对光OFDM系统中的三种同步算法进行了理论分析, 基于仿真实验, 针对定时度量、定时偏差、丢帧率三个性能参数进行了对比分析。仿真分析表明T.Schmidl & D.Cox算法由于循环前缀(Cyclic Prefix, CP)缘故使得其定位精确性欠佳, 偏差较大, 无法进行精确定位, 但算法稳定性及丢帧率方面具有很好的性能。而H.Minn和Lixun Huang & Lijun Sun两种同步算法定位精确性相对较好, 偏差较小, 但在稳定性及丢帧率方面, 后者优于前者。因而在这三种同步算法中, Lixun Huang & Lijun Sun算法性能较好, 适应性较强, 更具实际应用价值。

The theoretical analysis and simulations on three kinds of synchronization algorithms for optical orthogonal frequency division multiplexing (OFDM) systems were performed. Based on simulation experiments, comparative analysis on the performance of the three algorithms for the timing metric, timing estimator and frame loss rate was carried out. The simulation results show that T.Schmidl & D.Cox algorithm can not obtain an accurate timing estimator and has a greater offset because of the existence of the cyclic prefix(CP), but it has a good performance on the stability and frame loss rate. H.Min algorithm and Lixun Huang & Lijun Sun algorithm do better on accuracy of timing estimator, in meantime, Lixun Huang & Lijun Sun algorithm is better than H.Min algorithm on performances of stability and frame loss rate. As a result, Lixun Huang & Lijun Sun algorithm among the three synchronization algorithms has better performance and adaptabilities, so it has better practical application value.