Revealing Spectrum Allocation Scheme and Temporal Transmission Behavior of IoT Devices using Passive Packet Sniffing

The growing prevalence of Internet of Things (IoT) technologies in many outdoor applications has led to an increase in the deployment of popular LoRaWAN (Long Range Wide Area Network) networks. As a result, there is increasing interest in tracking end nodes and detecting device transmission behavior. In this paper, a passive packet sniffer is demonstrated to receive LoRa (Long Range) packets in the Campus Beaulieu area of Rennes, France, over a period of two months. After processing the acquired packets, the inter-arrival times between packets of each identified device are estimated properly using a proposed technique, i.e., based on Kernel Density Estimation (KDE). Another algorithm is also employed to detect any pattern in the inter-arrival times. The study is then extended to reveal the spectrum allocation technique of a device and detect any periodic structure in the sequence of the used frequencies. Over and above, the spectrum allocation scheme is modelled as a Markov chain, allowing for the prediction of the next selected frequency using the estimated trajectory. The proposed algorithms are validated through statistical analysis of the inter-arrival times, temporal patterns, and spectrum allocation schemes' characteristics from the actual acquired data.