'Rogue Sensor Identification for LoRaWAN deployments'
Mark Beach, Professor of Radio Systems Engineering, University of Bristol
With the ever-increasing use of low-cost wireless IoT technology for telemetry applications, including use cases within critical infrastructure, ensuring the ongoing security and resilience of these technologies is of paramount importance. However, as with many of the wireless networks that underpin modern life, these technologies rely on radio frequency (RF) interfaces which can be vulnerable to cyber-attacks or other failures. The detection and mitigation of potential RF cyber intrusions forms one of the research priorities of the UKRI Prosperity Partnership in Secure Wireless Agile Networks (SWAN). Within SWAN, we have adopted LoRaWAN, a proprietary low-power wide area technology developed for the Internet of Things (IoT), as a candidate technology for detailed evaluation given the increasing prevalence within remote telemetry applications. In this presentation, a novel unsupervised machine learning (ML) algorithm is presented for the expeditious RF fingerprinting of LoRa modulated chirps, as rogue device identification based on received signal strength indicator (RSSI) alone is unlikely to yield a robust means for sensor authentication within critical infrastructure deployments. Specifically, an unsupervised ML algorithm is used to rapidly train an artificial neural network (ANN) matrix creating self-organizing maps (SOMs) for each authentic transmitter and a potential rogue node. The SOMs are then used to train a general classifier which will be shown to reliably profile each transmitter as friend or foe.
