2019 IEEE 11th International Conference on Communication Software and Networks

A Novel Generalized Modeling Mechanism for VHF and UHF Mobile-to-Mobile

Channels

Haiyu Ren, Longfei Gao Xiaoxin He

State Key Laboratory of Integrated Information System State Key Laboratory of Integrated Information System
Technology, Technology,
Institute of Software, Chinese Academy of Sciences Institute of Software, Chinese Academy of Sciences
Beijing, China Beijing, China
e-mail: [email protected]. [email protected], e-mail: [email protected]

Abstract—An important link in the analysis of communication J et al. proposed a path loss model based on VHF and UHF
systems is the communication channel modeling, especially in bands using onboard transceivers. In [4], a delay extended
the two frequency bands of VHF and UHF. In this paper, distributed model for VHF and UHF bands is investigated. A
based on the existing measured data, we have established a channel modeling method under high speed railway
channel model that can be widely used in terrestrial scenarios environment is proposed in [5]. Zhang et al. [6] studied a
in both VHF and UHF frequency bands. We set up this model channel modeling method in a tunnel environment, and Tang
to solve the problem that the channel model under a single et al. [7] established a channel model in a vehicle-vehicle
scenario cannot meet the actual environmental requirements. communication environment. We refer the reader to [8], [9],
Meanwhile, we improved and modified the Okumura-Hata
[10] for a channel modeling method in a low-altitude
model to be suitable for scenarios where the BS antenna height
is less than 30 meters. In our analysis, we mainly discuss the
environment.
large-scale fading and small-scale fading under the four The channel models in these studies are close to the
scenarios of cities, mountains, jungles, and low altitude. The actual environment, they also provide a large amount of
results show that the channel models under different scenarios measured data, but the conditions considered are relatively
greatly increase the universality and authenticity. The simple. Fischer J et al.[3] only discusses the channel
modeling results and experimental data are well-fitted and can modeling based on large scale fading; Fischer J et al.[4] only
be widely applied in engineering practice. discusses the channel modeling based on small scale fading;
[5]-[10] only consider channel modeling under certain
Keywords-channel modeling; VHF; UHF; various scenarios scenarios. Moreover, the above-mentioned established
channel model does not take into account the actual
I. INTRODUCTION influence of the height of the transceiver antenna, or only
applies to a certain kind of transceiver antenna height.
With the rapid development of wireless communication The main contributions of this paper are: First, this paper
services, modern military communications have developed systematically discusses the influencing factors of large-scale
for more maneuverable network, higher speed transmission fading and small-scale fading, making the model more
and far distance transmission. Although channel modeling universal and real. Meanwhile, we improved and modified
has been studied more and more widely, in 5G the Okumura-Hata model to be suitable for scenarios where
communication [1], wireless channel modeling is still very the BS antenna height is less than 30 meters. Secondly,
important. Wireless ultra-shortwave communication system under the support of the existing measured data, this paper
has gradually become the focus of research in this field by its elaborates the channel modeling methods for different
excellent communication quality and high data transmission heights of transceiver antenna in urban, mountainous, jungle,
rate. The ultra-short wave [2] includes two frequency bands and low-altitude scenarios. It supports not only theoretical
of VHF and UHF, commonly used frequencies are in the formulas but also the actual fitting of measured data and
range of 30 to 900 MHz, and a large number of civilian and modeling results, has a very high engineering value. Third,
military radios operate in this frequency range. At present, although only four scenarios are discussed in this paper, all
many countries have conducted extensive research on ultra- kinds of environments related to land can be nested using the
shortwave technology. For example, the United States, channel model of this paper.
France, and the United Kingdom have developed various The rest of this paper is organized as follows. In the
models of ultra-shortwave radio stations; domestically, ultra- Section II, the system model and related descriptions will be
shortwave technology has also been put into practical use to introduced; general problems such as large-scale fading and
make troops command more mobility, diversity, real-time. small-scale fading will be solved in the Section III. In the
Therefore, modeling the ultrashort wave wireless channel is Section IV, we will analyze the modeling results under the
very critical. four scenarios, followed by concluding remarks in Section V.
Presently, there are many literatures on the channel
modeling methods for ultrashort-wave bands. In [3], Fischer

