WEBVTT - The DARPA Grand Challenges 0:00:04.120 --> 0:00:07.160 Get in touch with technology with tech Stuff from how 0:00:07.200 --> 0:00:13.640 stuff Works dot com. Hey there, and welcome to tech Stuff. 0:00:13.680 --> 0:00:16.959 I'm your host, Jonathan Strickland. I'm an executive producer with 0:00:17.040 --> 0:00:19.279 how Stuff Works in I heart radio and I love 0:00:19.360 --> 0:00:23.000 all things tech. And in our last episode, if you 0:00:23.040 --> 0:00:24.720 haven't heard that, you should probably go and listen to it. 0:00:24.760 --> 0:00:28.560 But I left off with Ernest Dickmans, the German engineer 0:00:28.960 --> 0:00:32.879 who had done amazing work with dynamic computer vision and 0:00:33.000 --> 0:00:36.760 vehicle automation in the nineteen eighties and nineties. But that 0:00:36.840 --> 0:00:39.760 work would only go as far as the sophistication of 0:00:39.800 --> 0:00:43.239 the technology of the time would allow, and funding for 0:00:43.280 --> 0:00:47.640 AI work had gotten pretty darn scarce in the eighties 0:00:47.640 --> 0:00:51.080 and nineties. So while he and his team were prepping 0:00:51.159 --> 0:00:55.319 for their ninety four demonstrations that were so impressive, there 0:00:55.360 --> 0:00:58.480 was similar work going on in the United States. From 0:00:59.640 --> 0:01:04.720 to nine. Research teams took advantage of six hundred fifty 0:01:04.760 --> 0:01:08.400 million dollars of funding created by the US government through 0:01:08.400 --> 0:01:13.039 the Intermodal Surface Transportation Efficiency Act. That act was meant 0:01:13.080 --> 0:01:18.840 to support planning for several transportation projects, uh including the 0:01:18.880 --> 0:01:21.840 creation of new rail lines between various cities. At the 0:01:21.840 --> 0:01:26.280 autonomous research was really just one piece of this larger legislation. 0:01:26.360 --> 0:01:28.240 And also just to let you guys know, in case 0:01:28.280 --> 0:01:31.600 you're curious, many of those projects did not receive enough 0:01:31.600 --> 0:01:35.199 funding to really accomplish their goals, so they never really 0:01:35.240 --> 0:01:38.839 self fruition. It's one of the downsides of government projects 0:01:38.880 --> 0:01:42.600 is that the budgets are too low. Nothing of real 0:01:42.760 --> 0:01:45.720 import tends to happen. People might get paid by some 0:01:45.760 --> 0:01:50.640 of that money, but the actual projects don't necessarily, you know, happen. 0:01:51.920 --> 0:01:55.360 On the driver list car side, nine organizations began to 0:01:55.360 --> 0:01:58.840 work on technologies to make driverless cars a reality, but 0:01:58.920 --> 0:02:02.480 the deadline for meeting those goals in a suitable demonstration 0:02:02.920 --> 0:02:06.919 was and from what I can tell, not a whole 0:02:06.960 --> 0:02:09.520 lot of progress was made in that time, or at 0:02:09.600 --> 0:02:12.640 least not enough to end up with an impressive demonstration. 0:02:12.760 --> 0:02:16.280 That marked a point in history that we would say, 0:02:16.360 --> 0:02:20.720 this is really important in driver lest cars. But we're 0:02:20.880 --> 0:02:24.720 continued on the various systems that would be necessary to 0:02:24.840 --> 0:02:28.799 make driverless cars. But you would argue that driverless car 0:02:28.880 --> 0:02:32.280 research and development as a whole, as a as a 0:02:32.320 --> 0:02:35.720 goal in of itself, was kind of shelved for a while. 0:02:36.320 --> 0:02:39.880 There were people working on various parts of the problem. 0:02:39.880 --> 0:02:42.239 Most of them were working in fields that ultimately would 0:02:42.280 --> 0:02:45.520 prove important to the development of driverless cars. But again, 0:02:45.960 --> 0:02:49.480 they were doing this for other reasons, and it wasn't 0:02:49.480 --> 0:02:53.680 necessarily as uh an in goal of creating an autonomous vehicle. 