Episode 48: Developing Automated Buses

Lisa Peña (LP): An automated 14 passenger shuttle is using algorithms, sensors, cameras, and software to drive itself around the SwRI campus. And this is just the beginning. Soon, passengers in big cities could be commuting on automated buses. We're hitting the road, destination autonomous driving next on this episode of Technology Today.

[MUSIC PLAYING]

We live with technology, science, engineering, and the results of innovative research every day. Now, let's understand it better. You're listening to the Technology Today podcast presented by Southwest Research Institute. Transcripts and photos for this episode and all episodes are available at podcast.swri.org.

Hello and welcome to Technology Today. I'm Lisa Peña. Today, we're talking about automated vehicles, specifically a new shuttle on our San Antonio, Texas campus. We'll also discuss a SwRI automated technology that could offer some relief for New York City commuters. We explored autonomous driving on episode five back in 2019. But over the past 3 and 1/2 years, there have been some new developments in autonomous technology. Our guest today, Jerry Towler, engineer and assistant director of Robotics in our Intelligence Systems Division, is sharing automated driving updates and telling us more about that self-driving shuttle. Thanks for being here, Jerry.

Jerry Towler (JT): Yeah, I really appreciate the opportunity to come give some updates on what we've been doing for the last couple of years.

LP: Well, we're excited to hear all about it. And so, let's start with our new SwRI 14 passenger shuttle. It is a Ford Transit outfitted for autonomous driving. And fun for me, I got to take a ride. So, it is hands off driving. You just kind of see the wheel moving by itself. It accelerates and brakes on its own, a really cool firsthand experience for me. So, thanks for setting that up. But what is the overall purpose of having a shuttle like this on our campus? How is it being used?

JT: Well, I think there are really two uses for this shuttle. The first one is that our team has done a really great job of integrating a lot of our state-of-the-art capabilities onto this one vehicle. So that's everything from motion execution through perception and machine learning, artificial intelligence, and connected vehicle technology all onto the same platform so that we can not only test each of them individually, but also all of them together. So having that kind of demonstration and test platform is really important. However, it's also intended to actually give tours of campus. So, we have people who want to visit all of our different facilities. And this is a way for them to get around and kind of take that automated windshield tour.

LP: Yeah, a really cool way to showcase our technology and what we're working on in those labs that sometimes are really secretive, and no one knows what's going on. But it's really nice to see something like this out on the roads and say, hey, this is what we've been doing all this time. So, a little bit about my perspective from-- my perspective as a passenger on the shuttle, I saw Logan, one of our engineers, push some buttons and off we went, at least that's what it felt like. The shuttle brakes at all intersections. And Logan told it when it was safe to go through the intersection by pushing a button on a screen. And it also brakes when it senses something is getting too close. So that was neat. It really felt like it's really aware of its surroundings. So that was my experience as a passenger. But from an engineering perspective, there's really so much more going on. So, what does the shuttle do? Walk us through operating the shuttle.

JT: Sure. So, operating the shuttle is exactly as simple as it looked once everything has already been set up. You select a destination, and you hit Go. And then the safety driver there, in this case, Logan, is aware of everything the shuttle can do and everything that it can't do. So, he can pay attention. And if necessary, he can control the vehicle, he can take over control, he can stop it to make sure that we don't have any issues that go wrong while you're taking a ride.
 

Courtesy of SwRI

SwRI’s automated shuttle is a 14-passenger, self-driving transport vehicle equipped with lidar, camera sensors and machine vision algorithms that enable automated driving. The shuttle autonomously navigates campus tours on pre-programmed routes, but a human driver is always behind the steering wheel for added safety.

But a lot of the work, in fact, almost all of it, comes before you get on the vehicle there. We have to build maps of the area that we are going to be driving in, we have to understand what kind of the what we call the operational driving domain, or kind of the context of the driving is going to be. And then what you saw with Logan telling the vehicle when to go through intersections is kind of an interesting switch on switch off functionality where the vehicle actually can execute those intersections by itself. We developed this a few years ago. It understands four way stops. It understands when it's the vehicle's turn and when it's somebody else's turn. And it knows how to avoid other vehicles going through those. However, especially going around SwRI's campus as we're doing this testing, it's a little bit simpler to let the safety operator take control of the vehicle at those times.

LP: So, let's talk a little bit more about the technology that this shuttle is outfitted with. So, it looks great from the outside, really shiny, black, roomy, inside leather seats, the whole-- all the works there. But that's just the start of why it's really an impressive vehicle. It's equipped with, as we mentioned, state of the art autonomous technology. So, let's get a little bit more into it. How does it drive itself?

JT: Well, there are really three components to this vehicle. And all of them are really pushing what is possible to get done. The first one is kind of what I call core autonomy. That's the part that actually drives the vehicle. It knows where the vehicle is in the world, it knows where it is on the map, it's planning your path through the world. It seems like just driving down a lane should be easy. But in fact, there are a lot of things that go on to identify where the lane is, and how to drive into center of it, and understanding what the speed limit is, and how to stay within that, and actually actuating the controls itself, the throttle, the brake, the steering wheel. And understanding stopping for stop signs, navigating intersections, like we talked about, identifying crosswalks, and stopping for pedestrians, slowing down when you come past pedestrians as you experienced. So that's kind of core autonomy. The second level is the perception components that actually identify the world around it. So, the world is going to change from the map that we built. The world is never going to be exactly the same. And so, we add cameras, and LiDARs, and other sensors to identify what the world looks like and then respond to it appropriately.

And that's a combination of classic computer vision algorithms and very state of the art deep neural networks that tell the vehicle exactly what the world is going to look like so it can respond appropriately. And then finally, and crucially for the bus project that you were talking about, there are connected vehicle components that communicate to the larger world. Most autonomous vehicles that you kind of hear about really only care about themselves and what they are doing. But we've put a lot of work here at SwRI into what you can do when the vehicles can talk to each other, or talk to the world around them, not just observe it. So, we've included that technology in there as well.

LP: All right. And that takes us to our next topic, which, as you said, as we just mentioned, this is something your team is working on a larger scale as well. So, tell us about the problem with New York City buses.
 

Courtesy of SwRI

The operator demonstrates the self-driving capability of the SwRI automated shuttle as it ferries passengers on a tour of the Institute’s campus in San Antonio.

JT: Well, I'm from San Antonio. So, I'm sure someone from New York could tell you much more about the problems with New York City buses. But in this case, it's actually not so much the buses but the roads. That is, there are too many buses. And the buses are too big for the roads. And in particular, on the exclusive bus lane that's leading into Manhattan, there are so many buses. And the buses are so big that even professional operators can have trouble making every run perfect. And so that's what this project is intended to help out with.

LP: All right. So how is SwRI helping out? The problem is too many buses. They are too big. The roads are too narrow. So, what is the solution here?

JT: There are a number of technologies developed at SwRI that we were able to bring to bear against these problems. And collectively, we call them connected autonomous buses. So just like the shuttle, they fuse vehicle to vehicle communication and autonomous technology and perception together to resolve these particular challenges, increase the capacity of their roadway, and kind of decrease the slowdowns and stoppages that are sometimes experienced right now.

So, there are three particular challenges. One of them is that because of the nature of this road, there are two merge points that are somewhat challenging on a road that is already crowded. Everybody is experienced with trying to merge on a road that is not moving very quickly and is completely crowded with cars. Now imagine that your vehicle is a 40-foot bus and so are all the other vehicles. And that's going to get a lot harder. So, we developed some work using internal research at the institute to allow the vehicles to communicate with each other and execute this merge completely autonomously without necessarily communicating with every other vehicle on the road, just the ones that are important to this particular maneuver. And merging isn't the only maneuver it can do, but it's the one that's relevant for this application. The second one is there are too many buses on the road. And so, we need to push those buses closer together. But of course, the closer the buses get, the less safety margin you have. So, we added what's called cooperative adaptive cruise control to these buses to allow us to move them closer together without losing any of that safety margin.

And then finally, the roads are just not big enough for modern buses. They were built a long time ago. And we need to help these operators drive the bus down exactly the same perfect path every time. So SwRI's Ranger technology that I'm sure you've talked about here in the past allows us to do centimeter level precision driving. And we can drive exactly the correct path down this roadway and not have some of the challenges that these operators, who are very good drivers, still can't quite nail that level of precision.

