Blind Joint Estimation for OFDM Time-Frequency Parameters

摘要

A novel blind joint estimation of orthogonal frequency division multiplexing (OFDM) time-frequency parameters is conceived in low signal-to-noise ratio (SNR) and multipath channel conditions. The oversampling factor is estimated by employing a new-designed cosine roll-off filter whose cut-off frequency is considered as the approximation of the sampling signal's spectrum. The useful data length is estimated by means of the improved varying time-delay autocorrelation, and its final estimated value is defined as a power of 2, which has the minimum Euclidean distance with the peak position. The estimated useful data length is further utilized to refine the previous estimation of the oversampling factor. The bandwidth of OFDM signals is estimated by adopting the spectrum approximation method based on the estimated time-parameters. The simulation results demonstrate that the proposed method is robust to noise and multipath channel conditions without any requirement of a prior knowledge.