Blind Estimation Algorithm Research For PN Sequence In DSSS Signals

Direct sequence spread spectrum (DSSS) signal has been widely used in militarycommunication, satellite communication and satellite navigation system. The uniqueproperties, such as low interception rate, anti-jamming, inhibiton of multipath effects, andlow SNR margin make it hot in electronic warfare field. The corresponding communicationscountermessure technology has been attented more and more. Because the blind estimationof PN sequence in the condition of knowing little amount of signal parameters is not yetmature, which restrict the performance of the entire system, so it is important and urgent toresearch blind estimation of PN sequence in lower SNR rate and more complex conditions.This thesis established DSSS system model firstly, and two types of non-cooperativePN sequences estimation have been studied in several usual blind estimation methods,including Eign Value Decomposition (EVD) approach and Projection approximationsubspace tracking (PAST) approach. What’s more, a new PCM-based blind estimationalgorithm is proposed based on EVD and PAST approaches. Partitioned Correlation Matrix(PCM) can estimate PN sequence in lower SNR condition and worse environment, whichdivided the data matrix into many sub-matirx, and calculate each partitioned matrixautocorrelation matrix to achieve noise reduction pretreatment. Finally, PN sequence can beestimated blindly from new combined matrix using PASTd technology. The computerexperimental simulation analysis the convergence, learning spead, computationalcomplexity and bit error rate (BER) of PCM, and evaluates the performance characteristicsand advantages of PCM algorithm.In order to solve the problem of long PN sequence and non-peridic signals, we proposea sub-section subspace tracking approach to blind estimate PN sequence. Because the PNsequence has its special features in non-peridic DSSS signals, the received signals will bedivided into partitioned matrices, synchronization by sliding window approach andestimation by subspace tracking approach can be made within each sub-section. Computerexperimental results are given to show that this algorithm need much less computationalcomplexity and storage space, what’s more, a good estimation can be obtained even thesignal is far below the noise level.