Identification Of Source And Information Processing Strategy Of Wireless Radio Frequency Signal

With the rapid development of wireless communication technology,the growing information processing capability of wireless device,and the increasing demand of wireless service,wireless spectrum is shared by a large number of different types of wireless devices and wireless services.Wireless spectrum resource has become more and more scarce,and its efficient use has high economic and social benefits.Reasonable allocation of radio spectrum resource is an important influencing factor to realize wireless communication technology convergence,wireless services coverage,and wireless devices cooperation.Therefore,scientific supervision and management of wireless radio frequency signals in the environment is very important.Carry out real-time monitoring of the wireless radio frequency signal,determine its type,and then take corresponding action to improve the efficiency of wireless communication technology.This dissertation achieves signal source and information processing method identification in the radio frequency signals coexistence environment based on data mining algorithms.The main research work and innovative points are as follows:(1)For IEEE 802.1 In radio frequency signals from different sources,the modulation types are both BPSK.Use Sora software radio platform and spectrum analyzer to capture power spectral density of the signal and study the identification method.Analyze the actual signal power spectral density data,and then use the improved k-nearest neighbor algorithm to accomplish signal source recognition combined with the data characteristic.The improvements include principal component analysis dimension reduction pre-processing and cross-validation option to select k value.(2)For IEEE 802.11a and IEEE 802.11n radio frequency signals from the same source,analyze the actual power spectral density data.There exists large difference between the two kinds of instantaneous power spectral densities,so linear classification method in data mining can make accurate identification.Select the linear discriminant analysis method,as it has no limit to the choice of variance and distribution.Fisher criterion can achieve categorization on the basis of feature dimensionality reduction.(3)For IEEE 802.11n BPSK and QPSK of two transmission rates radio frequency signals from the same source.Conduct any two types of wireless radio frequency signal classification respectively.First of all,use linear classification method of logistic regression model to make analysis,and the results show that IEEE 802.11n radio frequency signal data are nonlinear separable.Therefore,use non-linear support vector machine to identify the signal type,selecting model and parameters by cross-validation.