GPS Accuracy Calculation

[bc_metco]

in the gps example, there is a struct

nrf_modem_gnss_pvt_data_frame

in which contains the accuracy in meters, by documentation definition, it's /** Accuracy (2D 1-sigma) in meters. */

What is the formula to calculate this accuracy value? I'm confused when we mention 1-sigma (which is the standard deviation), usually the sigma is calculated by average value of a couple of samples. however, in GPS coordinates, this value is calculated even when the 1 fix data is received. if we use average data, then the initial value of sigma should always be zero, apparently, this is not the case. 

[Albrecht Markus Schellenberger]

Hello Bo,

What is the formula to calculate this accuracy value? I'm confused when we mention 1-sigma (which is the standard deviation), usually the sigma is calculated by average value of a couple of samples. however, in GPS coordinates, this value is calculated even when the 1 fix data is received. if we use average data, then the initial value of sigma should always be zero, apparently, this is not the case. 

The provided accuracy in the PVT information is not calculated as the standard deviation of the actual fixes. It is calculated from the DOP and the CNOs for each satellite. So, it is an estimate of the accuracy, not an actual measurement. This makes it possible to give an accuracy for the first (possibly single) fix already.

Regards,

Markus

[bc_metco]

could you share the formula or algorithm of the calculation based on DOP and CNO?

[Albrecht Markus Schellenberger]

bc_metco said:

could you share the formula or algorithm of the calculation based on DOP and CNO?

The formula is quite complex,so it is not that straight forward I’m afraid. But you should be able to find what you need using Google, like e.g. this research document.

Regards,

Markus