DJI Mavic, Air and Mini Drones
Friendly, Helpful & Knowledgeable Community
Join Us Now

Poor GPS - please help to analyze and make me a better pilot

lee82gx

Well-Known Member
Joined
Nov 19, 2019
Messages
681
Reactions
444
Age
41
Location
Malaysia

Kindly see the flight data, and I had a section of flight near an apartment block, which resulted in ATTI mode. I can say I've flown this path quite a number of times, maybe roughly at the same altitude even.

This is the first time my Mavic Air 2 has gone into ATTI mode and it did give me quite a scare at first. I encountered this on the way out and back, near the same location.

Questions:
1. Am I right that being so close to tall buildings will occlude the GPS signals? Is this what happened?
2. Next time should I land immediately? Or should I just stay calm like I did and maneuver home via the FPV screen?
3. Should I increase altitude?

Thank you
 
Weak GPS warnings were issued whenever the drone was close to those buildings. The height of your drone was over 100 meters and those buildings do not look like that tall so weak GPS was unlikely to be caused by those buildings obstructing the signals. I can only make a wild guess here. May be there are drone jammers in those apartments ? I can only suggest staying away from those buildings in such circumstance.


1606556784196.png
 
Last edited:
1. Am I right that being so close to tall buildings will occlude the GPS signals? Is this what happened?
Your drone needs to have a clear view of the sky to receive satellite signals.
When obstacles block the drone's skyview, the number of sats it can receive is reduced.
Less sky view = less satellites.

The number of sats your drone was receiving signals from decreased from about 17 down to 4 on the way out and zero on the way back.

The periods where satellite numbers dropped and GPS Health was rated 0 or 1 were:
6:36.8-6:49.7
6:53-7:16.2
7:44.2-8:10.1

The buildings you flew past appear to be 30 stories high (approx 100 metres) and built on land 20 metres lower than your launch point.
The drone was at 100-115 metres, which would have put it 20-35 metres higher than the buildings.
At that height, it would have had a 100% skyview so the buildings were not blocking satellite signal.

I've seen a lot of flight's data but never seen any indication of GPS jamming but it does look like there's something happening there that significantly reduces GPS satellite reception.
2. Next time should I land immediately? Or should I just stay calm like I did and maneuver home via the FPV screen
This kind of GPS interference is very rare and almost anywhere else that you fly, you're not likely to run into it.
The drone is still completely controllable without GPS but it loses the ability to hold position, so it has no "brakes".
If you take your hands off the sticks it will keep going so it's easy to crash if you are flying close to obstacles.
But away from obstacles, there's no real need to be concerned.
Just use gentle joystick inputs and stay clear of obstacles.
3. Should I increase altitude?
 
Weak GPS warnings were issued whenever to fly close to those buildings. The height of your drone was over 100 meters and those buildings do not look like that tall so weak GPS was unlikely to be caused by those buildings obstructing the signals. I can only make a wild guess here. May be there are drone jammers in those apartments ? I can only suggest staying away from those buildings in such circumstance.


View attachment 117987

Thanks for your reply boblui, I always look forward to your sensible analysis. When I look a little more detailed, the areas when I was on the way back with zero sat signal is nearer to the whitish building in the far right of the picture you posted. Which is actually unoccupied as of right now!
Your drone needs to have a clear view of the sky to receive satellite signals.
When obstacles block the drone's skyview, the number of sats it can receive is reduced.
Less sky view = less satellites.

The number of sats your drone was receiving signals from decreased from about 17 down to 4 on the way out and zero on the way back.

The periods where satellite numbers dropped and GPS Health was rated 0 or 1 were:
6:36.8-6:49.7
6:53-7:16.2
7:44.2-8:10.1

The buildings you flew past appear to be 30 stories high (approx 100 metres) and built on land 20 metres lower than your launch point.
The drone was at 100-115 metres, which would have put it 20-35 metres higher than the buildings.
At that height, it would have had a 100% skyview so the buildings were not blocking satellite signal.

