Re-identifying of LoRa Devices in an Actual Passive Packet Sniffer

The growing prevalence of Internet of Things (IoT) technologies in various applications has led to an increase in the deployment of popular LoRaWAN (Long Range Wide Area Network) networks. Since LoRa (Long Range) packets can travel long distances while carrying potentially sensitive information in their metadata, it is important to investigate the security vulnerabilities that can be revealed to an eavesdropper. In this paper, a passive packet sniffer is demonstrated to receive LoRa (Long Range) packets in the Campus Beaulieu area of Rennes, France. After processing the acquired packets, an algorithm is proposed based on machine learning to link the join request packets of a DevEUI to its corresponding data packets. This is done primarily by leveraging the proximity between their mutual radio profile characteristics, then, the proper data packets are identified based on their extensive set of features. The proposed algorithm is validated by applying it to the actual acquired data and then presenting a statistical analysis of the associated DevEUIs.