Design Of Blind Equalizer Based On

Since the1960s，with the high-speed development of computer technology andthe techniques of fast Fourier transform, digital signal processing has been widelyused in communications, radar, sonar and other fields. Equalization is actually a partof demodulating the signal in the digital signal processing, and it means to reduce oreliminate the inter-symbol interference in the signal processing or filteirng techniques,which generally requires training signal to help. The blind equalizer is proposed in1975b’y Japanese scholars Y.Sato. It neednt training signal, and adjust the receiverthrough statistical characteirstics of the transmitted signal to adapt transmissionenvironment, in order to avoid the receiver situation of a temporary failure of thebranch channel to be re-balanced branch, otherwise it will lack training signals or askthe central station to carry training signal all the time. Then the transmissioneiffciency will be affected. Therefore, the blind equalizer research is of very importanteconomic and scientific value for the actual rapid development in modern digital life.This paper is mainly focused on the blind equalizer of Maximum LikelihoodSequence Estimation (MLSE) and Pre-Survivor Processing algorithm (PSP),especially for the emergent signal tracking of PSP algorithm in the fast varyingchannel. Compared with traditional Maximum Likelihood Sequence Estimation(MLSE) blind equalization, the improved Maximum Likelihood Sequence Estimation(MLSE) blind equalization of PSP algorithm can combine channel estimation with thesequence detection at the same time，which avoids the delay in quick judgmentsvarying channel signal, and improve the eiffciency of blind equalizer.The main researches as follow:1.Firstly, the paper introduces several common blind equalizer and researchstatus at home and abroad. And then it gives a detailed analysis of the MLSE blindequalizer, including basic working pirnciple, key technology and application methods.2.Secondly, the channel estimation using PSP in the signal models of BPSK,QPSK,8PSK and QAM with noise are shown. According to the output of the mean square error map, the analysis and discussion is expanded. The main purpose is todetermine whether the PSP can run atfer each signal fast.3.The channel estimation of PSP is applied to MLSE blind equalizer. Thediscussion is shown through the observation of degree of convergence in each ievel ofthe signal constellation.4.Finally, The fifth chapter studies the effect factors of results in blindequalization based on PSP. The main method is to observe and compare the meansquare error map and constellation diagram with different step size and channel lengthin each signal model.