论文部分内容阅读
Orthogonal Frequency Division Multiplexing (OFDM) system is sensitive to Carrier Frequency Offset (CFO). In this paper, traditional Maximum Likelihood Estimators (MLE) for CFO of OFDM are introduced first. Then, averaging method and α-filter are introduced as Low Pass Filter (LPF) to improve the performance of cyclic prefix estimator. The bandwith of LPF is determined by the coherence time of radio channel. Estimation performance in multipath channel is analyzed. Outlier picking-out scheme is proposed to improve performance further. Performance of close-loop structure is presented briefly, which is worse than that of open-loop structure. Finally, a parallel switch structure of frequency synchronizer is proposed for mobile OFDM systems. The scheme exploits training sequence and cyclic prefix. The proposed synchronizer has a wide acquisition range. It is accurate and robust in both AWGN channel and multipath channels. The complexity is low due to functionality of α-filter. A better performance of frequency synchronization is obtained comparing to that of existing Maximum Likelihood Estimator(MLEs). We achieve these advantages without loss of bandwidth efficiency.
Orthogonal Frequency Division Multiplexing (OFDM) system is sensitive to Carrier Frequency Offset (CFO). In this paper, the traditional Maximum Likelihood Estimators (MLE) for CFO of OFDM are introduced first. Then, averaging method and α-filter are introduced as Low Pass Filter (LPF) to improve the performance of cyclic prefix estimator. The bandwith of LPF is determined by the coherence time of radio channel. Estimation performance in multipath channel is analyzed. Outlier picking-out scheme is proposed to improve performance further. The proposed synchronizer has a wide acquisition. The proposed synchronizer has a wide acquisition It is accurate and robust in both AWGN channel and multipath channels. The complexity is low due to functionality of α-filter. A better per formance of frequency synchronization is obtained comparing to that of existing Maximum Likelihood Estimator (MLEs). We achieve these advantages without loss of bandwidth efficiency.