Artificial Intelligence for UAV-Enabled Wireless Networks: A Survey

Mohamed-Amine Lahmeri,Mustafa A Kishk,Mohamed-Slim Alouini

doi:10.1109/ojcoms.2021.3075201

Abstract

Unmanned aerial vehicles (UAVs) are considered as one of the promising technologies for the next-generation wireless communication networks. Their mobility and their ability to establish line of sight (LOS) links with the users made them key solutions for many potential applications. In the same vein, artificial intelligence (AI) is growing rapidly nowadays and has been very successful, particularly due to the massive amount of the available data. As a result, a significant part of the research community has started to integrate intelligence at the core of UAVs networks by applying AI algorithms in solving several problems in relation to drones. In this article, we provide a comprehensive overview of some potential applications of AI in UAV-based networks. We also highlight the limits of the existing works and outline some potential future applications of AI for UAV networks.

Highlights

U NMANNED aerial vehicles, known as Unmanned aerial vehicles (UAVs), attracted a lot of research interest in the last decades due to many inherent attributes such as their mobility, their easy deployment, and their ability to establish line of sight (LOS) links with the users [1]–[4]
We study the implementation of intelligence at the edge of UAV networks by reporting some of the works done in federated learning (FL) for UAV-based networks
As we proposed FL as a solution to running machine learning (ML) on the edge at the ML discussion section, we should mention that the massive exchange of updates across the network will result in huge communication loads in the training phase, especially for neural networks, which will induce a scalability problem for FL

Summary

INTRODUCTION

U NMANNED aerial vehicles, known as UAVs, attracted a lot of research interest in the last decades due to many inherent attributes such as their mobility, their easy deployment, and their ability to establish line of sight (LOS) links with the users [1]–[4]. An optimization problem for joint transmission and content caching is formulated and solved via the use of deep learning techniques combining liquid state machine spiking neural networks and echo state networks In another context, authors in [44] provide a UAV VR simulation platform used to assess the performance of several DL-based solutions for UAVs such as autonomous path planning and obstacle avoidance. We noticed that there is a tendency to use supervised learning algorithms in solving UAV-based problems For this reason, we believe that unsupervised learning techniques, such as clustering and dimensionality reduction, could be further explored in the future for drone-related problems. Once the local update wkt+1 is received by the base station, it improves the global model and removes these updates because they are no longer needed

FEDERATED LEARNING FOR UAVS

OPEN ISSUES SUMMARY

Findings

CONCLUSION