Infrastructure Technology Podcast: The Mobility Data Revolution

• State DOTs are increasingly trusting of integrating crowd-sourced mobility data as it becomes more robust and comprehensive, offering greater temporal and spatial coverage than traditional sensors.
• New data types like origin-destination and volume data are emerging, but agencies must evaluate their credibility before relying on them for planning and funding decisions.
• Vehicle probe data, now collected from roughly one in three vehicles, offers significant insights into traffic flow, energy use, and emissions, especially valuable in high-density urban areas.
• Despite advances, rural areas still face challenges due to limited data penetration and connectivity, but emerging connected vehicle technologies may help close that gap.

What if your car’s data could help shape the future of transportation? In this episode of the Infrastructure Technology Podcast, Harlee Hewitt speaks with three leading mobility data experts about the evolution and future of crowd-sourced transportation data. Michael Fontaine of the Virginia Transportation Research Council discusses how mobility data has progressed and what it means for state DOTs. Then, Stanley Young and Jeff Gonder of the National Renewable Energy Laboratory join to share insights into how real-time vehicle probe data is transforming everything from energy efficiency to rural infrastructure monitoring. Tune in to hear how mobility data is shaping the future of infrastructure planning and public transportation.

Gavin Jenkins (00:00):

And welcome to another episode of Infrastructure Technology Podcast, the ITP brought to you by Endeavor, Business Media, Roads and Bridges, and Mass Transit. I'm Gavin Jenkins, your host and Senior Managing Editor of Roads and Bridges. And with me as always, we have the jack of all trades, Harlee Hewitt, Associate Editor of Roads and Bridges, and we also have the man, the myth, the legend coming from Cleveland, Ohio. Brandon Lewis, Associate Editor of Mass Transit Magazine, and we have a special episode for you today. We have two interviews from our jack of All trades, Harlee Hewitt, and before we dive into her first interview, let's get to know Harlee just a little bit more. But before we do that, let's recap what we already have learned about Harlee. She is a 25-year-old, right?

Harlee Hewitt (00:56):

Just turned 26, but keep going.

Gavin Jenkins (00:58):

Just turned 26. Oh my goodness. Alright. She is a five foot 11, 26-year-old former athlete from Oklahoma. She played softball, volleyball, basketball. She also has musical talents. She is a very good singer. I was about to say we've never heard her sing, but we are going to try to get her to sing a theme song. She has to write the lyrics. At some point we will get that done. Until then, we're just going to be using the free audio that we found online as our theme music, which I like. I think the Groove Funk music.

Harlee Hewitt (01:40):

Absolutely.

Gavin Jenkins (01:41):

Yeah. I wanted the cyberpunk dystopian beats and Brandon wanted heavy industrial metal music and we agreed on Harlee's love of funk and yeah, everybody loves little funk. So Harlee, tell us the story about how you landed this job because it has to do with music.

Harlee Hewitt (02:14):

It does. Well, I went to school to be a technical writer and then once I exited there, my dad--who just tiny bit of background on my dad. The day that he left high school, he jumped on a bus and toured the country with his hair metal band. This is 1985 for those listening. And then from there on he continued. He never really stopped. And about two years ago now, he was in an establishment we won't name on the podcast with our, just happened to be Endeavor Business Media's HR leader, Candia Field was in the audience. They started talking, they became friends. And pretty much one time that she saw him, she mentioned that there was a position open that would fit me. She also went to my university. So that was another connection--Go Pokes. And then she offered to bump me up in the resume line and here I am.

Gavin Jenkins (03:31):

It's all about who you know.

Harlee Hewitt (03:32):

All about who you know as they say.

Gavin Jenkins (03:36):

It's also about the power of hair metal.

Harlee Hewitt (03:41):

Yes, indeed. Actually it's strong. It's a strong force.

Gavin Jenkins (03:46):

Right? So earlier this year I tasked Harlee with finding some stories for Roads and Bridges' October issue, which is our technology issue. And when we were searching them together, we were looking at pane; discussions and lectures at conferences we didn't get a chance to go to. And one of the panel discussions that we both were really interested in and wish we had been able to attend was about mobility data. And this hits both of our audiences. So the mobility data ended up being a story for our October issue. It's called the Mobility Data revolution. But what Harlee did was really impressive to me because I just wanted her to reach out to one of the gentlemen on the panel and see if they would write a story. That's what happens with Roads and Bridges. We get submissions. So a lot of times I'm reaching out to people saying, Hey, can you write this?

