Research on Robust Cooperative Spectrum Sensing Based on Clustering Algorithms

• 1、School of Electronics and Information Engineering, Liaoning University of Engineering and Technology, Huludao, Liaoning 125105, China

Abstract：The complex communication environment and diverse malicious attack strategies pose challenges to cognitive radio systems. In response to these challenges, this paper proposes a robust cooperative spectrum sensing method based on clustering algorithms. The paper establishes a multi-antenna cooperative spectrum sensing network model, optimizing signal sensing accuracy and robustness through data sharing among secondary users. To address SSDF (Sybil-based Data Falsification) attacks by malicious users, two data fusion schemes are proposed: the DF-medoids method based on the K-medoids clustering algorithm, which iteratively determines the center points in the energy vector set to improve the accuracy of data fusion; and the DFMS-medoids method based on the Mean-shift clustering algorithm, which further optimizes data fusion through multiple iterations, enhancing the system's robustness. The cooperative spectrum sensing framework constructed based on these methods can significantly improve system detection performance in adverse signal-to-noise ratios and malicious attack environments. Simulation results show that the robustness of the DFMS-medoids method surpasses that of the DF-medoids method and the traditional mean fusion algorithm, with the Mean-shift clustering algorithm performing slightly better than Fast K-medoids in terms of classification effectiveness. Ultimately, experiments verify that this framework maintains high-precision spectrum sensing capabilities in complex environments.

Keywords： Cognitive Radio System； Cooperative Spectrum Sensing； Clustering Algorithm； K-medoids； Mean-shift； SSDF Attack