Does anybody know if the latest dual band GPS watches form Coros and Garmin actually use the dual frequencies to improve the Doppler speed calculations or is it only used to improve positional accuracy?
The primary purpose of dual frequency receivers is to model and remove the ionospheric delay, which improves the pseudorange observable for each satellite and thus improves positional accuracy.
Some of the newer signals are also broadcast at a higher power, making them easier to acquire and track. The higher chipping rate of some of the newer signals will only be of benefit to the pseudorange observable afaik.
Regarding improvements to the Doppler speeds, that is only really dependent on the observable Doppler shift which isn't directly affected by ionospheric delays, so can't be improved through the same modelling that leads to positional improvements.
There is a tiny effect that positional accuracy has on the speed calculations (re: line of sight vector between receiver and satellite) but we're likely talking thousandths of knot in this instance.
However, I think the principle of tracking two different signals from some of the satellites (two frequencies, two tracking loops) and feeding both doppler observables into the LSQ estimator (or Kalman filter) would effectively average them out and lead to a slightly more reliable result, with less jitter.
Whether this happens in practice is another matter entirely. I've not seen any papers on this specific topic and I've read quite a few on the topic of the implementation of speed within a GNSS receiver (maybe a couple of dozen).That's not to say that nobody is doing it though!
