WEBVTT - DARPA Robotics Challenge Review 0:00:04.200 --> 0:00:12.160 Get in touch with technology with tex Stuff from dot Com. 0:00:12.200 --> 0:00:14.440 Hey there, everyone, and welcome to tech Stuff. I am 0:00:14.760 --> 0:00:18.400 Jonathan Strickland, and today but I want to talk about 0:00:18.800 --> 0:00:21.840 the DARPA Robotics Challenge. You may remember I did an 0:00:21.840 --> 0:00:25.520 episode where I interviewed someone from DARPA to talk about 0:00:25.680 --> 0:00:29.280 the robots for US part of the DARPA Robotics Challenge, 0:00:29.880 --> 0:00:34.000 and since then the challenge actually occurred. It happened. I'm 0:00:34.040 --> 0:00:37.640 recording this on June eleven, two thousand fifteen, and the 0:00:37.640 --> 0:00:42.319 previous weekend is when they held the DARPA Robotics Challenge Finals. UH. 0:00:42.400 --> 0:00:45.560 This was the last step in a journey to create 0:00:45.600 --> 0:00:50.360 a robot that could respond to a simulated emergency situation, 0:00:51.360 --> 0:00:57.120 and it was a pretty interesting event. So to start off, 0:00:57.160 --> 0:01:01.080 we gotta talk about DARPA itself. Remember that's the research 0:01:01.080 --> 0:01:04.800 and development arm of the Department of Defense is technically 0:01:04.880 --> 0:01:10.280 the US Defense Advanced Research Projects Agency UH, and DARPA 0:01:10.319 --> 0:01:14.080 has been responsible for a lot of cutting edge research, 0:01:14.760 --> 0:01:18.640 largely in the mode of military use. But we have 0:01:18.760 --> 0:01:22.880 seen the benefits of that research hit us in other ways. 0:01:22.959 --> 0:01:28.160 Besides you know the military applications. So for example, our 0:01:28.240 --> 0:01:32.880 PA net, the predecessor to the Internet was the platform 0:01:33.000 --> 0:01:36.200 upon which the technologies that allow the Internet to work 0:01:36.440 --> 0:01:39.320 were developed, and that was a DARPA project that was 0:01:39.360 --> 0:01:43.200 back when they were first just called ARPA rather than DARPA. 0:01:43.440 --> 0:01:47.440 But we wouldn't have the Internet, at least not the 0:01:47.440 --> 0:01:50.000 way we have it today, if it were not for 0:01:50.040 --> 0:01:52.880 this agency. So the agency has done a lot of 0:01:52.920 --> 0:01:55.640 stuff that has benefited us in many ways. I'll talk 0:01:55.680 --> 0:01:58.480 about another one a little bit later on in the podcast. 0:01:58.600 --> 0:02:03.200 Kind of drop parallels to what happened with the robotics challenge. 0:02:04.160 --> 0:02:07.960 So here we have this agency. They're all about research 0:02:08.000 --> 0:02:11.720 and development. They don't actually have lots of labs with 0:02:11.760 --> 0:02:15.520 scientists working in them. Now, what this agency does is 0:02:15.560 --> 0:02:20.240 it sends out proposals to other groups and asks them 0:02:20.360 --> 0:02:25.120 to help work on various projects. Sometimes they hold challenges. 0:02:25.600 --> 0:02:28.520 In these challenges, it's kind of an open invitation for 0:02:28.639 --> 0:02:34.639 any group that can participate to uh to compete in 0:02:34.680 --> 0:02:37.640 some way, and the winner usually gets some sort of 0:02:37.639 --> 0:02:41.520 cash prize. And so often the cash prize will be 0:02:41.600 --> 0:02:45.200 less than the investment needed to actually participate in the 0:02:45.280 --> 0:02:48.520 competition to build out the thing that you need to 0:02:48.520 --> 0:02:52.639 make for us to to win, but it's still very 0:02:52.680 --> 0:02:57.600 important uh and prestigious event to be part of if 0:02:57.639 --> 0:03:00.520 you are capable of competing. Now we're talking king mostly 0:03:00.560 --> 0:03:05.160 about research departments engineering departments in universities, but there are 0:03:05.160 --> 0:03:10.000 also some private institutions, not just university type things, but 0:03:10.120 --> 0:03:13.840 actual research and development labs that will participate in DARPA 0:03:13.919 --> 0:03:19.440 projects as well. And it's collectively that these teams were 0:03:20.360 --> 0:03:23.040 kind of they're the ones that are pushing the technology forward, 0:03:23.080 --> 0:03:26.200 and DARK is kind of the facilitator. They're offering up 0:03:26.240 --> 0:03:32.440 the possibility for people to actually uh participate, and not 0:03:32.480 --> 0:03:35.960 only the possibility, but the goals. They define the goals 0:03:36.000 --> 0:03:38.320 that have to be met, and as it turns out, 0:03:39.120 --> 0:03:42.520 defining those goals is a very important part of these 0:03:42.600 --> 0:03:47.680 challenges because you have to tell the engineers what it 0:03:47.840 --> 0:03:51.000 is they need to be able to do before they 0:03:51.000 --> 0:03:53.400 can build the stuff that does it. If if you 0:03:53.440 --> 0:03:57.360 have a very vague description of what needs to be done, 0:03:58.120 --> 0:04:03.000 the possibilities of achieving it are so varied that it 0:04:03.040 --> 0:04:07.440 can often paralyze a project nothing gets done. But if 0:04:07.440 --> 0:04:11.200 you make it very specific, then it narrows down the 0:04:11.240 --> 0:04:14.840 options and it gives more focus to the project. So 0:04:15.240 --> 0:04:20.520 the Darker Robotics project itself was inspired by a real 0:04:20.680 --> 0:04:26.480 world disaster, specifically the Fukushima disaster in Japan. Now, this 0:04:26.680 --> 0:04:32.560 was the nuclear facility, the nuclear reactor that was flooded 0:04:33.040 --> 0:04:39.320 and then suffered some massive problems because of damage to 0:04:39.360 --> 0:04:45.000 the infrastructure. And we're talking about a nuclear facility, radiation 0:04:45.160 --> 0:04:49.320 is a factor, so it's a very dangerous environment to 0:04:49.400 --> 0:04:56.000 send responders to. Darpest approach was to create a scenario 0:04:56.680 --> 0:05:00.280 in which a robot would need to go through an 0:05:00.360 --> 0:05:04.440 environment similar to that that was at Fukushima and be 0:05:04.520 --> 0:05:09.640 able to perform the tasks necessary to help avert true 0:05:09.760 --> 0:05:14.800 devastation and catastrophe without putting people at risk. So the 0:05:14.839 --> 0:05:16.920 idea is that you can use a robot which is 0:05:16.960 --> 0:05:21.800 not going to respond to radiation the way a living 0:05:21.920 --> 0:05:26.080 organism would and be able to actually carry out the 0:05:26.120 --> 0:05:30.440 stuff you need to prevent a really huge problem from 0:05:30.480 --> 0:05:35.599 getting worse. So uh, some people have referred to the 0:05:35.680 --> 0:05:41.200 Darker Robotics Challenge as the Robolympics because it was a 0:05:41.200 --> 0:05:44.200 series of tasks that a robot has to complete and 0:05:44.240 --> 0:05:47.479 it had to all be the same. Robot teams could 0:05:47.520 --> 0:05:52.040 not build individual robots designed to complete specific tasks. If 0:05:52.080 --> 0:05:55.400 they could, it would be way easier of a challenge 0:05:55.720 --> 0:05:58.159 because most of the robots that are in use today 0:05:58.720 --> 0:06:03.360 have very narrow parameters. There's just maybe a single task, 0:06:04.000 --> 0:06:06.560 or perhaps just a couple of small, you know, a 0:06:06.600 --> 0:06:08.880 small range of tasks that the robot has to do. 0:06:09.760 --> 0:06:12.039 It doesn't have to do anything outside of that. So 0:06:12.080 --> 0:06:14.960 when you're designing the robot, you just say, all right, 0:06:15.120 --> 0:06:18.360 what design elements will allow the robot to do what 0:06:18.440 --> 0:06:21.440 it has to do. Anything that's outside of that is unnecessary, 0:06:21.480 --> 0:06:24.520 we will not do it. Take the roomba as an example. 0:06:24.560 --> 0:06:27.479 It's a great example. It's a robot the vacuums that 0:06:27.560 --> 0:06:31.599 cleans floors. Anything that's not necessary for the rumbo to 0:06:31.600 --> 0:06:35.680 do that is put aside. You'd want to make sure 0:06:35.720 --> 0:06:40.200 that all the design elements in your robot complement its purpose. 