A discussion page to allow me to add my experiences using a Raspberry Pi 2/3/Zero V1.2/1.3 as a companion computer for Ardupilot/APM/Pixhawk flight boards. I will update my experiences as I learn, feel free to comment and offer tips, it's all free here.
I'm building a github for my experiments: https://github.com/benb0jangles/Companion-Pi
Drone Unit: Raspberry Pi Zero V1.3 + Pi Camera Module
img file: https://drive.google.com/open?id=0B1cVb3uX0f0dQTZzSmVISFdYd3M
Ground Unit: Raspberry Pi 2
img file: <to be added>
Goggles Unit: Raspberry Pi Zero V1.3
Ardupilot Technical Questions:
Also, if you have any technical code/hardware questions which you feel may take ongoing contributions and help over an extended timeframe, feel free to ask over on Ardupilot.org technical discussion page here: http://discuss.ardupilot.org/t/companion-pi-2-3-zero/9460
I will update as I get by, please contribute by offering your 'plug & play' wifi adapters using Raspbian Jessie. Also, if you happen to pick up this project, and are working faster than me, then please send us your own .img file experiments. Thanks.
Have you ever just considered asking the companies to send you the things for free?
I guess this is the exact description of where we stand. For the first time I feel that drone industry is technically taking over DIY.
Paul Meier said:
Yes, somehow very sorry state of affairs when the best off-the-shelf-consumer-kits are better than the state of the art DIY (that same DIY where the consumer companies copied and (give credit improved) most of the concepts from the DIY folks......
Interesting comments and viewpoints. I gotta say I really agree with both sides. I'm fed up with spending inordinate amount of time tinkering and getting frustrated with cheap crap components trying to hobble them together into a safe, reliable, working drone, or a realtime video system, or a telemetry system. When the solo and then the mavic came out I was looking at them with envy and had to glue my credit card to the inside of my wallet to stop it coming out. If I was going skiing or doing research or having fun, at this moment in time I'd take a deep breath and buy a mavic. The form factor is awesome and the features and technology are just really cool, and it's probably a lot safer and more reliable than anything I can build.
However, in the long term this makes me sad and I can't bring myself to do it. Kind of like Apple technology which is really cool but completely proprietary and closed - you can't do anything other than what they want you to do, similarly with DJI. What the drone world really needs is a Linux or Android ecosystem, a more open (yet commercially viable) platform that allows you to build on it to do what you want. Ardupilot is a great 'Linux' - the kernel of this platform, but the surrounding components are sorely lacking. What it needs is 'recipes' like this one, wifibroadcast, teensy frsky telemetry etc, recipes that people can follow easily and put together a working, cheap component of the platform. The teensy frsky thing is a good example, as there are commercial variants that anyone can just buy and immediately get going with, or you can put it together bit by bit and reprogram it, hack it, improve it, if that's your thing. Patrick Duffy's projects are another great example, they use more commercial and existing opensource technology as a recipe to come up with a good, complete, working system.
So I applaud benbojangles, and Patrick, and Patrick, and everyone else doing DIY projects here, and say keep going! What you're doing isn't necessarily ready for mainstream yet, and you don't necessarily have the perfect complete answers, but you're contributing to and advancing the knowledge and experience necessary to build an open platform. This is the hard bit, the underlying theories, components and experimentations. I'm working on my own thing that hopefully will contribute a bit to companion computers. It will come together eventually into a fun, working, open platform that will provide an alternative to the locked down world of DJI.
To infinity.. and beyond!!!
Did you notice ? this is more and more becoming a kind of "put his faith in DIY to the test".
New RTF machines are so tempting and finally so affordable... Long range HD video and telemetry ? Check. Obstacle avoidance ? Check. Long flight time ? Check. Cool and clever design ? Check.
For your immediate flight & film pleasure : order one today and enjoy tomorrow.
But most of us get also pleasure in building, learning and testing stuffs. And expect great results with cheap components. So like Fnoop and Paul, and many others here, I would say thank you guys, who publish such posts to help us to keep faith in DIY !
