Self-Driving TechJust about every major player in Silicon Valley and the auto industry has efforts under way to develop autonomous car technology. Not only will self-driving cars change the way the construction industry builds roads, but some of that same tech will power the autonomous heavy equipment of the future.

In this episode, an update on who’s leading the autonomous charge from Alan Ohnsman, a senior editor at Forbes. Then, Ellice Perez from Waymo, formerly the Google self-driving car project, explains how equipment manufacturers can put LiDAR technology to work. And finally, what the advent of autonomous vehicles will mean for the roadbuilders of the future with Steve Vozar, the CTO of May Mobility.

 

Subscribe to the AEM Industry Advisor for regular updates in your email about industry news and insights. 

Subscribe on your favorite app: 

iTunes   Stitcher   Google   TuneIn  

 

Alan Ohnsman:

I would anticipate that deployments of autonomous vehicles for agricultural use, construction sites, mining, will be on a faster path than self-driving cars and trucks.

Dusty Weis:

Hello and welcome to another edition of the AEM Thinking Forward Podcast, advancing the equipment manufacturing industry.

Dusty Weis:

I'm your host, Dusty Weis, and in this edition, self-driving cars and the important technological takeaways for the heavy equipment industry.

Dusty Weis:

An update on who's leading the autonomous charge from a senior editor at Forbes, from the Google self-driving car project, Ellice Perez explains how equipment manufacturers can put LiDAR technology to work, and what the advent of autonomous vehicles will mean for the road builders of the future with the CTO of May Mobility.

Dusty Weis:

Some big ideas and big names in this episode, but it's these sorts of big ideas we work to bring you here on the AEM Thinking Forward Podcast.

Dusty Weis:

Each month, we explore a new subject area to help keep your business on the cutting edge of the industry. If you haven't yet, make sure you subscribe to our podcast feeds so that you get an update every time there's a new episode.

Dusty Weis:

We're always trying new things here on this podcast, and we'd love to hear what you think of it. Please do post a comment, rate, or review us in whatever your favorite podcasting app is. Not only does it let me know what's on your mind, it helps other industry pros like you find our show.

Dusty Weis:

To this month's topic, there was a time when self-driving cars seemed like a far-fetched super futuristic science fiction story, but these days, you're not a major player in the auto industry if you don't have some sort of autonomous technology in development.

Dusty Weis: 

Ford, GM, Toyota, Uber, Lyft, Google, these are just a few of the major players working in this space, and with that much automotive and technological expertise focused on the problem, some predict that major breakthroughs are right around the corner.

Dusty Weis:

Undoubtedly, these experts have autonomous technology insights to share with the equipment manufacturers considering their own self-driving tech, but the introduction of self-driving cars is going to reshape cities and transportation infrastructure as well, which could have particularly important implications for the construction equipment industry.

Dusty Weis:

And so we'll begin our discussion with the senior editor for Future Mobility at Forbes, Alan Ohnsman, who's been covering this topic for more than 20 years.

Dusty Weis:

Alan, welcome to the AEM Thinking Forward Podcast.

Alan Ohnsman:

Hi, Dusty, thanks for having me.

Dusty Weis:

This notion of self-driving vehicles, in the public's perception, I think that every single person that considers this question has a different idea of what qualifies and what doesn't, but for the purposes of level setting, what are we talking about when we say self-driving cars or autonomous technology?

Alan Ohnsman:

Well, that's a very good question, because it gets a little blurry at times, but when we talk about self-driving, we're talking about a vehicle that can operate without human intervention under almost all normal road conditions.

Alan Ohnsman:

The society of automotive engineers has set levels that are used, and we talk about levels four and five when we're talking about self-driving vehicles.

Alan Ohnsman:

A level four vehicle suggests a vehicle that can operate in known areas without human intervention at any point under most reasonable driving conditions.

Alan Ohnsman:

Level five is the next level up, and that means the vehicle can drive anywhere, anytime, under any conditions. That's much further out.

Alan Ohnsman:

Right now, when we talk about self-driving vehicle technology that's heading for deployment, we're really talking about that level four space. Level five would be a level rivaling, or perhaps even exceeding human capabilities, and no one knows exactly when that's going to show up.

Dusty Weis:

Now, a lot of different companies are in this space, and some are more vocal than others about touting their accomplishments as being a self-driving or a fully autonomous car. Who's really winning in this space right now?

Dusty Weis:

And on that scale of zero to five that you mentioned, what level are they really at? Is anybody really competing at a level four autonomous right now?

