Preamble-Based MMSE Channel Estimation with Low Pilot Overhead in MIMO-FBMC Systems

Abstract

As is well-known, high pilot overhead in channel estimation is a critical problem due to the imaginary interference in multiple input multiple output (MIMO) employing filter bank multicarrier (FBMC) (MIMO-FBMC) systems. In this paper, we propose a novel preamble design with low overhead and present a minimum mean square error (MMSE) algorithm for the channel estimation, which can improve the system performance in harsh environments. In the proposed preamble, 4 columns of pilots are required, including two columns of zeros to avoid the imaginary interference from data. It is shown that, the imaginary interference between non-zero pilots can be removed by adopting symmetry design. Therefore, pilot overhead is reduced compared with the conventional scheme with 5 columns of pilots. In addition, the noises at adjacent subcarriers are demonstrated to be correlated in MIMO-FBMC systems, and the correlation is considered in the proposed MMSE algorithm. Numerical simulations are conducted to verify the proposed scheme, and for comparison, we also give the performance of the conventional interference approximation method (IAM).