LP: So really involved, interesting technology here. You said it's like a three-pronged approach. The cooperative adaptive cruise control, the automated merge negotiation, and Ranger precision navigation. So, I want to talk a little bit more about Ranger. That is the SwRI navigation software. And it's really interesting because it takes a fingerprint of the road. And I got to see it in action. And it's so precise. You look at-- I saw the-- a square, a piece of road it was looking at. And it looks like all the same to the human eye. But for Ranger, it really picks apart those details that we just can't pick up on. And that's really-- so tell us a little bit more about how Ranger is doing its thing on these automated buses.
 

Three connected automated vehicles demonstrate distributed cooperative merge technology. The vehicles “talk” to each other to determine which will merge first, second and third.

JT: Sure. So, there are a couple of old things and a couple of new things with respect to Ranger on the automated buses. Ranger is a technology that we developed, again, using internal research funding a number of years ago to resolve the localization problem for automated vehicles. Figuring out where you are is outrageously more challenging than you really think it is when you're driving. Because of course, you know where you are most of the time. But the vehicle has to figure it out. And it needs to be extremely right about where it is.

So, Ranger solves this by taking a downward facing camera and taking images of the road surface and then matching those images with extreme precision to a map that was previously built of the road. Now sometimes when you think of maps, you think about these maps you see in the news of these ultra-complex what's called high-definition LiDAR maps that take enormous amounts of time and resources to build and optimize. But Ranger maps can be collected simply by driving down the road. And then the optimization is automatic.

And it works exactly as you said. You take an image of the road, you extract these key points, or these computer vision features out of the road surface, and then you match them to your map. And while asphalt kind of all looks the same to you and to me, Ranger actually can see statistically a few billion miles of roadway before it expects to see any kind of repetition in that texture. And that works on asphalt, or concrete, it works on airport tarmacs, and cement roads. It works even on packed dirt and gravel roads sometimes depending on exactly the nature of the surface.

So, this is a really capable technology. And of course, when we needed precision driving on the buses, we could immediately turn to it. But for the buses, they're much larger than most vehicles. And so, we actually added a second Ranger system, one to the front, and one to the back, so that when we're going around this extremely tight turn near the end of the roadway in question, we can make sure we know exactly where both ends of the bus are at all times, and that we are staying within kind of our designated lane.

LP: All right. Amazing technology there with Ranger. So, let's talk a little bit more about the other two legs here, cooperative adaptive cruise control. So that really lets the-- lets the shuttle or the bus pick up speed or drop speed as needed. What else does it allow the bus to accomplish?
 

Courtesy of SwRI

Jerry Towler, SwRI engineer and assistant director of robotics, gave a presentation on automated technology for buses at the SwRI 74th Annual Meeting.

JT: Sure. Maybe the best way to approach cooperative adaptive cruise control is by building up from something that everybody is familiar with, which is the standard cruise control you find in every vehicle, or almost every vehicle these days. So, in that cruise control, you set a speed and your vehicle maintains that speed, and you still have to steer. But your car is capable of staying at, let's say, 65 miles an hour. The challenge there is that if the car in front of you slows down, your vehicle doesn't know that. And so, you run right into the rear of that car.

So, vehicle developers created an adaptive cruise control, which adds a radar to the front of your vehicle so that we can tell when that car in front of you is slowing down. This is an offshoot of autonomous technology research. But it is simplified to this safety application. And now the car in front of you slows down and you also slowdown, which allows you to operate in much more closely packed traffic with the same kind of cruise control capabilities.

But even then, you need to maintain enough space between you and the next car to completely slow down when that car does. On this road, we need to get even closer than that. So, we add cooperative adaptive cruise control, which means that instead of just a radar, we also have a radio that is actively communicating from one bus to the next, and to the next, and to the next. So, we can see the entire road worth of vehicles at all times. And that means that as soon as the very first vehicle in the queue starts braking, we can also start braking.

As a result, we don't need nearly as much space to slow down as you would in a passenger vehicle. And so, we can bring those vehicles much physically closer together and still maintain the same safety margins for stopping.

LP: I do want to talk about the third part here, which is automated merge negotiation. So, tell us a little bit more about that.

