Hello,
You can also have at the multi-uav navigation modules already
existing:
http://wiki.paparazziuav.org/wiki/MultiUAV
They are very basic collision avoidance based on a simplified TCAS
logic and a simple formation flight controller.
You also add a "follow"
(http://wiki.paparazziuav.org/wiki/Flight_Plans#Follow) and a
"potential" modules not really documented (3D avoidance based on
potential fields)
The major issue is that for now, all communications are relayed by
the ground. It is not only laziness, we are using two messages
class, one for downlink, one for uplink, so air-to-air
communications is not completely obvious there...
But at some point, we also have to go in this direction.
Gautier
Le 16/02/2015 18:31, Hector Garcia de
Marina a écrit :
Thanks a lot for your exhaustive response!
Those features are quite hidden in the wiki I would
say.
I already have some algorithms in 3D formation
control that I would like to implement in PPZ.
I will take a more in depth look and ask for
support/suggestion here about how to proceed.
Thanks again!
On Feb 16, 2015 5:28 PM, "David Conger"
< address@hidden>
wrote:
My
apologies, forgot to include the video URL (also you can
YouTube
search for "Paparazzi the free autopilot")
http://wiki.paparazziuav.org/wiki/GCS
Cheers,
DC
On 2/16/15, David Conger <address@hidden>
wrote:
> Hello,
>
> From at least 2005 (when I first found Paparazzi)
multi-drone control
> with a single GCS ground station has been working very
well.
>
> Here is a video from 2007 showing a drone in Germany and
another in
> France being controlled from Berlin Germany over a mobile
Internet
> link.
>
> 2007!!!! ... can you imagine that's quite a long time ago
in Drone
> years. No other open source project at the time even flew
properly
> much less in a configuration anything like this.
>
> If there is a single drone to single GCS limit is it not
Paparazzi GCS
> you are using. Possible one of the others? QGround
Control?
>
> Even more interesting info about multi-drone flights:
> Paparazzi airframe configuration tool assigns an aircraft
ID which is
> used by GCS to plot each aircraft with a different
colored trace. As
> you fly each drone will have it's own color assigned.
Full control of
> each is maintained via the same buttons and commands and
a nice
> colored trace of the flight path is seen.
> Even more interesting is the security built in via a
shared (aircraft
> and GCS) hash code. A hash is done at compile time that
assures only
> the GCS with the correct hash code will send data to the
aircraft that
> is accepted. If someone else runs Paparazzi and tries to
control your
> drones the aircraft code will reject the messages as
invalid without
> having the proper hash code. At the time even some (maybe
all) of the
> military drones did not have this security.
> Note: For the flight you see in the video they had to
share the hash
> code to the operators in Berlin.
>
> Imagine all the new features you find inside Paparazzi
today (if you
> search) if this was circa 2005 I'm talking about now. ;)
>
> It is expected few people outside Paparazzi (and many
inside) still
> are not aware of so many powerful features inside.
Everyone in a truly
> open project like this is more interested in coding and
building vs
> marketing and selling. It's a good sign I think of a
truly open
> project that you do not see so much marketing and hype to
get you to
> buy something.
>
> What I have seen is the Paparazzi foundation is very
cutting edge even
> today. Cutting edge things like Ivy, OCaml, run the risk
of lack of
> adoption. Like Beta vs VHS for video tape wars a superior
way to do it
> can fail to gain wide adoption. I encourage everyone to
learn Ivy,
> learn OCaml vs trying to pull these powerful things out
from under
> Paparazzi. They are not trivial to substitute or replace
and are IMHO
> key components.
>
> Do not be misled by Apache IVY. Here is the Ivy inside
Paparazzi:
> http://www.eei.cena.fr/products/ivy/
>
> GCS is written in this (Caml):
> http://caml.inria.fr/resources/doc/index.en.html
>
> There are so many cool, useful, features in GCS. Sure it
does not look
> like an F35 cockpit but look at the troubles complexity
brings (joking
> about F35 problems). From day 1 I have followed other
ground control
> stations and feel they are trying to be too much like an
airplane
> cockpit. Drone control should not be like airplane
control. You are
> interacting with it not flying it. GCS is what you need
how you need
> it and flexible an extensible. Please review this:
> http://wiki.paparazziuav.org/wiki/GCS
>
> I hope this reply helps you to realize you have a lot of
exciting new
> features to learn about with Paparazzi. Ones that exist
now and are
> just waiting for you to learn and use now.
>
> I hope for any newbies or those on the fence watching
Paparazzi that
> you have learned Paparazzi is the best project for real
work
> yesterday, today and tomorrow. Maybe someday the others
will catch up
> but I doubt it. Paparazzi is constantly evolving and
adding new
> features. Now if only Parrot would adopt it and mention
Paparazzi
> instead of QGround Control their product would be
perfectly aligned
> for the new FAA rules. Imagine BeBop and ARDrones doing
gas pipeline
> surveys, following miles of power lines and looking at
the top of cell
> towers. Parrot has the capacity to delivery huge volumes
quickly so
> should a major telco or power company adopt these they
could get them
> in their hands quickly. :)
>
> Cheers,
> David B Conger
>
> On 2/16/15, Hector Garcia de Marina <address@hidden>
wrote:
>> Hi guys,
>>
>> I guess this is not new here. I was wondering if
there was/is any effort
>> for integrating to the system more than one vehicle
at the same time.
>>
>> With integration to the system I mean monitoring and
communicating to
>> multiple vehicles in the GCS and communication among
vehicles without
>> passing to the GCS.
>>
>> The objective is to perform some formation control. I
guess this would be
>> interesting for certain scenarios where flying in
formation is a
>> requirement or an improvement/advantage. The
formation is not only
>> restricted to a platoon with a particular shape, but
surveillance of an
>> area following a trajectory where the distance
between vehicles has to be
>> constant or another applications.
>>
>> This year I will have time for making efforts in this
direction. But
>> since
>> I never took a look at the GCS, I do not know how
difficult will be to
>> implement a "multivehicle" feature (or if there
exists one already!).
>>
>> In addition, what is the status of PPZ with ChibiOS?
A new stable version
>> of ChibiOS is going to be released (3.0) with a lot
of important changes
>> but I guess the PPZ branch is using 2.6.x right?
>>
>> Cheers
>>
>> --
>> Héctor
>> Webpage: http://mathtronics.wordpress.com/
>>
>
>
> --
> address@hidden
> http://www.ppzuav.com
>
--
address@hidden
http://www.ppzuav.com
_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel
|