0:06:41.360 --> 0:06:45.320 But if you have to complete a lot of different tasks, 0:06:45.400 --> 0:06:49.120 and those tasks are not always similar to one another, 0:06:50.040 --> 0:06:57.120 you complicate the whole design process by incredible factors. It's 0:06:57.160 --> 0:07:02.160 hard to even express how much more difficult this is. Now, 0:07:02.240 --> 0:07:05.440 if you've watched any of the videos that came out 0:07:05.480 --> 0:07:08.960 of the Darker Robotics Challenge, you probably saw at least 0:07:08.960 --> 0:07:12.880 one or two that were montages of robots falling over. 0:07:13.640 --> 0:07:15.640 And it's a little funny to watch. You see these 0:07:15.840 --> 0:07:21.880 big robots just topple over sometimes apparently for no reason. 0:07:22.080 --> 0:07:24.480 They just they just they're standing and then they collapse. 0:07:25.240 --> 0:07:27.720 And there's something comic about that, you know that you 0:07:27.800 --> 0:07:32.200 can't deny that it's funny, but it also demonstrates that 0:07:32.280 --> 0:07:34.880 these robots are trying to do something that while we 0:07:35.000 --> 0:07:38.640 as human beings might find easy, it's a real challenge 0:07:38.920 --> 0:07:44.960 from a robotics standpoint. So this Darker Robotics Challenge had 0:07:45.040 --> 0:07:49.880 three phases. The first one was a virtual challenge, which 0:07:50.560 --> 0:07:54.440 in which teams were to design software that would allow 0:07:54.560 --> 0:07:58.640 a virtual robot to complete um certain tasks within a 0:07:58.720 --> 0:08:03.000 virtual environment. So there were no real robotics here, but 0:08:03.080 --> 0:08:05.840 it was a test to see if if the teams 0:08:05.840 --> 0:08:09.320 could actually build the software necessary to make the robot 0:08:09.400 --> 0:08:13.560 do the tasks that needed to be done. The next 0:08:14.200 --> 0:08:17.400 had a trial where it was a physical trial where 0:08:17.440 --> 0:08:22.120 where teams actually had to uh place a real robot 0:08:22.240 --> 0:08:25.600 through too, you know, go to the next level. And 0:08:25.640 --> 0:08:27.480 then you had the finals, which were the ones that 0:08:27.880 --> 0:08:31.440 happened most recently. UH and I'll go through each of 0:08:31.480 --> 0:08:33.800 the tasks that they had to do in just a second. 0:08:34.720 --> 0:08:37.000 If we look back at the virtual challenge, there were 0:08:37.040 --> 0:08:40.199 twenty six teams that were part of it to build 0:08:40.200 --> 0:08:44.079 that software. The robot they were controlling was a virtual 0:08:44.240 --> 0:08:50.000 Atlas robot, which is a humanoid robot and was a 0:08:50.080 --> 0:08:52.400 popular choice for a lot of the teams. When they 0:08:52.440 --> 0:08:56.720 moved on to the more physical trials, they still would 0:08:56.840 --> 0:08:58.880 use the Atlas robots. The the challenge, by the way, 0:08:58.880 --> 0:09:01.600 did not require of the teams to build their own 0:09:01.679 --> 0:09:05.600 robots from scratch. They could use pre existing robots, but 0:09:05.600 --> 0:09:08.160 they had to program them so that they could complete 0:09:08.160 --> 0:09:11.600 the necessary tasks. So not every robot out there would 0:09:11.600 --> 0:09:13.600 have been a good fit for this. Let's go back 0:09:13.640 --> 0:09:16.000 to the roomba. The roombo would have been a terrible 0:09:16.080 --> 0:09:19.280 choice to enter into the Darbo Robotics Challenge. Would have 0:09:19.320 --> 0:09:23.880 been funny, but it wouldn't have worked. So uh. Here 0:09:23.920 --> 0:09:27.960 are some of the virtual tasks that the virtual robots 0:09:28.160 --> 0:09:31.880 had to complete using this software. They had to walk 0:09:31.920 --> 0:09:37.280 through an area of uneven ground that had debris included there, 0:09:37.280 --> 0:09:40.480 so simulated debris had to have a software that would 0:09:40.480 --> 0:09:43.880 allow the robot to maintain its balance even while tracking 0:09:43.880 --> 0:09:48.160 a path to a specific destination. Through this this uneven terrain, 0:09:48.840 --> 0:09:51.480 they had to attach a hose to a spigot and 0:09:51.640 --> 0:09:56.560 turn a valve. Uh. They also were given some artificial 0:09:56.640 --> 0:10:01.400 limits because and an emergency WHO situation, you cannot always 0:10:01.480 --> 0:10:07.560 count on your communication lines