A WaveGAN Approach for mmWave-Based FANET Topology Optimization.

Area of Science:

• Computer Science
• Electrical Engineering
• Network Engineering

Background:

• Flying Ad hoc Networks (FANETs) integrated with millimeter Wave (mmWave) technology offer high data transmission for data-intensive applications.
• Effective mmWave communication in dynamic FANETs requires precise antenna alignment between Unmanned Aerial Vehicles (UAVs).
• Optimizing network topology and antenna alignment is crucial for robust FANET performance.

Purpose of the Study:

• To propose a novel approach for rapid optimization of FANET topology and antenna alignment for mmWave communication.
• To maximize network throughput by selecting optimal communication paths with superior channel conditions.
• To address the challenge of precise antenna alignment in dynamic UAV-based networks.

Main Methods:

• A Generative Adversarial Network (GAN)-based approach, named WaveGAN, is proposed for FANET topology optimization.
• The WaveGAN model learns to generate optimized network topologies from a supervised dataset.
• A beam search algorithm refines the generated topologies to meet mmWave-based FANET structural requirements.

Main Results:

• Simulation results demonstrate the proposed WaveGAN approach's ability to rapidly determine optimized FANET topologies.
• The method achieves a very small optimality gap across different network sizes.
• WaveGAN effectively identifies communication paths with the best channel conditions for maximizing throughput.

Conclusions:

• The proposed WaveGAN approach provides an efficient solution for optimizing mmWave-enabled FANETs.
• This method enables quick determination of network topologies, enhancing overall network performance and throughput.
• WaveGAN is a promising technique for managing dynamic antenna alignment and topology in UAV communication networks.