(04:47):

Can you write that? And Harlee ended up booking a meeting with the entire panel from this conference. And the entire panel consisted of Stan Young, Michael Fontaine, Jeff Gonder, Michael Pack, and Sean Turner. And so one morning earlier this year I get on a call with this entire panel, we start rehashing the entire panel from the conference and they just ended up agreeing to all write the story for our magazine where they each took little snippets of what they said at the conference and just turned it into a 500 or 600 word section for a feature story. It was the cover story for our October issue. And they also agreed to be on this podcast. And so today we have interviews with Michael Fontaine and Jeffrey Gonder and Stanley Young. And first up is Michael Fontaine, and let me tell you a little bit something about him. He's the associate director at Virginia Transportation Research Council. Without further ado, here is Harlee's interview with Michael Fontaine.

Harlee Hewitt (06:00):

Welcome to the Infrastructure Technology Podcast. My name is Harlee Hewitt, Associate Editor of Roads and Bridges Magazine. Today I'm joined by Michael Fontaine, who is the Associate Director for the Safety Operations and Traffic Engineering team at the Virginia Transportation Research Council, which is the research division of the Virginia Department of Transportation. In this role, he manages VDOTs research program across a broad spectrum of areas including traffic control devices, emerging data sources, highway safety, intelligent transportation systems and connected and automated vehicles. Prior to becoming associate director, Michael served as a researcher at VTRC and at the Texas a and m Transportation Institute. Michael also currently serves as the chair of the Transportation Research Board Urban Data and Information Systems Committee, and was previously a member of the TRB Traffic Monitoring Committee. He was a registered professional engineer in Virginia. Michael, thanks so much for joining us and we're happy to have you on the program.

Michael Fontaine (07:05):

Thanks, Harlee. I appreciate you all having me on the call or on the podcast.

Harlee Hewitt (07:09):

Awesome. So if you would, Michael, just give us sort of a 30,000 foot view, if you will, of what mobility data is and what it's doing specifically at VDOT.

Michael Fontaine (07:25):

Sure. So mobility data is something that transportation agencies have collected for a long time. It's just we used to always be reliant on infrastructure based detectors, so loop detectors or radar units put on the side of the road. And I think what's really interesting about the data landscape today is that there is a lot of data providers that are mining crowdsource data, whether that be from cell phones and location-based services or fleet automatic vehicle location systems in order to give us data about speed, travel time, origin, destinations of people, and even down to the level of individual vehicle paths that are going on there. So it's become a very rich data set because of this crowdsource data that's available today.

Harlee Hewitt (08:14):

I see. So 15 years ago or so there was a private sector probe and travel time data that was used widely amongst transportation agencies, but sort of face challenges with validation in its use. So how has that kind of changed over time and how is mobility data viewed nowadays?

Michael Fontaine (08:38):

Yeah, so I would say 15 years ago, I wouldn't say that the crowdsource probe data was widely used. I'd say it was starting to be used by a lot of agencies, but as I think many of the listeners probably are aware of state dots tend be very conservative in their approaches and so they're very cautious. And so when these new data sets came on board, people really wanted to make sure that they were really representing conditions on the roadway because if you're using this data to make funding decisions about which projects to move forward with or if you're using it to provide real-time traveler information to people, you want to make sure it's accurate and credible and that you're making good decisions and you're not going to get a black eye with the public and telling them something that's not correct. So a lot of the initial efforts when these data sources first became available was really trying to establish credibility.

(09:35):

Could you use it and was it believable? And so what used to be the I-95 Transportation Coalition, now the Eastern Transportation Coalition really did a lot of work to try and establish trust in these data sets where they collected some benchmark data to show that it really was representing travel and speed and they identified areas where the data sets didn't work as well because we're relying on essentially people driving through the network with their cell phones on or large trucks that are using fleet AVL systems for report positions and then mining that to estimate speeds and travel times. And so what we've seen is over the last 15 years, the data providers have become much more robust and they've improved their analytical methods, the quality has really improved. And so for example, in the early years of these data sets, they really had problems with signaled arterial roads where you had a lot of severe congestion. So maybe it took you two or three cycles to get through the traffic signal and the methods they were applying were massively overestimating how fast people were moving on these roads because they thought that these cars slowly inching through the red light, were pedestrians walking on the side of the road with their phone.