being perfectly sound, such as 0:10:07.559 --> 0:10:11.840 that the Fukushima disaster. You need to be sure that 0:10:11.920 --> 0:10:14.360 your robot can contend with the fact that there might 0:10:14.400 --> 0:10:17.160 be lag between when you can issue a command and 0:10:17.160 --> 0:10:20.720 when the robot is ready for the next command. UH. 0:10:21.000 --> 0:10:23.280 It meant that some of the teams started to really 0:10:23.320 --> 0:10:28.120 look at ways to create a semi autonomous robot. So 0:10:28.200 --> 0:10:31.440 some of the teams built robots that were capable of 0:10:32.480 --> 0:10:39.040 taking on some tasks autonomously. Uh. Some teams avoided that entirely. 0:10:39.280 --> 0:10:42.560 They focused on using the robots as a direct extension 0:10:42.840 --> 0:10:47.280 of the controls that the humans were behind. So, in 0:10:47.280 --> 0:10:51.120 other words, you might think of those robots as enormous 0:10:51.360 --> 0:10:56.920 technological puppets. The puppets would respond to direct human commands 0:10:57.640 --> 0:11:00.160 and wouldn't be able to do anything on their own. 0:11:00.200 --> 0:11:03.559 They didn't have any autonomy. There were other ones that, 0:11:03.679 --> 0:11:06.440 you know, you you could have semi autonomy. Only show 0:11:06.480 --> 0:11:08.240 it a set of stairs and you send it the 0:11:08.280 --> 0:11:12.240 command to climb those stairs, and it could do the rest. 0:11:12.280 --> 0:11:14.360 It could calculate how high it needed to put its 0:11:14.400 --> 0:11:18.240 feet and how to shift its weight, that sort of stuff. UH, 0:11:18.240 --> 0:11:20.600 and there were a lot of different strategies employed and 0:11:20.880 --> 0:11:24.599 to varying degrees of success. It wasn't like the autonomous 0:11:24.720 --> 0:11:28.720 robots automatically were better than the human controlled or vice versa. 0:11:28.760 --> 0:11:32.600 There was actually it all depended on the implementation. So 0:11:32.800 --> 0:11:40.440 that virtual trial ended up having one team winning pretty 0:11:40.480 --> 0:11:45.360 pretty decisively. UM. It was the Institute for Human and 0:11:45.440 --> 0:11:51.120 Machine Cognition, which is UH, an organization in Pensacola, Florida, 0:11:51.360 --> 0:11:54.800 and eight other teams qualified for the trials that we're 0:11:54.840 --> 0:11:59.920 going to be held in December. Now, by the time 0:12:00.120 --> 0:12:04.520 those trials happened, some of the teams had merged UH 0:12:04.520 --> 0:12:08.800 and some other teams were coming into it. Six of 0:12:08.840 --> 0:12:11.600 the teams would go on to compete at the trials 0:12:11.600 --> 0:12:16.040 that were held at the Homestead Miami Speedway, and they 0:12:16.080 --> 0:12:19.000 had eight tasks they had to complete that were similar 0:12:19.040 --> 0:12:23.280 to the ones in the Darker Robotics Final and those 0:12:23.320 --> 0:12:28.679 tasks included manning a vehicle, um walking through uneven terrain 0:12:28.920 --> 0:12:32.960 or otherwise moving through uneven terrain. Robots did not have 0:12:33.080 --> 0:12:36.640 to be bipedal. They could move around on however many 0:12:36.679 --> 0:12:38.760 limbs they had. They just had to be also able 0:12:38.800 --> 0:12:41.760 to do these other tasks. Uh. They had to be 0:12:41.840 --> 0:12:43.800 able to climb a ladder. They had to be able 0:12:43.840 --> 0:12:46.400 to remove debris from a door and then open the door. 0:12:46.640 --> 0:12:49.520 They had to be able to break through a wall. Uh, 0:12:49.559 --> 0:12:51.719 they had to be able to handle valves and to 0:12:51.880 --> 0:12:55.960 attach a hose to a spigot. The winning team of 0:12:56.120 --> 0:13:00.679 that set of trials was a group called Shaft spelled 0:13:00.960 --> 0:13:05.199 s C H A f T. This was a group 0:13:05.200 --> 0:13:09.160 out of Japan. Now, when this group was competing, they 0:13:09.200 --> 0:13:12.880 were this little independent group. But then they got acquired 0:13:12.960 --> 0:13:16.199 by a little company that we've talked about several times 0:13:16.200 --> 0:13:19.520 on this podcast. You know him, you may or may 0:13:19.520 --> 0:13:23.240 not love him Google. So just before the trials, Google 0:13:23.240 --> 0:13:28.679 had acquired uh, this this this group, this organization. Now 0:13:28.760 --> 0:13:32.320 Shaft did really well, uh, which you know makes sense. 