WiSIDS: A Lightweight Machine Learning-Based Intrusion Detection System for Securing Wireless IoT Networks

Abstract

The rapid proliferation of Internet of Things (IoT) devices in wireless environments has introduced significant security challenges due to their limited computational resources and susceptibility to cyberattacks. Traditional Intrusion Detection Systems (IDS) are often unsuitable for IoT because of their high processing overhead, memory consumption, and latency. This paper proposes WiSIDS (Wireless IoT Secure Intrusion Detection System), lightweight IDS that leverages machine learning to provide accurate and efficient attack detection in wireless IoT networks. WiSIDS employs feature selection and model optimization to minimize computational complexity while maintaining high detection performance. Experimental evaluation is conducted on benchmark IoT security datasets, including NSL-KDD and Bot-IoT, to assess accuracy, precision, recall, F1-score, processing delay, and energy consumption. Results demonstrate that WiSIDS achieves a detection accuracy of over 95%, reduces latency by up to 60%, and lowers resource utilization compared to conventional IDS solutions. These findings highlight the feasibility of deploying WiSIDS in resource-constrained IoT devices to enhance wireless network security without compromising efficiency.