Mavic Air 2 - Changes in RTH Behaviour

[Deleted member 94047]

While going through the User Manual for the MA2 today, these changes (or modifications) to RTH behaviour caught my attention

1. The time it takes for Failsafe RTH to kick in after RC signal is lost is now bumped up to 11 seconds. The gap between RC signal loss and RTH activation is a mere 2 seconds for the Mavic 2 and 3 seconds for the Mavic Air and the Mavic Pro.

2. Retrace procedure upon Failsafe RTH activation has also changed in the Mavic Air 2. The Mavic 2 and the Mavic Air retraced back their path for 60 seconds when RC signal is lost before activating RTH. If signal is re-acquired inside the 60 second retrace window, they stop and hover in place and wait for pilot command for 10 seconds. If link with the RC is not re-established inside the 60 second retrace window or if no command is received from pilot within 10 seconds of re-acquiring signal, they then proceed with RTH. The Mavic Air 2 does it a little differently. Upon RC signal loss, it retraces the last 50 meters of its flight path, flying backwards, and initiates RTH if signal is not re-acquired. There is no indication in the user manual as to what happens when signal is re-acquired within the re-trace window.

3. There is now a totally new sub-mode of RTH which the manual describes as "Power saving RTH". The manual is a bit difficult to understand here but it looks like this sub-mode is automatically triggered when the AC presumes it is both too far AND too high from the Home Point to reach it in time using the regular RTH path. So it calculates a best path taking into account distance as well height and descends gradually as it travels back to the home point.



So my question is - what do you make of these changes? And is DJI likely to introduce these modifications/changes/improvements to how the RTH procedure works into other models like the MA1 and Mavic 2s.

 

[Prismatic]

While some details remain sketchy, overall these seem like improvements.

• Allowing 11 seconds before initiating action on loss-of-connection gives the pilot a chance to, for example, reorient the RC antenna if that's the problem.
• I'm less certain of the 50m retrace threshold (vs. a timed threshold), but in most cases it's probably a wash.
• In open terrain, the Power-Saving RTH is a clear winner, but ... what if that "glide path" intersecting a hill? Presumably it's in an Auto-APAS mode to forestall that sort of crash. Even then, a power line could cause trouble I suppose. Still, it's a pretty cool idea, shaving up to 20% off the return travel distance by following the hypotenuse of a right-triangle, as illustrated.

I guess I'd be a little surprised to see RTH behavior change on existing models, but as it's just a firmware change, it could be done. Still, the M2 firmware can auto-correct for a badly initialized compass--which is a truly major, drone-saving improvement--yet that firmware change has not percolated into previous models.

 

[Deleted member 94047]

Agree with you; there is some improvement in these changes (the 11 second threshold and the, as you aptly put it, "glide path" - even with its potential risk I think it is an improvement).

I guess I'd be a little surprised to see RTH behavior change on existing models, but as it's just a firmware change, it could be done.

I was thinking exactly along those lines; this new set of RTH procedures could indeed be implemented in existing drones through a firmware update; I was wondering if DJI will but I guess it is unlikely.

As for this

Still, the M2 firmware can auto-correct for a badly initialized compass--which is a truly major, drone-saving improvement--yet that firmware change has not percolated into previous models.

Could it perhaps be because the M2's have two compasses and the rest of the Mavics have just the one compass?

 

[DanMan32]

Involuntary glide slope on RTH is a bad idea. Too many obstacles are hard to detect and if signal is lost, you can't completely blame the pilot.

I have descended manually at times during RTH when I would be high, like 300 to 400ft, but I do so knowing I'm clear of obstacles during descent and how far I descend.

I hope MA2 rearward obstacle avoidance is excellent if it is going to fly backwards while you're blind.

 

[NLOL]

Shouldn't descent be a selectable option, that can be turned on in open/flat country or turned off where there are nearby objects? what happens in new system is drone is below point of launch?

 

[Deleted member 94047]

Shouldn't descent be a selectable option, that can be turned on in open/flat country or turned off where there are nearby objects? what happens in new system is drone is below point of launch?

Power saving RTH only activates when the aircraft is too far and too high from the home point, according to the manual. So if AC is below the home point elevation when RTH activates, it follows normal RTH procedure - climb to the set RTH altitude and return home.

 

[Deleted member 94047]

Involuntary glide slope on RTH is a bad idea. Too many obstacles are hard to detect and if signal is lost, you can't completely blame the pilot.

I have descended manually at times during RTH when I would be high, like 300 to 400ft, but I do so knowing I'm clear of obstacles during descent and how far I descend.

I hope MA2 rearward obstacle avoidance is excellent if it is going to fly backwards while you're blind.

Yeah, I agree. There is risk here but I am guessing (since the manual doesn't say how high too high is) there would be a height threshold where by AC will not enter this mode if it is below that height already. It wouldn't make sense for an AC 5km out and only 50m high to get into this glide slope. But all this is conjecture; wish DJI were a little forthcoming with the details of how this works.

 

[z28lt1]

Yeah, I agree. There is risk here but I am guessing (since the manual doesn't say how high too high is) there would be a height threshold where by AC will not enter this mode if it is below that height already. It wouldn't make sense for an AC 5km out and only 50m high to get into this glide slope. But all this is conjecture; wish DJI were a little forthcoming with the details of how this works.

It's not well worded, but I think what the manual is trying to say say is that it will not descend below 50 meters above home point until it gets to the home point. So, seems like 50 meter minimum.

 

[Deleted member 94047]

It's not well worded, but I think what the manual is trying to say say is that it will not descend below 50 meters above home point until it gets to the home point. So, seems like 50 meter minimum.

Yeah, you may be right about that.