0:13:32.320 --> 0:13:35.040 I mean, you've you've heard the theme song. But out 0:13:35.120 --> 0:13:38.240 of thirty two points that it could possibly score, it 0:13:38.520 --> 0:13:43.320 earned twenty seven, So pretty good score for for a 0:13:43.440 --> 0:13:46.920 robot and also that was seven points more than the 0:13:46.960 --> 0:13:51.040 next closest competitor, which was from I H M C, 0:13:51.520 --> 0:13:55.800 the same group that won the virtual trial earlier. Third 0:13:55.880 --> 0:14:01.520 place was Tartan Rescue from Carnegie Melon And that will 0:14:01.559 --> 0:14:03.200 be important in a second to when we get into 0:14:03.200 --> 0:14:08.120 the finals. Now, Shaft did not go on to compete 0:14:08.120 --> 0:14:11.719 in the DARPA Robotics Finals because Google announced that they 0:14:11.720 --> 0:14:17.679 were refocusing the team, dedicating it to actual commercial Google projects, 0:14:18.240 --> 0:14:23.320 So uh, they were no longer being dedicated to this challenge. 0:14:23.360 --> 0:14:26.200 They were being dedicated to a real world product of 0:14:26.240 --> 0:14:34.440 some sort. Google's also had other involvement with some DARPA stuff, 0:14:34.480 --> 0:14:37.680 at least after the fact. Google does not tend to 0:14:37.720 --> 0:14:42.360 get involved in these challenges directly, um probably because of 0:14:42.400 --> 0:14:47.440 the military association of DARPA and Google is very careful 0:14:47.480 --> 0:14:51.680 to avoid those kind of associations. Now, the finals took 0:14:51.720 --> 0:14:56.440 place over June five and June six and Pomona, California, 0:14:56.680 --> 0:15:00.960 and uh, it was at the Fair Plex, and there 0:15:01.040 --> 0:15:04.840 was quite a crowd watching these events, and the reports 0:15:04.840 --> 0:15:07.120 I read were really interesting, and also the videos I 0:15:07.160 --> 0:15:11.280 watched were interesting because the crowd was extremely enthusiastic cheering 0:15:11.360 --> 0:15:15.840 on the robots, groaning every time a robot failed a 0:15:15.960 --> 0:15:20.240 task or fell over, gasping when that sort of stuff happened, 0:15:20.640 --> 0:15:24.280 because everyone wanted to see these robots succeed. No one 0:15:24.360 --> 0:15:28.520 wanted any group to fail. And uh. Also, the other 0:15:28.600 --> 0:15:31.840 interesting thing was that it was really hard to tell 0:15:32.040 --> 0:15:34.920 if you were in the audience when a robot was 0:15:35.040 --> 0:15:39.960 operating autonomously as opposed to being controlled by humans. Um. 0:15:40.040 --> 0:15:44.320 That says two things, One that the line between these 0:15:44.360 --> 0:15:48.040 two is getting further and further blurred, and too that 0:15:48.280 --> 0:15:51.600 sometimes the robot would behave in a weird way, and 0:15:51.640 --> 0:15:54.520 you couldn't be sure if it was because the autonomous 0:15:54.560 --> 0:16:00.360 programming was lacking or because the control mechanisms were limited. So, 0:16:00.400 --> 0:16:03.840 in other words, it's not that necessarily that the art 0:16:03.880 --> 0:16:06.280 has gotten so advanced that we can't tell the difference. 0:16:06.320 --> 0:16:08.720 It may be that the art has not advanced enough 0:16:09.200 --> 0:16:13.600 that we can't spot what is the cause of incompetence. 