Table 6.

Some efforts addressed in A2G modeling.

Cite	Approach		Scenario			Method		Aim	Contributions
	ML	GM	UMa	UMi	RMa	St	N-St
[95]	X		X	X	X	X		PL and Delay Spread prediction for mmWave channels.	Low computational complexity. Full feature selection scheme. Frequency/scene-based transfer learning model.
[96]	X		X	X	X	X		PL and Shadowing effects analysis in 3D-LOS/NLOS Channel.	Unsupervised learning clustering technique to derive a 3D temporary channel.
[97]	X		X			X		PL empirical prediction with environmental parameters.	Location-based method by using 3D-GPS coordinates. Learning phase includes atmospheric conditions.
[98]	X		X	X	X		X	Collaborative algorithm to solve communication overload by achieving 1.5x throughput.	Optimization of Multi-UAV user deployment based in modified K- means distribution and POO.
[99]		X	X			X		3D non-stationary geometry-based stochastic channel model for A2G.	3D arbitrary trajectories. 3D antenna arrays for 5G. Computational Methods for time-variant channel parameters.
[100]		X	X			X		A MIMO wideband truncated ellipsoidal-shaped method with scatterer consideration.	Statistical derivation of space-time-correlation function and Dopler power spectrum density.
[101]		X		X		X		Geometrical model for UAV flight’s Multi-Path Components evolution.	Geometrical parametrization for the main MPCs. Simulation under non-intuitive effects of propagation.
[102]		X	X	X	X		X	Spatial-temporal correlation in function of UAV’s hover radius, flight altitude, and elevation angle.	Numerical approach of PL, Multi-shadow fading, Doppler shift, and channel correlation. Fixed-Wings UAV-BS Model.