Hierarchical anomaly based intrusion detection and localization in IoT

In IoT systems, WSNs and Gateways are exposed to many attacks. WSNs are usually exposed to different types of intrusions like node compromise and denial of service attacks. IoT gateways that connects WSN to the Internet are exposed to all conventional IP attacks. In this paper, we propose an anomaly detection approach using support vector machines (SVM) for WSN intrusion detection, and deep learning technique for gateway intrusion detection. We propose a detection protocol that dynamically executes the on-demand SVM classifier in a hierarchical way whenever an intrusion is suspected. We combine machine learning classification with a statistical approach for malicious node localization. This novel approach allows finding a compromise between intrusion detection efficiency and resource overhead for WSN and gateway security.