GNSS Jamming Clustering Using Unsupervised Learning and Radio Frequency Signals

Global Navigation Satellite Systems (GNSS) provide vital position and timing information to receivers on the ground. This service is relied upon worldwide for many industries including telecommunications, online banking and developing technologies such as driverless cars. GNSS signals are vulnerable to interference and low-cost devices called jammers purchased easily online create an interference signal so that the genuine signal cannot reach the receiver. Incidents of this nature are increasing in frequency with a report showing European interference incidents to have increased 20 times in the two-year period from 2018 to 2020. Timely identification of unwanted signals is paramount in dealing with this global issue. This paper shows that clustering graphical representations of the signal and utilising convolutional neural network (CNN) feature extraction with transfer learning produces a higher V-measure score than without the feature extraction. Further, CNN feature extraction reduces the processing time of the clustering. Overall, this paper shows that GPS jammer detection classes can be clustered using an unsupervised learning algorithm such as k-means clustering.