Modality-Decoupled Federated Learning for Privacy-Preserving Embodied Intelligence in 6G
作者: Zhuodong Liu, Xiangyu Li, Chunhong Yuan, Hongyang Du, Bodong Shang, Qingqing Wu, Tony Q. S. Quek, Mohsen Guizani
分类: eess.SP, cs.AI
发布日期: 2026-09-09
备注: This article has been accepted for publication in IEEE Wireless Commnunications Magazine
💡 一句话要点
提出FedMVLA框架以解决6G环境下隐私保护问题
🎯 匹配领域: 支柱一:机器人控制 (Robot Control) 支柱九:具身大模型 (Embodied Foundation Models)
关键词: 联邦学习 隐私保护 模态解耦 具身智能 6G网络 机器人协作 通信效率
📋 核心要点
- 现有的联邦学习方法未能有效处理视觉、语言和动作路径之间的内在差异,导致隐私保护和通信效率不足。
- 本文提出FedMVLA框架,通过模态感知的聚合、隐私分配和通信压缩机制,解决了多模态学习中的隐私和效率问题。
- 实验结果表明,FedMVLA在任务成功率上达到84.8%,比FedAvg提高22.2个百分点,同时显著减少了每客户端的上传负载。
📝 摘要(中文)
第六代(6G)无线网络预计将为大规模的具身智能提供关键基础设施,其中异构机器人通过低延迟连接、边缘智能和分布式感知进行协作。视觉-语言-动作(VLA)模型通过将视觉感知、语言理解和动作生成整合为统一的闭环策略,为这一目标奠定了基础。然而,将VLA模型训练和适应于分布式机器人代理面临隐私保护、通信效率和模型异质性等挑战。现有的联邦学习(FL)方法忽视了视觉、语言和动作路径在参数规模、隐私暴露、更新动态和对压缩或扰动的容忍度等方面的内在差异。为了解决这一问题,本文提出了FedMVLA,一个用于6G网络中隐私保护的模态解耦FL框架。FedMVLA结合了模态感知的联邦聚合、隐私分配和通信压缩机制,并设计了模态切片传输,以保护精确关键的动作流。
🔬 方法详解
问题定义:本文旨在解决在6G环境下,分布式机器人在训练视觉-语言-动作模型时面临的隐私保护、通信效率和模型异质性等问题。现有的联邦学习方法未能充分考虑不同模态之间的差异,导致隐私风险和效率低下。
核心思路:FedMVLA框架通过模态解耦的方式,分别处理视觉、语言和动作模态的聚合、隐私分配和通信压缩,从而提高隐私保护和通信效率。这样的设计使得每种模态可以根据其特性进行优化,避免了模态间的干扰。
技术框架:FedMVLA的整体架构包括三个主要模块:模态感知的联邦聚合(MAFA)、模态感知的隐私分配(MAPA)和模态感知的通信压缩(MACO)。此外,设计了模态切片传输,以确保关键动作流的低延迟和高可靠性。
关键创新:FedMVLA的主要创新在于模态解耦的联邦学习框架,能够针对不同模态的特性进行优化,显著提高了隐私保护和通信效率。这与传统的联邦学习方法形成了鲜明对比,后者通常将所有模态视为统一的处理对象。
关键设计:在FedMVLA中,模态感知的聚合机制允许不同模态的参数以不同的方式进行更新,隐私分配机制则根据模态的隐私风险进行动态调整,通信压缩机制则通过优化数据传输路径来减少上传负载。
🖼️ 关键图片
📊 实验亮点
实验结果显示,FedMVLA在任务成功率上达到84.8%,比基线FedAvg提高了22.2个百分点。同时,在扩展到128个客户端的情况下,上传负载减少了约96%,并且关键上传完成时间保持在1.5秒左右。
🎯 应用场景
该研究的潜在应用领域包括智能机器人、自动驾驶、远程医疗等需要高效隐私保护的场景。FedMVLA框架能够在保证隐私的前提下,实现多模态数据的高效协作,为未来的6G网络应用提供了重要的技术支持。
📄 摘要(原文)
Sixth-generation (6G) wireless networks are expected to provide a key infrastructure for large-scale embodied intelligence, where heterogeneous robots collaborate through low-latency connectivity, edge intelligence, and distributed sensing. Vision-language-action (VLA) models offer a foundation by integrating visual perception, language understanding, and action generation into a unified closed-loop policy. However, training and adapting VLA models to distributed robotic agents introduce challenges in privacy protection, communication efficiency, and model heterogeneity. Existing federated learning (FL) methods overlook the intrinsic differences among vision, language, and action pathways in parameter scale, privacy exposure, update dynamics, and tolerance to compression or perturbation. To address this issue, this article proposes FedMVLA, a modality-decoupled FL framework for privacy-preserving embodied intelligence in 6G networks. FedMVLA incorporates three mechanisms: modality-aware federated aggregation (MAFA), modality-aware privacy allocation (MAPA), and modality-aware communication compression (MACO), together with a modality-sliced transport design that routes the precision-critical action stream through a protected ultra-reliable low-latency slice. A case study on federated robotic manipulation over the Third Generation Partnership Project (3GPP)-based wireless substrate, covering fading, co-channel interference, and malicious jamming, shows that FedMVLA achieves an 84.8% task success rate, exceeds FedAvg by 22.2 percentage points, sustains a widening margin when scaling to 128 clients across eight cells, and reduces the schedule-averaged per-client uplink model-update payload by 95.6% (approximately 96%), while keeping the 95th percentile (p95) of the round-critical uplink completion time near 1.5s.