Research On Trust Model In Heterogeneous Wireless Integrated Networks

Heterogeneous Wireless Networks (HWNs) are expected to be the trends of next generation wireless network, which can meet users'demand mostly. However, it also brings some new challenges to the security. The trust ideas provide a new way to address the challenges. At present, there are three main problems in the trust researches. Firstly, the subjective characteristics of trust are not comprehensively described in the trust models which evaluate the trustworthiness of the service networks. Secondly, it is hard to efficiently predict the creditability of the access users. Thirdly, there is a short of effective study in restraining the false evaluation from the vicious entities. In order to solve these problems, this dissertation is devoted to research on the trust model in HWNs.Based on the known work, the dissertation presents two kinds of trust models, and proposes the enhanced security technology of these models. The main achievements of this dissertation are as follows:1. A trust model based on cloud model (TMBCM) is presented, in which the subjective characteristics of trust is comprehensively described. With the established trust level clouds, multi-attribute evaluation clouds and the main trust cloud, the trustworthiness of service network is effectively assessed. Finally, the dissertation presents a service scheduling mechanism which is applicable to the TMBCM model.2. A dynamic trust model based on user behavior prediction (DTMBP) is proposed. Firstly, the dissertation presents a Bayesian network model. Through the model, user's possible behavior can be predicted based on the statistic characters of user's historical behavior. And then, taking the dynamic character of user's behavior into consideration, the DTMBP model makes the access decision based on the Markov chain which improves the veracity of the user behavior prediction.3. The enhanced security technology based the closeness is proposed. The closeness of the evaluation results between the recommenders and the trust evaluation entity is used to judge whether the recommendation trust is advisable. Simulation results show that the enhanced security technology restrains the false evaluation from the vicious entities, and improves the stability of the trust models.