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

Abstract

The Global Navigation Satellite System (GNSS) refers to a constellation of satellites that provide position and time signals from space that allow a receiver to determine location. These GNSS signals can be used to determine the position and time of an object such as a human, a vehicle, and many other objects. Moreover, in this study, GNSS signals were used to determine the soil geophysical movement based on the signal positioning of the receiver station. A signal processing method has been applied to find the best feature extraction method for normal and abnormal signal classification. A set of GNSS signals has been acquired from a land deformation monitoring station in China that contains 3 months of the continuous GNSS signal. The signal is then processed and extracted using time series analysis and statistical approach to determine the normal and abnormality of the soil movement. There were three (3) periodic terms involved in the GNSS signal processing which velocity, distance, and acceleration. These 3 properties were used to analyze the normal and abnormality of soil movement that will be contributed to a landslide. Finally, the experimental results have proved that the feature extraction method on GNSS signal was found to be able to produce significant features to detect normal and abnormal land deformation.