Efficient Deployment with Throughput Maximization for UAVs Communication Networks.

Area of Science:

• Wireless communication networks
• Network optimization
• Artificial intelligence in networking

Background:

• UAVs offer flexible deployment for enhancing ground network coverage and capacity.
• Congestion in ground networks leads to performance degradation, including increased delay and packet loss.
• Dynamic optimization of UAV trajectories is crucial for efficient network performance.

Purpose of the Study:

• To develop and validate a throughput maximization approach for UAV-assisted ground networks.
• To minimize network delay and packet loss through intelligent UAV trajectory optimization.
• To enhance network performance by reinforcing congested nodes and transmission channels.

Main Methods:

• Characterizing network elements (nodes, links, topology) using metrics like delay, loss, throughput, and distance.
• Employing a position-aware graph neural network (GNN) for network characterization, prediction, and dynamic UAV trajectory enhancement.
• Validating the proposed approach against Optimized Link State Routing (OLSR) driven networks.

Main Results:

• The proposed approach significantly outperforms classical methods in terms of throughput and packet delivery ratio.
• Demonstrated notable reductions in network delay and packet loss compared to existing solutions.
• Performance analysis against Software-Defined UAVs (U-S) and UAVs as Base Stations (U-B) confirms consistent throughput gains and minimized latency.

Conclusions:

• The developed UAV trajectory optimization strategy effectively maximizes throughput in assisted ground networks.
• The position-aware GNN approach provides a robust and scalable solution for dynamic network enhancement.
• This method offers significant improvements in key performance indicators, making it a promising advancement for wireless network efficiency.