GPS horizontal accuracy

[pavloblindnology]

What is the reason of using HDOP value for computing horizontal accuracy of GPS_RAW_INT message instead of using h_acc values contained in the message? HDOP value is a scale factor, so, at least it should be multiplied by some metric accuracy estimation. Otherwise, it gives unrealistically small values.

mavros/mavros/src/plugins/global_position.cpp

Line 175 in 38c6fff

fix->position_covariance[0 + 0] = fix->position_covariance[3 + 1] = std::pow(hdop, 2);

[vooon]

Just because that code done for a very long time before GPS_RAW_INT have extended with that fields.
And note, extension do not work with mavlink v1.0.

[pavloblindnology]

Understood.
Can we use extension values if on mavlink v2.0 and HDOP otherwise?
Also, I'd add some metric accuracy estimation parameter to config file to be scaled by HDOP, VDOP.
Is it feasible?

[TSC21]

Understood.
Can we use extension values if on mavlink v2.0 and HDOP otherwise?
Also, I'd add some metric accuracy estimation parameter to config file to be scaled by HDOP, VDOP.
Is it feasible?

@pavloblindnology that seems reasonable to me. Can you please do a PR with it? Thank you!