978-1-7281-2184-0/19/$31.00 ©2019 IEEE 215

 II. SYSTEM MODEL This model provides a wide range of data and is widely
Most of the frequency bands of 30MHz to 512MHz applied to VHF and UHF bands.
belong to the ultra-shortwave band. As the main component The model is suitable for point-to-point and broadcast
of the wireless communication system, the frequency band communications, covering 1-10 meters of mobile personnel
has the advantages of high communication quality, large and vehicle antenna height, BS antenna height of 30-200
communication capacity, anti-interference and stable meters, and low altitude flight distance of 1-10 kilometers.
communication performance. Okumura-Hata is mainly modeled by several parameters
Generally, in the 30MHz to 512MHz channel modeling, such as path loss, BS antenna height, mobile station antenna
the effect of physical layer factors is realized by empirical height, transmission frequency, antenna height correction
model formula, and the accuracy of the simulation is limited factor, and distance between base station and mobile station.
to the empirical model formula used. However, these The path loss formula under the urban environment is as
empirical model formulas are only based on the empirical follows:
statistics or theoretical calculation when considering the
influence of environmental factors. It is impossible to LU =69.55+26.16lgf-13.82lghB-CH+(44.9-6.55lghB)lgd  
accurately predict the actual wireless channel transmission
characteristics without calculating the specific scenario. The For small or medium-sized cities:
main content of our research is to set up a wireless channel
model from 30MHz to 512MHz with different scenarios. CH = 0.8+(1.1lgf-0.7)hM-1.56lgf  
Low attitude
For large cities:

­
°8.29[lg(1.54hM )]  1.1 if 150 d f d 200   
2

CH ®
°̄3.2[lg(11.75hM )]  4.97 if 200 d f d 1500
2

Mountain
Among them, LU is the path loss for the urban area, the
unit is dB, hB is the height of the base station antenna, the
unit is m, hM is the height of the mobile station antenna, the
unit is m, f is the transmission frequency, the unit is MHz, CH
City
is the height correction factor of the antenna, d is the distance
between the base station and the mobile station, the unit is
km.
Jungle 2) Edge diffraction
Edge diffraction is a useful model for calculating the
Figure 1. System model.
propagation loss of electric waves in mountain
communications.
As shown in Figure 1, this paper focuses on four scenes There are many types of propagation modes for the
in cities, mountains, jungles and low altitude. In the four mountain diffraction, including unimodal diffraction and
scenarios, the possible wireless communication devices may multimodal diffraction.
be both base stations (BS), or vehicle antennas, aircraft and The unimodal diffraction loss is calculated by the
mobile personnel. Based on this system model, we will following formula [11]:
discuss large- scale fading and small-scale fading which are
crucial to channel modeling in the next part. A 6.9  20lg v  0.1  1  v  0.1 (v ! 0.7)   
2

III. LARGE-SCALE FADING AND SMALL-SCALE FADING

In this section, we first discuss the influencing factors of In this formula, A is the diffraction attenuation (dB), and
large-scale fading and related application models. Next, we is called Fresnel Kirchhoff loss, ν is the edge diffraction
discuss small-scale fading and related measured data. Finally, parameter.
we briefly introduce the basic knowledge of Rayleigh
distribution and Rice distribution. v 2h / F1   
A. Large-Scale Fading
The main factors affecting the large -scale fading include In the formula, h is the height of the obstacle vertex to
path loss and shadow fading. The more comprehensive the receiving point and the connection point of the receiving
models for path loss include the Okumura-Hata model and point. F1 is the first Fresnel radius of the obstacle point.
the blade peak model.
1) The Okumura-Hata model O d1d 2   
F1
d

