Elevation- and Tilt-Aware Shadow Fading Correlation Modeling for UAV Communications

TL;DR

This paper introduces an elevation- and tilt-aware shadow fading correlation model for UAV communications, demonstrating improved accuracy in signal strength prediction by incorporating UAV orientation and elevation effects.

Contribution

It proposes a novel correlation model that accounts for UAV pitch and elevation, enhancing the accuracy of shadow fading characterization in UAV communication channels.

Findings

A 10-degree tilt-angle separation reduces SF correlation by up to 15%.

A 20-degree elevation-angle separation reduces SF correlation by up to 40%.

The new model improves median RMSE by approximately 1.5 dB in signal prediction.

Abstract

Future wireless networks demand a more accurate understanding of channel behavior to enable efficient communication with reduced interference. Uncrewed Aerial Vehicles (UAVs) are poised to play an integral role in these networks, offering versatile applications and flexible deployment options. However, accurately characterizing the shadow fading (SF) behavior in UAV communications remains a challenge. Traditional SF correlation models rely on spatial distance and neglect the UAV's 3D orientation and elevation angle. Yet even slight variations in pitch angle (5 to 10 degrees) can significantly affect the signal strength observed by a UAV. In this study, we investigate the impact of UAV pitch and elevation geometry on SF and propose an elevation- and tilt-aware spatial correlation model. We use a real-world fixed-altitude UAV measurement dataset collected in a rural environment at 3.32 GHz with a 125 kHz bandwidth. Results show that a 10-degree tilt-angle separation and a 20-degree elevation-angle separation can reduce the SF correlation by up to 15% and 40%, respectively. In addition, integrating the proposed correlation model into the ordinary Kriging (OK) framework for signal strength prediction yields an approximate 1.5 dB improvement in median RMSE relative to the traditional correlation model that ignores UAV orientation and elevation.