Attacker Identification In LoRaWAN Through Physical Channel Fingerprinting

In today's environment, wireless networks and IoT technologies have been integrated into a variety of applications. One of the most essential IoT technologies is the Long-Range Wide Area Network (LoRaWAN) because of its capacity to combine long-range communication with energy-efficiency. Assuring network traffic safety and security is a critical issue in a wide range of today's industries, and security is a critical feature now that the digitization is accelerating. The "Man in the Middle" attack is one of the most dangerous and difficult to identify threats to wireless networks, and may become a critical threat for digitized assets and industrial processes. In this study, we look at an alternative approach to resolving this issue in order to assure the security of wireless networks. The technique is based on the "Behavioural Security Method," in which we employed a Feed-forward neural network model to build a binary classifier that can discriminate and recognize the original desired data while detecting the attacker's data. As a result, we discovered that as long as the attacker and the targeted node are not colocated, the model can discriminate between data coming from the source and data coming from the attacker with 99.6 percent accuracy. If, on the other hand, they are co-located, the model will fail to identify the source of data and the model's accuracy will drop to roughly 50%.