Beta Hebbian Learning for Intrusion Detection in Networks of IoT Devices

This research paper is focused on security in IoT devices network, by providing a visual tool based on Beta Hebbian Learning (BHL) to easily identify attacks in the network to human experts. Contrary to Artificial Intelligent-driven solutions based on supervised learning BHL does not require labelled information. A testing environment of IoT devices and web clients is created and attack by using a Sybil Attack type, recording all traffic information in separating the fields of the captures by the most relevant, these fields are taken from "Wireshark Display Filter Reference" tool. Results obtained by BHL algorithm provide clear projections where most of the attacks can be easily identified by human expert through visual inspection in real time, proving a powerful tool to be easily implemented in IoT environments.