S1E4: Forward to the future: connected vehicles

[00:00:00] Announcer: The following research is part of the National Institute for Congestion Reduction funded by the United States Department of Transportation through the University Transportation Center program. Learn more at www.nicr.usf.edu.

[00:00:24] Wayne Garcia: Welcome to Out of my Lane, a podcast of the Center for Urban Transportation Research at the University of South Florida in Tampa. I'm Wayne Garcia, your host for this podcast, which is about all things transportation, and how to get from point A to point B safely and without losing your mind. Always, a tough question in transportation. Today we want to talk about self-driving vehicles, autonomous vehicles. I know you've seen them in the science fiction movies, and now some of them are out on the streets today. They're being, piloted, perfected all of those kinds of things. And with me today from the Center for Urban Transportation Research or CUTR are two of their scientists and researchers who are working on those matters both separately from, but also as part of the NICR program that is out there, the National Institute for Congestion Reduction, which USF administers that program for a number of universities that are involved. With me today are Dr. Sisinnio Concas, Dr. Concas welcome. 

[00:01:41] Sisinnio Concas: Thank you. 

[00:01:42] Wayne Garcia: And Dr. Shaw Lee. Shaw, good to have you back. 

[00:01:46] Xiaopeng Li: Thanks. 

[00:01:46] Wayne Garcia: They have worked both separately and together on various research studies dealing with these kinds of vehicles. Dr. Concas is a program manager at CUTR, and he runs the Autonomous Connected Mobility Evaluation group program, ACME, for those out there, keeping track of the acronyms. And Dr. Lee is the director of the NICR program and grant and is the founder of the Connected and Autonomous Transportation Systems Lab that's CATS for those keeping track. So, we've got NICR, ACME, CATS and CUTR out there. There will be a test at the end of the podcast for all you listeners. Thank you so much for coming and being part of our discussion as we talk about autonomous vehicles. So, let me start off with, because we talked a little bit before we came on microphone about what's the difference between an autonomous vehicle and a connected autonomous vehicle, Sisinnio? 

[00:02:54] Sisinnio Concas: Sure. I would like Dr. Lee, elaborate a little bit more on the autonomous vehicles. Perhaps you probably have discussed this in previous podcasts, but in terms of connected vehicles, these are a technology that allows vehicles to communicate with each other and with the surrounded infrastructure. The communication happens over the air vehicles can communicate using their own language. Think of that as English for us right now. So, we can talk and acquire information. This information is very useful when exchanged between vehicles and the infrastructure to prevent both accidents and improve the flow of the vehicles on the road network. So, we will be talking in detail how the communication happens through this discussion. 

[00:03:36] Wayne Garcia: Gotcha. And so, Dr. Lee, the autonomous vehicles, they would seem to be like, to me, I would think, oh, these are all the same, but they're not, they're really different skill sets that the vehicles need to have than a connected system. So, talk about autonomous vehicles. What does that category mean? 

[00:03:52] Xiaopeng Li: Yeah. As Sisinnio mentioned, connected the vehicles are about communications, whereas autonomous vehicles are about vehicle control. So, we're talking about the regular vehicles that we are driving. So basically, each vehicle needs a human driver whereas an autonomous vehicle is essentially to replace the human driver with a robot driver. You're going to have computers to control the autonomous vehicles and you are going to have sensors equipped to perceive the surrounding environment. And you're going to have actuators and automated mechanical control to drive the vehicles. So that's ultimately about what an autonomous vehicle is. It's basically using a robot driver to replace a human driver. 

[00:04:44] Wayne Garcia: So, is it fair to say one piece of this study, the connected piece is how the vehicle connects to the rest of the world? So, when I first started researching for this podcast, I was like, connected. I like, like physically connected. "Are we talking like a train of cars now? It's the connection to the data in the world. 

[00:05:04] Sisinnio Concas: That's right. And I think this is probably in the future when we will have, hopefully, autonomous vehicles on the road, it will also be connected. They will be self-driving, but it will be communicating with each other and also with the surrounded infrastructure.

[00:05:19] Sisinnio Concas: So, a communication means exchange of information, mostly is data currently the connected vehicles they exchange. Position paid information, heading all sorts of, information related to their movement on the road. The communication happens because there are devices that they broadcast over the air using high frequency signal, this information, nearby vehicles, they can start and engaging this communication and along the road, and this is happening across the US in the world at large; there are also investments, large investments, in infrastructure in terms of having other devices installed at intersections along the road that can broadcast information from the vehicle to the vehicles. And also, vehicles can relay this information back to these boxes. Altogether, putting a framework for an ecosystem where in the future, when we will have autonomous vehicles, information will happen seamlessly and very efficiently.