Harlee Hewitt (10:56):

I see.

Michael Fontaine (10:56):

And so they were too aggressive in filtering out some of those methods, but over time the data vendors had really gotten a lot better at that. And so over the last 15 years, I think the speed and travel time data has been widely accepted by agencies and is very commonly used for us. But I think one of the things that we've seen recently is the spectrum of data that's available from the private companies has really expanded beyond speed and travel time. So now we have vendors that are selling volume data that are selling origin destination data that are selling waypoint information where we can track essentially individual vehicles driving through the road system.

(11:41):

So those datas haven't necessarily been validated as well by agencies. And so I think a lot of these new emerging data sets really are kind of being looked at the same way the speed and travel time data was 15 years ago, where do we trust the data enough to use it to make project decisions and evaluate projects. And so that's where I think there's a lot of interest right now is really the spectrum of the kinds of data available has really expanded a lot, especially in the last five years with the advent of a lot of more advanced AI analytic techniques to deal with these large

Harlee Hewitt (12:16):

Services. Sure. So along those lines, would you say then that something we cover a lot is infrastructure based sensors. So would you say then that this mobility data, you're speaking of the spectrum, it offers a much wider view maybe than those can necessarily?

Michael Fontaine (12:37):

So I would say the great thing about infrastructure sensors is that you get very detailed, precise data, but it's only at that specific location where you've got that detector. And so one of the issues with these more crowdsourced private sector provided data sets is that you've got much larger spatial coverage and temporal coverage. So rather than putting a temporary counter out for 48 hours, you're collecting data potentially 365 days a year, 24 hours a day. But the challenge though becomes that since we are getting essentially a sample of people driving through the network that have cell phones, then if we're talking about maybe 5% of the vehicles are reporting data back to us, if we're driving at 2:00 AM on a low volume rural road, the chances of having any data there are also going to drop off.

(13:34):

So we have the potential to really cover the entire network 24/7, 365, but in reality it's a function of how many cars are on the road and things like that. So we do a pretty good job on the higher volume roads as we get down to lower volume facilities, the data can become a little less reliable, but that ability to really have that comprehensive network coverage, especially in a big urban area offers a lot of benefits. And one of the challenges because there's still data that we can't do with the crowdsource data, if I want vehicle classification data, I want to look at trucks and how many axles are out there and be able to separate it out a tractor trailer, I can't get that from the crowdsource data. I need to have some sort of infrastructure sensor out there that can look at the configuration of that vehicle on the road. But for certain things it's very powerful and really provides us with a lot more information than we can get from a traditional sensor.

Harlee Hewitt (14:29):

I see. So what does that mean then for that comprehensive data set that you're able to get now? What does that mean for transportation agencies you might be hoping to maybe improve safety on their buses or other modes of transport? How does that data help there?

Michael Fontaine (14:48):

So I think one of the things helps us is first of all with some of the new data sets that are out there, we can get true or we can get an estimate of origins and destinations of travelers. And so that can be very useful for us if we're planning a public transportation line, where are people coming from and going to? So where do we need to provide service to get people from their home to work or shopping or something like that.

(15:16):

One of the other things that's really been interesting and is emerging for us is that agencies have always had a hard time counting or estimating travel for a lot of our vulnerable road users, our bikes and pedestrians. And so there's data providers out there right now that are again leveraging things like cell phone apps in order to estimate the degree of pedestrian bicycle travel. And so that's going to be important for us if we're looking at where do we need to improve our facilities to provide access to something like a subway station or a bus stop or something like that. As well as looking at where we might have potential for dangerous interactions between HighSpeed traffic as well as where we're getting a lot of those vulnerable road users out traveling right next.

Harlee Hewitt (16:06):

So before we were talking about 15 years ago compared to now and how the data has progressed. So what do you think with those newer forms is being done to kind of ensure trust in it or what can be done to ensure trust in the new data?