I've seen a lot of flight's data but never seen any indication of GPS jamming but it does look like there's something happening there that significantly reduces GPS satellite reception.

This kind of GPS interference is very rare and almost anywhere else that you fly, you're not likely to run into it.
The drone is still completely controllable without GPS but it loses the ability to hold position, so it has no "brakes".
If you take your hands off the sticks it will keep going so it's easy to crash if you are flying close to obstacles.
But away from obstacles, there's no real need to be concerned.
Just use gentle joystick inputs and stay clear of obstacles.
Hi meta4, thanks for your reply. I also look forward for your analysis as usual. Yes, i did fly higher than the buildings which is my usual practice. I can attach another flight that is similar whereby I fly a little bit further away from the buildings with no gps issue, on another day.
Could it be possible that most of the satellites still above my drone was still occluded being closer to the horizon?
 
Could it be possible that most of the satellites still above my drone was still occluded being closer to the horizon?
I don't think so.
The orbits of the sats are arranged to provide a good worldwide coverage, so they don't end up clustered together.
With 17 sats visible in the sky, they would have been well spread out.
The sats orbits are like this,
Except there are a lot more up there now
ConstellationGPS.gif
 
I don't think so.
The orbits of the sats are arranged to provide a good worldwide coverage, so they don't end up clustered together.
With 17 sats visible in the sky, they would have been well spread out.
The sats orbits are like this,
Except there are a lot more up there now
ConstellationGPS.gif
Very nice representation and explanation...sweet! (do people still use that phrase?) I just don't see what would cause his Sats to bottom out like they are? That is a apartment building, very hard pressed to believe ANY type of Jamming let alone signal is causing issue coming from there. Less maybe a Ham operator with a Yagi trying to "blast him" :p
 
Last edited:

Kindly see the flight data, and I had a section of flight near an apartment block, which resulted in ATTI mode. I can say I've flown this path quite a number of times, maybe roughly at the same altitude even.

This is the first time my Mavic Air 2 has gone into ATTI mode and it did give me quite a scare at first. I encountered this on the way out and back, near the same location.

Questions:
1. Am I right that being so close to tall buildings will occlude the GPS signals? Is this what happened?
2. Next time should I land immediately? Or should I just stay calm like I did and maneuver home via the FPV screen?
3. Should I increase altitude?

Thank you
The MA2 .DAT log file contains information about GPS being jammed or spoofed.
1606572172506.png
I don't know much about this data. It might be worth a look if you could retrieve the .DAT and post it. Information on retrieving the .DAT can be found here
Mavic Flight Log Retrieval and Analysis Guide
See section 3. Log File Retrieval and then in subsection
Mobile device DAT files (DJI GO 4 & DJI Fly)
 
The MA2 .DAT log file contains information about GPS being jammed or spoofed.
View attachment 117999
I don't know much about this data. It might be worth a look if you could retrieve the .DAT and post it. Information on retrieving the .DAT can be found here
Mavic Flight Log Retrieval and Analysis Guide
See section 3. Log File Retrieval and then in subsection
Mobile device DAT files (DJI GO 4 & DJI Fly)
Capture.JPG

Hi BudWalker,

In my version of CSViewer it does not show the GPS_hw_stats group, is this a special setting inside of CSViewer that I have to enable ?


Here is a copy of the raw dat file....
 
  • Like
Reactions: BudWalker
View attachment 118002

Hi BudWalker,

In my version of CSViewer it does not show the GPS_hw_stats group, is this a special setting inside of CSViewer that I have to enable ?


