Access to signals from Global Navigation Satellite Systems (GNSS) is crucial in many applications like navigation or exact time-stamping of transactions. Recently, however, radio jammers blocking GNSS components became more popular. These devices, which are readily available on any online market platform, are predominantly used to circumvent tracking like road user charging systems. The possession of such jammers is not illegal; only their operation is. Therefore, the police must match the recorded signal with a signal from a confiscated jammer to prosecute the use. In this thesis, we propose an approach to compare two jamming signals and assess its feasibility to identify jammers. The main characteristic of a jammer is the way the frequency of its signals varies over time. Therefore, we translate each signal into time-frequency space using a combination of the Short-Time Fourier Transformation and the Wigner-Ville Distribution. Features are subsequently identified based on the Singular Value Decomposition of the resulting time frequency matrix. This approach generates an input vector for a random forest or a neural network classifying the signal. Two signals from the same jammer class are then compared using a siamese neural network architecture. The outputs of this network are used to decide whether two signals correspond to the same jammer.      «

Access to signals from Global Navigation Satellite Systems (GNSS) is crucial in many applications like navigation or exact time-stamping of transactions. Recently, however, radio jammers blocking GNSS components became more popular. These devices, which are readily available on any online market platform, are predominantly used to circumvent tracking like road user charging systems. The possession of such jammers is not illegal; only their operation is. Therefore, the police must match the reco...     »