re heading down a path where we will get more and more vehicle automation as we head towards level 5 and less and less attention by a driver if circumstances arise similar to this. What have you done with respect to learning the lessons from this particular unfortunate fatal accident in terms of public policy?
Nathaniel Beuse, Associate Administrator for Vehicle Safety Research, National Highway Traffic Safety Administration, United States Department of Transportation: Certainly that particular crash highlights two things that Mr. Fulton talked about. One is in the guidance document released by Secretary Chao. There is an element that focuses on human factors. This is the overreliance issue that you mentioned. This is really the future for L3, L4 and L5 vehicles, the ones where the vehicle really is responsible for performing the driving task versus something like the level 2 vehicle that you mentioned, where the driver is responsible for performing the driving task.
We’ve done two things from a policy perspective. One is to ensure that our research program is focused on looking at these issues, but the other is also working directly with the vehicle manufacturers to make sure this is an area that they’re paying attention to.
With respect to how that system is captured and governed, it is different than how we’re thinking about the higher levels of automation. Certainly we’ve been on record saying that there are certain elements of the guidance that could be applied to such a system like that, and that human factors issues are different, but companies need to focus on them regardless.
That report just came out from the National Transportation Safety Board. We’re looking at our program right now to see if there are additional changes we want to make.
Senator Eggleton: Could one of those changes be a requirement that there be somebody who is still behind a wheel and ready to take over? And at what level would there be enough confidence to be able to forgo that requirement?
Mr. Fulton: Part of that is going to be on the levels of automation of each automobile. Auto manufacturers are competing on safety. They all want to have the safest product that people understand the best. How they educate the public on the actual technologies of their vehicles will go a long way to the success of their product. They know they have to sell people accurately the level of safety and attention they need to take, so there may not need to be an actual requirement. The car companies know this is one of the prime things they will be competing on.
We don’t want to be heavy-handed in regulating specifically how exactly they should do each step. We want it to be performance-based. They know that they need to have the highest level of safety and they have to communicate the exact requirements to the drivers as much as possible. This is something they’ve been telling us they need to focus on and something we will help them focus on.
Senator Eggleton: Governments have to be concerned about safety and security. In your opening statement, you said that a vision for safety encourages entities engaged in testing and deployment to publicly disclose voluntary safety self-assessments of their systems that demonstrate their varied approaches to achieving safety. That follows on from what you’ve just said.But where does voluntary end and government requirements come in? Where is the dividing line there? Governments have to be concerned about safety and security of people. Yes, you would hope that the industry would as well, and they’re competitive and they want to have safe vehicles. But where do you draw the line between what is voluntary and what would be government safety regulations?
Mr. Fulton: This is a good point to point out that in our guidance we put forth, NHTSA retains all enforcement and recall authorities. If something unsafe is on the road, NHTSA can still take complete action to recall those or take steps for safety. What we don’t need to do is proactively prescript what we think the approach should be. We should be ready to jump in where we think it’s necessary and not where we can hinder innovation.
Senator Griffin: Thanks for your presentation today. It’s very interesting. I understand that in September, your department released an automated vehicles policy and that you received input for it during the congressional hearings and other means from the public. What were the major themes of that feedback that you received during this process?
Mr. Beuse: There were a couple of key themes. One is the positive outcome where folks appreciated the voluntary approach, the flexible framework that provided for deployment of these vehicles, in particular, the testing of them. It gave companies a way to talk about safety beyond saying, “We’ve driven 5 million miles and we didn’t have an accident.” It was more like we have these concrete areas of safety that we can focus on. So that was a really positive outcome. Even in the sense that it was voluntary, there seemed to be a recognition in the community that this approach is one that they can support and also one they believe is important in explaining to consumers how they’re addressing safety.
Some of the more negative comments were more on clarifications and things that needed to be changed in some of the words to make it clearer. One element that’s worthwhile talking about is ethics. Originally, we had proposed ethics as one of the safety elements that companies should consider. Those elements originally came from a whole bunch of work we had done, that the Transportation Research Board in the United States had done and other international partners, and no matter who you talked to about self-driving vehicles, ethics was always one of the things folks would talk about.
As we put that out there and received comments back, all the commenters said ethics are important, but we’re not sure what the standard is to say what is ethical and what is not ethical. There wasn’t guidance, something for the company to rely on to answer that question for themselves or for the public. Some argued that it might send industry down a path where that’s not how vehicles are programmed in the first place. That’s one we removed in this latest iteration. That’s an example of one where it just wasn’t ready for adoption.
I would say another key area was to make it very clear, and Mr. Fulton touched on this, that it’s the responsibility of the federal government to regulate motor vehicles and motor vehicle equipment here in the United States and the responsibility of states to regulate the licensing, insurance and other aspects of motor vehicles. In our earlier document, while we attempted to make that very clear, there were some opportunities that we missed, frankly, so in this latest iteration we tightened those up and really actually added a piece to focus on state legislatures. A lot of the comments that came in said, “That’s great that you’re focusing on the department of motor vehicle folks, but you really missed an opportunity to talk directly to state legislatures as they’re considering legislation in this area.” So we’ve added some guidance and best practices for them as well.