Alan Ohnsman:

Yeah. There are a couple companies that are moving into that space. Obviously, the best known would be Waymo, which most people would know as the Google self-driving car project. It's the oldest and by-far the most mature.

Alan Ohnsman:

It's in its 10th year, and it has now advanced to a standalone unit within Alphabet, the parent company of Google, and by every measure, that particular program has racked up more on-road test miles, more simulated miles, and they currently operate a limited robo-taxi project in suburban Phoenix, and are generating small amounts of revenue.

Alan Ohnsman:

For the most part, those vehicles are maintaining a human safety driver at the wheel just in-case, and they are operating at what would be that SAE level four.

Alan Ohnsman:

I've been in the vans several times, twice in the last six months, and at no point did the safety driver need to intervene, so that would be the real measure. How are they actually handling real world conditions?

Alan Ohnsman:

To be fair, they're also testing in a place that is not as complex as San Francisco, or New York City, or Los Angeles. It is a suburban area with nice streets and excellent weather, and not quite as much pedestrian activity, but they do show a lot of progress and continue to improve month-by-month.

Alan Ohnsman:

Waymo is also planning to get into the heavy trucking space using their technology on autonomous Class 8 semi trucks, but they would certainly seem to be the program to match or to beat right now.

Alan Ohnsman:

There are some very intriguing other ones that are coming on. General Motors Cruise unit. Of course, you have a company like Tesla, and Elon Musk certainly makes fairly bold claims about the sophistication of Tesla's technology, but we don't have a lot of hard verification, unfortunately.

Dusty Weis:

I was going to say, from the way he talks, it sounds like they're winning this race, but by all indications, at least what's been seen publicly, they're still in the level three autonomous arena. Right?

Alan Ohnsman:

That's correct. Yeah, the autopilot system that Tesla uses is in that level two, level three, in which the vehicle can handle driving under fairly standard highway conditions, and can even make lane changes, or even take an exit, but a human must be paying attention and be ready to take over when circumstance get a little too complicated, and it's just very hard to evaluate Tesla's claims, because they are on a different development path, and what they are trying to perfect is slightly different than what the other major programs are working on.

Dusty Weis:

My ears perked up when you mentioned Waymo, because a little bit later in the program, we're going to talk to Ellice Perez from Waymo about their efforts to adapt LiDAR technology to the construction and agriculture fields.

Dusty Weis:

Of course, that's very, very pertinent to our members who build the equipment that you see on construction sites and in farm fields, but what other opportunities do you see for these different autonomous technology to sort of cross-pollinate from self-driving cars to autonomous heavy equipment?

Alan Ohnsman:

I think, that space, autonomous heavy equipment for both agricultural and industrial uses in construction, that is probably going to be on a faster path. A lot of the public activity centers on self-driving cars and trucks, but that will take longer to perfect, and there are regulatory considerations that have to be resolved.

Alan Ohnsman:

In the cases of the ag industry, construction, and mining, there really aren't any barriers to introducing the technology. One of the companies is a big player in the equipment space here, would be Velodyne, and Velodyne is one of the companies that first commercialized LiDAR for autonomous vehicles.

Alan Ohnsman:

A big customer for many years, Velodyne's LiDAR has been caterpillar, because LiDAR gives you this very detailed 3D 360-degree image if your surroundings, and it's a very useful sensor.

Alan Ohnsman:

It's ... Most companies find it absolutely essential for autonomous technology, but we've already had applications of some of the sensors going into these industrial uses, and there are a number of companies that are working on applications of autonomous tech.

Alan Ohnsman:

Honda at CES in Las Vegas this year, Honda doesn't make heavy vehicles, but they had some very intriguing, very small sort of modified ATVs using autonomous tech. You could make a little wagon train with these robotic autonomous vehicles that would carry your supplies and materials if you're a farmer, or if you're doing construction work.

Alan Ohnsman:

There's just no barrier to introduction of a technology like that in the way there would be for public roads, so there are going to be a lot of exciting applications for self-driving vehicles, because you don't have the same legal or regulatory hurdles at a mine, or at a farm, or at a huge construction site, that you would on a highway, or on city streets.

Alan Ohnsman:

It's a faster path to getting the technology out there.

Dusty Weis:

Whether we're talking about roadways, or construction sites, or farm fields, there are still a number of hurdles that autonomous technology has to clear before it's widely adopted.

Dusty Weis:

Some of them are technology, and some are a little bit more societal in nature, but let's start with the technology. Where is the technology? Not quite there yet.

Alan Ohnsman:

On vision systems combined with computing power, there has been remarkable progress over the last few years, but it has to get better in terms of the ability to instantaneously interpret all of the information that's flowing in.

