Over-the-Air Aggregation for Federated Learning: Waveform Superposition and Prototype Validation

TL;DR

This paper presents an OFDM-based over-the-air aggregation method for wireless federated learning, addressing synchronization challenges with a pre-equalization technique, and validates it through hardware experiments achieving high prediction accuracy.

Contribution

The paper introduces a novel waveform pre-equalization approach and a customized protocol for OTA aggregation in federated learning, validated with real hardware experiments.

Findings

Achieves comparable performance to offline training

Effectively mitigates frame timing and frequency offsets

Demonstrates high prediction accuracy in real-world tests

Abstract

In this paper, we develop an orthogonal-frequency-division-multiplexing (OFDM)-based over-the-air (OTA) aggregation solution for wireless federated learning (FL). In particular, the local gradients in massive IoT devices are modulated by an analog waveform and are then transmitted using the same wireless resources. To this end, achieving perfect waveform superposition is the key challenge, which is difficult due to the existence of frame timing offset (TO) and carrier frequency offset (CFO). In order to address these issues, we propose a two-stage waveform pre-equalization technique with a customized multiple access protocol that can estimate and then mitigate the TO and CFO for the OTA aggregation. Based on the proposed solution, we develop a hardware transceiver and application software to train a real-world FL task, which learns a deep neural network to predict the received signal strength with global positioning system information. Experiments verify that the proposed OTA aggregation solution can achieve comparable performance to offline learning procedures with high prediction accuracy.