基于可信度计算的分布式共识合作感知算法

对于分布式的认知无线网络，由于不存在融合中心节点，次级用户节点之间的合作感知往往 采用信息交互的渠道进行，其中一种基于共识合作的感知机制受到广泛的研究，但这种机制 在鲁棒性方面存在一定的缺陷，当恶意节点存在时，其错误信息将影响局部的感知判决结果 ，为此，提出一种基于感知节点可信度的共识合作感知机制。在该机制中，各节点对邻居节 点的可信度进行计算，并把计算出的可信度值发送给其他节点，通过对各节点可信度的累加 计算，最终计算出各节点的可信度，各节点以此决定是否与其邻居节点合作以及如何合作。 仿真结果证明，在恶意节点存在的情况下，该算法在感知性能和收敛速度上都较未改进算法 有不同程度的提升，减轻了不可靠节点对合作感知结果的影响。

A Reputation-computed Consensus Algorithm for Distributed Cooperative Spectrum Sensing

In distributed cognitive radio (CR) networks, the cognitive users usua lly actualize cooperative sensing through exchanging information with neighbors. In the existing schemes, a consensus-based cooperative scheme has been promised , but the cooperative scheme has a weak robustness when there are suspect users i n the networks. So this paper proposes a reputation-based scheme to classify the CR nodes for successful consensus cooperation. In the proposed scheme, each user will co mpute the reputation values of its neighboring users and send the values to othe r users, simultaneously each user will receive the computed result from other us ers, and then the reputation values will be accumulated for each user. Finally the uniform reputation value for each user is obtained, and the reputation value s are used by each user to decide whether to cooperate with its neighbors and how to cooperate. The simulation results show that when there are suspect users, the detecting performance of the proposed scheme outperforms the common scheme, and also the converging speed, so the proposed scheme can mitigate the harmful e ffect from the suspect nodes.