We tried some formation flying with 2 different APM equipped planes.
Both were loaded with AP v2.76 and the same exact mission except for the chase plane flying 20m higher than the target plane. Both planes had cameras in, a Mobius HD Cam and 1.3GHz video link in the chase plane.
In the video I first flew in FBW-A mode then in full AUTO.
We found that the distance between waypoints was too close at 600m as there was not enough time to catch up before the next turn when a plane fell behind or pulled ahead. Due to the variability of the airframes, getting the airspeed synchronized was troublesome as was their turnrate.
We did manage to get some air to air video but our fun was cut short by Phil's plane stalling on the downwind turn with Ardupilot ramping up the throttle too slowly and not being able to recover from the resultant spin, the crash is in the video.
Will try again soon.
Comments
Wow, im starting to surf on this topic of swarming and its awesome. For instance i see that it would be very very useful in geographic survey, if you can swarm to aircraft yo can adquire the terrain data far more quickly that with one aircrat.
Of all that i am reading i would propose my workflow as follow.
1.- Documentation first.
2.- Simulation with the hil firmware version. First on multicopters then on fixex wing aircraft
3.- Flight testing. Same first on multicopters then on fixed wing aircraft.
Im sure that if you can integrate a rtk GPS on a aircraft (both rotary and fixed wing) the precision will increase drastically.
By curiosity what companies offer aircraft swarming management as to a final consumer?
I think the answer to this lies in Mission Planner.I would rather rely on the computing power of the ground station using and controlling the positional reporting of the APM/Pixhawk.Members of the flying formation would VPN their notebooks with the lead aircraft ground station /notebook by 3G or Wifi or even with a hardwire ethernet connection. Not to different to kids playing lan games.
The Mission Planner of the lead aircraft would have a template of the proposed mission with every members position in the formation and function would be predefined The lead aircraft would take off and muster all the members to take up their positions and execute the mission.Sould the leader fail the number 2 would take over. should any plane fail the whole formation would simply move one place up.You would use the Auto / Manual mode function to join or leave the formation.
This could put another spin on our hobby. Just emagine 40 identical 1.8m planes flying in formation.I hope our software developers read this.Maybe Chris can get some sort of sponsorship from a transmitter manufacturer for our software boys.We know how they like to brag about integrity of the equipment they want us to buy. My bias is fixed wing but I can see this working in planes quads and surface vehicles.
If they can use 30 Quads do the Star Trek theme in London shurely we can do the same or more.As we know the word drone is very big today world wide and every one wants a part the rub.Corporates will pay top dollar for doing their logo at a football or soccer stadium, or for flying in formation at a field event with their livery and or banners.There might be issues in some parts of the world with drones in civilian use. However corporate sponsorship can cut through a lot of red tape get permissions and buy the insurance.
I would love to have a follow me feature for APM/pixhawk based craft (though I suspect the processing will mean the later is best suited).
Re: formation flying - conventional formation flying is considerably easier I suspect, because 1:1 craft have far better penetration, are more stable, less prone to needing to turn (they don't have to worry about FPV range!), more likely to maintain a steady course and probably most importantly - the human's brain in the cockpit has an easier job. When we see occulus rift-style HD goggles with non-fisheye 170' visibility in HD and more intuiative head tracking will we see more natural formation flying at the RC scale.
Ryan probably forgot to mention that approaching at 0' bearing means you're in the dirty air.
Excellent video thanks for sharing reminds me of the pre programmed quads flying the" Bat" formation over London
Try closing on the plane on the 30-45 degree bearing line. Flying form is really hard from the 6 o'lock position. If you close on the 45 deg bearing line you can more accurately judge your closer speed. Never just drive straight towards the plane. Always work on getting on the bearing line and then use throttle to work your way closer on the bearing line. If you get off of the bearing line get back on using roll before changing throttle.
-Beall
@Graham: Instructions for swarming/formation flying with the Mission Planner are in the manual here.
We will test the theory as soon I can put a plane back in the sky.Will leave the how bit to Graham.
@Graham: Theoretically, yes, but I think it's only been tested on copters.
@Chris, I didn't realize the swarm function existed, would it work for aircraft?
I did consider 'follow me' but putting an Android device plus extra telemetry radio into the plane seemed like a lot to achieve something we could achieve fairly easily without it, we also only had two telemetry radio's and wanted to keep in touch with both planes, plus we were actually interested in how the planes would fly together with independent autopilots.
What would be better was if the two groundstations could talk and have one plane follow the other that way. Or even have one ground station control two planes...
On that note, I haven't tried it yet but one can run two instances of the Mission Planner on one laptop with two radio's using different ports but the two MP's couldn't talk to each other without skills that I don't have.
Anyway I'm sure the planes could be tuned to fly very closely in formation with a little bit of effort.
-
1
-
2
of 2 Next