Technical Research Of Spectral Sensing Based On Compressed Sensing And Covariance

The utilization of television broadcasting, aeronautical telemetry and other authorized spectrum resources are very low. Cognitive radio technology is considered a new method to solve the limited available spectrum and very low spectrum efficiency. Spectrum sensing is the key technology of cognitive radio. Spectrum sensing technology in the future needs to sense a wide band of frequencies quickly and accurately, in the order of several GHz, which is a huge challenge for working in Nyquist sampling rate of the traditional spectrum estimation method. Based on analyzing the existing spectrum sensing technology indepth, it mainly including the following three aspects.First of all, covariance absolute value sense algorithm use fixed threshold and fixed false alarm probability, it can not guarantee at any time to achieve minimized spectrum sensing error. This paper proposes the optimization method of the adaptive threshold in order to overcome this shortcoming. Simulation results show that the proposed scheme effectively reduce the spectrum sensing error and increase the detection probability, particularly in the case of low SNR, the spectrum sensing performance improvements are significant.Secondly, most current collaborative detection assume that all the cognitive users have the same average SNR, in fact, cognitive user have different signal to noise ratio because of different geographic location and different environmental. In order to overcome this shortcoming and combine ouble threshold collaborative detection this paper proposed based on optimal SNR the double threshold collaborative detection algorithm. Experimental results show that when received signal energy in the range of judgment delay interval is lager, the performance collaborative detection improvement is more significant.At last, wideband sensing need desire a very large sampling rate, which is a challenge for existing spectrum sensing method. This paper utilize the sparse characteristics of the wideband signal spectrum and practical spectrum group sparsity characteristics, the compressed sensing algorithm is applied to wideband spectrum sensing, reconstruct sub-nyquist signal then perform spectrum sensing, wideband spectrum covariance sensing algorithm based on compressed sensing is proposed. Compared to SCS, not to affect the detection performance at the maximum, this method decreases the sampling rate, eases the pressure of the signal collection side, reduces the system costs.