Is intelligent infrastructure the key to autonomous cars? That’s the question we ask in the latest episode of The Six Five Connected with Diana Blass as we explore the use of LiDAR and smart cameras on the roads in Chattanooga, TN, home to the largest smart intersection projects in the United States. Hear from William Muller of Seoul Robotics, J Ed Marston of EPB, Dr. Mina Sartipi of the University of Tennessee Chattanooga with perspective from Daniel Newman of The Futurum Group.
Watch the video here:
Or listen to the full audio here:
Listen to the episode on your favorite streaming platform:
Disclaimer: The Six Five webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.
Diana Blass: Autonomous cars aren’t here yet, but the technology enabling the reality is quickly advancing. Just look at Chattanooga, Tennessee. Home to the largest smart intersection project in the United States. Here, the city is rolling out cameras and LiDAR in over 80 intersections, tracking the behaviors of cars and pedestrians.
Dr. Mina Sartipi: Between the intersections was a long distance, so there was a lot of crossing the street on the undesignated areas. So one of the things, was we wanted to understand where are the pedestrians crossing the street? And the other thing that the mayor’s office wanted us to look into was near misses.
Diana Blass: One of the top challenges facing autonomous cars is the unpredictable nature of our roads, which is why critics say a human touch is critical for self-driving cars to see. Could that touch come in the form of this? Sensors and edge devices, perception software, all technology that you expect to see inside an autonomous car.
Daniel Newman: LiDAR offers a depth of field that’s very unique.
Diana Blass: Now, increasingly found on the outside.
William Muller: I believe that the infrastructure being intelligent is the key missing step.
Diana Blass: But here’s the thing,
William Muller: There’s a lot of infrastructure that’s struggling, that’s aged, that’s got technology that’s 20, 30 years old.
Presiden Biden: But now we’re coming back because we came together and passed a Bipartisan Infrastructure Law.
Diana Blass: The Infrastructure Law passed under President Biden is dishing out billions for states to modernize their infrastructure. This as automakers say self-driving cars could hit the market in the next decade. Will cities be ready to adopt them? Let’s find out. It’s time to get connected to one of the nation’s smartest cities, Chattanooga, Tennessee.
When you think I’m a tech hub, you probably think of Silicon Valley. But have you been to the Silicon Valley of the South? Less than 200,000 people call Chattanooga, Tennessee home, yet the city has become a tech test bed thanks to its utility owned fiber network, which began serving customers in 2010.
William Muller: What we have is basically a 100% fiber to the premise network. It serves every home and business in a 600 square mile area, about 9,000 miles of fiber optics. And the population it serves is just under half a million people. We use it to deliver connectivity services including internet up to 25 megabits per second.
We have the fastest community-wide internet service available in the United States. And we also use the fiber optic network as the communications backbone for one of the most advanced smart power grid systems in the United States. So we manage our electric system using the fiber optic network as well.
Diana Blass: So let’s talk about the return on investment here that you’ve seen. Let’s talk about the economic value and also just the idea that you’re attracting more businesses to settle in Chattanooga as well.
William Muller: Certainly. We actually commissioned a projection of the community benefit as we were doing our business planning before we rolled anything out. And at the 10-year mark, actually at the five and the 10-year mark, after the deployment, we went back and said, “Okay, well this is what we projected. How did it actually turn out?”
And what we found at the 10-year mark was that the fiber optic system and smart grid had generated about $2.7 billion in community value. And it consists of everything from, I mentioned that the smart grid allows us to reroute power around damage to the system, that’s led to a 55% reduction in outage durations and minutes.
Diana Blass: Now interestingly, the network almost didn’t happen. It faced lawsuits and outcry from critics who claim that the network would eat away at the telecom market. An argument that’s quite common across the country. In fact, 16 cities are currently banned from building their own networks.
What kind of impact has that had upon innovation? A question that comes to mind as we dig into the smart city projects that have come alive inside the City of Chattanooga, home to a smart intersection project that’s grown massively in recent years from a mile and a half radius to now about a hundred intersections.
Dr. Mina Sartipi: The fact that we had the fiber, it made it easier for us, more efficient and potentially more cost-effective to deploy that.
Diana Blass: Dr. Mina Sartipi is a founding director of the Center for Urban Informatics and Progress at the University of Tennessee at Chattanooga, which initially funded the Smart Corridor project with support from the city and partners. It later got federal funding to expand the project. Today you’ll find LiDAR, cameras, audio and air quality sensors deployed at the intersections.
So before the smart intersection was expanded, what outcomes did you see from your what was like a mile and a half or so where you initially tested?
Dr. Mina Sartipi: So there was about almost nine months between the time that we got the approval from the university, from the city from EPB to deploy this until all the equipment were installed in our first intersection. That time was spent on community engagement. We did a lot of surveys, a lot of focus groups. And one of the things that came up very high was pedestrian safety. So one of the things was we wanted to understand where are the pedestrian crossing the street? And there were areas that we understood that between the intersections was a long distance, so there was a lot of crossing the street on the undesignated areas, which obviously it would be risky.
