UAV Communications with Machine Learning: Challenges, Applications and Open Issues

Unmanned aerial vehicles (UAV) have recently proved their ability to afford reliable and cost-effective solutions for many real-world scenarios. The autonomy, mobility and flexibility nature along with communications interoperability have made UAV able to provide a large variety of services. Awareness of context changes and adaptability to current services requirements are the key to UAVs' deployment success. For this reason, machine learning (ML) has been widely used in overcoming the challenges that UAV faces in mobility, communication and resources management. This paper will mainly focus on the proposed UAV-centric ML solutions and their satisfaction with network requirements taking into consideration UAVs' roles, collaboration, cooperation and network changing contexts. Solutions proposed in air to air, air to ground and ground to air communications as well as UAVs-enabled mobile edge computing are investigated for possible future insights. Future works will indeed highlight the need of UAVs' cooperation in the emergent 5G/6G networks.

Automated summary

This paper focuses on the application of machine learning in UAVs and explores the challenges they face in mobility, communication, and resource management. Possible future research directions include investigating solutions for air to air, air to ground, and ground to air communications, as well as mobile edge computing, and highlighting the importance of UAV cooperation in 5G/6G networks.