216
 In the formula, λ is the wavelength of work, d1 and d2 are 1 ª <2 º
the distance from the obstacles to the sending and receiving p < dB exp «  dB
2 »
  
points respectively, and d is the length of the lines. The 2SV< dB ¬« 2V< dB »¼
parameter diagram is shown in Figure 2.
The shadow fading ΨdB obeys the Gauss distribution
whose mean is zero and variance is σΨdB.
B. Small-Scale Fading
d1 d2 In mobile communications, since obstacles block the data
h
path, the emitted electromagnetic waves usually cannot reach
the receiving antenna directly. In fact, the received
T a1 R electromagnetic waves are superimposed by waves from
d different directions caused by reflection, diffraction and
scattering caused by buildings, trees and other obstacles.
Figure 2. Schematic diagram of signal wave propagtion parameters.
This phenomenon is called multipath propagation. Due to
The method of main obstacle method to calculate the multipath propagation, the received signal is superimposed
path loss of multi peak diffraction is to first clear the primary by an infinite number of transmitted signals after attenuation,
obstacle on the line, the obstacle with minimum delay and phase shift. In addition to multipath propagation,
normalization clearance or the minimum Finel order is the the Doppler effect also has a negative impact on the
primary obstacle, and the algorithm is calculated according transmission characteristics of mobile communications. Due
to the algorithm of the single peak diffraction loss, and then to the motion of the mobile unit, the Doppler effect will
the minimum normalization is found between the transmitter cause a frequency shift of each incoming wave.
and the primary obstacle. For the two-stage main obstacle, 1) Delay extension
the corresponding diffraction loss is the diffraction loss Because the multipath arrival time is different, the signal
caused by the two-stage main obstacle between the emitter received by the receiving end includes not only the signal
and the primary obstacle. The calculation method is also a that is desired to be received, but also various delay signals.
single peak diffraction method. This phenomenon is called delay extension, which causes
That is, the bimodal diffraction loss can be seen as the crosstalk between codes and seriously affects the
sum of the two unimodal diffraction losses, as shown in communication quality. The delay spread can be defined as
Figure 3. the time difference of the last resolvable multipath signal to
the first multipath signal. The parameters [12] used to
describe the delay spread mainly include the average
additional delay and the root mean square delay.
The power delay distribution can generally be calculated
by the following formula [4]:

Ph(τ) = Ah(τ, 0) = Ah(τ)  

where Ah(τ, Δt) is the autocorrelation function of the time-

Figure 3. Bimodal diffraction pattern. varying impulse response h(τ, t). The root mean square τRMS
of the delay spread can be calculated by
The highest peak height of the main obstacle is h1, the
height of the second obstacle peak from the vertex to the
f
receiving point is h2, d1 is the distance from the start point to
the main obstacle point, d2 is the distance from the main W RMS
³ 0
(W  W 0 )2 Ph (W )dW
  
f
obstacle point to the end point, d3 is the distance from the
main obstacle point to the second obstacle point, d4 is the ³ 0
Ph (W )dW
distance from the second obstacle point to the end point.
3) Shadow fading where τ0 is the average delay and can be written as:
Shadow fading is the electric field shading characteristic
of the wireless signal in the propagation path that encounters f
changes in the terrain, house construction, or other obstacles
W0
³ W P (W )dW 
0 h
  
to the transmission path of radio waves. The signal level of f

shadow fading is relatively slow, which is related to the ³ P (W )dW

0 h

terrain distribution on the signal propagation path and the

height of the building. Through measurement statistical Compare our measurements with the results of previous
studies, it has been found that shadow fading is a fading that work in the 900MHz band. The comparison found that the
obeys a lognormal distribution, which can be expressed as: median and standard deviation of the delay given by the

