I'm told you guys may be the one that might be able to help. Took my drone up to take some pics of fireworks a few nights ago. GPS signal dropped out and whenever it came back, it appears there are multiple error messages in the log of, "GPS Position NoMatch". I tried to get it back to me via attitude mode but the wind was picking up and I quickly lost visual sight of it. Tried to find my way back using the video feed but it was dark and couldn't orient myself. I'm hoping it is possible to recreate the flight path without the GPS by using the vertical, horizontal and other information. At the very least, I'm hoping to get an idea of what direction it was headed when the signal with the remote was lost. You can reach me on the forum but more quickly at [email protected].
Lost GPS and Lost my P4P
- Thread starter thomasbode1980
- Start date
GPS Position no match is seen quite often and is rarely an issue. If you want assistance we will need to see the logs from your device. If you use Airdata, share the log link and make it fully sharable such that the .csv file can be downloaded. OR, upload your device .txt log to the link below and share a link back here to that.
DJI Flight Log Viewer
Start by uploading your flight log here: DJI Flight Log Viewer. Once that's complete, come back and post a link.
There are a number of pilots here that may be able to help determine what happened.
There are a number of pilots here that may be able to help determine what happened.
I've seen the GPS Position Nomatch before and I've heard it's a false alarm. Nevertheless, you completely lost all satellites. This one's out of my league. More questions coming, for certain.
Yeah, that's what really threw me for a loop. All the satellites dropping out completely seems really questionable but I don't honestly know if that's typical or not. I had just received it back from DJI after a coworker crashing it into a power pole so maybe the GPS unit was damaged?
Losing GPS like that is not typical. My guess is hardware failure, possibly a result of the previous crash. DJI could have missed that as it appears to be intermittent. Hopefully your name and phone number is on the AC. Reminder to myself to do just that!
Yeah..... you'd think I would have done that but.... The location I wrote it on previously was replaced and I didn't think about that until after the fact.
So, is it a lost cause to try and figure out where it went without the the GPS info?
I agree with @LarBear360 . This is not typical at all. From this it appears to be either a faulty connection, and/or a bad GPS antenna. The intermittency noted until complete loss would point to something similar. As for your original question non re-creating the flight path, this can be done but with great difficulty. Really the only way IMO would be to use the Yaw, Speed and Altitude data to very roughly simulate this. However, being the fact that you were flying in ATTI and all over the place in direction, heading, speed, altitude and other categories this would take some time to figure out even a rough location. Even then it would only be a WAG. Below is a chart of your GPS loss. Still looking at the other data. But I do have a question. When you lost GPS completely at approx 2:15 seconds and only 120ft away at only 63 feet in altitude, why did you not simply quickly lose altitude. You could (And should) have landed immediately. Included with the chart is the GPS loss point.
Other than what I mentioned regarding the difficulty of attempting to simulate the direction of travel, no. There is really no other way, and doing it that way would be very rough and not very accurate. What would have to be done in detail, as I mentioned is to utilize the Yaw angles, speed, and altitude data and collect these values every 10 seconds ( Or less ) for the duration of the log for a very rough estimate. That said, this would not include any wind factors which would throw the estimate off even farther.
Is that approximation something a retail place would be able to do? I'm thinking someone like Fortress UAV in Plano, TX which is in the DFW, TX area about an hour from where I live. I'm still trying to see if I can locate it before I try and fight with DJI that I think I am entitled to a replacement due to manufacturer defect. Just trying to do my due diligence.
It is possible, however it could be quite pricey, depending on the complexity. I did do a bit myself, but it became quite time consuming. This requires going through the .csv file, line by line and extracting the data that way starting at the last GPS point, then using the airspeed and directional data very roughly calculating the distance traveled at any particular direction. As I mentioned, you were spinning and drifing all over the place in many different directions, occasssionally at full elevator and other times at basically a hover. This is why it would be so difficult to even get a close approximation not to mention the fact that once complete signal loss occurs there is no more data to go by and the aircraft could have auto landed most anywhere at critical battery.
Sorry to hear about your loss. Hopefully someone will find it and return it to you using your contact information on the Phantom. I wish you success getting DJI to replace it.
It's relatively straightforward to use the IMU velocity data to determine position - you just integrate the Vx (north) and Vy (east) velocity components with respect to time. The complication is that without any GPS data the sensor fusion scheme that corrects for accelerometer and rate gyro drift and bias has no absolute data to work with, and so the errors will accumulate, in this case for the final 291 s of the flight. That aside, the IMU data yield the following position data when integrated, with the GPS location data for the first 135 s of the flight included (in green) for comparison:
The GPS data are poor even for the first part of the flight, hence the somewhat tepid agreement even then. If we ignore the unknown accumulated bias and drift errors, the final recorded point is 98 m south and 51 m east of the home point. On the ground that shows as:
As a sanity check, transforming the frame of reference to convert Vx and Vy to velocity forwards and right with respect to the aircraft, and comparing velocity forwards with pitch (should be inversely correlated), we get:
That has some obvious issues, especially between 200 and 300 s where the significant negative pitch is not producing much of a forwards velocity, so I'm not sure how much we can trust the IMU velocities once sensor fusion failed. Probably worth a look though.
I had same problem last week, never had a gps problem ever before, so I sat down and thought ‘what have I changed with my set up since last flight?’
Answer: I started using a new lightening cord, it’s a nice short 300mm cord with the Apple lightning end at a right angle which I got on eBay
I removed it and tried the old apple cord and gps problem was gone. Don’t ask me why, I just telling you as it is
