Finite-alphabet based blind carrier frequency offset estimation for differentially coded OFDM

A novel blind carrier frequency offset estimator is proposed for differentially coded OFDM. Exploiting differential encoding and finiteness of signal constellation, our carrier-frequency offset estimation is carried out without requiring either channel knowledge or training. It is proved that under certain mild assumptions, the proposed estimator is capable of identifying the carrier-frequency offset with probability one. Different from some existing estimators, our proposed estimator does not rely on virtual subcarriers. More importantly, it is able to achieve reliable carrier frequency offset estimation by using only two received data blocks, which makes it very attractive especially for short-burst communications. The merits of our estimator are confirmed both by theoretical analysis and by simulations.