Research On Blind Identification Technology Of Key Parameters Of Communication Signals

In the non-cooperative communication system,the unauthorized third party attempts to gain access to the cooperative communication system to obtain communication information.With the rapid development of communication technology,the wireless communication exhibits a trend of higher complexity and more diversification,which causes the increasing complexity of analyzing signals at the receiving end of non-cooperative communication.In this context,the blind estimation of the signal symbol rate and modulation mode is the key to demodulate the signal at the receiving end of non-cooperative communication.However,there are few works on symbol rate estimation for the signals with some special modulation modes,and most researches related to modulation identification only consider the case without frequency offset.In view of these problems,this thesis investigates the blind identification technologies of some key parameters of communication signals for a variety of modulated signals with frequency offset.The main work and contributions are as follows:Firstly,to separate the useful signal from the wideband receiver,this work designs the detection and separation scheme of the received signal.The power spectrum is preprocessed based on Welch method and segmental averaging strategy,and the signal detection is completed based on the fluctuation of the power spectrum amplitude.In addition,the signal bandwidth and carrier frequency are both estimated.Then the baseband signal is filtered out based on these estimation results,and the starting point of the signal in the time domain is obtained by using the double sliding window method to extract the useful signal.Simulation results show that the proposed scheme can effectively detect and separate the target signal in systems corrupted by additive Gaussian white noise,and the estimations of bandwidth and carrier frequency are accurate.Secondly,a symbol rate estimation scheme which is insensitive to frequency offset is proposed in this work.To ensure the universality of the scheme,this work coarsely classifies 14 different modulation modes of digital signals into four categories and designs corresponding symbol rate estimation schemes.Especially for OQPSK modulated signals,the delay between two IQ channels invalidates most current rate estimation algorithms.This work proposes a rate signal scheme based on the multiplication of two IQ channels.The simulation results show that this scheme can accurately estimate the symbol rate of OQPSK signals at a signal-to-noise ratio over 4dB.Besides,for signal-to-noise ratio over 6dB,the proposed scheme can accurately estimate the symbol rates of 14 modulated signals such as MPSK,QAM,MAPSK,MFSK,GMSK,and OQPSK.Finally,this work designed a modulation mode classifier based on decision tree to classify 14 modulation modes.Due to the serious impact of frequency offset on the higher order cumulants and constellation characteristics of modulated signals,by analyzing the probability density functions(PDF)of signal amplitude and differential phase,this thesis proposes signal features with frequency offset resistance based on the number of peaks and the ratio of the value of PDF at different peaks.Subsequently,the extracted features together with instantaneous features and high-power spectral line features are utilized by the decision tree to achieve accurate modulation recognition.The simulation results show that the modulation recognition scheme proposed in this thesis can achieve a recognition accuracy of 99%when the signal-to-noise ratio is no less than 8dBIn summary,this thesis proposes practical symbol rate estimation scheme and modulation recognition schemes for multiple modulated signals with frequency offset.The proposed schemes can effectively identify signal parameters and have significant practical application value.