0:02:54.680 --> 0:02:58.120 Much of the excitement and money around the concept had 0:02:58.200 --> 0:03:02.520 died down by the mid to late nineties. Some of 0:03:02.560 --> 0:03:05.679 those related research projects would end up growing out of 0:03:05.720 --> 0:03:08.680 a US backed R and D project called the Strategic 0:03:08.840 --> 0:03:12.200 Computing Initiative or s c I now that actually got 0:03:12.200 --> 0:03:17.440 started in ninety three. The goal of the initiative was 0:03:17.560 --> 0:03:21.160 incredibly ambitious. That's a very kind way of putting it. 0:03:21.160 --> 0:03:26.720 I'd actually say it was unrealistically ambitious. It was to 0:03:26.840 --> 0:03:30.520 develop the technology to create a machine learning system capable 0:03:30.639 --> 0:03:35.520 of running quote, ten billion instructions per second to see here, 0:03:35.840 --> 0:03:39.440 speak and think like a human. The degree of integration 0:03:39.880 --> 0:03:43.560 required would rival that achieved by the human brain, the 0:03:43.600 --> 0:03:47.720 most complex instrument known to man. End quote. That was 0:03:47.720 --> 0:03:51.400 actually written in an account by Alex Rowland and Philip 0:03:51.440 --> 0:03:54.720 Shiman for the m I T Press about the s 0:03:54.720 --> 0:03:58.840 c I project. The goal was to achieve that in 0:03:59.040 --> 0:04:02.160 just a deck aid so starting in nineteen eighty three, 0:04:02.240 --> 0:04:05.120 the plan was that by ninetee, we're gonna have a 0:04:05.120 --> 0:04:10.840 computer that can think and and experience like a human can. Wowsers. 0:04:11.680 --> 0:04:14.160 So it's pretty obvious by the late eighties that this 0:04:14.280 --> 0:04:16.560 was not gonna happen. They were not going to hit 0:04:16.600 --> 0:04:20.600 that goal, and in retrospect, we would probably call this 0:04:20.680 --> 0:04:24.080 the result of hubrists. The human brain is far more 0:04:24.240 --> 0:04:28.800 complicated and technology is far more limited than we gave 0:04:28.960 --> 0:04:33.440 either credit for back in nineteen eighty three. Nevertheless, the 0:04:33.480 --> 0:04:38.120 Defense Advanced Research Projects Agency or DARPA and the Department 0:04:38.160 --> 0:04:42.159 of Defense, which is the department that oversees DARPA, poured 0:04:42.200 --> 0:04:46.320 about a billion dollars into funding various programs throughout the 0:04:46.360 --> 0:04:50.400 United States in an effort to achieve this goal. And 0:04:50.440 --> 0:04:52.880 while we did not get a computer that thinks like 0:04:52.920 --> 0:04:57.760 a person out of this whole process, many influential computer 0:04:57.880 --> 0:05:02.560 scientists and research projects were able to advance our capabilities 0:05:02.640 --> 0:05:07.280 and understanding through their work which got funding from this project. 0:05:07.320 --> 0:05:12.080 So we didn't get what the project was aimed to produce, 0:05:12.800 --> 0:05:16.480 but we did benefit from it. Some of that work 0:05:16.480 --> 0:05:19.800 would become really important for the next big DARPA initiative 0:05:19.880 --> 0:05:22.840 that I'm going to cover, and that is the Grand Challenge. 0:05:23.560 --> 0:05:27.000 The history of the Grand Challenge dates back to two 0:05:27.000 --> 0:05:31.040 thousand one. At that time, the United States Congress had 0:05:31.080 --> 0:05:34.520 its own challenge for the various branches of the U. S. Military. 0:05:34.839 --> 0:05:38.440 The US Congress said, we want you to develop the 0:05:38.520 --> 0:05:43.400 technology necessary to allow one third of all military ground 0:05:43.600 --> 0:05:48.400 combat vehicles to be unscrewed by two thousand fifteen. In 0:05:48.400 --> 0:05:56.039 other words, to have autonomous ground combat vehicles one third 0:05:56.080 --> 0:05:58.880 of all of them. The purpose, obviously was to keep 0:05:59.040 --> 0:06:01.760 soldiers out of harm's way as much as possible. That 0:06:01.839 --> 0:06:05.920 if you could have these ground combat vehicles operate autonomously, 0:06:06.480 --> 0:06:09.920 then if one gets destroyed, that's a huge amount of 0:06:09.960 --> 0:06:12.280 money gone down the drain, but no one dies, at 0:06:12.360 --> 0:06:16.359 least no one on your side dies. Achieving this goal 0:06:16.839 --> 0:06:21.400 would also be super difficult to do. Technology just wasn't 0:06:21.440 --> 0:06:24.760 where it would need to be by itself. The defense 0:06:24.839 --> 0:06:28.159 contractors that DARPA would work with on these sort of 0:06:28.200 --> 0:06:32.599 solutions weren't making the progress necessary in order to meet 0:06:32.640 --> 0:06:36.000 that two thousand fifteen goal, and DARPA recognized this early on. 0:06:36.080 --> 0:06:40.000 They