Thanks everyone for this little ''group therapy'' session, this is exactly what I need theses days, and please @Ben allow me to go off subject for a moment , while I will try to explain why I am in this kind of limbo.
As Pascal said we might be in a little state of shock since DJI released the Magical Mavic.
My pet project is an integration on a quad 450 of Randy's ''Balloon_Finder project'' (look for my post in this group) that I tried to implement on a RPI3 but it is lacking power so (step 2) I have upgraded to Odroid XU4 so it can to steer the vehicle using OpenCV and DRONEKIT. So far so good, now the tricky part is to use a neural network to extend the chase and send comand to seek for commons object or do visual follow-me. The simplest neural network (DeepBelief) is running too slow on the Odroid , so I bought (step 3) a Chinese micro PC running Intel Baytrail. So rebuilding and switching the whole system for a third time from 2 different flavors of ARM to ATOM takes time and lot of energy, and yesterday evening, when I was expecting the ''Grande Finale'' after endless hours of compile-crash- Make parameters change- recompile ... yes nearly there !!..POUF... black screen... the system died.... cant get it back to life....Have to arrange shipping to China for this little clown.
So YES..these days I wish I could magically transform this pile of useless silicon into a Mavic ;-)
@Fnoop,Pascal P, Some excellent points. I actually get as much enjoyment designing and building as the actual flying. It's the creative process that drives some of us, not just buying something somebody else mass produced in China. Buying a drone for us is like buying a Lego set already assembled and glued together - no fun in that.
Unfortunately, with all these new drones hitting the market, there will be thousands of novice operators with no real skin in the game, not having paid their dues in really understanding the machine, will start reeking havoc, and the end-result will be some bureaucrats and government overlords ruining our hobby. I keep waiting for some big incident to take place that will open the gate and let the dogs out. I hope it doesn't happen, but you know Murphy's Law - anything that can happen will.
@Patrick Poirier, Just remember you are trying to do the same thing as a team of engineers and millions of dollars of capitol are doing, so if you happen to pull of transforming your useless pile of silicon into something similar, what does that say about you compared to them?
The OpenCV stuff sounds interesting. I am using OpenCV at work to monitor car dashboards during EMC testing for new cars and I was thinking about this very subject a couple days ago. Sounds fun to use it on drones. Can you start a blog about this subject?
....But if you made it yourself from parts, then a ban/incident happens, you can dismantle and use for other projects - subs, boats, rovers etc. DIY is knowledge, which is more valuable than money. Even DJI Mavic is going to be leeched for ideas - folding frame joints 3d-printed, tiny pi gimbal 3d-printed parts. Open-source Digital video is almost done and will allow people to experiment more easily with captured video like all the usual robotics stuff - Opencv, Caffe, etc on a receiving laptop. Open source is just getting good, creative processes are more understood than ever, and faster to implement.
@Poirier regarding companion computers for development, you should really try contacting all the board providers with your project plans and asking for sample boards, they do actually provide. Sometimes takes a couple of responses an replies, but they do send them (Express post too).
I also like that my DIY learning experiences can crossover too, for example, i'm looking at my ebike and wondering if I can mod it (it's basically a brushless motor, esc, 2x6s lipos. I pulled the bike from a skip). Maybe even have opencv running on a front/rear camera when i'm on the road.
Don't fear the mavic :) they have an army of people all over the internet gleaming for what people want. They also have teams of peeps who are paid to discredit rivals and ideas, and paid teams to promote. That stuff doesn't come cheap, and I laugh at it knowing it's easy to detect when you know it's being done. I don't mind.
and the sad part of it: they (governement / legislators) will outlaw all the diy drones and make sure that only drones from the like of Parrot and DJI are allowed...
not going to happen. You can't stop diy :)
P.S is anyone going to this in London next week? http://www.terrapinn.com/exhibition/the-commercial-uav-show/index.stm
They provide free expo passes, and i'm going (anonymously) for fun
@Ben: I don´t fear the Mavic, I wish I could actually build something with same specs for similar money