Automatic Equatorial GPS Amplitude Scintillation Detection Using a Machine Learning Algorithm

Abstract

This paper presents a machine learning algorithm to detect ionospheric scintillation and classify scintillation events based on training data in the frequency domain. The detector input is the signal intensity. Validation using data from Ascension Island, Hong Kong, and Peru shows 91–96% accuracy of scintillation detection. Different combinations of training data, observation matrices, and learning algorithms are investigated to obtain performance measures. Testing results on data from Singapore demonstrate the general capabilities of the detector.