said, this is just not gonna happen. So in 0:06:40.040 --> 0:06:42.800 the final report for the two thousand four Grand Challenge, 0:06:42.839 --> 0:06:46.240 which was the first of the three Grand Challenges around 0:06:46.400 --> 0:06:52.279 autonomous cars, the logic that the agency laid out to 0:06:53.160 --> 0:06:58.120 justify this challenge was this, while there have been a 0:06:58.240 --> 0:07:03.240 number of significant technical breakthroughs leading to robust unmanned air 0:07:03.360 --> 0:07:08.200 vehicles that US forces used today, progress in unmanned autonomous 0:07:08.240 --> 0:07:12.080 ground vehicle technology has not occurred at a similar rate. 0:07:12.680 --> 0:07:15.920 Vehicle operations in a ground environment are a much more 0:07:16.040 --> 0:07:20.600 difficult challenge due to terrain, man made obstacles, and weather 0:07:21.520 --> 0:07:25.760 that's just scratching the surface. They actually were pretty you know, 0:07:25.920 --> 0:07:29.400 generous in that regard, because as it turns out, there 0:07:29.400 --> 0:07:32.000 are a lot of other factors that make this a 0:07:32.120 --> 0:07:37.120 really difficult problem. And I think one other thing quick 0:07:37.160 --> 0:07:40.760 tangent that I think is important here is DARPA pointing out, yes, 0:07:40.840 --> 0:07:45.200 we came up with autonomous systems for aircraft ages ago. 0:07:45.560 --> 0:07:48.440 We have lots of them. We have unmanned drones that 0:07:48.480 --> 0:07:51.760 we can fly, but we haven't really done that with cars. 0:07:51.880 --> 0:07:55.480 I think that's also a good reminder that technology does 0:07:55.520 --> 0:07:59.600 not all progress at the same accelerated rate. We have 0:07:59.720 --> 0:08:02.000 more is law, which has kind of conditioned us to 0:08:02.040 --> 0:08:06.840 think about our technology advancing rapidly over the course of 0:08:06.880 --> 0:08:11.120 every two years. But that doesn't apply to every technology. 0:08:11.160 --> 0:08:17.360 It applies specifically these days to computational power, and originally 0:08:18.200 --> 0:08:20.600 only dealt with the number of transistors you could fit 0:08:20.640 --> 0:08:24.840 on a square inch of silicon wafer. So this is 0:08:24.840 --> 0:08:27.240 good to remember because I think a lot of futurists 0:08:27.280 --> 0:08:30.680 just kind of apply More's law to all technologies and 0:08:30.760 --> 0:08:34.120 just assume that everything is accelerated at that saint or 0:08:34.160 --> 0:08:37.400 everything is moving at that same accelerated rate. Back to 0:08:37.600 --> 0:08:42.520 autonomous cars. Jose Negron, who worked at DARPA at the 0:08:42.559 --> 0:08:46.320 time of the Grand Challenges, went to the director of 0:08:46.400 --> 0:08:49.839 DARPA with an idea, and he said, there are a 0:08:49.880 --> 0:08:53.080 lot of people out there who could potentially contribute to 0:08:53.160 --> 0:08:58.920 the advancement of technology we're going to need for autonomous cars. 0:08:58.960 --> 0:09:03.120 But we normally would work with them because they're independent innovators, 0:09:03.160 --> 0:09:05.880 They work with smaller groups, they aren't part of defense 0:09:05.960 --> 0:09:09.600 contractor companies, and so they don't often get a chance 0:09:09.640 --> 0:09:11.800 to do any sort of work for the Department of Defense. 0:09:11.840 --> 0:09:14.000 But we could take advantage of them if we are 0:09:14.080 --> 0:09:16.400 if we give them the chance to participate, we could 0:09:16.480 --> 0:09:19.760 benefit from that and they might create the breakthroughs that 0:09:19.800 --> 0:09:24.280 we need to meet that two thousand fifteen deadline. And so, 0:09:24.520 --> 0:09:27.960 after a couple of years of funding various research projects, 0:09:28.000 --> 0:09:31.319 the then director of DARPA was a guy named Tony Teather. 0:09:31.679 --> 0:09:35.160 Anthony Tather announced a new initiative and it was for 0:09:35.240 --> 0:09:39.480 an open challenge to any teams that wanted to participate. 