Re: [Paparazzi-devel] Extending Paparazzi

[Vincent Breider]

Hello Hector,

Thank you for this opportunity. Though to avoid any misunderstandings the 14th is a Sunday on my calendar. Did you mean the 15th?

I have been told the airframe is designed for high speeds, and is naturally unstable. As I am a software engineering student, i could only imagine that it would be like trying to control a fixedwing aircraft as unstable as a rotorcraft and gains increased stability as its speed increases.

Though it would be nice for us to gain advice from someone who has experience with paparazzi.

Sincerely,

Vincent

On Fri, Apr 5, 2013 at 11:58 AM, Hector Garcia de Marina <address@hidden> wrote:

Hi Vincent,

in my opinion you have to specify your requirements with more detail. The airspeed of 238m/s for a nice stable vehicle probably does not require a "ultra fast controller", and just an innocent 

"very unstable because its mechanical configuration" quad-rotor probably does.

As you are in Delft, I will be in Utrecht Monday 14th from, let me know if you want to discuss further in person at lunch time.

Best regards,

Héctor

On Fri, Apr 5, 2013 at 11:50 AM, Vincent Breider <address@hidden> wrote:

Dear Paparazzi Community,

Please allow me to introduce myself.
My name is Vincent Breider I am a student from The Hague University of Applied Sciences in Delft, The Netherlands.

For an internship I have to research the possibilities of extending the Paparazzi Software.

The main goal is to create a testing platform for an adaptive control system that is capable
of controlling highspeed anhydral delta airframes.

I managed to contact Felix who then pointed out the questions I have could better be asked in the mailing list so the rest of the community can also help answer them.

So here is the mailing conversation between me and Felix so far. I like to invite the rest of you to elaborate his answers as you see fit.

These are the questions from my email to Felix:

Is Paparazzi scalable to control unstable anhydral delta airframes at speeds of approx. 238 m/s?

How did the Paparazzi Design Philosophy originate? And what should I keep in mind when extending, changing or implementing a Paparazzi module or subsystem? And how does this philosophy differ from using Object Oriented programming and C++ as a language?

Is Paparazzi a hard or soft real-time system? And where in the source code can I find routines that will be executed if a deadline is not met?

The tasks are executed at a certain frequency/period is there any known limit to which they could be increased. i.e. the sensor task of 60Hz to 512Hz?

Finally I will be visiting the Dortmund Intermodellbau 2013. Will the Paparazzi Team be attending so I could chat with some of the team members in person?

And here is his reply:

Hi Vincent,

first a general remark: We (or I) prefer if questions like this are asked on the mailing list, so others can help out as well and benefit from the answers.
Also I'm a bit short on time, so it is usually less work for me ;-)

A few short answers:

    Is Paparazzi scalable to control unstable anhydral delta airframes at speeds of approx. 238 m/s?

That depends on which sensors, estimation algorithms and controllers you use. So in principle, yes... but I'm pretty sure it hasn't been done/tested so far.
You should also make sure that

    How did the Paparazzi Design Philosophy originate? And what should I keep in mind when extending, changing or implementing a Paparazzi module or subsystem? And how does this philosophy differ from using Object Oriented programming and C++ as a language?

This is a very broad question that I can't answer properly due to lack of time right now...

    Is Paparazzi a hard or soft real-time system? And where in the source code can I find routines that will be executed if a deadline is not met?

There is no real operating system or real scheduler in Paparazzi. It is all "bare metal"... Separate tasks in paparazzi are called in a cooperative manner by the very simplistic scheduler (no preemption).
There is interest in integrating ChibiOS though...

    The tasks are executed at a certain frequency/period is there any known limit to which they could be increased. i.e. the sensor task of 60Hz to 512Hz?

 That depends on your main PERIODIC_FREQUENCY (can't be higher than that) and of course mainly on how much computation your task needs (it mustn't delay other tasks since we don't have preemption).

    Finally I will be visiting the Dortmund Intermodellbau 2013. Will the Paparazzi Team be attending so I could chat with some of the team members in person?

Not that I know of.

Cheers, Felix

 I like to thank all readers in advance for their time.

Sincerly,
Vincent Breider

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

--
Héctor

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel