Ofdm Systems In Time Domain Synchronization Algorithm Research

With the increasing demands for better mobile communications, the research on the next-generation mobile communications is forging ahead at the premise of the good construction of the 3G system. At the period of transition from 3G to 4G,3GPP starts the Long-Time Evolution (LTE) plan. To provide higher transmission rate, higher frequency spectrum efficiency, larger system throughout and cheaper cost, LTE exploits many new techniques, such as multiple-input and multiple-output (MIMO), orthogonal frequency division multiplexing (OFDM), software defined radio (SDR), etc.OFDM is one of the multi-carrier modulation techniques. It divides the high-rate data stream into many low-rate orthogonal subcarriers, so better resistance against time delay spread and improved efficiency can be achieved by the OFDM systems. Moreover, the cyclic prefix in OFDM systems can avoid the inter-symbol interference (ISI) and inter-carrier interference (ICI) to a certain extent. However, the OFDM technique is sensitive to the synchronization error, including timing error and frequency offset error, which is also a hot topic in communication field.The causes of the symbol timing error and frequency offset error are firstly discussed, and the influences of the errors on the OFDM systems are analyzed in this thesis. Then the Maximum Likelihood (ML) algorithm based on cyclic prefix, the Schmidl & Cox (SC) algorithm based on training sequence, and the synchronization algorithm based on the preamble sequence are studied and simulated by MATLAB. An improved algorithm is used to against the uncertainty of symbol timing synchronization in ML and SC algorithms, and good performance is obtained. A joint ML estimation based on preamble sequence is also introduced. This estimation can obtain the frame timing and carrier frequency offset using the correlation of any pair of repetition patterns. Simulated by MATLAB, this method can provide optimized performance with higher accuracy. A near-ML estimation algorithm is used to reduce the implementation complexity involved in the joint ML estimation. Compared to the conventional estimators, the near-ML estimation algorithm also has a better performance.