Data-aided CFO estimators based on the averaged cyclic autocorrelation

Wireless communication systems typically employ a repetitive preamble in each slot which is used for parameter acquisition. The repetitive preamble is useful for estimating the carrier frequency offset (CFO), usually based on the autocorrelation of the received signal. In this paper, we derive a family of novel data-aided CFO estimators. The proposed estimators are based on a new autocorrelation function which is defined using cyclostationary properties of the repetitive preamble. In contrast to previous approaches, the new estimators make use of high-order noise terms leading to an improved performance. We present a detailed analysis of the proposed estimators and provide closed-form expressions for the variance of the estimators. The new estimators are shown to outperform the existing estimators obtaining a moderate improvement at high signal to noise ratio (SNR) and a considerable improvement at low SNR, by means of a reasonable increase in computational complexity.

► We derived a family of CFO estimators based on the cyclic autocorrelation function. ► The proposed algorithms improve the performance for low and high SNR. ► In order to use all available information high-order noise term cannot be discarded.