We are using the APM 2.6 in 4m wingspan motorised gliders, some fitted with an electric motor and others fitted with a gasoline powered engine. We are trying to use the aircraft, flying at 120 m Above Ground Level to perform a raster scan to enable us to photograph the underlying terrain.
Our results to date are not good:
[-] even with a scan line separation of 145m, the plane performs an even wider turn at the end of each scan line (Waypoint Radius = 8m)
[-] the straight line sections are not that straight, deviating from the pre-set path by more than 50 m
So, I would like to suggest a Fixed Wing Raster Scan Challenge to anyone who thinks they can perform a raster scan under autopilot control.
Define and fly a raster scan pattern:
[-] 800 m scan line length, with an additional 100 m at each end for turning
[-] 100 m scan line separation
[-] 8 scan lines
[-] 120 m flight path above ground level
On completion, please post the:
[-] type of fixed wing aircraft used (no multi-rotors or helicopters)
[-] the take off weight and wingspan of the aircraft used
[-] the kmz format file for the pre-defined flight path
[-] the kmz format file for the actual flight path
[-] a copy of the parameter list used in the flight
[-] the wind speed and the wind direction at the time of the flight
The "Top Dog" will have an actual path that is closest to the planned flight path in the central 600m of the raster scan. I hope to periodically suggest who the Top Dog in this area currently is.
I hope many people find the time to try their skills. It's not that easy! Give it a go, Andrew!
Best Regards,
Joe.
Replies
very impressive
This is the best raster scan I have ever seen
I agree, it is a very good series of passes which also exhibit 'essing' in the X,Y plane (Z being altitude).
For an area of that size, a laser guided system would be impractical. The work that is being done with ground sourced, RF positional references might raise the bar on the accuracy. I do think that the aircraft will reach a plateau of flight stability that Gary M predicts but the post processing may include flight data from the aircraft that the image processing software uses to fix the image positional issues.
A 'system engineering' approach would combine all the elements to produce the product.
-=Doug
THIS blog/build/series starts with sketches and works through many of the issues with mapping.
Perhaps it could be used to compare your issues with the author's?
-=Doug
We might consider for the next T3 round but as Andke rather suggests this is a little to simple especially with the grid tools in MP available. Sounds like a little more tuning is required on the airframe.
I'm sure if you put up some prize money somebody will come and tune your airframe for you.
http://www.barnardmicrosystems.com/L2_unmanned_air_systems.htm
ok, so you have problems and want a quick-solution, make others do the research for you?
All I can say this is not a problem with a well tuned APM.
I've flown perfect mapping missions in wind so strong that the nose were almost 74degrees to the flight direction, and still on path, with completely assembled results.
There is no shortcuts to proper configuration, it depends on the airframe and servo/control surface deflection too.
So ... the challenge is on you :)