Cognitive Radio Spectrum Detection Based On Compressed Sensing

Radio spectrum is a valuable and non-renewable resources, authorized used by the State Radio Regulatory Commission in our country. This authorized spectrum allocation can avoid interruption among systems, because of incomplete use of spectrum among systems, it can waste spectrum resources easily. Cognitive Radio’s presentation is based on the above consideration. It can achieve dynamic spectrum allocation and alleviate the spectrum constraint’s problems effective. Cognitive user communication using spectrum detection technology at authorized channel by search "spectrum holes" in cognitive radio, so spectrum detection technology is the key technology of cognitive radio systems.In the broadband spectrum sensing, traditional programs is limited by A/D sampling rate, each sensing tip can detect signal within a narrow frequency range, this affects the broadband spectrum sensing detection performance greatly. However compressed sensing technology proposed in recently years, can solve this problem effectively. Compressed Sensing theory’s premise is the signal is compressible or is sparse under a transformation group, spectrum resource’s sparsity in the frequency domain is just to meet this condition. So we can combine the compressed sensing technology and spectrum detection technology to improve the efficiency of spectrum detection.In this paper, we study the cognitive radio spectrum sensing based on compressed sensing. First of all, to make the signal in the frequency domain has sparsity by the discrete Fourier transform; then, make the Gaussian random matrix to observe this sparse signal; the cognitive user use BP, MP, OMP, CoSaMP, ICoSaMP algorithm reconstruct original signal; finally, spectrum detection by energy detection method.Simulation results show that, ICoSaMP algorithm can accurately recover the original signal in noise and no-noise environment, compared with other recovery algorithms, ICoSaMP algorithm can obtain ideal reconstruction error. ICoSaMP algorithm can achieve an ideal compromise between computational complexity and probability of detection in the spectrum detection. And compared with BP, MP, OMP and other traditional recovery algorithm, ICoSaMP algorithm has no need to know the signal’s sparsity in the beginning, This is more suitable for the actual communication environment.Multi-node cooperative spectrum sensing can solve the drawbacks of single-node detection, it can not only improve the detect performance but also reduce the requirements of sensing conditions. But it can also cause the additional overhead, such as latency, energy loss and so on. We analysisd the "hard" decision algorithm in this paper.