A Comprehensive Survey of Threat Detection and Mitigation in Layered IoT Security Frameworks

The exponential growth of the Internet of Things (IoT) is transforming both daily life and industrial systems by enabling smart environments through interconnected devices. However, this expansion has also increased the surface for cyber threats such as denial of service, spoofing, data injection, and sinkhole attacks, which pose serious challenges to data integrity and system reliability. This survey presents a comprehensive review of intrusion detection strategies aligned with the layered IoT architecture. It examines traditional approaches such as signature-based detection, behavioral analysis, and honeypot systems and explores enhancements through intelligent methods like machine learning (ML), deep learning (DL), and reinforcement learning (RL).Going beyond isolated techniques, the paper proposes a hybrid, multilayered framework that integrates conventional and adaptive mechanisms to improve detection accuracy and real-time responsiveness. It also highlights recent benchmark evaluations, emphasizing the advantages of incorporating RL to develop scalable, robust, and energy-efficient security systems suitable for dynamic IoT environments.