Alan Ohnsman:

I mean, there are so many terabytes of data that are being collected by these vehicles because the vision system for an autonomous vehicle would include LiDAR, which we mentioned earlier. There's all sorts of digital cameras, and the vehicles are collecting vast amounts of visual data.

Alan Ohnsman:

We mentioned Waymo. Waymo also has microphones. You get auditory data so you can hear emergency vehicles coming by. There's a lot of information flowing in that then has to be processed very, very quickly, and that takes enormous computing power, and the computing power continues to increase whether it's the system that Waymo has designed, whether it's NVIDIA's self driving computer system, just remarkable amounts of data processing that are happening.

Alan Ohnsman:

And humans, we have our eyes and our ears giving us sensory information when we're driving. Our brains are pretty good at interpreting that, and it turns out that's not a simple thing to replicate.

Dusty Weis:

Yeah, it turns out we've got thousands of years of experience of interpreting inputs and making decisions in split-seconds, and computers just aren't there yet.

Alan Ohnsman:

That's right. AI is actually built to do this, to replicate the same way in which a human brain interprets and reacts to sensory information, and so that's a long way of saying that a lot of progress has been made, and autonomous vehicles, under certain tight circumstances work today, but they're not as good as humans yet, and it will be a little while before that happens.

Alan Ohnsman:

That's why we're seeing kind of a slower roll-out of autonomous vehicles, and we're probably, maybe two years or so, away from really solid reliable level four capability that day in and day out can contend with the basic obstructions and hurdles that we experience as human drivers.

Dusty Weis:

What about the social hurdles, then? And I bring this up because I feel like in the first few years of this journey into the autonomous future, safety has been a huge deal in this space.

Dusty Weis:

We've seen driverless vehicles rack up thousands and millions of miles, and very few fatalities so far, but when there is a death involving one of these self-driving cars, it's a huge deal, it's all over the news.

Alan Ohnsman:

Yeah.

Dusty Weis:

Are people more critical of self-driving technology because it's new? Because, let's face it, 10s of thousands of Americans die in human-operated vehicles every year.

Alan Ohnsman:

And there have been some studies looking at this, but our tolerance for mistakes, and accidents, and fatalities caused by machines is not what it is for human drivers. We will be more forgiving of human error than we will with machine or computer error, and this has already happened.

Alan Ohnsman:

We have had fatalities. Uber, last year in Tempe, Arizona, a test vehicle struck a pedestrian at night time. That absolutely should not have happened. There was a human safety driver. The human safety driver does not appear to have been paying attention at the time, but there was also system failure, so both things failed in that case, and that was unacceptable.

Alan Ohnsman:

That will remain an issue for a time, because this is a new technology, and most people are not yet exposed to it, have not ridden in self-driving vehicles, and the idea can feel a little intimidating if you've had no first-hand experience with it.

Alan Ohnsman:

And until this technology is more broadly available, more familiar to people, I wouldn't expect that to go down anytime soon.

Dusty Weis:

Of course, autonomy faces similar safety challenges as a job site solution in construction equipment. You had mentioned before that it has the advantage of functioning within a closed environment, or at least a more controlled environment than out on the open road.

Dusty Weis:

Which do you think we'll see first? Widespread adoption of self-driving cars which have a little bit of a headstart technologically right now, or widespread adoption of autonomous construction equipment?

Alan Ohnsman:

I would actually for for the ladder. I think that, already, there are applications and uses of autonomous technology that could be commercialized for industrial purposes.

Alan Ohnsman:

For many, many years, we've had AGVs, automated guide vehicles, inside of factories. I mean, I first saw one 20 years ago at a Honda plant in Ohio. Many factories, Amazon facilities, all sorts of factories that already use very sort of simple autonomous vehicles.

Alan Ohnsman:

And so yeah, I would look for a lot of activity in that space over the next two years, and I would certainly expect that that will be available sooner than you or I purchasing our own self-driving car.

Dusty Weis:

I think when you talk to most people, they still see self-driving cars true level five autonomous technology as this sort of far-fetched concept, something that we won't necessarily see in our lifetimes, but there are a lot of major players investing billions of dollars in this technology right now.

Dusty Weis:

What would you say is a realistic timeline for self-driving cars, and or heavy equipment to come into widespread use?

Alan Ohnsman:

Right now, you'll read some articles that say there was too much enthusiasm too early, and that this is a harder thing to solve.

Alan Ohnsman:

And that's true, but I think many of the major programs were talking about deployment of the technology in the early 2020s, and growing from there, and I think that still sounds right.

