An Analysis Of Anomaly Detection Techniques for IoT Devices: A Review

Abstract

There is an increasing demand for effective intrusion detection systems to protect the privacy and security of IoT devices and the data they collect and transmit. IoT devices are susceptible to a broad spectrum of security risks, such as unauthorised access, assaults, and breaches, which can put in danger the network's integrity, confidentiality, and availability. This study investigates the various methodologies used in IoT device anomaly detection. The paper discusses possible solutions, such as using lightweight algorithms, distributed intrusion detection systems, and adaptive security mechanisms. Some of the frameworks reviewed in this paper are BFA-PDBSCAN, B-Stacking, Two stream neural network and Hybrid (Anomaly-based +Specification-based) detection. Effective anomaly detection in IoT devices requires a multi-layered security approach that incorporates various intrusion detection techniques and best practises in order to protect the confidentiality and anonymity of IoT devices or the collected and transmitted data.