Research On Wide-band Spectrum Compressive Sensing Algorithm In Cognitive Radio Network

The contradiction between the growing spectrum demand and limited spectrum resources becomes one of the main factors that restricting the development of wireless communication systems. As the approach to overcome this contradiction, cognitive radio technology has the ability to perceive and adapt to the surrounding environment, so it can realize the highly reliable communication at anytime and anywhere, and the effective utilization of the spectrum resources. Spectrum sensing, dynamic spectrum assignment and wireless spectrum management are three key technologies of cognitive radio technology, and as spectrum sensing technology is the foundation of these three technologies, it becomes the focus of cognitive radio technology researches.Most current researches about cognitive radio are centered on the spectrum sensing technologies of narrow wideband, but considering that wideband spectrum sensing has more advantages in meeting cognitive radio network system's hard real-time and high reliability requirements, we do some researches about wideband spectrum and propose three wideband spectrum compressive sensing algorithms that suitable for multiple cooperating users.The main research content and results are as follows:Firstly, it introduce the research background, research purport and the status of cognitive radio technology and spectrum sensing technology. And the classification of cognitive radio spectrum sensing technologies is made based on sensing object and sensing users, the comparison is given between the typical sensing technologies and sensing schemes.Then, taking the strict demands of detection performance in cognitive radio network, this paper proposes the independent compressive sensing algorithm, cooperative compressive sensing algorithm and clustering-based cooperative compressive sensing algorithm. After a thorough discussion about the methodology, mathematic description and execution steps of the algorithm, and based on the details of compressive sampling, data fusion, spectrum reconstruction, channel decision and other steps, it gives the theoretical analysis of the advantages and disadvantages of these three algorithms.Finally, as the optimization problem of compressive sensing is a NP-hard problem, we use the CVX toolbox embedded in matlab to simulate the proposed algorithms. The results show that the clustering-based cooperative compressive sensing algorithm is optimal in increasing detection ration, detection efficiency and lowering false-alarm probability of spectrum sensing. This algorithm can be more effectively prevent the interference to main users and increase spectrum utilization, so it can meet the hard real-time and high reliability requirements on spectrum sensing of cognitive radio network to a certain extent.