A Fine-Tuned Positioning Algorithm for Space-Borne GNSS Timing Receivers

• Published in: Sensors (Basel, Switzerland) Vol. 20; no. 8; p. 2327
• Main Authors: Chen, Xi, Wei, QiHui, Zhan, YaFeng, Ma, TianYi
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI 19.04.2020; MDPI AG
• Subjects: GNSS; LING QIAO; positioning algorithm; space-borne; timing receiver; positioning algorithm; timing receiver; LING QIAO; space-borne; GNSS
• ISSN: 1424-8220; 1424-8220
• DOI: 10.3390/s20082327

To maximize the usage of limited transmission power and wireless spectrum, more communication satellites are adopting precise space-ground beam-forming, which poses a rigorous positioning and timing requirement of the satellite. To fulfill this requirement, a space-borne global navigation satellite system (GNSS) timing receiver with a disciplined high-performance clock is preferable. The space-borne GNSS timing receiver moves with the satellite, in contrast to its stationary counterpart on ground, making it tricky in its positioning algorithm design. Despite abundant existing positioning algorithms, there is a lack of dedicated work that systematically describes the delicate aspects of a space-borne GNSS timing receiver. Based on the experimental work of the LING QIAO (NORAD ID:40136) communication satellite's GNSS receiver, we propose a fine-tuned positioning algorithm for space-borne GNSS timing receivers. Specifically, the proposed algorithm includes: (1) a filtering architecture that separates the estimation of satellite position and velocity from other unknowns, which allows for a first estimation of satellite position and velocity incorporating any variation of orbit dynamics; (2) a two-threshold robust cubature Kalman filter to counteract the adverse influence of measurement outliers on positioning quality; (3) Reynolds averaging inspired clock and frequency error estimation. Hardware emulation test results show that the proposed algorithm has a performance with a 3D positioning RMS error of 1.2 m, 3D velocity RMS error of 0.02 m/s and a pulse per second (PPS) RMS error of 11.8ns. Simulations with MATLAB show that it can effectively detect and dispose outliers, and further on outperforms other algorithms in comparison.