.png){kind=icon}
.png){kind=icon}
Return
Federated Learning for Secure Multi-UAV Coordination
.png){kind=icon}
PRE
AI
DOI：10.1002/spy2.70134.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Drones (unmanned aerial vehicles [UAVs]) have significantly advanced technology. They are extensively used across various fields, including search and rescue operations, agricultural monitoring, industrial applications, environmental studies, and safety measures. Additionally, drones support search and rescue operations, crisis management, and other critical tasks. Due to their exceptional characteristics and capacity for carrying out significant tasks such as photography, videography, and data collection, UAVs play a crucial role. The primary concern regarding privacy in the context of drones pertains to data security. Conventional UAVs have utilized various machine learning (ML) techniques to address privacy issues; however, the concept of Federated Learning (FL) has emerged as a superior approach. The FL methodology is employed to enhance security, serving as a recognized strategy for maintaining confidentiality in drone operations while minimizing the risk of data breaches. This research proposes a framework for deploying multiple UAVs that integrate the FL concept. Additionally, the proposed methodology incorporates Convolutional Neural Networks (CNN), which are known for their computational efficiency and superior performance compared to alternative neural networks. Secure Aggregation (SA) techniques have also been implemented to safeguard model parameters. Simulation results demonstrate significant improvements in metrics such as F1-score, precision, accuracy, and recall when compared to existing ML techniques. Furthermore, the paper includes a comparative analysis of both conventional ML and FL methodologies.
Keywords:
accuracy
CNN
FL
ML
security
training
UAV
Journal
S
IF:
2.1
Papers:
125
Citations:
717