Tim Lord for Slashdot: Rich, you are holding what is called the HooperFly, I guess, one instance of a HooperFly. Can you explain what that's about?
Rich: So, this is indeed a HooperFly and it’s – we call it the Racer PEX Quad because there are four motors. This is an open source DIY UAV running sophisticated open source autopilot software called Paparazzi UAV. Also as a part of it, there is a very sophisticated – when I mean sophisticated advanced alien technology from 1 Bit Squared and that is the Lisa/MX.
Slashdot: Okay. Why don’t you start talking about Lisa/MX a little bit? What features does that bring to your UAV?
Rich: Well, Lisa/MX is a true autopilot in that it has an IMU with 10-Degrees-Of-Freedom. And hopefully the hardware engineer can correct me if I make a mistake. It also has a STM32 F4 ARM processor on it. So, it’s a fully capable autopilot and what makes it even more capable is the software that runs on top of that which is Paparazzi UAV has been around since 2003. It evolved out of university projects where they were doing mission-based autonomous flights. So these things do very well in that arena and that was kind of the goal with these frames was to find software and hardware where the common person could go out and make a highly reliable robot--in this case, the HooperFly.
Slashdot: Give a quick rundown of what that means--mission-based as opposed to what?
Rich: So, when I mean mission-based... a lot of people know what a waypoint is. Well, waypoints are just a variable in this environment. Paparazzi is really about code blocks. It has code generators in it. So, you can describe new behavior and plug in new modules. So, if you have new sensor types, you can go ahead and write those and plug right in and have new behavior. So, if you want to give your robot eyes, you could give your robot eyes. If you want to give your robot a nose, you can give your robot a nose.
Slashdot: Could you say something like ‘hover by this waypoint until it’s 80 degrees’
Rich: That's right. So, in a mission-based, so you can like in any program where you're having flow control, Paparazzi gives you these blocks of behavior that you can then can connect and link to, like in a go-to, or if some conditions get satisfied, do this. For instance, one of the conditions that I play with often is, if my GPS telemetry or accuracy goes down below certain level, I will automatically land the craft given that I'm flying in an environment where that happens.
Slashdot: So, I should point out too that what you’re flying here, this is since it is a HooperFly, it’s not the only way it could look. Could you talk about the construction a little bit?
Rich: Right. So the idea was to build multiple flying robots in many configurations because like in the bird world, there are different birds that are very efficient and different, hummingbirds do hovers, albatrosses do greater soaring. So, the idea was if you want to build a – I’m going to call it the unoccupied collective so if you want to build an unoccupied collective AKA a bunch of robots in the air, then you can do that. You can have varying designs so, for instance, this is a quadcopter with a fairly mid-sized prop on it, which you can get a similar performance from this hexacopter, so smaller props, more motors but similar performance. A little more stable because you’ve got more moments of inertia--I'll use that word.
So anyway this is a little more stable. Fights wind a little better, a little lower wind profile, but yet it’s the same software and it’s the same hardware, it allows you to experiment. Because one of the goals of this was you wanted to fly more than fix. So, the idea behind this is, you have a little bit of efficiency in having these tools out here, but what you gain is flight time. You're not repairing your craft, it can fall out of the air from a great height. As you can see, this one is flying for the last year and it's got dings and nicks. And sometimes it bends and you just need to bend it back with a pair of pliers and you are back flying.
Slashdot: Talk about the expensive exotic materials that lets you fix it so easily.
Rich: Okay, so the material, the space age material we're using is called PEX, which is basically a hot cold water pipe that you would use in your homes. And the reason for that was we use PEX and zip-ties because it's readily available. And it's amazingly strong and you can bend it into all kinds of geometries that once bent, it becomes even stronger. So the concept is to build lifting rings with geometric designs, for a lack of a better word, that allows the craft to be sustainable and survive crash.
Slashdot: What's the largest that you've actually pushed this design to get to?
Rich: I have an octocopter that's about 6 feet in diameter once you put it on the ground, so imagine this frame –not with six rotors, but eight, that was this high.
Slashdot: Talk a little bit about the place where you actually get to experiment with these copters because not everybody has a good spot to experiment.