JT: Sure. So, this is another thing that is really unlocked by being able to connect the autonomous vehicles to one another. You can imagine how a person does a merge. You look around, you identify the cars that are going to be potentially challenging for you to merge with, and then you slow down, or you speed up. And if you're lucky, they also slow down or speed up. You-- eventually, you match speeds, and you can merge in, or they can merge into the highway. And if everybody does their job well, then nobody has to really slow down. And the whole highway keeps going. This technology does a very similar thing using radios. The vehicles identify which vehicles are going to be part of the merge operation, that is, which ones are going to have to potentially speed up or slow down to allow the merge to occur smoothly.

Then they negotiate among themselves, which vehicle should go first, which one should go second, if necessary, maybe third or fourth. And then the vehicles that are on the roadway that's being merged into will create that space exactly like you would on a highway. And they'll either slow down or speed up as necessary with staying within the speed limits, of course, to open up a big enough space for that vehicle to merge into. And just like human drivers, if we do our job well, we don't slow down the road at all, or no more than necessary.

LP: So, you have some really neat footage of this really cool video of how this works. And hopefully, we can post that on our episode 48 web episode page if you don't mind sharing that with us.

JT: Yeah, we'll do our best to get you something that looks cool.

LP: Yeah. So how soon do you envision these automated buses hitting the roads in New York?

JT: Well, I have to be careful not to let my kind of inner tech company CEO come out and say that we've performed a proof of concept here with real vehicles on real roads. And it does everything that we wanted it to do. It works the way that we want it to work. But like with all automated vehicle deployments, there's a lot more work to do before you're really going to see this in operational action. Excellent technology is only one little piece of the puzzle of getting automated vehicles onto the roads.

There's also cooperation of all of the various governments and policymakers involved, the insurance companies, the organizations that own the roads and the infrastructure themselves, the bus operators, the bus manufacturers. And of course, passengers also have to be willing to take their morning commute on an automated bus, which is not necessarily something that everybody is about to do tomorrow. So even beyond just the AV research, which is itself quite challenging, there's a whole host of things between here and putting these things on the road. So, I certainly look forward to being able to put these through more advanced trials with larger amounts of vehicles and more independence. But I don't think we should be looking for these any time soon.

LP: All right. A lot of parts there that have to coordinate to make this happen. So right now, you're looking at the New York City area as a potential first spot for this. But could we see this rolled out nationwide one day? I know you're careful, cautious to say-- give us an exact timeline. But if you're starting there, it's reasonable to think that this could go fan out to other big cities at some point.

JT: Sure. I think that's actually completely reasonable. And one of the things that SwRI has worked very hard to do is to make sure that nothing we've done for this project is specific to that area. Nothing we've done for the shuttle on campus is specific to our campus. Of course, we've built a map of that road. We've encoded things like speed limits and stop signs. But none of the technology relies on being in San Antonio or in New Jersey somewhere. So there's no really good technical reason that the same kind of technology can't help the same kinds of problems in other major cities, or highways, or even in bus maintenance yards where this kind of precision navigation and autonomous driving can be really helpful to move buses around the yard during refueling, and maintenance, and that sort of thing where right now, you need an operator to do all of that for you. So, I see no reason that this can't be spread out really across the country. And in fact, you see that there are a few, not a huge number, but a few organizations that are trying to do exactly that.

LP: All right. So SwRI can bring this technology to you. Give us a call. Little commercial there. All right. So, any creative process requires trial and error. So, what does trial and error look like when you're creating an automated bus? How do you test it?

JT: That is the question that keeps up every good researcher at night, and probably even some of the bad ones. Because here's the thing. With a 40-foot bus, even at low speeds, error is pretty catastrophic. So, step one, it's not a trial and error. It's really just a trial. And so, we take a couple of steps here. The first is that of course, there's plenty of software testing that has to go on ahead of time. We can simulate these vehicles. We can build virtual environments. We can make sure that all the components work together. But here at SwRI, we're fortunate to have private test facilities and space to store the vehicles here on campus. So, we actually are able to get out on the roads much earlier than some other organizations have been able to just because we have the facilities here in San Antonio to do so. After that though, a lot of the trials are actually incredibly boring to people who aren't AV engineers. In the beginning, it's mostly just data collection, not really testing the autonomy at all.