Here is a copy of the raw dat file....
CsvView (and DatCon) need to be told to include all signal groups. This includes gps_hw_status.
1606578591991.png

As I mentioned I've not looked too much at this data. It's just guess work that is somewhat dependent on the labels that DJI has given to the data. gps_hw_status:jammingState seems to be a state variable indicating the degree to which the GPS is jammed. Maybe something like 1=normal, 2=kinda jammed, and 3=really jammed. When jammingState (black trace) reaches 3 the sAcc (blue trace, combination of hDOP and pDOP) starts to rise and numSats (red trace) drops to 0
1606579259907.png

Also noted is that these two occurrences were in the same vicinity where the MA2 was facing east. There were another two occurrences in that vicinity but there was no jamming indicated. The difference seems to be that the MA2 was facing west instead of east.
1606579915500.png 1606579948336.pngJammed

1606579991241.png1606580016134.png Not jammed

If anything, it's probably something rather than someone jamming the GPS.
 
Last edited:
Hi BudWalker,
Yes, I am able to extract and display the GPS hw state now. I did not realize I need to select that before loading the dat file.

I highly skeptical right now that at that particular space and time, something jammed GPS signals. It was even not too near the apartments. Looks like 100m+ from ground level and almost 100m from apartment. Could there be something powerful enough at that range, save for some dedicated equipment?

Today I will for the sake of science fly nearby. I might take off right there.
 
Hi BudWalker,
Yes, I am able to extract and display the GPS hw state now. I did not realize I need to select that before loading the dat file.

I highly skeptical right now that at that particular space and time, something jammed GPS signals. It was even not too near the apartments. Looks like 100m+ from ground level and almost 100m from apartment. Could there be something powerful enough at that range, save for some dedicated equipment?

Today I will for the sake of science fly nearby. I might take off right there.
I'm also somewhat skeptical. But, for now it's all we have. Supposing there is some source of GPS interference I suggest you try several laps between W and E
1606658711750.png
The GPS problems only occurred when travelling from W to E and not when travelling from E to W. I suspect that's because the GPS antenna is more exposed to any interference coming from the front of the MA2. You might also try several abrupt start/stops in this interval. Going from E to W an abrupt stop will cause the MA2 to briefly tilt back towards the interference source.
 

Attachments

  • 1606658551562.png
    1606658551562.png
    108.2 KB · Views: 6
Screenshot_20201129-225522.jpg

I did a few laps around today. Even reduced altitude to lower than the apartments. Between 15 to 18 Sats, full navhealth all the way....

Could it be the "KPI" interference ? But that does not happen over a small specific part of the flight..right?
 

Attachments

  • Screenshot_20201129-225522.jpg
    Screenshot_20201129-225522.jpg
    981.7 KB · Views: 7
View attachment 118070

I did a few laps around today. Even reduced altitude to lower than the apartments. Between 15 to 18 Sats, full navhealth all the way....

Could it be the "KPI" interference ? But that does not happen over a small specific part of the flight..right?
Can you post the .DAT?
 
  • Like
Reactions: Thomas B
Could it be the "KPI" interference ? But that does not happen over a small specific part of the flight..right?
The K-index hasn't ever caused GPS issues for drones.
If it did, it wouldn't be for such a short time anyway.
 
View attachment 118070

I did a few laps around today. Even reduced altitude to lower than the apartments. Between 15 to 18 Sats, full navhealth all the way....

Could it be the "KPI" interference ? But that does not happen over a small specific part of the flight..right?
I have to agree - there aren't any of the GPS problems seen in the first flight.

I don't know much about the K-index so can't offer an opinion.
 
@BudWalker here is the txt log file.
Thanks. I had requested this .txt for reasons unrelated to the GPS mystery. Specifically, there is a flyCState that DJI has mysteriously labelled as "other" instead of something like GPS_ATTI, GoHome, etc. From the eventLog stream in the corresponding .DAT it is apparent that you were doing time lapse photography after you had flown the test. Is that right?
 
Last edited:
Lycus Tech Mavic Air 3 Case

DJI Drone Deals

New Threads

Forum statistics

Threads
130,411
Messages
1,552,346
Members
159,418
Latest member
jamoke