A Spectrum Sensing Method for Cognitive Radio Networks Based on Random Tensor Theory

Spectrum sensing is very important for improving the spectrum usage of cognitive radio (CR) communications. As for the current spectrum sensing approaches face issues such as high dependence on primary user (PU) signal information, low detection accuracy under low signal-to-noise ratio (SNR) conditions, and poor reliability, etc., a new approach based on the random tensor theory (RTT) is proposed in this paper. Firstly, the spectrum sensing of CR networks is formulated as a hypothesis testing problem. Then the RTT-based detection algorithm is proposed, which involves spatio-temporal data set formulation and partition, data dimensionality increment, tensor version sample covariance matrix calculation and comparison, etc. And on this basis, the linear eigenvalue statistic (LES) is designed as the statistical index to indicate the primary user behavior. The simulation experiments validate the effectiveness and advantages of the proposed approach.