This repository consists of the source codes that allow one to realize over-the-air computation for federated edge learning by using Adalm Pluto SDRs. The details are given the following paper.
The compiled pluto.frm is also available above.
A Demonstration of Over-the-Air Computation for Federated Edge Learning - IEEE GLOBECOM 2022 Workshops
Abstract: In this study, we propose a general-purpose synchronization method that allows a set of software-defined radios (SDRs) to transmit or receive any in-phase/quadrature data with precise timings while maintaining the baseband processing in the corresponding companion computers. The proposed method relies on the detection of a synchronization waveform in both receive and transmit directions and controlling the direct memory access blocks jointly with the processing system. By implementing this synchronization method on a set of lowcost SDRs, we demonstrate the performance of frequency-shift keying (FSK)-based majority vote (MV), i.e., an over-the-air-computation scheme for federated edge learning, and introduce the corresponding procedures. Our experiment shows that the test accuracy can reach more than 95% for homogeneous and heterogeneous data distributions without using channel state information at the edge devices.
Paper: https://arxiv.org/abs/2209.09954, https://ieeexplore.ieee.org/document/10008778