A Deep-Learning Approach to a Volumetric Radio Environment Map Construction for UAV-Assisted Networks

Table 3

WinProp simulation and deep-learning model training parameters.

Parameters	Values
WinProp-related parameters
Areal resolution,	1 m
Height resolution, ,	3 m
Network	LTE—advanced
Bandwidth	10 MHz
Center frequency,	2600 MHz
GBS/ABS transmission power	43 dBm/33 dBm
GBS/ABS height	35 m/150 m
GUE height	1.5 m–45 m
AUE height	75 m–90 m
Deep-learning-related parameters
Adam optimizer learning rate	0.002
Training/test/validation data split (for the 80 transmitters in each scenario)	80/10/10%
Epochs	30–80
Batch size	2, 4