0:16:15.040 --> 0:16:18.640 Uh And and I use the word incompetence mainly as 0:16:18.680 --> 0:16:22.960 a joke. It's just a joke because truthfully it displays 0:16:23.000 --> 0:16:27.280 how difficult robotics as a discipline actually is, how how 0:16:27.400 --> 0:16:30.960 challenging it is to design a robot is capable of 0:16:31.000 --> 0:16:33.440 doing the same sort of things that humans can do. 0:16:34.360 --> 0:16:36.800 It shows that the things that we take for granted 0:16:37.440 --> 0:16:42.520 as being pretty simple are fiendish lee difficult when you 0:16:42.560 --> 0:16:45.800 get to a design and programming level when you're actually 0:16:45.800 --> 0:16:47.920 building a robot that's going to be capable of doing 0:16:47.960 --> 0:16:51.800 the same thing. So here are the eight tasks that 0:16:51.840 --> 0:16:55.400 the robots had to complete, and one of these tasks 0:16:55.760 --> 0:16:58.680 was called Surprise. I'll get to it, but the reason 0:16:58.720 --> 0:17:01.960 it's called surprise is that DARPA told all the teams, 0:17:02.680 --> 0:17:07.200 your robot will encounter a task similar to these other 0:17:07.280 --> 0:17:10.520 ones that you already know about, but we're not going 0:17:10.560 --> 0:17:13.320 to tell you what it is. And it meant that 0:17:13.359 --> 0:17:16.440 the teams knew that there was going to be something 0:17:17.320 --> 0:17:21.680 on that list of tasks that was not not defined, 0:17:22.359 --> 0:17:24.040 and that the robot would have to be able to 0:17:24.080 --> 0:17:26.600 contend with it in order to get a point for 0:17:26.720 --> 0:17:30.919 that particular task. And so the tasks were um there 0:17:30.920 --> 0:17:33.800 were eight, Like I said, eight of them. You were 0:17:33.880 --> 0:17:37.240 awarded a point if your robot could successfully complete that task. 0:17:38.640 --> 0:17:42.680 So the score, the final scores once everything was done, 0:17:43.000 --> 0:17:46.439 were determined by how many tasks were completed. Successfully and 0:17:46.560 --> 0:17:51.280 what was the time of the robots performance. Now, when 0:17:51.520 --> 0:17:55.320 they were first designing the robots, teams were told they 0:17:55.320 --> 0:17:58.480 would have half an hour per task, So you have 0:17:58.560 --> 0:18:01.280 eight tasks half an hour per task, that's four hours total. 0:18:01.800 --> 0:18:04.200 But the actual competition, they were told they would have 0:18:04.240 --> 0:18:09.240 one hour to complete all eight tasks. Um. And while 0:18:09.320 --> 0:18:11.560 you might think that how is that fair as a 0:18:11.560 --> 0:18:14.639 as a bait and switch, you also have to understand 0:18:14.760 --> 0:18:20.000 this was all about simulating an emergency response. So under 0:18:20.080 --> 0:18:23.840 those emergency conditions, you can't expect to ask for more time. 0:18:24.080 --> 0:18:27.679 That's not realistic. Uh. And it added an extra layer 0:18:27.720 --> 0:18:32.280 of pressure on the teams. So those eight tasks, the 0:18:32.320 --> 0:18:35.240 first one was to drive, so the robots had to 0:18:35.359 --> 0:18:39.919 drive a Polaris Ranger XP nine vehicle. These if you 0:18:39.920 --> 0:18:42.000 haven't seen them, they look like sometimes you might call 0:18:42.119 --> 0:18:44.520 mcgator or you might think of it as a golf 0:18:44.600 --> 0:18:49.359 cart on steroids. These are um vehicles that are similar 0:18:49.400 --> 0:18:53.040 to golf carts. They're they're largely open, they've got heavier 0:18:53.040 --> 0:18:56.200 wheels than golf carts, do more warm horsepower than a 0:18:56.280 --> 0:19:00.520 golf cart would. But it's definitely in that rain between 0:19:00.520 --> 0:19:04.439 golf cart and real car teams had five minutes to 0:19:04.520 --> 0:19:09.600 alter the vehicle without using tools to make sure that 0:19:09.640 --> 0:19:12.760 their robots could actually operate it. So what the robots 0:19:12.800 --> 0:19:15.760 had to do was drive from the starting point to 0:19:16.480 --> 0:19:21.080 a destination, and it was only considered a success if 0:19:21.080 --> 0:19:24.720 the vehicle had completely moved past a finishing mark. The 0:19:24.800 --> 0:19:29.119 vehicle also had to go through essentially a driving course 0:19:29.400 --> 0:19:33.200 with obstacles and cones set up, and it was determined 0:19:33.200 --> 0:19:36.560 that if the robot were to collide with one of 0:19:36.560 --> 0:19:38.600 those obstacles or one of those cones and cause it 0:19:38.640 --> 0:19:42.119 to move as a result of that collision, the robot 0:19:42.160 --> 0:19:47.639 would not receive a point for that task uh the robot. Also, 0:19:47.960 --> 0:19:51.200 the team could choose for the robot too, instead