Alan Ohnsman:

As far as personal vehicles, our own cars and trucks that we own and drive, I'm not sure about that. I mean, Tesla claims they're going to have it soon, and we'll see. I think for most major OEMs, that will be slower.

Alan Ohnsman:

I think fleet deployments are definitely the path to look to over the next few years, vehicles that are well maintained and well monitored by the operators versus personally owned vehicles.

Alan Ohnsman:

Maybe don't look for a car that you can buy that has level four autonomy until maybe the mid 2020s, but I think fleets will certainly be available and on the road in advance of that.

Dusty Weis:

Well, it's a fascinating look into the future here. It's been real pleasure tapping into your expertise in this field.

Dusty Weis:

Alan Ohnsman, you're the senior editor for Future Mobility at Forbes. Also a good follow on Twitter, by the way, @AlanOhnsman.

Dusty Weis:

Alan, thank you for joining us on the AEM Thinking Forward Podcast.

Alan Ohnsman:

Great, thanks Dusty.

Dusty Weis:

You heard there about the potential this technology has to cross over into the construction and agricultural equipment fields. We now get to explore that in detail with one of the leaders in the autonomous driving space, Ellice Perez.

Dusty Weis:

She's the head of Laser Bear, a subsidiary of Waymo, Google's self-driving car project. Ellice, welcome to the program.

Ellice Perez:

Why thanks, Dusty. Nice to be here, thanks for having us.

Dusty Weis: 

Everyone knows about Google, but for those who may not be as familiar with Waymo, this is a project that's been going on for around 10 years now. It's considered a front-runner in the race to full-blown autonomous cars.

Dusty Weis:

Catch us up on the history a little bit here. What's Waymo's goal with this, and what are the crowning achievements so far?

Ellice Perez:

Chauffeur started a project, one of Google's original moonshots about 10 years ago here in the building where we're talking to you from.

Ellice Perez:

We were addressing some of the statistics that we've seen. Every year, 1.3 million fatalities happen in cars worldwide, and 90% of those are due to human error. There's a lot of lives that we shouldn't be losing, and so we started Waymo as Chauffeur 10 years ago, addressing this very problem.

Ellice Perez:

Waymo's mission is to help people and things move around safely, and we have been working on this, and now have accumulated over 10 million miles in our autonomous vehicles.

Dusty Weis:

One of the great debates in the field of autonomous technology, right now, sort of splits it into two camps. There are those who say that it will depend on a technology called LiDAR, and those who say it won't.

Dusty Weis:

This is going to become an important debate in the field of heavy equipment as well, as we have more companies rolling out autonomous product lines. I'm pretty sure it's a safe bet that you land on team LiDAR.

Dusty Weis:

What is LiDAR, and why is it important?

Ellice Perez:

LiDARs send out electric pulses which bounce off of objects, and really sense and feel the world in 3D, and so a camera, which is kind of like the human eye, we can't see as well at night, and nor can a camera, and the LiDAR basically feels things.

Ellice Perez: 

The LiDAR can feel what the world looks like, and then with the data that we capture from the LiDAR, we can classify objects, and then use that information to then connect it with software to make predictions, et-cetera, et-cetera.

Ellice Perez:

Yeah, so certainly, at Waymo, we are very much in favor of our LiDAR, and our LiDAR, what we sell at Waymo, we have a vertical field of view of 95 degrees, which is unparalleled in the industry. We have multiple returns for shot, and we have a minimum range of zero, so our LiDAR has been proven over the 10 million miles that we've driven over the 10 years, and we're really excited because the interest that we're seeing from the industries, including some of the AEM members, has been fantastic.

Dusty Weis:

And it's exciting because it allows you to essentially feed more data into an autonomous vehicle as it's going down the road. Not just the visual data, but the spacial data as well, and paint a more complete picture.

Dusty Weis:

Now, Elon Musk famously said this spring, that LiDAR is unnecessary for autonomous operation, that you only need the cameras and AI, and I know that among your community, he tends to gobble up more than his share of headlines, but how do you respond to his points?

Ellice Perez:

Well, he was speaking about the use of LiDARs on autonomous vehicles, and what I work on is our LiDARs or selling to other industries, and robotics, and warehouses, and agriculture, and mining, and so, yes, we've absolutely proven, in the sensor suite that we have built for our cars and for our trucks is, the LiDAR is very, very important.

Dusty Weis:

Where are you at in the development process of Waymo's LiDAR technology there? And how is this cross-pollinating into other fields?