0:09:39.559 --> 0:09:44.559 Anyone could apply to be part of this challenge. Specifically, 0:09:45.240 --> 0:09:49.320 the challenge was to develop autonomous vehicle technology and incorporate 0:09:49.400 --> 0:09:53.240 it into a vehicle that could traverse a one forty 0:09:53.280 --> 0:09:58.240 two mile course without human intervention. So from the start 0:09:58.280 --> 0:10:00.280 of the course to the end of it, the vehicle 0:10:00.320 --> 0:10:03.320 would have to be operating under its own abilities, its 0:10:03.320 --> 0:10:07.000 own power. The team to complete this course in the 0:10:07.040 --> 0:10:12.760 shortest amount of time would take home a million dollars. Now, initially, 0:10:13.360 --> 0:10:16.440 Anthony Tather was skeptical that this announcement was going to 0:10:16.520 --> 0:10:19.520 draw much interest. He said, we'll get a shot, but 0:10:19.559 --> 0:10:21.640 I don't think very many people are going to respond. 0:10:22.240 --> 0:10:25.840 DARPA scheduled a kickoff event at the Peterson Automotive Museum 0:10:25.840 --> 0:10:30.400 in Los Angeles, California for anyone interested in participating. On 0:10:30.480 --> 0:10:33.599 the day of that event, Anthony Tather arrived at the 0:10:33.679 --> 0:10:37.200 venue a half hour before the scheduled time to start, 0:10:37.720 --> 0:10:40.360 and he discovered there was a line wrapping around the 0:10:40.400 --> 0:10:44.240 block of people who were interested in participating. According to 0:10:44.240 --> 0:10:46.920 the final report for the two thousand four Challenge, more 0:10:46.960 --> 0:10:50.479 than four hundred people showed up to that first conference. 0:10:51.360 --> 0:10:55.120 The race course for this two thousand four event would 0:10:55.200 --> 0:10:59.760 end up being incredibly challenging. A guy named Salve Fish 0:11:00.120 --> 0:11:03.280 actually designed the course and he made it really tough. 0:11:03.720 --> 0:11:07.640 Vehicles would travel through the desert, including trips up and 0:11:07.760 --> 0:11:11.760 down hills and through switchbacks. Sometimes the road would narrow 0:11:11.800 --> 0:11:14.600 down to about ten feet wide. There are points where 0:11:14.640 --> 0:11:17.520 the road would drop off several feet to one side 0:11:17.800 --> 0:11:21.520 or the other. The course crossed railroad tracks. There was 0:11:21.559 --> 0:11:25.320 the chance that the cars might encounter animals along the course. 0:11:26.040 --> 0:11:29.920 DARPA had some professional drivers actually go down this course 0:11:30.000 --> 0:11:33.040 after it had been established to kind of give their 0:11:33.160 --> 0:11:36.560 estimation of it, and they said it was fairly challenging 0:11:36.600 --> 0:11:40.640 even for a professional trained off road driver. Any autonomous 0:11:40.640 --> 0:11:42.880 car that was to complete this course was expected to 0:11:42.920 --> 0:11:46.960 do so within ten hours of starting, so there's a 0:11:47.000 --> 0:11:51.880 ten hour time limit essentially from start to finish. So 0:11:51.920 --> 0:11:54.920 why did they make it so challenging. Well, keep in mind, 0:11:55.040 --> 0:11:58.360 the ultimate goal of this and the following challenges was 0:11:58.400 --> 0:12:01.840 to encourage the development of technology that would allow the 0:12:01.920 --> 0:12:06.439 US military to build autonomous military vehicles. And the US 0:12:06.720 --> 0:12:10.440 was and still is, heavily involved in operations in the 0:12:10.440 --> 0:12:13.400 Middle East, and so the course was partly designed to 0:12:13.480 --> 0:12:17.440 mimic the conditions that military vehicles might regularly encounter in 0:12:17.480 --> 0:12:21.320 that part of the world. And while on a test 0:12:21.440 --> 0:12:24.680 we might say, while that's really challenging, in a real 0:12:24.720 --> 0:12:30.319 world scenario, we don't really. We can't change the parameters. 0:12:30.400 --> 0:12:33.280 It's it is what it is. So it needed to 0:12:33.320 --> 0:12:35.880 be really hard. It needed to be hard also to 0:12:35.960 --> 0:12:40.319 spur on that innovation. The course would begin in Barstow, California, 0:12:40.360 --> 0:12:43.640 and it would end in prem Nevada, passing through the 0:12:43.640 --> 0:12:47.360 Mojave Desert. All right, I'm gonna talk about how this 0:12:47.480 --> 0:12:50.080 challenge unfolded, but first let's take a quick break to 0:12:50.160 --> 0:13:03.200 thank our sponsor. The teams had just one year to develop, build, test, 0:13:03.400 --> 0:13:06.800 and complete their technologies. So, like I said, this was 0:13:07.240 --> 0:13:13.560 an enormous