Slashdot Mirror
Best Way To Build A DIY UAV?
Shojun writes "I am very interested in building my own UAV. Not just one that can fly around happily, but one that I can program to say, take photos every second as it does a barrel roll under a bus (ok, that part may be a pipe dream). I have enough embedded programming experience — it's the hardware which I'm uncertain about. I can go the kit way, and then build the remaining stuff, or get some Dollar Tree Foam boards and build it all. I'm in favor of ease, however. Once the plane is built, buying a dev board seems like a possibility, but I wonder whether it's overkill. Alternatively, if there was a How-to-build example on the net for such an activity that I could adapt, to the degree that I could then program in even completely hardcoded flight instructions, I can certainly take it from there. Thoughts? Has anyone here tried something like this before?"
I'd make sure the Feds have no problem with you running something like this around. Best to make sure you won't get shot down/at.
I shouldn't have to look up acronyms because an editor fails at adding one to the summary. Since I had to look it up anyway -- for those as clueless as me, UAV means Unmanned Aerial Vehicle.
http://paparazzi.enac.fr/wiki/Main_Page
Open source autopilot/software/hardware design for small UAVs. Check succes stories and links on their webpage for a quick overview of what (quite a lot!) can be reasonably easily achieved.
One that hath name thou can not otter
If I were to attempt this, I'd probably just get a regular RC aircraft to start with and then rig something like this into the airframe. I'm sure there are cheaper solutions, but it would probably be one of the easiest.
Enemy of the Sun
Link to old contest stuff
Less is more.
go to the http://www.rcgroups.com/forums/index.php RC forums there is alot of info in what you want to do. and here is the forums you want http://www.rcgroups.com/uav-unmanned-aerial-vehicles-238/ Note that if you live in USA it is illegal to make UAV. Even first person view flying is illegal. But first you need to learn how to make stuff fly before you even attempt to do the UAV stuff.
Too boring, already been done: New Zealand man builds cruise missile in his garage.
You want to visit DIYDrones.com
It's a very active community that has a lot of resources for people entering the UAV scene.
tasty and delicious
I'm a just-graduated aerospace engineer from Notre Dame. For our senior design project, we build uav's... well, really RC planes. Everything had to be constructed from scratch, except for the electronics (motor/battery/GPS/receiver/etc). This year's goal was to have a mothership-daughtership configuration where the daughtership would detach mid-flight and maneuver on its own. Believe me, it's loads of fun to build everything from scratch, but it is a lot of work. And I definitely think it is doable by anyone, not just aerospace engineering majors.
Here was my team's plane: http://www.youtube.com/watch?v=eW68B3DnNWA
If you're interested in actually constructing the structure by yourself, I'd definitely suggest picked up a book on model airplane construction. Hobby shop dudes are also a big help, just go in and throw some ideas out and most hobby store owners will be very enthusiastic. And, if you're _really_ interested, I'd suggest Aircraft Design: A Conceptual Approach by Daniel Raymer. Link: http://www.aiaa.org/content.cfm?pageid=360&id=1396
Oh, also, flying a model aircraft requires a hell of a lot of skill- we get the awesome dudes down at the South Bend RC Plane Club to fly ours.
Have a look at this project: http://www.mikrokopter.de/ucwiki/en/Mikrokopter-Get-started
They offer assembly instructions and software.
Some pictures: http://gallery.mikrokopter.de/main.php
and videos: http://www.mikrokopter.de/ucwiki/VideoListe
I hope you have a few spare thousand dollars.
From your post you clearly know nothing about r/c aircraft. Learn to fly an r/c aircraft well without crashing. Go find a club and an instructor who'll teach you. Also get hold of a good simulator unless you want to spend thousands. That'll take you at least 6 months, probably closer to a year. (Longer if you don't have any aptitude for it). Flying r/c planes takes more practice and skill than you might think. It'll also cost more than you think. Once you have an appreciation for the difficulties of flying R/C you might stand half a chance programming one with a robotic interface. You'll also want to be able to take over manually from time to time when you're programming the thing so if you get something slightly wrong you've got some chance of saving it.
You could also learn about the robotics more simply with an r/c car. R/c cars can move slowly without any risk of falling out of the sky. Some of what you learn will translate to air, other parts won't.
If you want something off the shelf, I did read about robotised r/c helicopters for commercial applications like security but I think they cost in the 10's of thousands. I think you STILL need to know how to take over manually.
