Ionospheric delay prediction based on online polynomial modeling for real-time cycle slip repair of undifferenced triple-frequency GNSS signals

Abstract

In real-time cycle slip repair of undifferenced triple frequency Global Navigation Satellite Systems (GNSS) phase signals, ionospheric delay prediction is crucial in estimating the narrow lane part of the cycle slips. A quadratic model is recursively constructed using estimated ionospheric delays and used to perform one-step prediction. The model errors are distinguished from the data introduced errors and the model error variance is adaptively tuned online. In estimating the float narrow-lane cycle slips, a statistically rigorous error model is followed where the correlations between the predicted ionospheric delays and the epoch-differenced phases are given full considerations. An efficient algorithm for implementing the proposed method is developed where matrix inversion is avoided. A real Beidou System (BDS) data set is employed to check the performance of the proposed method. The manually introduced cycle slips have been all correctly repaired.