An efficient coarse frequency offset estimator for OFDM

A novel scheme of integer frequency offset (IFO) estimation for OFDM systems is proposed based on an OFDM training symbol with several identical parts. In this scheme, the received OFDM training symbol is first reshaped into several sub-symbols by selecting the samples with the same index in each part. It shows that the reshaping process introduces time diversity. The IFO is estimated by finding the maximum magnitude of the frequency domain signals obtained by taking a modified FFT algorithm on some sub-symbols. To efficiently use the multiple sub-symbols, an adaptive time diversity combining method based on estimate reliability decision is presented. Simulation results show that the proposed IFO estimate scheme has an improved estimate performance compared to conventional estimate scheme while maintaining a lower complexity.