Publication

Abstract : In the era of the Internet of Things (IoT), smart homes have become a cornerstone of modern living, offering convenience, energy efficiency, and security. The extensive connectivity within these networks brings forth security challenges, making the detection of anomalies a critical aspect of IoT network security. Anomalies encompass a wide array of unexpected events, from irregular data patterns to communication anomalies, power fluctuations, behavioral deviations, security breaches, and environmental variations. This paper underscores the pivotal role of anomaly detection in securing IoT networks, highlighting its significance in identifying both known threats and emerging risks. Anomaly detection in IoT relies on machine learning and statistical techniques, encompassing data collection, preprocessing, feature extraction, model training, real-time anomaly detection, and responsive actions. Reinforcement Learning(RL), a subset of machine learning, emerges as a formidable tool in enhancing IoT network security. RL models continuously observe and adapt to device behavior, identifying anomalies and security breaches in real-time. RL offers adaptability in rapidly changing IoT environments, proactive responses to anomalies, and the ability to learn and evolve against new threats. RL strengthens network security without compromising individual device data. Our advanced security system that refines its understanding of the smart home environment, collectively learns from anomalies, and bolsters network security. Synergy of IoT and RL is reshaping IoT network security, promising not only seamless device functionality but also the safeguarding of sensitive data and personal spaces. © 2024 IEEE