RE: nRF9160 GNSS accuracy

Albrecht Markus Schellenberger — Wed, 27 Apr 2022 11:21:15 GMT

Hello again Arik,

[quote user=""]Is the nRF9160 GPS implementation using Doppler for velocity measurement or is it based on position or something else?[/quote]

The GPS receiver always has to deal with movement: Movement of the satellites, movement of the Earth, movement of the user, movement (slipping) of the clocks in SV and equipment.

All of the above (and more) have to be figured in when calculating a position, and as a side product, also the user movement vector can be calculated in pretty much the same way as the position is calculated. So it is not calculated directly from a measured Doppler, nor is it calculated from (position B - position A)/time, it is actually solved from a complex equation system.

Similarly to the position, the calculated velocity varies slightly due to a number of circumstances (number of SVs, DOP, ionospheric delay, multipaths, noise, ...), therefore velocity is also included in a Kalman filter state vector and thereby improved (averaged).

The notion that velocity is measured (calculated) more accurately than position is probably just an illusion stemming from different reference frames: As a purely fictional example, we think that a 10 meter position error is a big error, whereas 82 km/h (instead of a correct 80 km/h) is pretty much a negligible error (and looks even smaller when expressed in m/s).

Regards,

Markus

RE: nRF9160 GNSS accuracy

Albrecht Markus Schellenberger — Tue, 26 Apr 2022 11:26:20 GMT

Hello Arik,

I have to check this with our modem team. I will come back to you as soon as I have more information.

Regards,

Markus