Channel-Adaptive Robust Aggregation for Over-the-Air Federated Learning in Heterogeneous Networks

2026-07-05

AI Summary

This work addresses the degradation in model accuracy and convergence difficulties in over-the-air federated learning (AirFL) over heterogeneous wireless networks, where fixed aggregation schedules fail to account for channel noise, fading, and client heterogeneity. To overcome these challenges, the paper proposes CHARGE-FL, a novel framework that introduces, for the first time, a joint channel-aware and task-driven adaptive aggregation mechanism. CHARGE-FL dynamically schedules aggregation rounds, tailors optimization strategies per client, and employs a dual-purpose precoding technique to jointly optimize communication and learning processes. Evaluated under real-world wireless conditions, CHARGE-FL significantly enhances model accuracy, training stability, and convergence speed, with pronounced advantages particularly evident under high client heterogeneity and adverse channel conditions.

Abstract

The growing demand for privacy-preserving, data-intensive applications such as IoT, augmented reality, and autonomous systems positions Federated Learning (FL) as a key enabler in 6G networks. Over-the-Air FL (OTA-FL) leverages the superposition property of the wireless multiple access channel for efficient aggregation via simultaneous transmissions. Existing methods rely on fixed aggregation schedules and do not jointly address noise, fading, and client heterogeneity. We propose CHARGE-FL (CHannel-Adaptive Robust agGrEgation), a framework that adaptively schedules aggregation based on channel dynamics and application readiness. By combining a tailored optimization strategy with a dual-purpose precoding mechanism, CHARGE-FL mitigates channel distortion and bias from partial updates, achieving superior accuracy, stability, and convergence under realistic wireless conditions. Empirical results under realistic wireless conditions show that CHARGE-FL significantly improves accuracy, stability, and convergence over state-of-the-art OTA-FL methods, particularly in straggler-prone and noisy scenarios.

Problem

• Over-the-Air Federated Learning
• channel noise
• fading
• client heterogeneity
• aggregation scheduling

Innovation

• Channel-Adaptive Aggregation
• Over-the-Air Federated Learning
• Robust Precoding
• Client Heterogeneity
• Wireless FL