The 2007 Urban Challenge

a series about driver less cars. In the first two episodes, I covered the early concepts, which mostly involved building out an infrastructure that's outside of cars, like a smart super highway that would guide traffic to where it would need to go. And I also talked about the first two Grand Challenges that were sponsored by DARPA, that's the R

and D division of the Department of Defense. Darpest goal was to create an incentive for engineers, researchers, computer scientists to to innovate in the space in that autonomous vehicle space, because the United States Congress had decided that a third of all military ground combat vehicles should be autonomous by the year and progress was just not going at the

speed that was needed to make that happen. So dark As assessment was that the usual defense contractors just weren't making progress fast enough that that they needed to have some other way of getting new ideas, new blood into all of this, So the Grand Challenge was a way to invite numerous other engineers to solve these really difficult engineering problems. The two thousand four challenge saw no winner. No one was able to take home the prize money.

The best performing team made it less than seven and a half miles through a one forty two mile course. But the two thousand five challenge turned out much better. Five teams completed the full course and one Stanford's racing team took the top prize that year, which was more

than a million dollars. Now, there was no question that the challenges had encouraged invention and innovation, from creative computer algorithms that would guide an autonomous car through decisions on wind and how to act, to designing all new hardware that would be used to sense the environment. These challenges had pushed technological development considerably. But the driverless cars of two thousand five were meant to cross a desert course

sort of an off road course. There was tricky terrain to cross, and the cars had to make some really complex maneuvers in a few cases, but there was no need to compensate for the types of stuff that your

typical human driver encounters every day. Namely, there was no need to worry about traffic laws or traffic for that matter, and it was important for the cars to be able to detect obstacles in order to work their way around them, but there was less worry about more specific challenges like navigating through a city that has lots of potential moving

obstacles pedestrians, bicyclists, other cars. So for the next challenge, which would take place in two thousand seven, DARPA wanted to put teams to the test and have them design a car that can navigate through a simulated urban environment. Just as the agency had increased the prize money from two thousand four to two thousand and five, the the prize kitty was a million dollars for two thousand four,

it was two million for two thousand five. The urban version of the challenge up to it again to three and a half million dollars in prize money total. DARPA offered two different routes to enter the Urban Challenge. You

could apply using one of two different approaches. They called it Track A and Track B. A team submitting to participate in Track A would be eligible to receive funding from DARPA of up to a million dollars, but to qualify, the team would have to submit a detailed proposal for review and then, if selected, they would have to demonstrate the progress of the experiment of the prototype at four different milestone events in order to still be eligible for

that program. If the team could not meet the standards that DARPA had set, then they would be removed from Track A and all financial support would stop. Sixty five teams submitted for Track A consideration, DARPA would select eleven of them to participate. Track B teams were self supported, meaning they were responsible for securing the funding necessary for technology development, execution, that sort of stuff. Those teams could

seek out funding from sponsors and from other sources. So Track B teams also had to participate in certain qualifying events in order to continue the competition. There were site visits, and then there was a big qualifier that a lot of teams had to participate, and I'll talk about that a little bit later. Now, those qualifiers were meant to demonstrate that the respective teams vehicles could have a decent chance of completing a competition's task on the actual day

of the final event. So It's really all about whittling down the competitors to the ones most likely to achieve success. Track B also had several other additional rules they had to follow. This is from DARPA's actual official rule book for the two thousand seven Urban Challenge. So, for example, the team leader of a Track B team would be ineligible to participate in any other Urban Challenge team. They could only work on their own team that they were leading.

