Optimized Design Of Wideband Compressed Spectrum Sensing In Cognitive Radio Network

With the rapid development of wireless communication technology, the demand for wirelessspectrum resources increased dramatically, while the fixed spectrum allocation system has caused ascarcity of spectrum resources globally. Technology of cognitive radio improves the wirelessspectrum utilization by accessing "spectrum holes" dynamically. Cognitive radio can reuse thewireless spectrum. Spectrum sensing technology is not only a powerful tool to identify "spectrumholes", but also the foundation of the entire system to achieve cognitive radio spectrummanagement and sharing. The compressed sensing approach is adopted to accomplish compressivesampling at lower than Nyquist sampling rate while significantly reduces the pressure of thehardware of broadband spectrum sensing.This thesis studies the problem of broadband compressed spectrum sensing, mainly in thefollowing two aspects with depth study.Firstly, in order to reduce the complexity of the broadband spectrum sensing, and achieve fastperception, this thesis presents a new compressed spectrum sensing scheme. The new scheme is aprediction way based on the Pseudo difference signal to recompress the observed signal. Thecognitive network predicts the occupancy probability of authorized channel by learning. Based onthe probability of the occupancy of authorized channel, the authorized channels are divided into twoparts, and been sensed twice. Finally, the computational complexity are reduced twice.TheSimulation results of the new scheme show that it can reduce the computational complexity of thebroadband spectrum sensing to improve the broadband spectrum sensing efficiency and achievesrapid perception without reducing the accuracy of the spectrum sensing.Secondly, in order to solve the problem of block-based spectrum sensing without the boundaryconditions of the non-zero signals. This thesis first introduces the block-based compressive theory,then in order to solve the problem of the sparse signals of unknown the conditions of thenon-zero signals, a modified spectrum sensing scheme for wideband cognitive radio based onblock-sparse orthogonal matching pursuit is proposed, which combines traditional single-pointreconstruction algorithm and block sparse reconstruction algorithm together. Simulation resultsshow that the modified scheme can still detect the occupancy of the channel in a very short timeunder a low detection error probability without known the boundary of the non-zero sub-block.