Blind Estimation Of Multiple Pseudo-random Codes Of Direct Spread Spectrum Signals

In recent years,the pseudo-random(PN)code blind estimation of direct sequence spread spectrum(DSSS)signals in non-cooperative communication has drawn much attention.The pseudocode estimation is difficult in non-cooperative communication,because direct sequence code division multiple access(DS-CDMA)signals contain spread spectrum signals of multiple users and the structure is complex.In this paper,we mainly study the pseudo-code blind estimation problem of Periodic Long Code Direct Sequence Code Division Multiple Access(PLC-DS-CDMA)and the nonperiodic DS-CDMA signals.The non-periodic CDMA signal includes the non-periodic long code(NPLC-DS-CDMA)signal and the non-periodic long code DS-CDMA(NPLSC-DS-CDMA)signal.First of all,the basic working principle of DSSS communication system and DS-CDMA signal classification,characteristics and practical applications are introduced.The current situation of research on different kinds of DSSS estimation methods in non-cooperative communication is briefly introduced.Secondly,a blind estimation method of the PLC-DS-CDMA signals based on BCH error correction and Fast-ICA is proposed.The third order correlation function(TCF)of m sequence,the equivalence principle of BCH code and m sequence and the basic principle of independent component analysis are introduced.Based on the above theory,the algorithm for long codes blind estimation of PLC-DS-CDMA using Fast-ICA method,third-order correlation theory and BCH error correction is proposed.The performance of long code estimation is superior to the existing algorithms.Then,in order to reduce the complexity of long codes estimation of PLC-DS-CDMA signal and ensure the performance of estimation,a blind estimate of long codes of multi-antenna PLC-DSCDMA signal based on tensor decomposition is proposed.The multi-antenna PLC-DS-CDMA signal is constructed into tensor and blocked by using the spatial information of multi-antenna receiving signal.The TUCKER factorization factor matrix is calculated by variable step gradient descent algorithm,and the permutation matrix is estimated by using receiving gain matrix.Then the long codes of PLC-DS-CDMA signal are estimated.Finally,the pseudo-code blind estimation of aperiodic DS-CDMA signal is studied.In the case of unknown types of non-periodic DS-CDMA signals,Fast-ICA algorithm,twice delay multiplication method and third-order correlation function are used to calculate the long-code primitive polynomial.After eliminating the long codes,the signal type is judged by comparing the variance of the sequence and the eigen decomposition method is further used to estimate the short codes.Simulation results show that this algorithm can effectively estimate the long codes and the codes and short codes of NPLC-DS-CDMA signal and NPLSC-DS-CDMA signal.