And even when you really start testing the autonomy, it's very slow. You take it one step at a time. Can we bring up the system? Can we engage the system? Can we drive in a straight line slowly? Can we follow a path? And then finally, you build up to, can these vehicles at speed merge together smoothly, and correctly, and perfectly every time? But then, that's when the really boring work begins. Because you don't need it to work once. You need it to work every single time. So, then an engineer, possibly Logan, gets to sit in that vehicle as it drives the same path over, and over, and over, and over. And so, these trials are completely necessary and somewhat not exciting at all.

LP: And when it finally, finally does what it's supposed to do when you get that bus that's driving itself, like I saw, what does that feel like? Can you tell us about that moment for you?

JT: Well, unfortunately, it's been a few years since I got to have that particular moment for myself, but I do remember it pretty well. My experience is that the moment of triumph really comes later because the first time it works exactly correctly, your first thought seems to be, wait. Why isn't anything failing? Hold on. Can I do that again? And then it's maybe the 50th run, or the 500th run that you finally-- it finally kind of dawns on you that you've done something really amazing, that you've gotten something really fantastic, and that you can kind of relax a little bit and start working on what's next. And it really is a fantastic feeling when the vehicle doesn't really need your help anymore and it does what it's supposed to do. And you can be suddenly a safety operator and not a test engineer.

LP: Well, as someone who-- I took my first ride in that shuttle. And it really felt like the bus was thinking about things if that makes any sense. I mean, there were times where it would break suddenly. But and Logan said, oh, it saw something. And it's just kind of so aware of its surroundings. And it was just really cool. So, I mean, congratulations to you and your team. I know that's years in the making. And that experience of me sitting in it was just the culmination of all your hard work. And there's bigger moments ahead, of course. But just it really does give you that sense of this bus knows what's going on.

JT: Well, thank you. I certainly appreciate it. We don't always like to anthropomorphize the vehicle too much. But it really is paying attention to the entire world all the time. And we would rather have it be too conscious than not enough.

LP: Yeah, I definitely felt it. So, as I mentioned, this is not our first automated driving episode. A lot has developed since then, 3 and 1/2 years ago. Certainly, the shuttle on campus is a new development. But what would you say are the biggest updates since then?

JT: Ooh, that.

LP: Loaded question.

JT: Huge question. Wow. Well, as with anything in kind of the high-tech world, everything seems to change every couple of years. And in 3 and 1/2 years-- here's the thing about automated vehicle technology. It is not one technology. It is many, many really challenging technologies that all have to come together into one place. So, we've got people with expertise in localization and others with path planning, and mapping, and the sensor technologies, and not just, hey, I know what a LiDAR is, but what are the newest kinds of LiDARs, and what can they do for us? And how do we integrate the very newest things into these vehicles? And then there's all of the incredible advancement in computer vision and machine learning over the last couple of years, and in neural networks, and even in the hardware to run those neural networks. And so, I'm not sure anybody can really keep up with everything here. But here's kind of what we've been focusing on at SwRI in the last few years.

And they are the tasks that we know that AVs can do, that they are perfectly suited for, and that solve real world problems. Because it's kind of fun to imagine a world where all of the vehicles on the road are automated, and everybody has a robotaxi that's just as easy to pick up as your Uber or Lyft is now. But that could be decades away. And we're seeing some neat progress. But you could really be looking at decades before that sort of thing is anywhere close to ubiquitous.

So instead, we've been looking at places where we can help our clients today. So, in practical terms, what that means is taking these proven technologies and applying them to interesting new fields, closed environments that are relatively easy to control like our campus, or an airport let's say, or a seaport, or an exclusive bus lane. And so we've been taking these technologies and making sure that they are adaptable and expandable to other domains, other contexts, so that we're not reliant on, oh, there must be lane lines, or our vehicle can't work, or we must have GPS, or our vehicle can't work. Or we must have this incredibly detailed expensive map or our vehicle can't work. We want to make sure that autonomy, our autonomy, is available to all of our clients regardless of what their domain is.