Michael Fontaine (16:25):

So that's a great question. One of the challenges we've got right now is that when you're talking about speed or travel time, there's ways to go out and collect a ground truth benchmark of how fast traffic was going to show that the crowdsource data is working right. And even the same thing with traffic volume. If I'm trying to count cars and I'm getting an estimate from a private company that's looking at cell phone data to estimate volume, I can do that. Now, one of the challenges we have though is if I'm using cell phone data to look at origins and destinations in a metropolitan area, how do I really benchmark that there is no ground truth? It's not like I can go out and track every individual. Even if you could, I'm not sure you'd want to really determine that. And so one of the challenges we have is that there's this data that's very useful for us for transportation planning and programming purposes, but there's not necessarily a great ground truth for us to compare against.

(17:27):

And so some of the challenges that we see now is agencies want this data but they don't a hundred percent trust it. But there's really a matter of getting agencies to understand this data is fundamentally different than speed and travel time. And so we can compare it to our old traveler surveys, which we know have errors in 'em, and we can compare vendors to each other to look at degrees of consistency. But ultimately that's really one of the challenges I think a lot of us are trying to solve is how do we really show that the data is useful and consistent and providing reasonable results and the absence of anything that could be sort of looking at ground truth. So a lot of agencies were trying to build this up in different ways. So we're looking at okay, maybe we can really get true origins and destinations on the corridor level and we can use that to assess it there. We can look at degrees of consistency with census data or the national household transportation survey. And so we're not necessarily able to say, yes, this is good or accurate, but we can say, is it consistent? And by doing that it provides an opportunity for us to show, okay, vendors have a place in the data ecosystem for some of these use.

Harlee Hewitt (18:45):

So say that you are able to kind of prove that consistency and usefulness of it. What does it look like then to have kind of wide scale use of that mobility? How does implementation work?

Michael Fontaine (19:01):

Yeah, so that's something where it's really interesting for agencies that are just beginning to use these data sets because you're moving from a situation where if you're using infrastructure based sensors, the agency owns all that data and they can do whatever they want with it. Whereas these data sets that you're buying from a private company, there's a data use agreement that limits how you can or cannot distribute the data. Then to some degrees it may be a black box to the agency, there's something that comes out the back end that tells you, okay, this was the speed or the travel time, but it's not always clear to them how that was generated. And so one of the things that we see here is that getting the users and the agency to really understand the characteristics of the dataset that they're buying so that they're using it for appropriate applications is very important for us.

(20:00):

And then the other challenge is these data sets are huge. We're talking about potentially we purchased in my state six months of connected car data that gave us vehicle traces and it was almost two terabytes. And so getting an agency to have staff with the right skills and expertise as well as computing resources to analyze and manipulate the data to get actionable information out of it is something that agencies have to think about or are you leveraging consultants or academia to help you with some of this? And then another thing that we see that's a potential challenge is that an agency may contract with a provider to provide a particular type of data and you run potential challenges with vendor lock-in at that point. And so at some point if an agency starts building a lot of processes around that particular dataset, it becomes potentially cost prohibited to switch to another vendor even if the other vendor has attributes or data down the line that might be more beneficial to the agency. So how do you create your own internal agency systems to answer the questions you want to answer without feeling like you are locked into a particular vendor for the next 20 years and you want to have something that hopefully is vendor agnostic and how you're adjusting these kinds of data sets. So those can all be challenges that you see from the agency perspective.

Harlee Hewitt (21:30):

I see. So with that in mind, a big glossing question for you, what is the future of mobility data from your perspective? What do you think it's best applications are? What are you excited for? Share some of that.

Michael Fontaine (21:47):

Yeah, so I think one of the things that's going to be really interesting for us is I think we've got sort of two competing forces that are happening. We're seeing really an explosion in vehicle connectivity out there and there are vendors out there that are leveraging the connected car data as a way to generate a lot of very detailed information, which is tremendously huge, but also can potentially really offer a lot of opportunities for us from an analytical perspective to answer transportation questions. Now on the other hand, if you look at this data, there's potential privacy concerns in the data that's being generated and resold. And so we've already kind of seen this in some cases where privacy regulations have impacted the kind of data that's available from location-based services. And so while the connected car data is emerging and potentially very rich, how are the privacy regulations going to impact what's available to agencies to make those decisions?

