Research On Spectrum Sensing In Cognitive Radio Networks

Wireless spectrum resource is becoming increasingly scarce with the rapid development of wireless communications services. Wireless spectrum is allocated for use by the nation today, and Wireless spectrum is divided into two parts: the licensed frequency band (LFB) and unlicensed bands (UFB). The fixed spectrum allocation approach has seriously affected the development of wireless networks. The utilization rate of licensed spectrum is very low in the time domain, frequency domain and space domain, while the unlicensed band is overcrowded seriously. There is a pressing need for a new wireless spectrum use pattern due to the limited spectrum resources and low utilization of the spectrum. Cognitive radio technology is proposed to solve this problem.Spectrum sensing technology is a basic function of cognitive radio systems, and the premise of spectrum management and spectrum sharing. This paper mainly studies spectrum sensing in cognitive radio. We mainly study a single cognitive user's local spectrum sensing and multi-user collaborative spectrum sensing. We design the spectrum sensing algorithms respectively.Firstly, the case of a single cognitive user's local spectrum sensing is considered. In this part, we propose a spectrum sensing algorithm based on frequency-domain entropy. The main idea is based on the different spectrum distributions of communication signals and environmental noises, and information entropy is introduced to describe the characteristics of the spectrum distribution. The frequency domain entropy of communication signal is smaller than that of environmental noise. We use the threshold method to distinguish whether the primary user is present. Furthermore, we design dynamic threshold selection algorithm to set detection threshold dynamically.Secondly, we further study multi-user cooperative spectrum sensing. In this part, we propose the collaborative spectrum sensing algorithm based on signal correlation and spectrum consistency in cognitive radio networks. When the primary user is active, the signals received by cognitive users from primary user are correlated. Otherwise the signals received by cognitive users are uncorrelated. Furthermore, we analyze the eigenvalue distribution of cognitive user's correlation coefficient matrix, and design the spectrum sensing algorithm based on the eigenvalue distribution of cognitive user's correlation coefficient matrix. We also give dynamic threshold selection algorithm.