If you were not team leader, you could switch teams. You could work on one team and then switch to another one. Let's say that you're working on one team and it is unable to meet the qualifications that are necessary at a certain milestone, then you would be allowed to switch to another team if another team wanted to. Also, if you were a team leader, you had to be at least twenty one years old, and you had to be a US citizen. Non citizens could participate on teams,

but they could not be team leaders. This is another reminder that ultimately this competition was keyed into national defense. The spirit of the competition was very jovial from what I can understand, it was very cooperative. But when you take a step back and you look at the overall picture. You remember, Oh, this is so that DARPA can develop the technologies or can can encourage the development of the technologies that will be needed to power the future ground

combat vehicles. Unlike the previous Grand Challenges, DARPA even allowed teams to secure funding from government sources the previous ones. If you listen to my last episode, you know they said you couldn't get money from federal sources, but all such funding had to be explicitly approved and authorized by whatever the respective government department was for the use of

this Grand Challenge. Government sponsored teams were, however, prevented from using any technology or information that was under any kind of classified status. So you couldn't use any top secret stuff because that was against the rules. I mentioned briefly what the overall goal of the Urban Challenge was, this idea of maneuvering through a simulated urban environment. But I think it's more helpful if we actually go through all the elements that DARPA spelled out in the official rules.

This is what was sent to any interested party that was applying to be part of the Urban Challenge. This illustrates exactly how challenging the whole thing ended up being and the specific objectives of the Grand Challenge were and these are all quotes from the rule book. Complete a mission defined by an ordered series of checkpoints in a

complex route network. The vehicle will have five minutes to process a mission description, but for attempting the course, interpret static lane markings as in white or yellow lines provided with the Route Network Definition file that's R in DF, and behave in accordance with applicable traffic laws and conventions. Darp as intent is for the R and d F lane boundary descriptors to match the physical lane markings on

the ground. DARPA cannot ensure that this will be the case in all areas, and as such, the R n DF will take precedence over the physical ground markings in conflicting areas. So, in other words, DARPA would provide to each team a digital file that would represent the area

the features of the course. It wouldn't give information about where anyone was expected to travel, but it would give a layout of all the different roads in that course, and according to that file, you would program your autonomous cars behaviors to follow within those guidelines, and if the car were to come upon part of the road that was not in accordance with that same digital file, it would defer to the digital file as opposed to the

actual conditions the road. So if in one case a lane is is shown as being in one orientation and it turns out it's slightly off in reality, you go with the file, not with what reality is. This reminds us that this is in fact a simulation, because that does not fly in a real world situation. You have to conform to what is really there for all drivers. Back to the rules, the vehicles to exhibit context dependent speed control to ensure safe operation, including adherents to speed limits.

It's to exhibit safe following behavior when approaching other vehicles from behind in a traffic lane, which includes maintaining a say following distance. Is to exhibit safe check and go behavior when pulling around a stopped vehicle, pulling out of a parking spot, moving through intersections, and in situations where collision is possible, and is to stay on the road and in a legal and appropriate travel lane while en route,

including around sharp turns, through intersections, and while passing. The route definition or Route network definition file will specify the GPS coordinates of the stop signs. The R and d F specifies the location of stop lines on the ground. On paved areas. Such stop lines will be represented by a painted stop line on the pavement. Physical stop signs, however, may or may not be present at the stop line locations. Navigate safely in areas where GPS signals are partially or

entirely blocked. Uh there to follow paved and unpaved roads and stay in the lane with very sparse or low accuracy GPS way points very important because there for a long time GPS receivers were of questionable reliability, especially if you're going, like through a city with a lot of skyscrapers, for example, you might have signals blocked and then you would lose connectivity with GPS receivers. This still happens to this day, but not as frequently, at least not in

my experience. It turns out that um whether there are receivers that are using other systems to supplement that data, or the actual receivers are just more sensitive, it seems to work more reliably these days. The vehicle was to change lanes safely when legal and appropriate, such as when passing a vehicle or entering an opposing traffic lane to pass a stopped vehicle. Vehicles must not pass other vehicles

queued at an intersection. Very important. You can go around the car that's stopped in the middle of the road because of some possible mechanical failure. But you can't go around the car just because it's come to a stop at an intersection, because it might be waiting for its turn to go. Merge safely with traffic moving in one or more lanes. After stopping at an intersection, the vehicle would be able to pull across one lane of moving traffic to merge with moving traffic in the opposing lane.