Senator Bovey: Thank you. I want to build on the question my colleague Senator Eggleton asked and your responses. I certainly appreciate that you’re not wanting to drive the research and development, and I respect that. But I guess as we take a look at the cross-border activity between Canada and the United States at a time when our federal government is looking to sort out what are the appropriate regulations at a federal level versus the responsibilities at a provincial level, I want to dig a little deeper on whether you would recommend that the Canadian government have voluntary guidelines as opposed to mandatory regulations. I appreciate the flexibility that comes with voluntary, but I’m also concerned for public safety. At some point, voluntary surely has to become regulatory or regulated.
Mr. Fulton: Certainly the approach we’ve taken is that we’ve built upon each year iteratively the guidance that we have with and work with our stakeholders in Congress, in the industry and the public writ large to make sure we take steps iteratively so we can help everyone get on the same page.
We feel like for us it would be premature to put forth too many mandatory regulations in this area. There will be a great deal of work and a great deal of conversation going forward, but at this point in time, we don’t feel comfortable trying to make mandates that will affect these technologies decades out when they’re still in their infancy and there’s still so much more to learn and so much more to do. We want to shape the direction and the path forward, but we do not want to mandate the path forward because we may be choosing wrong parts of the technology or wrong paths forward that may not be the best outcome.
Mr. Beuse: If I can add to what Mr. Fulton just said, I think one of the things we’ve found in the U.S. and the nature of the technology is being very clear on the specifics of exactly what we’re talking about. Currently there are no L3 vehicles on the road, which is the focus of the guidance. The guidance is focused on vehicles that are not yet here. Even the first manufacturer who has publicly said they’re going to produce an L3 vehicle keeps pushing that date back. Originally it was supposed to be this year; now they’re saying next year.
The issue that the senator brought up before about — that’s a very specific, we’ll call it, issue with L2 vehicles, the sort of driver confusion. How much am I responsible for the driving tasks? Are the icons big enough, et cetera? That’s a Human Factors 101 issue versus what we’re talking about with L3, L4 and L5 vehicles, where we’re talking about the software driving the vehicle, cybersecurity issues and vehicles that might have alternative designs. Those areas, as Mr. Fulton mentioned, keep changing. Every six months someone comes up with a different version of what was out the previous six months. If you look at Waymo as a company as a whole, when they started out, they were supposedly building an L3/L4 highway vehicle that morphed into a low-speed vehicle that now they’re saying we’re not even doing a vehicle; we’re just going to sell software.
Those are some of the considerations we’re looking at. Rather than prejudging where everybody is going, it’s more like how do we get people to focus on safety and have safety be part of the conversation in addition to all the great stuff they’re doing because of the potential? I think it’s worth reiterating that we’re not giving anything up — we still have all of our tools and enforcement ability — and we don’t have to wait for a large number of crashes or any crashes at all to actually use that authority. We’re very cognizant of our role in overseeing the safety of these vehicles.
Senator Bovey: Very quickly on the ethical and legal side, the other side of safety is cybersecurity and the safety of data. Can you address that briefly for us?
Mr. Beuse: Sure. In the U.S. we’ve taken an approach on cybersecurity that involves a lot more collaboration with vehicle manufacturers. In the U.S. over the past 12 months there’s been a couple of activities probably worth noting. One is that the auto industry has formed what’s called an information sharing and analysis centre. This essentially started out with a group of light vehicle manufacturers that has now grown to over 40-some entities that are sharing information about cyber events in an anonymous fashion. So it’s not “GM saw this”; it’s “Company A saw this.” There’s dramatic learning that happens within that environment, and that’s borrowed from other sectors like banking, energy, electrical grid,et cetera. It proves to be a successful model to capture when cyber events happen and to have a rapid response from the industry.That same group has also developed a set of best practices for the industries that are in that collection.
We’ve actually mirrored that and published our own set of best practices that we’re now in the process of determining how we will finalize. This is focused more on the design aspects of the vehicle. At the same time, we strongly encourage SAE to develop a set of industry standards for cybersecurity, and that work is also well on its way. That first standard published late last year. We’ve taken an approach to identify what the issues are and have industry work towards solving them.
On the data side and data privacy in particular, we think that’s an area where in the U.S. there is a clear delineation of responsibility between us at the Department of Transportation and a sister agency called the Federal Trade Commission. The Federal Trade Commission is largely responsible for privacy issues in the United States, including whatever theft would happen from a vehicle, but we’ve said we’re responsible for safety and they’re responsible for the privacy aspects. I hope that answers your question.
Senator Galvez: Thank you very much. We’ve been hearing a lot of witnesses saying that we have to leave the technology and the research relatively free so that new things come, as long as everybody has safety in mind. I sort of agree with that.