LP: So let's think about three to five years down the road. What are the seeds you're planting today that we'll be talking about in the next few years with this technology?

JT: Yeah, well, unfortunately, the one that I guarantee everybody will be talking about I can't say out loud. But trust me, you'll want another episode when we can talk about it.

LP: All right. Way to pique our interest.

JT: Yeah.

LP: Now that's a tease.

JT: Oh. I wish it weren't. I wish I could talk about it. But--

LP: And we have a lot of that here at SwRI where we're just working on stuff that we just can't talk about yet. And hopefully one day, we can. So what can you tell us?

JT: But here are a few things that I think are specific research topics that I think kind of suit the themes of this episode. So the first one is doing more effective simulation for challenging environments. As complicated as the road network is, the-- it's relatively simple compared to the entire rest of the world. And so being able to do autonomy in the rest of the world environments, all of these off road environments, is just as useful, or more useful, than doing it on the road. And unfortunately, it's much, much, much harder to test if I could deploy this vehicle potentially anywhere than just trying it out on a couple hundred acres on our facility in San Antonio isn't good enough. And so building up the capabilities to do that testing and simulation effectively, and importantly, cost efficiently. Not everybody has billions of dollars available to exhaustively test every possible thing on massive supercomputers.

How can we bring that sort of thing to all of our clients? I think that's something that we're really pushing. And in fact, we've recently filed a patent that starts us down a path of building simulations specifically for those cases where we can't possibly test the vehicle before it gets deployed. So I think that's an interesting place to go. The second one is really evaluating the performance of automated vehicles. As AV developers, we have a really good sense of how these vehicles are operating, whether they are doing what we expect them to do, how to collect the data, and the metrics, that tell us this is doing exactly the right thing, or it's doing the right thing but for the wrong reasons, or it's not doing what we expect it to do at all.

So you rode in the vehicle. But as it was driving, it was collecting an enormous amount of data that we can go back and analyze and figure out whether it was actually doing what it ought to be doing or not. And so putting those things together is actually very challenging. And a lot of institutes around the country have been looking at, how do you really build good verification and validation techniques for automated vehicles? And of course, the institute has been doing this as well. So what we're working on doing is providing that kind of capability to organizations that don't necessarily have AV experts on staff. For example, if you were running a city, you have people on staff who are experts at running cities. You don't necessarily have your autonomy engineers just sitting there to ask questions to. And that's where the institute comes in where we're able to provide that kind of analysis for others.

And the only way to do that is to draw on our many years of developing these vehicles ourselves. You can't just come in and say, oh, well, I think this should be true. It requires all this work that we've already done. And then finally, we're really looking at vehicles that do more than just drive. The shuttle and the buses we're talking about, their primary job is to drive. They take passengers from point A to point B. But there are a lot of fields where the vehicle is actually just a transportation for something else, some other implement, maybe in mining, or in agriculture, or in construction, or in other fields where maybe you want to put a robot arm on the back of your vehicle, and you need all of that to work in concert together, and maybe even in the presence of humans.

So taking these vehicles and putting them into these positions where they're not just driving but also performing tasks as they go, that's something that has been a focus and will continue to be, I think, a really rich research area for us over the next several years.

LP: All right. Really interesting. Not just driving and taking people from place to place, but shuttling goods, transporting items. So the million dollar question here. We know some cars on the road today have automation capabilities. But when will automated driving be the norm? I have to ask. I know it's a difficult question to answer. But when do you think we'll be out buying autonomous vehicles like we buy smartphones now? No big deal. Everyone has one. Jump in your automated car. Let's go.

JT: All right. First, I want to say I love the way you asked that question because saying that vehicles on the road today have automation capabilities is exactly right. There are not autonomous vehicles on the roads today with some very, very few exceptions. Instead, what we have is vehicles that take these autonomy technologies, these super high end, challenging, autonomous technologies and incorporate them as safety systems. So we talked already about adaptive cruise control. And you can add to that list things like lane keeping assistance, or lane departure warnings, depending on just how fancy your vehicle is, that can help you stay on the road, stay in your lane. And so I think that the path to ubiquitous autonomy is really through what we call advanced driver assistance systems, or ADAS. So as those systems become more powerful based on this autonomy technology, every vehicle is going to become more and more capable of driving with less and less human input.