And the other thing that we realized and the mayor’s office wanted us to look into was near misses, or near hit depending how you want to say it, that it would be like, “Okay, so if an accident had happened, there’s a record of it, but where are the areas that they, maybe there hasn’t been an accident, but there’s a lot of near misses happening?”
So that was one of the things that we looked into very early on in different intersections in terms of the traffic control timing, and in terms of where the crosswalks will be and such. So some of those has been implemented. And the other thing that currently we are working on is funded project by Department of Energy that we are optimizing the traffic controllers along the corridor and this is basically making it adaptive.
Diana Blass: And so just to confirm, when you say adaptive, you mean real time like analyzing that data in real time?
Dr. Mina Sartipi: Yes. The final, final goal is to be able to look at the data and based on the data, based on the current occupancy, predict the future occupancy and the next cycle occupancy and based on that adapt the traffic controller. So a lot of those would be based on data that will be collected in real time and the decision will be made and the traffic controller will be adjusted.
Diana Blass: The project is interesting for many reasons, particularly its use of LiDAR, a sensing technology that uses light in the form of a pulse laser to measure ranges.
LiDAR kind of got a bad wrap because it was so expensive. That seemed to be one of the main hurdles to get over. Is that still an issue when you think about further expanding this or deploying more LiDAR in your city?
Dr. Mina Sartipi: That’s great question because when we deployed our test bed, we didn’t even think about LiDAR because it was so expensive. The price has came down quite a bit since the three, four years ago when we were starting the purchasing process and everything, the price has came down. And as of now, still is our plan to have all hundred intersections have LiDARs and cameras. And we are hoping, and we have great partners in the LiDAR, both from the hardware, which we work with Ouster, and from the software part we work with Seoul Robotics. Those are our industry partners that we work with them that we’re going to continue working with them and making sure that we will be able to deploy it as it gets expanded. But the need for technologies like lidar is there really for improving the accuracy.
Diana Blass: Typically, we associate it with the tech found inside an autonomous car where it helps the car to see its surroundings, but not everyone’s convinced of its value there.
Elon Musk: It’s expensive and unnecessary. And as Andrea was saying, once you solve vision, it’s worthless.
Diana Blass: Musk has led Tesla’s crusade against the technology. He says cameras and computer vision are just as effective as LiDAR and come at a lower cost. He’s not wrong. LiDAR is expensive, but that’s changing. Let’s check in with The Six Five co-host and The Futurum Group CEO, Daniel Newman.
Daniel Newman: LiDAR offers a depth of field that’s very unique. And by the way, I’ve actually experienced this. I rode in the back of a Tesla and then I rode in the back of a Lexus that was powered by Luminar’s LiDAR out at CES and there was some different experiments. And for instance, they did the kind of static stop, which the traditional sensing with cameras worked really well. But then they did something like where a child comes out from behind a vehicle that was parked along the side of a busy road and LiDAR was able to stop very promptly, whereas traditional camera radar technology missed it altogether and ran over the child. In fact, one of the most shared ever social videos I saw was that exact demonstration.
Diana Blass: Wow. Yeah, it’s interesting because I remember when I first started covering LiDAR, this was 2019, 2020, and Elon Musk at the time famously was like, “Oh, our cars aren’t going to use LiDAR.” And it’s a big debate of whether it was needed or not. And now it seems, at least in the smart city spaces, the idea of using cameras and LiDAR together to have the best vision zero path forward for pedestrian safety and stuff like that.
Daniel Newman: It’s an interesting debate that the industry is having because we’ve seen with Tesla, for instance, how a company can largely use vision only on camera to create a safe and secure near autonomous driving experience, except in those cases where it didn’t work, and there have been some of them. And when you can add LiDAR, and of course price has always been a problem, it was multi thousands of dollars per vehicle. But of course we’ve seen that come down.
And then anyone that’s seen some of the LiDAR driving around on Waymo and some of the driverless vehicles saw the size and the profile of LiDAR and having a large computer in your trunk and then having a large sensor on your roof, and it was pretty instantaneously recognized that that’s not a realistic modality for having a vehicle. But we’ve seen the profile now. If you look at what Luminar did with its XC90 and Volvo, it’s a very low profile, sits right above the mirror in front of the windshield and offers the same LiDAR capabilities. And I’ve had a number of conversations with Luminar’s CEO, Austin Russell, and he’ll actually talk about the fact that it’s now under $1,000 a vehicle. And of course in very economical vehicles, that could be a challenge. But you also have to think about what is the value of a life?
Diana Blass: Now, just to play devil’s advocate here, let’s take another look at that smart intersection project in Chattanooga. Chattanooga is currently collecting massive amounts of data about its road conditions, data that new technologies like V2X and V2V will be able to communicate to cars on the road. I would think that takes a lot of pressure off of the cars, which maybe don’t need to be as smart as we thought.
William Muller: So LiDAR is just one piece of the puzzle. So absolutely the more technology and the more connected things you have in a space, the better your success potential going to be on anything you deploy.
