Generative AI-enhanced Low-Altitude UAV-Mounted Stacked Intelligent Metasurfaces

TL;DR

This work tackles the challenge of achieving high-capacity uplink communication in low-altitude economy networks by integrating mobile UAV-mounted stacked intelligent metasurfaces (UAV-SIMs). It formulates a joint optimization problem (USBJOP) over user association, 3D UAV positioning, and multi-layer SIM phase shifts, proving NP-hard and non-convex. An alternating-optimization framework decomposes USBJOP into AUUOP, ULOP, and USPSOP, solving the first two with convex CVX-based methods and the third with a hybrid generative approach (HGPSO) that combines layer-by-layer optimization (LBL-IPSO) and a CVAE-based model (CVAE-M) using a capacity-aware loss. Simulation results show approximately 1.5× gains in network capacity over benchmarks, with SIM layers and meta-atoms effectively mitigating multi-user interference, and the GAI component reducing runtime by about 10% while preserving solution quality. The proposed framework demonstrates practical viability for real-world UAV-SIM deployments in LAE networks, offering substantial performance and efficiency benefits for dynamic wireless environments.

Abstract

Wireless communication systems face challenges in meeting the demand for higher data rates and reliable connectivity in complex environments. Stacked intelligent metasurfaces (SIMs) have emerged as a promising technology for advanced wave-domain signal processing, where mobile SIMs can outperform fixed counterparts. In this paper, we propose a novel unmanned aerial vehicle (UAV)-mounted SIM (UAV-SIM) assisted communication system within low-altitude economy (LAE) networks, where UAVs act as both cache-enabled base stations and mobile SIM carriers to enhance uplink transmissions. To maximize network capacity, we formulate a UAV-SIM-based joint optimization problem (USBJOP) that integrates user association, UAV-SIM three-dimensional positioning, and multi-layer SIM phase shift design. Due to the non-convexity and NP-hardness of USBJOP, we decompose it into three subproblems, which are the association between UAV-SIMs and users optimization problem (AUUOP), the UAV location optimization problem (ULOP), and the UAV-SIM phase shifts optimization problem (USPSOP). Then, we solve them through an alternating optimization strategy. Specifically, AUUOP and ULOP are transformed into convex forms solvable via the CVX tool, while USPSOP is addressed by a generative artificial intelligence (GAI)-based hybrid optimization algorithm. Simulation results show that the proposed approach achieves approximately 1.5 times higher network capacity compared with suboptimal schemes, effectively mitigates multi-user interference with increasing SIM layers and meta-atoms, and reduces runtime by 10\% while maintaining solution quality, thereby demonstrating its practicality for real-world deployments.

Figures (7)

• Figure 1: An uplink communication system assisted by UAV-SIM within the LAE networks paradigm, in which multiple UAV-SIMs are deployed at a fixed altitude while maintaining a safe distance, to process data from multiple ground users.
• Figure 2: The structure of the AO strategy framework in LAE networks.
• Figure 3: Algorithm convergence effect diagram. (a) Loss curve diagram of CVAE model under different learning rates. (b) Convergence curves of different algorithms.
• Figure 4: Network performance versus different loss functions.
• Figure 5: Network capacity $R$ versus SIM configuration. (a) Network capacity $R$ versus the number of metasurface layers $L$ with $K=5$ and $M=3$. (b) Network capacity $R$ versus the number of meta-atoms on each layer $N$ with $K=5$ and $M=3$. (c) Network capacity $R$ versus the number of metasurface layers $L$ with $K=8$ and $M=5$. (d) Network capacity $R$ versus the number of meta-atoms on each layer $N$ with $K=8$ and $M=5$.