Adaptive Kalman filter for GNSS receiver

In GNSS satellite navigation system receivers, Kalman filtering algorithms are typically used. As a rule, Extended Kalman Filters (EKF) are applied, designed under the assumption of constant pseudorange accuracy. In practice, pseudoranges have variable accuracy and, additionally, can occasionally be subject to errors significantly exceeding typical values (anomalous errors). For this reason, it is advisable to estimate the measurement error covariance matrix within the filter, rather than assuming a constant value. This paper presents an Adaptive Kalman Filter for GNSS receivers (ADKF), developed by the author, in which this matrix is estimated in real-time, allowing the filter to adjust its weights to the changing accuracy of measurements. The design model and the filter algorithm are provided, and selected research results are presented.GNSS (Global Navigation Satellite System) navigation receivers determine position using the ranging method [3, 6, 12]. The measured quantities are the time differences between the moments of transmission and reception of radio signals broadcast by navigation satellites. Satellite signals contain, in the so-called navigation message, information necessary to determine their position and the signal transmission time, among other things.
The GNSS receiver uses its own clock to determine the signal reception time. (…)

Adaptive Kalman filter for GNSS receiver – PDF file for download