GLONASS inter-frequency phase bias rate estimation by single-epoch or Kalman filter algorithm

Abstract

GLONASS double-differenced (DD) ambiguity resolution is hindered by the inter-frequency bias (IFB) in GLONASS observation. We propose a new algorithm for IFB rate estimation to solve this problem. Although the wavelength of the widelane observation is several times that of the L1 observation, their IFB errors are similar in units of meters. Based on this property, the new algorithm can restrict the IFB effect on widelane observation within 0.5 cycles, which means the GLONASS widelane DD ambiguity can be accurately fixed. With the widelane integer ambiguity and phase observation, the IFB rate can be estimated using single-epoch measurements, called the single-epoch IFB rate estimation algorithm, or using the Kalman filter to process all data, called the Kalman filter-based IFB rate estimation algorithm. Due to insufficient accuracy of the IFB rate estimated from widelane observations, the IFB rate has to be further refined with L1 and L2 observations. A new reference satellite selection method is proposed to serve the IFB rate estimation. The experiment results show that the IFB rates on L1 and L2 bands are different, that an accurate IFB rate will help us to obtain more fixed solutions at places with serious occlusion, that the single-epoch IFB rate estimation algorithm can meet the requirements for real-time kinematic positioning with only 8% extra computational time, and that the Kalman filter-based IFB rate estimation algorithm is a satisfactory option for high-accuracy GLONASS positioning.