[00:00.000 --> 00:09.380] ...head of Johannes from Monochrom, because he has this awesome thing called RoboExotica, which is the International Festival of Cocktail Robotics, okay? [00:09.760 --> 00:16.000] So, Chassie has kind of an unusual capability, right? [00:16.120 --> 00:22.020] That is Chassie, he rolls around, he's got stuff, and he's kind of fully equipped, right? [00:23.140 --> 00:27.840] Because he is, in fact, a beverage delivery device, right? [00:27.840 --> 00:32.000] Okay, so he is essentially a mobile kegerator, right? [00:32.240 --> 00:35.420] So, finally, a robot that does something useful! [00:37.060 --> 00:42.600] Okay, so we took Chassie to all sorts of parties, we shipped him around or something. [00:43.320 --> 00:48.260] We can serve all kinds of liquids from inside his case. [00:48.560 --> 00:58.340] We have a corny keg, we pressurize it with CO2 for beer, we can use nitrogen, we can serve wine, just pressurize it with air if you want to serve lemonade or just about anything. [00:59.340 --> 01:02.240] And Chassie is completely controlled by remote control. [01:02.540 --> 01:05.660] There's actually very little intelligence in Chassie, which is alright. [01:06.140 --> 01:08.820] So, there he is, he is willing to serve you. [01:10.020 --> 01:21.840] We also put an intercom in Chassie's head, so we can have like a little man behind the curtain, who can kind of hide behind a column, and this person serves as Chassie's personality. [01:22.020 --> 01:35.360] And it's funny, when you give the person the microphone to control Chassie, and I think Johannes can testify to this, that their personality changes a little bit, because Chassie is actually kind of a little obnoxious, and he'll kind of run into you, and maybe over your toes and stuff like that. [01:36.740 --> 01:38.360] So, how did we build Chassie? [01:39.860 --> 01:44.560] My pal Al did a lot of the mechanical stuff, and this is something I'll get back to later. [01:44.680 --> 01:46.560] If you're doing robots, it's like sex. [01:46.680 --> 01:48.800] You can do it yourself, but you want to do it with other people. [01:48.940 --> 01:49.620] It's a lot more fun. [01:50.800 --> 01:52.960] Okay, so, this is Chassie. [01:53.060 --> 02:00.100] He was basically something from an old service station, and he had this wonderful kind of Art Deco look to him. [02:00.740 --> 02:19.020] He just used to have casters down there, and the way they would use this in old service station is they put basically a barrel of lubrication oil in it, and there was a little fitting on his head for an air fixture, and you would basically wheel it up to your car that you're servicing and pump in new oil out of Chassie into the car. [02:20.340 --> 02:22.440] So, we kind of improved on that. [02:22.800 --> 02:23.760] We took him apart. [02:23.900 --> 02:25.140] We took out all the oil stuff. [02:25.300 --> 02:26.320] We powder-coated him. [02:26.380 --> 02:30.820] We stripped him down to bare metal and powder-coated him, so that he was nice and shiny and red. [02:31.580 --> 02:39.760] This is all his locomotion stuff is in his base, which leaves the whole inside of him free for the good stuff, the beer. [02:40.360 --> 02:42.040] So, these are two motors. [02:42.220 --> 02:44.900] These are from electric scooters. [02:44.980 --> 02:46.800] You can get them from China, super cheap. [02:47.000 --> 02:47.400] They're great. [02:48.260 --> 02:50.720] These are some batteries, pretty standard batteries. [02:50.880 --> 02:51.840] You get them out of... [02:51.840 --> 02:57.020] We got some surplus ones out of emergency lighting systems by code. [02:57.120 --> 03:03.200] They have to rotate them so often, so you get ones that are basically unused for not very much. [03:03.680 --> 03:05.260] Here's the circuitry stuff. [03:05.460 --> 03:06.600] This is a motor controller. [03:06.760 --> 03:12.180] You might notice that some of the transistors are there, are kind of in an unfortunate shape. [03:12.400 --> 03:14.340] You know, most transistors are supposed to have three legs. [03:14.440 --> 03:16.280] This one kind of got blown apart. [03:18.600 --> 03:24.780] And what happened to that poor speed controller was, some high school students got their hands on it. [03:24.920 --> 03:26.740] And I'll talk about that towards the end. [03:26.740 --> 03:30.740] Well, no, this is awesome, because this is how you learn how to do things. [03:30.900 --> 03:40.240] And so, I scavenged these motor controllers, just threw in some new MOSFET transistors, power transistors, and had nice new speed controllers for not much money. [03:40.280 --> 03:41.980] So, most of chassis is recycled. [03:42.100 --> 03:43.400] We're a little bit proud of that. [03:44.260 --> 03:51.600] Okay, so we took him to Vienna for the International Festival of Cocktail Robotics, which is an awesome, awesome party. [03:51.600 --> 03:55.280] If you're ever in Vienna in December, do not miss it. [03:56.680 --> 04:01.680] Johannes and monochrom and Schiffz, if I'm pronouncing that right, put this on. [04:01.760 --> 04:10.660] I think they get some funding from the Viennese city government, because in civilized countries, the government actually supports the arts and things like that. [04:10.840 --> 04:18.040] So, it was a wonderful time, and everybody enjoys chassis, because what's not to like? [04:18.740 --> 04:30.120] It was kind of amazing to me, and it still amazes me, we take chassis to parties and things, that people will treat this piece of junk like it's a person, or it has some kind of intelligence, and it's not, right? [04:30.200 --> 04:41.740] It's just this, there's somebody driving it around with a remote control, and maybe somebody talking through a little headset microphone, giving it the impersonation of a personality, but there's nothing there. [04:43.660 --> 04:49.620] Okay, which kind of leads me to my next set of robots, and these are not humanoid. [04:49.740 --> 04:52.340] They look nothing like chassis. [04:52.540 --> 04:54.740] They are spheres, and they roll by themselves. [04:55.000 --> 04:56.800] This is an art project called Swarm. [04:56.940 --> 04:59.020] I did, it's kinetic art robots. [05:00.100 --> 05:01.860] I did this not by myself. [05:01.860 --> 05:05.700] I did this with a whole bunch of other people, and I can't give them all credit. [05:05.700 --> 05:09.720] There are probably maybe 40 or 50 people who worked on this project with me. [05:11.100 --> 05:13.680] So, this was a Connecticut art project. [05:13.900 --> 05:15.360] We are interested... [05:15.360 --> 05:21.420] This gives you an idea about the scale of these things, or about the size to crush a small child, right? [05:25.980 --> 05:28.800] We were interested in developing this project. [05:28.960 --> 05:34.620] There's been a lot of research in academic labs about swarming behavior and nanobots and things like this. [05:34.720 --> 05:37.120] And this was really influential to us artistically. [05:37.280 --> 05:43.400] What can we do with robots that might be autonomous, might navigate by themselves? [05:44.660 --> 05:49.160] Maybe we can use them as a platform for choreography, doing dancing and things like that. [05:49.680 --> 05:52.540] Maybe we can get