Can the altitude AGL update dynamically?

[sar104]

I'm sure the others in this thread will come along to expand on this, but the Mavic resets its barometric altimeter to zero every time you take off. It's very accurate after that.

Yes - while it's true that barometric pressure varies significantly with atmospheric conditions, when used in relative altitude mode it should be good to a couple of meters over the maximum duration of a flight.

That may not be true in stormy weather however, when barometric pressure can change by up to 0.8 kPa/hr, which would be 0.4 kPa over a 30 minute flight. That equates to a 40 meter change in altitude at sea level.

 

[sar104]

correct, but thats not what my reply was to. It was to whether the Mavic uses GPS for any kind of altitude and whether barometric altitude is accurate in a commercial aircraft, of which it is, when properly matched as you increase in altitude and waypoints. The barometric on the Mavic is accurate because the altitudes are relatively small

That is also a factor. The FC almost certainly assumes a standard atmosphere for barometric calculations, which is only an approximation. If the air density is significantly different from the model then the rate of change of pressure with altitude will be incorrect, and so the altitude error will increase approximately linearly with change in altitude.

 

[FoxhallGH]

you can't really call a drone an "aircraft" in this case for pressure altitude monitoring. Accurate pressure monitoring from pitot tubes are only at about every 1000-1500 feet. Since no one ever really flies a drone above half that height, barometer by itself is not going to be accurate. I had to dig through a manual, but the drones do indeed use barometer and GPS combined to maintain an altitude. So it could recalculate based on absolute or MSL, but there are a lot of reasons why you wouldn't (RTH, not accurate with regard to where you are/were/are going)

from manual
View attachment 53317

There is a big difference in sensing barometric altitude - and sensing a change [delta] in barometric altitude! When you get into an aircraft, one of the first things you do is calibrate the altimeter by setting a 'QNH' - which is the local barometric pressure reported by the control tower. If the Mavic is simply sensing it's own 'local' barometric pressure, it can easily tell if it's going up or down and compensate for that - but that 'delta-altitude' is not the same as actual altitude ... In other words, the barometric reading can be used to maintain height/altitude, but is useless to tell you what height the drone is at!

 

[DanMan32]

Useless in determining MSL, but not useless for altitude from launch point within the 30 min of flight.

 

[FoxhallGH]

Useless in determining MSL, but not useless for altitude from launch point within the 30 min of flight.

Correct - because it's setting and using its own reference point ...

 

[Meta4]

Standard GPS, non WAAS corrected, vertical accuracy is plus or minus 9 meters, or about +/- 30 feet. GPS altitude is a calculation based on the height above a surface defined by an ellipsoid that models the earth’s shape at sea level. WG84 is the name of the model and is close to a sphere that bulges slightly at the equator. In some places the model is off by small amounts up to 100 feet, but this is corrected for in the GPS by using a correction database."

the Mavic as with all of the advanced drones use a combination of GPS and Baro for measurements, you can argue all day long, but it is what it is.

And what it (GPS altitude accuracy) is, is woeful.
I gave an example in post #7, there was a supporting post in #11.
Here's what Garmin have to say about the matter:
How accurate is the GPS elevation reading?
GPS heights are based on an ellipsoid (a mathematical representation of the earth's shape), while USGS map elevations are based on a vertical datum tied to the geoid (or what is commonly called mean sea level).
Basically, these are two different systems, although they have a relationship that has been modeled.

The main source of error has to do with the arrangement of the satellite configurations during fix determinations.
The earth blocks out satellites needed to get a good quality vertical measurement.
Once the vertical datum is taken into account, the accuracy permitted by geometry considerations remains less than that of horizontal positions.
It is not uncommon for satellite heights to be off from map elevations by +/- 400 ft.
Use these values with caution when navigating.

DJI understand this and their drones do not use GPS at all for flight altitude data.
Flight altitude data comes 100% from a barometric sensor.

 

[sar104]

And what it (GPS altitude accuracy) is, is woeful.
I gave an example in post #7, there was a supporting post in #11.
Here's what Garmin have to say about the matter:
How accurate is the GPS elevation reading?
GPS heights are based on an ellipsoid (a mathematical representation of the earth's shape), while USGS map elevations are based on a vertical datum tied to the geoid (or what is commonly called mean sea level).
Basically, these are two different systems, although they have a relationship that has been modeled.

The main source of error has to do with the arrangement of the satellite configurations during fix determinations.
The earth blocks out satellites needed to get a good quality vertical measurement.
Once the vertical datum is taken into account, the accuracy permitted by geometry considerations remains less than that of horizontal positions.
It is not uncommon for satellite heights to be off from map elevations by +/- 400 ft.
Use these values with caution when navigating.

DJI understand this and their drones do not use GPS at all for flight altitude data.
Flight altitude data comes 100% from a barometric sensor.

To be fair - that statement from Garmin is very old and accuracy has improved. It's relatively straightforward to show that the expected vertical accuracy is around a factor of two worse than horizontal. The other factor to consider is the variation of the WGS84 ellipsoid from theoretical MSL based on the old earth geoid. Off the east coast of Australia the difference is particularly large - around 44 meters. Add tidal effects and that probably accounts for a significant amount of the 90 meter discrepancy that you noted, rather than it all being random error in the GPS altitude.