It was supposed to stop safely within one meter of the stop line at a stop signed intersection and proceed without excessive delay, so less than ten seconds delay according to intersection precedents rules. It was to exhibit proper que behavior at an intersection, including stopping at a safe distance from other vehicles and stop and go procession to the

stop line without excessive delay. It was to navigate a destination toward a destination in a large open area where minimal or no GPS points are provided, as in loading dock areas or parking lots. These areas may contain fixed obstacles such as parked vehicles and moving obstacles, including other vehicles. It was to safely pull into and back out of a specified parking space in a parking lot. It was to safely execute one or more three point turning maneuvers

to affect a U turn. And it was to dynamically replan and execute the route to a destination if the primary route is blocked or impassable. These are all things that human drivers can do, and these are all things

that human drivers do on a regular basis. But as you sit there and think about what it's needed to program a machine to be able to do these things, you start to recognize the challenges here because you have to be able to detect these scenarios and then you have to be able to react appropriately based upon the information that's available to you. So it's not just you know, if this, then that there are a lot of other considerations, and it gives you a hint at the challenge that

was ahead of these teams. However, DARPA did point out a few things that were outside the scope of this challenge that teams would not need to worry about for the purposes of this competition that included the recognition of external traffic signals like traffic lights and stop signs through the use of sensors. You didn't have to worry about that.

So because they were providing information about where stop signs were based on GPS coordinates, you could program that directly into your system, so that your car quote unquote knows to stop at a certain point because it quote unquote knows there's a stop sign, there doesn't have to detect it. It's been programmed with that info, so there was no need to develop any kind of optical system that would

detect a stop sign or a traffic light. This works great if you're operating in a very limited area, right in this case, in this simulated urban environment, if you're operating in something where you've got to find border, and you know quote unquote everything about that's everything that's inside that border, you can program this kind of stuff into your vehicle directly. But this approach it's more and more

impractical the larger your area of services. So while you might be able to program this in for an area that might consist of a few simulated blocks of a city, it doesn't really work so well for a full city. You would have to go in and program the GPS coordinates for every single stop sign and every single traffic signal. Not only that, but for traffic signals, you would also have to figure out a way for your vehicle to understand when the signal had changed. But for the purposes

of the competition, they said this isn't necessary. They also said it's not necessary to be able to recognize pedestrians or build in pedestrian avoidance in your vehicle because there were not going to be any pedestrians on the simulated course. Clearly, that would also be something that would be necessary if we were to use driver lest cars in everyday life, but for the purposes of this challenge not not pertinent.

Also outside the scope or behaviors necessary for high a driving, such as high speed passing or high speed merge at an on ramp. Speed limits for the urban challenge will be thirty miles per hour or less, because again, it was supposed to be going through surface streets in a city environment, not on a highway. Driving in difficult off

road terrain is outside the scope of the program. Off Road navigation in an unpaved area, travel along roads with potholes, and travel along a dirt road are within scope, so no off road travel would be necessary, But they didn't guarantee that all the roads would be in perfect condition and that a vehicle might have some difficult terrain that that a regular driver might encounter if uh, the area they are in has different types of roads like gravel

roads and that kind of thing. No, uh, obviously we would need more than what DARPA was rolling out to have, like a driverless car that that a person could safely eide in. But you have to crawl before you can walk. We'll talk a little bit more about the Urban Challenge, but first let's take a quick break to thank our sponsor. All Right, so I've talked about the objective of the