choreographers to come to work with us. [05:52.540 --> 05:57.780] Each orb has lights and sounds that you can also program or control remotely. [05:58.080 --> 06:03.020] So, we thought this was an interesting platform to work on. [06:03.220 --> 06:05.820] Now, it didn't actually work quite as smoothly as that. [06:05.820 --> 06:06.640] I'll get into that. [06:07.560 --> 06:10.620] So, we took it up to the desert, basically. [06:10.620 --> 06:12.600] We got some funding from Burning Man. [06:12.600 --> 06:17.600] If you don't know about it, at the end of the summer, it's this giant art party out in the Nevada desert. [06:18.340 --> 06:21.160] And people just go there, and they do a lot of stupid stuff. [06:21.360 --> 06:23.000] And it's just a heck of a lot of fun. [06:23.720 --> 06:25.200] Consider going if you haven't been. [06:26.240 --> 06:32.520] But to go to Burning Man, you buy a ticket, and a small fraction of the ticket cost actually goes to supporting art. [06:32.720 --> 06:40.480] And so, we got basically the seed money for this project from Burning Man to basically buy us parts, which was awesome. [06:40.480 --> 06:41.980] And we put in a lot of volunteer time. [06:42.860 --> 06:43.300] All right. [06:43.540 --> 06:46.240] So, you can see out there, it's on a dry lake bed. [06:46.400 --> 06:47.240] It's hella hot. [06:47.380 --> 06:48.080] It's hella dusty. [06:48.260 --> 06:49.500] They say hella in California. [06:49.500 --> 06:50.260] It's kind of slang. [06:50.440 --> 06:50.920] It's cool. [06:52.480 --> 06:56.720] So, one of the things we had to deal with in our design is dust. [06:56.840 --> 06:58.680] And this dust is not just any kind of dust. [06:58.820 --> 07:00.020] It's alkali dust. [07:00.180 --> 07:02.240] And that, with moisture, turns out to eat electronics. [07:02.280 --> 07:04.520] So, that's something we had to think about. [07:05.760 --> 07:06.280] All right. [07:07.300 --> 07:07.660] Exactly. [07:07.980 --> 07:09.320] So, who are we? [07:09.320 --> 07:12.240] Well, basically, we are just a bunch of geeks. [07:12.380 --> 07:13.220] And we got together. [07:13.420 --> 07:15.120] We've worked on other Burning Man projects. [07:15.120 --> 07:17.980] We know each other in San Francisco, in the Bay Area. [07:18.120 --> 07:20.440] There's this wonderful scene of kind of hackers and geeks. [07:20.600 --> 07:22.740] And we did this before. [07:22.740 --> 07:24.400] There was Noise Bridge, the hackerspace. [07:24.560 --> 07:27.640] But we've been having some meetings at Noise Bridge. [07:28.740 --> 07:31.280] And just a wonderful place to do projects like this. [07:31.420 --> 07:36.900] And I'm sure there's equal centers of energy I'm seeing with the hackerspaces around here, things like that. [07:36.900 --> 07:40.480] So, definitely, if this is your kind of thing, zero in on that. [07:40.600 --> 07:41.540] I'll talk about that later. [07:41.860 --> 07:42.300] Okay. [07:42.380 --> 07:43.340] We're kind of open source. [07:44.500 --> 07:45.960] There's no particular boss. [07:46.160 --> 07:47.580] If you want something done, you do it. [07:48.560 --> 07:50.560] Decision-making is kind of distributed. [07:50.940 --> 07:53.040] Everybody helps out doing stuff they want to do. [07:53.140 --> 07:56.500] Here's Lisa helping to weld a shell of the orb. [07:56.760 --> 08:00.420] And none of us were expert welders when we started this. [08:00.420 --> 08:02.420] By the end, we were because we got a lot of practice. [08:02.920 --> 08:06.900] And it turns out this is welding aluminum, which turns out to be hard and takes a lot of practice. [08:08.060 --> 08:08.580] All right. [08:09.760 --> 08:12.620] Like I said, if something wants to be done, you can do it. [08:12.720 --> 08:14.420] This is Eric, and he's working on... [08:15.300 --> 08:18.720] This is a pedal-powered carrier for the robots. [08:20.120 --> 08:21.800] It's like a conference bike. [08:22.000 --> 08:23.140] Four people can ride it. [08:23.220 --> 08:27.060] We built that just to schlep them around to various performance spots. [08:27.060 --> 08:27.920] That was fun. [08:29.100 --> 08:33.980] And the question is, where do we get our money from to build these things? [08:34.220 --> 08:35.620] Because we're not an organization. [08:35.620 --> 08:36.540] We're not a company. [08:38.360 --> 08:42.160] We're not incorporated in any way. [08:42.260 --> 08:43.140] We're just a bunch of geeks. [08:43.820 --> 08:48.460] So we get some money from the art festivals, like I mentioned. [08:49.160 --> 08:52.160] We took them to a music festival in Southern California. [08:52.980 --> 08:54.540] We do tech festivals. [08:54.540 --> 08:57.360] We took them to a tech festival in India. [08:57.540 --> 08:58.380] I'll talk about that later. [08:59.060 --> 08:59.940] Corporate functions. [09:01.160 --> 09:06.100] Local museums have these evenings where they serve drinks, and we roll the spheres around. [09:06.280 --> 09:08.260] That's a good times. [09:08.480 --> 09:09.920] And then we pay for stuff ourselves. [09:10.240 --> 09:10.980] Not a lot. [09:12.380 --> 09:12.820] All right. [09:12.920 --> 09:14.520] So how the heck do these things work? [09:14.620 --> 09:15.960] They roll, but they don't have wheels. [09:16.060 --> 09:16.860] They roll by themselves. [09:17.020 --> 09:22.920] And the secret is basically, they just have a big old heavy weight in the bottom, and they have some motors to move them around. [09:22.920 --> 09:32.860] And what you basically do is you take this weight, and you move it off-center, and the orb will roll in the direction because you moved the center of gravity off the center. [09:33.020 --> 09:34.980] It will move in that direction. [09:35.560 --> 09:43.260] And so when it's rolling in this direction, you can tilt the weight back and forth, and if you tilt it, the whole thing will tilt on its axis, and you can steer it that way. [09:43.280 --> 09:48.060] So you have two degrees of freedom, forward, back, and kind of left-right steering. [09:49.040 --> 09:52.360] And that's a little mock-up we did with a hamster ball. [09:52.520 --> 09:54.520] It turns out we're not the only ones who thought about this. [09:54.640 --> 10:01.120] There are about five or six other people doing spherical robots, most of which we learned about when we were well into the project, as one does. [10:01.340 --> 10:03.840] But it's kind of an interesting thing. [10:04.140 --> 10:14.500] And apparently NASA is talking about using things like this, maybe made of basically airbags to explore Mars, and they're called tumbleweeds. [10:14.600 --> 10:15.880] They would blow around on the winds. [10:16.000 --> 10:24.800] And we can tell them that, from our experience, the spheres, they don't do great with hills, and they certainly don't do great with little depressions. [10:24.820 --> 10:25.880] They tend to get stuck. [10:26.400 --> 10:32.100] Although there are new designs which have gyroscopes, which give you a little reactive force so you can pop yourself out of things. [10:33.300 --> 10:33.740] All right. [10:33.820 --> 10:37.940] So this is what they look like in the desert at night with our carrier behind them. [10:38.900 --> 10:42.780] I probably should have showed some video, but that