Research On Wideband Spectrum Sensing Based On Compressed Measurement And Signal Reconstruction

Cognitive radio(CR)is a dynamic spectrum access(DSA)technology that can detect the frequency band quickly and accurately,which is not occupied by the primary users(PU)by sensing the surrounding spectrum environment.Without interfering with the normal communication of PU,CR allows the secondary users(SU)to access the primary band to work temporarily to achieve the purpose of improving the spectrum utilization.Spectrum sensing(SS),as the key technology to discover the free spectrum and avoid interfering with the normal communication of PU,is the premise and key of the CR technology.With the rapid development of wireless communication technology,traditional narrow-band spectrum sensing cannot meet the needs of users of the spectrum resources gradually,but wideband spectrum sensing(WSS)can provide more choices of spectrum access for cognitive users due to its wide range of perception and has become a hot research topic in the field of CR.At present,the biggest challenge of WSS technology is the problem of high sampling rate.Compressed Sensing(CS)theory shows that utilizing the sparsity of wideband spectrum can achieve WSS technology with low sampling rate.The paper is focused on the key of Wideband Compressed Spectrum Sensing(WCSS).The main research work and innovations are as follows:According to the non-unique signal reconstruction results due to the high column coherence of existing random measurement matrix,a optimization algorithm for measurement matrix is proposed in the paper.On the one hand,the algorithm reduces the global column correlation of the measurement matrix by averaging the eigenvalues of gram matrix;On the other hand,the threshold function is used to shrink the larger non-diagonal elements in the Gram matrix to reduce the maximum column correlation.And then the adequate measurement matrix is obtained by iteration.Finally,the effectiveness of the algorithm is proved by simulation experiments.And by comparing with the same algorithms,the algorithm has better optimization effect on the measurement matrix.A sparsity adaptive signal reconstruction algorithm——Differetial adaptive matched pursuit(DAMP)is proposed to solve the problem about the blind sparsity of spectrum sensing technology in practical application.Firstly,the algorithm is used to estimate the signal sparseness by atomic matching test.Secondly,the signal measurement is sorted and differential operated.The best "fault" position is found to estimate the signal support set by referring to the initial sparseness.Then the iterations are used to reduce the residuals successively until the requirements are met,and then the accurate sparseness estimates areobtained,and finally the signal reconstruction is completed.The simulation results show that the DAMP algorithm is effective in the case of unknown sparsity,and the reconstruction performance is better by comparing with the similar algorithms.A selective measurement denoising reconstruction algorithm based on noise filter matrix is proposed in the paper to solve the problem about the signal reconstruction performance which is influenced by the phenomenon of noise folding seriously.Firstly,the noise filter matrix is constructed by using the observed data and the knowledge of probability theory;Secondly,the noise filter matrix is used to assist the measurement matrix to realize the selective measurement of the sparse signal,so as to achieve the purpose of suppressing the noise folding from the source and improve the reconstruction performance.Finally,utilize the existing reconstruction algorithm to reconstruct the signal.The simulation results show that the proposed algorithm has better performance of noise reduction than the similar algorithms.