A Novel Trust Computation Scheme for Internet of Things Applications

Quality of service and trustworthiness of data is of high importance for decision making in Internet-of-Things (IoT) applications. Malicious nodes and devices may compromise the quality of service and experience for other nodes through providing invalid data and evaluations. Hence, a trust management system to assess the trust level of users and gathered data is deemed to be essential to every IoT system. The current approach in the literature for computing the trust level is entity-centric trust in which the trust level of end users are estimated. However, the trustworthiness of data is equally important in many applications. In this paper, we propose, Trusty, a hybrid trust computation approach, aiming at trust assessment for both entities as well as data. In our proposed approach, a Bayesian learning method is used for computing the entity trust, while Dempster-Shafer theory is exploited to data fusion and data trustworthiness assessment. We implement Trusty in a smart parking system scenario to investigate the performance of our model in the different settings for misbehavior nodes and faulty sensors. As shown by the extensive simulation experiments, Trusty outperforms the competing approaches in terms of convergence for both data trustworthiness and entity trust.