would take all night, so it sucks to be you. [10:44.340 --> 10:49.700] Here's drive mechanics, how they work, a little bit more about what's going on in the motors. [10:50.180 --> 10:50.240] Okay. [10:50.380 --> 10:56.340] So here's the center axle, which goes through basically the center of the shell, is stationary. [10:56.340 --> 10:57.520] It doesn't move. [10:57.700 --> 10:59.420] The shell turns around the axle. [10:59.540 --> 11:07.700] And this is a drive shaft, which basically takes the motion of this motor and spins the entire shell around the axle. [11:09.680 --> 11:16.360] This is the steering motor, which basically shifts that heavy battery pack left and right to steer it with. [11:18.020 --> 11:18.600] All right. [11:18.800 --> 11:21.080] So robots will kick your ass. [11:21.520 --> 11:28.240] This is us getting our asses thoroughly kicked by an early prototype in public nonetheless. [11:28.780 --> 11:38.880] We made this shell out of thin sheet steel, turned out to be too thin, which was good because we then went back to the drawing board or the drawing computer in any case. [11:39.160 --> 11:43.740] And my friend Michael, who's a mechanical engineer, does this wonderful design. [11:43.900 --> 11:47.160] Turns out you just can't buy spheres, giant metal spheres. [11:48.300 --> 11:53.480] Like you can't just like Google it and find Spheres R Us on the Internet or something that will sell you spheres. [11:54.400 --> 12:00.380] There are things around, but they're either too heavy to be easily driven by motors or they're too thin. [12:00.380 --> 12:02.180] If they're light enough, they'll dent. [12:02.380 --> 12:08.620] So Mike made this really nice design, which you can basically make out of a flat sheet of metal. [12:08.920 --> 12:10.780] In our case, one eighth inch aluminum. [12:11.120 --> 12:16.860] And we got this water jet cut from some friends of ours who were not supposed to say who they are, but they were very kind. [12:17.080 --> 12:19.400] They have a startup and they have a water jet cutter. [12:19.560 --> 12:23.440] If you haven't seen a water jet cutter, it's a computer controlled water jet. [12:23.440 --> 12:27.140] It's 45 kilopascals, which is just an enormous amount of pressure. [12:27.300 --> 12:30.220] It will cut through an inch of steel just using pressurized water. [12:30.680 --> 12:31.680] They're kind of amazing. [12:32.160 --> 12:32.920] And they're around there. [12:33.000 --> 12:33.620] You can find them. [12:33.660 --> 12:34.860] A lot of machine shops have them. [12:34.940 --> 12:35.700] They're kind of expensive. [12:36.760 --> 12:38.980] But if you want to do stuff like this, you can do it. [12:39.060 --> 12:40.100] There are places on the Internet. [12:40.260 --> 12:41.860] If you want stuff like this, cut it out. [12:42.140 --> 12:45.880] You can basically do up the design file in CAD, send it off. [12:46.100 --> 12:47.920] People will cut this stuff out for you. [12:48.760 --> 12:50.440] So the design is pretty awesome. [12:51.000 --> 12:52.100] Here it fits together. [12:52.100 --> 12:57.920] If you've ever seen like those plywood dinosaurs that fits together like a jigsaw, like that. [12:58.360 --> 13:00.360] And we made a little jig and it all fits together. [13:00.520 --> 13:02.040] And then we weld it so it's totally solid. [13:02.180 --> 13:02.860] So it was pretty great. [13:04.460 --> 13:07.260] So this is what's inside the sphere. [13:07.260 --> 13:09.260] And this is what's called a chassis. [13:09.540 --> 13:11.820] And we have lots of stuff here. [13:12.420 --> 13:15.700] At the top, we have an RF communication. [13:15.820 --> 13:19.280] This is how the spheres tell where they are. [13:19.380 --> 13:22.000] And this is how we send them commands through this. [13:22.100 --> 13:24.460] We have four LED illuminators. [13:24.720 --> 13:27.060] They have LEDs on them. [13:27.220 --> 13:28.460] We can address each one. [13:28.580 --> 13:32.460] We can turn them any color we want to because they have RGB LEDs on them. [13:32.620 --> 13:34.040] We have a sound module. [13:34.160 --> 13:38.500] This is basically a little car stereo with a remote controlled MP3 player. [13:38.500 --> 13:46.740] You just put in a thumb drive with whatever sounds you want to do or music or sound effects and away you go. [13:49.000 --> 13:58.480] I heard a lot of people explain when we're running the orbs, they explain earnestly to each other about how there's a camera in this dome. [13:58.480 --> 14:00.560] And that's how it avoids running into people. [14:00.840 --> 14:02.500] Turns out there's no camera at all. [14:02.680 --> 14:05.680] In fact, there are no real sensors on this to detect people. [14:06.000 --> 14:08.520] We are doing Wizard of Oz with remote controls. [14:08.780 --> 14:11.400] So we're faking it a little bit when people are around. [14:12.100 --> 14:15.140] What this dome is here is to keep the dust out of the computer stuff. [14:15.140 --> 14:22.700] So basically, it's just this nice dome and we got to put a gasket and a flat thing and you can just unscrew these things to get at the computer. [14:22.900 --> 14:23.880] You can screw that dome on. [14:24.040 --> 14:25.600] Keeps most of the dust out. [14:26.120 --> 14:36.900] Okay, so there's a drive shaft I was saying and this rotates around the axle to transmit the motor to spin the shell. [14:37.920 --> 14:40.560] Steering servos is right there. [14:40.980 --> 14:43.740] Motor controllers are these aluminum boxes. [14:43.740 --> 14:50.840] We kind of rolled our own and at the very bottom is a battery pack that's full of those same surplus lead-acid batteries. [14:51.040 --> 14:53.440] And lead-acid battery was a good choice for this because they're nice and heavy. [14:55.000 --> 14:58.420] Okay, so let's talk about the geek stuff. [14:59.740 --> 15:05.060] We wanted some fairly sophisticated capabilities, so we got some fairly sophisticated CPU. [15:06.200 --> 15:07.940] This was a Linux ARM processor. [15:08.400 --> 15:12.560] This was pretty much state-of-the-art when we started building this three years ago. [15:15.560 --> 15:20.340] Basically, all the software and hardware, all our designs are open source. [15:21.280 --> 15:23.360] We have a wiki where all this stuff is on. [15:23.520 --> 15:25.740] It's not necessarily well organized, but it's all there. [15:26.380 --> 15:35.120] We did a custom AVR daughter board that sits on this and generates a lot of real-time signals like pulse width modulation from motor controls. [15:35.120 --> 15:39.980] This is basically the same class as an Arduino board, the same complexity. [15:40.320 --> 15:44.480] You can program it using the Arduino tools, although we're using GCC. [15:46.580 --> 15:49.100] So pretty straightforward stuff. [15:50.180 --> 15:57.860] Sensors, our fundamental goal is kind of autonomous behavior, so the orbs have to know where they are and where the other ones are. [15:57.860 --> 16:01.020] And we do this using navigation sensors. [16:01.200 --> 16:05.600] And the global positioning thing is a big part of that. [16:05.740 --> 16:11.720] And it turns out we had to put that outside the shell on a little ear because the shell interfered with the radio signals because it's a