Abstract
Meta-material-based antenna designs, such as large intelligent surface (LIS), are expected to be a game changer in future wireless cellular systems, since they provide a simple yet effective mean of drastically improving the wireless propagation environment. This paper investigates the ergodic capacity of LIS-aided multiple input multiple output (MIMO), a.k.a. MIMO-LIS, systems. To this end, the derivation of the probability density function (pdf) of the cascaded channel, i.e. the transmitter-to-LIS-to-receiver channel, is studied. Moreover, both high signal-to-noise ratio (SNR) asymptotic expression and closed-form approximations of this ergodic capacity are provided. Monte-Carlo simulations graphically validate the correctness and accuracy of our various expressions, for different antenna configurations. Furthermore, our performance analysis shows that the MIMO-LIS system outperforms both MIMO-AF and MIMO systems (by more than 60% and 15% respectively, at a 30 dB SNR) from an ergodic capacity point of view, which confirms that LIS can be beneficial for improving the propagation environment.