217
existing results are largely different from the fitted values The correction factor α is based on actual measurements,
given in the previous literature, as shown in Table I. 1<α<5. The experiment found that by introducing the
correction factor, the modified Okumura-Hata model can
TABLE I. COMPARISON OF MEASUREMENTS AND PREVIOUS WORK adapt to ultrashort wave wireless communication between
mobile stations with antenna heights of 1-10 meters.
Scenes Measurement results τmed σ
2) Mountain scene
Measurement result in this paper 0.9-1.4μs 1.2-2.3 dB Through the analysis of the actual measurement data, we
City
Cos [5] 1.1μs 2.7 dB can directly use the blade peak model for channel modeling.
Measurement result in this paper 0.5-1.0μs 1.0-2.5 dB
The calculation formula for the propagation loss of the
Rural scattering near the mountain is as follows:
Van Rees [6] 0.1μs 4.2 dB
Measurement result in this paper 1.0-2.0μs 1.9-2.5 dB LS = L(q)+Lr+LA+Lc-1.6+30log(θ)  
Jungle
- - -
Measurement result in this paper 5.8-6.2μs 1.9-2.3 dB
where L(q) is the basic transmission loss, Lr is the ground
Mountains reflection loss; LA is the atmospheric absorption loss; Lc is
De Weck [7] 3.8sμs 3.3 dB the antenna dielectric coupling loss; 1.6 is the difference
between the annual median transmission loss and the annual
2) Doppler shift average transmission loss; 30log(θ) is Correction of
Doppler shift is often referred to as Doppler spread, tropospheric scattering propagation loss under large
resulting from relative motion at the transmit and receive elevation angle conditions [13].
ends. This is a characteristic of any wave process. When modeling single-peak diffraction, the calculation
Specifically, the Doppler shift can be expressed as: method of single-peak diffraction loss can be directly used,
as shown in the following equation:
fd = fcνcosθ/c  
A 6.9  20lg §¨ v  0.1  1  v  0.1·¸ v ! 0.7   
2
where fd is the Doppler shift, fc is the carrier frequency, ν is © ¹
the relative speed, c is the speed of light, and θ is the angle
between the incident wave and the direction of relative Multi-peak diffraction modeling is similar to single-peak
motion. In this paper, due to the fast flight speed of low- diffraction modeling.
altitude scenes, the Doppler shift is large. The movement 3) Jungle scene
speed of people or vehicles in mountains, jungles, and urban The jungle scene mainly considers the influence of trees,
scenes is small. The Doppler shift is relatively small. while the loss of trees is mainly calculated by vegetation loss.
IV. THE RESULT OF CHANNEL MODELING The vegetation loss in this scenario mainly talks about two
situations: one is that the propagation path is all inside the
This part mainly models the four scenes of urban, dense forest; the other is that the propagation path passes
mountainous, jungle and low-altitude, and starts with large- through or near the medium dense forest, or there are
scale fading and small-scale fading. Large-scale fading in moderately dense trees around the antenna, and the antenna
urban, mountainous, and jungle scenarios will be explained height is lower than the height of the tree top.
by theoretical formulas, and small-scale fading will be given For the case where the propagation path is entirely inside
in measured data tables. In particular, large-scale fading and the dense forest. When the receiving and transmitting
small-scale fading will be additionally expressed in low- antennas are all inside the forest, the entire propagation path
altitude scenarios. passes through the trees or through the jungle, and the height
A. Large-Scale Fading in Urban, Mountainous and Jungle of the receiving and transmitting antennas is lower than the
Scenes average height of the forest, the additional loss caused by the
blocking and absorption of trees and jungles The measured
1) Urban scene map [14] shows that when the frequency is as low as 30
From the above analysis, the modified Okumura-Hata MHz and the horizontal polarization is used, the attenuation
model can be used to describe the large-scale fading of urban rate inside the forest is 5 dB/km. Adding vegetation loss to
scenes. But it cannot be applied to the scenarios where the the topographic path loss, which is related to the distance
BS antenna height is less than 30 meters. In order to correct factor d of the transceiver, the path loss in this case is:
the model, the correction factor is redefined, which should be
determined based on the measured data, so the path loss is: LF = LO -5d.  
LU =69.55+26.16lgf-CHB-CH+ (44.9-6.55lghB) lgd   LF is the jungle path loss, the unit is dB, d is the distance
between the base station and the mobile station, the unit is
­13.82lg hB hB ! 30m km. LO = LU-4.78(lgf)2+18.33lgf-40.94.
Among them, CHB ® . The propagation path passes through or near a
¯ D  13.82lg hB hB d 30m moderately dense forest, or there are moderately dense trees

