Lightweight AI for Smart Home Security: IoT Sensor-Based Automated Botnet Detection

Abstract

The rapid expansion of IoT devices in smart homes has led to an increase in cyberattacks, particularly IoT-based botnets, which pose significant security risks. Traditional security mechanisms often struggle with scalability, efficiency, and adaptability, leaving critical vulnerabilities unaddressed. Existing methods like SH-SecNet focus on network-level threat detection but lack the efficiency needed for real-time, resource-constrained environments. To address these challenges, the method proposes a lightweight AI-based model for automated botnet detection using IoT sensors in smart homes. The model utilizes advanced classification techniques to ensure accurate, low-latency detection with minimal resource consumption. The novelty of this approach lies in its optimized use of AI for real-time botnet detection while maintaining energy efficiency—critical for IoT devices. The model achieves 88% security threat detection, 91% accuracy, and 93% throughput, outperforming SH-SecNet with a 13% improvement in threat detection, a 6% boost in accuracy, and a 13% increase in throughput. Additionally, it shows superior scalability (82% vs. 65%) and network protocol support (89% vs. 78%) while reducing round-trip time by 24%. The proposed method offers significant advantages over existing models, demonstrating superior performance in handling varying IoT traffic and integrating with smart city infrastructures. This model presents a robust, scalable, and energy-efficient solution for enhancing security in smart home environments. Future work will focus on adapting the model to address emerging botnet attack vectors in more complex IoT ecosystems.