[00:06:13] Wayne Garcia: Yeah, I hadn't thought about it this way, but the autonomous vehicles that are out being tested on the road today, they don't talk to each other. They're like independent players. They're trying to sense the world as they individually see it. So, they're trying to figure out is that light red or green? Where's the curb, where's my turn lane, that kind of thing. So, then the next level is to connect them. So, everyone can talk and understand like you're going to be coming up and the light's changing instead of trying to sense that the light's changing 

[00:06:44] Sisinnio Concas: It, it is correct. It is correct. I think that Dr. Lee can elaborate because he has done a lot of work in terms of autonomous vehicles and developments in terms of the sensors that are used to obtain this information. And when we speak about sensors in vehicle sensor, for example, reading lane marks on the road, but in terms of connected vehicle, the information can also be broadcasted, for example, by these roadside units at intersections, then they can broadcast information on the timing of the signal, the layout of the intersection, everything digital, the connective vehicles absorb this information over the air. They can process. They can compensate for some of the shortcomings of the own sensors, like the capacity of reading an intersection, color of the signal, signal that could be red, but in front of the signal, you could have the sun coming up and providing a problem to the camera or other sensor capacity to read the color of that light. Using information instead broadcasted by the intersection directly to the vehicle would provide more seamless and efficient way of, obtaining this information. 

[00:07:49] Wayne Garcia: So then going back to the autonomous vehicles, Dr. Lee, how much do they need to know? I'm thinking now in terms of sensing, processing, understanding, that's got to be really complex. 

[00:08:02] Xiaopeng Li: Yeah. So, do you mean, how much do they need to do the autonomous vehicles need to perceive? Is that a question?

[00:08:09] Wayne Garcia: Yeah. Like how many things are going on inside an autonomous vehicle in terms of its ability to navigate a roadway? It's not just some kind of scanning and stay between the lanes. There's so much more going on; give our listeners a sense of the complexity of getting that 

[00:08:27] Xiaopeng Li: yeah, sure. So, I want to make an analogy to human drivers, think about, we all know, like when we drive what we do and what organs of our body we need to use. First of all, we probably want to use our eyes to see the situations around. Sometimes we want to use our ears to hear sirens and honks that kind of thing. So, these are like our perception organs, and there will be like an equivalent, components in autonomous vehicles about perception. They have camera sensors just to like our eyes and they may have more advanced sensors using LIDAR, radar, you know, and other techniques that can perceive faster potentially better than human, eyes. And next, when we drive, after we perceive the surrounding information, we need to identify which ones are objects, which ones are human beings. What, where is my rights of land and that kind of thing. So, our brain's going to work and according to that, we're going to decide the driving decision of autonomous vehicles.

[00:09:41] Xiaopeng Li: And that's where the computer component, as well as complicated software some of that is related to artificial intelligence is for. So, you do complicated data analytics to recognize objects and classify them into different types and then make their driver this decide which, space is drivable and what route or pass I need to follow.

[00:10:05] Xiaopeng Li: And next, think about as a human driver, we want to use our hands and feet to execute the decisions, right? Steering, steering wheels, and also step on gas or brake pedals. So autonomous vehicles also have that function with automating the control. We can convert the computer control signals into the very fast reacting mechanical controls to execute, the control decisions to follow the plan pass. Yeah, that's about overview of the autonomous vehicle functions. You can really just sync autonomous vehicles' robot driver that has pretty much equivalent in body functions as a human driver. 

[00:10:47] Wayne Garcia: Now you have one of these vehicles, correct? 

[00:10:50] Xiaopeng Li: We have them in our lab. 

[00:10:53] Wayne Garcia: And so, you've ridden in one. Tell us what is that like?

[00:10:57] Xiaopeng Li: Yeah. Actually I. I feel like I enjoy it, it's cool. You're going to see vehicles drive themselves. And of course, we put the driver behind the wheels to control the vehicle just in case something happens. But most of the time it functions pretty well. You can. It's pretty cool.

[00:11:15] Xiaopeng Li: And yeah. And when I, especially when I wave to people with my hands open and hands up, and that, that looks pretty cool. And also. It's actually not, driving behavior is very smooth. We actually did lots of work to adjust to the vehicles, controllability and comfort-ness, and therefore actually driving pattern actually it's a better than a good driver. 

[00:11:40] Wayne Garcia: You feel like the public has some level of a split to it of people who go, wow, this is great. I can't wait. And others who are like, I will never not be in control of my car. That sort of gets us into the behavioral questions. And Dr. Concas, I know your training is in economics. And so, you look at a lot of the things that you're looking at for, these research studies in terms of wants needs human behaviors, even. 

