• Federated learning framework enables mobile transmission scheduling while preserving device privacy. • Three energy‑constrained problems tackled: minimize transmission time, fixed‑rate scheduling, maximize data upload. • Scheduler interacts only with digital twins, generating global fractional solutions without revealing private data. • Dependent rounding converts fractional solutions into practical channel transmission schedules for physical devices. • Experiments show consistent makespan reductions, negligible bandwidth/energy violations, millisecond runtime on edge servers. • First framework to share channels across digital twins without exposing sensitive information.
Article Summaries:
- A new study introduces a federated‑learning framework that lets mobile devices schedule transmissions while keeping sensitive data-such as mobility patterns and channel conditions-private. By interacting only with each device’s Digital Twin, the scheduler solves three energy‑constrained problems: minimizing total transmission time under fixed‑power or fixed‑rate schemes, and maximizing data uploaded in a fixed period. The method first generates fractional solutions via federated optimization, then applies dependent rounding to produce a concrete channel schedule. Experiments on typical edge‑server hardware show consistent reductions in makespan, negligible bandwidth or energy violations, and millisecond‑scale end‑to‑end runtime, marking the first privacy‑preserving channel‑sharing solution across Digital Twins.
Sources: