A novel predictive algorithm for double difference observations of obstructed BeiDou geostationary earth orbit (GEO) satellites. Issue 5 (1st March 2019)

• Record Type: Journal Article
• Authors: Du, Yuan; Huang, Guanwen; Zhang, Qin; Tu, Rui; Han, Junqiang; Yan, Xingyuan; Wang, Xiaolei
• Is Part Of: Advances in space research. Volume 63:Issue 5(2019)
• Page Start: 1554
• Page End: 1565
• Publication Date: 2019-03-01
• Subjects: BDS GEO -- Double difference observation prediction -- RTK -- Douglas-Peucker; Space sciences -- Periodicals; Astronautics -- Periodicals; Geophysics -- Periodicals; 500.505
• DOI: 10.1016/j.asr.2018.11.009
• Languages: English
• ISSNs: 0273-1177

Abstract

Abstract: Transmission link disturbances and device failure cause global navigation satellite system (GNSS) receivers to miss observations, leading to poor accuracy in real-time kinematic (RTK) positioning. Previously described solutions for this problem are influenced by the length of the prediction period, or are unable to account for changes in receiver state because they use information from previous epochs to make predictions. We propose an algorithm for predicting double difference (DD) observations of obstructed BeiDou navigation system (BDS) GEO satellites. Our approach adopts the first-degree polynomial function for predicting missing observations. We introduce a Douglas-Peucker algorithm to judge the state of the rover receiver to reduce the impact of predictive biases. Static and kinematic experiments were carried out on BDS observations to evaluate the proposed algorithm. The results of our navigation experiment demonstrate that RTK positioning accuracy is improved from meter to decimeter level with fixed ambiguity (horizontal < 2 cm, vertical < 18 cm). Horizontal accuracy is improved by over 50%, and the vertical accuracies of the results of the static and kinematic experiments are increased by 47% and 27% respectively, compared with the results produced by the classical approach. Though as the baseline becomes longer, the accuracy is weakened, our predictive algorithm is an improvement over existing approaches to overcome the issue of missing data.