Study On LTE Downlink Primary Synchronization Algorithm

As the standard of the next generation of communication systems,LTE has been accepted by the major operators and research institutions. In order to obtain higher communication speed,greater coverage and system capacity,LTE uses OFDM,MIMO and other key technologies to meet the requirements for mobile communication quality of services.LTE cell search is the first physical layer process when mobile devices access to the cell.In the cell search process,the first step is the coarse timing synchronization of PSS. Timing shift will affects the synchronization, and primary synchronization performance has a direct impact on subsequent communication process, thus the primary synchronization algorithm research has important significance.Firstly, the background of LTE and research status of the downlink timing synchronization algorithms are discussed in the thesis.Thestructure of radio frame and time-frequency resource in the LTE physical layer are introduced.The principle of the OFDM systems and synchronization techniques, the effect caused by the error of timing and frequency offset are also discussed in the paper. Then this paper analyzes the cell search process, as well as the characteristics of the downlink synchronization sequence.S&C algorithm,Minn algorithm which performance is improved than S&C and Yong algorithm are existing autocorrelation synchronization algorithms.These algorithms belong to second-order autocorrelation algorithms,which adopt training different from PSS.Therefore they cannot be applied to LTE downlink PSS synchronization.In addition,it is proven that second-order statistics are not sufficient statistics.therefore,the proposed method adapts to the stucture of PSS and is based on fourth-order statistics.A new timing metrics based on self-related is proposed which is immune to CFO and the method can achieve excellent anti-noise performance.Based on simulation in channel models of LTE and AWGN channel,this paper adopts the mean square error(MSE) to evaluate timing performance.The simulation results show that in AWGN channel and multipath channels, the CFO would not effect the timing performance of theproposed method.However the performance of traditional algorithms experience degradation in heavy CFO. The simulation also proves that these two proposed methods can both increase the adjustable lenths of correlation and use threshold-based detection to improve anti-noise performance under lower SNR.The existing cross-correlation algorithms including Kang algorithm, M-part algorithm,etc. can achieve good performance in small CFO.However the results of simulation show that timing performance will experience degradation in heavy CFO.The proposed method based on cross-correlation is immune to CFO and improves the anti-noise performance.The proposed method can increase the adjustable lenths of correlation and use threshold-based detection to improve anti-noise performance.The simulation shows that the proposed method based on cross-correlation improves anti-noise performance and anti-offset property.In comparison, the proposed method based on cross-correlation achieves a better timing performance than the proposed method based on self-related.