[00:12:08] Sisinnio Concas: Yeah, definitely as, the training of an economist, is based on, the study of human behavior, rationalizing, decisions.

[00:12:16] Sisinnio Concas: Why do you make such decisions? And for what purposes and ultimately, what the, what is the outcome. In the case of transportation, the demand for transport, it is an indirect demand. So, we end up using vehicles or whatever mode to get from A to B either to, pick up somebody or to acquire good or services.

[00:12:34] Sisinnio Concas: So, the element of economics plays a great role in terms of studying human behavior. But I would like to bring it back to what, in terms of the realm of our work of, research work, autonomous and connected vehicles in, in the case of CUTR, for example, for the last four to five years, we have engaged in large and complex studies of human behavior, as it pertains to, connected vehicles.

[00:12:58] Sisinnio Concas: The Tampa downtown is a very lively area, is growing very fast and happens to be one of the three sites in the United States where the largest scale deployment of connected vehicle has been happening for the last five years. We have the Tampa Hillsborough Expressway Authority – THEA because we have a THEA downtown Tampa CV pilot.

[00:13:18] Sisinnio Concas: We, at one point in time, had recruited up to 1000 participants and equip their vehicles with connected vehicle tech. The exchange of information between vehicles happens at a very high frequency to ensure that safety is retained. And when I say high frequencies -up to 10 times per second. So, when you think about studying behavior, relying on a panel of 1000 individuals over the course of two three years, the amount of information to be processed is humongous in the order of billions and billions of observations. This is the work of CUTR in terms of analyzing ultimately, and understand, what are, if any constant in terms of human behavior, as the, these participants behind the wheels are exposed to connective vehicle technologies. The human behavior what we have found through the study of connected vehicles, bear in mind the connected vehicle technology allows to provide information directly to the drivers behind the wheels. And in the case of the CSVV pilot, this happens by way of a human machine interface by displaying information on the rear-view mirror. So, what we have done, we have started studying, and we've been doing this for the last five years, in terms of understanding what happens when individuals are receiving information that is relevant as they move through the transportation network. That's their behavior change, or if it doesn't why? And this is what we've been doing basically for the last five years. And, it is a, a very simulating and challenging field of work. 

[00:14:51] Wayne Garcia: Do we know anything yet even preliminarily about their behaviors change with this information or not. And how much more does this study have to go in terms of really mining that data?

[00:15:04] Sisinnio Concas: Sure. We have found that, in some of the connected vehicle application, there is, there are really relevant in terms of ensuring increased safety to travelers in the system. There is a positive response as the humans’ travelers digest this information provided by the collective connective vehicles they can in, in one of the, probably the most important applications that rely heavily on connected vehicle technology it is the electronic, emergency brake light. So, you are in stream of traffic, there's a lot of vehicles in front of you. And this is probably a striking difference so far with respect to autonomous vehicles, which is always the autonomous vehicles is scanning for the nearby environment. But when there are cars in front of you, it could be a few hundred feet and there is a car ahead of you, you don't see that car, but if the car is connected it is transmitting signals. And if the car is in distress and is having by distress a conflict with another vehicle that information can be digested by your vehicles. You might be distracted, but you have enough time. If you receive that warning on your rear-view mirror to take proper action.

[00:16:14] Sisinnio Concas: So, we found out the participants in this study using these connective vehicle applications, they ended up increasing the safety of the road sections where these applications had been deployed in downtown. 

[00:16:25] Sisinnio Concas: Tampa 

[00:16:26] Wayne Garcia: Wow, that, that, that's just amazing. And how do these two research projects then come together? Ultimately does connected vehicles then obviate the need for some of the sensing and computing ability or really need both and we're both moving down the track at the same time. 

[00:16:41] Xiaopeng Li: Sure. So, I think, in my opinion, if they can get married, the world will be better. Think you can't. Connectivity alone or have automation alone. But ultimately, I think that if they work together, it's going to be better. For example, with automation, the connected information communication could be a lot more complicated and a richer than human drivers. Our human drivers, if we drive and connect to the vehicles because of the limits of the human beings, we can only process so much information. If it's more than that, it could be a distraction. But think about the computers, it can process like megabytes of data in a split of second.

