Al-Based Automated Attack Detection in IOT Networks using Threat Analysis

This review paper provides a comprehensive evaluation of the Internet of Things (IoT) security landscape, emphasizing the pivotal role of Artificial Intelligence (AI) in automated threat detection. Given the limitations of traditional signature-based systems in handling the massive data volumes and sophisticated zero-day attacks of modern networks, this study categorizes common IoT threats, including DDOS, MitM, and botnet infiltrations. The primary focus is a synthesis of research findings on Machine Learning (ML) algorithms like SVM and Random Forest, and Deep Learning (DL) architectures such as CNN, LSTM, and GAN, delineating their respective strengths in identifying network anomalies. Furthermore, the article addresses critical technical and ethical challenges, specifically data heterogeneity, resource constraints in edge devices, and the vulnerability of AI models to adversarial attacks. Finally, emerging trends including Federated Learning, Edge AI, and Blockchain integration are analyzed as futuristic solutions for achieving a resilient and autonomous security framework in the ۶G era. This synthesis aims to bridge the gap between high-detection accuracy and the hardware limitations of decentralized IoT ecosystems.

کلیدواژه ها:

Artificial Intelligence (AI) ، Internet of Things Security (IOT) ، Instrusion Detection Systems (IDS) ، Deep Learning ، Federated Learning ، ۶G networks