Mission Completed - Not

[bsartist]

Yes. However, when it says "disconnected" instead of "mission ended", the preprogrammed gimbal movements don't work. When it's truly lost contact, the gimbal movements don't work and speed changes between waypoints don't work. This is a limitation of the DJI SDK and not Litchi's fault. Litchi programmers can only work within the confines of what DJI allows in their SDK (Software Development Kit).

Here is a repost of one of my detailed explanations from above:

"I'm aware of the connection requirement for automated gimbal control. This behavior as documented does not match my scenario.

1) I never got a disconnected message as you see on the screen recording and mission logs. I only got a very brief signal drop to 8% and it shot back up to solid signal.

2) even if we want to call that a disconnect it was very short lived. Very. A s cond or two. The AC and RC were solidly in connection for the remaining waypoints for which gimbal control is ignored."

 

[SkyHog]

Here is a repost of one of my detailed explanations from above:
"I'm aware of the connection requirement for automated gimbal control. This behavior as documented does not match my scenario.
1) I never got a disconnected message as you see on the screen recording and mission logs. I only got a very brief signal drop to 8% and it shot back up to solid signal.
2) even if we want to call that a disconnect it was very short lived. Very. A s cond or two. The AC and RC were solidly in connection for the remaining waypoints for which gimbal control is ignored."

I have flown about about twenty Litchi waypoint missions on a course that I have been using to characterize Litchi, but I use the Android version, which operates somewhat differently than the ios version. In any case, my course runs about the same shape and length and treed environment as yours. There are two spots where I dip down below the tree line over two separate ponds in my neighborhood, which often causes a very degraded signal while over the first pond, and a complete signal loss over the second pond. Very good for tests.

I often see a link with zero bars (ios shows a link quality number, as well as bars, but android shows bars only). My missions often show zero bars for long stretches, with no apparent problems, continuing through waypoints, but with poor video, (I don't think the signal indicators are very well calibrated to true signal strength). At some point, I run out of control signal completely, and then I get a "Signal Lost" voice message, plus a red display message display that says "Disconnected" "Signal Lost".

I have never heard the "Mission Ended" message, even with a complete signal loss. Instead, I just hear "Signal Lost". When I hear the "Signal Lost" message, the AC will continue to fly the rest of the mission normally but eventually it will reconnect automatically. It won't control the gimbal, nor change speeds between waypoints while it is disconnected. However, it will never resume control. Once it loses connection, all the smooth interpolation of speeds and gimbal pitch not only stop, but they remain frozen for the rest of mission, at whatever pitch and speed that existed just before the disconnect, whether you regain a good signal or not.

I wonder if your "Mission Ended" voice message in ios is essentially the same as "Signal Lost" in the android version, because the behavior you described matches the same behavior I observe with "Signal Lost". I know you said you only had degraded signal for a brief time down to 8%, BUT, I doubt that displayed signal indicator is really accurate and updated frequently enough, especially under poor signal conditions. I'm not certain you can rely on that number to be sure you didn't lose link. But if you did lose link, there should have been a notification other than "mission ended". I would think.

Hope that helps.


EDIT: I found a configuration that does indeed allow Litchi to resume control of speed and gimbal after a signal reconnect. See my Feb 8 2016 post below.

 

[bsartist]

However, it will never resume control. Once it loses connection, all the smooth interpolation of speeds and gimbal pitch not only stop, but they remain frozen for the rest of mission, at whatever pitch and speed that existed just before the disconnect, whether you regain a good signal or not.

This part is very interesting. I thought it was supposed to regain gimbal control upon reconnect. Maybe I'm wrong. I'll have to run some smaller tests. Maybe shutting off the controller to force a disconnect.

I hear what you are saying about not relying on the signal % on the iPhone. But I also read the log.

So far I've sidestepped that issue on this project my slightly changing my position where I stand with the controller to keep the house from interfering. Now I'm just battling some erratic gimbal movements.

 

[SkyHog]

This part is very interesting. I thought it was supposed to regain gimbal control upon reconnect. Maybe I'm wrong. I'll have to run some smaller tests. Maybe shutting off the controller to force a disconnect.
I hear what you are saying about not relying on the signal % on the iPhone. But I also read the log.
So far I've sidestepped that issue on this project my slightly changing my position where I stand with the controller to keep the house from interfering. Now I'm just battling some erratic gimbal movements.

