论文部分内容阅读
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,本文提出一种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,本文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,本文所提方案具有更好的检测与防御性能。
When there is a cooperation between malicious users in the cognitive users, especially malicious users, it will bring great harm to cooperative spectrum sensing. In order to effectively resist malicious behavior, this paper proposes a distributed intelligent intrusion prevention scheme that combines the credibility model with the consistency fusion. In each iteration, rewards and punishments should be used to reward or punish the creditworthiness of cognitive users, and then combine the credit value with the fusion factor in the consistency fusion and work together in the process of consensus fusion. Under the program, intelligent malicious users will eventually choose to abandon malicious attacks. The simulation results show that the proposed scheme can effectively resist the attacks of multiple malicious users (collaborative / non-cooperative). Compared with the three existing defense mechanisms, the proposed solution has better detection and defense performance.