KlooGee
Second Officer
Flight distance : 16783757 ft
United States
Offline
|
AlansDronePics Posted at 3-5 05:18
I have carried out some practical experiments regarding magnetism and its effect on the drone system.
Although the video is thrown together, the content clearly shows this preoccupation with the compass is misguided.
Firstly, I tried to read through this whole thread, but it seemed like it just devolved into the normal bickering. I'm sure there was some good information hidden somewhere in there, but I can't make myself suffer through all of it to try to see what everybody was trying to say.
Okay, having said that, here are my comments on the 3 videos that you provided as proof that magnetic interference isn't an issue..
#1 - Just having magnetic interference does not cause a fly away. It can cause the drone to drop into ATTI mode which many GPS camera drone operators are not skilled or experienced in handling and thus cause a "fly away".
#2 - The main issue with magnetic interference comes from the fact that it can cause 2 different systems to have different readings. This creates a situation where the algorithms don't know which system to trust and thus stops controlling the system algorithmically and drops into ATTI mode allowing the operator to have full control over positioning.
#3 - the videos...
- The first video "Magnetic influence on Mavic Pro" only shows how magnet interference can be grossly detected by the system, but does not illuminate any details in regards to the behavior during flight.
- The second video "DJI Drone trying to pick up a magnet 1st try" seems to be completely unrelated because it never shows a source of magnetic interference anywhere close to the drone.
- The third video "Mavic AC -- Induced Magnetic Interference during Flight" is the most interesting of the 3 videos supplied. It does show some interesting scenarios, but I don't think he put together a test scenario that would help to come to any sort of conclusion.
- In the first experiment, the drone was already on and had already initialized its compass and all of its systems. This means that at this point, the compass and the IMU were on the same page and assuming that straight ahead was a given direction. Then he introduced the magnetic interference by placing a magnet near the back of the drone. Of course the drone immediately detected this sudden shift and gave immediate errors and warnings and verbally said it changed to ATTI mode. This is because (as I described above), now the compass and internal accelerometers have a disagreement on the direction the drone is facing. The accelerometers say it hasn't changed headings, but because of the magnetic interference, the compass now is showing a vastly different heading. At this point, the operator launches the drone. We can see that it is in OPTI mode and using the downward based vision sensors for its positioning algorithms and keeping it in its current location instead of actually being in ATTI mode. After the drone is up in the air away from the magnetic interference, the compass and the accelerometers now agree again and can thus switch back to GPS positioning mode. Also, in this test scenario, the drone never got higher than about 3ft, so the vision sensing system would still be able to assist with its positioning if it was choosing to ignore the GPS system. In my opinion, everything worked just like I would have expected and doesn't help to come to any sort of conclusion to say that magnetic interference is not a problem.
- In the second experiment, the operator had the drone turned on for a period of time, then taped a magnet to the drone. Once the magnet was taped to the drone, the system immediately detected the magnetic interference and gave on screen errors as well as a verbal message it was switching to ATTI mode. He talks (but doesn't actually show) the fact that the software now thinks his drone is pointed the wrong direction from where it is actually physically pointed at. As soon as he starts the motors, you can see that the system switches to OPTI mode to use the vision system for positioning. If you look closely, you can also briefly see in the DJI GO 4 interface that there are no GPS bars shown. This gives us insight into the fact that it is not using GPS positioning at all. He also mentions that it is drifting around a little bit. This is likely also an indication of using the visual optical flow for positioning instead of GPS. The problem with this test scenario is that it was still using the OPTI system for positioning. Had he gone higher than the visual sensing system could work, then it likely would have dropped into ATTI mode. Again, in this test scenario, everything worked just like expected and doesn't help us to come to any sort of conclusion to say that magnetic interference is not a problem.
Not sure anyone will take the time to read or understand that wall of text, but if you did, then thumbs up to you!
Cheers!
|
|