Classification of GNSS Signals for Land Deformation Detection Based on Signal Processing Techniques

Abstract

Global Navigation Satellite System (GNSS) technology encompasses various satellite navigation systems, including GPS, GLONASS, Beidou, and Galileo, providing positioning with global coverage and enabling users to determine their location and timing. GNSS operates through three main components: the spatial segment, consisting of satellites that transmit signals from space; the control segment, which manages and monitors the satellite operations; and the receiver segment, where GNSS receivers process the signals to derive precise location and timing data. Moreover, GNSS can also be used for deformation monitoring, particularly landslide monitoring. In this paper, we propose a signal processing technique for the classification of GNSS land deformation abnormalities. Six GNSS stations at different polar coordinates were used and a time series analysis based on a statistical approach was proposed for the feature extraction process. There are three periodic terms involved in time series analysis including distance, velocity and acceleration which become important in detecting ground movements through GNSS geographic coordinate signals. These terms were later on combined with statistical parameters, i.e. minimum and maximum values, as well as feature vectors for input to the threshold classifier. Classification results were obtained above 96% for normal and abnormal ground deformation GNSS signals which implies the encouraging performance of the proposed classification technique.