Research On Single Node Spectrum Sensing Method In Cognitive Radio

With the rapid development of computer technology, semiconductor technology and mobile communication technology, the focus of wireless communication services has transferred on the services that need broadband and high speed download, and it made that spectrum resource scarcity issues became more serious. Under the spectrum fixed allocation system, the utilization of most spectrum resources is low, and thus we need to improve the utilization of spectrum resources. Cognitive radio is a radio that can change its transmitter parameters based on interaction with the environment in which it operates, and thus it can improve the utilization of spectrum resources and realize high reliable communications. The key technologies of cognitive radio physical layer include spectrum sensing, spectrum analysis and decision, spectrum sharing. Spectrum sensing is the core technology of cognitive radio and is the premise and foundation of spectrum management and sharing. The purpose of spectrum sensing is finding spectrum holes, while avoiding interference to primary user.Spectrum sensing can be divided into user transmitter sensing, cooperation sensing, spectrum sensing based on interference temperature and reception. User transmitter sensing also known as single node transmitter sensing, it contains energy sensing, matched filter sensing and cyclostationary sensing, and most research are based on the three traditional methods. In order to solve the traditional detection's problem that poor property of anti-noise, highly computational complexity and difficult to determine the decision threshold, the thesis proposed two new improved methods of spectrum sensing as below:Firstly, in order to solve the traditional energy detection algorithm's problem that the property of anti-noise is poor, this thesis proposes a novel spectrum sensing method for cognitive radio based on singular value decomposition (SVD). Theoretical analysis and computer simulation results show that the detection probability of the proposed spectrum sensing technique is higher under the same signal to noise ratio (SNR) and false-alarm probability, and the ability of anti-noise has been improved.Secondly, in order to solve the traditional cyclostationary detection's problem that highly computational complexity and difficult to determine the decision threshold, this thesis proposes a novel spectrum sensing method in cognitive radio of joint cycle correlation spectral coefficients based on multiple cyclic frequencies. The method applies the Chair-Varshney data fusion criteria to jointing cycle correlation spectral coefficients at multiple cyclic frequencies. Theoretical analysis and computer simulation results show that this method make full use of the cycle correlation spectral coefficients'information at multiple cyclic frequencies, so the detection probability is high under the lower SNR, the ability of anti-noise and anti-noise uncertainty has been improved, and the detection performance improves with the increase of the number of cyclic frequencies.