As more pilot tests and more exercises and demonstrations take place, and now with the fact that some government offices like yours, like the DOT, are coming up with examples of evolving policies, I was wondering, are we going to have for a while passengers who are going to be covered in a different way? We are saying that this mode of transport will reduce crashes and be safer. Are you saying some people will have some type of insurance and other people will have another type of insurance? Will the driver’s licences and the permits to circulate on the highways, the regulations, be compatible or incompatible and for how long? What do you expect?
Mr. Fulton: I think this is one bit that was actually eloquently put by somebody that previously testified before your committee. I don’t have his name off the top of my head, but he was with the Arizona State DOT. The way he put it, and I thought it was pretty straightforward, was when he buys a car, he gets insured. If his wife drives his car and gets in a wreck, it’s still his insurance that covers the vehicle and he’s still responsible, even though he was not in the car and she was driving it.
That seems to be a fundamental way that we can look at how our insurance can work on a state level. Of course, in the United States, insurance is a state issue, so this is something that we will be working with our state and local partners to make sure we’re on the same page. It’s part of the conversation that we’re already having, and it’s part of the best practice of policymakers that we’re putting forward.
Senator Galvez: Okay. I’ll come back. Thank you.
The Deputy Chair: Before we go to the second round, I’d like to ask you about vehicle-to-vehicle communications. Late last year the U.S. Department of Transportation issued a notice of proposed rule-making for the establishment of a new federal motor vehicle standard to mandate vehicle-to-vehicle communications for new light vehicles and to standardize the message and format of vehicle-to-vehicle transmissions. I wonder if you could give us an update of the status of this proposed rule. What actions have been taken in this regard?
Mr. Fulton: That is certainly still pending. Part of the issue is when you get a new administration, you have to set up a number of different people in a number of different offices. This is one that we are coordinating with the Office of Science and Technology Policy, the Federal Communications Commission, and there are a number of different stakeholders at the table.
This is one where we certainly have not lost sight of it. It’s under way. There are ongoing discussions about the future of that NPRM and about the future of these technologies.
Senator Eggleton: I understand that the department has announced 10 pilot sites to encourage testing and information sharing of AV technology. Can you tell me how that’s coming along? Are you working with any Canadian counterparts in that regard?
Mr. Fulton: The 10 proving grounds that were put forth obviously are still testing routinely. They coordinate with each other, with us and provide information to us.
What I think we’ve learned coming in is that the testing is not limited in any way to these proving grounds. There are tests running all over the country and in different ways. I think we even saw a study that said there has been $80 billion spent on testing and developing these technologies over the last two to three years. What we’ve learned is those 10 proving grounds can prove to be a useful tool, but we should in no way limit our research and operations to just those 10 proving grounds.
Do you know anything about any Canadian testing in our partnerships on that?
Mr. Beuse: On the Canadian front, typically the way we’ve coordinated with PMG and those folks is through Transport Canada. Right now my office is involved in a study where we’re looking at our capabilities and what capabilities we think going forward, and we’ll be sharing that, as soon as that study is done, with Transport Canada sometime next year when it’s supposed to be completed.
The Deputy Chair: Before we finish, I want to bring up one question in regard to what Senator Eggleton referred to. Those 10 sites have been designated, with one in Michigan. Earlier, we had people in here from Michigan we were speaking to. Are they all replicating the same tests? Are those 10 different sites operating fully independently of each other in terms of the types of tests they’re applying?
Mr. Fulton: They are fully independent. They do work to coordinate with each other. They routinely have phone calls and they email. They have no benefit to duplicating work, so they’ve been doing that in a decentralized manner. It hasn’t been the DOT that says, “You do this research.” It is very much up to each proving ground.
Mr. Beuse: To add to what Mr. Fulton said, within Michigan, if you look at something like Mcity and something like the other proving grounds that are in the area, there are certain capabilities that some have that are different from the capabilities that the others may have. Maybe in stark contrast you can think of a city environment versus a high-speed environment.
What we’ve seen happening, even at the test facility that I’m responsible for in Ohio, is they are expanding because there is an overwhelming need to capture different aspects. We’re really talking about maybe a level-four or level-five vehicle as opposed to a level-two or level-three vehicle. These are the more advanced vehicles that have yet to come and the recognition that there needs to be some additional laboratory capabilities in addition to whatever might be accomplished on simulation and road testing, which is very important. I cannot underscore that enough.
Senator Eggleton: I want to follow up on that. Are these pilot sites operated by industry or are they university-based research kinds of facilities? What’s the characterization of them?
Mr. Beuse: It’s all of those and then some for the 10 that were under the previous designation. You have cities involved, normal test facilities involved. It really spans the gamut.
The Deputy Chair: I’d like to thank Mr. Fulton and Mr. Beuse for their participation here today.
Honourable senators, for our meeting tomorrow, we will hear from representatives of Uber.
(The committee adjourned.)