Urban Challenge, but but what about the actual procedure. Well, every team would receive a detailed layout of the test area that would be that digital file I mentioned that would have all the accessible roads, all the parts of the course that the vehicles could drive through without penalty. That would help teams define exactly where their vehicles could

and could not go. So this is sort of like getting a really really good map of a specific part of a city that you're working with there's no information about the route on that map. It's just a map of the area. Immediately before the test, vehicles would be given a series of check points that they were told

they had to visit. The pathway between those checkpoints was undefined, however, so the vehicles could actually plot their own course from start to finish, which is kind of like getting in a friends car and you say, hey, let's go to that that pizza place that we like to go to, and you find out that they take a different route

than you would to get to that specific destination. So each vehicle would be able to plot its own route that that made the most quote unquote sense to the respective vehicle, and they'd have to do this a few times. There'd be a few different checkpoints, but the actual uh direction that the vehicle would travel would be up to the vehicle. In addition, there could be road blockages on the course and they would not be indicated by the

r n d F information. They wouldn't be on that map, so you wouldn't get a map that said, oh, by the way, this little road here is actually going to be blocked so you can't go down it. And that was important. They want to have the blockages there, but they didn't want the teams to have advanced knowledge of which routes were blocked. Just like a static map, you wouldn't know if a road was temporarily closed because of

a wreck or flooding or whatever it might be. So when you encounter that sort of obstacle, if you're a human driver and let's say you've just got a paper map, you don't have a GPS receiver, you're forced to reevaluate your plan. You're forced to look at the map and consult it and figure out a different way to get to where you're going, and you have to change things on the fly. Well, that would happen in the Urban

Challenge as well to these autonomous cars. In addition to the Route Network Data File, DARPA would provide each team with the Mission Data File MDF. This is the one that would actually have the information about the checkpoints that the vehicle was meant to visit, as well as information about the minimum and maximum speed limits for each road segment, saying, along this section of this road, you can go up to thirty miles per hour. For example, on the a section,

twenty miles hours your maximum. That kind of thing and as the rules had stated, DARPA expected all vehicles to operate within those speed parameters for each road segment. Actually, to be more specific, each car was to behave according

to the driving laws of California. Vehicles could be teleoperated, that is, they could be controlled remotely, but only for the purposes of staging the vehicle at the start of the competition, so a team could take over manual control of the vehicle in order to move it to its starting point, but then they would have to disengage from control, and at that point, each vehicle had to be under complete autonomous control with no input from the teams until

the completion of the course or if the vehicle experienced some sort of failure that would remove it from consideration. All vehicles had to be built on top of a chassis that had a documented safety record, so you couldn't just build an entirely new vehicle on a custom chassis and be eligible for the competition. Most teams would retrofit existing vehicles for their entries. In addition, all vehicles had specific parameters they had to fall inside, including weight requirements.

Um they had to be at least two thousand pounds that's about nine seven rams. They could be no heavier than a whopping thirty thousand pounds or thirteen thousand, six hundred krams. Now, for reference, the heaviest consumer car on the market that I could find is the Mercedes may BACHS six hundred Pullman Guard at eleven thousand, two hundred forty four pounds or five thousand one ms. Still very

shy of that thirty thousand mark. If you want to go with pickup trucks, that's the heaviest class of commercial, well consumer vehicle. I should say not commercial. I mean you've got massive dump trucks and stuff. They're super heavy. But if you wanted to go out and by the heaviest pickup truck out there, you would look for the International x T, which weighs in at fourteen thousand, fifty one pounds or six thousand six d. You're still less than half of what the upper weight limit is at

that point. All other heavy consumer vehicles like SUVs and vans are actually in between those two extremes. So it would have to be a really hefty vehicle to max out the weight limit of DARPA. But this reminds us again the ultimate golfer DARPA is to develop technology to devote toward automating military combat ground vehicles, which tend to

be a bit on the hefty side. In addition, the vehicles had to have a minimum wheelbase of seventy two inches, which is about one point eight meters from the front axle to the back axle, and it could have a maximum width of nine feet or two point seven four meters and a maximum height of twelve feet or about three point seven meters. Vehicles had to be able to move autonomously, both forward and in reverse. They had to