Faraday cage. [16:11.920 --> 16:12.060] Duh. [16:13.620 --> 16:17.540] We have accelerometers and Yarray gyros. [16:18.020 --> 16:20.920] And it turns out you can put all that information together. [16:21.360 --> 16:36.180] And then we are using RF and actually Wi-Fi for debugging, but we use this Zigbee Pro, which is, I think, 60 milliwatts, which gives you something like a mile line of sight communication ability, which is pretty great. [16:39.180 --> 16:40.700] So what do we do with them? [16:40.700 --> 16:45.840] Well, we took them to the desert and the desert is really nice because a super receptive audience out there. [16:45.920 --> 16:47.040] They are really into it. [16:47.680 --> 16:52.700] As I said, when there are people around, they don't have good sensors to not run over you. [16:52.780 --> 16:55.080] So we wizard of Oz it with remote controls. [16:55.220 --> 16:58.440] We're a little bit of a shame to this because that's not our final goal. [16:58.700 --> 17:04.420] But as far as where people are going, safety is of primary importance. [17:04.540 --> 17:07.600] We just can't guarantee that these things won't roll over you and they're heavy. [17:07.740 --> 17:09.740] And I have scars on my shin to prove that. [17:10.680 --> 17:11.280] All right. [17:13.120 --> 17:14.720] You can do fun stuff in the desert. [17:14.900 --> 17:16.520] Everything is better with pyro, right? [17:16.740 --> 17:23.100] So one year we strap, zip tied some pyro to it as part of a big show with some friends of ours. [17:23.240 --> 17:25.900] It had a giant art piece that spewed fire and everything. [17:26.060 --> 17:26.600] It was awesome. [17:29.200 --> 17:31.660] And so we were pretty happy with how they worked. [17:31.900 --> 17:32.640] You can see some of the... [17:32.640 --> 17:34.020] This is taken at Burning Man. [17:34.100 --> 17:35.960] You can see some of the fire in the background. [17:36.300 --> 17:38.700] And this is what it looks like. [17:38.900 --> 17:43.360] And as I mentioned, we use a remote control when there's crowds around. [17:43.520 --> 17:50.760] And so we can kind of interact with people and dance with them and steer around them and usually not steer into them. [17:52.540 --> 17:55.520] So we had a lot of fun with these... with the remote controls. [17:55.680 --> 17:57.120] We took them to a lot of places. [17:58.660 --> 18:00.160] Here's one place we took them to. [18:00.680 --> 18:02.940] Coachella is this giant music festival. [18:02.940 --> 18:04.860] It's in the desert of Southern California. [18:06.000 --> 18:07.740] The year we went was 2008. [18:07.740 --> 18:10.960] They had Roger Waters and Prince and all sorts. [18:11.040 --> 18:11.840] And so it's a big deal. [18:12.100 --> 18:16.820] And the organizers of Coachella know that they can get art on the cheap from Burning Man. [18:17.020 --> 18:21.580] So they paid us probably what they spent on the buffet for Prince. [18:22.100 --> 18:22.940] I'm guessing. [18:23.160 --> 18:23.720] Maybe less. [18:24.360 --> 18:30.200] And so we were happy to go down there and sleep in tents and roll our stuff around and entertain the ravers. [18:30.880 --> 18:35.380] So you can't see from this photograph. [18:35.460 --> 18:41.140] But the band that we took our robots to see on this particular evening was none other than Kraftwerk. [18:41.400 --> 18:41.480] Right? [18:42.080 --> 18:47.020] And I'm like, is this some kind of nerdgasm taking your robots to see Kraftwerk? [18:47.200 --> 18:47.280] Right? [18:48.240 --> 18:51.540] If you don't know Kraftwerk, they basically invented techno. [18:51.820 --> 18:51.920] Right? [18:52.160 --> 18:54.760] So they are old school electro. [18:55.400 --> 18:59.880] They sing cute songs like, you know, I'm the operator with my pocket calculator. [18:59.880 --> 19:04.940] But, you know, our robots were actually less impressed with their robot song. [19:05.900 --> 19:06.420] All right. [19:06.540 --> 19:08.720] So that's not the only place we took them. [19:08.800 --> 19:10.880] We also took them to India. [19:11.780 --> 19:14.600] There was a tech festival in India called Tech Reedy. [19:14.720 --> 19:17.460] They said, you know, do you want to come and show off your robots? [19:17.500 --> 19:18.500] And how do you say no? [19:18.840 --> 19:21.440] Of course you say, yes, we'll go to India and show off our robots. [19:21.640 --> 19:26.160] And so they paid for our airfare and they put us up and they were really nice. [19:26.260 --> 19:28.200] And they showed us around. [19:28.680 --> 19:32.360] So I'm here with my two pals. [19:32.580 --> 19:33.840] That's Lee and that's Marnia. [19:34.040 --> 19:36.780] And these are some of our hosts from IIT Kanpur. [19:37.260 --> 19:39.300] IIT is like the MIT of India. [19:39.620 --> 19:41.660] Those guys are way smarter than you or me, man. [19:41.780 --> 19:42.500] Those guys are scary. [19:42.500 --> 19:47.600] If you've ever seen the entrance exam to get into IIT, it is a fearsome thing. [19:47.740 --> 19:50.860] And they were super nice to us and they hosted us. [19:50.860 --> 19:54.920] So we rolled them around at this tech fair. [19:55.080 --> 19:56.860] And this is basically what we saw. [19:57.340 --> 19:58.640] People were very into it. [19:58.680 --> 19:59.740] We were pretty much mobbed. [19:59.780 --> 20:00.520] We were swarmed. [20:01.860 --> 20:04.260] A lot of people didn't really know what to think. [20:05.340 --> 20:10.620] This kind of mechanical art is not a big thing in a lot of places. [20:11.220 --> 20:16.500] A very common question was, sir, can you tell me what business purpose these things have? [20:16.620 --> 20:20.140] And that was a little tough to answer. [20:20.180 --> 20:23.980] I'm not sure I really have a good question about that other than maybe, oh, it draws a crowd. [20:26.360 --> 20:29.200] So that's what it was like outside. [20:29.420 --> 20:32.400] This is right outside the IIT Kanpur campus. [20:32.680 --> 20:34.640] And it's really like that. [20:34.760 --> 20:39.720] It was great seeing a different culture and how they react to things. [20:41.400 --> 20:44.920] And basically, we had a good time. [20:45.920 --> 20:46.460] All right. [20:47.840 --> 20:54.740] And so one of the things, you know, we put the crate together with tools. [20:56.160 --> 20:57.680] And we took our own tools. [20:57.720 --> 20:59.860] And we had to take them back to take the tools back. [20:59.860 --> 21:01.940] We had to put them in the crate, which leaves a problem. [21:02.100 --> 21:05.620] How do you close the crate with the tools to close the crate inside the crate? [21:05.720 --> 21:11.260] So this was a cricket bat that just happened to be lying around, which was a good thing. [21:13.060 --> 21:13.520] All right. [21:13.520 --> 21:20.380] So I'm going to geek out a little bit, basically put up the slide just to show you we're badass about this autonomy thing. [21:20.560 --> 21:21.380] We're still working on it. [21:21.500 --> 21:23.500] We have made some significant progress. [21:24.760 --> 21:30.560] The Kalman filter is a statistical method of sensor fusion, basically. [21:30.840 --> 21:33.300] Kalman filter got the Apollo rocket to the moon. [21:33.780 --> 21:35.160] It's a well-known