基于信誉与共识的分布式智能入侵防御方案

当认知用户中存在恶意用户,特别是恶意用户间存在协作行为时,将给协作频谱感知带来极大危害。为有效抵御恶意行为,提出将信誉模型与一致性融合相结合的分布式智能入侵防御方案。每一次迭代过程,需要依据奖惩机制对认知用户的信誉值进行奖励或惩罚,进而将信誉值与一致性融合中的融合因子相结合,并共同作用于一致性融合过程。该方案下,智能的恶意用户将最终选择放弃恶意攻击。仿真结果表明,所提方案能有效抵御多个恶意用户(有协作/无协作)的攻击。与现有的3种防御机制相比,该方案具有更好的检测与防御性能。

Distributed intelligent intrusion prevention scheme based onreputation and consensus

In cooperative spectrum sensing, some cognitive users may be malicious. More seriously, malicious users may adopt cooperative behaviors, which will bring great harm to spectrum sensing. In order to effectively resist malicious behaviors in cooperative spectrum sensing, a reputation and consensus based distributed intelligent intrusion prevention scheme is proposed. During each iteration in the proposed scheme, the reputation values of cognitive users are updated, in which honest users will be rewarded and malicious users will be punished. Then, the reputation values of cognitive users will be combined with the fusion factor in the consensus algorithm to achieve the consensus process. Intelligent malicious users will eventually choose to abandon malicious attacks. Simulation results show that the proposed scheme can effectively defend against non-cooperative attacks and cooperative attacks with multiple malicious users. Compared with the existing three defense mechanisms, the proposed scheme has better detection and defense performance.