Robust Equalizer for Multicell Massive MIMO Uplink With Channel Model Uncertainty

We study the robust linear equalizer for the uplink of massive multiple-input–multiple-output (MIMO) systems with multicell pilot reuse. We use the worst-case approach for robust design to combat pilot contamination. When the number of base station (BS) antennas is large, we build a novel model to indicate the relationship between the instantaneous channel matrix and the imperfect channel estimation in the presence of channel model uncertainty. Based on this model, we formulate the robust equalizer design problem into a min–max problem. Furthermore, we transform the min–max problem into an unconstrained one, and the optimality conditions are derived. With the resulting optimality conditions, two iterative algorithms and a simple approximation method are proposed to compute the optimal robust equalizers. Simulations are adopted to evaluate the performance of the proposed algorithms and the approximation method. Compared with the conventional equalizers, the proposed robust equalizers achieve a better bit error rate (BER) performance, particularly in the regime of high signal-to-noise ratio (SNR), where pilot contamination is significant.