Ellice Perez:

We started building our LiDAR in-house in 2011 because we didn't find a sensor that fit the needs that we have with our goal of getting to L4 autonomy, and so we started to build it out, and you can imagine after various iterations, we're really excited to see it proven over 10 million miles in the last 10 years.

Ellice Perez:

Now, one of the things that we're looking at is talking to various industries, which is what we started in March of this past year, and seeing and learning more about how our LiDAR can be helpful for other industries as well.

Ellice Perez:

We know it works fantastically for our use-cases of trucking and autonomous vehicles, and we're learning, as well, the other use-cases are also great applications for our LiDARs.

Dusty Weis:

We've certainly seen that among AEM's members. I know a couple of months ago we did a podcast with Volvo construction equipment, talking about an autonomous project that they're piloting in a mine in Sweden.

Dusty Weis:

But where do you see the opportunities for construction and agriculture equipment to put LiDAR to work?

Ellice Perez:

We're already working with some customers in construction, and mining, and agriculture, and as you know, and your members know, there's a lot of debris, and foliage, and dust, and things like that, and with our LiDAR, we have multiple returns per shot, which means that we can perceive through the foliage and the dust, and that, in particular, is really interesting for our customers in construction.

Dusty Weis:

And ultimately, it's again, just about painting a better picture of the environment in which this equipment is operating, and I would imagine, increasing the safety of everybody on that job site.

Dusty Weis:

But what do heavy equipment manufacturers need to do to leverage this technology for themselves and their customers?

Ellice Perez: 

Like I said, our LiDAR just connects onto the sensor suite or the application that these companies and our customers are already using, so call us and we'll help you work through your application.

Dusty Weis:

In a lot of ways, you personally get a strong hand in shaping the future of mobility, construction, agriculture, and through that, the future of human society.

Dusty Weis:

What's the coolest part of doing the job that you do?

Ellice Perez:

That's a fun question. I think the coolest part is ... Well, I guess there's a couple of different things that are cool. I get to ride in our cars all the time. Right?

Ellice Perez:

I realize that not everyone gets to ride in an autonomous vehicle everyday, and think that that's normal, and we here at Waymo, we take our cars all the time, and they work fantastically, and it's fun to be riding in the future and thinking that the future is actually now.

Ellice Perez:

I would say the second part is really crafting and working through the growing business that we have here at Laser Bear selling our Honeycombs to the world, and really figuring out as we develop and mature our business, and learning more about the business and how we can help others make use of our LiDAR in addition to what we have learned over the last 10 million miles.

Dusty Weis:

If someone wants to learn more about you and Waymo's Honeycomb technology, where can they go?

Ellice Perez:

You would go to waymo.com/lidar. Sorry, there's a robot walking by with a ... one of our LiDARs on it. If you can hear that in the background, it's an actual use-case.

Dusty Weis:

You have the coolest office, that's incredible.

Ellice Perez:

That background noise is our LiDAR at work.

Dusty Weis:

Well, Ellice Perez, head of Waymo Laser Bear, thanks for joining us on the AEM Thinking Forward Podcast.

Dusty Weis:

Finally, with self-driving cars on the horizon, cities, roads, and infrastructure themselves may have to evolve in order to accommodate this new technology. Since AEM's members build the equipment that builds those things, it's a pretty big deal for us.

Dusty Weis:

To explore these implications, we're joined now by Steve Vozar, co-founder and CTO of May Mobility.

Dusty Weis:

Steve, welcome to the program.

Steve Vozar:

Thank you very much for having me.

Dusty Weis:

May Mobility is another big player in the self-driving technology stack, and you've deployed self-driving shuttles in cities around the Midwest. Can you tell me a little bit more about the company's history and what it is that you do?

Steve Vozar:

Sure. At May Mobility, our vision is to transform cities through autonomous technology to create a safer, green, and more accessible world, and the way that we're doing that today is by re-imagining transportation by developing and deploying autonomous shuttles that get people where they need to go safely, easily, and with a lot more fun.

Steve Vozar:

We are really focused today on the autonomous shuttle market. That is Geo-fence environments, the places that are pre-mapped, we know where there shuttle's going to go. Moving people anywhere from one to five miles around where they live and work, and the reason we're doing this today is because there is a long path towards autonomy, and we want to be able to put a product out in the word today that actually solves a need, and so that means that we not only develop the technology ourselves, and we also run the operations ourselves, which gives a lot of control over how these shuttles are used, and it makes us really confident they're going to be safe and easy to use.

Dusty Weis: 

When you talk about the shuttles that you build, you're talking about these smallish vehicles that ... I think it's five or six people can ride in? Correct me if I'm wrong.