Yes, I scared myself a couple of times as the AC descends to my last waypoint just above the ground at a much higher descent speed than I had planned, even with a full five bars of link signal. It is unnerving to see the AC descending straight down at you at 15 mph, when it's planned speed was only 1 mph. Maybe it would have stopped at the last waypoint, just 20 feet above the ground, but I sure didn't trust it to stop on a dime with that kind of descent speed. I quickly hit P-mode, and then landed it myself. It took a couple of flights to figure out the bit about frozen speeds, as well as frozen gimbal. Now I always land manually in P-mode rather than trust an Autoland in Litchi, unless I see the approach speed is consistent with my plan. I know it works well if the link is never disconnected.

BTW, if you did observe a slow approach to autoland, consistent with a slow approach that you planned in your mission waypoint plan, that would indicate to me that you probably did NOT lose link during the mission.

EDIT: I found a configuration that does indeed allow Litchi to resume control of speed and gimbal after a signal reconnect. See my Feb 8 2016 post below.

 

[SkyHog]

Yes, I scared myself a couple of times as the AC descends to my last waypoint just above the ground at a much higher descent speed than I had planned, even with a full five bars of link signal. It is unnerving to see the AC descending straight down at you at 15 mph, when it's planned speed was only 1 mph. Maybe it would have stopped at the last waypoint, just 20 feet above the ground, but I sure didn't trust it to stop on a dime with that kind of descent speed. I quickly hit P-mode, and then landed it myself. It took a couple of flights to figure out the bit about frozen speeds, as well as frozen gimbal. Now I always land manually in P-mode rather than trust an Autoland in Litchi, unless I see the approach speed is consistent with my plan. I know it works well if the link is never disconnected.

BTW, if you did observe a slow approach to autoland, consistent with a slow approach that you planned in your mission waypoint plan, that would indicate to me that you probably did NOT lose link during the mission.

But the other possibility, is that you toggled to P-Mode somewhere in the middle of the mission, which would then trigger an RTH if you later lost link, or if it did never lose link, you might have manually commanded an RTH while in P-mode (I have forgotten the details of your story). Either of those scenarios would explain the nice smooth landing at home that you observed as well.

 

[The Springs]

I have been running a similar mission where I will lose connection due to a group of trees near the house .I can avoid this by walking a 100m to get a clear LOS. When I say I lose connection the telemetry stops as does the map and video but I don't get a message or voice telling me any thing .
I too thought that once you regained connection the mission would resume (gimbal angle and speed) If speed between WPs is not restored this could dramatically effect the overall mission time ?
My main problem is the instability of the app on the s tab ,very difficult to connect and dodgy in FPV.
Also be a bit cautious with RTH I had it set to hover but it returned instead and kept going at high speed and wasn't going to stop.

 

[bsartist]

Yes, I scared myself a couple of times as the AC descends to my last waypoint just above the ground at a much higher descent speed than I had planned, even with a full five bars of link signal. It is unnerving to see the AC descending straight down at you at 15 mph, when it's planned speed was only 1 mph. Maybe it would have stopped at the last waypoint, just 20 feet above the ground, but I sure didn't trust it to stop on a dime with that kind of descent speed. I quickly hit P-mode, and then landed it myself. It took a couple of flights to figure out the bit about frozen speeds, as well as frozen gimbal. Now I always land manually in P-mode rather than trust an Autoland in Litchi, unless I see the approach speed is consistent with my plan. I know it works well if the link is never disconnected.

BTW, if you did observe a slow approach to autoland, consistent with a slow approach that you planned in your mission waypoint plan, that would indicate to me that you probably did NOT lose link during the mission.

This switching to manual mode always reminds me of Neil Armstrong taking over control of the LEM during final descent to avoid crashing into boulders.

 

[SkyHog]

I have been running a similar mission where I will lose connection due to a group of trees near the house .I can avoid this by walking a 100m to get a clear LOS. When I say I lose connection the telemetry stops as does the map and video but I don't get a message or voice telling me any thing .
I too thought that once you regained connection the mission would resume (gimbal angle and speed) If speed between WPs is not restored this could dramatically effect the overall mission time ?
My main problem is the instability of the app on the s tab ,very difficult to connect and dodgy in FPV.
Also be a bit cautious with RTH I had it set to hover but it returned instead and kept going at high speed and wasn't going to stop.

If you don't get the Disconnected message and/or Signal Lost message, it's hard to know exactly what the link state is, particularly when my signal strength shows zero bars quite a lot with no apparent effect on the mission. So I haven't been able to correlate the marginal link state with gimbal or speed freezing; only the full link loss. I think the link has to be completely gone for three full seconds before the Disconnected message is announced, and then it's easy to see that the gimbal stops moving after that, and the speeds stop changing, even when the link comes back.