be able to make that you turn. They had to be able to turn on a typical urban street which would be about thirty feet wide or nine meters, and not climb up on the curb on either side. They had to be vehicles that traveled on tires. There it could be no treads, no tracks, nothing like that. Had to be on tires. Uh. They were not to damage the surface of a street with their passage, so it couldn't be a means of getting around that could actually tear up the road. And all the vehicles, sensors, and

technologies had to be self contained. The teams would not be allowed to set up any sort of additional sensors in the area to aid in the vehicle navigation or operation. All data processing would similarly have to take place inside the vehicle, So it was against the rules for the sensors to send data to some external computer for processing

and then have that external computer beam back instructions. And when you consider the fact that vehicles need to make split second decisions to avoid a potential accident, you probably come to the same conclusion that this is the most logical approach. You don't want to insert delay if you can avoid it. It was also against the rules to operate a vehicle that was a hazard to its environment.

DARPA stated that except for the quote normal byproducts of power generation end quote, vehicles would not be allowed to jettison any other material from them. So you couldn't do anything other than put out the normal kind of exhaust that vehicles tend to put out. Actually, DARPA specifically prohibited quote smoke screen or any other obscurant intentionally discharged end quote.

I guess they were keeping in mind that a lot of people participating in the challenge had built robots for robot battles, and maybe they were taking the competition parts super serious, but I find that pretty amusing. The idea that no no smoke screens, and it was specifically pointed

out in the rules. Teams were allowed to create a backup vehicle to the one day intended to race, and that way, if the primary vehicle should suffer some sort of setback on race day, the backup vehicle could take its place, but they had to be absolutely identical in operation and in systems, and DARPA would put any second vehicle through the same inspection and safety demonstration processes as they would the primary vehicle. All vehicles had to have

a wireless emergency stop system built into them. An emergency stop was the only permitted outside interference DARPA would allow during the final event, and it would mean removing the vehicle from consideration if you had to activate it. But if you're at the point when you need to hit the emergency stop on a vehicle, you've pretty much concluded that you are no longer in the running. You're just

trying to minimize damage at that point. So DARPA supplied every team entering into the qualifier event with a government owned e stop system, and it was up to the teams to integrate that system into their respective vehicles and designs. As for wireless signals in general, DARPA did allow vehicles to receive wireless signals, but really only for the purposes

of position determination. So like GPS satellite data, they could receive that, and vehicles could omit and sense signals as part of sensing the environments, just using li dar, you know, the laser based variant of what radar is. But teams would not be allowed to communicate with the vehicles and say, hey, hey, you want to turn left up ahead. You couldn't do that. The equipment on the vehicle had to be able to

do one other really important task. The systems on board each vehicle had to be able to accept a USB two point oh flash drive, because that's how DARPA would transfer the mission data file over to the car's system. Cars had to be able to go from the MDF loading process to full autonomy within five minutes. That was the five minutes start, so you get the MDF, the timer would start ticking and your car had to leave

the starting area within five minutes. There are a few other things that teams had to take it to consideration, you know, I mentioned earlier that cars were not allowed to stop for more than ten seconds and to stop sign once you got through the process of making sure the right of way was given. Uh. This was to avoid the situation where someone gets at a stop and they're afraid to go forward, they just keep waving other

drivers through. It's the equivalent of that. And uh, you know, making sure that the car could still operate even if GPS reception was lost, that sort of stuff. So we understand what the parameters were. How did it all turn out? Well, I'll tell you, but first I'm gonna take a sip

of tea and thank our sponsor. After all the applications, DARPA selected fifty three teams to move into the next phase of the competition, and out of those fifty three teams, DARPA would authorize thirty six of them to participate in the National Qualification Event, the n QE. That's kind of the precursor to the final race, and like the final race, the n QUE took place in a simulated urban environment. DARPA would actually make use of a retired Air Force

