Research On OFDM Symbol Timing Synchronization Algorithm Based On Serial Correlation

With fast transmission data rates, high spectrum efficiency and strong anti-interference ability,Orthogonal Frequency Division Multiplexing (OFDM) technology has been widely used especiallyin the wireless LAN standard and fourth-generation mobile communication technology.However, OFDM technology is very sensitive to the synchronization error as it will cause thesystem to generate inter-symbol interference, destroy the orthogonality of the subcarriers and affectsystem performance. The effective implementation of the synchronization technology plays a keyrole in the OFDM system performance. So the synchronization algorithm research has become animportant field and mentor is also agree with this field has very good research value. This thesiscarries research about the OFDM symbol timing synchronization technology:Firstly, the symbol timing synchronization algorithms based on training sequence are analyzedin the thesis, such as Schmidl&Cox algorithm and the algorithms proposed by Minn and Park.However, both of them have vice peaks in timing synchronization. A new algorithm is proposedaccording to the characteristic of the Minn and Park algorithms. Simulation results show that thetiming position has found spikes and no vice peaks, the mean square error is reduced in AddictiveWhite Gauss Noise (AWGN) channel and the new algorithm has better timing performance andaccuracy. And simulation also founds that new algorithm has better performance in multi-pathchannel than traditional algorithms.Secondly, the symbol timing synchronization algorithms based on cyclic prefix are analyzed inthe thesis, such as the classical Maximum Likelihood (ML) algorithm. Simulation founds that MLalgorithm degrades greatly in multi-path channel by comparision. Finally, this thesis notices that theensemble correlation algorithm and difference correlation algorithm are used in multi-path channel.For the aboved combined algorithms, a novel algorithm which improves the difference energy incombined algorithms by delaying several sampling points is proposed. Simulation results based onIEEE802.11a standard channel parameters show that the mean square error of the novel algorithmis reduced, and the research is in line with the prospective aim.