 

[Meta4]

To be fair - that statement from Garmin is very old and accuracy has improved. It's relatively straightforward to show that the expected vertical accuracy is around a factor of two worse than horizontal. The other factor to consider is the variation of the WGS84 ellipsoid from theoretical MSL based on the old earth geoid. Off the east coast of Australia the difference is particularly large - around 44 meters. Add tidal effects and that probably accounts for a significant amount of the 90 meter discrepancy that you noted, rather than it all being random error in the GPS altitude.

But if I look at the GPS altitude in exif info from photos taken at different times, I see the error is all over the place.
It's nothing like a constant error.

 

[sar104]

But if I look at the GPS altitude in exif info from photos taken at different times, I see the error is all over the place.
It's nothing like a constant error.

That is a bit unexpected then, if it is random and consistently more than twice the horizontal error. The variation that I see is really quite small - typically less than 15 m.

 

[droneworx]

And what it (GPS altitude accuracy) is, is woeful.
I gave an example in post #7, there was a supporting post in #11.
Here's what Garmin have to say about the matter:
How accurate is the GPS elevation reading?
GPS heights are based on an ellipsoid (a mathematical representation of the earth's shape), while USGS map elevations are based on a vertical datum tied to the geoid (or what is commonly called mean sea level).
Basically, these are two different systems, although they have a relationship that has been modeled.

The main source of error has to do with the arrangement of the satellite configurations during fix determinations.
The earth blocks out satellites needed to get a good quality vertical measurement.
Once the vertical datum is taken into account, the accuracy permitted by geometry considerations remains less than that of horizontal positions.
It is not uncommon for satellite heights to be off from map elevations by +/- 400 ft.
Use these values with caution when navigating.

DJI understand this and their drones do not use GPS at all for flight altitude data.
Flight altitude data comes 100% from a barometric sensor.

this is a photo of a fairly modern (couple years) handheld GPS unit that I can guarantee is only using GPS to determine altitude, and the exact same spot on a top map. Essentially the GPS determined elevation is pretty spot on.

 

[droneworx]

And what it (GPS altitude accuracy) is, is woeful.

DJI understand this and their drones do not use GPS at all for flight altitude data.
Flight altitude data comes 100% from a barometric sensor.

we can argue this all day, but I can tell you I know for a FACT that GPS is used to assist in determine altitude, nothing is the result of a single setting. If you open up your images and edit the tags, you will see a exif tags that reference GPS pulled altitude and position.

GPS Version ID :
GPS Latitude Ref :
GPS Longitude Ref :
GPS Altitude Ref
GPS Altitude :
GPS Latitude :
GPS Longitude :
GPS Position :

 

[DanMan32]

That doesn't mean the FC uses it for flight info.

 

[sar104]

we can argue this all day, but I can tell you I know for a FACT that GPS is used to assist in determine altitude, nothing is the result of a single setting. If you open up your images and edit the tags, you will see a exif tags that reference GPS pulled altitude and position.

GPS Version ID :
GPS Latitude Ref :
GPS Longitude Ref :
GPS Altitude Ref
GPS Altitude :
GPS Latitude :
GPS Longitude :
GPS Position :

Yes - both GPS and barometric relative altitude are in the EXIF data, but I haven't seen any evidence that the GPS vertical data are used in the sensor fusion process. That said, I'm not sure that we can rule it out, even if the barometric data are the primary or weighted more heavily.

 

[Meta4]

this is a photo of a fairly modern (couple years) handheld GPS unit that I can guarantee is only using GPS to determine altitude, and the exact same spot on a top map.
Essentially the GPS determined elevation is pretty spot on.

How can you guarantee it was only using GPS for altitude?
What model is the unit?
Is it WAAS enabled and was that in North America?

we can argue this all day, but I can tell you I know for a FACT that GPS is used to assist in determine altitude, nothing is the result of a single setting. If you open up your images and edit the tags, you will see a exif tags that reference GPS pulled altitude and position.

Why do you say that you know this for a fact?
What evidence is there?

I'm well aware what's in the exif info and I know that DJI have put GPS altitude data in there since about a year into the run of the Phantom 3 series.
But having the GPS altitude data stored in exif doesn't tell you anything about the Phantom using GPS data at all for flight.
That's simply camera data.
If the exif info was your proof, it's proved nothing.
I showed you the exif info for one of my pix up in post #7.
How much use would that GPS altitude data be for flight purposes?

 

[FoxhallGH]

Considering that the US Military put its GPS system in place with the intention of using it to deliver bombs and missiles onto targets - you can be pretty sure that the system is going to be pretty accurate in all 3 dimensions!!

 

[Meta4]

Considering that the US Military put its GPS system in place with the intention of using it to deliver bombs and missiles onto targets - you can be pretty sure that the system is going to be pretty accurate in all 3 dimensions!!

Look at post #7 and tell me that again

 

[FoxhallGH]

Look at post #7 and tell me that again

Look at all the holes in Afghanistan! :-(

 

[sar104]

Considering that the US Military put its GPS system in place with the intention of using it to deliver bombs and missiles onto targets - you can be pretty sure that the system is going to be pretty accurate in all 3 dimensions!!

