Computationally efficient frequency offset estimation for flat-fading MIMO channels: performance analysis and training sequence design

The paper deals with carrier frequency offset estimation for a flat-fading multiple-input multiple-output (MIMO) channel using a training sequence. The resulting maximum likelihood (ML) estimation entails solving a maximization problem with no closed-form solution. Since numerical calculation of the estimate is computationally hard, we propose a sub-optimal closed-form solution. In contrast with single-input-single-output (SISO) systems, however, self-noise arises in MIMO closed-form frequency offset estimation. Through proper training sequence design, we show how to avoid this self-noise and achieve a performance close to ML-performance and the Cramer-Rao bound (CRB).