LEVERAGING EDGE AI FOR REAL-TIME ANOMALY DETECTION IN INTERNET OF THINGS NETWORKS

Abstract

The rapid growth of the Internet of Things (IoT) has enabled unprecedented connectivity but has also exposed networks to large-scale security and reliability risks. Traditional centralized anomaly detection methods struggle with latency, bandwidth constraints, and privacy concerns. This paper proposes an Edge AI framework for real-time anomaly detection in IoT networks, moving intelligence closer to data sources to reduce latency and preserve privacy. The system combines distributed edge devices running lightweight deep learning models with a cloud-based aggregation server for coordination and model updates. Experimental results show improved detection accuracy, faster response times, and reduced network load compared to centralized systems. These findings highlight the promise of Edge AI as a scalable and privacyconscious solution for securing IoT infrastructures