Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

EPFL-REMNet: Efficient Personalized Federated Digital Twin Towards 6G Heterogeneous Radio Environment

Peide Li, Liu Cao, Lyutianyang Zhang, Dongyu Wei, Ye Hu, Qipeng Xie

TL;DR

This work tackles the challenge of building a high-fidelity Radio Environment Map (REM) digital twin for 6G under strong data heterogeneity and limited uplink bandwidth. It introduces EPFL-REMNet, a co-designed framework that couples a shared backbone with client-specific heads and a multi-stage communication compression pipeline to achieve superior REM fidelity while drastically reducing uplink traffic. Across light, medium, and heavy Non-IID scenarios, EPFL-REMNet outperforms standard FedAvg and recent baselines, demonstrates Pareto-optimal fidelity-communication trade-offs, and shows robust fairness across base stations and long-tail clients. The findings highlight the value of jointly optimizing personalization and communication efficiency to enable scalable, reliable digital twins in heterogeneous wireless environments.

Abstract

Radio Environment Map (REM) is transitioning from 5G homogeneous environments to B5G/6G heterogeneous landscapes. However, standard Federated Learning (FL), a natural fit for this distributed task, struggles with performance degradation in accuracy and communication efficiency under the non-independent and identically distributed (Non-IID) data conditions inherent to these new environments. This paper proposes EPFL-REMNet, an efficient personalized federated framework for constructing a high-fidelity digital twin of the 6G heterogeneous radio environment. The proposed EPFL-REMNet employs a"shared backbone + lightweight personalized head" model, where only the compressed shared backbone is transmitted between the server and clients, while each client's personalized head is maintained locally. We tested EPFL-REMNet by constructing three distinct Non-IID scenarios (light, medium, and heavy) based on radio environment complexity, with data geographically partitioned across 90 clients. Experimental results demonstrate that EPFL-REMNet simultaneously achieves higher digital twin fidelity (accuracy) and lower uplink overhead across all Non-IID settings compared to standard FedAvg and recent state-of-the-art methods. Particularly, it significantly reduces performance disparities across datasets and improves local map accuracy for long-tail clients, enhancing the overall integrity of digital twin.

EPFL-REMNet: Efficient Personalized Federated Digital Twin Towards 6G Heterogeneous Radio Environment

TL;DR

This work tackles the challenge of building a high-fidelity Radio Environment Map (REM) digital twin for 6G under strong data heterogeneity and limited uplink bandwidth. It introduces EPFL-REMNet, a co-designed framework that couples a shared backbone with client-specific heads and a multi-stage communication compression pipeline to achieve superior REM fidelity while drastically reducing uplink traffic. Across light, medium, and heavy Non-IID scenarios, EPFL-REMNet outperforms standard FedAvg and recent baselines, demonstrates Pareto-optimal fidelity-communication trade-offs, and shows robust fairness across base stations and long-tail clients. The findings highlight the value of jointly optimizing personalization and communication efficiency to enable scalable, reliable digital twins in heterogeneous wireless environments.

Abstract

Radio Environment Map (REM) is transitioning from 5G homogeneous environments to B5G/6G heterogeneous landscapes. However, standard Federated Learning (FL), a natural fit for this distributed task, struggles with performance degradation in accuracy and communication efficiency under the non-independent and identically distributed (Non-IID) data conditions inherent to these new environments. This paper proposes EPFL-REMNet, an efficient personalized federated framework for constructing a high-fidelity digital twin of the 6G heterogeneous radio environment. The proposed EPFL-REMNet employs a"shared backbone + lightweight personalized head" model, where only the compressed shared backbone is transmitted between the server and clients, while each client's personalized head is maintained locally. We tested EPFL-REMNet by constructing three distinct Non-IID scenarios (light, medium, and heavy) based on radio environment complexity, with data geographically partitioned across 90 clients. Experimental results demonstrate that EPFL-REMNet simultaneously achieves higher digital twin fidelity (accuracy) and lower uplink overhead across all Non-IID settings compared to standard FedAvg and recent state-of-the-art methods. Particularly, it significantly reduces performance disparities across datasets and improves local map accuracy for long-tail clients, enhancing the overall integrity of digital twin.

Paper Structure

This paper contains 13 sections, 5 equations, 3 figures, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. The EPFL-REMNet Framework
  3. System Architecture
  4. Problem Formulation
  5. Federated Optimization and Communication Method
  6. Experimental Evaluation
  7. Experimental Setup
  8. Main Results: Performance Across Heterogeneity Levels
  9. Convergence and Communication Dynamics
  10. Accuracy-Communication Pareto Frontier Analysis
  11. Fairness and Robustness Analysis
  12. Ablation Study: Deconstructing EPFL-REMNet
  13. Conclusion

Figures (3)

  • Figure 1: EPFL-REMNet System Architecture.
  • Figure 2: Convergence curves and cumulative communication volume curves of each method in three-level Non-IID scenarios.
  • Figure 3: Precision-communication Pareto frontier graph of EPFL-REMNet in three-level Non-IID scenarios.