AutoTrust: A privacy-enhanced trust-based intrusion detection approach for internet of smart things

Abstract

Internet of Cloud Things (IoCT) is a new era of technology introduced as an extension to Internet of Things (IoT). IoT itself is a complex and heterogeneous network that allows access to the cloud that raises numerous privacy and security challenges. When devices access the cloud for storage or services that are provided by other cloud services, it is important to identify and eliminate malicious service providers. Different encryption approaches have been proposed, however, these cause more energy consumption. In the IoT scenario, several nodes are located in a remote area where a consistent supply of energy is impossible. To address this challenge, a recurrent neural network (RNN)-based autonomic trust management approach, named AutoTrust, is proposed in this paper, which can predict the malicious behaviors of nodes and eliminate them. The proposed mechanism maintains a set of standards provided by the trustors offering an additional tier of security. A novel dataset consists of 70,000 values has been utilized to train and test the model. The AutoTrust is evaluated and compared with the existing mechanisms, where the results show that the proposed mechanism can provide better privacy and security.