Physical Layer Identification and authentication of electronic devices

In this thesis, I have investigated the problem of identification and authentication of electronic devices through their physical layer intrinsic features or fingerprints. The concept is that small differences in the electronic components of electronic devices leave small but significant traces in the digital output generated by the electronic device. Then, an analysis of the digital output provides the capability to identify and/or authenticate an electronic device from its digital output with a degree of accuracy, which is based on various factors including environmental effects. This research area has become more prominent in recent times due to the increasing computing power available for signal processing and analysis, which allows a more efficient and accurate extraction of the fingerprints. Even if there is considerable research in this area, which has proven the concept both with theoretical analysis and experimental results, there are still many aspects to be investigated both for the different types of electronic devices and for the analysis of the digital output through signal processing and machine learning techniques. The PhD activities have investigated various novel aspects in comparison to the existing literature. This thesis describes most of the results and describes the novelty in comparison to previous research literature. Three specific use cases were considered: identification of wireless devices, microphones and magnetometers.