(22:52):

Now the other thing that I think is really potentially beneficial for us is the size and complexity of the data has been a barrier to your average transportation agency in using these in the past. And I think that the emergence and maturation of machine learning and artificial intelligence tools is really providing a lot of opportunity for us where hopefully the application and use of these data is going to be easier by agencies, but the development of AI is also improving the quality of the data we're getting from these vendors as well. And so what we're seeing is a lot of the vendors are using AI-based training tools in order to improve the quality of their estimates. And so I think it is helping us on data quality, but it's also helping us in terms of utilization of the data as well.

Harlee Hewitt (23:47):

Got it. Well, with that, I really want to thank you Michael for being on with us today and thank you for your insights.

Michael Fontaine (23:57):

Alright, thank you Harlee. Appreciate the invitation to join today.

Harlee Hewitt (24:01):

Thank you

Gavin Jenkins (24:04):

And welcome back to the ITP, that was Harlee Hewitt's interview with Michael Fontaine. Harlee, really good job on that. What did you find to be the most interesting thing? Because he's talking about the impacts on the agencies.

Harlee Hewitt (24:21):

Yeah, I mean I think what was so interesting to me is that he talks about how this isn't new. Mobility data isn't new by any means. It's only been in the last 15 years though that they've started to have this crowdsource data from all these third parties that are really helping improve the quality and helping drive these innovations that well we talk about on the podcast. And he talks about trying to break through to these transportation agencies, some of which haven't traditionally been as forward thinking in terms of moving on to the next thing. Sometimes the mentality is if it's not broke, don't fix it. So he talks about establishing trust in the data and how a can actually really improve our transportation, which is always something obviously that we're trying to do. And just hearing him talk about how they're trying to do that, it's definitely a long road ahead, but also we're seeing the results of that data coming to light. I think one example are these kind of smart corridors that we keep hearing about, that we've covered those types of things and even EVs and such would not be possible would without the data that they're collecting and dishing out to these different agencies. So that's what I found.

Gavin Jenkins (26:01):

It's an exciting time for mobility data for sure. So up next we have your interview with Stan Young and Jeffrey Gonder who leads the mobility behavior and advanced powertrains group in the Center for Integrated Mobility Sciences at NREL. And Stan is a professional engineer and PhD who serves as the mobility innovations and equity team leader at the NREL, which is the National Renewable Energy Laboratory. Stan is also the chief data officer for the Eastern Transportation Coalition. And when we come back after that, we'll get to hear Brandon's take on today's episode. But without further ado, let's go into Harlee's second interview of the show

Harlee Hewitt (26:57):

Welcome to the Infrastructure Technology podcast. I'm Harlee Hewitt, associate editor of Roads and Bridges magazine. Today I'm joined by Jeff Gonder and Stanley Young. I thank you both for being on the program today.

Jeffrey Gonder (27:14):

Thanks for having us

Stanley Young (27:15):

Thank you Harlee, looking forward to the discussion.

Harlee Hewitt (27:20):

Thank you both. So I would appreciate it if we could start maybe with just an overview of today's topic. So mobility data, where it is today, maybe a bit about where it's come from, how did we get here, so and how is it becoming integrated today?

Stanley Young (27:41):

I'll take that one. I think I'm on here because I'm the old man and historian of the group. I'll tell a few stories and then hand it over to Jeff. It kind of is a narrative. Mid two thousands I was taking, I took a new job at University of Maryland Center for Advanced Transportation Technology and prior to that I was with the state DOT. I've always been in technology and data, but this was the first time I had started working in the operations space and I was kind of amazed when I got there that the eyes on the road, so to speak, were highly limited. This was pre smartphone days, this was the days that were going from two G to 3G. The philosophy at the time we're talking 2005 was to put as many sensors on the road as you could. There were several initiatives to do that and everyone was learning that that just doesn't scale.