These posts express my own personal views, not those of my employer
The last time some of my friends tried doing an automatic control system, the plane turned straight toward the flight line and tried to kill us all!
Unless you have extensive experience designing them, I would recommend going with a kit plane for hardware rather than trying to build one from scratch out of foam boards. The reason for this is that you will start out with a design you know is flyable and has the stability properties you want. One of the classic errors in model-scale UAV design I've seen people make is trying to design the craft from scratch only to discover that their control surfaces are poorly sized, the thing is dynamically unstable, and it requires hand-made spare parts after every flight.
I think an ideal platform for a UAV like you describe would be a foam flying wing with maybe a 3-4 foot wingspan. The flying wing design would at least in theory allow you to decouple some equations which would be difficult to do in traditional fused aircraft and impossible to do in helicopters. Also, unibody construction makes it easier to land without landing gear. Landing without some pretty complex rangefinding hardware is tough, even for a computer system. Doing a skid landing on that huge wing surface with a rear-facing prop will add some margin of error to your landing sequence. If possible, get an ARF (Almost Ready to Fly) model. They come with airframe, power system, and sometimes all the servos. All you need to add is the radio equipment (I assume you are going to have a manual override backup. No, really. You're going to want a manual override.). Expanded polypropylene foam is actually more durable than a lot of people give it credit for, and replacement parts for these aircraft are easy to find.
10 Bits= $.25
100 Bits= $.50
110 Bits= $.75
1000 Bits= 1 byte
I have thought about doing a similar project for a long time, one where you can just enter GPS coordinates at get the plane to fly to those coordinates and take a picture, maybe take some weather readings as well, and send it back to a base station. A big problem that I see would be that it's hard to know how much a finished board would weigh, and how much power consumption would the instruments impose on the battery pack? Would I get an advantage from a more powerful engine from more lift, or would it just lead to power waste for the sake of a bit of extra speed? I'd also need to know that I have enough lift from the planes wings to carry the UAV circuit too or else it will be bogged down or not fly at all. The project seems to be more mech eng heavy than I'd like to take on as an elec techy.
You might try instructables.com.. They have a section with this kind of project.
It seem like the obvious approach would be a fairly large RC plane and mount a second the camera (perhaps on a servo) and a tv transmitter on it. You downlink the video to a laptop that then uses some sort of usb connection to a gutted rc controller, either with servos moving the sticks directly, or better yet, bypassing the potentiometers and variably outputting voltage directly to the control board.
It seems like the hardest thing is avoiding (auto)pilot error. I don't have any experience with RC planes, but from what I've heard you have to go into with the attitude that you're going to spend a thousand dollars for 10 seconds of entertainment. You just have to assume that the plane is going to be destroyed on its first flight. Anything after that is bonus.
I have always been interested in the same thing. The problem I have always encountered is that you would want this thing to fly on its own, to other states, territories, etc, maybe with a camera. Ideally you would be able to go to your PC, bring up an app, and see (out of the cameras on your UAV) where it is (flying over a beautiful mountain peak, etc). You would also want to be able to send to it new coordinates.
But how do you keep in communication with it? Military UAV's most definitely use satellites. Without the use of satellites, I find it hard for a UAV you build to go beyond your own visual range (= ~10 miles) from the launch site.
Does anyone have any solutions for this, or does one have to rent time/frequency sharing with a satellite provider (read:expensive)?
There were some people who built an autonomous glider which could perform many of the things you mention (with the notable exception of powered flight), including flying pre-programmed routes while taking photos (as well as navigating to specified coordinates autonomously). The process of building and testing it is documented in a fair amount of detail, including information on choices made for the on board electronics.
I have no particular interest in building aircraft, and still thought that page was a good read.
# cat
Damn, my RAM is full of llamas.
Not sure if you already have radio-controlled airplane experience. If you do not, I have a very solid recommendation for you:
A world-class starter platform for both learning to fly and lifting is the Slow Stick. It is one of the most popular planes with RC hackers, is cheap as dirt, has solid lifting potential (and upgrades can make it a real monster), and has lots of commercially available upgrade parts.
I'd go with a slow stick glider, and add a cheap brushless motor for starters (in fact, that's precisely what I have about six feet behind me for my first aerial photography platform). That will give you a good mix of cheap and solid lifting potential.