218
around the antenna, and the antenna height is lower than the B. Small-Scale Fading in Urban, Mountainous, and Jungle
height of the tree top. Under this condition, Bollington gives Scenes
a prediction curve in the range of 30MHz to 1000 MHz [15]. Small-scale fading can be described by the relative delay,
When communicating in the 30MHz to 512MHz band, the relative power, carrier frequency, moving speed, angle of
vegetation loss is about 10dB, so the path loss in this case is: incidence, Doppler shift, and delay spread of each path. The
delay spread can be obtained from the previous fit model.
LF = LO -10   Separately, Table II is the small-scale fading of urban scene,
Table III is the small-scale fading of mountain scene, and
where LF is the jungle path loss in units of dB and d is the Table IV is the small-scale fading of jungle scene.
distance between the base station and the mobile station, in
units of km.

TABLE II. SMALL-SCALE FADING OF URBAN SCENE

Trail Relative Average Average Carrier Movement Angle of Doppler Delay
marking delay (us) power (w) power (dB) frequency(MHz) speed (m/s) incidence(°) shift Hz) extension(us)
1 0 0.01 -20 512 11 0 18.77
2 1 1 0 512 11 15 18.13
3 2 0.631 -2 512 11 30 16.25
4 3 0.1585 -8 512 11 45 13.27
5 4 0.0447 -13.5 512 11 60 9.38
6 5 0.0141 -18.5 512 11 75 4.86
1.4513
7 6 0.0028 -25.5 512 11 30 16.25
8 7 0.0063 -22 512 11 60 9.38
9 8 0.005 -23 512 11 75 4.86
10 9 0.0126 -19 512 11 45 13.27
11 13 0.0056 -22.5 512 11 30 16.25
12 14 0.0063 -22 512 11 15 18.13

TABLE III. SMALL-SCALE FADING OF MOUNTAIN SCENE

Angle of
Average Average Delay
Relative Carrier Movement incidence Doppler
Trail marking power power extension
delay (us) frequency speed (m/s) (°) shift (Hz)
(w) (dB) (us)
(MHz)
1 0 0.1 -10 512 5 0 8.53
2 3.5 0.01 -20 512 5 15 8.24
3 9 1 0 512 5 30 7.39
9.4634
4 12 0.2512 -6 512 5 45 6.03
5 18 0.0316 -15 512 5 60 4.27
6 20 0.0063 -22 512 5 75 2.21

TABLE IV. SMALL-SCALE FADING OF JUNGLE SCENE

Average Average Delay

Relative Carrier Movement Angle of Doppler
Trail marking power power extension
delay (us) frequency speed (m/s) incidence shift (Hz)
(w) (dB) (us)
(MHz) (°)
1 0 0.1 -10 512 5 0 8.53
2 1 0.3162 -5 512 5 15 8.24
3 2 1 0 512 5 30 7.39
4 3 0.6310 -2 512 5 45 6.03
5 4 0.1585 -8 512 5 60 4.27
6 5 0.0316 -15 512 5 75 2.21
3.9005
7 7 0.01 -20 512 5 30 7.39
8 8 0.1585 -8 512 5 60 4.27
9 10 0.3162 -5 512 5 75 2.21
10 12 0.1 -10 512 5 45 6.03
11 14 0.1585 -8 512 5 30 7.39
12 15 0.0032 -25 512 5 15 8.24

during the flight, the channel can be divided into navigation,

C. Large-Scale Fading in Low-Altitude Scenes stop, takeoff and landing, and over-the-top state for
Low-altitude airspace usually refers to airspace with a discussion. Different parameters such as aircraft Doppler
true height of less than 1000 m. Due to the different states