of driving, 0:19:51.280 --> 0:19:54.359 to walk to the destination, but in that case they 0:19:54.400 --> 0:19:56.879 would not be awarded a point for that task. They 0:19:56.880 --> 0:20:01.920 would forfeit the point um. But ideally the robot would 0:20:01.960 --> 0:20:04.359 be able to operate the accelerator and the steering wheel 0:20:05.200 --> 0:20:08.800 and maybe even the break and the shift. The cars 0:20:08.840 --> 0:20:12.240 were already or the vehicles were already started, the engines 0:20:12.240 --> 0:20:16.760 were already running, and they were already in high gear. 0:20:17.160 --> 0:20:20.000 Because it was considered to be the smoothest way for 0:20:20.040 --> 0:20:24.479 the robots to to maneuver the vehicle. But if teams 0:20:24.480 --> 0:20:26.919 wanted to, they could design a robot that would be 0:20:26.920 --> 0:20:30.520 capable of shifting gears. It was not a requirement, it 0:20:30.600 --> 0:20:32.400 was just something they could do if they wanted to. 0:20:33.560 --> 0:20:36.040 At any rate, the robot needed to be able to 0:20:36.119 --> 0:20:40.280 drive safely from the starting point to the destination and 0:20:40.320 --> 0:20:44.119 then halt the vehicle, either by letting off of the 0:20:44.160 --> 0:20:46.760 accelerator so it coasted to a stop or actually applying 0:20:46.760 --> 0:20:50.720 the brake. The next task. That was just task number one. 0:20:51.280 --> 0:20:54.080 The next task was called egress, which is really just 0:20:54.280 --> 0:20:57.800 getting out of the car. It sounds incredibly simple, and 0:20:57.840 --> 0:21:00.359 for humans, for most of us, it really is pretty simple, 0:21:00.960 --> 0:21:04.680 you know it. We we intuitively know how to maintain 0:21:04.720 --> 0:21:07.439 our balance and to shift our weight so that we 0:21:07.560 --> 0:21:10.800 go from a seated position to standing. But that's not 0:21:10.880 --> 0:21:13.119 the case with the robots. You have to design the 0:21:13.200 --> 0:21:15.399 robot to do that. You have to program the robot 0:21:15.440 --> 0:21:17.320 to do that. You have to the robot be able 0:21:17.359 --> 0:21:21.760 to anticipate what a shift and its weight will do, 0:21:21.880 --> 0:21:25.359 what how it's momentum will carry it forward. As it 0:21:25.359 --> 0:21:28.240 turns out, this was one of the trickier tasks that 0:21:28.840 --> 0:21:31.600 teams had to complete to get out of a vehicle. 0:21:31.720 --> 0:21:36.600 It was not an easy thing to do um but 0:21:36.800 --> 0:21:38.200 they were told that they could get out of the 0:21:38.280 --> 0:21:40.800 vehicle in either direction. They didn't have to exit out 0:21:40.800 --> 0:21:42.879 all the left side versus the right side. That the 0:21:42.960 --> 0:21:46.840 robot could get out of either side, and again the 0:21:46.880 --> 0:21:50.200 task was considered complete if the robot could get out 0:21:50.200 --> 0:21:52.720 of the vehicle and maintain its balance and move on 0:21:52.800 --> 0:21:55.920 to the next challenge, and if it could, it would 0:21:55.920 --> 0:22:00.760 receive a point. The next one was door. Now, egress 0:22:00.840 --> 0:22:04.800 was tough. Door was surprisingly tough. You would think that 0:22:04.920 --> 0:22:07.919 opening a door would be a pretty simple task. This 0:22:08.000 --> 0:22:10.960 was a push door operated by a lever style handle. 0:22:11.040 --> 0:22:13.560 You had to push up or pull you know, pull 0:22:13.640 --> 0:22:16.640 up or push down rather on the handle and then 0:22:16.680 --> 0:22:20.320 push the door to open it, then step through the door. 0:22:20.720 --> 0:22:22.600 The door did not have a threshold, so there was 0:22:22.640 --> 0:22:25.600 nothing that you had to step over. However, the door 0:22:25.800 --> 0:22:29.480 was a standard thirty six inch wide doorway and with 0:22:29.600 --> 0:22:31.840 the door jam that's closer to thirty three and a 0:22:31.840 --> 0:22:34.960 half inches. And some of these robots were too wide 0:22:35.240 --> 0:22:38.520 to walk through the door, you know, heading with their 0:22:38.560 --> 0:22:41.679 torso facing forward. They actually had to turn sideways and 0:22:41.720 --> 0:22:45.280 then shuffle through the doorway, and there were a lot 0:22:45.320 --> 0:22:48.600 of robots that fell down at this stage of the 0:22:49.080 --> 0:22:54.800 the challenge. They would lose their balance either when pushing 0:22:54.840 --> 0:22:57.640 the door or when trying to maneuver through the doorway. 