Page 59 - ITU Journal: Volume 2, No. 1 - Special issue - Propagation modelling for advanced future radio systems - Challenges for a congested radio spectrum

ITU Journal: ICT Discoveries, Vol. 2(1), December 2019

20 ⋅10 −2
RTS 1
RTS (VC)
DACM
15 0.8 RTS (Melb)
Simulated results
DACM
( m ) 10 ( m ) 0.6
5 0.4
0.2
0
0
0 0.2 0.4 0.6 0.8 1 −180 −135 −90 −45 0 45 90 135 180

m (degree)
m
(a) ℳ = 2, = 2 GHz
(a) Veri ication on m , = 2 GHz
40
RTS 20
DACM RTS (VC)
30 RTS (Melb)
15 Simulated results
DACM
( m ) 20 ( m ) 10
10
5
0
0 0.2 0.4 0.6 0.8 1 0 0 0.2 0.4 0.6 0.8 1
m

m
(b) ℳ = 3, = 28 GHz
(b) Veri ication on m , ℳ = 3, = 28 GHz
Fig. 7 – Examples of the curve itting for ( m ).
Fig. 8 – Examples of the channel model veri ication on m and m .
tion. For the latter method, we simulate the channel fol- ability of the cluster number if ℳ ≤ 3, but these clus-
lowing Algorithm 1. The receiver number is also and ters are more concentrated at the LoS direction, and the
the simulation follows the steps in section 7. A compari- power distribution is more uneven. Future works can be
son of these verifying cases, RTS (VC), RTS (Melb), simu- extended from the 2D model to a 3D air-to-ground chan-
lated model and the channel model can be found in Fig. 8. nel model, which brings more interest to the elevation an-
In both of the sub- igures, the scattered dots are the veri- gle. Also, the clustering algorithm and the modeling of the
ication results, and the colors, red, yellow and green rep- root mean square delay will be studied to make the DACM
resent the RTS in a virtual city, the RTS in Melbourne CBD more completed.
and the numerical results respectively. The solid blue line
is the proposed directional channel model. The carrier REFERENCES
frequencies are 2 GHz for Fig. 8(a) and 28 GHz for Fig.
8(b). The three veri ication cases match the model very [1] NGMN Alliance, “NGMN 5G white paper,” accessed
well in both igures. Actually, we have veri ied every case, 2019-07-17. [Online]. Available: https://www.
including , conditioned by each ℳ at two frequen- ngmn.org/work-programme/5g-white-paper.html
m
m
cies. All the cases have shown a high level of matching.
However, only two examples are depicted in this paper. [2] Z. Li, K. Magowe, A. Giorgetti, and S. Kandeepan,
“Blind localization of primary users with sectorial
antennas,” in IEEE International Conference on Com-
9. CONCLUSION
munications Workshops (ICC Workshops), Kansas
In this paper, we presented a statistical clustering DACM City, MO, USA, May 2018, pp. 1–6.
for a general urban scenario. It can bene it the directional
wireless network design, optimization and utilization in [3] R. T. Rakesh, G. Das, and D. Sen, “An analytical model
substance since the channel features can be estimated by for millimeter wave outdoor directional non-line-of-
merely a few distance-independent probability distribu- sight channels,” in IEEE International Conference on
tion parameters. The model is proposed in three key as- Communications (ICC), Paris, France, May 2017, pp.
pects, the number of clusters, the cluster center distribu- 1–6.
tion and the cluster power ratio distribution at two car-
rier frequencies, 2 GHz and 28 GHz. We have concluded [4] M. Sha i, J. Zhang, H. Tataria, A. F. Molisch, S. Sun, T. S.
that for both frequencies, these three parameters share Rappaport, F. Tufvesson, S. Wu, and K. Kitao, “Mi-
similar features but with notable differences. Compared crowave vs. millimeter-wave propagation channels:
with 2 GHz, 28 GHz tends to bring relatively equal prob- Key differences and impact on 5G cellular systems,”

© International Telecommunication Union, 2019 43