Revolutionising IoT Network Security By Assessing ML Localisation Techniques Against Jamming Attacks

Abstract

The Internet of Things (IoT) revolutionises data flow management by interconnecting various devices, from sim-ple sensors to complex systems, for varied applications, including smart homes and industrial automation. However, this extensive integration also renders IoT networks vulnerable to cyber threats, mainly jamming attacks, which can disrupt wireless communications and risk total network failure, significantly impacting critical sectors such as healthcare and industrial au-tomation. This research introduces a novel approach to IoT network security, transitioning from conventional detection meth-ods to a refined strategy emphasising the accurate localisation of jamming sources. Integrating machine learning with network security to counter jamming attacks establishes a foundation for future exploration. It effectively utilises Fuzzy Logic-based Intrusion Detection Systems (FLIDS) for the initial detection of jamming threats, addressing a vital need for heightened security in IoT networks. Central to our research is adopting advanced machine learning models, including Recurrent Neural Network - Bidirectional Long Short-Term Memory (RNN-B-LSTM), Temporal Convolutional Network (TCN), Recurrent Neural Network (RNN) and Convolutional Neural Network (CNN). These models are carefully selected for their capacity to process complex data patterns and suitability for real-time IoT environments. We conduct a thorough evaluation of these models using various metrics. Consequently, the RNN-B-LSTM model, in particular, demonstrates exceptional accuracy in jammer localisation.