0:22:58.800 --> 0:23:02.399 One robot fell down at this stage and was able 0:23:02.440 --> 0:23:06.480 to pick itself back up very very slowly and deliberately, 0:23:07.640 --> 0:23:12.720 which is remarkable because it's the only robot that managed 0:23:12.840 --> 0:23:15.439 to pick itself back up after falling over, and that 0:23:15.520 --> 0:23:19.080 was the Carnegie Melon robot, the same one we mentioned 0:23:19.080 --> 0:23:22.080 in the Trials part in just a few moments ago. 0:23:23.600 --> 0:23:25.200 But anyway, that was it. You just had to open 0:23:25.200 --> 0:23:27.000 the door and walk through the doorway, and that was 0:23:27.040 --> 0:23:30.200 the end of that task, and yet it was deceptively difficult. 0:23:30.680 --> 0:23:35.560 The next one was Valve, which was not about the 0:23:35.760 --> 0:23:39.359 game company, but rather about turning a control as if 0:23:39.400 --> 0:23:42.720 you were turning a valve to operate a fire hose 0:23:42.960 --> 0:23:46.159 or to shut down water to a specific part of 0:23:46.160 --> 0:23:49.040 the system, which makes sense in a nuclear facility you 0:23:49.119 --> 0:23:52.520 might have to shut off or open a valve. In 0:23:52.560 --> 0:23:57.359 that case, the valve had to be turned counterclockwise in 0:23:57.480 --> 0:24:00.600 three sixty degrees in order for the task to complete, 0:24:01.080 --> 0:24:03.679 and the team was only told that the valve would 0:24:03.920 --> 0:24:07.240 have a size between four and sixteen inches in diameter, 0:24:08.000 --> 0:24:10.520 which meant that you had to create a robot that 0:24:10.560 --> 0:24:13.959 would be capable of gripping anything within that range and 0:24:13.960 --> 0:24:19.080 then turning it in that counterclockwise direction three degrees again 0:24:19.160 --> 0:24:22.840 trigger than it sounds the next step. The next task 0:24:23.000 --> 0:24:25.160 was called wall, in which a robot had to pick 0:24:25.240 --> 0:24:28.080 up a cordless drill. There were two of them available, 0:24:28.520 --> 0:24:31.520 They were not automatically on. The robot had to operate 0:24:31.560 --> 0:24:35.040 the cordless drill and turn it on by squeezing the trigger, 0:24:35.920 --> 0:24:39.000 and use the cordless drill to cut through some drywall 0:24:39.480 --> 0:24:43.720 in a shape that was drawn on the drywall. The 0:24:43.800 --> 0:24:47.080 drywall was a half inch thick, and the robot had 0:24:47.119 --> 0:24:49.840 to hold the drill, operate the drill, and move in 0:24:49.840 --> 0:24:53.000 this shape and then remove the rubble. The idea being 0:24:53.040 --> 0:24:56.320 that sometimes the robot might have to cut through a 0:24:56.480 --> 0:25:00.239 surface in order to get access to certain controls uh 0:25:00.440 --> 0:25:04.119 that might otherwise be blocked by a collapse of a 0:25:04.200 --> 0:25:08.399 room or something along those lines. Also pretty tricky, you know, 0:25:08.560 --> 0:25:11.680 using tools that were designed to be held in human hands. 0:25:12.280 --> 0:25:14.720 You know, humans, we we have the ability to detect 0:25:14.960 --> 0:25:19.360 how much pressure we're using and to um to change 0:25:19.440 --> 0:25:22.919 that based upon what's happening. Robots a little more tricky. 0:25:22.960 --> 0:25:26.399 I mean, you can have sensors that alert the the 0:25:26.400 --> 0:25:30.800 operator what's happening to what's happening, but it's not intuitive. Again, 0:25:30.840 --> 0:25:35.760 you have to program it. After wall was the surprise. Now, 0:25:35.840 --> 0:25:39.080 the surprise was something that DARPA could change out from 0:25:39.160 --> 0:25:41.600 day to day. There were two days of this series 0:25:41.640 --> 0:25:45.200 of challenges, and I know that on one day, I'm 0:25:45.320 --> 0:25:47.640