Cooperative Spectrum Sensing Method against Spectrum Sensing Data Falsification Attack

Spectrum sensing is a key technology of cognitive radio, by which Sensing Users (SUs) are able to scan and detect a specific spectrum to obtain patterns of spectrum usage and discover spectrum holes. Cooperative Spectrum Sensing (CSS) can effectively improve the efficiency and accuracy of spectrum sensing in Flying Ad-hoc NETwork (FANET). However, the dynamic topology and the lack of authentication management make it difficult for FANET to resist Spectrum Sensing Data Falsification (SSDF) attack, thus one distributed CSS method against SSDF is proposed. In this method, each SU is regarded as an agent, which performs spectrum sensing independently and interacts with neighbor nodes through reinforcement learning, where a clustering method of Density-Based Spatial Clustering of Applications with Noise (DBSCAN) is introduced to calculate rewards and guide neighbor selection. Each SU is used as a local fusion center for data fusion where a sliding window reputation mechanism is proposed to assign consensus weights. Experimental results show that in various SSDF attack scenarios, the proposed method can effectively identify and resist the interference of malicious users, as well as improving the efficiency and accuracy of CSS.