Synchronization Issues In Ofdm Transmission Systems

Orthogonal Frequency Division Multiplexing(OFDM) is a highly-efficient multi-carrier modulation technology. The great advantage of OFDM is that it improves frequency utilizing rate simultaneously and that it is robust against frequency selective fading, So it is promising to become the core technology of the next generation mobile communication. However, OFDM systems are more sensitive to synchronization errors than single carrier. So synchronization is the key problem for OFDM system. Lack of necessary synchronizer, the OFDM performance will be worse.This paper mainly research synchronization algorithms. This paper first introduces the basic theory and characteristic of OFDM system. There are three kinds of synchronal modes in OFDM system: symbol timing synchronization, carrier synchronization, sampling-bell synchronization, I discuss that why these synchronization error forms and how synchronization error affects system performance. I divide synchronization algorithms into two kinds—Data-Aided(DA) and NonData-Aided(NDA),then I mainly research Schmidl&Cox algorithm based Training Symbol and Maximum Likelihood(ML) algorithm based Cycle Prefix(CP). I improved them. To the shortage of high estimation error in multi-path channel and high implementation complexity of conventional ML algorithm, an improved symbol synchronization algorithm is proposed, which based on autocorrelation property of part OFDM signal.. The estimation error of improved algorithm is much lower than that of the conventional ML method in multi-path channel. Good performance and low complexity of the improved algorithm are shown in the analysis and the simulations. To Schmidl&Cox algorithm, I give an improved algorithm by adding calculation and smooth maximum. The algorithm can reduce fluctuate of symbol synchronization and improved performance in low SNR.