Research On Cognitive Radio Spectrum Sensing Method Based On Information Geometry

Cognitive Radio?CR?is a new kind of intelligent radio technology,which can realize reliable wireless communication and improve the utilization of spectrum resources by continuously sensing changes in the external wireless environment and autonomously adjusting transmission parameters to adapt to changes in the external environment.Spectrum Sensing plays a central role in cognitive radio technology.It requires precise,efficient,and reliable detection of available spectrum in a complex external wireless environment.It can improve the utilization of spectrum resources by making full use of the idle periods of licensed frequency bands.The theory of information geometry is an emerging theoretical system in recent years that can be used to deal with statistical detection problems.Based on the analysis of the research status of spectrum sensing,in order to improve the performance of spectrum sensing in different environments,this paper studies the spectrum sensing technology based on the information geometry theory.By analyzing the statistical characteristics of wireless spectrum signals,the statistical detection problem is mapped to geometric problems on the manifold to achieve spectrum sensing.The main research contents and innovative results achieved are as follows:The paper first briefly explained the basic concepts and research status of cognitive radio systems,and introduced some existing spectrum sensing technologies in cognitive radios.For the types of spectrum sensing technology,the basic principles of several classical single-user spectrum sensing methods are analyzed and their advantages and disadvantages are compared.Then some multi-user spectrum sensing solutions and data fusion methods are discussed.Finally,some basic concepts of information geometry theory are introduced,and the feasibility of spectrum sensing using information geometry theory is analyzed.A method of constant false alarm rate spectrum sensing based on information geometry is proposed.By explaining the measurement methods,calculation formulas,and matrix mean values on statistical manifolds,the distances between different probability distributions on statistical manifolds are defined,and the statistical characteristics of perceptual signals are mapped to geometric characteristics on manifolds to solve the spectrum sensing problem.The method needs to estimate the noise environment first,and then realize spectrum sensing by calculating the difference between the corresponding two points on the manifold that perceived wireless spectrum signal and the noise environment,respectively.Finally,the spectrum sensing performance of the method is analyzed by simulation experiments.Two improved spectrum sensing methods based on information geometry are further proposed to improve spectrum sensing detection performance.One is the spectrum sensing method based on distance detector.Based on information geometry theory,we analyze the statistical characteristics of two kinds of obs erved data,₀?noise only?and₁?noise mixed with signals?,and correspond to two pointsand+on statistical manifold S respectively.By calculating the difference between the corresponding points of the perceived spectrum signal on S and the distance between these two points,the purpose of spectrum sensing is realized.The other is an improved spectrum sensing method based on the k-medoids clustering algorithm.In the process of spectrum sensing,whether there is a difference in the s tatistical characteristics corresponding to the presence or absence of the primary user signal sense data corresponds to points at different positions on the statistical manifold.Firstly,the distance features of the spectrum signal on the statistical man ifold are calculated by statistical metric tools.Then the k-medoids clustering algorithm is used to train the distance features to obtain a classifier.Finally,the distance features calculated by the perceived signal are put into the classifier to determine whether the primary user signal exists to realize the spectrum sensing.