Understanding information in flight logs

[Phantom3User]

Hi All,

I'm sort of a newbie and just figured out that it is possible to analyze Phantom 3 flight records. I was using datcon to do the same. I was wondering if anyone knew where some of the information in these logs comes from? Specifically, I was wondering if anyone knew how the velocity is measured, and what magX, magY and magZ stand for? Any help would be appreceated! Also, does anyone have any idea of what the best way to, re-generate the actual flight coordinates would be, (for an indoor flight with no GPS) using the flight log?

 

[BudWalker]

Hi All,

I'm sort of a newbie and just figured out that it is possible to analyze Phantom 3 flight records. I was using datcon to do the same. I was wondering if anyone knew where some of the information in these logs comes from? Specifically, I was wondering if anyone knew how the velocity is measured, and what magX, magY and magZ stand for? Any help would be appreceated! Also, does anyone have any idea of what the best way to, re-generate the actual flight coordinates would be, (for an indoor flight with no GPS) using the flight log?

Documentation on the .csv columns can be found here. magX, magY, magZ are the magnetometer values; they are the values you see when you look at the compass page in the DJI Go App. velN, velE, and VelD come from the P3 itself. I'm not sure how it computes these values. vel is computed by DatCon from velN, velE, and velD.

 

[Phantom3User]

Documentation on the .csv columns can be found here. magX, magY, magZ are the magnetometer values; they are the values you see when you look at the compass page in the DJI Go App. velN, velE, and VelD come from the P3 itself. I'm not sure how it computes these values. vel is computed by DatCon from velN, velE, and velD.

Thanks! Do you have have any idea of what the best way to, re-generate the actual flight coordinates might be? Also, would you happen to know if there are there any other places I could search to figure out how P3 is computing the velocity?

 

[BudWalker]

Thanks! Do you have have any idea of what the best way to, re-generate the actual flight coordinates might be? Also, would you happen to know if there are there any other places I could search to figure out how P3 is computing the velocity?

Not sure what you mean by "actual flight coordinates ". The lat-long coordinates you see in the .csv from DatCon come from the flight controller.

If you find more info on how the P3 is computing it's velocity lemme know. I suspect that it is from successive GPS values. But, you may have noticed that the .csv has a field called velGPS-velH(M/S). This is computed by DatCon using successive GPS values and isn't always 0.0. I.e. there is sometimes a difference between DatCon's velocity computed from successive GPS values and the P3's velocity. Never the less, I think the P3 uses successive GPS values to compute velocity. The difference in velocities are probably caused by sampling differences and/or numerical routine differences.

 

[Phantom3User]

Not sure what you mean by "actual flight coordinates ". The lat-long coordinates you see in the .csv from DatCon come from the flight controller.

If you find more info on how the P3 is computing it's velocity lemme know. I suspect that it is from successive GPS values. But, you may have noticed that the .csv has a field called velGPS-velH(M/S). This is computed by DatCon using successive GPS values and isn't always 0.0. I.e. there is sometimes a difference between DatCon's velocity computed from successive GPS values and the P3's velocity. Never the less, I think the P3 uses successive GPS values to compute velocity. The difference in velocities are probably caused by sampling differences and/or numerical routine differences.

The thing is I needed this information for indoor flights without GPS. So I was wondering if there was any numerical way to compute coordinates using other information.

 

[BudWalker]

The thing is I needed this information for indoor flights without GPS. So I was wondering if there was any numerical way to compute coordinates using other information.

Not that I know of.

 

[Phantom3User]

Not that I know of.

I've been digging a little more and I found this in the SDK documentation Flight Controller - DJI Mobile SDK Documentation. It says "Downward facing cameras can help determine relative position and velocity more accurately than consumer satellite positioning systems. They can also be used to hover in GPS denied environments such as inside buildings". It may then be the case that these velocity measurements are being computed by the vision positioning system. Any thoughts on whether this is a plausible explanation?

 

[BudWalker]

I've been digging a little more and I found this in the SDK documentation Flight Controller - DJI Mobile SDK Documentation. It says "Downward facing cameras can help determine relative position and velocity more accurately than consumer satellite positioning systems. They can also be used to hover in GPS denied environments such as inside buildings". It may then be the case that these velocity measurements are being computed by the vision positioning system. Any thoughts on whether this is a plausible explanation?

The P3 doesn't have the downward camera. Just an ultrasonic sensor. But, even if you switched to a Mavic, P4, or P4Pro it would be kinda tough. since DatCon doesn't know anything about these vision systems.