可信的反馈声誉保护协作频谱感知软判决算法

协作频谱感知中信任机制的引入，起到了抑制恶意用户频谱感知数据伪造( SSDF)攻击行为的作用。然而，数据融合中心不加区分地接收协作感知结束后的反馈信息，为恶意用户带来了实施“掺沙子”攻击的机会。恶意用户向数据融合中心反馈错误的主用户频谱状态，使信任机制不能得出准确的信任值。为此，提出了一种基于反馈声誉的信任机制，考虑反馈中的个体性特征，引入反馈声誉的思想来量化认知用户信任值。同时，将信任值量化结果用于权重经典软判决算法———序贯概率比检测( SPRT)算法，消除SSDF恶意用户参与软判决数据融合的影响，形成可信序贯概率比检测算法( FSPRT)。仿真结果表明FSPRT算法的性能优于传统SPRT算法，能有效降低网络信任值计算误差，并保持较好的感知性能。

A trusted cooperative spectrum sensing soft-decision scheme using feedback reputation protection

The trust mechanism can suppress spectrum sensing data falsification( SSDF) attack caused by malicious users in cooperative sensing spectrum. However,the data fusion center receives the feedback in-formation indiscriminately at the end of cooperative spectrum sensing,which offers the chance for malicious users to launch“sanded” attack. By using the wrong feedback of malicious users about the status of prima-ry users,trust mechanism cannot get an accurate trust value. A novel trusted cooperative spectrum sensing soft-decision scheme using feedback reputation protection is proposed to avoid“sanded” attack. By consid-ering the individuality characteristic in the feedback,the idea of feedback reputation is introduced to quan-tify cognitive user’s trust value,which can be used to weight classic soft-decision sequential probability ra-tio test(SPRT) algorithm,and thus eliminating SSDF attack in the soft-decision data fusion. Simulations show that the novel scheme outperforms the conventional SPRT soft-decision schemes. The FSPRT scheme can reduce trust value error effectively,and maintain a better sensing performance.