Intensive Investigation of RPL Attacks Based on Artificial Neural Network in Internet of Things (IoT)

Abstract

The restricted Internet of Things (IoT) devices are well-suited to the energy-efficient processes and accessible, confined, secure modes of operation of the Routing Protocol for Low Power and Lossy Networks. Nevertheless, we feel compelled to investigate, evaluate, and provide probable explanations for routing vulnerabilities due to the pressing need for security in RPL-based IoT networks. In order to study, evaluate, and quantify network topology assaults in RPL, also known as rank and local repair, this research introduces a lightweight model of an artificial neural network. This study builds a small artificial neural network (ANN) model with a low false positive rate, high accuracy (up to 0.99), and a significant effect by merging two datasets of attack scenarios. The suggested anomalous method is based on this strategy, which involves researching, evaluating, and capitalizing on the most crucial aspects of the RPL protocol.