Research On Spectrum Sensing Algorithm In Cognitive Radio

Fixed spectrum allocation strategy leads to shortage of spectrum resources.Cognitive radio is an efficient spectrum management technology,which can effectively improve spectrum utilization.Spectrum sensing is one of the key technologies of cognitive radio,and it is a hot point in the field of wireless communication.The research on this thesis is focused on the spectrum sensing algorithms based on random matrix theory and power spectrum.Firstly,spectrum sensing algorithm based on random matrix theory is studied.In order to reduce the complexity of spectrum sensing algorithm based on random matrix theory,a spectrum sensing algorithm(2EDC)is proposed,which combines the maximum and minimum eigenvalue(DMM)algorithm with energy detection algorithm.The DMM algorithm is used for spectrum sensing at low SNR,and the energy detection algorithm is used at high SNR so as to reduce the complexity of the algorithm.The spectrum sensing performance of 2EDC algorithm is better than energy detection algorithm and DMM algorithm.In order to improve the spectrum sensing performance at low SNR,the background noise signal is collected and its maximum eigenvalue is obtained.And then combining it with the maximum eigenvalue of the received signal,the double eigenvalue spectrum algorithm is presented by using a bistable stochastic resonance(BSR)system.Simulation results show that the algorithm can obtain higher detection probability at low SNR.Secondly,the spectrum sensing algorithm based on power spectrum is studied.According to the properties of power spectrum function and the concept of Rayleigh entropy,the principle of spectrum sensing algorithm based on power spectrum is analyzed theoretically,and the spectrum sensing algorithm(PMMD)based on the difference of the maximum and minimum of power spectrum is proposed.And then the mathematical expressions of detection threshold and detection probability are derived.Simulation results show that PMMD algorithm has good spectrum sensing performance in additive Gauss white noise(AWGN)channel and Rayleigh fading channel.Because of noise influence,the minimum point of power spectrum is not fixed,so the mean value of the middle part of power spectrum is used to estimate the minimum,and the minimum estimation accuracy of PMMD algorithm is improved.Then the spectrum sensing(PSMA)based on the mean of power spectrum is proposed.The effect of minimum volatility on algorithm performance is eliminated.And the mathematical expressions of detection threshold and detection probability are given.Simulation results show that the spectrum sensing performance of PSMA algorithm is betterthan PMMD algorithm.Finally,the spectrum sensing algorithm with combating the noise uncertainty is studied.Noise power estimation error reduces the performance of spectrum sensing algorithms based on random matrix theory(such as DMM),so the spectrum sensing algorithm(DED)based on double eigenvalue and energy is proposed.The mathematical expressions of detection threshold and detection probability are given.Simulation results show that the DED has good performance against noise uncertainty.Because the ratio between the maximum and minimum of power spectrum of the received signal is different when the primary user signal is present or absent,the spectrum sensing algorithm(PSRA)based on average of the middle part of power spectral is proposed,to improve the performance against noise uncertainty of PMMD algorithm and PSMA algorithm.The PSRA uses the ratio between the maximum and middle part of power spectrum as detection statistic.The detection threshold of the algorithm is not correlative to the noise power,and the robustness of the algorithm to the noise uncertainty is improved.Simulation results show that the performance of PSRA algorithm is better than PSMA algorithm and presented the spectrum sensing algorithm based on power spectrum in AWGN channel and Rayleigh fading channel.