219
shift and delay extension in different scenarios require 2) Stop state
different channel models. The stop state [17] refers to the situation where the
1) Navigation state aircraft is parked on the tarmac or taxied on the ground. The
The state of navigation [12] means that the aircraft is in state of the stoppage is due to the fact that the aircraft is on
normal navigation. The channel in this scenario can be the ground, and the surrounding buildings are densely
described by a two-path model. In the navigation state, due covered and the fading is the most serious. At this time, the
to the fact that the aircraft is far away from the ground and channel has no direct path and is composed of a reflection
transmitted in the open space, due to the influence of the scattering diameter. We can use the urban model modeling
ground environment, there may be obstacles such as above.
buildings, mountains, vegetation, etc. Therefore, large-scale Since the state of the stop is similar to the ground state,
fading is mainly caused by path loss and shadow fading. The the large-scale fading is as large as the above-mentioned
total attenuation [16] can be expressed as: urban, jungle, and mountain scenes.
3) Take off and land
PF[dB] = PL[dB] +Ψ[dB]   The take-off and landing state [12] is the state of the
aircraft after it has taken off the ground and before landing.
where PF[dB] is the total attenuation, PL[dB] is the path loss, This scene is a combination of navigational and parking
Ψ[dB] is a shadow fading of the mean value is zero, and the conditions.
variance is σΨ[dB]. Among them, the path loss can be This state and stop state are consistent in large-scale
expressed as follows: fading.
4) Over the top state
The over-the-top state value flies over the base station
­ d
° 10J 1 ˜ log d d db during the voyage of the aircraft. This channel has only one
° d0 direct component, and large-scale fading cannot be ignored.
PL > dB @ ®   
Different from the navigation state, the direct component
°10J log db  10J log d d t db
°̄ 1 2 intensity at this time is very strong, and the related large-
d0 db scale fading reference equations (17), (18), but the specific
parameters are set as shown in Table VI:
This expression expresses the segment path loss when the
transceiver has different signal-to-noise ratios at different TABLE VI. PARAMETERS IN OVER THE TOP STATE
distances, d represents the distance between the transmitter
and the receiver, d0 represents the reference distance, here Scenes 1 2
we take 1km, db represents the segmentation Point distance, Transceiver distance 0.6km 1km
γ1, γ2 represents the measured signal to noise ratio of the
segment. Signal to noise ratio 0.74 2.29
The measured parameters in this scenario are as shown in Shadow fading 1.23dB 2.15dB
Table V:
D. Small-Scale Fading in Low-Altitude Scene
TABLE V. PARAMETERS IN NAVIGATION STATE
In the navigation state, take-off and landing state, over
Scenes 1 2 the top state, see the Table VII, Table VIII and Table IX for
the measured data for small-scale fading.
Transceiver distance 0.6km 1km The small-scale fading in the stop state refers to the
Signal to noise ratio 1.52 3.15 small-scale fading of cities, mountains, and jungles above.
Shadow fading 5.5dB 3.9dB

TABLE VII. NAVIGATION STATUS IN LOW ALTITUDE SCENE

Carrier Doppler
Trail Relative Average Average power Movement Angle of Delay
frequency shift
marking delay (us) power (w) (dB) speed (m/s) incidence (°) extension(us)
(MHz) (Hz)
1 0 0.01 -20 512 540 0 921.6
1.4513
2 1 1 0 512 540 15 890.2

TABLE VIII. TAKEOFF AND LANDING CONDITIONS IN LOW ALTITUDE SCENE

Carrier
Trail Relative Average Average Movement Angle of Doppler Delay
frequency
marking delay (us) power (w) power (dB) speed (m/s) incidence (°) shift (Hz) extension(us)
(MHz)
1 0 0.1 -10 512 78 0 133.12
1.2015
2 0.8 1 0 512 78 45 133.11