[00:17:26] Xiaopeng Li: And you can and with that, you can expand the communication bandwidth of connected vehicles and take a greater benefit of that. From the automation point of view, if you only have automation, you can only make decisions like Sisinnio said, based on, what you just immediately perceive. That's just a few vehicles surround you, but if you have connected vehicle information, like you can see further, you can see after vehicles that block your view and you can see miles down there is going to be an incident. Okay. There is going to be a signal light. That's going to turn red on half mile down the road, things like that. So, with that automation will make a much better decision and that can improve the mobility, reduce the energy consumption, emissions, and much improve safety. 

[00:18:16] Wayne Garcia: Dr. Concas, you mentioned Tampa has a very good system, especially downtown of these connections, these things that have been put into place to tell vehicles things, and then ultimately to tell a driver or to tell an autonomous vehicle, it could be either way, how embracing is our local governments at this point to really building these into transportation systems that, that this connectivity is out there. Are we at just at the start of doing this, embracing this, or are we further along than most people realize?

[00:18:51] Sisinnio Concas: I think distinguishing between states within the United States are large. There's a lot of heterogenetic differences, but we are in Florida. And I think we're on the forefront in terms of, now at the, the latest technologies and, the state of Florida is very active. So, they've been embracing changes in technology for quite a few years now. And I think both in realm of autonomous and connected vehicles, there are some challenges for agencies in terms of the infrastructure necessary to implement connected technology; although the more you deploy, the more on average, you think that there is economic efficiency in terms of lowering the unit cost of the infrastructure necessary for connected vehicle technology to be deployed on a large scale. And in the case of Tampa, the agencies are very well open. There's a low communication between, the different partners and stakeholders in, in terms of those managing the road network for the state as the entire state of Florida. Both Dr. Lee and I, for example, are working on a larger scale deployment, the I 4 or interstate four frame network, which sees the deployment of connected and, potentially, also autonomous connected vehicle technology in the future, a tapestry of connected vehicle infrastructure between here in Tampa and all the way to and pass and well beyond Orlando. And that in the next few years will mean a lot in terms of what we expect. And as Dr. Lee was mentioning in the different benefits associated with this technology, which on the forefront, on the most importantly are in terms of safety. But also, the efficient movement of goods and people in the state and the ancillary benefits that come from efficiency and reduction accidents, which are environmental and, energy related as well.

[00:20:32] Wayne Garcia: And that's interesting, you bring up not just individual use, but the commercial use. Do you see the commercial industries, the trucking industry adopting these kinds of solutions before the public at large does? Do we know anything about, what that might look like in terms of embracing...? 

[00:20:51] Sisinnio Concas: I'm going to focus on the connected element, which is a lot of the work that I do. And I think Dr. Lee can elaborate in terms of freight autonomous vehicles, which is a very interesting and fast developing realm of work in terms of the private sector, taking care of scaling up the deployment of autonomous vehicles as it pertains to freight movement. But in terms of also autonomous connected, vehicle technology related to the movement of heavy vehicles- yes, it is very important because the connected vehicle technology can allow the freight tracks to digest information from nearby infrastructure and move a little bit more seamlessly. For example, you're approaching an arterial, which is, layered with several intersections. If these intersections, each one of them equipped with a light can give a little bit more priority, considering all the other vehicles, it could be emergency vehicles and so forth, but a freight vehicle approaching the intersection can ask a little bit more priority in terms of going by and extending the green light so that at the very end, it can get from A to B a little faster. And that will mean a lot for us in terms of having goods moving little bit more seamlessly in the network. That happens for freight. It can happen for public transportation, definitely, but it's an area of work that, that, that is fast developing, and I think in terms of the movement of, heavy vehicles, relying on autonomous technology there's a lot going on. 

[00:22:14] Wayne Garcia: Has the driver shortage and the whole supply chain issue driven more interest in trucking companies that say, "Wow, this would be really cool to have no driver shortage?"

[00:22:26] Xiaopeng Li: Yes. That could be a factor. Basically, if you know the cost structure of the freight industry, driver costs or a significant component, if, they could reduce or completely eliminate that portion of a cost, the freight industry can benefit a lot. And potentially you could do that by automating the vehicle.

[00:22:48] Xiaopeng Li: There is some transition to technology. Some are referred as plateauing or, connected adaptive cruise control, but these basically mean that trucks, they can sort of, via communication automation, they can follow each other autonomously with the smaller gaps and, some, current technologies just to having one driver in the lead track, it can guide the several autonomous tracks without driver and along long-haul transportation journey. And in the future, they envision that they will, make all these tracks autonomous and that way, they can potentially completely eliminate the driver cost and worry of the driver shortage possibly. But of course, there is going to be another issue of the workforce impact to the driver. The truck drivers are a significant sector of, workforce, in our society and, how to take care of them after that is going to become an issue, we should be mindful for, for that as well. So 

[00:23:54] Wayne Garcia: What's the timetable here? I'll ask each of you because you're in connected, but separate fields. What's the timetable for widespread appearance of connections for vehicles, driver, or driver less. And the timetable for seeing a substantial number of self-driving vehicles. I'll start with Dr. Lee first. How far is the future away on this? 

