Dual Kalman filtering based GNSS phase tracking for scintillation mitigation

Abstract

This paper treats the problem of tracking the global navigation satellite system's (GNSS) signal phase under ionospheric scintillations and to generate an estimate of the scintillation phase and amplitude. In the case of ionospheric scintillations traditional phase locked loop (PLL) based GNSS receivers have problems to track the signal phase accurately. Moreover, these receivers do not estimate the scintillation phase and amplitude, which is evaluated in ionospheric monitoring applications. Therefore, Kalman filter based phase tracking algorithms have been proposed, which track line-of-sight (LOS) phase and scintillation amplitude and phase separately with different state space models. For modeling the scintillation phase and amplitude auto-regressive (AR) models can be used. In this case the parameters for the AR model have to be estimated together with the LOS and scintillation phase and amplitude. We propose to use a second Kalman filter for the estimation of the AR model parameters. This Kalman filter takes the scintillation phase and amplitude estimates from the first Kalman filter as measurement and the AR model parameters as system state. Simulation results for the global positioning system (GPS) show that the proposed algorithm achieves the same estimation performance for scintillation phase and amplitude as an algorithm with a priori knowledge on the AR model parameters.

Survey on signal processing for GNSS under ionospheric scintillation: Detection, monitoring, and mitigation

“…Recently, to increase robustness in time‐varying scenarios, adaptive AR parameter estimation techniques have been considered in Vilà‐Valls, Fernández‐Prades, Closas, and Arribas (2017) and Fohlmeister et al. (2018), showing that the approach in Vilà‐Valls, Closas, Fernández‐Prades, and Curran (2018) is the latest trend on single‐frequency scintillation mitigation, and must be seen as the performance benchmark for the derivation of new scintillation mitigation strategies.…”

Traditional and Kalman Filter-Based GNSS Receiver Structures for Ionospheric Scintillation Monitoring and Mitigation

“…The process noise covariance matrix Q D for the LOS dynamics is typically defined by σ 2 a D = 0.2 rad 2 /s 5 [16]. The variances σ 2 φ S and σ 2 ρ S are experimentally adjusted based on the analysis of the data.…”

“…This concept is generalized in [15] with an extended Kalman filter with increased complexity including online identification of the AR model parameters and online updating of the process noise covariance matrix based on the identification statistics. Other techniques also consider adaptive Kalman filtering with online adaptive estimation of the AR scintillation model parameter, as in [16], where a dual Kalman filter was applied. The scintillation phase and amplitude are estimated by a first Kalman filter.…”

Kalman Filter-Based Carrier Tracking with Ionospheric Scintillation Mitigation for GNSS

“…The process noise covariance matrix Q D for the LOS dynamics is typically defined by σ 2 a D = 0.2 rad 2 /s 5 [5]. The variances σ 2 φS and σ 2 ρS are experimentally adjusted based on the analysis of the data.…”

“…The introduction of models representing the scintillation dynamics in the received signal into the Kalman filter formulation enabled decoupling both scintillation and LOS contributions to overcome the limitations of those indirect techniques that adjust the level of uncertainty in models accounting for states related to the LOS dynamics only. The LOS dynamics are typically modeled by a kinematic process [3] in the Kalman PLL and scintillation amplitude and phase are typically modeled by autoregressive (AR) processes [4], [5]. In this case, the parameters of the AR model also have to be estimated.…”