Research On Ofdm Signal Modulation Recognition Based On Hierarchical Iterative Support Vector Machine

Orthogonal Frequency Division Multiplexing(OFDM)is a special multi-carrier modulated signal.It has become the first choice for wireless broadband communication systems because of its low cost,high spectrum utilization and strong anti-multipath fading capability.Globally,it is widely used in many fields such as civil communication and military communications,such as digital video broadcasting(DVB)systems,fourth-generation ultra-wideband mobile communication systems,wireless local area networks(WIFI),and various tactical/strategic communication systems.However,in the process of spatial transmission,due to the interference of receiving channel noise and other signals in space,the influence of modulation recognition on OFDM is relatively large.Therefore,a layered iterative structure using support vector machine is proposed to solve OFDM.Modulation identification of signals in complex electromagnetic environments.The main contents of this article include:1.The basic theory and key techniques of OFDM system modulation and recognition are introduced.The single carrier signals including MPSK signal,MFSK signal,MQAM signal and multi-load such as Wavelet Packet Modulation(WPM)signal are analyzed.Modulation characteristics.2.The basic classification principle of Support Vector Machines(SVM)is introduced.The characteristic parameters of OFDM signals and other signals are analyzed,including the distinction between high-order cumulants of multi-carrier and single-carrier,and the distinction between multi-carriers.The correspondence between the two spectra.3.In the background of Rayleigh channel,combined with the classification tlheory of SVM and the difference of eigenvalues of various modulated signals,a hierarchical iterative SVM classifier OFDM signal modulation recognition structure is proposed,and the high-order cumulant and double of each modulated signal arc analyzed.Spectral characteristics,which are used as training sample parameters for the classifier.A three-layer support vector machine classifier based on Radial Basis Kernel Function(RBF)is designed to separate OFDM signals from single-carrier and wavelet-packet signals,and then hierarchical iterative method is used to train classifiers.Parameters,first training each layer,and then training the trained three-layer classifier to optimize the parameters of the classifier,and finally recognize the OFDM modulation signal.The research results show that the three-layer classifier iterative structure can effectively identify OFDM signals,which can be classified not only with single carrier but also with multi-carrier wavelet packet signals.