Mavic 2 Pro RTH not accurate

gnirtS · Aug 31, 2018

GringoLoco said:
Well, at least I don't have to buy a landing pad now.

You do...But instead of one thats 1m across you now need a 10m one.

offtheback · Sep 15, 2018

Kilrah said:
5-8ft is very accurate actually for GPS... hard to understand people's expectations sometimes.

Expectations are based on an understanding of what the RTH promises.I too am experiencing issues with the drone returning 10-30 feet away.I'm looking into it.

ruggb · Nov 28, 2018

I have just done my first flight (M2zoom) with 01.00.0200 firmware which supposedly has precision landing. It did not want to land from where it took off. I made sure the home point was set b4 takeoff. When I cancelled and retried. it went even farther off.

Toddmoone · Nov 28, 2018

Kilrah said:
5-8ft is very accurate actually for GPS... hard to understand people's expectations sometimes.

He said he expects the higher accuracy of the optical sensor. He didn't say that was bad for GPS.

ruggb · Nov 28, 2018

5-8 ft is NOT good for GPS. My P3S landed within a 3ft radius-4' max- all the time. Other M pros land within 2' consistently. My 1st flight of my M2Zoom wanted to land 6 ft from home. Something is not right.

DanMan32 · Nov 28, 2018

ruggb said:
I have just done my first flight (M2zoom) with 01.00.0200 firmware which supposedly has precision landing. It did not want to land from where it took off. I made sure the home point was set b4 takeoff. When I cancelled and retried. it went even farther off.

How did you take off? You have to take off straight up at least 10ft (might be 10m).
If you make certain flight deviations during RTH, it will cancel PL even if it doesn't cancel RTH.

ruggb · Nov 28, 2018

I went straight up to 120M. When it started landing it was far enough off that I cancelled it and moved it, then did RTH again. It was still off. When I moved it again farther away it started landing where I moved it to. Battery was still not critical. I'll try again when it gets warmer outside.

sar104 · Nov 28, 2018

ruggb said:
5-8 ft is NOT good for GPS. My P3S landed within a 3ft radius-4' max- all the time. Other M pros land within 2' consistently. My 1st flight of my M2Zoom wanted to land 6 ft from home. Something is not right.

Actually 5 - 8 ft is well within normal accuracy for high-quality civilian GPS (single frequency).

http://www.nstb.tc.faa.gov/reports/PAN96_0117.pdf

ruggb · Nov 28, 2018

It depends on how many sats you are tracking. My Garmin Handheld will get to 3'-5' with 9 -12 sats. Bottom line is my P3S was very consistent within 3' so a Mavic 2 should be at least as good.

sar104 · Nov 28, 2018

ruggb said:
It depends on how many sats you are tracking. My Garmin Handheld will get to 3'-5' with 9 -12 sats. Bottom line is my P3S was very consistent within 3' so a Mavic 2 should be at least as good.

No it doesn't depend on that, and it really doesn't matter how much anecdotal evidence you have - GPS is not consistently that accurate, even in the US with WAAS. Did you read the report that I linked to? Or if you don't like that - just do a search - you will find plenty of documentation and evidence.

Phoenix Drone · Nov 28, 2018

82 flights with Mavic Pro Platinum, always about 6” away from exact takeoff point using a landing pad. It is a precision takeoff with RTH.

Meta4 · Nov 29, 2018

ruggb said:
It depends on how many sats you are tracking. My Garmin Handheld will get to 3'-5' with 9 -12 sats. Bottom line is my P3S was very consistent within 3' so a Mavic 2 should be at least as good.

No consumer GPS can consistently get you within 3 ft.
Try reading this from the people that run the GPS system: GPS.gov: GPS Accuracy
Or look at this graph from them to see the variable inaccuracy that's inherent in GPS.

ruggb · Nov 29, 2018

So the majority of samples are under 3' with a rapid drop approx 4'.
from FAA site

0.7m is not too bad, wouldn't you say?
Bottom line is still that my P3S consistently landed within 3' from home point.
Since I only flew my M2 once, I can't make a conclusion, but something wasn't right on that flight.

sar104 · Nov 29, 2018

ruggb said:
So the majority of samples are under 3' with a rapid drop approx 4'.
from FAA site

View attachment 54584

0.7m is not too bad, wouldn't you say?
Bottom line is still that my P3S consistently landed within 3' from home point.
Since I only flew my M2 once, I can't make a conclusion, but something wasn't right on that flight.

