Frequency Offset Estimation for OFDM Systems Using ICI Self-Cancellation Schemes

This paper presents a novel frequency offset estimator for inter-carrier-interference (ICI) self-cancellation coded orthogonal frequency division multiplexing (OFDM) systems, which map data symbols onto adjacent or symmetric subcarriers with weighting coefficients of +1 and -1. There are two major drawbacks for traditional ICI self-cancellation schemes. First, the common phase error (CPE) caused by the carrier frequency offset (CFO) can not be removed. Second, the performance is substantially degraded when CFO is large. To remedy these drawbacks, pilot sub-carriers are inserted and the maximum likelihood estimator (MLE) proposed by Moose (1994) is adopted to eliminate CFO. If residual CFO is small enough, traditional ICI self-cancellation scheme can operate efficiently. In this paper, a second order estimator (SOE) is proposed to exploit the second order statistical properties of the demodulated pilot signals at the output of ICI self-cancellation demodulation. The SOE is combined with MLE to estimate CFO and the combined estimator is termed second order with maximum likelihood estimator (SO-MLE). The optimal weighting coefficient is determined by conducting simulation experiments. It is demonstrated that the proposed SO-MLE has the advantage of tolerating larger CFO and CPE than traditional MLE, and the ICI can thus be significantly diminished.