Steve Vozar:

Yeah, that's right. The shuttles themselves are six seater electric vehicles, and they're classified in the United States as a low-speed electric vehicle, which means that they're street legal on roads up to 35 miles an hour, and they have a top speed of 25 miles an hour.

Steve Vozar:

They're a six seater vehicle, and today on public roads, we reserve one seat for our fleet attendant. They're ambassador for May, they're ambassador for the technology, and if something happens or something goes wrong, they can always take over.

Steve Vozar: 

We liken them to sort of an elevator operator. People were really scared of elevators for a long time. It was a new technology. They felt safer with a person in there with them that knew how it worked, and eventually they weren't needed anymore.

Steve Vozar:

That's how we feel about fleet attendants. They serve a really vital role today in getting people up to speed with technology, and we want to make sure that people adopt this technology, so we're okay with having an attendant onboard for now.

Dusty Weis:

This should be an interesting topic for you, as an autonomous shuttle company that focuses on that sort of last-mile transportation hurdle, but historians very often point out that the car has played a major role in the development of Americas society and infrastructure.

Dusty Weis:

It influences where people live and work, it determines how we vacation, it commands the use of vast acres of real-estate for parking lots and freeways, and the like.

Dusty Weis:

What sort of broader implications does self-driving technology have just beyond how we get around?

Steve Vozar:

I think, in my view, what I would love to see is more beautiful, greener, better land use in urban cars.

Steve Vozar: 

I don't think anybody gets excited ... Well, maybe not anybody, but I think most people don't get excited when they hear that a new parking lot is going up and they're downtown.

Steve Vozar:

That could be a restaurant, or that could be an art gallery, or that could be a park, and so what really excites me is transforming the way that cities are built to consider that autonomous technology may transform the way that that urban environment's maybe built on.

Steve Vozar:

And so we love to collaborate with people who are doing renovations in their downtown area, or building up new central business districts to talk about where we would like to see parking, and where we would like to see shuttle stops, and if people build this in, inherently, in early days, than it can be a much different type of downtown area that prioritizes pedestrians and cyclists, and deprioritizes cars, and that's really what we'd like to see in sort of the downtown central business district areas.

Steve Vozar:

In terms of broader societal implications, if we can get to a place where you don't have to concentrate on driving for an hour a day, two hours a day, in some people's cases, three hours a day, that really opens up different places where people can live. It will change real-estate, it will change all sorts of things.

Steve Vozar:

But there's also a danger that we could pollute more. We're seeing that as people live further away from where they work, and they want more and more comfortable cars, they want bigger cars, and they use more gas, and so there's dangers and risks as well, and I think we need to consider that as we sort of develop this technology and what the second order results will be.

Dusty Weis:

From an infrastructure perspective, we're all pretty familiar with what cars need to get around roads, traffic control devices, parking, emergency response services, but since AEM's members build the equipment that builds a lot of that infrastructure, I want to pay particular attention to how those infrastructure needs will change with the advent of autonomous technology.

Dusty Weis:

Firstly, how is it going to change the roads themselves? Is a four lane highway still going to be just four lanes and some white paint?

Steve Vozar:

It's an interesting thing. I think early days, a lot of self-driving cars queued their localization or where they thought they were based on the road pane.

Steve Vozar:

From vision systems they did it with LiDAR, and it makes sense because humans do this, too. But I honestly believe that the way of the future is that it's okay to sort of add extra queues to the road that are subtle, don't require a lot of work, just 3D structure to the environment that will yield much more robust localization solutions.

Steve Vozar:

And so as soon as you get a dusting of snow, for a AV that's basically going off of road lines, then it's blind to where the lanes are. And so humans have this extra layer of context. Whether it's imbuing cars with that extra layer of context, or just queuing off of different features, I think if we're ever ... if we're going to have mixed-use roads, we're going to have to see a four lane highway exist in the way that it exists today.

Dusty Weis: 

Well, and that raises this notion, then, of vehicle to infrastructure communication. V2I is the hip term for it, I believe.

Dusty Weis: 

What is that? Why is it important? And how is it going to change the way that roads get built for autonomous vehicles?

Steve Vozar:

V2I communication is a way of having infrastructure that's embedded in the world, whether it's information about a traffic light, a cross-walk, whether it has camera sensor, or a radar sensor, to be able to communicate that with vehicles that are nearby.

Steve Vozar:

And again, this is not a ubiquitous technology today, and so we develop our cars with the notion that this does not exist right now, but it may exist in the future to help as a redundant means of seeing what's going on in the environment.