base in California. Is an Air Force base that was no longer in active use had been essentially abandoned, and they ended up taking over it and turning a very small part of this very large base into a couple of city blocks or simulated city blocks. The n QUE had three major components to it. In one, autonomous cars were to essentially circle a block by making left turns

at each corner. So go down the street, stop at an intersection, make a left, go down to the next intersections, stop, make a left, etcetera, etcetera, and it would try and circle the block as many times within a given amount

of time. Meanwhile, human drivers professional stunt drivers would be driving along in both directions of traffic, so the car would have to integrate itself into traffic, so not just turning left, but turning left at a point that was appropriate based upon the movement of traffic at that time. Although all cars were supposed to behave as if it were a four way stop is from from what I understand.

The next part of the inquee required cars to navigate through a suburban environment and demonstrate that the basic functions of an autonomous car navigation rerouting if it encounters an obstacle parking that kind of stuff. Uh. The third part required the vehicles to navigate through a four way stop several times as human drivers would move through the area at different speeds and that sort of thing, so to make sure that the car was behaving consistently and safely

and in accordance with law. Now, the original concept had DARPA selecting the top twenty teams from that event to move onward into the final competition, but based on the performance of the vehicles at the inn quee and some concerns about what the safety issues might be if DARPA were to put too many autonomous vehicles into the test area at the same time, they ultimately selected only eleven

teams to continue on to the final event. They felt that if they had gone with twenty, there may have been too high a concentration of autonomous cars to human controlled cars and that it would increase the likelihood of accidents happening. So the final event of the two thousand

seven challenge happened on November third, two thousand seven. Teams had six hours to complete the challenge objectives while human drivers and the other challengers were also on the course, So you had people stunt drivers driving vehicles, and you had other autonomous cars on the course while your autonomous car was trying to complete subjectives and all the same at the same time, you have to obey all the

California traffic laws. Now, out of those eleven finalists, six completed the challenge, and three of them in less than six hours. Now that doesn't sound like a lot, but it's actually an incredible achievement. Remember, in the two thousand four Grand Challenge there were no winners. In the two thousand five Grand Challenge, five teams were able to complete the course. In the two thousand seven Challenge, there were only eleven finalists in total, and more than half of them.

Six of them completed the course, three of them within the time frame. Coming in first place was Carnegie Mellon University's Boss vehicle, which had the reputation for being a bit of an aggressive driver, but the CMU team said, well, yeah, it's a race. A little less than twenty minutes after Boss's final task, after it crossed the finish line, the

Stanford Racing Team's Junior vehicle finished. So this was a flip of the two thousand five Grand Challenge, where Stanford's team came in first and then c m US team came in second and third in third place of the two thousand seven challenge was Virginia Tech's vehicle Odin. That one finished less than ten minutes after Stanford's vehicle, so it was hot on the the tail of Stanford's car.

The other three that finished were M I. T. S. Tallos vehicle which ended right around the six hour time limit, and then the little Ben Toyota Prius from the Ben Franklin Racing team that had team members from the University of Pennsylvania and Lehigh University from Philadelphia, Pennsylvania. And then you had in the sixth place Cornell's team which was called sky Net really Cute Cornell. The Cornell team in the Philadelphia team, their their cars finished after the six

hours had passed. Of the five vehicles that got a d n F as in did not finish, one of them collided with another vehicle in a minor collision, two of them ran into stationary objects, and two of them froze after hitting either an intersection or a traffic circle, and so they were disqualified. The outcome of the Urban Challenge had a lasting legacy. While the autonomous cars that competed, we're still far too limited to function on normal roads.

You would never want to unleash these on your average city street. The incredible progress that had been made in the field showed the driver less cars might actually be a possibility, and maybe not that far off. But you could also argue that this demonstration caused some people to project an unrealistic expectation of when we might see autonomous cars get a wide rollout on roads around the world. And you can kind of understand where that would come from.