Don't forget that the military receivers decode both the L1 and L2 frequencies, which allows them to do much better correction for the variable ionospheric delays that are the major source of error. So that's not a good comparison. And simply from geometric considerations, as I mentioned earlier, one can show that vertical accuracy, on average, should be around a factor of two worse than horizontal.

 

[droneworx]

How can you guarantee it was only using GPS for altitude?
What model is the unit?
Is it WAAS enabled and was that in North America?


Why do you say that you know this for a fact?
What evidence is there?

I'm well aware what's in the exif info and I know that DJI have put GPS altitude data in there since about a year into the run of the Phantom 3 series.
But having the GPS altitude data stored in exif doesn't tell you anything about the Phantom using GPS data at all for flight.
That's simply camera data.
If the exif info was your proof, it's proved nothing.
I showed you the exif info for one of my pix up in post #7.
How much use would that GPS altitude data be for flight purposes?

GPS Elevation Accuracy Test: Smartphone Apps vs. Dedicated GPS -
"Hardware
Many high-end GPS units feature a barometric altimeter in an effort to improve elevation accuracy, but truthfully it doesn’t make much difference. As one GPS company rep confirmed, a barometric altimeter is pretty useless unless it’s regularly calibrated. That’s because it uses atmospheric pressure to determine your altitude–but atmospheric pressure changes not just with elevation, but also with the weather. Only one of the top 4 units listed above features a barometric altimeter. The worst-performing unit includes a barometric altimeter as well. And FYI, our test happened a few hours after a massive storm system rolled through."

while I dont remember the exact model number, being a few years old, I can tell you it was a 59 dollars special, not a high end one. But even the "high end" GPS units use GPS with baro assist only. But again its irrelevant, because unless you have an aircraft grade pitot tube fed barometric system, they are not accurate enough to determine altitude alone, as was repeated by the GPS company rep above.

Im not understanding why your so deadset in proving your point on this? The reason I say I know for a fact the Mavic does, is 1. why have it available to the unit to record elevation for pictures, but it not be used for the unit? that's just stupid. 2. During one of my phantom (not Mavic, but why would the system be different in how they measure systems, becuase they both use DJI4 it has to read the same sensors and settings) repairs, I broke the barometric daughter board by accident, and GPS readings were still available. 3. GPS is more accurate.

 

[sar104]

GPS Elevation Accuracy Test: Smartphone Apps vs. Dedicated GPS -
"Hardware
Many high-end GPS units feature a barometric altimeter in an effort to improve elevation accuracy, but truthfully it doesn’t make much difference. As one GPS company rep confirmed, a barometric altimeter is pretty useless unless it’s regularly calibrated. That’s because it uses atmospheric pressure to determine your altitude–but atmospheric pressure changes not just with elevation, but also with the weather. Only one of the top 4 units listed above features a barometric altimeter. The worst-performing unit includes a barometric altimeter as well. And FYI, our test happened a few hours after a massive storm system rolled through."

while I dont remember the exact model number, being a few years old, I can tell you it was a 59 dollars special, not a high end one. But even the "high end" GPS units use GPS with baro assist only. But again its irrelevant, because unless you have an aircraft grade pitot tube fed barometric system, they are not accurate enough to determine altitude alone, as was repeated by the GPS company rep above.

Im not understanding why your so deadset in proving your point on this? The reason I say I know for a fact the Mavic does, is 1. why have it available to the unit to record elevation for pictures, but it not be used for the unit? that's just stupid. 2. During one of my phantom (not Mavic, but why would the system be different in how they measure systems, becuase they both use DJI4 it has to read the same sensors and settings) repairs, I broke the barometric daughter board by accident, and GPS readings were still available. 3. GPS is more accurate.

None of that is actually conclusive. The aircraft has GPS, and thus a 3D GPS solution, so the altitude data are going to be there anyway. The question is whether the sensor fusion algorithm (Kalman filter or similar) uses the altitude data. And it is a bit more complicated even than that because the GPS supplies 3D position and 3D velocity data independently. We know that horizontal position data are fed into the fusion algorithm. I think that the consensus is that horizontal velocity data are also used. In principle there is no reason not to use both vertical datasets too.

The argument over GPS vs. barometric accuracy also needs to be put into context. Neither has very good absolute accuracy - barometric because atmospheric pressure varies and GPS because the vertical solution is worse than the horizontal and because the reference ellipsoid deviates from the MSL geoid. Barometric relative altitude is generally pretty good over short time periods because, most of the time, local atmospheric pressure is not changing fast enough to make a big difference over a 30 minute flight. GPS-derived altitude errors from the ellipsoid/geoid issue don't change at all with time, while random errors do change as the constellation moves around, and at a rate faster than atmospheric pressure. The relative temporal rates of change are probably why barometric is better over the course of an entire flight but, for instantaneous detection of altitude changes, GPS, via either vertical velocity or position data, is quite good, and certainly could be used in sensor fusion. The issue is that, as far as I'm aware, it has not been confirmed.