Noncooperative DSSS Signals Detection And Parameter Estimation

With the rapid development of wireless mobile communication,direct sequence spread spectrum communication(DSSS)has increased its demand in military and civil communication with its advantages of anti-narrowband interference and multi-path interference and high security.However,in non cooperative communication with low SNR,direct sequence spread spectrum communication system has the following problems.On the one hand,in non-cooperative communication,with low SNR,the detection effect of the widely used direct-expansion signal detection method becomes worse and cannot meet the detection performance requirements,resulting in the failure of the directexpansion system to carry out subsequent parameter estimation and demodulation.direct sequence spread spectrum signal detection algorithm with better performance has become an urgent problem for scholars.On the other hand,in the case of non-cooperative communication,the performance of the widely used multi-user signal pseudo-code estimation method is poor in the case of non-integral digital cell rate sampling.In order to ensure the demodulation performance of subsequent noncooperative DSSS signals,a more effective pseudo-code estimation algorithm becomes an urgent task.Therefore,this paper starts from these two aspects,and studies the DSSS signals detection algorithm and multi-user signal pseudocode estimation algorithm with better performance under low SNR in non-cooperative communication.In non-cooperative communication,the presence detection of DSSS signals becomes very difficult at low SNR.To solve this problem,this paper proposes a detection method based on the combination of signal eigenvalue histogram and residual network.First of all,according to the characteristics of direct expansion signals,the signal matrix model to be detected is constructed,and the eigenvalue of the covariance matrix of the signal to be detected is obtained by eigenvalue decomposition.The eigenvalue vector is superimposed according to the principal feature analysis method,and then the histogram of eigenvalue is drawn.To make it more distinguishable than the widely used FFT spectrum diagram and covariance matrix diagram;Then,a residual network model introduced by the channel attention mechanism was built and two kinds of data set images were made.The channel attention mechanism made the network model pay more attention to the local feature extraction of the image itself and improved the feature extraction ability of the residual network model.Theoretical analysis and experimental simulation show that the detection performance of the proposed algorithm is superior to the traditional FFT spectrum detection method,the convolutional neural network method based on FFT spectrum and the convolutional neural network method based on covariance matrix.In the case of non-cooperative communication,the pseudo-code estimation of multi-user DSSS signal is often inaccurate in the case of non-integer multisymbol rate sampling.To solve this problem,a pseudo-code estimation method based on delayed autocorrelation and PCA is proposed in this paper.Firstly,the carrier frequency of the multi-user signal is estimated by the frequency-domain square spectrum method.After down-conversion processing of the multi-user signal,the pseudo-code rate of the multi-user signal is estimated by the time-domain delay multipliation-autocorrelation method.Then,according to the pseudo-code rate estimate,the multi-user signal is converted to sample rate,and estimation of pn code period of multi-user signals using time-domain time delay autocorrelation Finally,according to the estimated pseudo-code period,the pseudo-code delay autocorrelation method was used to find the starting point of the pseudo-code information segment in the multi-user signal after sampling rate conversion,and the pseudo-code information segment was extracted to construct the blind source separation model.The modeled pseudo-code information matrix was separated by the principal feature independent component analysis method,and the pseudo-code of multiple users was obtained.Theoretical analysis and experimental simulation show that the performance of the proposed method is better than that of the sliding window method and ICA based multi-user pseudo-code estimation method.