Researches On Spectrum Sensing And Spectrum Applicability Analysis In Cognitive Radio

Spectrum sensing and its applicability analysis are the key technologies in cognitive radio,have been attracted as a great research attention of wireless communications in recent years.The researches of cognitive radio in the paper,supported by the National High Technology Research Development Program of China(863 Program)and others findings,have some important and valuable prospects in theories and practices.The paper contributes the following innovative achievements based on the deeply researching of cognitive radio,i.e.,spectrum sensing technologies,sensing channel performance estimation,data fusion strategies,spectrum feature and spectrum applicability,etc.For the problem that some existed spectrum sensing technologies may need to know prior knowledge about wireless environment and primary user,this paper proposes a local spectrum sensing algorithm,called energy-autocorrelation detection algorithm(E-A detection),which eliminates the effect of noise based on energy statistic and autocorrelation statistic,the functions of false probability and detection probability are deduced by comparing the numerical characteristics and distributed functions of this two statistics.According to Neyman-Pearson criterion,the expression of detection probability is attained under the conditions of the threshold of false probability.Simulations testify that the algorithm is only sensitive to channel gain of sensing user,and the detection performance can be dynamically adjusted by changing the sampled parameters of sensing users.The algorithm realizes the spectrum sensing without any prior knowledge about wireless environment and primary user.For the problem that local spectrum sensing can not process shadow fading and hidden terminal,this paper proposes a cooperative spectrum sensing model and channel estimation algorithm based on cross-correlation matrix.The algorithm,called cross-correlation matrix(CCM)estimation,which analyzes and compares the features of off-diagonal elements and diagonal elements based on the cross-correlation matrix of sensing users,estimates the ratio of channels gains,signal to noise ratio and other channel characteristics.When the estimated values of channel characteristics are substituted into the detection probability expression of E-A detection,the problem,which is sensitive to channel gain,will be overcome.Simulation testifies that the performance of CCM estimation is superior to other estimation algorithms.The realization of CCM estimation is only based on the sampled signals of sensing users without any prior knowledge of primary user.For the problem that traditional fusion strategies do not achieve the selectivity of secondary users in data fusion,the paper proposes a data fusion strategy based on the selectivity of sensing users,which includes two realized algorithms.Algorithm(1)is that all of sensing users are allowed to participate in data fusion in fusion center,but they are given different weights in term of their channels gain;Algorithm(2)is that the sensing users with little attenuation are allowed to participate in data fusion in fusion center,while others with serious attenuation are excluded from fusion center.Simulations testify that the performance of two algorithms are greatly superior to "AND" criterion and "OR" criterion,and the performance of Algorithm(2)is superior to that of Algorithm(1).For the problem that how to evaluate the idle spectrum feature and applicability of business type by means of cooperative spectrum sensing network,the paper exploringly proposes a mechanism which analyzes the applicability of idle spectrum to meet different users requirements for the quality of service.An index table of idle spectrums performance(ITISP)is constructed,which is composed of three categories:(1)spectrum parameters:center frequency,bandwidth;(2)band occupied features of primary users:spectrum availability,requirement of service time;(3)channel characteristics:Doppler spread,channel capacity and loss.The matrix,which is combined with the type of business and determine whether the idle spectrums to meet secondary users demand,is constructed,and the number of available idle spectrum of secondary users are attained by comparing the demand values of secondary users and available values of ITISP.A search strategy of spectrum applicability without considering spectrum policies is established in the light of the matrix elements and the number.Simulations testify that the strategy is superior to the random spectrum assignment in term of collision rate and spectrum efficiency.Finally,the paper summarizes the whole research,the system model of spectrum sensing and spectrum applicability is given.The future research directions and prospects of cognitive radio are discussed in the last part of the paper.