Re: [Paparazzi-devel] Paparazzi Autopilot for Linux

[Gareth Roberts]

> As part of the Rutgers University team entering the AUVSI UAS
> competition, Bradley Lord and I ported the Paparazzi airborne code to
> run on a Linux-based single board computer.
Excellent work!  I believe it's important for paparazzi to be able to
run on a multitude of platforms, especially given the shrinking size of
such devices and the increasing complexity of payloads (I sometimes fly
with a separate beagleboard anyway for monitoring sensor packages or
vision tasks).

Congratulations on your achievement, and good luck in the competition.
Gpsd integration would be fantastic if it helped remove the dependency
on hard-to-source U-Blox chips.

> To replace a Tiny v2, we used a Beagleboard and many devices...

I'd also be interested in developing some firmware to allow Tiny's to
act as simple IO boards as a transitional measure.

Finally, regarding extensions please consider using something like
XML-RPC - it's pretty mature, the slightly slower speed shouldn't matter
on an 500MHz ARM localhost and most importantly there are good quality
and easy to use implementations for almost all programming languages
(our university considers Java the only language worth teaching...).

I'll hopefully be able to try the code out in a few weeks, and I look
forward to it.

--G

On Sun, 2010-02-21 at 16:21 -0500, Pat Hickey wrote:
> Hi all,
> 
>  We needed a powerful
> onboard computer to develop our imaging system, and in order to
> consolidate hardware and make the software more extensible, we decided
> to run the autopilot on that computer as well.
> 
>  The few
> major changes had to do with PPM input and PWM output: instead of
> processing any signals from the safety R/C radio, we leave the input
> selection up to a Pololu 4-channel servo multiplexer. We used a Pololu
> Micro Maestro device for servo output from a USB device. The  Micro
> Maestro has a programmable failsafe mode if servo updates from the
> flight computer have timed out.
> 
> We haven't had to worry about real-time performance- Paparazzi is very
> reliable on the 500MHz processor, and took up only 5% of the available
> power. We trimmed out much of the code that was specific to the arm7
> and real-time responsibilities. We're also working on two extensions
> (an IMU, and integration with gpsd) that will extend Paparazzi using
> higher level languages (like C++, Python) using sockets and message
> queues. All of these changes make our platform a bit more flexible,
> and it might be a good option for a research, sensing, or imaging
> platform.
> 
> We currently use a thermopile based attitude tracking system just like
> the Tiny uses, but we're moving to an IMU soon.
> 
> I've posted our code on Github
> (http://github.com/pchickey/paparazzi-linux-release/), and I prepared
> a website describing our project that has more details about our work:
> http://moreproductive.org/autopilot/.
> 
> We owe a great deal of thanks to the Paparazzi developers and
> community for providing the excellent software we built on. We look
> forward to continuing work on this project. I'm happy to correspond
> with anyone who needs help using our code.
> 
> Pat Hickey
> 
> 
> _______________________________________________
> Paparazzi-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/paparazzi-devel