GNSS receivers are nowadays used in many applications, ranging from personal navigation devices to spacecraft navigation and weapon guidance. Therefore, reliable GNSS positioning is fundamental and failure to provide accurate positioning could cause  catastrophic effects threatening vital infrastructures and even causing loss of human lives. There may be different reasons for unreliable GNSS signals, among which malicious attacks. Malicious attacks can be of different types, namely jamming and spoofing.

In this study, we aim at exploring the possibility of using machine learning techniques for detecting GNSS attacks. In particular, we would like to apply different feature selection techniques on publicly available GNSS signal databases with the purpose of identifying
features and correlations among them that are relevant in discerning legitimate from spoofed and/or jammed signals.