challenge. Competing teams included universities, there were research facilities, 0:13:13.880 --> 0:13:17.040 There are some company teams. There were even some high 0:13:17.080 --> 0:13:23.559 school teams that applied. DARPA received one hundred six applications. Now, 0:13:23.600 --> 0:13:26.760 not everyone who attended the kickoff event went through with 0:13:26.840 --> 0:13:30.200 the step of applying, and the rules did restrict who 0:13:30.200 --> 0:13:34.199 could participate. They said US federal government organizations could not 0:13:34.400 --> 0:13:38.760 participate in this event. Federal employees could participate, but only 0:13:38.840 --> 0:13:41.600 as a private citizen on a team. They could not 0:13:41.840 --> 0:13:46.160 represent a federally backed team. Federal funding wasn't allowed either, 0:13:46.520 --> 0:13:50.600 and teams were prohibited from using government owned equipment, with 0:13:50.679 --> 0:13:53.880 one exception, which was that if the government were to 0:13:54.080 --> 0:13:58.439 offer equipment to all teams, that's fine, but no team 0:13:58.440 --> 0:14:02.000 could just take advantage of government owned equipment if they 0:14:02.000 --> 0:14:04.800 were the only team that had access to it. DARPA 0:14:04.880 --> 0:14:07.920 set up a website for the competition that included a 0:14:07.920 --> 0:14:11.360 forum and which interested parties could post questions. They could 0:14:11.400 --> 0:14:15.080 share advice, They could share strategies for how they were 0:14:15.080 --> 0:14:19.640 going to solve really hard problems like obstacle detection, path finding, 0:14:20.040 --> 0:14:24.200 position location, the control software that would be needed for 0:14:24.280 --> 0:14:28.720 the vehicle itself. Teams could also help each other find sponsors, 0:14:29.080 --> 0:14:32.400 to help get money and funding to cover expenses, and 0:14:32.480 --> 0:14:35.200 it sounds to me like it was a fairly collaborative 0:14:35.240 --> 0:14:39.480 space despite the competitive nature of the objective. The general 0:14:39.520 --> 0:14:43.800 strategy for most teams was to pair several technologies together 0:14:44.000 --> 0:14:48.360 to create an autonomous car, so they weren't necessarily developing 0:14:48.440 --> 0:14:51.000 all these tools themselves. In some cases, they were making 0:14:51.120 --> 0:14:53.840 use of existing tools, so a lot of them would 0:14:53.920 --> 0:14:57.480 use stuff like a GPS receiver. I've talked about GPS 0:14:57.600 --> 0:15:00.360 extensively not too long ago, so I'm not going to 0:15:00.440 --> 0:15:05.200 go over that again. Also optical sensors, so essentially cameras 0:15:05.200 --> 0:15:10.200 and similar sensors, laser range finders, lidar systems, things like that. 0:15:10.680 --> 0:15:13.880 Some teams spent most of their time developing the computer 0:15:13.920 --> 0:15:18.000 programs that would make decisions based on the incoming data 0:15:18.080 --> 0:15:23.680 from these various sensors, such as whether a vehicle should 0:15:23.680 --> 0:15:26.640 accelerate to climb a steep hill or to break to 0:15:26.760 --> 0:15:29.400 avoid going off the edge of the road, or to 0:15:29.560 --> 0:15:33.200 steer in a particular direction. One team created a system 0:15:33.200 --> 0:15:36.120 that used a voting mechanism to decide on what to 0:15:36.200 --> 0:15:39.880 do at any given point, and it would wait various 0:15:39.920 --> 0:15:43.080 factors in that decision making process so that the computer 0:15:43.160 --> 0:15:46.280 could arrive at what it thought was the best course 0:15:46.320 --> 0:15:50.480 of action for any given set of circumstances. DARPA chose 0:15:50.880 --> 0:15:55.320 twenty five of the teams after reviewing their technical applications, 0:15:56.120 --> 0:15:59.720 and immediately upon looking at the technical applications, they were 0:15:59.760 --> 0:16:03.080 able to select nineteen teams by itself, They said, all right, 0:16:03.160 --> 0:16:08.880 these nineteen really have a solid application. For the other six. 0:16:08.920 --> 0:16:12.760 They visited several of those teams for site visits, kind 0:16:12.800 --> 0:16:14.600 of to get an idea of how far along each 0:16:14.640 --> 0:16:18.360 project was, how realistic the chances were that the team 0:16:18.440 --> 0:16:21.680 was going