Research On Modulation Classification Algorithm Of Integrated Radar Communication OFDM Signals

With the development of synaesthesia integration technology,radar communication integration system based on OFDM signal has been widely used in the field of vehicle networking.As the number of vehicles equipped with such systems increases,the problem of interference between vehicles becomes more and more prominent.In order to solve the problem of interference between systems,spectrum sensing is needed,and the difficulty lies in the accurate modulation classification.In particular,it is necessary to consider the influence of channel fading in practical engineering applications,such as the common multipath channel in urban environment.To solve the above problems,this paper proposes a modulation classification algorithm for radar communication integration OFDM signals based on cyclic stationarity,which realizes the estimation of demodulation parameters of OFDM signals,rough sorting of radar communication signals,modulation classification of OFDM radar signals and subcarrier modulation classification of OFDM communication signals.Specific work is as follows:1.Firstly,the radar communication integration system and signal model are introduced.The estimation algorithm of OFDM signal parameters based on cyclic spectrum and cyclic autocorrelation characteristics is studied,including carrier frequency,chip rate,effective symbol length and total symbol length of OFDM signal.The simulation results show that the accuracy of each parameter estimation is about 99%when the signal-to-noise ratio is greater than 2 dB in the white Gaussian noise and multipath channel environment.2.A modulation classification algorithm of OFDM radar signal based on time-frequency characteristics and cyclic spectrum is proposed.The algorithm makes grayscale processing and normalization of the time-frequency and cyclic spectra of the signal to form two layers of image features.Then,taking the image features as input parameters,a convolutional neural network is designed to realize modulation classification of OFDM radar signals.The simulation results show that the algorithm has a recognition accuracy of more than 90%for PSK-OFDM,LFM-OFDM and MLFM-OFDM radar signals in multi-path channel when the signal-to-noise ratio is greater than-4 dB.3.A modulation classification algorithm for OFDM communication single carrier signals is proposed based on second-order to eighth-order cyclic cumulants of grid cyclic frequency.The algorithm preprocesses the OFDM communication signal and obtains the singleton carrier signal.Then,high-order cycle cumulants are selected to form signal features according to two dimensions of grid cycle frequency and order,and machine learning classifier is trained to realize modulation classification.The simulation results show that this feature has better anti-noise performance than the traditional high-order cumulant.When the signal-to-noise ratio is greater than 6 dB,the overall recognition accuracy of the 6-seed carrier modulation mode is more than 80%.4.Finally,based on the above algorithm,the processing flow of radar communication integrated signal data set is designed.Firstly,the radar communication signal is selected by the duty ratio,and then the radar/communication signal is preprocessed by using the information such as parameter estimation.Then the modulation classification is carried out on the OFDM monadic carrier signal and the pulse part of the radar signal respectively.USRP equipment is used to build a semi-physical simulation platform to verify the performance of the proposed modulation classification algorithm for radar communication integration system.