(28:35):

As soon as you get a sensor out there within 18 months, it probably would break and needs attention. And pretty soon you were too busy fixing sensors rather than watching the road. And even then, if you could do it, you could only watch your major interstates. And that's when the phenomenon, just a whole new way to think about it came along. It started with long haul trucking. Even in that day and age, those were expensive assets and you wanted to track them. So, they would use satellite telecom and the truck would report every 30 minutes, every 60 minutes where it was at with GPS. And that data stream started coming in and the logistics company said, Hey, we can sort of see how the interstate highway and the freeways are running. So, they took that data and started talking with the road operations folks. And as I should say, the rest is history. About 2006

(29:30):

To 2008 working through University of Maryland, we put together a program that we ended up calling the vehicle probe project. And that idea of vehicle probes is really key. Instead of monitoring traffic from some type of asset along the roadside to say, ‘Hey, these cars can talk to us’, really turned the corner. So, we put together a traffic monitoring program that spanned much of the eastern seaboard and the first generation really relied on freight movement that was reporting its position periodically, again maybe every 15, 30, 60 minutes. And then as time went by, telecoms just kept getting better and local fleets came on board and they got GPS and they reported every one to five minutes. And then everyone remembers the smartphone revolution. Everyone had one in their pocket that could talk a data language and not just an audio language. And pretty soon we had data coming from people and from cars from every direction.

(30:34):

There's various flavors of it. Even today it is rare that a new automobile is produced that doesn't have a digital tether back to its OEM. And so we have data coming in from freight, we have data coming in from smartphones, we have data coming in from cars. There's various varieties and flavors of it. And recent estimates would pretty much show that about one in three vehicles on the road today is contributing data so that we have some form of perception of how fast traffic is moving, how many cars are on the roadway. I like to tell the story, it was about 2008, the biggest app at the time was putting these little changeable message signs along the roadway saying, ‘Hey, congestion ahead’. Or ‘Hey, it's going to take you two hours to get from DC to Baltimore, Maryland’. Every state was doing this in some form or another.

(31:30):

And Maryland was just put together a statewide plan and hey, we're going to do this. We're going to do it on all the major freeways in the state of Maryland. And at the time they were still in the physical infrastructure do it with some type of radar. Lidar had yet to proliferate so as either radar or loops or something in the ground. And so they laid out a plan, this is how much it's going to cost umpteen million dollars so many years. And that's when that first Vehicle Pro project came online. I was working with the lead of that project and they were hoping that it would work, but they were highly skeptical as were many people saying, how in the world could we ever get data from vehicles to let us know what's going on the roadway? So, the first system became live in about 2008 and everybody was in this state of does this really work?

(32:23):

Is can we trust the data? So I was talking to my contact there at Maryland and he called me up one day and he said, Stan, I think we got a problem. I go, what are you talking about? I says, well, this new data feed from the vehicle pro project, it's showing a congestion point in western Maryland. For those of you not from Maryland, Western Maryland is highly rural. It's unusual that you have any sort of traffic maybe on I-70 out there during holidays and stuff. But it was showing a slowdown on a state route in western Maryland. And he goes, that can't be right. There's no slowdowns on rural roadways. And then there was a long pause and I was thinking, uh oh the thing screwed up. And then he goes, wait a minute, Stan, hold on, let me check the construction records. And there was a pause and he went back and he came back.

(33:10):

He goes, wow, that's the exact spot where they're doing a road overlay construction project that was, I can cite validations in all the things that we did. But that was the proof in the pudding saying, wow, this system picked up on road congestion due to construction in the middle of a rural area in Maryland. Something that even the state DOT had a hard time correlating its construction plans with its operations and road maintenance. We had another situation where the procurement officer I was working with at the time took a trip to Western Maryland and got stuck in a queue and had no idea why it was stuck in a queue and backed up for miles. So he just called me, he says, Stan, you've got that new system I help you procure, what the heck is going on? So I got online on the interstate and at the time it didn't tell me what was going on, but I could say, well that queue you're in if you're at this location is about three miles long and be patient, it should clear.

(34:12):

And that was that first aha moment. There's a different way of going about looking at congestion across the network. And from 2008, we've done that program three times. Every time we have more data, different data, different data aspects coming in, not only do we see now where people are at and how fast they're traveling, the data networks and the embedded computers can tell us a little bit more like, is it raining? How much energy are you using? Kind of bridging over to Jeff's comments on how that can be used in other domains other than just simple congestion management.