Yes, I think you are right about the need for caution with all the end-of-mission actions (RTH, Hover, and AutoLand). Some testing has been done to confirm that AutoLand does NOT work unless the controller signal is present for the landing. The AC will attempt RTH instead of AutoLand if no RC signal is present, which could be very risky if you haven't planned enough battery for a round trip. Hover and RTH apparently do work without an RC signal present, however, RTH might be engaged with an undesirable frozen speed, as I mentioned earlier, so I wouldn't trust RTH in any case where the signal might have been lost on the way to the destination. That is why I only do long distance Litchi waypoint missions that bring the AC back to the starting point, where I can always take over and land it myself if I need to. Some others have tried one-way long distance missions, far from the starting point, but I'm not sure how successful those have been, and I just don't know how to do that safely, myself. But I'm comfortable running very long two-way missions, now that I understand the limitations.

 

[bsartist]

I contacted Litchi regarding the specific question of whether gimbal movements and speed changes will resume after reconnection. Their answer was "Yes, they will resume execution." So now I'll send them the complete scenario.

 

[SkyHog]

Interesting. They clearly do not resume execution, in my tests. Maybe that's an artifact of the latest firmware 1.8.10? Maybe it's a configuration issue of some sort. I'll try a few more things, but I'm heading to the islands and I won't be able to do any more tests for about a week or so.

EDIT: I've done more tests and found a new configuration that does indeed work correctly.

 

[Pharm]

Here is a repost of one of my detailed explanations from above:

"I'm aware of the connection requirement for automated gimbal control. This behavior as documented does not match my scenario.

1) I never got a disconnected message as you see on the screen recording and mission logs. I only got a very brief signal drop to 8% and it shot back up to solid signal.

2) even if we want to call that a disconnect it was very short lived. Very. A s cond or two. The AC and RC were solidly in connection for the remaining waypoints for which gimbal control is ignored."

Upon further reflection, I could be wrong. The reason I say this is because the mission I was thinking about involved mapping a large plot of land. My gimbal was set to point the camera straight down at every waypoint and there was never a camera angle change. If it DID lose the ability to change the gimbal and "stuck" at the last gimbal position prior to the "mission ended" warning, I wouldn't have known it because the angle was always the same. Hmmm... Now I'll have to try it again when it warms up a bit. My Phantom may fly ok in colder weather but I tend to personally have very limited tolerance to standing outside freezing my bagettes (sp?) off!

I noticed a Litchi update this morning that says, among other things, "improvements and bug fixes". Maybe that's one of the bugs.

 

[SkyHog]

I'm looking forward to testing with the new Litchi update, but unfortunately, I won't be able to do that for another week. I hope it fixes the reconnect problem.

 

[bsartist]

Here is what Litchi said about this:

Hi,

The mission ended is a bug from the DJI dev kit which we already reported to DJI. Litchi will announce “mission ended” whenever it is told that the mission has ended, which wasn’t the case here. It should be rare however as we have not received a lot of similar reports, the reports we did receive in the past were all for the P3 Standard however and would happen when signal is lost (even if just for a split second). It is possible that after the DJI dev kit reports that the mission has ended, it will also stop reporting other changes to Litchi such as when the drone reaches a waypoint for example, this kind of information is required for the gimbal control to work.
We’re not sure when (or if) DJI will fix this bug, we reported it a long time ago for the P3 Standard.

 

[SkyHog]

Well I'm back from my trip, and I've recently completed more testing. Now I've got new results. Now the Litchi app does indeed resume gimbal control and speed control after RC reconnecting on a waypoint mission. This is great news, because it had behaved differently for a long time in my system. It used to freeze after an RC disconnect/ reconnect, and Litchi would not resume gimbal or speed control even after reconnecting with a solid signal. But it behaves correctly now.

So what changed? Unfortunately, two things happened since my last tests a couple of weeks ago. First, I uninstalled my DJI Go app, to eliminate any potential conflicts between the two apps. Second, Litchi updated the app to version 3.10.3, since I flew last, and I had auto updates enabled, so I have a new version of Litchi installed now. So I'm not sure which of these changes is responsible for the improvement in Litchi behavior after RC signal re-connection. I suppose it could be both or either one. In any case, I'm glad to get a configuration that does work properly now.

My next step is to reinstall DJI Go, to see if Litchi still behaves well. However, I'll probably reinstall ver 3.0.1 because I have had a lot of concern about the newest DJI Go ver, 3.1.2 anyway, because of the HomePoint setting issue in the latest version of DJI Go. That HomePoint location error issue almost got me once already.