Very Disconcerting During Litchi Waypoint Mission

[dwallersv]

I'm guessing this could be simulated by flipping from F to A and back to F. Let the battery run down and see if it comes back. Doing this in an open field with the AC nearby.

I'd like to think this is the sort of functionality that's not under the control of third party developers through the SDK at all. However, DJI has goofed before.

Really, fail-safe low batt should be something that happens regardless. Of anything.

I may goof around with this with a minimission in a sports field while I wait for my camera gimbal parts to arrive. Gives me reason to fly the P4, and I've really been jonesin'.

 

[dwallersv]

Interesting theory. Thanks. Too bad the logs don't contain DOP

That would be nice, wouldn't it?

What day was that wiggly flight? I was having some major GPS inaccuracies yesterday and the day before here in the Santa Cruz, CA area. Hubsan H502S has been 20-30'+ off the last few days on RTH and it's usually within a few feet. I've been puzzled, wondering if the military was doing something with the GPS for testing.

 

[bsartist]

That would be nice, wouldn't it?

What day was that wiggly flight? I was having some major GPS inaccuracies yesterday and the day before here in the Santa Cruz, CA area. Hubsan H502S has been 20-30'+ off the last few days on RTH and it's usually within a few feet. I've been puzzled, wondering if the military was doing something with the GPS for testing.

It was two days ago (Friday) that I had the three missions in a row end with switching from F-WP to P-Atti then to F-GPS and hover. The first two were within a couple feet of each other it seems. The other was probably 20 to 30 feet further on the leg. These occurred earlier in the day before the wiggly line mission. But occurred in the same area of the wiggly line.

What troubles me about the GPS inaccuracy theory is that the problem seemed so very localized and so transient. Yet repeatable.

I should also point out that I had this same problem a day or two before that. Relatively same place but even further down that leg.

 

[dwallersv]

Something was messing with location sensing in that area, there's no question about that. ATTI, wandering path, etc. -- it clearly was having a hard time knowing accurately where it was.

Why is the scary thing... It's a bit of a relief that it seems location-specific, which points more to interference or bad GPS there in some way rather than something wrong with the drone.

 

[skyeboysteve]

Is it? I'm curious regarding the difference between interactive mode and flying a mission disconnected.

Even if you have RTH enabled, the aircraft will continue the mission to completion when disconnection rather than initiate a failsafe RTH. So obviously RTH can be overridden by Litchi under some circumstances. The question is, how is low-battery RTH affected by this? Thata "failsafe" RTH mode too.

I'm hoping you know with some evidence (the Litchi folks, DJI, official docs you read, experts you've spoken to, etc.). If it's just that it makes sense, I agree -- but engineers do make mistakes when designing complex systems.

OK, you've missed the nuance of SMART RTH. SMART RTH activation has nothing to do with radio reception connection; but simply the running calculation that the "smart" battery is performing based on distance from your home, the altitude you've set your RTH to, and thus the energy needed to return home in relationship to the the voltage of the battery. This calculation is represented by the yellow line on your power indicator graph at the top of your screen. I have plenty of first hand "evidence" or at least experience when flying long waypoint missions in the past, and it initiates RTH simply due to the function of battery voltage, RTH height and distance from home; before the mission is complete. The only glitch Litchi has in this matter (that's annoying) is that if you're in radio range it will give you the chance to ignore SMART RTH initiation with a 10 second timer; but even when cancelled it will do it anyway. So, that's why I've known of others that disable SMART RTH before flying Litchi missions, because SMART RTH tends to be rather conservative (mostly because it's calculations are based on return to home speeds that are relatively low compared to the most efficient speed of 31mph--which is a good margin to have, if you're downwind of home at the time); however if you're confident that you have the juice to make it, disabling SMART RTH lets the drone fly till completion.

 

[skyeboysteve]

If Skyboysteve was answering my scenario then we are talking about the AC stopping the mission and hovering out there beyond range sitting in F-GPS mode (no longer F-WP mode). True that it's not in P-GPS mode either. But it's no longer flying a mission. But I suppose maybe it does need to be tested in order to be sure.

I'm guessing this could be simulated by flipping from F to A and back to F. Let the battery run down and see if it comes back. Doing this in an open field with the AC nearby.

This has already been tested by myself. Granted I personally haven't flown a P3S (just P3P and P4P); but I know the "SMART" RTH works the same way--I've seen YouTube videos of Litchi missions flying P3S' with SMART RTH disabled on purpose.... and it works in ATTI mode too. However, if you're in ATTI mode and have no radio connection, you're at the mercy of the winds and have to hope that it doesn't drift into something before SMART RTH kicks in.

