Multi-User MISO with Stacked Intelligent Metasurfaces: A DRL-Based Sum-Rate Optimization Approach

TL;DR

This work addresses interference and hardware complexity in multi-user MISO by introducing SIM-enabled, wave-domain precoding. A DRL framework based on Deep Deterministic Policy Gradient (DDPG) with whitening and a residual CNN jointly optimizes SIM phase shifts and transmit powers to maximize the sum-rate, addressing the non-convex joint design. Results show significant gains over AO and baselines, with improved robustness from hyperparameter tuning and the whitening process. The approach offers scalable, low-complexity interference cancellation suitable for large SIM-enabled wireless systems and motivates further exploration of broadband and discrete-space optimization in SIM contexts.

Abstract

Stacked intelligent metasurfaces (SIMs) represent a novel signal processing paradigm that enables over-the-air processing of electromagnetic waves at the speed of light. Their multi-layer architecture exhibits customizable computational capabilities compared to conventional single-layer reconfigurable intelligent surfaces and metasurface lenses. In this paper, we deploy SIM to improve the performance of multi-user multiple-input single-output (MISO) wireless systems through a low complexity manner with reduced numbers of transmit radio frequency chains. In particular, an optimization formulation for the joint design of the SIM phase shifts and the transmit power allocation is presented, which is efficiently tackled via a customized deep reinforcement learning (DRL) approach that systematically explores pre-designed states of the SIM-parametrized smart wireless environment. The presented performance evaluation results demonstrate the proposed method's capability to effectively learn from the wireless environment, while consistently outperforming conventional precoding schemes under low transmit power conditions. Furthermore, the implementation of hyperparameter tuning and whitening process significantly enhance the robustness of the proposed DRL framework.

Figures (13)

• Figure 1: The considered SIM-aided multi-user MISO transmission system with $L$ layers of metasurfaces.
• Figure 2: The proposed DRL-based SIM design framework adopting DDPG.
• Figure 3: Architecture of the actor network and the critic network.
• Figure 4: Simulation setup of a SIM-assisted multi-user MISO system.
• Figure 5: Average sum rate versus the transmit power $P$ for SIM-assisted schemes considering $M=4,\, L=4,\,\text{and}\, N=81$.