M2P activetrack crash into tree

[Altitude000]

So in a very similar situation to the thread posted by Chuampis yesterday, I was doing an activetrack orbit (with me as the subject, so wasn't watching the controller, that'll teach me). I was doing some tracks where it would be fine for a while then lose track of me in an odd way despite ideal conditions with the subject wearing dark clothing against a white background (frozen water).

On the crash flight, it started and from looking at the flight log on a map the orbit it was going in for that flight seemed much wider than I'd expect which meant it ended up going over the only small piece of land jutting into the water and therefore managed to fly into the only place where it could hit something and of course it did. Because of the camera angle I can't tell if it hit a tree or a sign that was there, but it sustained some proper damage in a 45 degree bent front left leg and the gimbal is out of alignment.

It'll need to go back to DJI for repair either way, but I'm curious what might have happened. Certainly I could have been further away from the tiny point where it could hit something, but the orbit should have put it well out of range. Also I feel the proximity sensors should have been able to cope with that. But having said that, there have been a couple of occurrences recently where it detected an "obstacle" whilst right up in the air nowhere near anything at all so I'm not sure if something was already wonky.

Thanks for any insight that can be provided.

 

[Deleted member 103366]

Please someone correct me if I am wrong but in an orbit doesn't the drone fly sideways in a circle with the nose pointed at the subject ( I have only done one orbit with an M2P and this was how it was flown)? If so, in such a flight would you not be 'relying' on the sideways looking OA sensors? I have the impression they are only active in tripod mode, am I mistaken? (I haven't delved into the manual concerning this).
I suppose it might be possible the turn the M2P camera so that it was looking sideways relative to the drone, which might mean that forwards looking or rearwards looking sensors were available, but I haven't tried that.

Is from 8 minutes onward in

the situation you encountered?

 

[slup]

So in a very similar situation to the thread posted by Chuampis...

On the crash flight... the orbit it was going in for that flight seemed much wider than I'd expect which meant it ended up going over the only small piece of land jutting into the water and therefore managed to fly into the only place where it could hit something and of course it did....

... I feel the proximity sensors should have been able to cope with that.

No, your mishaps is not similar to to the other case you refer to ... you misjudged the path for the auto mission you launched, & you had to high expectations on the lateral OD system.

As seen here where the red path symbolize the flight & the green bar the AC with the bar pointing into the heading direction ... you took out those trees full side on.




In the chart below I have plotted all 3 gyro axis to mark the crash ... where the markers are placed the OD sensors actually starts to see something & brake (light brown field), followed by the green field where the AC tries to avoid the obstacles. But note that this happens in the middle of the crash (seen by the gyro graphs). Just before the AC starts to brake the speed is 7,7m/s ... actually something hits the AC already a bit earlier, already 585sec as seen from the gyro graphs.




This is what the manual says about the lateral sensors ...



Sorry but this was only a pilot error ... nothing more.

 

[Altitude000]

Sorry but this was only a pilot error ... nothing more

I can accept that, hell I kinda expected that. I almost never use the automated modes such as activetrack so wasn't aware of the difference in how the sensors work with that mode - though to be honest I was intentionally out there so I could be far away from anything that it could hit. Well, had to happen sometime, it's my first major incident and got lucky at least that it happened in a way that I could easily retrieve the drone and no one else is impacted.

Tack så mycket for your time and indepth reply. Now I need to figure out the Swedish postal/courier system on how to get it sent to NL without me paying silly money!

Is from 8 minutes onward in
the situation you encountered?

Yeah pretty much. You were right about the orientation and because I've only ever really flown manually I haven't relied on those sensors before and wasn't aware of the limitations. In that vid you linked at about 10 mins he does have a similar thing where the camera seems to continue to track but the orbit widens quite considerably - though in his case a lot wider than mine was.