An Enhanced DFT-Based Channel Estimator for LTE-A Uplink

In this paper, we investigate the popular discrete Fourier transform (DFT)-based channel estimator (D-bCE) for Long-Term Evolution (LTE)-Advanced uplink. Due to limitations of the current pilot pattern, orthogonality among channels of multiple antennas at the time domain does not exactly hold. This degrades the performance of a conventional D-bCE severely. Therefore, we propose a novel enhancement of the D-bCE by using the frequency-domain weighting scheme. By analyzing the mean square error, we present a heuristic way to design weights and reformulate the resulting optimization into a second-order cone programming problem, which can be easily solved. The proposed scheme does not rely on the channel covariance matrix, and the weights can be generated offline. Thus, it is easy to use this scheme in practice. Simulation results validate the effectiveness of our proposed method.