Abstract

The IoT is widely used globally to automate operations, enhance system functionality, and enable remote monitoring and diagnostics. IoT devices incorporate sensors that continuously collect real-time information and relay it to a central system or controller. This central system typically employs machine learning models to process incoming data and regulate system behavior. Despite these benefits, IoT systems present several challenges. Security is a major concern, as most IoT devices are connected to the Internet or local networks and may be vulnerable to attacks such as Denial of Service (DoS), Distributed Denial of Service (DDoS), and Man-in-the-Middle (MitM). Energy efficiency is another important concern, since many existing intrusion detection models are computationally intensive and consume substantial energy, limiting their use in resource-constrained IoT devices. In view of these challenges, this work presents a lightweight machine learning model for IoT threat detection. The proposed solution classifies network traffic using selected features to detect malicious activity while maintaining low energy consumption. The TON-IoT dataset is used to train and evaluate the model, achieving a balance between detection accuracy and computational performance.