Approaching Massive MIMO Performance with Reconfigurable Intelligent Surfaces: We Do Not Need Many Antennas

TL;DR

This paper considers an antenna structure where a (non-large) array of radiating elements is placed at short distance in front of a reconfigurable intelligent surface (RIS) to be able to approach massive MIMO performance levels in a cost-effective way with reduced hardware resources.

Abstract

This paper considers an antenna structure where a (non-large) array of radiating elements is placed at short distance in front of a reconfigurable intelligent surface (RIS), herein nicknamed reconfigurable intelligent base station (RIBS). We firstly derive a closed-form expression for the channel between the array of radiating elements and the RIS that captures the near-field effects, and give some considerations on the channel hardening and favorable propagation in this scenario. Focusing on both active and passive RIS, we describe channel estimation and downlink signal processing techniques suitable for the RIBS structure. Additionally, we formulate and solve an optimization problem aimed at maximizing the fairness among the users with respect to the downlink power coefficients and RIS configuration both in the cases of active and passive RIBS. Numerical results show that the proposed structure is effective and capable of outperforming conventional non-RIS aided MIMO systems, especially in the case of active RIBS. The proposed antenna structure is thus shown to be able to approach massive MIMO performance levels in a cost-effective way with reduced hardware resources.

Figures (9)

• Figure 1: \ref{['subfig:RIBS']} A UPA with $N_{\text{A}}$ active antennas is mounted at close distance from an RIS with $N_{\text{R}}$ reflective elements. The relative positioning of the active array with respect to the RIS is such that no UE is obstructed. \ref{['subfig:back_view']} The BS antennas have size $\Delta_{\text{A}}^{\text{h}}$ along the $x$-axis and $\Delta_{\text{A}}^{\text{v}}$ along the $y$-axis, and are spaced by $d_{\text{A}}^{\text{h}}$ and $d_{\text{A}}^{\text{v}}$ in the $x$- and $y$-axis, respectively. The RIS elements have size $\Delta_{\text{R}}^{\text{h}}$ along the $x$-axis and $\Delta_{\text{R}}^{\text{v}}$ along the $y$-axis, and are spaced by $d_{\text{R}}^{\text{h}}$ and $d_{\text{R}}^{\text{v}}$ in the $x$- and $y$-axis, respectively. \ref{['subfig:side_view']} View of the RIBS in the $yz$-plane, and characterization of the cascaded channel. $D$ is the distance between the RIS and the BS, while $\alpha$ denotes the downtilt angle of the BS with respect to the RIS. The heights of RIS, BS and UE are denoted by $h_{\text{RIS}}$, $h_{\text{BS}}$ and $h_{\text{UE}}$, respectively.
• Figure 2: Squared Frobenius norm of the BS-to-RIS channel $\mathbf{H}$ obtained via \ref{['eq:Hnear']} and \ref{['eq:H_ij_planar']} versus $D/\lambda$. Here, we assume $N_{\text{A}}\!=\!16$, $N_{\text{R}}\!=\!64$ and $\alpha\!=\!\pi/6$. The Fraunhofer distance corresponds to $D\!=\!20\lambda.$
• Figure 3: CDF of the DL SE per UE achieved with MR, under pCSI and iCSI assumption. Performance comparison include RIBS with passive and active RIS and co-located MIMO (mMIMO) with $M$ antennas. Here, $K\!=\!8$, $N_{\text{A}}\!=\!16$, $N_{\text{R}}\!=\!64$, $\alpha\!=\!\pi/6$ and $a_{\text{max}}\!=\!5$.
• Figure 4: CDF of the DL SE per UE achieved with MMSE, under pCSI and iCSI assumption. Performance comparison include RIBS with passive and active RIS and co-located MIMO (mMIMO) with $M$ antennas. Here, $K\!=\!8$, $N_{\text{A}}\!=\!16$, $N_{\text{R}}\!=\!64$, $\alpha\!=\!\pi/6$ and $a_{\text{max}}\!=\!5$.
• Figure 5: Power split between RIS and BS in an active RIBS, assuming MMSE and MR with the two initialization strategies proposed in \ref{['eq:p-adjusting:active']} and \ref{['eq:eta-adjusting:active']}. The simulation settings are the same as those used to obtain Figs. \ref{['Fig:MRT_pCSI_ipCSI']} and \ref{['Fig:MmSE_pCSI_ipCSI']}.

Theorems & Definitions (1)

• Lemma 1