http://www.tu-ilmenau.de

Logo TU Ilmenau


Ansprechpartner

Prof. Dr.-Ing. habil. Andreas Mitschele-Thiel

Fachgebietsleiter

Telefon +49 3677 / 69 2819

E-Mail senden


INHALTE

Veröffentlichungen

Veröffentlichung

Experimental Validation of Air-to-Ground Propagation Models for Low-Altitude Platforms

Autoren:
M.Sc. Oleksandr Andryeyev
Umut Onus
M. Sc. Victor Casas Melo
Prof. Dr.-Ing. habil. Andreas Mitschele-Thiel
Typ:
Konferenz
Status:
akzeptiert
Veröffentlichungsdatum
31.05.2019
Abstract:
As traffic demands are steadily growing, demanding for a densification of the mobile network infrastructure, the use of aerial base stations is an important option in order to serve areas with temporarily high data demand.
Aerial base station have the advantage of being able to follow the traffic demand and to supplement the ground infrastructure with extra capacity.
Due to their elevation and proximity to the user, path loss is typically smaller in comparison to ground infrastructure which enables high capacity services.
However, up to now there was a small progress in experimental validation of Air-to-Ground (A2G) models for low-altitude aerial platforms.

The main objective of this work is the selection and validation of the propagation model for computer simulation.
We use the micro-aerial vehicle with software-defined radio onboard to collect the signal strength measurements for two different user groups.
We select the most appropriate model by applying regression analysis on the collected data and compare state-of-the-art A2G models using Root Mean Square Error (RMSE) and R-squared metrics.

We show that the best fit is presented by the analytical model with the lowest RMSE (8.29 dBm) and the highest R-squared value (0.67), which shows accurate model prediction properties.
We also compare two different user groups located in building courtyards and outdoor areas and show that the first user group begins to observe non-line-of-sight effects at higher elevation angles (< 46 degrees) than the second user group (< 30 degrees), which proves that the model can successfully distinguish between different environments.
Bibtex