Your P3S definitely doesn't use LPV. 2.9 m is the relevant number.

Phoenix Drone · Nov 29, 2018

Optical position is used in combination with GPS. GPS to get in general vicinity considering WASS accuracy. Don’t know how many tests you guys are doing with landing pad, but the contrast sensed by the downward camera is greatly increased. As I mentioned, I am getting very close with the M1P.

DanMan32 · Nov 29, 2018

M2 didn't have PL at all until 11/22

ruggb · Nov 29, 2018

I updated it to .0200 b4 my flight. Maybe I just do not know how to use it.
Must I have a pad to make it work? - my P3S did not and did not have sensors. The M2 was far enough off long b4 it could see the pad - if there was one. So it has to be a GPS issue.
It is 51° now so I guess I will try it again.

gfer66 · Nov 29, 2018

ruggb said:
I updated it to .0200 b4 my flight. Maybe I just do not know how to use it.
Must I have a pad to make it work? - my P3S did not and did not have sensors. The M2 was far enough off long b4 it could see the pad - if there was one. So it has to be a GPS issue.
It is 51° now so I guess I will try it again.

In a YouTube video a guy tried to land his M2Z using his pad and, in a second test, a circle of stones. The drone landed on the pad but failed on the stones. So I guess Phoenix Drone is right.

Meta4 · Nov 29, 2018

ruggb said:
I have just done my first flight (M2zoom) with 01.00.0200 firmware which supposedly has precision landing. It did not want to land from where it took off. I made sure the home point was set b4 takeoff. When I cancelled and retried. it went even farther off.

ruggb said:
I updated it to .0200 b4 my flight. Maybe I just do not know how to use it.
Must I have a pad to make it work? - my P3S did not and did not have sensors. The M2 was far enough off long b4 it could see the pad - if there was one. So it has to be a GPS issue.
It is 51° now so I guess I will try it again.

Precision Landing isn't magic and needs certain conditions for it to work.
Note particularly B,D & E

Precision Landing
The Mavic 2 Pro/Zoom automatically scans and attempts to match the terrain features underneath during Return to Home. When the current terrain matches Home Point terrain, the Mavic 2 Pro/Zoom
will start landing.
The DJI GO 4 app will show a terrain feature mismatch prompt if matching fails

Precision Landing performance is subject to the following conditions:
a. The Home Point must be recorded upon takeoff and must not be changed during flight, otherwise the aircraft will have no record of the Home Point’s terrain features.
b. During takeoff the aircraft must ascend vertically 7 m before moving horizontally.
c. The Home Point terrain features must remain largely unchanged.
d. The Home Point terrain features must be sufficiently distinctive.
e. The lighting conditions must not be too light or too dark.

The following actions are available during Precision Landing:
a. Throttle down to accelerate landing.
b. Move the control sticks in any other direction to stop Precision Landing. The Mavic 2 Pro/Zoom will descend vertically after the control sticks are released

gfer66 said:
In a YouTube video a guy tried to land his M2Z using his pad and, in a second test, a circle of stones. The drone landed on the pad but failed on the stones. So I guess Phoenix Drone is right.

Unless that video was made in the last week and with the latest firmware, his M2 didn't have Precision Landing and would have been using GPS only.
If PL was available, points B, D & E above must also be met.

gfer66 · Nov 29, 2018

Unless that video was made in the last week and with the latest firmware, his M2 didn't have Precision Landing and would have been using GPS only.
If PL was available, points B, D & E above must also be met.

The video was published yesterday. It was originally to test new firmware's Precision Landing and compare it with Mavic Pro's PL, and you can see the results, but when the guy was testing a group of youngsters tried to assault him, so in the video he talks more about the assault than the test.

I didn't provide a link to the video because of this and because it's in spanish (the youtuber is colombian), but here it is...