State-of-the-art survey of artificial intelligent techniques for IoT security

Abstract

The data protection problem concerning the Internet of Things (IoT) paradigm has drawn the innovation community’s considerable attention. Several surveys have covered different IoT-centered issues, namely vulnerability simulation, intrusion detection systems, and state-of-the-art techniques were put forward for this purpose. In comparison, we concentrate exclusively on the emerging IoT vulnerabilities and related Artificial Techniques in the current research. This paper initializes the detailed categorization of recent research works, which explore different Machine Learning and Deep Learning techniques for the IoT paradigm. Additionally, a novel taxonomy is included based on IoT vulnerabilities, corresponding attackers, and effects, threats that explore weak links, effective remedies, and organizational authentication technologies that are currently available to recognize and track such deficiencies. This seeks to offer a multidimensional analysis viewpoint on IoT vulnerabilities to the reader, including the technological specifics and effects, which are intended to be leveraged for remediation goals. Inspired by the lack of IoT paradigm-related scientific (and malicious) evidence, the current study provides an emphasis on IoT manipulation from passive measurements. The current research illustrates the seriousness of the IoT problem while offering organizational knowledge resources that will inevitably assist in the mitigating mission in general. In addition to open issues and research concerns, informative conclusions, inferences, and results are revealed in the current research, which will lead to future research initiatives to resolve scientific concerns relevant to IoT security.