Study On The Method Of Intelligent Detection And Parameter Estimation Of Polyphase Code And Combined Modulation Radar Signals

In the current complex battlefield environment,low-intercept probability polyphase code and combined modulation radar signals have been widely used for their advantages such as large time-bandwidth product,high resolution,and anti-jamming.Traditional radar reconnaissance methods are difficult to effectively detect and estimate the parameters of polyphase code and combined modulation radar signals.In order to solve the problem of detection and parameter estimation of polyphase code and combined modulation radar signals,this thesis combines artificial intelligence to carry out research on intelligent detection and parameter estimation methods of polyphase code and combined modulation radar signals.This thesis analyzes the time-frequency characteristics and waveform parameters of polyphase code and combined modulation radar signals.On this basis,the intelligent detection method based on the feature extraction of the auto-encoder,the intra-pulse modulation recognition method based on the convolution neural network radar signal and the radar signal parameter estimation method based on the BP neural network are proposed.The main contents are as follows:1.An intelligent detection method based on auto-encoder feature extraction is proposed,which improves the detection probability of polyphase code and combined modulation radar signals in a low SNR environment.The method uses an auto-encoder to extract the deep characteristics of the signal spectrum,and then uses a classifier to complete the intelligent detection of the signal.The determination of the dimension of signal spectrum feature extraction and the influence of different classifiers on signal detection probability are analyzed,it is finally determined that the frequency spectrum of the signal is extracted to two dimensions by the encoder and classified by the decision tree,so as to effectively detect the polyphase code and the combined modulation radar signal in the low SNR environment.2.The method of intra-pulse modulation recognition based on convolutional neural network radar signal is studied.Firstly,the time-frequency image of polyphase code and combined modulation radar signal is extracted,then carries on the down-samples and the Otsu method preprocessing to the image.The time-frequency image of the signal is transformed into a binary image.Finally,the intra-pulse modulation recognition of radar signal is realized by using convolutional neural network.This method can also recognize 9 kinds of radar signals with low SNR.3.The method of parameter estimation of polyphase code radar signal based on BP neural network is studied.Firstly,the best order of FrFT transform of polyphase code signal is estimated by IQPF method,and then the FrFT domain spectrum under the best order is calculated.Finally,the intelligent estimation of time-frequency ridge interval parameter is completed by BP neural network.At the same time,the method of parameter estimation of FSK BPSK signal based on time-frequency image is studied.This method uses Otsu method and morphological close operation to fill in the discontinuous pixels of time-frequency image in a FSK symbol caused by BPSK modulation,and then the timefrequency image is projected horizontally and vertically to realize the estimation of FSK modulation parameters.The above two methods can also accurately estimate the signal parameters in the case of low SNR.The effectiveness of the above methods have been verified by simulation experiments.The results show that the intelligent detection and parameter estimation of polyphase code and combined modulation radar signals are realized.