algorithm. [21:35.720 --> 21:42.720] And we have these sensor inputs, like I mentioned, the GPS, the accelerometer, the gyroscopes and the odometry. [21:42.720 --> 21:48.180] We can measure how far we've rolled, basically, just by counting the revolutions of the sphere. [21:49.600 --> 21:52.680] So Kalman filter devotees will recognize this. [21:52.700 --> 21:58.520] In fact, the extended Kalman filter that Jacobian there kind of linearizes this stuff because it's nonlinear. [21:58.520 --> 22:03.020] And this is actually the Kalman smooth thing. [22:03.600 --> 22:04.360] Let's see. [22:04.560 --> 22:13.860] I think the the GPS is in red and the Kalman filter is in blue with the with the inclusion of all the other data. [22:14.500 --> 22:23.660] For this test, we're basically running around the central circle of a soccer field or football pitch, depending on what you want to call it. [22:23.660 --> 22:26.420] And that's, I think, 12 meters in diameter or something. [22:26.560 --> 22:29.240] So it is it was pretty good. [22:29.420 --> 22:31.680] And we're working on path planning. [22:31.740 --> 22:33.100] And this is actually real data. [22:33.520 --> 22:36.300] Red is the desired path and green is the actual path. [22:36.480 --> 22:40.440] We're kind of being gnarly on how tight we can make some of those turns. [22:40.440 --> 22:46.180] But we are working on doing some more autonomous stuff and working on you can do. [22:47.060 --> 22:50.360] We hope that very simple algorithms will have some interesting behavior. [22:50.580 --> 22:58.240] Like there's some swarming algorithms where you can say, OK, gang, everybody go to the center of gravity of the group. [22:58.280 --> 22:59.740] But don't get to any close. [22:59.840 --> 23:02.660] Don't get too close to any particular one of your neighbors. [23:02.660 --> 23:12.560] And there are flocking algorithms with which people use for CGI to do, you know, troop movements or birds or fish, schools of fish. [23:12.700 --> 23:14.160] We want to do this in real life. [23:14.160 --> 23:15.920] And we're working on this right now. [23:16.420 --> 23:17.860] This stuff is kind of working. [23:18.120 --> 23:21.740] We're kind of a little bit stumped on where to do this. [23:21.880 --> 23:25.020] This works really nice on AstroTurf, but you go to a real park. [23:25.020 --> 23:33.340] Basically, there's, you know, there's those little divots where the ground has been extra fertilized by dogs and the grass is maybe a little stiffer. [23:33.700 --> 23:35.040] Stuff like that messes us up. [23:35.200 --> 23:38.940] So we need to do a little bit better job on the feedback and actually doing navigation. [23:38.940 --> 23:42.540] So this will work in less controlled scenarios. [23:43.600 --> 23:44.200] All right. [23:44.480 --> 23:48.360] So that's about it for the technical portion of the talk. [23:48.360 --> 23:56.720] And maybe you're thinking, wow, you know, rock festivals and travel to exotic places. [23:57.340 --> 23:59.120] This sounds like the lifestyle for me. [23:59.340 --> 24:01.420] So John, tell me, how do I get started with robots? [24:01.620 --> 24:05.040] And the next last couple of slides are on some resources to do that. [24:05.860 --> 24:06.240] Okay. [24:06.320 --> 24:08.420] First thing, it's a lot of work, right? [24:08.620 --> 24:15.780] I'd say if you run a robot for five minutes, you're going to spend an hour fixing it, building it, tuning it up. [24:16.660 --> 24:19.400] They break these particular ones. [24:19.540 --> 24:21.700] We've never built anything like this before. [24:22.260 --> 24:25.900] So some things we might have done differently now that we've done it already. [24:26.080 --> 24:27.980] But it was a wonderful learning experience. [24:29.100 --> 24:29.700] All right. [24:29.860 --> 24:32.300] So ways to get started with robotics. [24:32.540 --> 24:35.080] I hear some cheers. [24:35.320 --> 24:37.960] Who knows about FIRST Robotics? [24:38.440 --> 24:38.920] Okay. [24:39.160 --> 24:42.840] If you don't know about FIRST Robotics, I'm going to tell you because it is full of awesome. [24:42.980 --> 24:44.800] This is founded by Dean Kamen. [24:44.800 --> 24:49.320] It's a program for students in high school and even younger. [24:50.380 --> 24:51.840] Basically, the high school... [24:51.840 --> 24:54.900] This week, over the winter, every high school... [24:54.900 --> 24:59.820] I think there are about more than 1,000 schools participating in this all over the country, all over the world. [25:00.060 --> 25:03.120] Every high school gets a box of parts and they build a robot. [25:03.400 --> 25:09.040] And at the end of the building time, they compete with other high schools in a non-violent challenge. [25:09.040 --> 25:11.200] It's like robot soccer or this one. [25:11.300 --> 25:13.200] I think you have to lift these balls over the thing. [25:13.840 --> 25:14.840] It is awesome. [25:15.060 --> 25:23.880] If your high school experience was like mine, you know the value of some cool geeks coming in to you and say, Hey, there is life after high school. [25:24.040 --> 25:25.180] You will survive, right? [25:26.280 --> 25:27.640] And Dean Kamen... [25:27.640 --> 25:30.900] You should really watch this on the FIRST. [25:31.200 --> 25:36.880] USFirst.org website has this video introduction where he speaks absolutely from the heart. [25:36.880 --> 25:44.280] And he is saying, you know, in this culture, we reward the sports stars, the music stars, you know, the athletes and people. [25:44.460 --> 25:47.000] What about the people who make sure you have clean water to drink? [25:47.140 --> 25:49.800] What about the people who make sure that your elevators don't crash? [25:50.000 --> 25:54.640] You know, what about the scientists and engineers who are responsible for all this awesome stuff we have? [25:54.720 --> 25:57.920] You know, we have computers in our pants now, right? [25:58.880 --> 26:01.080] Why aren't those guys rock stars too, right? [26:01.180 --> 26:01.760] They should be. [26:03.380 --> 26:06.100] So, he started this program. [26:06.340 --> 26:18.860] Dean Kamen, he did the Segway or stuff, so that you can get a taste of engineering in high school and hang out with nerds and older nerds and it's just the best thing in the world. [26:19.200 --> 26:23.200] And if you want to learn to do robots, go volunteer for one of these things. [26:23.380 --> 26:24.920] Let the kids make all the mistakes. [26:25.100 --> 26:27.400] Let the kids let out the magic smoke, right? [26:27.580 --> 26:27.880] Right? [26:28.700 --> 26:30.020] It's an awesome way to learn. [26:30.200 --> 26:32.340] And that's actually one of the ways I got into it. [26:32.520 --> 26:36.980] And I can't say enough good things about it. [26:38.900 --> 26:44.040] Was anybody... actually, it's been around for 10 years, so I've actually met some people who were in this... [26:44.040 --> 26:45.080] in high school. [26:45.240 --> 26:45.560] Is anybody? [26:45.780 --> 26:46.320] Yeah, alright. [26:46.620 --> 26:47.700] So, power to you. [26:47.780 --> 26:49.640] You guys know how awesome this is. [26:49.760 --> 26:51.360] I wish it was around when I was in high school. [26:52.520 --> 26:53.980] Definitely good stuff. [26:54.580 --> 26:54.720] Alright. [26:55.980 --> 26:58.160] This is kind of Bay Area centric. [26:58.260 --> 27:01.040] I apologize, but there's something called the Homebrew Robotics Club. [27:02.840 --> 27:06.520] This... you may remember something called the Homebrew Computer Club. [27:06.640 --> 27:17.920] Also in the Bay Area, this is basically where Steve and Woz brought their first prototypes to and hung out with people kind of, I think, in the same state of development. [27:17.920 --> 27:19.860] Robots are not ready for prime time. [27:20.260 --> 27:26.040] Back in the 70s, when the Homebrew Computer Club was going, computers were not ready for prime time. [27:26.140 --> 27:27.320] But look how far they've come. [27:27.520 --> 27:29.760] Look how far we might be able to go with robots. [27:30.020 --> 27:32.960] And right now is the time to get into it, because you can do it as a hobbyist. [27:34.140 --> 27:34.520] Okay. [27:35.400 --> 27:36.740] Maybe you're not in the Bay Area. [27:37.060 --> 27:44.460] There are all sorts of little nuggets, little nexuses of excitement around technology and building stuff. [27:44.460 --> 27:47.300] Support your local hackerspace, hackerspace.org. [27:47.360 --> 27:49.800] I think there's at least two in New York City. [27:49.960 --> 27:51.300] There's, what, three or four in Boston. [27:51.360 --> 27:52.360] They're all over the place. [27:53.220 --> 27:55.840] There's all sorts of awesome open source hardware. [27:57.660 --> 27:58.060] MakerBot. [27:58.200 --> 28:02.140] You can build your own CNC deposition machine. [28:03.300 --> 28:04.100] Adafruit, SparkFun. [28:05.080 --> 28:06.260] Limor and Phil were here. [28:08.040 --> 28:09.320] I didn't put them on the slide. [28:09.420 --> 28:09.960] I should have. [28:10.140 --> 28:12.040] Mitch and Jimmy will teach you how to solder. [28:12.040 --> 28:14.340] If you're a soldering noob, go do it. [28:14.420 --> 28:14.820] It's easy. [28:14.960 --> 28:15.220] It's fun. [28:15.340 --> 28:16.260] You get to melt metal. [28:16.740 --> 28:17.440] It's great. [28:19.160 --> 28:21.800] They're doing that, I think, all weekend down in the Mezzanine. [28:22.220 --> 28:23.320] Tell them I sent you. [28:23.640 --> 28:24.380] They're awesome. [28:24.580 --> 28:25.180] It's really easy. [28:25.260 --> 28:25.860] It's really fun. [28:26.480 --> 28:27.740] Everybody was a noob once. [28:27.760 --> 28:28.340] I was a noob. [28:28.440 --> 28:30.220] Don't let that stop you, because it's really fun. [28:32.020 --> 28:38.720] So anyway, so maybe you've seen all this kind of community stuff and this non-violent competition. [28:38.720 --> 28:40.120] Well, maybe that's not your thing. [28:40.280 --> 28:43.700] It's like, maybe this is your thing. [28:44.520 --> 28:45.160] Robo games. [28:45.420 --> 28:45.580] All right. [28:45.720 --> 28:46.520] Combat robotics. [28:47.060 --> 28:52.700] They may not be on TV, but you can bet your ass they are still around and they are knocking the shit out of each other. [28:52.820 --> 28:53.520] It's amazing. [28:53.740 --> 28:54.060] All right, right. [28:54.340 --> 29:01.440] If you ever want to build something and put it in the path of something that looks like a lawnmower with titanium blades or something, this is the place to do it. [29:01.520 --> 29:02.380] Or a flamethrower. [29:03.280 --> 29:18.180] Okay, if you go to robo games.net and you're into this, there is something like a 400 page PDF on their homepage of their site, which tells you there's a whole ecosystem of people building combat robot parts. [29:18.340 --> 29:20.480] You can buy kick ass speed controllers. [29:20.660 --> 29:21.420] You can buy wheels. [29:21.540 --> 29:22.300] You can buy chassis. [29:22.540 --> 29:27.120] If you want to throw a lot of money at it, you can do that and build a kick ass bot. [29:27.260 --> 29:30.460] I wouldn't recommend that the first time, because you're going to get your ass handed to you. [29:30.900 --> 29:32.480] But you can definitely start up. [29:33.360 --> 29:36.360] Every year in the Bay Area is robo games. [29:36.540 --> 29:41.180] And people come from all over the world to do this robot combat competition. [29:41.880 --> 29:42.360] All right. [29:42.700 --> 29:45.920] I don't know how I'm doing on time, but I'm pretty close to done. [29:46.140 --> 29:47.740] So thank you for your attention. [29:48.040 --> 29:49.120] I appreciate it. [29:58.790 --> 30:00.990] Got some time for questions, if anybody's got any. [30:08.820 --> 30:10.540] Hey, I really love your work. [30:10.540 --> 30:21.580] I was curious, if somebody wanted to get into this commercially, like not just as an art hobby, if you... is there a professional applications for this sort of work? [30:21.600 --> 30:26.620] Or are you more going to be doing like a research in a lab or... [30:27.200 --> 30:28.500] There's all sorts of things. [30:28.660 --> 30:31.300] There's a number of startups that are doing robots. [30:32.720 --> 30:33.300] Anybots. [30:34.280 --> 30:37.780] A couple of things which I'm forgetting the names of. [30:38.220 --> 30:39.860] Some of those places are hiring. [30:40.540 --> 30:41.540] There's people... [30:41.540 --> 30:45.320] As far as commercial applications, there's a lot of medical robotic applications. [30:47.580 --> 30:50.620] There's a lot of research in a lot of the universities. [30:51.180 --> 30:53.600] Carnegie Mellon's been doing robots for a long time. [30:55.260 --> 30:59.580] Cornell, Berkeley, Santa Cruz are the ones I'm aware of. [30:59.640 --> 31:01.940] I'm sure there are lots more doing badass stuff. [31:02.380 --> 31:03.780] Look on YouTube. [31:03.780 --> 31:07.120] You can see the autonomous helicopters and things like that. [31:07.860 --> 31:08.440] Great stuff. [31:09.340 --> 31:10.620] Did that answer your question? [31:11.000 --> 31:15.020] Well, I was thinking more in line with the specific projects that you've done. [31:15.200 --> 31:17.340] I guess it's the same question the people in India asked. [31:17.560 --> 31:18.380] Right, right, yeah. [31:18.560 --> 31:21.460] And I don't have a good answer. [31:21.460 --> 31:29.860] There is a Danish company, I think, in Scandinavia, which is making something called Rotundus, which is a spherical robot. [31:30.080 --> 31:33.680] And I think their application is for security. [31:33.980 --> 31:41.700] Like this thing will replace a security guard and roll around your perimeter and crush intruders and stuff like that. [31:41.840 --> 31:42.120] Like Rover? [31:42.660 --> 31:44.180] Exactly, exactly, exactly. [31:44.340 --> 31:45.900] Like the prisoner. [31:46.420 --> 31:53.480] I'm not sure what state that is in, but they seem to be existing and they have pictures of their thing rolling around. [31:54.020 --> 31:56.460] And their thing is cool because it's sealed, it will actually go on water. [31:57.460 --> 31:57.780] Oh. [31:58.800 --> 31:58.940] Yeah. [31:59.440 --> 32:01.060] Do you have any robots here? [32:01.320 --> 32:02.060] I do not. [32:02.180 --> 32:02.500] I'm sorry. [32:04.640 --> 32:04.980] No. [32:05.360 --> 32:05.960] I'm sorry? [32:06.200 --> 32:07.420] I'll bring a robot on Sunday. [32:07.460 --> 32:08.360] Bring a robot on Sunday. [32:08.460 --> 