Standard Propagation Channel Models for MIMO Communication Systems

Table 5

Description of critical aspects of the 3GPP 3D channel model.

S/N.	The aspect of 3D channel model	Brief description
1.	Applicability of the 3D channel model	The height of the eNB in the 3D-UMa is around 25 m higher than the buildings’ heights in the vicinity. In this case, the prevalent propagation technique for indoor and outdoor UEs is built over the roof’s diffractions. Similarly, for 3D-UMi, the eNB is considered to be 10 m below the surrounding buildings. Consequently, the signal intensity obtained at the UE is a combination of signal propagation mechanisms over the rooftop and the surrounding structures. The building density, building heights, and street orientation were considered to derive this model for proper ray-tracing, simulations, and field measurements [199].; For carrier frequencies in the 2-6 GHz range, with up to 100 MHz bandwidth, the 3D channel model is valid. The model has validity in the neighborhood of 1.5 m to 22.5 m for UE heights. The 1.5 m height is essentially the level of a street. As reported in [199, 213, 214], an indoor UE can be linked with a height denoted by , where refers to the floor number, which has a uniform distribution between 1 and , and is the building height in floors. More specifically, this is uniformly distributed between 4 and 8. The height of the outdoor UEs can typically be taken to be 1.5 m. This aspect demonstrates the extension of the current 3GPP and ITU-R specifications, where a street-level UE is normally modeled.
2.	Antenna modeling	For each antenna elements in the 2D channel model, corresponding channel responses are generated in the 3D channel model, among other details. It is worthy to note that 3D channel modeling helps address some significant limitations of the 2D channel model.; The polarization is another parameter worth noting. It is a crucial feature required in the modeling of antennas. Suppose the cross-polarized transmit antenna pair of ±45/45 degrees is considered. In that case, the constant polarization model tends to bring about an equal power split for all UE positions in the vertical and horizontal directions. On the other hand, equal power split at the antenna boresight in both the vertical and horizontal directions is seen via the slanted dipole polarization model. However, the power split ratio depends on the UE position in both azimuths and elevation measurements. It can be inferred that the antenna patterns of the simulated eNBs affect the use of a particular polarization model in any simulation scenario.
3.	LoS probability and path loss modeling	Line of sight likelihood and path loss are primary parameters often considered in MIMO channel modeling. The effect on LoS likelihood and path loss modeling of various UE heights ranging from 1.5 m to 22.5 m for the 3D-UMa and 3D-UMi scenarios has been documented [199] [215]. Simulation of the likelihood of LoS, path loss, and the fast fading channel’s characterization requires the inclusion of the channel model’s building dimensions. In 3GPP, a decision was reached to adopt the stochastic modeling method used in the SCM and WINNER II, independent entirely, on the building/street orientation dimensions. In the development of the 3D-UMa and 3D-UMi, some of the 2D channel model parameters were used. This results in enormous savings in propagation measurement costs and reduces computational complexity for simulations at the machine level [199].
4.	Path loss modeling (LoS and NLoS)	In 3GPP, the LoS path modeling is achieved by applying the 3D distance separating the eNB from the UE. The coefficients for 3D-UMa and 3D-UMi by using LoS path loss equations were specified in the ITU working documents. For the modeling of NLoS path loss, in a 3D-UMa environment, the dominant propagation paths appear to move over rooftops via multiple diffraction, by diffraction occurring at the edge of a building, following the report in [214]. Measured data recorded by 3GPP [214] shows that the path loss increases with the diffraction angle as the UE moves from a high floor towards a low floor. A linear term for the height-gain expressed as is implemented for functional modeling, and is referred to as the gain coefficient.
5.	Fast fading model	For a cellular downlink, fast fading can be represented such that the AoA and the AoD are specified on the UE side and eNB side, respectively. The time variance of wireless channels coordinated by a mix of multipath and UE motion is modeled by fast fading channel coefficients.; Further information on fast fading is available in the 3GPP recommendation document [199, 214], explaining how the radio channels are formed. As shown in Figure 16, the channel realizations followed a hierarchical protocol. It is worth noting that there is no description of the propagation between the first and the last interaction. This method could aid in modeling several interactions with the scattering media. For example, this is an indication that the geometry cannot achieve the delay of multipath. Parameters of arrival and departure have to be exchanged for the uplink case, and presumption is made for the downlink scenario. The channel coefficient generation process from steps 4 to 11 for the LoS O2I scenario is the same for the NLoS scenario illustrated in Figure 16.; In general, the channel generation equation is correctly updated in the last step of the channel generation coefficient to account for the ZoAs and ZoDs. A structure for studying 3D channel model extensions and features is given in [216], and preliminary results on 3D channel modeling are presented in [217]. Further to this, the reports in [218, 219] also include implementing, validating, and deploying the open-source simulation software model.