Blind estimation of residual timing error in OFDM receivers: a non-data-aided maximum-likelihood approach

This paper proposes a novel non-data-aided maximum likelihood (ML) approach for the estimation of the residual timing error in OFDM receivers. The novel approach effectively utilizes the finite alphabet property of the received symbol constellation to perform a near perfect residual timing error estimation. Unlike some of the current techniques, the proposed approach requires no pilots and therefore is bandwidth efficient. Moreover, the reduced complexity version of the post-FFT ML algorithm minimizes the receiver computational burden. Simulation results show that the BER degradation due to residual timing error can be almost completely recovered for both AWGN and Rayleigh fading channel scenarios