32:08.700] Excellent. [32:08.700 --> 32:08.720] No. [32:09.080 --> 32:09.760] Where's it going to be? [32:12.600 --> 32:13.480] Ham Radio Place. [32:13.480 --> 32:14.320] Ham Radio Place. [32:14.460 --> 32:16.540] Robots on Sunday by the Ham Radio Place. [32:17.600 --> 32:18.340] Other questions? [32:19.020 --> 32:19.180] Yeah. [32:19.500 --> 32:22.800] So what's the fastest you've taken the swarm so far? [32:23.440 --> 32:24.340] The fastest? [32:24.660 --> 32:25.060] The fastest. [32:25.220 --> 32:25.360] Yeah. [32:25.440 --> 32:26.200] This is a good question. [32:26.400 --> 32:27.960] People ask, how fast do they go? [32:28.140 --> 32:30.520] And my answer to that is, faster than you can run. [32:33.620 --> 32:39.280] So, yeah, to actually be serious, we actually took down the top speed quite a bit. [32:39.540 --> 32:39.940] Yeah. [32:39.940 --> 32:43.000] Because we kind of over designed the motor controllers, which is not a bad thing. [32:43.400 --> 32:48.340] But, you know, for some reason, when one gets away from you, you want to be able to track it down. [32:48.480 --> 32:49.940] And there's a funny story. [32:50.160 --> 32:57.280] We have kill switches on the axle where you can turn off the power to the motor just in case one starts misbehaving. [32:57.960 --> 33:01.340] They don't do it often, but now and then they will. [33:01.860 --> 33:06.840] And we took it to the desert and turns out those switches managed to get some dust in them. [33:06.980 --> 33:09.620] We thought they were more dust resistant than they were. [33:09.760 --> 33:15.220] So we were reduced to begging for condoms to cover these switches to dust proof them. [33:16.600 --> 33:17.880] Have they broken any bones? [33:18.540 --> 33:19.600] Not to my knowledge. [33:20.140 --> 33:21.780] Although we have run into a number of people. [33:22.640 --> 33:23.080] Thanks. [33:23.520 --> 33:23.940] Sure. [33:26.260 --> 33:28.120] This is a bit off topic, I guess. [33:28.640 --> 33:31.820] That cricket bet you mentioned, was it in case of a zombie invasion or something? [33:32.320 --> 33:33.080] I guess so. [33:33.180 --> 33:34.860] It was just lying there, but it was certainly handy. [33:40.380 --> 33:42.740] Thanks, John, for the shout out for Rob Exotica. [33:42.860 --> 33:43.280] Thanks a lot. [33:43.720 --> 33:47.080] So my question would be, you had two projects there. [33:49.000 --> 33:55.140] Like an anthropomorphoid robot, like a man... [33:55.140 --> 33:56.320] Right, like a mechanical man. [33:56.520 --> 33:59.500] A mechanical man, and you had something like a sphere rolling around. [33:59.760 --> 33:59.880] Right. [34:00.500 --> 34:03.460] In general, there are robots all around us, all the time. [34:03.640 --> 34:05.260] Like an ATM is a robot. [34:05.500 --> 34:07.820] Like a subway is a robot, to an extent. [34:08.220 --> 34:10.760] Most of the robots out there are kind of like enslaved. [34:10.760 --> 34:12.660] They have to carry people around. [34:12.660 --> 34:13.980] They have to give out cash. [34:14.120 --> 34:14.220] Right. [34:14.340 --> 34:15.000] Stuff like that. [34:15.160 --> 34:17.980] They have to weld stupid cars that no one wants to buy. [34:18.280 --> 34:18.640] Okay. [34:18.800 --> 34:18.880] Yeah. [34:19.260 --> 34:22.200] So my basic idea would be... [34:22.200 --> 34:28.720] And the popular reception of robots is still more like at the chassis and not at the rolling sphere end. [34:28.720 --> 34:37.280] So I think more and more robots should be out there who do not fulfill this stupid like R2-D2 like ass cancer kind of thing. [34:37.520 --> 34:37.540] Yeah. [34:38.340 --> 34:47.800] So I think we all should like get our heads together and think of robots that are not like the stupid like 1950s future robots out there. [34:47.960 --> 34:50.540] That's just like my five cents. [34:50.860 --> 34:51.440] Absolutely. [34:51.620 --> 34:52.920] That's an excellent point. [34:53.080 --> 34:55.660] Robots are only limited by your imagination. [34:55.660 --> 34:58.800] If you can think up a cooler one, and I bet you can, do it. [34:58.940 --> 34:59.380] It would be awesome. [34:59.920 --> 35:00.240] Thank you. [35:01.140 --> 35:01.380] Question. [35:02.420 --> 35:07.300] So I keep hearing in the news about all these unmanned military vehicles. [35:08.160 --> 35:16.400] Does anyone in the hobbyist community know of their... how related these are to the kind of stuff that you can build at home? [35:16.760 --> 35:18.660] Or are we not allowed to know that? [35:20.920 --> 35:21.400] Yeah. [35:21.600 --> 35:24.840] Well, there's a company out of Boston called iRobot which does... [35:26.740 --> 35:28.940] Basically, not too different from chassis. [35:29.100 --> 35:30.440] I mean, obviously much more sophisticated. [35:30.680 --> 35:35.440] Remote control things with a remote actuator arm and a video. [35:35.900 --> 35:38.320] And that's for disarming IEDs. [35:38.580 --> 35:42.100] And they're actually using those in combat things. [35:42.300 --> 35:44.560] But this is under complete human control. [35:45.040 --> 35:49.820] And then, of course, the Predator drones and things which are also supposedly under human control. [35:49.820 --> 35:53.140] I personally don't know about things which are autonomous. [35:53.840 --> 36:00.180] I'd be a little scared about that because, you know, things are not that clever to begin with. [36:00.440 --> 36:05.700] And I'm not sure if that answers your question. [36:05.800 --> 36:09.780] There are robots with military applications on the battlefield right now. [36:09.780 --> 36:13.880] I don't know anything about autonomous ones. [36:14.100 --> 36:17.060] The DARPA Grand Challenge has some good examples of universal [36:21.190 --> 36:21.690] cars. [36:21.950 --> 36:22.810] Excellent point. [36:23.030 --> 36:30.270] The DARPA Grand Challenge where they have the cars that basically have to navigate autonomously over a long distance. [36:30.270 --> 36:32.750] And that is funded by DARPA. [36:32.950 --> 36:34.450] So guess what they're going to do with that? [36:34.910 --> 36:35.510] Let me get that working. [36:36.050 --> 36:39.050] Also, there is a DIYdrones.org. [36:39.130 --> 36:46.730] It's a website that is basically an amateur community for doing autonomous or self-guided aerial vehicles. [36:46.930 --> 36:53.790] You're about $500 and some elbow grease away from having a miniature, very lightweight cruise missile, if that's what you want to do. [36:55.830 --> 37:08.490] And the capabilities are actually surprisingly good because it's this wonderful intersection of the model aircraft community that have long since worked out how to fly many, many pounds of things for an extended amount of time. [37:08.910 --> 37:10.990] And cheap, miniaturized electronics. [37:11.490 --> 37:12.830] And right now you can... [37:12.830 --> 37:21.730] It's literally, I think, $300 from SparkFun to buy a six-degree-of-freedom GPS-guided IMU unit that you can plug into your choice of model aircraft. [37:22.090 --> 37:23.410] So there's a lot of fun there. [37:24.010 --> 37:24.490] Thank you. [37:24.570 --> 37:25.090] Excellent point. [37:25.230 --> 37:26.810] And a glaring omission from my presentation. [37:27.010 --> 37:27.690] I will add that