[00:24:21] Xiaopeng Li: Yeah, that, that's a good question, but I don't have the crystal ball, but what I can tell you is a little bit of the history. Back in the 1930s, people were predicting that by the year of 1960s, all the vehicle fleets would be autonomous by that time. We did have some ups and downs along the journey. We did, complete accomplish great technology advance from a 1930s to 1960s, but vehicles were apparently not fully automated. So, there would be some hypes and, high hopes driven by either good wishes from the academia or the industry interests, right? The venture invests companies. You're going to see different, predictions of the timelines, but I would say that it's a variable and it depends on what we do here actually. What governments do in the next 10 years or 20 years will impact its timeline and what, we as researchers do, maybe what Dr. Concas and CUTR, NICR do may have also a minor impact to this process as well. And what do you do as a podcast host? A host could also have impact after people hear this episode, and maybe they all think more of getting to this, field and this realm to advance the technology. A low-level automation is already there. It's it happened and I think maybe in it's widely happened in, oh... 

[00:26:01] Wayne Garcia: Yeah. We have lane assist and emergency side. We have a lot of things that are starting to be built into every single. Vehicle, just like Dr. Concas on your timeline for fully connected vehicles. We're somewhat connected now. If I go on Waze, it's going to tell me, oh, there's a train crossing here, you should slow down or there's a police officer ahead, you should definitely slow down. So, what's the timeline for that?

[00:26:26] Sisinnio Concas: I think that, and I'm going to relate it to our specific deployment the most current phase that we have in downtown, but simply because we're engaging with the OEMs, the original equipment manufacturers. Specifically, we're working closely with Toyota Hyundai and Honda in the deployment of connected vehicle technologies. Now the different car makers in this instance, their vehicles, they are communicating with each other. So, they're starting to speak the same language, which I think is a substantial evolution in terms of the widespread adoption of one way of communicating between different car makers. So that means that if, and this is coupled with, Dr. Lee's comments about the regulatory framework, the changes that are in the horizon will be substantial. Then there will be a faster adoption of this technology because we're in terms of the technology itself, we're really quite well ahead, both in terms of autonomous and connected via technology, but the regulatory framework pays a key role in here, where in the midst, as it pertains to connected vehicle technology waiting for resolution from the Federal Communication Commission regarding the exchange of information between vehicles, rules. Clarity is obtained hopefully by the end of this year, then it will mean that the OEMs will affect effectively make a strategic decision in terms of commercializing on their own vehicles, this technology. So right now, we might have a small-scale deployment and acceptance, but it's like breaking the ice once it happens, it happens. 

[00:28:01] Xiaopeng Li: Yeah. I just want to add that low level automations, as I mentioned earlier, already there, right? Many of you already drive vehicles with autopilot, they can follow the proceeding vehicle and they can keep you in the same land. Or sometimes they can do some cool maneuvers, like lane changings. These are already level one and level two automations. And recently if you watch news, you might see that Mercedes-Benz would issue a level three automated vehicle and be probably in this May, they're going to start to operate that, in Germany that's already level three and then there is only going to be level four and level five to go to reach the highest level of automation. So, I'm still very hopeful despite the ups and downs of the history. It’s just like stock market. It grows in a long run. The overall picture is growing, but it has ups and downs if you're looking to a particular timeframe. 

[00:29:00] Wayne Garcia: Excellent. Gentlemen, we're at the end of this episode. We could talk forever because these are really things that people, once they hear this, they're going to realize how much their feet are already in this new world and what's coming. And it's not one of those things where they'll say, in my lifetime, I'll see this. No, I think sooner than later, obviously. So again, Sisinnio Concas and Shaw Lee from the Center for Urban Transportation Research. And both of them are researchers on the NICR program, trying to reduce congestion, make us all safer on the roadways. Thank you for joining us. I'm your host Wayne Garcia. We'll see you at the next podcast. 

[00:29:43] Announcer: The National Institute for Congestion Reduction, NICR, is a Transportation Research Center focused on innovative congestion strategies. The center is composed of researchers from the University of South Florida, the University of California, Berkeley, Texas A & M University, and the University of Puerto Rico at Mayagüez and funded by the United States Department of Transportation. For more information, please visit www.nicr.usf.edu.