sensors
Article
Doppler Modeling and Simulation of Train-to-Train
Communication in Metro Tunnel Environment
Pengyu Zhao 1 , Xiaoyong Wang 2 , Kai Zhang 3 , Yanliang Jin 1, * and Guoxin Zheng 1
Citation: Zhao, P.; Wang, X.; Zhang,
K.; Jin, Y.; Zheng, G. Doppler
Modeling and Simulation of
Train-to-Train Communication in
Metro Tunnel Environment. Sensors
2022, 22, 4289. https://doi.org/
10.3390/s22114289
Academic Editor: Omprakash
Kaiwartya
Received: 9 May 2022
Accepted: 2 June 2022
Published: 4 June 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Sensors 2022, 22, 4289. https://doi.org/10.3390/s22114289
1 Key Laboratory of Specialty Fiber Optics and Optical Access Networks, Joint International Research
Laboratory of Specialty Fiber Optics and Advanced Communication, Shanghai Institute for Advanced
Communication and Data Science, Shanghai University, Shanghai 200444, China; 18321445156@163.com (P.Z.);
gxzheng@staff.shu.edu.cn (G.Z.)
2 CASCO Signal Ltd., Shanghai 200071, China; wangxiaoyong@casco.com.cn
3 Research Institute of Intelligent Networks, Zhejiang Lab, Hangzhou 311121, China; 13262733921@163.com
* Correspondence: wuhaide@shu.edu.cn
Abstract: The communication system of urban rail transit is gradually changing from train-to-ground
(T2G) to train-to-train (T2T) communication. The subway can travel at speeds of up to 200 km/h
in the tunnel environment, and communication between trains can be conducted via millimeter
waves with minimum latency. A precise channel model is required to test the reliability of T2T
communication over a non-line-of-sight (NLoS) Doppler channel in a tunnel scenario. In this paper,
the description of the ray angle for a T2T communication terminal is established, and the mapping
relationship of the multipath signals from the transmitter to the receiver is established. The channel
parameters including the angle, amplitude, and mapping matrix from the transmitter to the receiver
are obtained by the ray-tracing method. In addition, the channel model for the T2T communication
system with multipath propagations is constructed. The Doppler spread simulation results in this
paper are consistent with the RT simulation results. A channel physics modelling approach using an
IQ vector phase shifter to achieve Doppler spread in the RF domain is proposed when paired with
the Doppler spread model.
Keywords: train-to-train (T2T) communication; doppler spread; channel simulation; metro tunnel
1. Introduction
With the development of vehicle-to-everything (V2X) communication technology,
intelligent transportation systems consisting of vehicle-to-vehicle (V2V) communication,
vehicle-to-infrastructure (V2I) communication, vehicle-to-pedestrian (V2P) communication,
and vehicle-to-cloud network (V2N) communication have been introduced [1,2]. V2X
communication technology mainly includes dedicated short-range communication (DSRC)-
based vehicle networks and cellular-based vehicle networks [3]. DSRC is a wireless access
solution based on the IEEE 802.11p standard with good distributed characteristics but
shortcomings including a limited coverage, low transmission rate, and inability to be used
in high-speed mobile environments. The cellular network-based wireless access technology
C-V2X developed by the Third Generation Partnership Project (3GPP) enables vehicles to
operate normally in scenarios with and without cellular network coverage [4]. In order to
improve the performance of V2X communication systems, 3GPP has successively released
the LTE-V2X network based on Long Term Evolution (LTE) and a 5G-V2X network for
fifth-generation new radio (5G NR) [5]. However, as the number of communication devices
and digital applications grows, communication systems face new challenges in terms of
data rate, latency, coverage, spectral efficiency, and security [6–8]. Industry 5.0 will have
access to sixth-generation (6G) wireless communication networks with data rates up to
0.1 TB/s and massive ultra-high-speed wireless access [9]. Wireless data sharing between
entities such as vehicles, trains, drones, infrastructure roadside units, traffic data centers,
https://www.mdpi.com/journal/sensors
Sensors 2022, 22, 4289 2 of 13

etc., can be further achieved with industry 5.0 [10]. Combine artificial intelligence, cloud
computing, block chain, and other emerging technologies to develop reliable autonomous
driving applications [11], reasonably plan vehicle routes and driving speeds, reduce vehicle
energy consumption and promote green development of traffic.