As for the forum, Slashdot is a good place to start for all things geeky, but the specialist forums you're looking for are at RCGroups:
http://www.rcgroups.com/
Here's the main starter thread for Slow Sticks:
http://www.rcgroups.com/forums/showthread.php?t=122951
Admit your noob-ness, ask for advice, be respectful, weather the occasional ornery response with good humor, and you can learn everything you want to know at RC Groups.
Stop-Prism.org: Opt Out of Surveillance
if your in the US your getting into a legal shit storm, look here:
http://www.faa.gov/aircraft/air_cert/design_approvals/uas/reg/media/frnotice_uas.pdf
and here
http://rgl.faa.gov/Regulatory_and_Guidance_Library/rgAdvisoryCircular.nsf/0/1ACFC3F689769A56862569E70077C9CC?OpenDocument&Highlight=91
other than that, it is an interesting controls project, most interesting part will be getting accurate sensor information without spending a ton on a decent gyro...
build a simulator or you will wreck a lot of airplanes before you get it working 100%
use the cell phone network for comms if your going outside ~5 miles, 900mhz radios should reach that far line of sight with a decent antenna.
Q) What countries have more relaxed UAV regulations?
A) Australia and New Zealand are famously progressive in their UAV policies. Other countries, such as Mexico, have been know to be relatively friendly, too.
http://www.rcgroups.com/forums/showthread.php?s=31fcf01ee166e7be6375a4830cd4fd5e&t=831627
-- I have a private email server in my basement.
I had intended to build one, but my working budget went from a little something to less than nothing due to job changes. I still keep my eyes open to what can be done though. Right now, it's a mental exercise.
The most important thing to remember is, as a hobby toy, unless you want to get in serious trouble with the FAA, you must follow a few rules. This is probably not all inclusive. It's just what I can think of off the top of my head from my own research. Find a local R/C group, and reference the FAA pages for more information.
The FAA has a notice on UAV's here.
1) It can not go over 400 feet.
2) You must stay out of any airspace that an aircraft may be flying in. That is, stay out of the approach and departure areas of any airport. Someone just got in trouble for this, where they had an R/C airplane with a camera that filmed a commercial airliner flying by. It was several seconds between the time the aircraft passed, and the wake turbulence knocked his R/C plane out of the air, which would imply a decent separation, but still, stay away from aircraft.
If you haven't gone through private pilot flight school, you may not be aware of the airspace restrictions. Stop by a local small airport and ask. There will always be someone with time on their hands that will love to talk to a newbie.
3) It can never leave your sight,
4) You must have control of it at all times. That is, your remote control must be able to override anything it wants to do.
6) Watch the frequencies that you're using. If you're on R/C frequencies and TX power, you're safe, but play nice with other people who may be flying. Don't hog a bunch of frequencies because you need them for additional controls. If you're working with other frequencies, check the licensing on those. You don't want to piss off the FCC too.
Now I'll go into the territory of ignoring FAA and FCC rules. Don't do it. Don't get caught doing it. Don't tell random strangers that you're doing it. Sure as hell don't post youtube videos of it, because you'll have feds in your livingroom with a no-knock warrant and a one way ticket for you to Southeastern Cuba.
For mine, I looked at a variety of options. If you search around enough, you'll find people mid-sized R/C airplanes (say about a 3' wingspan) with embedded PC's to do their dirty work. I didn't find this totally practical both from the OS standpoint and the interfaces. I want lots and lots of standard interfaces, and I want flexability to use anything I can. I intended to use a small x86 platform machine, running from a flash card (SD/MMC/CF). Delicate parts will get broken quick. Embedded or x86, you'll probably want several onboard to handle different functions. They'd need to be networked together so you can exchange data. For example, one reading your sensors, one to control the servos, one for comms, etc, etc.
I wanted to have the ability to carry at least a couple camcorders, and USB webcams. Every ounce of weight you add means you need the aircraft to support it. That means it needs a good amount of lift and thrust.
I'm assuming you've flown before. If you haven't, go to a local small airport and go for your private pilots license. That will include both ground school (the book work on how things work), and flight (actually working an aircraft). To program an aircraft for perfect conditions is one thing. Making it takeoff, fly, and land in less than ideal conditions is another. What happens if the wind picks up, and you have to slip during your landing? If you haven't programmed for it, either you'll end up way off course if tracking to a GPS coordinate, or you'll get blown way off of the field, probably into something less tha
Serious? Seriousness is well above my pay grade.