A Study On DSSS Signals Blind Separation Technique

The Direct Sequence Spread Spectrum(DSSS)system is widely used in military,civillian and satellite communications,because of its strong concealment and strong antiinterference.In particular,the DS-CDMA(Direct Sequence-Code Division Multiple Access)signal inherits the advantages of the DSSS signal and also has the advantage of code division multiple access,which making it widely used in modern communication.The important parameters of the DSSS signal are unknown for the non-cooperative receivers,such as the spreading sequence.The information sequences cannot be directly obtained from the received signal before the parameters are estimated.Therefore,it`s meaningful to study the blind dispreading of the DSSS signal.This thesis is mainly researching in the blind despreading of the DSSS singal.The main contents of this theis are as follows:1.Detecting the DSSS signals needs a large amount of prior information.If there has the Doppler shift,the detecting of the DSSS signals will be more difficult.To solve the problems,this thesis designs a DSSS signal detection system.Even under low SNR conditions,the system can still detect the DSSS signal with the doppler shift and estmate the parameters.The priori information of the signals required by the detection system is less than before and can work for both the long-code DSSS signal and the short-code DSSS signal.2.based on the blind despreading of the single-user short-code DSSS signals,we start to study the blind despreading algorithm of the short-code DS-CDMA signals under the single-channel reception.According the model of the short-code DS-CDMA signal,we propose a blind despreading algorithm based on ILSP(Iterative Least Squares with Projection)algorithm,which can solve the blind despreading problems of the synchronous short-code DS-CDMA signal,asynchronous short-code DS-CDMA signal and the synchronous short-code DS-CDMA signal with multi-path.This thesis also deduces the CRB(Cramer-Rao Bound)of the short-code DS-CDMA signals in the three cases mentioned above.The simulation results of the algorithm are compared with the CRB and the simulation results of the traditional blind despreading algorithm based on EM(Expectation Maximization)algorithm,and the comparation proves the proposed algorithms is great.3.This thesis also studies the blind despreading of the synchronous long-code DS-CDMA signal under single-channel reception.In this thesis,we propose three ways which are a blind despreading algorithm based on the EM-ILSP algorithm,the blind despreading algorithm based on subspace segmentation estimation and the blind despreading algorithm based on the ILSP algorithm to solve the problem of the blind despreading of the long-code DS-CDMA signal.4.In this part,the blind despreading of the DS-CDMA signal under single-channel reception is extended to the blind despreading of the DS-CDMA signal under multichannel reception.The thesis presents a third-oder tensor model of Parallel Factor(PARAFAC)to sovle the problems of the blind depreading under this situation.The synchronous short-code DS-CDMA signal,the asynchronous short-code DS-CDMA signal,the synchronous short-code DS-CDMA signal under multipath propagation and synchronous long-code DS-CDMA signal received by multi-channel can be modeled as a three-order tensor models.According to the characteristics of different signal models,the models can be low-rank decomposed or rank F-F-1 decomposed by the ALS(Alternating Least Squares)algorithm or the interpolating ALS algorithm.