Research Of OFDM Time And Frequency Synchronizations Technologies Based On Training Symbol

With the increasing requirements of the data transmission rate in wirelesscommunication business, the effect of multipath becomes the main factor that limitsthe improvement of the data transmission rate. Fortunately, the OFDM (OrthogonalFrequency Division Multiplexing) technology has the characteristics that caneffectively resist the influence of multipath, and has high spectrum efficiency. So, thereare more and more researches and applications in the field of the OFDM. Moreover,the OFDM technology has become one of the key techniques for data transmission.The synchro error, however, leads to substantial decrease of OFDM systemperformance even if the value of error is small. Thus, the researches of synchronizationare extremely valuable to the OFDM technology.This paper mainly studies the algorithm of time and frequency synchronizations(TFS), and the main research contents are shown as follow:Firstly, some fundamental knowledge of the OFDM technology is given a detaileddescription, such as the theory, the merit and demerit and the significant technologies.And the influence of the wireless transmission channel on the characteristic of signal isgiven. Also, it is given that the analysis and simulation results of the influence onOFDM system caused by synchronization error. All above are the bases for the TFSalgorithm.Secondly, the merit, the demerit and the application condition of several classicTFS methods are given, through the detailed formula derivation and simulation. Someconclusions are as follows: The timing metric of Schmidl and Cox’s algorithm has aplateau. The timing metric of Park’s algorithm has the sidelobe interference. Under themultipath channel, the frequency offset estimation algorithm accuracy can be affectedby the structure of training symbols with PN sequence weighted, even can’t beobtained.Thirdly, in order to obtain the improvement, the technique based on a newtraining symbols and the synchronization algorithm is introduced. The new trainingsymbol consists of four CAZAC sequences which are partly weighted by PN sequence. The simulation results prove that the sharp peak can be achieved by using the goodself-correlation quality of CAZAC sequence and special structure of the trainingsymbol, not the plateau and the sidelobe achieved in several classic TFS methods.What’s more, training symbol structure will not have an impact on the frequency offsetestimation accuracy. Under the multipath channel or the AWGN, the accurate timingsynchronization and the accurate frequency offset estimation within the entirebandwidth of the OFDM symbol are obtained by using the new technique.Finally, concerning the condition that the first path is not the strongest path, a newmethod to improve the system performance of the timing synchronization algorithm isproposed. The new method can capture the starting sample point location of OFDMsymbol. Also, the relationship between the peak value of timing metrics and frequencyoffset is considered. According to the results, it is obtained that the new trainingsequence is not affected by the frequency offset. What’s more, the precise frequencyoffset estimation within the entire bandwidth of the OFDM symbol and the timingsynchronization can be achieved by using this method. At last a complete time andfrequency synchronization process is given.