Real-time prediction of GNSS epoch-differenced ionospheric delay

A new adaptive Kalman filter algorithm was proposed to perform the prediction of the epoch-differenced ionospheric delay. A dynamic model of the epoch-differenced ionospheric delay was constructed and its noise level was adjusted online based on the method of the variance component estimation. After analysis of the colored noise characteristics of GNSS carrier phase time differential observations, the real-time estimated and predicted values of epoch-differenced ionospheric delay were obtained by designing an adaptive Kalman filter with a finite impulse response colored noise. Combining the predicted value, real-time detection and repair of cycle slip in GNSS dual-frequency and tri-frequency non-difference phase observations could be presented in a stepwise manner. Namely: First, the wide and extra-wide lane cycle slips were calculated by direct rounding the initial float cycle slip. Second, the cycle slip on the third frequency of the carrier signal was determined by the least square estimation after taking into account the correlation between the predicted ionospheric value and the carrier signal and correcting the narrow lane cycle ship and the carrier signal of current epoch using the predicted value. Finally, the carrier signal repaired by the cycle slip was used to estimate the epoch-differenced ionospheric delay of the current epoch. The results of experiment with the BDS observational data at different frequencies show that: 1) the standard daviation of obtained predicted values, the estimated values, and the difference values between the two values of ionospheric delay are less than 1.4, 1.5, and 0.7 cm, respectively; 2) with the obtained predicted value of epoch-differenced ionospheric delay to aid the detection and repair of cycle slip in GNSS dual-frequency and tri-frequency signals, a success rate of 100% can be reached. © 2019, Editorial Board of Journal of CUMT. All right reserved.