A Robust Trust Model against Malicious Attacks in Heterogeneous Internet of Things

In recent years, the Internet of Things (IoT) has gained many interests and applications in various fields from the automation of industrial processes to smart environments. Despite the advantages of IoT, the reliability of connected objects is a serious challenge in these environments. The heterogeneity of objects in IoT and the imbalance of power and resources do not allow the use of a reliable trust management model in these contexts. On the other hand, the use of more general and weaker models that can be implemented in all connected objects also reduces the security of these environments. In this paper, focusing on the heterogeneity of the IoT, a two-layer trust model is proposed. In the first layer, the model allows objects to select and use a trust model appropriate to their capabilities. Then in the second layer of the model, the result of the calculations of the previous layer is aggregated and the final trust values of the objects are calculated by considering the accuracy of each object. The proposed model uses the EigenTrust algorithm to aggregate the trust from the first layer. The paper also evaluates the robustness of the proposed model compared to the base models against various attacks including dishonest behavior, on-off attack, value imbalance attack and oscillation, and the results show that the proposed model is able to identify attackers with 86% accuracy and has also improved the accuracy of the base models up to 30%.