Does RTH work with Compass Error?

[syotr]

This is getting dull. If you want your Phantom to swirl around all over the place in the off chance it can eventually get home, go for it. The success rate will be low. There's a reason why every AHRS based navigation system that is in the air and on the water uses a compass.

We flew to the moon without a compass or GPS. We used gyros and corrected errors on the way. The computer had less power than the one in your cell phone.

 

[Shammyh]

This is getting dull. If you want your Phantom to swirl around all over the place in the off chance it can eventually get home, go for it. The success rate will be low. There's a reason why every AHRS based navigation system that is in the air and on the water uses a compass.

Well, that's kinda the point. None of this is practical anyway... And I thought we we're talking about last ditch in-case-of-compass-failure type scenarios?

Sort of how a hex can still "fly" with a single prop/motor failure. It'll be spinning like crazy, and not able to serve as a camera platform anymore, but it'll get itself home.

That was what I was going for. Like, if you only have access to gps position and the imu, what's the most efficient way to use that data to get home? Sorry if that's too dull. ;-)

 

[eBoon]

My fixed wing Bix 3 does just fine with the compass disabled. That includes RTH.

 

[N017RW]

My fixed wing Bix 3 does just fine with the compass disabled. That includes RTH.

What flight controller does it use?

 

[MikeyOnline]

Read every post and since I posted a similar idea in another thread, I'll throw my two pennies in and see if they come out the coin return...

My first thought is: what causes a compass error? I think most times, it is the result of the bird trying to maintain course so it heads off in the direction dictated by the compass and finds out (via GPS) that it's not going that way. So it throws a compass error and stops using the compass. In such a case (probably nearly all of the cases out there), the compass isn't dead. It's just mixed up. So it should be possible to look at the compass, look at where it's going by GPS, and recalibrate the compass on-the-fly well enough to get home. In other words, regardless of what the compass is reading, you could probably tweak motor output a few times until you notice that the AC is headed in the general direction of home (via GPS) and then use the current compass reading (whatever that is) to hold that OR use the information you have in order to get a gross compass calibration good enough to get you near home. For that matter, you might be able to yaw it a full rotation while in-air in order to get a reasonable compass reading and recalibrate it while it is flying. You couldn't get the "face down" calibration but 360 upright might be enough to get you in range enough for it to fly.

On the "limp home" idea, I do think it could work but it would be messy. If the AC is moving, you have a GPS reading but you have no idea which way the AC is pointed so you don't know how to tweak the motors. So you'd have to just give one motor 10% more thrust and see how that affects the GPS movement. If it is moving away from home and that tweak makes it move away even faster, bring that motor back to where it was and give the opposite motor +10%. Is it at least slowing down? Did it change heading? Example: home is due north but the AC is moving south (again, you know this via GPS). Give motor 1 +10% thrust. Now it's moving south even faster so you put motor 1 back to where it was and give motor 4 (opposite side) +10%. Now it stopped going south and is going west. Now give motor 3 10% thrust. Going more north now? And continue that logic until you get a north heading and keep running the logic because wind and other factors could change it. I think this "extrapolation" could get you close enough to home to at least hear the beehive and take over in ATTI mode. I don't see altitude as a problem because the baro is separate from the compass.

Mike

 

[syotr]

There were many flyaways with the Pi and P2 models. Most certainly many were caused by compass errors. With the P3 and P4, there have been very few and now the aircraft reports back when it has compass errors and is still able to fly home. It would seem that DJI has figured out how to ignore the compass and still come home.

 

[eBoon]

What flight controller does it use?

Eagle Tree Vector. We purposely disable the compass on fixed wing with this and most fixed wing installations.

Not saying that the phantom doesn't have problems with a wonkie compass only that GPS RTH is not dependant on a working compass.

 

[Mimoid]

There were many flyaways with the Pi and P2 models. Most certainly many were caused by compass errors. With the P3 and P4, there have been very few and now the aircraft reports back when it has compass errors and is still able to fly home. It would seem that DJI has figured out how to ignore the compass and still come home.

This is in fact what this long (and now painfully drawn out) thread is all about.

When a compass error is detected, the P3 (and presumably also the P4) will still try to use the incorrect data. It will not switch into some kind of "limp mode" but will simply warn that the data from the sensors are not congruent and will recommend you to fly manually (it even switches to Atti mode by itself).

If the automatic RTH or Home Lock modes are nevertheless engaged in this situation, they may make the Phantom fly in a completely wrong direction.

//Tom

 

[Oso]

This has been an interesting thread. I've enjoyed reading the passionate debate and especially the arguments explaining in detail (with examples) how DJI could implement a no compass RTH if they ever chose to do so. Thanks guys. It has been so much better than repeatedly reading all of the threads asking easily searched for questions such as "does full left stick down stop the motors in mid flight" or "which tablet is best" or "which SD card is best".

 

[Mimoid]

Me too. But somehow I just cannot help but wishing that DJI had included some of these possible functions into the P4 instead of some of the more esoteric functions, like the "sport" mode. For most of us, a Phantom drone is quite a hefty investment, and any help that can save our precious toy from disappearing (not to mention all the possible collateral damage that can be the result of a malfunction or a pilot error) would be extremely welcome.

I would really suggest DJI that the calibration process (for both the IMU and the Compass) would be expained in much more detail in the manual, as it is definitely not just something that can (or should) be done on regular basis. You need to know WHEN, WHY and HOW to do it. It is not just a matter of doing the "DJI dance" every time you go out flying, "just to be safe". It is in fact a very delicate action that will affect the whole functionality of the drone.

