Simulation And Research On Channel Estimation And Synchronization Algrithms For OFDM System Based On Sequences

Recently, Orthogonal Frequency Division Multiplexing (OFDM) has gained increased interests of researchers in both academic research and project applications. This technology has become key technology of B3G because of its high special efficiency and robustness to selective fading. Channel estimation is a key technology of gaining channel information and lowing BER. Although OFDM has many advantages, it has drawbacks like sensitive to time and frequency offset, time and frequency synchronization is the precondition of OFDM correct processing. Channel estimation and synchronization is mainly based on algorithms of certain pilots, this paper discusses channel estimation and a synchronization technology based on training symbols and pilots, and compares the performance of each algorithm, and designs an improved method for Schmidl timing algorithm. The auto-correlation and correlation of zadoff-chu sequence are also discussed.After discuss basic OFDM principle and core algorithms in detail, this paper discusses the channel estimation algorithms firstly, which include Minimum Mean Square Estimate (MMSE), Maximum Likelihood Estimate (MLE), Least Square (LS) channel estimation algorithms based on pilots and Linear Interpolation, Time Interpolation algorithms, then simulates the system and compares the performance of each algorithm. As to synchronization, the paper analyzes influences of time and frequency offset OFDM system caused by frequency selective fading channel. After reading lots of paper, I chose three classic algorithms:Moose algorithm, Schmidl and ML algorithm based on CP, design simulation system for analyzing the BER performance and mean-square error, and design a new method to eliminate the peak platform phenomenon of Schmidl.From simulation results we can conclude, for channel estimation, in the aspect of implementation, MMSE and MLE are complex to achieve but the performance is better than LS; Time interpolation has a better performance than linear interpolation but implementation is complex. As to synchronization, MLE is the base algorithm, Schmidl algorithm gives the best performance of the three, in this paper, a new timing method based on Schmidl timing algorithm is designed, to eliminate the peak platform, the ML algorithm is based on CP, it is easily influenced by channel situation.