A Solution for Real-Time Ionospheric Delay Using an Adaptive Kalman Filter Based on Estimating the Variance Component

Abstract

Real-time solution of Global Navigation Satellite System (GNSS) epoch-differenced ionospheric delay (DID) is of great significance for real-time cycle slip detection and repair of multi-GNSS dual-frequency or trifrequency undifferenced measurements under high ionospheric activity. We construct a dynamic model of DID and perform a real-time estimate of the noise level of DID based on estimating the variance component. The estimated and predicted values of DID are obtained by designing a new adaptive Kalman filter algorithm with colored noise. Combining the predicted value and the detection method for cycle slips for Melbourne-Wübbena (MW) and Geometry-Free (GF) combination and taking into account the correlation between the predicted value and the carrier signal, we estimate the cycle slip, N2, on the second frequency of the carrier signal. The prediction and estimate of DID and detection and repair of dual-frequency cycle slip of multisystem undifferenced phase observations are measured with the GNSS multisystem observational data at different sampling rates (30 s, 15 s, 10 s, and 1 s). The results show that the DID model constructed in this paper is correct. The predicted value of DID has a high accuracy, which can effectively assist in dual-frequency cycle slip detection and repair. (1) The obtained predicted values, the estimated value, and the difference value between the two values of DID are less than 1.2 cm (STD), 1.2 cm (STD), and 0.6 cm (STD), respectively; (2) the precisions of the detection of cycle slip for MW, GF, and N2 are less than 0.083 cycles (STD), 0.4 cm (STD), and 0.071 cycles (STD), respectively; (3) with the obtained predicted value of DID to aid the detection and repair of cycle slip in GNSS double-frequency signals, a success rate of 100% can be reached.