Automatic GPS Phase Scintillation Detector Using a Machine Learning Algorithm

Yu Jiao, John Hall and Yu (Jade) Morton

Peer Reviewed

Abstract: Carrier phase scintillation is the main cause of large errors in GNSS navigation solutions. Timely detection of phases scintillation will enable adaptive processing to mitigate its effects on navigation solutions. This paper presents a machine learning algorithm to autonomously detect ionospheric phase scintillation without assuming probability distributions of the scintillation signals. The machine learns to classify phase scintillation events based on given training data in the frequency domain. Training validation using scintillation data from Gakona, Alaska shows the accuracy of phase scintillation detection at around 92%. Test results using data from Poker Flat (Alaska), Jicamarca (Peru), Singapore, and Hong Kong demonstrate the general capability of the phase scintillation detector. Different variations of the machine learning algorithm are also investigated to obtain performance measures, which reveal that carrier phase scintillation index ?? values may not be a good indicator of the scintillation activity. In addition, concurrent phase and amplitude scintillation detection using similar machine learning algorithms is further investigated with data from equatorial regions. Results show that although the values of S_4 and ?? indices are highly correlated at low latitudes, the detection of phase and amplitude scintillation events may not occur simultaneously.
Published in: Proceedings of the 2017 International Technical Meeting of The Institute of Navigation
January 30 - 2, 2017
Hyatt Regency Monterey
Monterey, California
Pages: 1160 - 1172
Cite this article: Jiao, Yu, Hall, John, Morton, Yu (Jade), "Automatic GPS Phase Scintillation Detector Using a Machine Learning Algorithm," Proceedings of the 2017 International Technical Meeting of The Institute of Navigation, Monterey, California, January 2017, pp. 1160-1172. https://doi.org/10.33012/2017.14903
Full Paper: ION Members/Non-Members: 1 Download Credit
Sign In