Noise Statistics Estimation Techniques for Robust GNSS Carrier Tracking

Carrier tracking is typically implemented by well-known phase-locked loops (PLLs), a receiver architecture that has shown its benefits for many decades. Equivalences between the PLL and the Kalman filter (KF) exists, being the former a particularization of the KF. Progressive substitution of PLL schemes by more sophisticated KF-based carrier tracking loops would allow receivers to tackle effects that are currently degrading PLLs performance and, consequently, carrier-phase estimates. This comes at the price of slightly larger number of parameters to tune. Particularly, this article points out a caveat as to the difficulties in tuning covariance matrices for the dynamical system, which turns to be a critical factor in KF performance. It is desirable to use techniques that, autonomously, select the appropriate values for such covariance matrices. The paper discusses the main covariance estimation techniques for KF adjustment and its applicability in the context of GNSS carrier-tracking with meaningful computer simulation results.