Distributed trust management in pervasive computing
An effective trust management technique plays a vital role in evaluating relationships among devices in pervasive computing. In this thesis, we propose two trust management schemes for pervasive computing environments. First, we propose a deterministic approach that aims at establishing trust relationships among devices using direct and indirect computation methods. Recommendations and trust updating mechanisms are used to increase the reliability of trust computations. Second, we propose a probabilistic approach that considers trust value as a probability that a device will provide satisfactory interactions. An iterative filtering method is employed to eliminate the effect of false recommendations, while a weighting method is used to capture the effect of time on the current behavior of devices. We have carried out performance evaluations using simulation experiments. The experimental results demonstrate that the proposed approaches improve performance and ensure the security of interactions. The results also show that the probabilistic approach adapts faster to changes in the environment than the deterministic one does, resulting in more accurate trust computations.