I’m building a fixed wing plane with onboard vision for fine-tuning the plane’s position. The main goal is to use GPS positioning to get within a small margin and then see if we can find a way for vision to make an appreciable impact on flight precision (any opinions?). I can’t seem to find much information on how I can get the plane to make automated adjustments to roll and pitch with commands sent from my companion computer. I’m using pymavlink right now and would prefer to stick with that if possible as I have a deadline coming up and want to avoid learning something new for now. I’m having trouble understanding the pymavlink documentation so apologies in advance if I’m missing something obvious.
The fixed wing will be physically interacting with a ground station, which we want to be as low-tech, movable, and inexpensive as possible (currently there is no tech on it at all). I think rtk would work great for our purposes and we have it on our plane but as I understand it we would need another rtk module on the ground station for us to know where to position the plane, so we would have to add a battery, control board, gps, etc. which we want to avoid.
That’s OK, it just sounded a little weird
Unfortunately I dont see the http://canberrauav.org.au/ link anymore, there was a lot of interesting information on the subject.
I found a helpful example here for overriding RC inputs, and I think it’s what I’m looking for. We are using an ArUco marker and a telephoto lens to detect the marker from as far away as possible, and our first iteration of the code will just modify the plane’s orientation based on where the marker is in the frame.
