Identification And Parameter Estimation For Non-cooperative OFDM Signals In Multipath Channel

In recent years,with the continuous progress of information society,wireless spectrum resource is getting more and more intense and the wireless electromagnetic environment is getting more and more complex.Thus,Orthogonal Frequency Division Multiplexing(OFDM)technology is widely used in modern communication system due to its efficient spectrum utilization and robustness against multipath interference.Many of the application scenarios are carried out under the premise of non-cooperation,such as electromagnetic spectrum monitoring,radio interception in the field of civil communications and communication reconnaissance,communication interference in the field of military communications.And in turn,it promotes the development of non-cooperative OFDM technology.As an intermediate step for signal detection and signal demodulation,noncooperative OFDM signal recognition and parameter estimation is of great significance to the research of non-cooperative OFDM signal technology.In turn,it also promotes economic development.OFDM is usually applied to multipath scenarios due to its superior ability to resist multipath delay.Combining the universality of multipath transmission in the wireless electromagnetic environment,in this thesis,non-cooperative OFDM signal receiving is selected as application background.The aim is to find a simple and efficient method of OFDM signal recognition and parameter estimation with strong channel adaptability.Starting with the characteristics of OFDM signal,in order to recognize {2FSK,4FSK,QPSK,8PSK,16 QAM,64QAM,OFDM} as a signal set,this thesis focuses on the recognition and parameter estimation of OFDM signal in multipath scene.The specific research work is listed as follows:In the first part,the characteristic of OFDM signal are analyzed.Starting from the characteristics of OFDM signal,the time domain statistical characteristics,autocorrelation characteristics of symbols and cyclostationary characteristics of OFDM signal under noise-free environment and additive white gaussion noise(AWGN)channel are discussed respectively.On this basis,the characteristics of OFDM signal in multipath channel are analyzed,which lay a theoretical foundation for subsequent identification and parameter estimation.Secondly,based on the statistical characteristics of OFDM signal in time domain,a recognition method based on high order cumulants is given,and the applicability of the method under multipath channel is analyzed.However,considering the high complexity of structural features and the poor recognition results under low signal to noise ratio,the recognition method based on high order cumulants under multipath channel is no longer applicable.The wavelet transform results of single carrier signal and OFDM signal are deduced and analyzed based on its excellent time-frequency localization.The OFDM signal recognition scheme based on wavelet transform is determined,which has good recognition performance under low signal to noise ratio.At last,the effective symbol length and symbol period of the signal are estimated based on the autocorrelation of OFDM signal.Through mathematical deduction,it is found that the correlation peak under multipath channel is influenced by the Doppler frequency shift of the channel.Therefore,in this thesis,a variable start point autocorrelation method is adopted.This method can improve the estimation accuracy of parameters by increasing the length of observation data,and does not depend on symbol timing synchronization.After that,based on cyclostationary characteristics of OFDM signals,two parameter estimation methods based on cyclic autocorrelation function and cyclic spectrum respectively are presented,compared and analyzed.After the above parameter estimation is completed,the subcarrier identification is modeled,and the OFDM signal subcarrier modulation recognition is completed based on the step-by-step classification by combination of high order cumulants.