A novel multiple-level trust management framework for wireless sensor networks

The distributed deployment nature of wireless sensor networks (WSNs) poses a challenge to the security of node cooperation in them as it is difficult for WSN to ensure that all nodes can recognise a huge number of other individual nodes and select appropriate and trustworthy nodes for cooperation. Node cooperation may therefore be launched in an unreliable environment and might be vulnerable to attacks. Consequently, the security of nodes is of paramount importance for the proper operation of WSNs. The distributed trust management scheme is a feasible solution. With a view to making improvement on the existing trust management mechanisms, we in this paper propose ML-TRUST, a multiple-level trust management framework for trust management in WSN in which three levels of trust are used to establish trustworthy relationships among nodes for their cooperation, namely, (1) a subjective trust, which is defined as belief and is proposed with respect to three aspects: past judgements, witness evidence, and capacity evaluation; (2) an objective trust, which is defined as reputation and is proposed with two factors, number of functioning communities and weighted judgements by rating nodes’ reputations, being introduced in reputation rating, and with several rules and fraud factor tests being given to prevent reputation rating from malicious attacks, and (3) the recommended trust method, which is proposed to obtain trustable impressions from strange recommendations with, in connection, several consistency factors being presented to determine the trustworthiness of a recommendation. Besides using a set of lemmas and theorems to back up our ML-TRUST framework, we also list the results of a series of simulation tests to further verify the performance of our mechanism.