Direction Estimation in 3D Outdoor Air-Air Wireless Channels through Machine Learning.

Area of Science:

• Electrical Engineering
• Computer Science
• Aerospace Engineering

Background:

• Unmanned Aerial Vehicles (UAVs) require robust 3D communication, but wireless transmission power dissipation limits range.
• Traditional Multiple-Input Multiple-Output (MIMO) systems require significant power resources on UAVs.
• Phased arrays offer directionality but involve complex beam sweeping and search delays, especially with dynamic UAV mobility.

Purpose of the Study:

• To develop an efficient machine learning-based method for estimating the direction of a transmitting node in aerial-to-aerial (A2A) links.
• To reduce search time and complexity in UAV communication by enabling direct directional beam selection.
• To validate the proposed method using real-world drone-to-drone measurements.

Main Methods:

• Analysis of multi-antenna channels between two UAVs in A2A links.
• Development of a machine learning model utilizing channel estimates from a 2x2 MIMO system (4 antennas).
• In-field validation through drone-to-drone measurements.

Main Results:

• The proposed machine learning method accurately estimates the transmitter's direction in A2A links with 86% accuracy.
• The method effectively reduces the need for beam sweeping in UAV communication.
• Demonstrated deployability for UAV-based massive MIMO systems.

Conclusions:

• The developed machine learning approach significantly improves directional communication efficiency for UAVs.
• This method offers a practical solution for optimizing wireless links in dynamic aerial networks.
• Accurate direction estimation paves the way for enhanced UAV communication performance without search delays.