to be ready by race time, and generally 0:16:21.680 --> 0:16:23.520 just get a gut feeling for who had a real 0:16:23.680 --> 0:16:27.840 chance to to do something special at this competition. The 0:16:27.840 --> 0:16:31.560 teams were invited to a qualifying event at the California 0:16:31.600 --> 0:16:36.160 Speedway in Fontana, California, and of the twenty five that 0:16:36.200 --> 0:16:40.320 were selected, twenty one showed up. Four teams did not 0:16:40.520 --> 0:16:44.440 make it to the qualifying event. The qualifying event tested 0:16:44.520 --> 0:16:49.960 vehicles in many ways, including responsiveness, speed, safety. They went 0:16:50.000 --> 0:16:54.320 through some areas that would simulate what the cars would 0:16:54.400 --> 0:16:59.320 encounter on the actual race. Only seven teams were actually 0:16:59.360 --> 0:17:03.000 able to navig Gate the full qualifier, but DARPA selected 0:17:03.040 --> 0:17:05.880 another eight on top of the seven that made it through. 0:17:06.400 --> 0:17:09.639 They were close enough for consideration, and so the on 0:17:10.160 --> 0:17:13.920 plus teams that had applied the competition was narrowed down 0:17:13.960 --> 0:17:18.000 to fifteen teams. For the big race itself, DARPA had 0:17:18.080 --> 0:17:22.159 some other special rules. One was that each team would 0:17:22.160 --> 0:17:25.840 not be given any information about the course until about 0:17:25.920 --> 0:17:29.520 two hours before the actual start time. Another was that 0:17:30.119 --> 0:17:33.960 a human driven chase vehicle would follow each of the 0:17:34.000 --> 0:17:37.560 autonomous vehicles, and inside that chase vehicle there would be 0:17:37.560 --> 0:17:40.800 a laptop operated by a team member that would allow 0:17:40.880 --> 0:17:44.959 them to activate a kill switch. So if an autonomous 0:17:45.000 --> 0:17:48.320 car or motorcycle more on that in a second where 0:17:48.359 --> 0:17:51.400 to go nuts, start careening off course and driving out 0:17:51.400 --> 0:17:54.399 of control, a human in the chase vehicle could hit 0:17:54.400 --> 0:17:57.680 the kill switch and that would shut down the target 0:17:57.800 --> 0:18:01.639 vehicle before any real calamity could occur. The cars weren't 0:18:01.640 --> 0:18:03.920 meant all to leave at the same time. It wasn't 0:18:03.960 --> 0:18:06.600 like that kind of race where they all lined up 0:18:06.640 --> 0:18:10.200 and then the flag dropped. There was a staggered departure 0:18:10.240 --> 0:18:13.119 time UH, and it was staggered by several minutes so 0:18:13.200 --> 0:18:16.399 the goal was to have the fastest time from start 0:18:16.440 --> 0:18:20.399 to finish, but not necessarily to physically race against another 0:18:20.480 --> 0:18:24.680 vehicle in real time, although there was the potential that 0:18:25.000 --> 0:18:27.600 a car could catch up with whichever vehicles were in 0:18:27.640 --> 0:18:29.639 front of it and pass it that way, that was 0:18:29.680 --> 0:18:35.320 a possibility. DARPA told CNN that about five people were 0:18:35.400 --> 0:18:38.960 working on the race in total, some of them were volunteers, 0:18:39.040 --> 0:18:42.160 some of them were DARPEST staff members, and that the 0:18:42.280 --> 0:18:47.600 race also employed several wreckers tow trucks. Essentially, so if 0:18:47.720 --> 0:18:50.760 vehicles got stuck, a wrecker could come in and pull 0:18:50.840 --> 0:18:54.440 them out of the way, and that vehicle would obviously 0:18:54.480 --> 0:18:59.720 be uh. It would be eliminated from the race. Uh 0:18:59.760 --> 0:19:02.960 and whoever pass that finished line in the shortest amount 0:19:03.000 --> 0:19:05.560 of time would walk away with a cool one million 0:19:05.640 --> 0:19:08.760 dollars in prize money. Now, one of the participants in 0:19:08.760 --> 0:19:11.720 this race, just as a moment of interest here was 0:19:11.760 --> 0:19:15.800 a guy named Anthony Lewandowski. And I've talked about Lewandowski 0:19:15.960 --> 0:19:18.120 in tech stuff. I think it was a year ago 0:19:18.160 --> 0:19:21.920 when I was talking about Lewandowski. His approach was unique 0:19:21.960 --> 0:19:25.520 in that he brought a self balancing motorcycle to the 0:19:25.600 --> 0:19:30.560 event rather than a four wheeled vehicle. Lewandowski became somewhat 0:19:30.680 --> 0:19:36.560 infamous for departing Google he worked in their autonomous vehicle 0:19:36.960 --> 0:19:40.120 department and then going to work for Uber and then 0:19:40.160 --> 0:19:45.840 allegedly bringing along with him some proprietary information from Google. 