Next-Gen Internet of Drones: Federated Learning and Digital Twin Synergy for Energy-Efficient Task Allocation and Seamless Service Migration

The computing-intensive tasks generated by Internet of Things devices cannot be handled alone by themselves due to limitations in battery and processing power. An appropriate approach to this problem is the Internet of Drones (IoDs) with edge computing capabilities, which can offload the created tas...

Full description

Saved in:
Bibliographic Details
Main Authors: Ahmad Arsalan, Tariq Umer, Rana Asif Rehman, Byung-Seo Kim
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10950141/
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The computing-intensive tasks generated by Internet of Things devices cannot be handled alone by themselves due to limitations in battery and processing power. An appropriate approach to this problem is the Internet of Drones (IoDs) with edge computing capabilities, which can offload the created tasks from IoT devices to IoDs. To improve sustainability by maximizing energy efficiency, minimizing duplicate service migrations, and guaranteeing dynamic task offloading in UAV-supported IoD networks, this paper proposes a Federated Digital aided Internet of Drones (FD-IoD) architecture. The proposed framework guarantees the long-term viability of IoD-based edge networks by combining digital twin technology with federated deep reinforcement learning. The FD-IoD framework integrates energy harvesting algorithms and optimizes mobility-aware resource management to extend drone lifespan and reduce unnecessary computational overheads. To adjust to various IoT environments, the framework uses a dual-layer optimization approach that combines local agent learning with global decision-making via digital twin. The framework outperforms current benchmarks by up to 40% in energy efficiency, lower service migration rates, and faster task completion rates, as shown by extensive simulations. Additionally, the proposed framework guarantees decreased latency, efficient resource use, and queue stability even in heavy demand.
ISSN:2169-3536