Right The two thousand five Grand Challenge was an off road desert course. You didn't have to worry about making your way through traffic or around stalled vehicles or anything like that. The advance you would see in two thousand seven suggested that, hey, maybe there's just a few tweaks that need to happen, and then we're gonna have driverless

cars everywhere. But that ignores the fact that the challenges were really, really difficult, and the teams that met those challenges we're still struggling with some really tough problems, and they didn't have to take into consideration stuff like pedestrians. So it would take a lot more than just refining this approach to make a vehicle roadworthy. Now, one of the technologies I want to mention before I close is lidar.

I've I've mentioned it once or twice, but I really want to take a moment to point it out because it was one of the standout developments the various grand challenges. So lighter works on the basic principle as radar. Right, radar sends out radio signals and then there's a receiver, and the receiver picks up the reflected radio signals, and through measuring the time difference between when they were sent out and when they were coming back, you can tell

how far away an object is. There's also Doppler shift and stuff like that, but we've talked about that before, so I'm not going to go into it here. Ldar does the same sort of thing, but uses lasers instead of radio waves. So you have a receiver, a sensor, a photo cell essentially that's looking for that frequency of light and can pick up those, uh, those reflections and then be able to project on a digital map where an object is. But the problem was that the early

use of lidar was very primitive, it was very limited. Uh. They really just could be sent out in in a single direction and you would know that there was an object out there, but you weren't getting very high resolution results from this. And that changed in two thousand five, and really changed in two thousand seven thanks to a guy named Dave Hall. Dave Hall had found out about

lidar from a guy named Jim McBride who worked with Ford. Now, when Dave Hall looked into lidar, he found it really interesting. And Dave Hall he was the founder of a company um still is actually that was called Vladine still is called Vladine, and it was in the business of doing audio equipment. So he was getting bored with doing audio equipment. He thought, well, I'll play with this autonomous car area for a while. He had also participated in various robot

war challenges and competitions. So he goes and looks into lidar and he thinks, well, this has got some potential, but it's really limited based on how people are using it right now. And while he hadn't even known what lidar was, in two thousand four and two thousand five he invented a lighter tool that would become incredibly important

for all the urban challenge teams. He had what was the equivalent of sixty four lasers that were scanning outward, and they were mounted on a spinning scanner, So the scanner would spin in a circle and you would get a three hundred sixty degree scan of the area around a vehicle, and that would allow him to gather information about all the objects surrounding a vehicle, their distance from the vehicle, even whether or not those objects are moving

toward or away from the vehicle. Now, that wouldn't make him win the two thou five challenge. He didn't, but his tool got a lot of attention. People noticed, Hey, this is a really cool technology. It's incredibly innovative. So in two thousand and seven, a whole bunch of the vehicles that were submitted as part of that Urban Challenge sported a lightar scanner that was created by Hall's company, Velo Dine, And in fact, Velo Dine is a leading

manufacturer of light oar systems speci scifically for autonomous car use. So, uh, he became a big shot billionaire from that. It was

a really amazing development. Now, in the next episode, I'm going to talk a little bit more about where driver lest cars went after the Urban Challenge figuratively speaking, And then on the episode following that, we're going to look more into the philosophy that uh that Thomas cars are great, and we're also gonna look at some opposing viewpoints, not necessarily that a Thomas cars are bad, but perhaps suggesting that we're not nearly as far along as some futurists

would like us to believe. So that's gonna be the next couple of episodes, But for now we're going to wrap this up. If you guys have suggestions for a future show topics, whether it's for a single episode or maybe it's an arc like the one where currently in let me know. You can send me an email. The address is tech stuff at how stuff works dot com, or you can head on over to our website that's tech Stuff podcast dot com and you can let us

know through the various links to our social media. Also, don't forget our merchandise store over at t public dot com slash tech stuff. Remember every purchase you make goes to help the show. We greatly appreciate it, and I'll talk to you again really soon for more on this and thousands of other topics. Because it how stuff works dot com