Diana Blass: Earlier I spoke with William Mueller of Seoul Robotics, the company behind the perception software within the LiDAR sensors, check it out.
William Muller: So what we do is our software sits on the edge in the cab, we have our own processor. What we find that most cities, including Chattanooga, do is they take the actionable events from us and use that to push function from the camera systems. Somebody crossing a pedestrian crossing or somebody jaywalk, then we can take an actionable event and say, “Okay, let’s bring up the video feed. Let’s try to understand what’s going on there.” To say, “Well, maybe it might be a mother and a stroller take little longer to cross the road. Or it might be somebody that’s handicapped.” Something like that. Those are some example use cases.
As well as the technology has very deep insights where you actually know the size, the speed, the heading, the trajectory of those objects. And then all the data, the process data, the data that’s got the events and that’s relevant that the city needs is all then pushed out of the fiber networks to, call it, their data center. And at their data center is where they do a lot of the fusion of data and information, overlaying of information. But actually in parallel because they have the bandwidth available, they also do real time data fusion and overlay of information.
Diana Blass: Okay, yeah, that makes complete sense. But just to confirm, and you probably said this and it’s going over my head, but your software only then dictates the LiDAR sensors or it takes all the footage from also the cameras and then pieces them together?
William Muller: Only from the LiDARs.
Diana Blass: Okay.
William Muller: We only take the LiDAR data and then we push the information to the camera systems. So the camera system still does its own thing. We do our own thing. So we’re kind of in the middle right now, we’re stuck in the middle where we’ve got two systems. Now there is some programs that are starting to take place where now there’s becoming like a sensor fusion approach, but you do sit with a risk there again, right? When suddenly you bringing in the image onto the 3D data and combining the two again, well then you’re saying, “Well wait a minute. Now we’re compromising the aspects of privacy.”
So in some instances you still need to keep them apart. But from a data processing in the background, there is some levels of sensor fusion going on when you might be doing things like speed enforcement or you might be doing things like flow tolling systems where vehicles are passing through and you need something to trigger where to take the plate photo and things like that. So those are some use cases where you’re seeing more of a data fusion versus a image fusion.
Diana Blass: I see. I think Velodyne’s CEO talked about how the smart city market is just going to be a top driver for LiDAR. Do you agree with that statement? And if so, how does then the smart city vertical further push the adoption of LiDAR into other areas like automotive?
William Muller: So this is a topic that I’m pretty passionate about. Seoul Robotics as a company, we ultimately originally developed our software to think it’s going to go on autonomous vehicles. That was that goal when we first started developing this technology. But you quickly realize a couple of things, right? The time of where things are and the time it takes to get to where you want to be. And to be able to decide between two paths, so there’s one path where you’ve got city to city connection, where you’ve got these long haul highways where interstates running from one state to the other. It’s a lot easier to run autonomous solutions on long hauls because you don’t have a lot of dynamics, right? It’s pretty much just a highway, pretty much the same structure, the way things work. It was all kind of standardized to some degree.
Now I try and do it in a city, it’s very dynamic. You add people, you add bikes, you add a lot of buildings, you add a lot of obstructions, occlusions, it starts making autonomy solutions really challenging. We see a lot of videos out there of these vehicles getting stuck at certain scenarios, not knowing what to do. So we took a decision a while back and said, “Wait a minute. We believe that the step before autonomous should be smart infrastructure.”
So having the infrastructure having a level of intelligence is actually going to make the path’s autonomy more successful. So now to make that happen and to catch up, because I almost feel like the developments of the vehicles are almost a little bit ahead of the infrastructure.
And as we all know, a lot of aging infrastructure is out there.
Take North America specifically, there’s a lot of infrastructure that’s struggling that’s aged, that’s got technology that’s 20, 30 years old that I don’t think they even make parts for the stuff anymore. But somehow we want to start seeing these autonomous vehicles cruising the streets there. We want to start seeing these last mile delivery robots running through there. All of that stuff is going to be a lot more successful when that infrastructure is in place to support that.
Diana Blass: It’s interesting to look at how Chattanooga has overcome those challenges. As we mentioned earlier, the University of Tennessee, Chattanooga has played a massive role in the deployment of smart city projects in Chattanooga.
Dr. Mina Sartipi: The majority of these works, if not 100% of them, are dominated by our students. So this is also an opportunity for the students to get experience with the real world scenario, to work with people outside of the university. The same thing for I think from the company’s perspective would be that some brilliant young minds would get to work on some of their challenges and alongside with them.
Diana Blass: In this episode, we also discussed the value of the utility owned fiber network, which has enabled the city to test new technologies in a cost-effective way. Many areas in the country are unable to do that, as evidenced by the digital divide. That, combined with our aging infrastructure continues to challenge the arrival of an autonomous future. The Biden administration’s Infrastructure Law could help as it dishes out billions to cities to invest in broadband, close the digital divide and modernize our infrastructure. Could that be the missing piece to the puzzle? Well, stay tuned because I’ll be sure to connect you with what happens next. Until next time, I’m Diana Blass.