220
 TABLE IX. OVER-TOP STATES IN LOW-ALTITUDE SCENE
Carrier
Average power Movement Angle of incidence
Trail marking Average power (w) frequency Doppler shift (Hz)
(dB) speed (m/s) (°)
(MHz)
1 0.1 -10 512 78 0 133.12

[7] Guanjie Tang, Modeling and Simulation of Vehicle-vehicle

V. CONCLUSION Communication Channel, Changchun University of Technology,
2013.
In this paper, we study a generalized channel modeling [8] X. Chu et al, “Channel Modeling for Low-Altitude UAV in Suburban
method for VHF and UHF bands. Also, we have improved Environments Based on Ray Tracer,” IET Conference Publications,
the existing Okumura-Hata model for scenarios with an 2018.
antenna height of less than 30 meters and establish a novel [9] Q. Zhihong et al, “Low Altitude UAV Air-to-Ground Channel
correction formula. Under the support of four scenarios: Measurement and Modeling in Semiurban Environments,” Wireless
urban, mountainous, jungle, and low-altitude, the Communications and Mobile Computing, vol. 2017, pp. 1-11, 2017.
experimental results show that the modeling results and [10] X. Cai et al, “Low altitude UAV propagation channel modelling,” in
measured data can Perfect fitting, which proves that this Proc. EUCAP, pp. 1443-1447, 2017.
generalized channel modeling method can be applied to most [11] Lijiang Liu, Zhiqun Song, Dayong Chen, “Engineering
Approximation Model of Single-Edge Peak Diffraction Loss in
land scenarios based on VHF or UHF bands and has high Mountain Communication,” Radio Communications Technology, vol.
engineering practice value. 41, no. 01, pp. 24-27, 2015.
[12] Xiaoyue Wang, Channel Modeling Analysis and Research of Low-
REFERENCES altitude Air-ground Communication System Based on TD-LTE,
[1] M. M. Lodro, N. Majeed, A. A. Khuwaja, A. H. Sodhro, and S. Shanghai Jiaotong University, 2014.
Greedy, “Statistical Channel modelling of 5G mm Wave MIMO [13] Po Lu, Baiwei Sun, Dan Wu, and Zhen Lu, “Analysis of Cross-
Wireless Communication,” in Proc. iCoMET, pp.1-5, April 2018. mountain Communication Channels,” Radio Communications
[2] Linqiang Ge, “Research and implementation of ultra-shortwave Technology, vol. 36, no. 06, pp. 35-37, 2010.
wireless channel access technology,” Fudan University, 2011. [14] Lingtao Kong and Hao Tang, “Research on Wireless Communication
[3] J. Fischer et al, “A measurement-based path loss model for wireless of Tactical Communication,” Radio computing Technology, vol. 37,
links in mobile ad-hoc networks (MANET) operating in the VHF and no. 3, pp. 26-29, 2011.
UHF band,” IEEE APWC, pp. 349-352. 2012. [15] BERTONIHL, Radio Propagation for Modern Wireless System, USA:
[4] J. Fischer et al, “A novel delay spread distribution model for VHF Pearson Education North Asia Limited: Published House of
and UHF mobile-to-mobile channels,” IEEE EuCAP, pp. 469-472, Electronics Industry, 2000.
2013. [16] X. Cai et al, “Low altitude UAV propagation channel modelling,” in
[5] Xinwen Zhang, Research on modeling method of high-speed railway Proc. EUCAP, pp. 1443-1447, 2017.
car-ground broadband wireless communication channel, Beijing [17] Qiang Chen et al, “Research progress in low-altitude wireless channel
Jiaotong University, 2016. modeling and its equalization technology,” Telecommunications
[6] Jing Zhang, Wireless channel modeling and capacity analysis in Technology, vol. 52, no. 1, pp. 117-124, 2012.
tunnel environment, Beijing Jiaotong University, 2014.

221