基于增强信誉度的联合频谱感知算法研究

Study on robust cooperative spectrum sensing algorithm based on enhanced reputation

在认知无线电多节点联合频谱感知中,当感知节点中存在恶意用户发起频谱感知数据窜改(spectrum sensing data falsification,SSDF)攻击时,联合频谱感知的性能将受到严重影响。针对上述问题,研究并建立一种具有潜伏期的攻击模型。该模型在信誉值偏低时将停止攻击并累积信誉值,并在此基础上提出基于信誉机制的加权序贯检测(weighted sequential probability ratio test,WSPRT)算法,在计算权值时,综合考虑单节点表现的可信度和稳定度。仿真比较了传统检测算法与该算法在不同攻击模式下的关键性能。仿真结果表明,与现有算法相比,该算法检测性能提升,适用于有大量恶意节点存在的环境,并可以有效抵御更加灵活的SSDF攻击。

In cognitive radio cooperative spectrum sensing,when malicious nodes exist,by sending wrong data to the fusion center,the performance of cooperative spectrum sensing will be impacted severely,this kind of attack is called SSDF.To solve this problem,this paper introduced a new smarter S-SSDF pattern,which could heal its reputation by performing like a normal node when its reputation was low.And it proposed a WSPRT algorithm based on reputation mechanism by considering both reliability and stability of the CR nodes when calculating the weights.In the simulation,it compared key indicators of traditional algorithm and the WSPRT algorithm.The simulation results show,the scheme proposed can adapt a large amount of malicious nodes,and has better robustness against SSDF attacks than traditional scheme.