next time. [37:28.390 --> 37:33.810] There's actually just this week, the Ministry of Defence in the UK announced that they have a prototype. [37:33.990 --> 37:38.750] They say it's not ever going to be used in a field of warfare, but it is supposed to be completely autonomous. [37:39.130 --> 37:40.690] But somebody at the end has a kill switch. [37:41.310 --> 37:43.350] How well that kill switch works, I don't know. [37:43.490 --> 37:44.570] Whenever they decided to... [37:44.570 --> 37:50.950] I apologize, I can't remember the actual name of it, but they named it after the Celtic God of War, which doesn't exactly fill me full of confidence. [37:52.210 --> 37:52.690] Okay. [37:52.850 --> 37:54.190] How are we doing on time? [37:54.510 --> 37:55.490] There's a time... [37:55.490 --> 37:56.050] 15 minutes. [37:56.190 --> 37:56.730] 15 minutes. [37:56.910 --> 37:57.030] Okay. [37:57.150 --> 37:59.670] Got some time for more questions. [37:59.870 --> 38:02.850] We can also take a little break before the next series of awesome. [38:03.310 --> 38:09.810] Do you have any information about the Korean robots along the border with North Korea that have just been deployed recently? [38:10.110 --> 38:10.790] I do not. [38:10.990 --> 38:11.390] I do not. [38:18.630 --> 38:22.450] What's the strangest failure mode that you hit with the spheres? [38:23.270 --> 38:24.590] Strangest failure mode? [38:27.350 --> 38:28.950] That's a good one. [38:31.430 --> 38:34.010] They seem to have lost their sensors. [38:34.010 --> 38:37.350] There's a servo feedback which sends... [38:37.350 --> 38:41.090] You tell the motor to put the weight at a certain angle. [38:41.210 --> 38:42.030] It's supposed to go there. [38:42.770 --> 38:48.690] And basically, sometimes, if you don't give it a big enough dead band, the thing will kind of wiggle there. [38:48.730 --> 38:53.190] So you see these spheres there, and they're kind of shimming there. [38:53.350 --> 38:56.070] It's kind of neat, but completely unintended and kind of wrong. [38:56.070 --> 38:57.510] And we finally fixed it. [38:57.610 --> 39:00.170] But it was kind of a cool, emergent behavior, nonetheless. [39:00.510 --> 39:02.050] Because it's kind of this nonlinear thing. [39:02.350 --> 39:04.970] And the PID loop doing control on it. [39:05.050 --> 39:07.650] And it wasn't a good solution to that control problem. [39:08.410 --> 39:09.310] But it was kind of cool. [39:10.750 --> 39:11.150] So... [39:11.150 --> 39:15.290] Have you done any work with artificial intelligence in your robots? [39:15.950 --> 39:16.730] No, no. [39:17.030 --> 39:20.090] Actually, I come up from a background of machine learning and things like that. [39:20.210 --> 39:25.770] And I think AI is a little bit of a misnomer, because I don't think things are really that smart at all. [39:28.230 --> 39:35.850] And I think, if you want to call it AI, this kind of autonomous navigation is about as close as you're going to find. [39:35.990 --> 39:37.410] And there may be other things that do that. [39:38.470 --> 39:44.390] But this is... the autonomous navigation is kind of where we've been focusing our time. [39:45.810 --> 39:49.830] How long do you think we have until the revolution comes and our robot overlords take over? [39:54.600 --> 39:57.360] I, for one, am not losing any sleep at all over. [39:57.360 --> 40:01.140] When you build these things, you realize how brittle, how fragile this stuff is. [40:01.260 --> 40:03.160] And just getting the work at all is kind of miraculous. [40:03.440 --> 40:03.800] So... [40:05.760 --> 40:07.360] I wouldn't lose any sleep. [40:07.560 --> 40:13.940] But, yeah, I can't... I can't prognosticate too far in the future, right? [40:14.020 --> 40:15.100] Because stuff changes so quickly. [40:15.300 --> 40:19.360] Also, do you have any plans for using the orbs for cooperative robotics research? [40:20.360 --> 40:22.580] Well, all our stuff is open source. [40:23.600 --> 40:29.000] If somebody... we're going to put our data sets... our data sets are online for that thing. [40:29.140 --> 40:35.120] So, we started this before a lot of the open autonomous vehicle stuff was done. [40:35.120 --> 40:39.920] We could probably use a lot of their code. [40:40.300 --> 40:42.460] So, we wound up reinventing the wheel. [40:43.180 --> 40:45.120] But, it's all there for sharing. [40:45.400 --> 40:48.940] And if there's something that will save us time, we're definitely on it. [40:49.100 --> 40:49.580] Thank you. [40:49.780 --> 40:50.100] So, thanks. [40:51.760 --> 40:52.780] Just a quick question. [40:53.240 --> 40:55.360] Just some of your next projects that are coming up. [40:56.000 --> 40:56.880] Next project. [40:57.280 --> 41:01.740] Personally, basically, I just got a little grant to do something I call... [41:02.920 --> 41:08.460] Cardo for completely awesome robotic autonomous desk lamps. [41:08.720 --> 41:09.680] And these... experiment. [41:10.240 --> 41:10.620] The E, right? [41:10.820 --> 41:13.460] And these are basically just desk lamps you get from IKEA. [41:13.720 --> 41:15.220] Basically, you put some servos on them. [41:15.280 --> 41:16.620] Going to make five or six of them. [41:17.020 --> 41:19.920] Going to put high brightness LEDs in them. [41:20.420 --> 41:22.560] So, you can change their color. [41:22.560 --> 41:25.220] And you also make them nice white for a nice work lamp. [41:25.880 --> 41:30.300] And these things will hopefully be able to track things around. [41:30.560 --> 41:31.600] Point in particular directions. [41:31.800 --> 41:32.500] Or you press a button. [41:32.740 --> 41:34.360] And it's party time. [41:34.580 --> 41:35.520] And they point in the ceiling. [41:35.700 --> 41:37.780] And they do cool colored things like this. [41:37.860 --> 41:38.380] And strobe. [41:38.480 --> 41:40.700] And maybe move in synchrony. [41:42.720 --> 41:43.160] Right. [41:43.280 --> 41:43.500] Right. [41:43.580 --> 41:43.740] Right. [41:43.740 --> 41:43.960] Right. [41:44.260 --> 41:48.160] That's going to be version two where they actually walk around like Luxo Junior. [41:48.420 --> 41:48.520] Right. [41:48.640 --> 41:48.740] Right. [41:49.440 --> 41:50.180] Voice activated. [41:50.480 --> 41:51.460] Voice activated. [41:51.620 --> 41:51.700] Yeah. [41:52.560 --> 41:53.360] That's a little tougher. [41:53.520 --> 41:54.040] But we can do that. [41:55.400 --> 41:55.560] All right. [41:56.140 --> 41:57.060] They smell the what? [41:58.600 --> 41:58.960] Right. [41:59.280 --> 42:00.840] The scent of blood sensor. [42:02.320 --> 42:02.880] All right. [42:03.040 --> 42:07.060] In the absence of more questions, why don't I stop here? [42:07.240 --> 42:10.960] And that gives people time to run and get a drink or something before the next panel. [42:11.180 --> 42:12.700] Which I believe is on hackerspaces. [42:12.860 --> 42:13.140] Is that right? [42:13.340 --> 42:13.480] Yeah. [42:14.080 --> 42:14.360] Right. [42:14.460 --> 42:15.560] Which is going to be totally awesome. [42:15.680 --> 42:16.460] And I don't want to miss it. [42:16.580 --> 42:17.760] So, I'm going to stop right here. [42:17.980 --> 42:18.980] Thank you for your time. [42:18.980 --> 42:19.080] Thank you, everyone. [42:19.620 --> 42:19.980] Thank you.