update on my aerial robotics paltform.

[dunk]

so i've been making some good progress with this recently.
i have a flying microcontroller platform that i can radio control from the ground.
it's still early versions of the firmware but it's showing a lot of promise.

the next step is to get 2 way communication working so i can monitor flight data from the ground.

building instructions here: http://sites.google.com/site/mrdunk/Home


dunk.


[edit] spelling typo.

[airman00]

nice work

what made you choose a remote control airplane over some sort of helicopter/tricopter/quadracopter

[brijesh]

Wonder if you have come across this site

http://diydrones.com/

[dunk]

what made you choose a remote control airplane over some sort of helicopter/tricopter/quadracopter

i admit it would be a cool technical challenge but i don't have any real interest in building a helicopter at this time.
range is very limited, the platform is unstable.
with a fixed wing aircraft i can have a range of kilometers rather than tens of meters.

even if i was thinking of building a UAV helicopter i'd still build a fixed wing aircraft first. many of the lessons learned are the same.
the electronic hardware i am building will work on either helicopter or airplane platforms.

Wonder if you have come across this site

yup.
diydrones.com is an interesting project using off the shelf RC equipment and as such has some design restrictions imposed on it by the off the shelf components.

i'm building radio control, processing and the ability to read sensor input into the same board.
it should be considerably more compact.

realistically i won't get as far as that any time soon though as i'll be setting off traveling for a year in 3 months time so won't be working on it for long.


dunk.

[Admin]

realistically i won't get as far as that any time soon though as i'll be setting off traveling for a year in 3 months time so won't be working on it for long.

be like admin and travel while making robots

As I type this I'm taking a vacation in Singapore to get away from my vacation in Thailand - and still working on robots

I spend most of my time planning and designing and programming however, as lack of tools/equipment makes actual building difficult . . .

[yankeefan238]

Nice job!

Your plane looks really good and your camera skills are better than you think lmao...
 

[dunk]

hmm. i've been neglecting you guys.
time for an update.

so i ran into problems with the Unigen RF modules i was using.
there were several undocumented aspects of the modules in the Datasheet, one of which lead to horrible transmission errors when the system was exposed to any sort of 2.4GHz interference. (of which there is a lot anywhere with wifi...)

2 months later i managed to work out the problem.
after chasing lots of red herrings, playing with Cypress CYRF6936 registers,
it turns out the module will only receive correctly is one of it's undocumented pins (PACTLn) is forced low.

so anyway, i now have a nice stable RC system that can relay information back from the receiver.

there has been a reasonable amount of interest in this on RCgroups.com and i am currently putting the finishing touches to firmware that will allow it to be used as just a RC system that reports battery life and signal strength at the receiver back to the transmitter.
additional sensors (airspeed, altitude, GPS , etc) will be added later and that data will also be transmitted back to the controller.

i also need to provide an option to allow the controller to be operated using a laptop instead of a PS2 controller....

anyway, building instructions here:
http://sites.google.com/site/mrdunk/


dunk.

[dunk]

cross-posted here: http://www.rcgroups.com/forums/showthread.php?t=971053&page=10

hey people,
so a lot of people have asked me about the range of these Unigen RF modules
although i can't seem to find any of the threads with the question in just now...

it's sort of relevant here though.

so i did a range test today on foot.
i sat in the park with my laptop connected to the receiver while my friend Ryan walked out across the park carrying the receiver.

the ground was flat and level.
the TX sat on the ground with a 7dBi antenna pointing straight up.
the RX was in Ryan's hand and had a 25mm wire antenna.

frames are sent every 20ms from the TX to the RX.
the RX replies every 20ms with the number of frames successfully received.
the TX records the number of frames received from the RX.

the TX outputs both these sets of statistics (along with signal strength data) to the laptop connected to the TX.

this graph shows the resulting transmission statistics:


i speculate this will be usable on a model aircraft out past 500meters and a ground based vehicle out to around 800meters. (ground vehicles by their nature being less inclined to hit planet sized obstacles when reception is lost....)

obviously the reception will probably improve on board a model aircraft as one of the antenna will be higher above the ground but this is as yet untested.


dunk.

[Admin]

What are the units on the Y-axis for 'signal at Rx/Tx'? dB?

And I'm amused that the 'very approximate meters' has an accuracy of 0.25m

[dunk]

o stop being so fussy.....
i guess i should have put the legend to the left of the graph instead of the right.


the Y-axis scale:
for the red and green traces are the number of successful frames per second (out of the 50 transmitted).
so on the top left of the graph where the red and green lines are up at 50, 100% of transmitted frames are being received.

the yellow and blue lines are the raw RSSI (Received Signal Strength Indicator) output of the CYRF6936. (maximum theoretically possible value is 31. background interference is around 6.)
in theory a multiplier could be applied to this figure to give a dB/m output but i don't have a way of accurately calibrating it.


the reason the horizontal scale (in approximate meters) is inaccurate is i forgot to attach an encoder to Ryan's legs.
even if i did there would always be some slippage as he's not very stable in the mud. i must replace him with a robot in future tests....
also the GPS utility he was using on his phone didn't record time information so we only know the time of the start and and end points accurately.

as for the millimeter precision on the scale, that's just how the software works. (using Google docs spreadsheet. functional but not much formatting control.)
i'm just trying to wow you all here with a pretty picture without actually spending hours formatting it.
switch off your engineering module for a moment and appreciate the Art.


dunk.

[dunk]

hey guys,
time to poke this sleeping thread...

so i've been making good progress with this project recently. (being unemployed for a few months earlier this year helped...)
the bi-directional radio link is working very well.
i have finally taught myself to fly RC aircraft so crashes are less frequent.

over the last month i have been working on the IMU (Inertial Measurement Unit) so i can eventually achieve my end goal of fully autonomous flight.
progress is good:

My ongoing self built UAV project.

the next step will be to add GPS to the aircraft so i can give GPS locations for it to fly to.


dunk.

[Admin]

dunk . . . not sure if you've already thought of this already or not, but my experience with our space balloon has taught me these really important things for bots that fly:

1) make sure your vehicle will still work (or at least be recoverable) if GPS fails
- either a backup GPS, and/or some other backup transmitter

2) in testing, intentionally cause each of these to fail and see how your vehicle reacts
- at low battery, mcu's act funny . . . plug in a low battery and see what happens
- unplug your gps and see what happens
- unplug any other important sensors and see what happens

3) in code, set up a warning system to let you know something isn't working before flight
- for example, don't let the code to allow an initial launch if less than 4 satellites are locked
- set up code that checks your pre-flight GPS way-points for being outside of bounds by accident (ie sending it to China)