Machine Learning Techniques for the Geolocalization of Jamming Sources in Indoor Wireless Networks

Jamming attacks produce malicious interference on wireless communication systems. The early and accurate localization of a jammer is extremely important in critical indoor environments, like airports and operating rooms. We employ electromagnetic simulations based on the ray tracing technique for quickly creating a dataset mapping the effects of the jammer on passive antennas in an indoor environment. Our simulation methodology is experimentally validated in a real scenario, using a commercial jammer and a multiport software defined radio transceiver. This dataset is further exploited using machine learning approaches to obtain an accurate localization of a jammer. The limitations of the proposed approach are also explored according to the measurement noise.