 

[Javiejav]

Hey I'm just gonna throw this out there in ref to GPS and possible flight patterns being a little off...

G2 (Moderate) Geomagnetic Storm Watch Issued for 28/29 March | NOAA / NWS Space Weather Prediction Center

Geomagnetic storm watch for this week. Maybe there's something there??

 

[skyeboysteve]

Hey I'm just gonna throw this out there in ref to GPS and possible flight patterns being a little off...

G2 (Moderate) Geomagnetic Storm Watch Issued for 28/29 March | NOAA / NWS Space Weather Prediction Center

Geomagnetic storm watch for this week. Maybe there's something there??

Thanks for the headsup. Around here, the Kp index has been really low the last couple days. However, as dwallersv mentioned... it could be military testing that you've encountered bsartist. I've gotten a couple email alerts from the FAA this past month or so of testing--but it mostly effected high altitude traffic unless you're within a relatively small radius of the testing target. I haven't seen any alerts recently, but I only get them for tests that affect the area where I live.

 

[dwallersv]

OK, you've missed the nuance of SMART RTH. SMART RTH activation has nothing to do with radio reception connection; but simply the running calculation that the "smart" battery is performing based on distance from your home, the altitude you've set your RTH to, and thus the energy needed to return home in relationship to the the voltage of the battery. This calculation is represented by the yellow line on your power indicator graph at the top of your screen.

Nope, didn't miss any of this. Know well how it functions.

Rather, if other RTH modes can be over-ridden by the SDK, for example failsafe RTH, it doesn't seem nuts to wonder if Smart RTH can be over-ridden too. Do you know, or are you going with what "makes sense" to you?

It makes sense to me that this would be a "firewalled" bit of functionality, inaccessible and unaffected by anything the aircraft is doing, whether interactive or running autonomously, regardless of where the commands are coming from (DJI GO or a third-party app).

This is just what makes obvious sense. DJI has done the opposite of what makes obvious sense before (witness the NFZ behavior), so I'm a little concerned without a verified, know-for-sure answer

 

[skyeboysteve]

Nope, didn't miss any of this. Know well how it functions.

Rather, if other RTH modes can be over-ridden by the SDK, for example failsafe RTH, it doesn't seem nuts to wonder if Smart RTH can be over-ridden too. Do you know, or are you going with what "makes sense" to you?

It makes sense to me that this would be a "firewalled" bit of functionality, inaccessible and unaffected by anything the aircraft is doing, whether interactive or running autonomously, regardless of where the commands are coming from (DJI GO or a third-party app).

This is just what makes obvious sense. DJI has done the opposite of what makes obvious sense before (witness the NFZ behavior), so I'm a little concerned without a verified, know-for-sure answer

Smart RTH can be over-ridden (as I mentioned earlier). It's a toggle switch, which stores the setting on-board the phantom. If you keep Smart RTH enabled, then the battery will over-ride anything else that the P3 was programmed to do... and yes, (as I stated earlier) I've tested this (sometimes inadvertantly) many times and Litchi doesn't over-ride (without user input) the Smart RTH feature regardless of radio connection status. The commands are not coming from the app you're using--other than setting your toggle switch initially. I've used mapping software that doesn't have the toggle switch, so whatever was stored on-board by the Go app or Litchi previously is what's used, and sure enough, when mapping and Smart RTH kicks in, the phantom stops its grid flying, and initiates RTH.

 

[RodPad]

So it appears the compass calibration fixed the heading problem and it switching to Atti?

I recall @Meta4 stating many times, the the Phantom will go into Atti if there is conflict of data between the gps and compass, besides not having enough GPS satellites.

Rod

 

[bsartist]

So it appears the compass calibration fixed the heading problem and it switching to Atti?

I recall @Meta4 stating many times, the the Phantom will go into Atti if there is conflict of data between the gps and compass, besides not having enough GPS satellites.

Rod

May be too early to tell on the Atti problem. It had gone away before. The heading it better. It's either due to compass calibration or disappearance of some kind of localized interference. If I had calibrated right after the three failures on the 24th I would have a more solid temporal relationship.