0:19:46.680 --> 0:19:49.080 So that was what I was talking about the last 0:19:49.080 --> 0:19:53.160 time I chatted about Lewandowski. The fifteen teams, in order 0:19:53.280 --> 0:19:57.080 of their start time were the Red Team from Carnegie 0:19:57.119 --> 0:20:03.160 Mellon University, PSI Autonics two from Thousand Oaks, California, Team 0:20:03.280 --> 0:20:09.080 cal Tech from California Technical University, Digital Auto Drive or 0:20:09.240 --> 0:20:14.760 Dad from Morgan Hill, California, Virginia Tech which is from Virginia, 0:20:15.240 --> 0:20:20.560 Axion Racing from Westlake Village, California, Team Cajun Bot which 0:20:20.640 --> 0:20:25.359 might have my favorite name there from Lafayette, Louisiana, Team 0:20:25.480 --> 0:20:30.439 Endsco from Falls Church, Virginia, Team SIMAR that's see I 0:20:30.800 --> 0:20:33.399 M A R from Gainesville, Florida, and Logan, Utah. It 0:20:33.440 --> 0:20:37.760 was a joint project the Palos Verdes High School Road 0:20:37.800 --> 0:20:42.280 Warriors from Palace Very Days Estates, California, PSI A Tonics 0:20:42.400 --> 0:20:46.760 one from Thousand Oaks, so they actually left the starting 0:20:46.800 --> 0:20:51.520 line well after PSI A Tonics too uh. Team Tera 0:20:51.600 --> 0:20:56.520 Max from Oshkosh, Wisconsin, Team Tara Hawk from Guardina, California, 0:20:56.840 --> 0:21:00.359 the Golem Group from Santa Monica, California, and the Blue 0:21:00.359 --> 0:21:04.080 Team from Berkeley, California. At six thirty in the morning 0:21:04.520 --> 0:21:09.439 on Saturday, March two thousand four, the Red Team's vehicle 0:21:09.640 --> 0:21:13.680 called Sandstorm so It's a Sandstorm from Carnegie Mellon became 0:21:13.720 --> 0:21:17.160 the first of the driver lest cars to tackle this course. Now, 0:21:17.240 --> 0:21:20.480 it would get about seven and a half miles down 0:21:20.520 --> 0:21:23.320 the road, but then it got stuck while trying to 0:21:23.400 --> 0:21:28.240 navigate a tight switchback. It was the most successful of 0:21:28.280 --> 0:21:31.119 all the vehicles, seven and a half miles out of 0:21:31.119 --> 0:21:35.119 a forty two So out of the fifteen cars that 0:21:35.680 --> 0:21:39.080 made it through to the competition, two of them withdrew 0:21:39.280 --> 0:21:41.199 prior to the start of the race. That would be 0:21:41.200 --> 0:21:44.520 Tara Hawk and Blue Team. Four of them didn't make 0:21:44.520 --> 0:21:47.360 it out of the starting area because of various problems 0:21:47.480 --> 0:21:51.280 ranging from steering abnormalities to colliding with wall, which I 0:21:51.280 --> 0:21:54.960 guess in its way as its own steering abnormality. Then 0:21:55.000 --> 0:21:57.480 you had Team Ends Goes vehicle which made it point 0:21:57.520 --> 0:22:02.000 two miles before it flipped over. Team Simar's car got 0:22:02.040 --> 0:22:05.800 tangled up by wire. Half a mile in the Golden 0:22:05.840 --> 0:22:08.960 Group's vehicle got stuck on an incline and couldn't provide 0:22:09.040 --> 0:22:12.600 enough throttle to overcome it. Team Terra Max's vehicle kept 0:22:12.600 --> 0:22:16.159 detecting bushes and eventually just stopped moving forward after it 0:22:16.480 --> 0:22:20.359 thought that there were bushes everywhere. I guess wouldn't budge. 0:22:20.800 --> 0:22:23.200 Team cal Tex car went off course and through a 0:22:23.200 --> 0:22:25.480 fence and could not get back on course, and so 0:22:25.520 --> 0:22:28.960 it was disqualified. So none of the vehicles were able 0:22:29.000 --> 0:22:32.399 to complete this on two mile course, not by a 0:22:32.520 --> 0:22:37.879 long shot. No one collected the million dollars. But DARPA 0:22:38.000 --> 0:22:40.600 wasn't about to give up there. For one thing, the 0:22:40.640 --> 0:22:46.280 agency had created an incredibly ambitious challenge, a really hard one, 0:22:46.720 --> 0:22:49.800 and for another, the teams were learning from their mistakes, 0:22:50.200 --> 0:22:53.600 and so DARPA chose to announce a second challenge that 0:22:53.600 --> 0:22:56.560 would take