A Survey on Machine-Learning Techniques for UAV-Based Communications.

Area of Science:

• Wireless Communication Networks
• Artificial Intelligence
• Network Engineering

Background:

• Unmanned aerial vehicles (UAVs) are crucial for future wireless networks, promising improved coverage and efficiency over ground-based systems.
• Integrating UAVs introduces unique network challenges that require advanced solutions.
• Machine learning (ML) is identified as a key technology to address these challenges.

Purpose of the Study:

• To provide a comprehensive survey of research applying ML techniques to UAV-based communication systems.
• To identify and categorize the various design and functional aspects of UAV communications improved by ML.
• To highlight the potential of ML in optimizing UAV network performance and reliability.

Main Methods:

• Systematic literature review of research papers employing ML in UAV communications.
• Categorization of ML applications based on functional aspects: channel modeling, resource management, positioning, and security.
• Analysis of existing studies to identify trends and challenges in ML for UAV networks.

Main Results:

• ML techniques are effectively applied to enhance channel modeling accuracy in UAV networks.
• ML optimizes resource management strategies, including spectrum and power allocation for UAVs.
• ML contributes to improved positioning accuracy and enhanced security protocols for UAV communication systems.

Conclusions:

• ML is a vital enabler for overcoming the complexities of integrating UAVs into next-generation wireless networks.
• The surveyed ML applications demonstrate significant potential for improving coverage, spectral efficiency, and overall network performance.
• Further research in ML for UAV communications is essential to fully realize their potential in future networks.