Research On Parameters Estimation For Direct Sequence Spread Spectrum Signals In Non-cooperative Communication

The direct sequence spread spectrum(DSSS)signals have been widely used in civil communications,military tactical communication and so on,owning to the good confidentiality,flexible channel allocation capability,strong ability of anti-multipath interference and multi-access interference,etc.In the non-cooperative communication,the research on parameters estimation of DSSS signals has great significance,among which the blind estimation of Pseudo Noise(PN)sequence is the difficult and hot issue.For the issue of parameters estimation of DSSS signals in additive white Gaussian noise(AWGN)channel,this dissertation has conducted some research and exploration.The main research work includes the following:(1)For the problem of parameters estimation of DSSS signals,an algorithm based on eigenvalue decomposition(EVD)of the covariance matrix is proposed.Firstly,the period of PN sequence is obtained by reprocessing of the power spectrum density.Then,the received signal is sampled and divided into temporal windows whose duration is twice the period of PN sequence,from which the covariance matrix is computed and decomposed.The eigenvector corresponding to the maximum eigenvalue contains the information of parameters of DSSS signals.According to the parameter characteristics,the line spectrums representing carrier frequency and chip rate respectively are obtained by doing nonlinear transform to the eigenvector.After the modulating synchronization point is available by applying the vector 2-norm criterion or matrix F-norm criterion to the eigenvector,the output sequence of chip timing algorithm can be used to restore the PN sequence.The suitable estimation algorithm for each parameter is selected according to its performance.(2)Based on singular value decomposition(SVD),an approach for joint estimation of PN sequence and information sequence of DSSS signals is proposed.A signal matrix is built by dividing the received signal into segments according to twice the PN sequence period.The left singular vectors and right singular vectors of signal subspace contain the information of PN sequence and information sequence,respectively.Thus,the PN sequence can be restored by analyzing the left singular vectors.Meanwhile,the information of right singular vectors is exploited to estimate the information sequence.Experimental results show that comparing with other algorithms,this algorithm can achieve the best performance in PN sequence estimation and performance approaching the best in information sequence estimation.(3)In non-cooperative communication systems such as spectrum surveillance and electronic interception,the received signal may be intentionally or unintentionally affected by narrowband interference(NBI).Focus on the problem of parameters and PN sequence estimation of DSSS signals under NBIs,an efficient algorithm,which can estimate the PN sequence and analyze the characteristics of NBIs at the same time,is proposed.The covariance matrix of signal segments whose duration is twice the period of PN sequence is computed and decomposed.The eigenvectors containing the information of NBIs and PN sequence are separated in signal subspace.According to the size distribution of eigen-value,the equivalent source number can be determined.Then,the eigenvector containing the whole information of PN sequence can be located with some custom functions.The parameters and PN sequence can be estimated by analyzing this eigenvector.Comparing with other algorithms,this algorithm performs best in carrier frequency and PN sequence estimations.(4)To solve the problem of PN sequences estimation for M-ary direct sequence spread spectrum(MSSS)signal,a method of estimating PN sequences in MSSS signals on K-means clustering algorithm is proposed by using the idea of unsupervised clustering analysis provides for reference.The received signal is sampled and divided into nonoverlapping segments according to PN sequence period.A detection statistic built from the correlation property of these signal segments is exploited to estimate the modulating synchronization point and thus the signal segments can be synchronized.The PN sequence set's scale can be estimated by maximizing the similarity difference function,which can measure the clustering result of asynchronous signal segments.With the estimated modulating synchronization point and PN sequence set's scale,the PN sequences can be restored from the clustering result of these signal segments.This simulation experiment results reveal that this algorithm has good PN sequences estimation in varied PN sequence set's scales and PN sequences.