Optimal detection in multiple-input multiple-output orthogonal frequency-division multiplexing systems with imperfect channel estimation

An optimal detector for orthogonal frequency-division multiplexing (OFDM) signals with pilot symbol assisted modulation in multiple-input multiple-output (MIMO) frequency selective fading channels with imperfect channel estimation is derived. This detector does not estimate the channel explicitly but maximises the likelihood function incorporating received data and pilot symbols to recover the data. The authors investigate and compare its performance with that of mismatched detectors treating maximum likelihood (ML), regularised ML or minimum mean squared error (MMSE) channel estimates as perfect channel information in spatially uncorrelated MIMO channels with binary phase shift keying (BPSK) and 16-quadrature amplitude modulation (QAM) by simulation. The authors use basis functions to model the channel frequency response and compare the performance of different basis expansion models in frequency selective fading channels with different delay spread. In single-input single-output (SISO) channels, the performance of the optimal detector is close to that of the mismatched detector with MMSE channel estimates, which provides the best performance among these mismatched detectors. However, the optimal detector significantly outperforms the mismatched detectors in MIMO channels.