Ambient Wind Reported?

[jclarkw]

I've been told that newer DJI drones, including the Mavic Air 3, report ambient wind speed, direction, and more in their log files! (For example, the latest versions of Mike Singer's "Flight Reader" software list the following 7 new columns of flight data in their log-file reports: WEATHER.windDirection, WEATHER.windRelativeDirection, WEATHER.windSpeed [MPH], WEATHER.maxWindSpeed [MPH], WEATHER.windStrength, WEATHER.isFacingWind, and WEATHER.isFlyingIntoWind.) So far there seems to be nothing about this new feature on the DJI Web site. Questions: 1) Has anyone already used these data? (I can imagine many applications including the field of boundary-layer meteorology.) 2) Is this information unique at every data sample (every 1/10 s for my old Mavic Air 2, which does not report these data), or (more likely) what is the averaging period for these readings? 3) What kind of accuracy is achieved in these data? Any information on this greatly appreciated! -- jclarkw

 

[Suren]

Airdata has been giving those wind speeds, directions etc for a while now.

 

[Suren]

 

[Yorkshire_Pud]

I've been told that newer DJI drones, including the Mavic Air 3, report ambient wind speed, direction, and more in their log files!

Excellent, just looked at a download Air 3 'crash' log from 2024-04-23 and weather is in the PH csv too.

 

[jclarkw]

Airdata has been giving those wind speeds, directions etc for a while now.

True, and they even appear there for the old Mavic Air 2, therefore not courtesy of DJI as with the newer models. However, for the Air 2 they appear only sporadically and don't appear to make much sense.

I quizzed Airdata about them but didn't get any of my questions answered.

What is your evaluation of the wind data, and from what model(s)? Do you have any sample time series of these 7 parameters from the Air 3 that you would be willing to share? -- jclarkw

 

[Suren]

I tend to find them quite accurate. This is from my Air 3. Below is a link to my last flight from it, happy to share it with you. Let me know if you need more.

 

[Suren]

 

[Yorkshire_Pud]

What is your evaluation of the wind data, and from what model(s)? Do you have any sample time series of these 7 parameters from the Air 3 that you would be willing to share? -- jclarkw

@sar104 etc. can possibly shoot holes in this idea but you might try the following.

On a truly windless day fly a set of straight line, only-forwards runs which are of sufficient length that the drone achieves steady speed in each run.
Do this at various speeds from 0 to full speed and in all three flight modes.

When trying to achieve intermediate speeds you might try using the cruise control or, if it is possible, adjusting the maximum speed setting in the app. I might have seen that the latter is possible in at least one of the recent DJI drone models.
Repeat the above for backwards flight and each sideways flight.

Once you have a reasonable set of various speeds runs, for each direction of travel, completed, look at the flight log/s and plot, for each direction of travel, a graph of ground speed vs drone angle.

You could then compare those graphs with logs from windy flights, look at the ground speeds for reasonably stable drone angles. The difference between the ground speeds for a given drone angle is, to my mind due, to wind speed and you could compare that value with the wind speed shown in the weather column of the log.

I realise that getting a steady drone angle on a windy day will probably be difficult and you might have to 'average' drone angles.
You can see the drone tilt 'live' whilst flying via the attitude indicator.

 

[jclarkw]

@sar104 etc. can possibly shoot holes in this idea but you might try the following.

On a truly windless day fly a set of straight line, only-forwards runs which are of sufficient length that the drone achieves steady speed in each run.
Do this at various speeds from 0 to full speed and in all three flight modes.

When trying to achieve intermediate speeds you might try using the cruise control or, if it is possible, adjusting the maximum speed setting in the app. I might have seen that the latter is possible in at least one of the recent DJI drone models.
Repeat the above for backwards flight and each sideways flight.

Once you have a reasonable set of various speeds runs, for each direction of travel, completed, look at the flight log/s and plot, for each direction of travel, a graph of ground speed vs drone angle.

You could then compare those graphs with logs from windy flights, looking at the ground speeds for reasonably stable drone angles. The difference between the ground speeds fo a given drone angle is, to my mind due, to wind speed and you could compare that value with the wind speed shown in the log.

I realise that getting a steady drone angle on a windy day will probably be difficult and you might have to 'average' drone angles.
You can see the drone tilt 'live' whilst flying via the attitude indicator.

Interesting, if tedious, idea that sounds as though it would work. Thanks -- jclarkw

 

[jclarkw]

I tend to find them quite accurate. This is from my Air 3. Below is a link to my last flight from it, happy to share it with you. Let me know if you need more.

Thanks! -- With the "original" version I can process into Flight Reader and proceed from there. -- jclarkw

 

[Yorkshire_Pud]

tedious,

Indeed but it is still flying the drone lol.

 

[eseex]

labdien , vai kāds ir mēģinājis sapārot air3 goggles integra ar rc-n2

 

[msinger]

Airdata has been giving those wind speeds, directions etc for a while now.

To clarify, Airdata has been calculating their own wind speeds/directions. My flight log viewers are actually extracting the wind data DJI stores in the TXT flight logs.

In either case though, both are calculated since DJI drones don't have instruments onboard to record that data. I'm not sure how DJI is calculating their data.

 

[Droning on and on...]

In either case though, both are calculated since DJI drones don't have instruments onboard to record that data. I'm not sure how DJI is calculating their data.

True.

Also true is that with GPS data and roll/pitch tilt angles, a pretty accurate wind speed and direction can be calculated.

I don't know the actual accuracy, but I'd bet it's as good as within a few mph.

 

[jclarkw]

Also true is that with GPS data and roll/pitch tilt angles, a pretty accurate wind speed and direction can be calculated.

I'm sure that's true, but I've been too lazy to attempt a derivation of the relevant equations. (In particular, the conversions between the A/C coordinate system and true [north, east, up] coordinates make my head spin, and the effect of acceleration/deceleration on the pitch/roll/tilt angles would have to be removed unless data during those conditions were flagged as probably bad!)

Have you ever seen these equations? Regards -- jclarkw

 

[msinger]

Have you ever seen these equations?

I had no luck finding anything like that. If it exists though, the size, weight, etc. of the drone must have some kind of impact on those equations.

 

[surfbird]

Yes it has internal wind calculation.
In MDSK5 its possible to read out these values. They are even shown in the example atti-widget.
It's pretty accurate. Wind direction is only in SW,S,SE and so on. And the direction is actually 180 wrong.Tested with mini3.

DJI Mobile SDK Documentation

 

[Droning on and on...]

I'm sure that's true, but I've been too lazy to attempt a derivation of the relevant equations. (In particular, the conversions between the A/C coordinate system and true [north, east, up] coordinates make my head spin, and the effect of acceleration/deceleration on the pitch/roll/tilt angles would have to be removed unless data during those conditions were flagged as probably bad!)

Have you ever seen these equations? Regards -- jclarkw

Heck no!

You're right, it's complicated. And as @msinger notes, there are other factors.

I was only pointing out it can be done.