Steve Vozar:

And so, I think if you're going to see lots and lots of self-driving cars out there, you're going to see V2I technology, whether it's 5G or DSRC. I'm sort of technology agnostic when it comes to that, but I think that this has a capability to help both human drivers and autonomous vehicles, and so I think it's a win-win.

Dusty Weis:

Let's use May Mobility shuttles, then, as an example of this. Your shuttles, where they're deployed. Kind of go out and they run this circuit. What do they do when they come to a stop light?

Dusty Weis:

Do they have visual sensors that see whether the light is red or green, and take their queues that way? And how would V2I change how that works?

Steve Vozar:

It's interesting that you ask that, because we actually deploy our own V2I infrastructure. It's on a proprietary network, so it's closed. It's not the open DSRC standard, because we want to have the redundancy specifically for traffic light detection.

Steve Vozar: 

We can do both onboard and off-board traffic light detection. The way we do it is completely passive so we're not tapping into the traffic light signal network. We are simply observing it visually with a camera, and then sending that data over to the car as a double check.

Dusty Weis:

And so there's basically a camera watching the stoplight, it sees whether or not it's red or green, and then it sends a wireless message to your shuttles that says, "This light is red, time to hit the breaks."

Steve Vozar:

That's exactly right, and there's more things that we can do if we were able to tap into the actual infrastructure, we can get more granular signal phasing and timing information, but we really launched our own infrastructure because DSRC and 5G are not ubiquitous, and we didn't want to have to wait for that technology to be everywhere for us to deploy, and we didn't want to make our customers install it where it may not be ready yet.

Dusty Weis:

You mentioned using that data to keep track of other things on a more granular level. One of those, I'm sure, is traffic congestion, and as more autonomous vehicles enter the roadways, there's sort of two schools of thought about how they're going to impact that.

Dusty Weis:

Some say that congestion is going to get better, some say it's going to get worse. What are your thoughts, and why?

Steve Vozar:

It all depends on how it's used. Right? If you have autonomous vehicles circling a block waiting for a fare, and basically using our roads as parking lots or places to stall, then we're going to have a problem.

Dusty Weis:

It's sort of ... My ... I look at it like the prisoner's dilemma where it's ... Say I have an autonomous car, and I'm in downtown Chicago, and I don't want to pay $40 to park it, I can just program it to, "Okay, drive around the block while I'm in the store running my errands, and come back and get me in an hour."

Dusty Weis:

And that would make it worse.

Steve Vozar:

Fortunately, I think personally owned self-driving cars a long way off, and so we can get ahead of this and control it a little bit.

Steve Vozar:

We believe that shared rides are the way to go, both in terms of sharing the actual physical car so that you wouldn't even think to program it to go drive around for 40 minutes, but rather, there would be an incentive to have it pick somebody up and actually use that physical asset.

Steve Vozar:

But additionally to how people actually share rides, and put less vehicles on the road for the number of people that need to go places.

Steve Vozar:

We're viewing this closer to a public asset that people can use in different ways, rather than personally owned vehicles, and I think that that's really the direction we're going. But yeah, if people are trying to circumvent traffic laws or parking tickets by sending their car off on wild goose chases, then we're going to have a problem.

Dusty Weis: 

Sorry, I just ... I always go to the absolute most selfish setting that I can when it comes to these sorts of things, just because human nature. I like to keep that in mind as we talk about this, but if vehicle autonomy leads to more ride-sharing, that's eventually going to have a significant impact on the construction of destinations like workplaces, schools, entertainment districts, and other high-traffic places.

Dusty Weis:

How will these institutions have to evolve to keep pace with the changing habits of their visitors?

Steve Vozar:

That's a great question. I think that these institutions are in an interesting bind, because they have to plan for 20, 30, 50 years out, and they don't know what the trends are going to be.

Steve Vozar:

There's a middle school near my house that, it was built pretty recently, and I think that they must not have taken into account that change trends of parents dropping their kids off rather than taking the bus, because every morning there is congestion that spills out onto the roadway, and clearly, this was not anticipated.

Steve Vozar:

And so I think that they need to get creative about how these new modalities, whether it's self-driving cars, or shared rides, or scooters, get used around their campus, and also point that in our deployment in Providence, Rhode Island, we go by a Dunkin' Donuts.

Dusty Weis:

Oh, no.

Steve Vozar:

And traffic from the drive-through spills out into the streets during rush hour. And so these are the kinds of urban planning things that we need to think about, and once it's built, it's almost too late.

Steve Vozar:

We need to think about ways that we can get around that.