place in two thousand five. It would double 0:22:56.760 --> 0:23:00.359 the prize money to two million dollars and the ameters 0:23:00.359 --> 0:23:03.320 of the challenge would change slightly. Now, according to the 0:23:03.359 --> 0:23:06.920 final report for the two thousand four challenge, the agency 0:23:07.040 --> 0:23:10.800 quote believes it prudent to continue with the prize authority 0:23:10.840 --> 0:23:14.960 approach and hold a second Grand Challenge for autonomous unmanned 0:23:15.200 --> 0:23:18.560 ground vehicles in two thousand five. The prize authority approach 0:23:18.680 --> 0:23:21.600 is meeting the goals of attracting new talent with new 0:23:21.640 --> 0:23:26.439 ideas and accelerating advancement in robotic vehicle research. Without the 0:23:26.480 --> 0:23:29.119 Grand Challenge, it is doubtful there would be much progress 0:23:29.119 --> 0:23:33.040 without substantial new investment in accelerating research on autonomous ground 0:23:33.080 --> 0:23:37.440 vehicles that could traverse difficult terrain at militarily relevant speeds 0:23:37.680 --> 0:23:42.320 end quote. So how did that turn out? Well, I'll 0:23:42.320 --> 0:23:46.320 tell you. First thing, get a drink of water, and 0:23:46.320 --> 0:23:56.880 we're gonna thanks O our sponsors here. As Anthony Tether 0:23:57.040 --> 0:23:59.760 would say later, the thing to remember was the two 0:23:59.760 --> 0:24:02.400 thous and four Grand Challenge was something new. It had 0:24:02.440 --> 0:24:05.080 never been done before, and it was an engineering challenge 0:24:05.119 --> 0:24:07.800 open to anyone, and teams that were able to meet 0:24:07.800 --> 0:24:11.920 the selection criteria were able to participate. So while there 0:24:12.119 --> 0:24:14.919 was disappointment when no car was able to make it 0:24:14.960 --> 0:24:18.040 over the big hill that Sandstorm got stuck on, it 0:24:18.080 --> 0:24:22.320 didn't diminish the excitement for a second Grand Challenge. Returning 0:24:22.359 --> 0:24:25.280 teams were energized by the need to outperform their previous 0:24:25.320 --> 0:24:29.440 attempts and new teams came forward inspired by the original challenge. 0:24:29.520 --> 0:24:32.480 So while it could have been the needle that would 0:24:32.560 --> 0:24:37.240 deflate the autonomous balloon could lead to another artificial intelligence winter. 0:24:37.440 --> 0:24:39.840 Like I said in the last episode, it actually got 0:24:39.840 --> 0:24:44.320 more people excited about those possibilities. So in August two 0:24:44.320 --> 0:24:48.560 thousand four, DARPA held another participants conference for parties that 0:24:48.600 --> 0:24:51.840 were interested in either participating or sponsoring a team. Now, 0:24:51.880 --> 0:24:55.120 remember it just held the Grand Challenge in March. Now 0:24:55.160 --> 0:24:57.680 it's August and they're ready to talk about the next one. 0:24:57.920 --> 0:25:01.000 The new challenge was scheduled for October eight, two thousand five, 0:25:01.440 --> 0:25:04.359 and the rules allowed for flexible starting time, So if 0:25:04.400 --> 0:25:07.359 the weather was bad or other conditions were such that 0:25:07.840 --> 0:25:10.160 you couldn't really start in October eight, you would start 0:25:10.200 --> 0:25:13.040 the next day or the next up to October eighth, 0:25:13.119 --> 0:25:15.600 two thousand five. If it kept missing it up to 0:25:15.600 --> 0:25:18.000 that point, the whole thing would be canceled. Now, according 0:25:18.040 --> 0:25:20.280 to the rules, they said the route would be quote 0:25:20.320 --> 0:25:23.879 no longer than one miles end quote. I guess that 0:25:23.960 --> 0:25:26.280 was a relief since no one got past seven and 0:25:26.320 --> 0:25:28.600 a half miles in the first one, and also said 0:25:28.600 --> 0:25:33.440 the route would include quote paved roads, unpaved roads, trails, 0:25:33.720 --> 0:25:39.240 and off road desert areas. Examples of obstacles include ditches, berms, washboard, 0:25:39.400 --> 0:25:45.160 sandy ground, standing water, rocks and boulders, narrow underpasses, construction equipment, 0:25:45.440 --> 0:25:50.280 concrete safety rails, power line towers, barbed wire fences, and 0:25:50.480 --> 0:25:54.480 cattle guards end quote. The rules also stated that DARPA 0:25:54.520 --> 0:25:58.680