Improving the Routing Protocol in the Internet of Things by Rule Extraction on Simulation Data

The modern world is witnessing a growing fascination with the Internet of Things (IoT), which has seamlessly integrated into our daily lives. This integration has resulted in the widespread adoption of Low-power and Lossy Networks (LLNs). These networks face limitations in terms of power and storage capacity. To address this, the Internet Engineering Task Force (IETF) has introduced RPL, an open standard routing protocol. Unfortunately, the RPL protocol is vulnerable to various attacks that can negatively impact the network's efficiency and resource utilization, ultimately leading to erroneous outcomes. In previous methods, the attacker carried out his attack based on fixed parameters, and the defender responded based on his understanding of the attack. In this study, unlike previous methods, an attempt has been made to implement a general structure for a comprehensive attack. Also, this attack is prevented based on a combination of data mining and decision trees. In this study, the attacker and defender are unaware of each other’s performance and conditions, and the defender determines the type of attacker detection based on data obtained from various simulations, data extraction, and modeling by decision trees through the Binary Relevance method. The simulation outcomes demonstrate that the employed approach exhibits favorable performance when compared to the latest techniques available.