FLSTRA: Federated Learning in Stratosphere

TL;DR

This paper introduces FLSTRA, a federated learning system using high altitude platform stations to improve convergence and reduce delays in large-scale terrestrial networks, with optimized client selection and resource management.

Contribution

It presents a novel FL system in the stratosphere, along with algorithms for client selection and resource allocation to optimize delay and accuracy trade-offs.

Findings

FLSTRA outperforms terrestrial benchmarks in delay and accuracy.

The proposed algorithms effectively balance delay and convergence rate.

Simulation results validate the system's efficiency in large-scale FL scenarios.

Abstract

We propose a federated learning (FL) in stratosphere (FLSTRA) system, where a high altitude platform station (HAPS) facilitates a large number of terrestrial clients to collaboratively learn a global model without sharing the training data. FLSTRA overcomes the challenges faced by FL in terrestrial networks, such as slow convergence and high communication delay due to limited client participation and multi-hop communications. HAPS leverages its altitude and size to allow the participation of more clients with line-of-sight (LOS) links and the placement of a powerful server. However, handling many clients at once introduces computing and transmission delays. Thus, we aim to obtain a delay-accuracy trade-off for FLSTRA. Specifically, we first develop a joint client selection and resource allocation algorithm for uplink and downlink to minimize the FL delay subject to the energy and quality-of-service (QoS) constraints. Second, we propose a communication and computation resource-aware (CCRA-FL) algorithm to achieve the target FL accuracy while deriving an upper bound for its convergence rate. The formulated problem is non-convex; thus, we propose an iterative algorithm to solve it. Simulation results demonstrate the effectiveness of the proposed FLSTRA system, compared to terrestrial benchmarks, in terms of FL delay and accuracy.