Cooperative Spectrum Sensing With Quantization Combining Over Imperfect Feedback Channels

Abstract

A practical fusion rule with quantization combining is proposed for cooperative spectrum sensing over imperfect feedback channels that are inherent in the wireless environment. Local sensor nodes generate decision symbols regarding the existence of a primary user signal by multilevel detection and send them over feedback channels to the fusion center, in which demodulated symbols are combined to draw the cooperative decision. The proposed quantization combining scheme over imperfect feedback channels is analyzed, based on which detection probability and false alarm probability are derived. Operating parameters of the fusion rule with quantization combining are optimally determined under Neyman-Pearson criterion by using a differential evolution algorithm. The proposed fusion rule is compared with other schemes in terms of the detection performance, the amount of system information required for operation, the sensitivity to the channel estimation error, and the computational complexity required for parameter optimization. It is shown that the proposed fusion rule offers the reliable sensing capability even over imperfect feedback channels with requiring small amount of system information and complexity.