Evaluation of MIMO Systems with Respect to Front-End Imperfections

In multiuser multiple-input multiple-output (MIMO) systems the performance of joint detection and joint predistortion techniques highly depends on the quality of radio-frequency (RF) components. Therefore, the focus of this paper is on front-end imperfections. We show the influence of erroneous calibration (which is essential for downlink predistortion), phase noise, and nonlinear amplifiers. For all issues we elaborate on the different disturbances that arise, analyze the impact on symbol estimation, and evaluate system performance in terms of coded bit error ratio (BER). A general result for all examined frontend imperfections is that their influence increases with the number of spatially multiplexed data streams culminating in unacceptable performance degradation in fully loaded systems.