|
| From: | Michaël Bos |
| Subject: | Re: [Paparazzi-devel] Multiple UAVs |
| Date: | Sat, 6 Feb 2016 11:58:59 +0100 |
Hi,
Thanks for your answers. This is very interesting and it give me a first insight of the work to do. The flying robot commander is a nice idea.
The aim of the project is providing a system for managing a uav fleet with few operators (without compiling knowledge!), something like the 100 drone record of Ars Futura. Would paparazzi be able to do that?
Also, I'm considering building a team in Paris area in the next months , do you know of anyone who would enjoy this kind of project around here?
Thanks again!
Michael
Hi,
Some additions to what Felix already said:
- We know that a big limitation to have multiple UAVs doing collision avoidance on their own is that the traffic informations are currently relayed from the ground (and tcas resolutions as well actually). Some people are working on workarounds to allow air-to-air communication, but a new version of the message protocol would help for this (and we have been thinking about this for a long time now...)
- The GCS (and a single operator with enough experience) can currently handle three aircrafts easily, and up to 5 or 6 if they are flying well and the flight plans don't need to much attention. More aircraft is probably a pain because of the number of interactions required (no group control), the performances of the whole system going down, the big mess of trajectories unless you reduce the tracks length...
- Hooperfly started an "flight robot commander" (https://github.com/paparazzi/flyingrobotcommander) to ease the control of multiple UAVs. You can find a few youtube videos as well.
- 2 years ago, I gave a student project for my student where they tried to handle groups of UAVs and send common commands. It sure is not that easy to do properly...
Beside that, I really believe handling multiple UAVs as easily as a single one is an important goal to achieve for us.
Can you tell us a bit more about the kind of applications your aiming for ?
Gautier
Le 03/02/2016 21:51, Felix Ruess a écrit :
Hi Michael,
I tested differents open source ground station and it appears none is really advanced. Paparazzi core is multi uav oriented, which makes it a great candidate. I read the limit is about 255.
Yes, that would be the (theoretical) limit of how many aircrafts you can control via the same network/groundstation as the aircrafts are identified by an 8bit ID.
1. what are the other contraints? (radio, ground management, take off, landing)
I think one of the biggest constraints (apart from actually managing all those aircrafts, charging, launching, safety pilot, etc) is the communication.
E.g. with most wireless modems you will probably have bandwidth and connection issues at some point...
2. anti collision : the module currently existing based on tcas computed on the ground is limited. Is there any work in progress about it? (maybe integrating last works of MIT about drones finding a path in a forest for exemple, if possible )
Hehe, that would be nice (although stuff like from MIT won't be able to run on a small microprocessor any more).
While integrating more advanced collision avoidance is of course possible, I'm currently not aware of any work in that direction.
IMHO this is also better done on a more capable separate computer (can still be onboard), which then interfaces with the regular autopilot.
3. hmi : the ground station is quite geek oriented. What are the option to make it more user friendly, so multiple operators could manage uav groups ?
What exactly do you have in mind?
I know that it's not the nicest looking GCS out there, but IMHO it's not bad functionality wise and lots of people start appreciating this when using it in the field.
A lot of people (including myself) would really appreciate a rewrite of the GCS, but AFAIK no one has taken up the task so far (it's not small). It works well enough for most of the core devs and the main point is probably lacking time and other stuff higher on the priority list. Personally I also suck at UI ;-)
Regarding multiple operators: what you can already do is to have multiple GCSs running, where each operator can control a subset of the aircrafts on his GCS.
You can set the aircrafts to manage via the `-ac_ids` option.
4. multiple uav management : I like the inflight updatable waypoints, is it possible for multiple uav to use the same wp ? (1 wp moved, n uav updated?)
That is a very interesting idea!
Something like that is currently not possible out of the box, but I think it wouldn't be too hard to script something like that in python (if you know that the aircrafts share waypoints).
But having a nice interface for this in the GCS is quite a bit more work... How would you specify which waypoints are common? How should they be displayed? etc...
Hope that helps...
Cheers, Felix
_______________________________________________ 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
| [Prev in Thread] | Current Thread | [Next in Thread] |