Harlee Hewitt (34:53):

Yeah. So you mentioned that probe data is being very key. So I'm wondering how that granular probe data can benefit energy and emissions applications, if you want to speak to that.

Jeffrey Gonder (35:08):

Yeah, so I can jump in there and Stan gave some good anecdotes on crashes and slowdowns and so forth. The granular data is useful for energy emissions backgrounds, you can tell is more my background because how vehicles are flowing and not just speeds, but their acceleration deceleration profiles correlate directly to how much energy they're consuming, what the emissions outputs of the vehicles are. And so with this type of higher fidelity vehicle data, you can get improved estimates from the energy and emissions. And so when we got into it, we would start using GPS speed traces initially from travel surveys that were starting to incorporate GPS in it. And so from the travel surveys or if you could get the information just from the big data probe data providers, that gives this unique window into energy and emissions, both from commissional vehicles, internal combustion engine vehicles that are the most prevalent on the road and then also how that can differ for vehicles of different powertrains. So we have modeling and simulation tools where we can estimate if these vehicles were hybrid electric vehicles or all battery electric vehicles. How that differs

Stanley Young (36:39):

Back to that story about Maryland putting in the changeable message signs. The end story is they switched from the concept of deploying sensors to using this probe data and it's the only government program that I can say finished in half the time and in half the budget and in many sense was wildly successful. It was looking back, there was some humorous moments. One is there's a very popular traffic report from a radio station around the DC area, eyes in the sky. And again, this was late two thousands, the system was working, the Maryland State Highway administration was putting up changeable message signs across the interstates and saying, so many minutes to Baltimore, if you're driving from Washington up to Baltimore, and the eyes in the sky weren't as pleased because they were taken away some of their business. But they did notice that those big changeable message signs were causing, at least they reported, were causing a slowdown because people were slowing down to read them.

(37:51):

And we never really could confirm whether that was true or whether that was kind of a biased observation. But that was the late 2000s, that's when smartphones began to proliferate and then eventually ways and Google map today we're in 2024 and we're getting asked. It's kind of like again, change of mindset that change the mind of the operations folks, we don't need to plant sensors. Now the pavement management folks are saying, Hey, wait a minute, there's a heck of a lot of data coming off of these new cars. Can we better monitor where potholes are or where the road is rough? Similarly, operations folks saying everything is so weather dependent and we got all these weather stations throughout, can't we use some of this sensing data coming in from the cars? And those are the growth areas. I mean, do we need any more location data?

(38:48):

I am kind of a pessimist. One in three cars is enough folks, we got that nailed. Are there benefits once those cars start reporting more than just their location and speed and are there significant benefits? Jeff spoke to energy after energy. You got two things, safety and congestion. And depending on which camp you are in, you can argue, but right now a lot of people, there's always going to be applications where we can save money and get better viewability into the state of the system. Are there any game changers like they were in the mid two thousands saying, wow, I didn't realize that with this new data stream that is pervasive here is a significant impact that we can have a measurable impact, particularly in the safety realm.

Harlee Hewitt (39:41):

So kind of in line then with safety, what might be kind of some of the challenges and considerations with these applications is various applications you've mentioned.

Jeffrey Gonder (39:56):

Sure. So given a current mix of different vehicle technologies that you estimate are in your traffic, you can quantify what those energy emissions outcomes are in the current situation. And then if you have different interventions that you're adding to the road, you're changing something. Changing speed limits or adding lanes, doing a roundabout or different things, quantified before and after impacts is something that you can more reliably do. Now all the way down to knowing the details on how vehicles are driving, you can advise on routing estimate what different routes might be less energy consuming to individual drivers. You can advise feedback to have them drive a little differently and improve lower their energy use from their driving all the way to looking at electric vehicles and not just the way that vehicles are driving, but where being able to potentially track vehicles a little bit more and understand origin, destination flow patterns that much better. Planning for charging infrastructure for electric vehicles is another application that we've been involved with.

