Abstract

In this paper, we address the problem of estimating channel impulse response (CIR) and carrier frequency offset (CFO) in OFDM systems utilizing the method of maximum likelihood. When the CFO is small, Moose's technique for CFO estimation is accurate and numerically efficient. However, if the CFO is high, there are no straightforward solutions for its estimation. We show that the log-likelihood function of the CIR and the CFO can be concentrated in one variable (the CFO), and that its estimation can be accurately carried out by first performing a coarse line search and then applying Newton's method at a low computational cost. We illustrate the benefits of our proposal via numerical simulations.