An improved robust Kalman filtering strategy for GNSS kinematic positioning considering small cycle slips

摘要

In GNSS (Global Navigation Satellite Systems) kinematic positioning, observations will be inevitably contaminated by cycle slips and gross errors, as the complex observation environment changes rapidly. These outliers will degrade the performance of classic Kalman filtering applied in GNSS kinematic resolution and eventually, the filtering may converge slowly or even diverge and thus the precision will be degraded. Therefore, a robust Kalman filter should be applied to resist the influence of these outliers that cannot be identified in the data preprocessing stage. Based on the conventional IGG (Institute of Geodesy and Geophysics) III equivalent weight method which addresses the outliers of the zero-weight segment with the same strategy, this paper proposes an improved robust Kalman filtering strategy that detects outliers by both posterior phase residuals and standardized residuals and handles the carrier-phase observation of zero weight segment as cycle slips. In addition, to avoid unnecessary ambiguity reinitialization caused by the detected cycle slips, only when the carrier-phase observation of the same satellite is classified in the zero-weight segment over two consecutive epochs should the ambiguity be reinitialized. Experimental results of relative positioning show that the improved method can not only mitigate the influence of unexpected outliers in the Kalman filter but also improve the fixing rate of ambiguity resolution as well as the accuracy and reliability of positioning. (C) 2017 COSPAR. Published by Elsevier Ltd. All rights reserved.