university project

[davidw]

i'm an engineering science undergraduate at a university in the UK, and i'm considering making a basic UAV as my final year project next year. i have been flying model planes for about 10 years, but i have no experience in the autopilot side of things.

i'm not trying to make a high performance UAV, or anything that hasn't been done before. it's the control techniques required, and their physical application of these that i'm interested in. basically i just need a fairly simple and doable project that's fun but also impressive! :)

my plan is this:

1. firstly i will create a small cheap electric powered UAV. a possibility is the foam tornado ( http://www.rcgroups.com/links/index.php?id=3940 ), though i have no idea if this will be able to carry enough weight. it's advantages are it is very cheap and robust. i will have the controls set up as follows:

- the elevator will be used to maintain a constant airspeed.
- the ailerons will used to maintain a desired bank angle.

this will obviously require 2 sensors on board (airspeed and bank angle), and the appropriate controllers that i will create myself. i will have a 2ch rx on board that control the desired bank angle (input to one of the controllers) and the motor speed. if i can do this then i should be able to then fly the UAV via my tx, controlling the motor for height and the bank angle to turn.

2. IF all this goes to plan and there is sufficient time remaining, i will add a GPS unit and another (higher) level of control. this will take over from the rx as the input to the bank angle controller and motor. i will program waypoints, and the new controller will from this give inputs to the bank angle controller and motor needed to get there. the penultimate waypoint in the flight plan will have a flag that disable the motor completely, then the final waypoint will be far away so that it will (hopefully!) glide down until it lands.


the idea is that it's the concepts that are important. the whole thing could obviously be scaled up to a high performance long range plane with a camera or any other sensor/device fitted, but that's NOT the idea of this project :)

as a side note, i am sponsored by QinetiQ (the recently privitised UK MoD). if i intend to go ahead with this project, i hope to speak to the relavant ppl there about current development in UAV's to see if there is anything specific i might try to achieve. maybe also they can give me a GPS unit!

any suggestions/advice/ideas would be very welcome please.

- David

[moon-dawg]

You are going to need a pretty large plane if you go electric and you want to carry any weight/batteries for endurance or driving your controller and sensors.

The SoLong project might be a good inspiration for you: www.acpropulsion.com

Here is an application note on implementing an inexpensive airspeed sensor using a PIC and a MEMS accelerometer:

http://www.analog.com/UploadedFiles/Application_Notes/49905351395455AN-596_a.pdf

Using a ADXRS Peizo Gyro will also give you an in expensive roll rate sensor.

[treehog]

http://www.rcgroups.com/forums/showthread.php?t=423667
In this competition it is only micro uavs


I dont think much for tornado ability to carry weight

suggest saucer depron indoor large wing area slow flight

personaly figure simulation is the way forward if you want to join in a simulato uav project i will probably start a thread soon

simulator crashes and equipment super cheap :D

Ralf

[kd7ost]

http://www.rcgroups.com/forums/showthread.php?t=423667
In this competition it is only micro uavs


I dont think much for tornado ability to carry weight

suggest saucer depron indoor large wing area slow flight

personaly figure simulation is the way forward if you want to join in a simulato uav project i will probably start a thread soon

simulator crashes and equipment super cheap :D

Ralf

Simulators leave a lot out. It's easy to develop technology to make micro simulations work that don't translate to the real world without SMM technology. Simulations are cool no doubt, but the "Real thing" is the only thing that is "Real".

Dan

[treehog]

I started this thread
http://www.rcgroups.com/forums/showthread.php?t=457281#post4792505

I first used a simulator called visio aero which could predict the flight charicteristics of a model on 486 computer in1996

From this i could remove my grossly over powered 1 metre wing span self dezined prototype with s600 and 10*500ars and get the suitable s400 geared to do the job better

Then getting the formulas from the local libary able to fine tune the visio aero to predict more correctly and texas intruments 500mbits memory 199x could do engine predictions before motor calc arrived

Using this proccess ensured I could dezine modify succesfully other craft to suit my needs better and I could try multipile solutions before commiting to buy the equipment

So putting the go faster brushless stickers wont fool me

Prediction programming can often prevent preventable events such as software conflicts in extreme manovers that freeze up the programs or responses

For mini uavs the arms seems to be the most popular and the small planes tend to bounce when the plot goes wrong but
My preference is to start bigger some 3 metres as I have a real reason to get that sized plane back to base in certian events
and so dont want to go down the try small arms rudder only build up system if I can jump that gap faster

Ralf

[hugo_vincent]

- the elevator will be used to maintain a constant airspeed.
- the ailerons will used to maintain a desired bank angle.
- David

Quick note: make sure you consider what will happen if/when the engine stops. For example, some people have used elevator to maintain altitude, but when their engine stops, the airplane keeps trying to pitch up until it stalls. You should be alright to use pitch to control airspeed, and throttle to control altitude, and just use a pitch damper (don't try and maintain any particular pitch, just whatever pitch you need to maintain airspeed).

Hugo.

[Vindication]

You shouldn't need that large of a UAV to go electric. We use 48" flying wings. I don't however recommend a flying wing for your project as you need a fairly advanced autopilot for them...

[tworley]

I mounted a Crossbow (www.xbow.com) unav and stargate board (Expensive!!!) to the bottom of a GWS slow stick ($34.00 USD). It few just fine on the stock motor and a two cell li-poly ($40.00 USD). The plane weighed in at 19 oz.

I logged data over a CF card 802.11b connection. The code is open source if you would like the link. If you need ideas for the control board design, I can offer up suggestions.

Tim Worley
www.trekaero.com

[hugo_vincent]

I mounted a Crossbow (www.xbow.com) unav and stargate board (Expensive!!!) to the bottom of a GWS slow stick ($34.00 USD). It few just fine on the stock motor and a two cell li-poly ($40.00 USD). The plane weighed in at 19 oz.

I logged data over a CF card 802.11b connection. The code is open source if you would like the link. If you need ideas for the control board design, I can offer up suggestions.

Tim Worley
www.trekaero.com

I would love to see the source code!

[Unterhausen]

I would love to see the source code!Try this:
http://sourceforge.net/projects/micronav

[slopedog]

any more info on it?

[Iraqigeek]

slopedog,

From my experience, you can easily get a mile of range if you use a high gain antenna (~12-15db) on the ground. I used to install long range wifi links and we could get over 3 miles (5km) operating at full speed (11mbps, we used 802.11b) by using a pair of 24db antennae with no amplifiers involved. Of course, we had those numbers when we had clear line of sight.