Aerial Semantic Relay-Enabled SAGIN: Joint UAV Deployment and Resource Allocation

TL;DR

The paper tackles a multi-cluster SAGIN scenario where UAV relays enable semantic communication to SemUsers while concurrently serving ConUsers with traditional bit-level transmission. It formulates a non-convex joint optimization problem over transmit powers, bandwidths, and UAV positions, and solves it via an auxiliary-variable based alternating optimization procedure that decomposes the problem into convex subproblems. The proposed AO framework yields substantial sum-rate and spectral-efficiency gains across diverse channel conditions and user distributions, validating the benefits of jointly optimizing UAV deployment and cross-layer resources in semantic-relay SAGINs. The findings underscore the practical impact of semantic-aware routing and cooperative UAV deployment for future 6G networks, enabling broader coverage and higher efficiency with heterogeneous devices.

Abstract

Space-Air-Ground Integrated Networks (SAGINs) are pivotal for enabling ubiquitous connectivity in 6G systems, yet they face significant challenges due to severe satellite-to-ground link impairments. Although Unmanned Aerial Vehicles (UAVs) can function as relay nodes to compensate for air-to-ground channel degradation, the satellite-to-UAV link remains a critical bottleneck. Semantic Communication (SemCom) emerges as a promising solution to enhance spectral efficiency by transmitting essential semantic information. This paper proposes a novel multi-cluster UAV-aided SAGIN SemCom architecture that supports both semantic users (SemUsers) and conventional users (ConUsers). While SemCom is employed in the satellite-to-UAV link to improve transmission efficiency, the UAVs implement an intelligent adaptive relay strategy, capable of either directly forwarding semantic data to SemUsers or converting it into bit-level data for ConUsers. Compared to existing similar schemes, this design guarantees the high-efficiency advantages of SemCom while enabling network access for larger coverage area. A joint optimization problem is formulated to maximize the system's sum-rate through coordinated allocation of power, bandwidth, and UAV positions. To address this non-convex problem, we develop an efficient alternating optimization (AO) algorithm, which decomposes the original problem into tractable subproblems. Numerical results demonstrate that the proposed algorithm significantly outperforms baseline schemes in terms of both sum-rate and spectral efficiency across various channel conditions and user distributions, underscoring the importance of joint resource allocation and intelligent UAV deployment.

Figures (11)

• Figure 1: Illustration of the downlink semantic relay system in the considered SAGIN architecture. A satellite serves as the base station in space, while $N$ UAVs function as aerial relays, each supporting a cluster of ground users.
• Figure 2: Illustration of dual-mode transmission schemes for semantic-capable and conventional users. The satellite encodes source data into semantic representations. Upon reception, the UAV implements two distinct processing strategies: for semantic users, it forwards semantic symbols directly; for conventional users, it performs semantic decoding and transmits reconstructed bit-level data.
• Figure 3: Sum-rate versus $\tilde{B_R}$ with different algorithms and $\tilde{B_S}$.
• Figure 4: Sum-rate versus $\tilde{P_R}$ with different algorithms and $\tilde{P_S}$.
• Figure 5: Sum-rate versus $\tilde{B_R}$ in different user scenarios.

Theorems & Definitions (12)

• proof
• Lemma 1
• proof
• Remark 1
• Lemma 2
• proof
• Remark 2
• Lemma 3
• proof
• Lemma 4