Dusty Weis:

We've looked, at a pretty granular level, at the impact that self-driving technology could have on urban centers, but getting from where people live to the downtown of cities, you're basically going to be on a freeway at some point or another, and self-driving technology has implications for freeways as well.

Dusty Weis:

One of those that I've heard debated online, and I want to get your take on it, is this notion of lane width, and if you can program cars to follow a very narrow track down the middle of a lane, you don't need to make the lane as wide to account for human margin of error.

Dusty Weis:

What does that mean, then, for the future of road building?

Steve Vozar:

That's an interesting question. I think that self-driving cars today need to know their location on earth within about 10 centimeters, and humans don't.

Steve Vozar:

It's just I don't know where I am within 10 centimeters, especially longitudinally. Laterally, I have a pretty good idea. Longitudinally, I could be a mile off, doesn't really matter. There's sliding scales of where you need to be localized.

Steve Vozar:

Just the way that some of the limitations of how self-driving cars are programmed, today, they really do need to know where they are with a pretty high level of precision. It doesn't necessarily mean that they need to be on rails, that they are going to drive over the same location over, and over, and over again, and in-fact, we can program them not to do that. We just need to know exactly where they are.

Steve Vozar:

Two things. Over time, I believe self-driving cars will loosen the restriction on how precise they need to know their location, and I also believe that even if we narrow the lanes, they can have a sense of where the road is worn, and how to even that out.

Steve Vozar:

And so I think that we're not going to see these sort of ruts in the road.

Dusty Weis:

From cars just sort of running over the exact same spot over, and over, and over again?

Steve Vozar:

Exactly.

Dusty Weis:

You think that you can program cars to account for that and keep from wearing the road straight down to the gravel underneath?

Dusty Weis:

Now, that's a fascinating notion, because if in the future we're able to narrow the width of lanes by 75-80%, all of a sudden, that opens up, potentially, more lanes of traffic using the same amount of space.

Steve Vozar:

Yeah, absolutely. However, I will say that, if you have a completely self-driving car highway, then you don't really even necessarily need lanes. They can communicate with one another, then we can do a lot of interesting things with evening out the wear on the road.

Dusty Weis:

Oh, my god. You just gave me so much anxiety with that notion, because I'm just trying to picture in my head thousands of cars crossing over the same space, and none of them in any lane whatsoever.

Dusty Weis:

But you're right, if they're all communicating with each other, that could potentially work.

Steve Vozar:

Yeah. I mean ... And you can imagine flexible highways, so if you need to maximize the number of vehicles that are in a wide highway, then you stack them five wide, and you do that during rush our, and then during lower peak times, then you stagger where the cars are going, and they can change lanes on the fly, and they don't even need to know what the nominal lanes are.

Dusty Weis:

Steve Vozar, the co-founder and CTO at May Mobility. If someone wants to learn more about you or what you do, where can they go on the web?

Steve Vozar:

They can visit us at maymobility.com, our our Twitter is may_mobility.

Dusty Weis:

Well, Steve, thanks for taking the time to talk to us. We really appreciate it.

Steve Vozar:

All right, thank you very much, and a pleasure.

Dusty Weis:

Steve, I should add, was one of the keynote speakers at AEM's Thinking Forward event at the Henry Ford Center in Detroit, in May. No surprises there, he crushed it.

Dusty Weis:

But if you haven't been to one of these events yet, do yourself a favor and make the time. There are three of them yet this year each, an excellent opportunity to do a little bit of professional development.

Dusty Weis:

The next one is September 10th at the Cisco Innovation Centre in Toronto. IoT, big-data, intergenerational workforces, a behind the scenes tour at Cisco, all on the docket for that one.

Dusty Weis:

There are also events in October and November, each with a different panel of experts carefully curated by AEM's education team. Milwaukee and St. Louis are the locations for those, but you've got to reserve your seat now.

Dusty Weis:

All the details and the sign-up forms at aem.org/think. That is going to wrap up this edition of the AEM Thinking Forward Podcast. For more valuable industry insights, make sure that you've signed up for the AEM Industry Advisor, our twice weekly e-newsletter. Visit aem.org/subscribe.

Dusty Weis:

If you need to get in-touch with me direct, shoot me an email at podcast@aem.org. The AEM Thinking Forward Podcast is brought to you by the association of equipment manufacturers, and produced by Podcamp Media, branded podcast production for businesses.

Dusty Weis:

Visit podcampmedia.com. Little Glass Men does the music for this program, and for AEM. Thanks for listening, I'm Dusty Weis.

×