My thanks to all of you who contributed to this thread!

//Tom

 

[ianwood]

This is in fact what this long (and now painfully drawn out) thread is all about.

When a compass error is detected, the P3 (and presumably also the P4) will still try to use the incorrect data. It will not switch into some kind of "limp mode" but will simply warn that the data from the sensors are not congruent and will recommend you to fly manually (it even switches to Atti mode by itself).

This is a good point. And it is this strategy that has the highest likelihood of succeeding. I've seen compass errors come up in weird places, thankfully not that often. But an open field at 100ft is either full of magnetite or the flight controller got it wrong. Either way, it kept flying just fine.

Other times near the 400kV power lines over the LA river, there are no compass errors. But it flies crooked, the camera horizon drifts and it does this weird bucking every once in a while. That's all because the compass is getting messed with by electromagnetic interference. Yet it's not detected by the flight controller.

And even if the FC could detect it and switch to some blindfolded attempt to go home, it would be in the papers the next day for causing a regional blackout when it stumbles liken a drunk into the pylons.

 

[syotr]

I have worked as an engineer and programmer for 44 years. I don't know how many times someone told me something was impossible when I could clearly see the solution. More than once, I built a working prototype or wrote a program just to prove them wrong. The last 16 years I worked for a company that designed and built custom, automated mining machinery. They joked that the "impossible" just takes an extra 2 weeks (unless it violated the laws of physics, then it was 3 weeks.) There were no problems, only opportunities.
For those of you who have not done programming for modern computers, I have to say you have no idea what is possible. Computers solve lots of problems where there is no direct solution, by iteration, that is a good guess followed by testing and refinement. They can recalculate thousands of times in a second. With artificial intelligence, they can look at trends in multiple streams of data and optimize a process that has so many variables that the programmers don't really know how it arrived at the solution, just that it works.
Who would have thought 10 years ago, that we would have an affordable copter that could fly overhead, track, follow and shoot high res video of a moving subject by computer vision? Navigating without a compass is trivial compared to this.

 

[MikeyOnline]

I agree, and I'll add one thought. I for one don't presume that DJI is too dumb to implement an extrapolated limp home without a compass and wonder why "we are smarter". DJI would have no problem implementing something like that but in the real world, it comes down to what do you put in a craft based on how likely you will be to need it and how much testing is involved. Then there's liability. They have to think about things we don't. What if you set your height to 30m because you know you'll be flying due south and won't come closer than say 200m to that cell tower to the west. But then when it decides to limp home like a drunken sailor when the compass fails and it clips it on the way home, is that a warranty claim? Lots to consider. So I suspect the fact that DJI didn't implement it has nothing to do with them not knowing how, and much to do with the practicality of such a feature and the amount of testing involved to make it happen.

Mike

 

[N017RW]

Yup MO.
All of these solutions assume there is ample computing power, memory, etc.
Phantoms are DJI's lowest tier products so I don't fault them for delivering a fine machine and features at the price point.

 

[HWCM]

Me too. But somehow I just cannot help but wishing that DJI had included some of these possible functions into the P4 instead of some of the more esoteric functions, like the "sport" mode. For most of us, a Phantom drone is quite a hefty investment, and any help that can save our precious toy from disappearing (not to mention all the possible collateral damage that can be the result of a malfunction or a pilot error) would be extremely welcome.

I would really suggest DJI that the calibration process (for both the IMU and the Compass) would be expained in much more detail in the manual, as it is definitely not just something that can (or should) be done on regular basis. You need to know WHEN, WHY and HOW to do it. It is not just a matter of doing the "DJI dance" every time you go out flying, "just to be safe". It is in fact a very delicate action that will affect the whole functionality of the drone.

My thanks to all of you who contributed to this thread!

//Tom

I believe the P4 has dual compasses like the pixhawk for redundancy.

 

[Erised]

I was reading my p4 manuel and it say's that RTH will not work if GPS doesn't have a lock but doesn't say anything about the compass.

 

[bryan_m]

Interesting read here. Seems now I would know what to do in the case of a compass error and avoid flyaway. My first instinct would have been to switch to ATTI - IOC and use either home lock or course lock to fly back. However it seems that this would also use the compass so the solution would be ATTI and just fly visually with video till it's close enough to determine fully orientation and land. I'm not sure why the Home Lock doesn't use the same method as those $50 drones which have no GPS or compass. I have a JJRC which will fly headless back to home point. I would guess it's just following the vector from the strongest radio signal for control. DJI could easily implement that as an update.

 

[warmbooter]

GPS is enough to know the direction, if the drone is moving. As soon as you have a vector from point A to point B, you know the direction. But I'm not saying Phantom will do this... my guess is that it depends on good compass calibration to fly correctly.

 

[JohnK]

Doesn't the green "tracking" path in the app still work even in ATTI mode and loss of video feed? Raining here so I can't go do a test for myself.

 

[dgd3]

If the compass error means you have no compass, the aircraft can't come home because it doesn't know where home is. Even if it knows its own GPS location and even if it knows the home point location, it still can't find home. It's like - what if you woke up and found yourself in a field in complete darkness. You know your house is exactly 100 meters due north, but you have no idea where north is. So you stumble off in a random direction hoping to get lucky...

What I don't understand is why the software wasn't (can't) have the GPS "drag" the phantom back home sideways or at whatever orientation it is pointing the same way we can control the direction of motion no matter which way front is pointing.