基于训练序列的改进OFDM同步算法

Improved synchronization algorithm for OFDM based on training sequence

针对正交频分复用系统对同步偏差的敏感性,设计了两种同时具备频偏估计能力的新训练序列,并给出了相应的定时尺度函数和频偏估计函数。利用训练序列的共轭对称性及部分数据的运算符号相反的特性,进行相关运算得到定时估计值和频偏估计值。理论分析和仿真结果表明,新训练序列的定时尺度非常尖锐,且不受循环前缀长度的影响;在加性高斯白噪声环境下,定时估计均值和均方误差均为零,实现了精确的定时同步;新频偏估计函数在保持了估计精度的同时扩大了频偏估计范围,且在信噪比较低时,新频偏估计算法仍具有很好的估计性能。

This paper designed two new training sequences to overcome the serious influence of synchronization offset on OFDM systems.In addition,this paper presented new timing metric functions and frequency offset estimation functions as well.The two training sequences have the property of conjugate symmetry,and the operation sign of some data is different.Thus,to get the value of timing estimation and frequency offset estimation,the new algorithm could conduct correlation calculation of training sequences.The theoretical analysis and simulation results show that the two new training sequences have an impulse-shaped timing metric for different cyclic prefix length and almost have no interference peaks.The mean and mean square error （MSE） of timing estimation are equal to zero in additive white Gaussian noise （AWGN） channel.The new frequency offset estimation algorithm could enlarge frequency offset estimation range while maintaining the estimation accuracy,and they still have superior performance for low signal-to-noise ratio （SNR）.