Fluid Antenna-Assisted MIMO Transmission Exploiting Statistical CSI

TL;DR

This work tackles rate maximization for a fluid antenna (FA)–assisted MIMO system under slow-varying statistical CSI, motivated by the difficulty of obtaining instantaneous CSI when antenna positions change. It develops an alternating-optimization framework that uses Jensen's inequality to form an analytical upper bound $\overline{R}$ and yields a closed-form transmit covariance $\mathbf{Q}^* \propto \mathbf{G}(\mathbf{t})^H \mathbf{G}(\mathbf{t})$, together with second-order Taylor updates for optimizing transmit and receive FA positions. Simulations show substantial rate gains of FA-based designs over fixed-position and moveable baselines across SNR and region size, with gains up to around 38% in favorable regimes; the results indicate that a region size around $A/\lambda \approx 2$ suffices to achieve near-maximum performance. Overall, the work demonstrates practical and scalable gains for FA–MIMO systems by leveraging statistical CSI and decoupled, tractable optimization of precoding and FA positions.

Abstract

In conventional multiple-input multiple-output (MIMO) communication systems, the positions of antennas are fixed. To take full advantage of spatial degrees of freedom, a new technology called fluid antenna (FA) is proposed to obtain higher achievable rate and diversity gain. Most existing works on FA exploit instantaneous channel state information (CSI). However, in FA-assisted systems, it is difficult to obtain instantaneous CSI since changes in the antenna position will lead to channel variation. In this letter, we investigate a FA-assisted MIMO system using relatively slow-varying statistical CSI. Specifically, in the criterion of rate maximization, we propose an algorithmic framework for transmit precoding and transmit/receive FAs position designs with statistical CSI. Simulation results show that our proposed algorithm in FA-assisted systems significantly outperforms baselines in terms of rate performance.

Figures (3)

• Figure 1: Convergence of Algorithm 1 under different values of $P_{\text{max}}$.
• Figure 2: Rate with respect to SNR.
• Figure 3: Rate versus the size of the normalized region $A/\lambda$.