Rule-Based System with Machine Learning Support for Detecting Anomalies in 5G WLANs

Abstract

The purpose of this paper is to design and implement a complete system for monitoring and detecting attacks and anomalies in 5G wireless local area networks. Regrettably, the development of most open source systems has been stopped, making them unable to detect emerging forms of threats. The system provides a modular framework to create and add new detection rules as new attacks emerge. The system is based on packet analysis modules and rules and incorporates machine learning models to enhance its efficiency. The use of rule-based detection establishes a strong basis for the identification of recognized threats, whereas the additional implementation of machine learning models enables the detection of new and emerging attacks at an early stage. Therefore, the ultimate aim is to create a tool that constantly evolves by integrating novel attack detection techniques. The efficiency of the system is proven experimentally with accuracy levels up to 98.57% and precision as well as recall scores as high as 92%.