Rich: So, it's really important when you're flying autonomous vehicles. First, you should learn to fly your vehicle. Everyone who flies anything should know how to fly manually. Once you've got that on log, then the next thing is you want to fly a robot, that's fine. Make sure you have a safe area to do that. So here in the city, we have the school grounds, is where you can fly and so we have a large abandoned school area. The school is now gone, but it's safe to fly, it's probably a half mile by a half mile. And it gives you good buffer so you can build algorithms so that if your craft flies outside that area that it will fail in a safe way. So, one of the things about robotic flight is you’re thinking about monitoring and executing safe flight. That's very important.
Slashdot: Is one the things that Paparazzi brings the ability to link some of these craft together?
Rich: Yeah, so Paparazzi like I said before is an open source platform, so anybody out there that has an inclination, you can go on to a GitHub site, paparazziuav.org, is a wiki, there is a GitHub there, all open source. You can design modules, so if you have custom radios that you want to talk craft to craft, that you want to coordinate with, you can do that. Most of the times, as I’m building robots, the goal is not to have to talk to them, they only talk to me so to speak. And the way they do that is through the ground control station. So, this is the ground control station for Paparazzi and as you can see I'm going to turn my little hexacopter, you can see the little red guy turning there. So all I'm doing when – after I program these to fly their missions is monitoring them and then I have a set of protocols that I use if something goes outside the mission plan that hopefully, it'll correct itself because they’re robots. That’s the goal. And the other part is if for some reason they don’t, there are enough fail safes, they can bring the aircraft out of the air. Worst case, you flip a switch, they all fall out of the air.
Slashdot: Now, another aspect of when a project is open source, if people want to build one of these, do you have a lot of information online that would help them do that? Or should they – what should they do?
Rich: So, in the evolution of this frame design and building robots so to speak, the key for me was to find get hardware and software and I've been very happy using Paparazzi and the Lisa/MX autopilot from 1 Bit Squared. That was important to me because I wanted to play around the other variables, the frame design. So to your point with respect to building it, I can build these for people and ship them to them. I have also wanted people just to look at it and build their own version of it. Because in my mind, it was a bunch of tangent circles and I thought that's obvious hopefully people will build these.
That being said, we can do a better job of putting more information up there on how to build these. I have software that I use that does all the calculations for me because one of the things – the first decisions are what motors are you going to use; how big of prop are you going to spin; and then that starts to drive the geometries, and how many of them do you want? Because as you can see, you don't want props to hit. So you want to be able to lay out your frame in a way that doesn't violate the actual function of a frame.
Slashdot: Any of the major laws of physics, you want to respect.
Rich: Yes, yes. And so I’ll go back to this one because it's bigger and brighter. So anyway, so you can see here that it basically uses _____ system that's collapsible, so it takes energy out of the system when it crashes. And the goal is that it just crashes and you pick it up and you walk back and you do more science. Because at the end of the day, you just want to do science.
Slashdot: And even if the piece breaks, you’re talking about pretty cheap material.
Rich: Actually, it’s a good point. So, I’ve been building this frame in this design for two and a half years. I have every single one I’ve ever built because they don’t break. They do bend on occasion. And as we showed earlier, if you bend one of these, usually you can deform it back into shape. This one has been flying for about 3 months and I’m getting better at building robots, so, it doesn’t have as many gains. So, my robots get much better – I don’t know what the right word is, but they just look better. You can tell that
Rich: Yes. Yes, they have less nicks in them. But the thing is all these hardware pieces are readily available. The goal was to build robots with this econo parts as you can, and still have stable flight. So, I tried to build with this econo parts– I use the word econo, some people might say cheap. I just mean econo, it’s value, whatever you want. And get a flying robot because if you can do that, then you can build more of them and you can do more science.
Slashdot: And what kind of sensor payload you send out?
Rich: What kind of payloads? Well, everybody on the planet – yeah. So, everybody on the planet knows about it--putting a camera on the bottom and so that’s a no-brainer. So you can put cameras on here, the kind of cameras I like to put on are any of your near infrared cameras. Some friends of ours who ran a public lab, I think it’s a dotorg, they have some great open source science experiments they can do and one of them is a point and shoot, near infrared camera and I love flying that thing because I can look at plant health with it. So, for me, if I’m flying a camera, I like to fly something that I’m going to do science with. You could fly sensors, air sensors, if you have them small enough, that would be giving a nose to your robot. You could fly – I’ve flown army men up with parachuters on them for kids, so they could get engaged in the project. They really enjoy seeing things parachute out of flying aircraft. Anything you can imagine. And really it’s just a function of how many motors you want to put on it, how big the props you want to swing, how much energy you want to put in the system, it really depe nds on what you’re going to lift.