Research On Low Processing Pelay Peceiver Pechnology In Burst Communication System

Since its excellent performance in anti-interference and anti-interception,burst communication has been widely and well applied in deep space communication,satellite communication and military communication;At the same time,spread spectrum communication technology has attractive performance in communication security.Carrier synchronization and channel estimation in burst communication are always the focus and hot topics in this field.In addition,the ideal of development of modern communication not only ensure the high reliability of communication,but also achieve the low time delay of communication system.Therefore,in allusion to the research focus of burst spread spectrum communication and the low delay requirement of modern communication,the feasibility of extrapolation method is studied in the field of carrier synchronization technology and channel estimation technology,and the processing delay of burst communication system can be intuitively reduced by the way of preprocessing.In this thesis,the extrapolation method is introduced in carrier synchronization technology.This thesis focuses on high dynamic scene deviation estimation technology,and find the difficulty of applying the extrapolation method to frequency offset estimation technique is the contradiction between estimate range and estimation precision of the algorithm,which greatly limits the historical sample size,especially in the high dynamic large frequency offset.For time domain frequency offset estimation algorithm,in practical systems,the estimation accuracy and estimation range are generally compromised according to the requirements.That is to say,when the estimation range is expanded,the estimation accuracy of the algorithm has to be lost.However,the estimation accuracy of the algorithm must be improved at the cost of the estimation range.Therefore,based on the traditional time domain frequency offset estimation algorithm——Fitz algorithm,this paper improves the estimation range of the original Fitz algorithm by compensating the phase reversal of the autocorrelation function,and gains a larger estimation range with additional computational complexity.Fitz algorithm based on compensation of phase reversal can estimate the frequency offset within the ±50% range of symbol rate when it meets the SNR threshold requirements,and the estimation error is close to CRLB.Compared with the original Fitz algorithm,it has both high estimation accuracy and wide estimation range,which effectively solves the problem of mutual limitation between estimation range and estimation accuracy.Then improve the feasibility of extrapolation method in frequency offset estimation algorithm.Then this thesis continues to discuss the application of extrapolation method in the field of channel estimation.Different from the problems caused by the extrapolation of frequency offset estimation algorithm,the time variability of channel itself and the randomness of fading signal are the difficulties and focuses of the research.In this thesis,from the perspective of autoregressive（AR）channel prediction model,the shortcomings of the traditional direct one-step prediction method and channel prediction algorithm based on RLS are analyzed,and the NLMS adaptive filtering algorithm is proposed to channel prediction.According to the historical channel information,the new algorithm can predict the channel state by iteration,and the predicted value can be directly used in the next burst frame,which can directly save the processing time of channel estimation module and reduce the system processing delay intuitively.The channel prediction algorithm based on NLMS has better prediction performance than direct one-step prediction,and its computational complexity is obviously lower than that of RLS based channel estimation algorithm,which meets the requirement of low processing delay in burst spread spectrum system.