Split Federated Learning for UAV-Enabled Integrated Sensing, Computation, and Communication

2025-04-02

AI Summary

To address computational overload, privacy leakage, and communication inefficiency in federated learning (FL) for resource-constrained unmanned aerial vehicles (UAVs) within integrated sensing, computing, and communication (ISCC) scenarios, this paper proposes Split Federated Learning for UAV–Server Collaboration (SFLSCC)—the first FL framework enabling collaborative model training partitioning between UAVs and edge servers. We innovatively establish a joint convergence bound theory that jointly incorporates deployment configuration, model partitioning points, sensing workload, and aggregation frequency. Furthermore, we design a low-complexity joint optimization algorithm to simultaneously maximize energy efficiency and model accuracy. Experimental results on target motion recognition demonstrate that SFLSCC achieves a 37% faster convergence rate and reduces UAV-side energy consumption by 52% compared to baseline methods, while strictly preserving data locality and ensuring privacy security.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• Security and Privacy: Security and privacy of machine learning and AI applications

Abstract

Unmanned aerial vehicles (UAVs) with integrated sensing, computation, and communication (ISCC) capabilities have become key enablers of next-generation wireless networks. Federated edge learning (FEL) leverages UAVs as mobile learning agents to collect data, perform local model updates, and contribute to global model aggregation. However, existing UAV-assisted FEL systems face critical challenges, including excessive computational demands, privacy risks, and inefficient communication, primarily due to the requirement for full-model training on resource-constrained UAVs. To deal with aforementioned challenges, we propose Split Federated Learning for UAV-Enabled ISCC (SFLSCC), a novel framework that integrates split federated learning (SFL) into UAV-assisted FEL. SFLSCC optimally partitions model training between UAVs and edge servers, significantly reducing UAVs' computational burden while preserving data privacy. We conduct a theoretical analysis of UAV deployment, split point selection, data sensing volume, and client-side aggregation frequency, deriving closed-form upper bounds for the convergence gap. Based on these insights, we conceive a joint optimization problem to minimize the energy consumption required to achieve a target model accuracy. Given the non-convex nature of the problem, we develop a low-complexity algorithm to efficiently determine UAV deployment, split point selection, and communication frequency. Extensive simulations on a target motion recognition task validate the effectiveness of SFLSCC, demonstrating superior convergence performance and energy efficiency compared to baseline methods.

Problem

• Reduces computational burden on resource-constrained UAVs
• Enhances data privacy in federated edge learning
• Optimizes energy efficiency for model convergence

Innovation

• Split Federated Learning for UAV-Enabled ISCC
• Optimal model partitioning between UAVs and edge servers
• Joint optimization for energy-efficient UAV deployment