SemSpaceFL: A Collaborative Hierarchical Federated Learning Framework for Semantic Communication in 6G LEO Satellites

TL;DR

The paper tackles bandwidth, privacy, and connectivity challenges in 6G LEO satellite networks by introducing SemSpaceFL, a mobility-aware two-tier hierarchical federated learning framework that jointly trains semantic encoding-decoding (DJSCC) models. It decomposes learning into local satellite updates, gateway-level sub-region aggregation, and cloud-level global aggregation, with novel, mobility- and energy-aware satellite association and dynamic, data- and loss-informed gateway weighting. The proposed approach yields faster convergence and superior reconstruction quality (PSNR and MS-SSIM) compared to traditional FL baselines, while efficiently utilizing satellite resources under realistic channel and energy constraints. This work advances practical, scalable, and privacy-preserving semantic communication in dynamic 6G LEO networks, with potential impact on bandwidth-efficient satellite AI services and network intelligence.

Abstract

The advent of the sixth-generation (6G) wireless networks, enhanced by artificial intelligence, promises ubiquitous connectivity through Low Earth Orbit (LEO) satellites. These satellites are capable of collecting vast amounts of geographically diverse and real-time data, which can be immensely valuable for training intelligent models. However, limited inter-satellite communication and data privacy constraints hinder data collection on a single server for training. Therefore, we propose SemSpaceFL, a novel hierarchical federated learning (HFL) framework for LEO satellite networks, with integrated semantic communication capabilities. Our framework introduces a two-tier aggregation architecture where satellite models are first aggregated at regional gateways before final consolidation at a cloud server, which explicitly accounts for satellite mobility patterns and energy constraints. The key innovation lies in our novel aggregation approach, which dynamically adjusts the contribution of each satellite based on its trajectory and association with different gateways, which ensures stable model convergence despite the highly dynamic nature of LEO constellations. To further enhance communication efficiency, we incorporate semantic encoding-decoding techniques trained through the proposed HFL framework, which enables intelligent data compression while maintaining signal integrity. Our experimental results demonstrate that the proposed aggregation strategy achieves superior performance and faster convergence compared to existing benchmarks, while effectively managing the challenges of satellite mobility and energy limitations in dynamic LEO networks.

Figures (7)

• Figure 1: Illustration of collaborative hierarchical federated learning framework for semantic communication in 6G LEO satellites.
• Figure 2: Illustration of the proposed association approach.
• Figure 3: The performance convergence of FL under single Gateway and the collaboration scenarios.
• Figure 4: The convergence line of HFL against single gateway FL scenarios for the MS-SSIM metric.
• Figure 5: A comparison of the average training epochs per global round for each satellite across two distinct association approaches.