Stanley Young (41:24):

I'll add another one in there. I'm waiting for bright young kids to come along and have the next killer app with this data. That'll always be fun. But one of the issues we run into is urban versus rural. As I said, this worked really good on interstates to begin with and as we got more and more data, it lit up metropolitan areas. One of the issues that we continue to work with states on is the rural areas. There's a couple of things we're fighting. One is data communications are not as pervasive in rural areas. A lot of times there's blackouts if you ever travel the upper Midwest, particularly between the towns, you have no cell coverage. And the other thing we're fighting is that these probe vehicles, with that one in three number, we can validate particularly in metropolitan areas, but once you get out into the rural areas that what we call the penetration rate really goes down for a couple of reasons. Older vehicle stock, less communications opportunities and less freight. Most freight these days are fully instrumented and most freight gravitate to the highly traveled roads. So that is an area of, there's this whole area out there called connected vehicles and my definition of a connected vehicle is a data link to every vehicle in every location. Usually an IP link, maybe start with your smartphone. And that is still a challenge in rural areas, but one I think will be addressed in time and investment.

Harlee Hewitt (42:53):

So we've talked obstacles and limitations. What might be then some opportunities, some key opportunities you've spoken to some that you're looking forward to on the energy side and just in general for this data?

Jeffrey Gonder (43:12):

Yeah. Well per my bio, something that's been exciting for our group to do is collaborate with Google Maps on helping them deploy eco-friendly routing. And so that's been a nice opportunity to do something with real world reach with a company that's so pervasively used for routing. I guess expanding a little bit on some of the broader obstacles and limitation side of things. On the traffic speed side you have traffic counters and well-established ways of getting ground truth, but from the energy and emissions side, that's less the case. Getting that ground truth validation can sometimes be a challenge. So you definitely have to have concerted efforts to work towards that. And while the data is great for lots of different applications, including the injury applications that as we've been talking about, it's not necessarily the only source with probe data, you don't get everything that you need for comprehensive modeling. Things that you might need or a travel survey, a dedicated travel to get, understanding the purpose of trips that people are making and are they traveling with others, did they consider alternate modes? So definitely complimentary resources for data collection are still valuable and helpful for energy as well as broader transportation analysis applications.

Harlee Hewitt (44:54):

Great. Alright. Well Jeff and Stan, I want to thank you for your time today and being on the program. We really appreciate

Stanley Young (45:01):

It. Thank you Harlee. Yes, thank you very much.

Gavin Jenkins (45:05):

Welcome back. That was Harlee's interview with Stan Young and Jeff Gonder. Brandon Lewis, give us your thoughts.

Brandon Lewis (45:13):

Well you guys haven't heard much from me yet today on the ITP. I know shocker because I feel like I never shut up when I get talking, so let me bring some energy here. But no seriously, Harlee, phenomenal job. Obviously even though these two interviews here with Michael, Stanley and Jeff were more Roads and Bridges related, they do relate to the mass transit side. For Stanley and Jeff, the interview that you guys just heard, I really thought that the big thing on the mass transit side was they talked about tools to show the difference between the battery electric buses and the hydrogen fuel vehicles. Obviously at mass transit we've covered a ton about the zero-emission transition, not only in the United States but in Canada as well. And with that a lot of these transit agencies and state dots are getting different kinds of those zero emission vehicles.

(46:17):

I also thought the interesting part about that second interview was they talked about advice on routing slash less energy consumption and obviously with what those battery watcher buses and hydrogen fuel vehicles do and obviously the less energy consumption they use, the better for the environment. And then just quickly on the account, just going back to Michael's interview, what I found most fascinating about that was mobility data and all the data that is used and what agencies are using to get the origin of destination travelers for public transit. But Michael also talks to how you have to find staff with the right skills to analyze that data because they're so huge. So I found that fascinating as well. But again, Harlee, great job with those interviews. And Harlee, Gavin, do you have any other thoughts on what we've heard today?

Gavin Jenkins (47:26):

I don't have anything other than to say you done brung the energy man, as

Harlee Hewitt (47:32):

As always. He’s great for that.

Brandon Lewis (47:35):

Thank you guys. And as for the jack of all trades of Endeavor Business Media, she is associate editor of Roads and Bridges, plus the Senior Managing Editor of Roads and Bridges Gavin Jenkins, Brandon Lewis editor, a Mass, the magazine. We would also like to thank Endeavor Business Media for allowing us to host the Infrastructure Technology podcast.