Now my opinion is that the answer to your question is never. I think that we are rapidly approaching a day when every new vehicle has some automation capabilities, and maybe the high end of vehicles have very advanced automation capabilities. But we ought to ask the question, does every vehicle need to be fully autonomous? Is that really the future that we're headed towards? Or are we going to put autonomy in the places where it makes a ton of sense, buses, and trucks, and some passenger vehicles, and leave the rest of them to adapt this autonomous technology into safety features of the vehicle so that you're looking at cars that you can actually afford instead of the million-dollar monsters that they are right now?

LP: Well, I didn't expect a never, but I understand the answer.

JT: I think that it is certainly possible that we will get to a point where every vehicle is automated, or autonomous. But I think that it may not be necessary. And these will always be driven by a million considerations. Some of them will be technology is cool. But some of them will be considerations about what job you need your vehicle to do.

LP: So our SwRI mission is to benefit government, industry, and the public through innovative science and technology. And part of that is research and development to better humanity. So how do you feel your work is contributing to a better world?

JT: I think that is a great question. And it's one that we should all probably think about more often. Here are some specific examples from the practical side. If you're running an army and your vehicles keep getting blown up, would you rather have four soldiers in it, or zero? Regardless of what you think of armies, and wars, and militaries, I think we can all agree that zero is a better answer to that. If you're running a bus company and every once in a while, your operator makes a very human error, and it ruins a couple of thousand people's mornings, you probably want to help them out. Same thing if you're running an airport and your worker population is aging out and you literally don't have enough people to drive all of the vehicles that you need.

Then you really want to build some automated technology into those things so that you can continue serving all of those customers. So how are we contributing to a better world? Well, there are three very practical problems that exist today where autonomous technology can save lives, can improve your day every single day, and can make the planes run on time. So I'm really excited about the kind of practical implications of this. But I think there's actually a better answer that is really specific to SwRI, and what we can do collectively. Back in the '50s and '60s, all anybody could talk about was the space race. It was the most exciting thing that anybody could even think about. Everybody wanted to think about the new technology that was going to get us into space and get us to the moon. And at SwRI, we still have those people who get us into space. And they bring us back these amazing pictures of things that are totally beyond our comprehension.

And I think in the last couple of years, maybe the last couple of decades, robotics has joined space as one of those things that is just so much fun to think about. And you're asking questions about how this technology can change the world and how we're solving different problems, and where is it going to go? And everybody has these questions. And I think building that kind of excitement, capturing people's imaginations like that is a really important part of the work that we do, not just as roboticists, but specifically at SwRI, we want to make sure that we are not just doing cool research, but also making sure that we are inspiring and enlightening kind of the next generation of scientists and engineers who are going to do whatever is next. And robotics, frankly, it kind of combines the best and the hardest problems in bits and atoms. And it lets us ask the question, hey, how can we harness some of this really cool stuff and set it loose in the world to help some people out?

LP: Well, your enthusiasm about the field is really palpable. I can really feel it. And that's-- it's wonderful to have you and your expertise working on this issue and creating these amazing solutions for us and for our world. So thank you so much for being here, Jerry, and offering your voice to this conversation.

JT: Well, I am really excited to get to talk about it. I love talking to anybody who will listen. I think that the field of robotics, and not just autonomous robotics, but robotics in general, is incredibly exciting. And the ability that we have to touch not only the cool technology but also all of these different fields for all of our different clients, it's just really exciting. And so I appreciate that I get to come and talk about a little bit more of it.

And thank you to our listeners for learning along with us today. You can hear all of our Technology Today episodes and see photos and complete transcripts at podcast.swri.org. Remember to share our podcast and subscribe on your favorite podcast platform.

Want to see what else we're up to? Connect with Southwest Research Institute on Facebook, Instagram, Twitter, LinkedIn, and YouTube. Check out the Technology Today Magazine at technologytoday.swri.org. And now is a great time to become an SwRI problem solver. Visit our career page at SwRI.jobs.

Ian McKinney and Bryan Ortiz are the podcast audio engineers and editors. I am producer and host, Lisa Peña.

Thanks for listening.

[MUSIC PLAYING]