Detection of Cyberattacks in IoT Networks Using Artificial Intelligence: A Comparative Study
Matheus Figueiredo, Dar'c Pabla Sodre, Renan Landau Paiva de Medeiros, Vicente Ferreira de Lucena, Iury Valente de Bessa · 2024
The use of Internet of Things (IoT) technologies has become more readily available with the advent of cyber-physical systems. This context motivates concerns about cybersecurity and the occurrence of malicious attacks in IoT networks. This paper investigates the problem of the automatic detection of cyberattacks in MQTT-based IoT networks by deploying artificial intelligence algorithms for processing traffic data and indicating whether an attack is occurring or not. Thus, this paper trains different artificial intelligence binary classifiers based on machine learning and compares their performance for malicious attack detection in cyber-physical systems with MQTT-based IoT networks. For training and testing the classifiers, we employ the MQTTset dataset which contains many labelled samples with both legitimate traffic and observations under attack occurrence. By analyzing data features and preprocessing, the algorithms achieved good performance in classifying network traffic, contributing to the security of cyber-physical systems.