So I am starting a new school term and I need a project for my Master's class in embedded real-time software systems. I am thinking of trying to convince the doctor to let me do a traditional heli UAV. In order to do this for the cheapest price I was going to try the Trex 450 clone with the Turnigy 9X transmitter and the APM2. I know on the scale of 1- 10 hardness of a semester project this might fall in the range of 12 but I am checking it out anyways. (Cause Helicopters are cool, man)
Is there anyone who has a lot of experience building these things that can give me a hint or two on a few things. I have read the wiki posts and the discussion board. I am however looking for a few suggestions.
I am willing to help update the wiki as I go through this process if there is any moderator that wants to approve my changes. I have over 10 years of commercial software development experience so it is not the code that is scaring me, it is the hardware.
Thanks in advance.
sorry I cannot answer any of your questions, however I am curious: What will the objective of your Master Thesis be? I thought that traditional helis are already flying with the apm, are they not?
Thanks for the reply
Right now I am just in the exploritory phase. I am not sure what the final project will be focused on. and yes helis are flying with existing tech. however I do want to expand existing tech and functionality of the autopilot as it stands now. If automated throttle is not possible I would like to look at other items.
Also to clarify this is a term project not the final thesis, Although I could use and prob would use this for a thesis later.
I am worried that hardware setup time would not leave me enough time for software changes.
Thanks for the reply
Ok I will try that. I was looking though his site and he has some really cool stuff though is company. Optical flow sensors would add a significant wow factor to the project. I am a little concerned about the quads because of the price point. it seems Most of them start about 500 USD before you get to the avionics.
I think I found a quad platform to start off cheap. I may have to upgrade engines and ESC. I will email you on what ideas you have for the Centeye.
Regarding your original question, it's possible to get a Trex flying with the APM2 (or 2.5 would be better) within 4 months. The code is mostly identical to the quad code including such things as autoland.
There is a MUX on the APM (channel 8) which you can engage from your radio and which allows the remaining 7 channels from your radio to pass through the APM2 and out to the servos untouched. This doesn't really help out with heli's though because the APM2 tradheli code requires that your input be unmixed (i.e. roll, pitch, throttle and yaw must be fed into the APM2's 1,2,3 and 4 input channels). So when you engage the mux unmixed output would be passed through to the servos which will just make your servos move in weird ways that certainly won't keep it airborn for more than a couple of seconds.
When it comes to cost, once you factor in mechanical failures (because they're more complex) and repair costs (from inevitable crashes on either platform), a quad is certainly cheaper and easier than a trad heli.
We could of course really use your help on the tradheli code and you'll find that Robert Lefebvre and I are around to answer your questions and give some guidance but the choice is all yours!
A couple more things..
I use an original trex450 because I wanted to reduce the possibility of hardware issues because when i started there was no tradheli support. Robert uses a trex600 clone. I suspect a clone is ok but you should probably make sure you get decent digital servos and I believe some parts will need to be replaced with real align parts but I can't tell you which ones exactly.
By the way a flybarless head is much better at least when it comes to using the autopilot for loiter and waypoints.
Yeah... honestly, learning to build and fly a Trad Heli in 4 months... I wouldn't recommend trying. Here are a few cliches about helicopters:
1) Helicopters don't really fly, they just vibrate so badly the ground rejects them. (and remember, the APM does not like vibes!)
2) You couldn't afford to fly a helicopter even if it was free.
I think if you were skilled, and could tackle this FULL TIME for 4 months, maybe. But if it's like 2 hours a day? No way.
I have two clones, a 450 and a 600. The 600 was kinda junky, but it worked. A few of the parts are no good. Well, basically anything that is supposed to be round and straight, needs to be replaced. ;) Actually one of the big problems I find is that the metal is too soft. Any little bump and things bend.
Still, when you are learning, you crash a lot anyway, and even an Align will break. That's how I look at it.
The 450 Pro FBL is actually quite good quality. The only real problem with it is the main shaft was not round or straight. But otherwise, very impressive. Likewise, I'd suggest the HK550GT as it's newer than the 600, and I'm guessing the parts are higher quality, like the 450 Pro which is also a newer kit.
I'm currently building up a 550, but basically it's a conglomeration of aftermarket parts.
Align kits are good, but the problem is you usually have to take their electronics as a package, and I don't like that.
Thanks for all the reply's I am still looking for the project but it looks like the Traditional Heli might be too big of an ask, unless I can convince the professor of an exploratory project. I really do appreciate your input. Does anyone know if the JDrones quad kit comes with the latest APM 2.5 or do I have to buy that extra?
I would still like to help out with the code set however I would have problems doing so until I get a platform to test on.
Harry and I could use some help on getting the ArduIMU V3 to function as a simple flight controller. The Tri-copter only uses 3 motors, props, ect so it's not hugely expensive and the ArduIMU is cost effective too.
I will check this out. Thanks,
Moved the thread to Autopilot software/algorithms since I will prob not do a traditional heli