Cooperative Spectrum Sensing Algorithm Based On Random Matrix

For 2020 and the future,the mobile internet and internet of things business will grow rapidly,applications with ultra-high flow density,ultra-high number of connections,ultra-high mobility will become more and more pervasive.In face of the increasing performance and capacity requirements for communications systems,inefficient static spectrum allocation is no longer applicable.Cognitive radio(CR)technology,which has the characteristics of dynamic sensing and opportunistic access,can meet this requirement well and is widely considered to be one of the key technologies to improve spectrum efficiency and solve the problem of spectrum crisis.Spectrum sensing is the prerequisite and basis for cognitive radio.The main task of spectrum sensing is to detect whether the authorized band is idle and its detection performance is critical to the entire cognitive radio system.Focusing on spectrum sensing problems in CR,the cooperative spectrum sensing algorithm based on the statistical distribution of the sample covariance matrix elements and the spectral distribution of wishart matrix are mainly studied in this thesis.The main contents are summerized as follows:(1)Starting from the practical application,the thesis first analyzes the shortcomings of single-user detection,and then according to the detection performance requirements in relevant standards of CR,pointing out the necessity of applying random matrix theory to cooperative spectrum sensing.(2)The related distribution characteristics of sampling covariance matrix and wishart matrix are studied,especially the the sum of absolute values of the non-diagonal elements and diagonal elements,in addition,the non-asymptotic spectral distribution of wishart matrix is studied.(3)For the multi-user cooperative sensing scenarios,the disadvantages of the existing detection algorithms based on sample covariance matrix are analyzed.According to the statistical distribution of sample covariance matrix,the probability density function(PDF)and the cumulative distribution function(CDF)of the sum of the absolute values of the non-diagonal elements and diagonal elements are derived.Then,using the ratio between the sum of non-diagonal elements and diagonal elements as test statistic,an improved covariance absolute value(ICAV)sensing algorithm is proposed.Theoretical and simulation results show that the ICAV algorithm has a more accurate decision threshold than the existing covariance absolute value based algorithm,and has better detection performance than the other algorithms.(4)Aiming at the shortcomings of the cooperative sensing algorithm based on the asymptotic spectral distribution of random matrices,the non-asymptotic spectrum theory of random matrices is applied to cooperative spectrum sensing,and an exact maximum minimum eigenvalue difference(EMMED)cooperative sensing algorithm is proposed.For any given numbers of cooperative users K and sampling points N,the exact PDF and CDF of the difference between the maximum and minimum eigenvalues are derived.Then,an accurate decision threshold is designed by using the distribution function.Theoretical and simulation results show that the decision threshold of the proposed algorithm is more accurate than the existing asymptotic threshold,and has better detection performance than the other eigenvalue-based algorithms.