 

[dwallersv]

Smart RTH can be over-ridden (as I mentioned earlier). It's a toggle switch, which stores the setting on-board the phantom. If you keep Smart RTH enabled, then the battery will over-ride anything else that the P3 was programmed to do... and yes, (as I stated earlier) I've tested this (sometimes inadvertantly) many times and Litchi doesn't over-ride (without user input) the Smart RTH feature regardless of radio connection status. The commands are not coming from the app you're using--other than setting your toggle switch initially. I've used mapping software that doesn't have the toggle switch, so whatever was stored on-board by the Go app or Litchi previously is what's used, and sure enough, when mapping and Smart RTH kicks in, the phantom stops its grid flying, and initiates RTH.

Thanks for sharing your actual experience.

I think where we're talking past each other is because there are different types of RTH, as outlined in the manual and based on what they do.

There is "Failsafe RTH", "Smart RTH", and plain old user initiated "RTH".

Failsafe RTH is triggered under several different critical circumstances. One is when the aircraft loses connection with the RC. This can be overridden by third-party software like Litchi, so that it doesn't activate. This is obviously needed if it is to be possible to program and fly missions where the connection may necessarily be lost.

I understand how Smart RTH works. I was simply concerned that, since Failsafe RTH can be overridden through the Software Development Kit Application Programming Interfaces, could this also be true for Smart RTH?

Your own tests are encouraging, but for me not definitive. I'd like to hear from someone who is familiar with programming the P4 through the DJI interfaces -- hear that this is simply not possible (disabling Smart RTH).

 

[bsartist]

Thanks for sharing your actual experience.

I think where we're talking past each other is because there are different types of RTH, as outlined in the manual and based on what they do.

There is "Failsafe RTH", "Smart RTH", and plain old user initiated "RTH".

Failsafe RTH is triggered under several different critical circumstances. One is when the aircraft loses connection with the RC. This can be overridden by third-party software like Litchi, so that it doesn't activate. This is obviously needed if it is to be possible to program and fly missions where the connection may necessarily be lost.

I understand how Smart RTH works. I was simply concerned that, since Failsafe RTH can be overridden through the Software Development Kit Application Programming Interfaces, could this also be true for Smart RTH?

Your own tests are encouraging, but for me not definitive. I'd like to hear from someone who is familiar with programming the P4 through the DJI interfaces -- hear that this is simply not possible (disabling Smart RTH).

Sounds like he said that there is an option to disable it in the interface. Therefore there must be an API call to do it.

 

[dwallersv]

Sounds like he said that there is an option to disable it in the interface. Therefore there must be an API call to do it.

Good catch!

So now I'm worried again...

 

[twowingers]

Calibrate it frist ?

 

[reversevector]

Jumping over from the P3S forum, I had an issue where pausing a Litchi waypoint mission and then losing remote signal causes the drone to get stuck. Upon reconnecting, I couldn't initiate RTH through Litchi or by toggling the remote switch. Toggling the mode switch didn't work and the drone didn't respond to any stick inputs.

I got it working by resuming the mission and then initiating RTH through Litchi. I wonder if the Litchi waypoint mode disables low battery RTH and the aircraft only makes a forced landing on critical low battery.

 

[HWCM]

Here are two separate runs of the curved turns version of the mission. Pretty amazing how exact they both are. If you know how to relax your eyes and do a stereogram you can watch in 3D

That is almost perfect 3D

 

[reversevector]

Well, I crashed the bird after a Litchi waypoint mission stopped halfway with a downlink disconnect. I managed to regain manual control but the video downlink never re-established itself, even after starting the app a few times and then restarting the remote.

Worst of all, the bird seemed to fly on Atti and it was drifting everywhere because of some wind. I toggled back and forth between GPS and Atti modes but it didn't work. I got it near the ground but the drift and a hard landing (plus the infrared camera mounted below the gimbal protector) made it flip over on its back. Almost gave me a heart attack.

Damage - two props were badly scratched on gravel, some minor scratches on the body, no issues with the gimbal and motors on a test flight the next morning. The lens filter on the infrared camera took most of the impact and ended up with a slight scratch. I swapped out all props just in case. The Litchi logs showed a downlink disconnect and nothing else... I'm wondering if this has to do with the latest Litchi and DJI SDK update.

 

[jndalee]

Here is the mission.

Mission Hub - Litchi

It stopped drizzling so I flew it again and it worked. The only thing is that the drone is not facing the POI like it should - heading seems to be off. I wonder if there is some compass interference at that point. But at 300'? Would that even cause this? But the heading is off in a reproducible way. I get no compass errors though.

I know this is an old thread but I have been looking for an answer to an issue I am having and noticed you are the only person that has given a link to you mission plan. looking at your mission WP 5 , 6 and 7 the POI focus is disabled, but you probably found that by now?