1.1. Related Literature

In the Orthogonal Multiple Access (OMA) technology used by existing networks,
the limitation of spectrum resources results in low access efficiency and low data rates
in dense vehicular networks. In the next generation of V2X communication, the efficient
utilization of radio resources by the multiplexing of transmitted signals and successive
interference cancellation (SIC) at the receiving end by using the Non-Orthogonal Mul-
tiple Access (NOMA) technique is proposed [12]. Reference [13] proposed a two-stage
scheme of centralized resource allocation and distributed power control strategy that
increased communication capacity by 5% while lowering transmission power usage by
36%. Reference [14] also verified that the system capacity can be improved by adopting
centralized resource allocation and distributed power control methods. In [15], a roadside
unit-based energy-saving power allocation scheme for NOMA was proposed, which an-
alyzed the energy efficiency maximization problem under the constraints of incomplete
channel state information (CSI), quality of service (QoS), and power constraints. Refer-
ence [16] proposed a cognitive radio-assisted NOMA-V2X system model and analyzed its
outage probability and throughput performance. Reference [17] provided a new approach
to improve spectrum utilization and system capacity, and it also addressed the near–far
problem of NOMA-enabled backscatter communication systems. Reference [18] proposed
an energy-efficient transmission optimization framework for V2X networks based on am-
bient backscatter communication and NOMA for imperfect CSI. Under the premise of
ensuring QoS, the total transmit power of the V2X network is minimized by optimizing
the power distribution of the base station and the reflection coefficient of the backscatter
sensor. Compared with the traditional NOMA-based V2X network, the V2X network using
the fusion of ambient backscatter communication and NOMA has advantages in terms
of energy efficiency. On the other hand, the deep integration of artificial intelligence (AI)
tools with wireless communication systems [19–21] can further improve network perfor-
mance [22,23] and plays an important role in channel estimation, scene recognition, velocity
prediction, channel characteristic prediction [24–26], and channel modeling [27–29]. The
control system based on T2T communication extends the T2G wireless communication
system between trains. The front and rear trains can interchange data immediately, al-
lowing the back train to acquire real-time position and speed information from the front
train. As a result, railway operations can be made safer and the safe spacing between trains
can be reduced. LTE wireless communication technology has the advantages of having
strong anti-interference ability, a high transmission rate, and low system delay. It can be
further applied to the train control system based on T2T communication [30]. However,
the Doppler shift induced by a high train speed and high signal frequency will have an
impact on the wireless communication system’s performance.

1.2. Motivation and Contributions

To verify the T2T communication’s reliability, the T2T communication channel model
must be implanted into a simulation platform of a physical channel with time varying traffic,
and the train control system must be combined to form a hardware-in-the-loop simulation
platform to complete the T2T communication performance verification. Therefore, it is
necessary to establish a channel simulation model of T2T communication, especially the
Doppler spread model caused by high-speed movement. Existing Doppler research on T2T
communication mainly focuses on ground scenarios [31–36] and V2V communication [37]
scenarios in highway tunnels. However, the subway tunnel is narrow and long, including
straight, curved, climbing, descending, and other scenarios, making it more challenging
than the ground tunnel. Trains using millimeter wave communications traveling at high
Sensors 2022, 22, 4289 3 of 13

speeds of more than 200 km/h along a given track in a tunnel environment cause higher
Doppler shifts. Thus far, the channel modeling work in subway tunnels has mainly studied
the influence of the LoS path, and the single-bounced and double-bounced signal on the
channel in the same coordinate space in the V2G communication scene [38–40]. In the
T2T communication system, the transceiver ends are located in different locations, so it
is difficult to apply the same coordinate space analysis. In addition, the movement of
the antenna causes rapid changes in the channel environment, complicating the signal
propagation process. The influence of multi-bounced signals and the movement of the
transceiver antenna on the Doppler shift must be considered, so a Doppler spread model
suited for the movement of two-terminal terminals in the tunnel must be established.
In this paper, the receiver (Rx) and transmitter (Tx) coordinate systems are established,
and the Doppler spread model of the multipath signals from the transmitter to the receiver
is established in their respective coordinate systems. The mapping matrix approach is
proposed as an innovative solution to the problem of signal matching at the receiver and
transmitter, as well as a method for obtaining the mapping matrix. In order to verify the
proposed Doppler model for T2T communication in the tunnel, the ray-tracing (RT) [41]
approach is utilized to obtain the angle and amplitude of the signals at the transmitting and
receiving ends. By comparing the simulation results of the RT approach with the simulation
results of the Doppler spread model, the validity of the Doppler spread model is proved. A
channel physical simulation method using IQ vector phase shifters [42,43] is proposed to
execute the T2T communication channel simulation in the tunnel environment, which can
be used for future tunnel environments when combined with the T2T channel model and
test data analysis in the tunnel [44–46]. It can provide a reference for 5G millimeter wave
physical channel simulation in the future tunnel environment.

1.3. Article Structure

The rest of this paper is organized as follows: In Section 2, the Doppler shift models for
transmitted and received signals are established, followed by the solution method for the
multipath signal mapping relationship at the transmit and receive ends and the multipath
signal’s Doppler spread model. In Section 3, the RT simulation method is used to obtain the
angle, amplitude, and mapping relationship of the transmitted and received signals, and
the Doppler spread simulation of the theoretical model is discussed. In Section 4, combined
with the analysis of the communication signal in the tunnel, a method to realize physical
channel simulation using an IQ vector phase shifter is proposed. Finally, Section 5 provides
the conclusion of this paper.

2. Modeling of Wireless Channel for T2T Communication in Tunnel Scenario

2.1. Wireless Channel Model of T2T Communication
Figure 1 depicts a T2T communication scenario in which the front and rear trains
run through a tunnel with a width W and a height H. The complicated propagation
process of multipath signals through straight or curved tunnels, such as reflection and
scattering, is represented by the multipath link between two trains. The space where the
multipath link is located represents the complex channel environment, such as straight or
curved tunnels. A three-dimensional coordinate system is established at the transceiver
antennas, with the tunnel depth as the x-axis, the height as the z-axis, and the horizontal
direction as the y-axis. The downlink of the T2T communication consists of the transmitting
antenna Tx located in the rear train and the receiving antenna Rx located in the front
train. In addition, vt and vr represent the moving speed of the transmitting antenna
and the receiving antenna, and the moving direction follows the same path as the x-axis.
The received signal consists of the NLoS signal that reaches the receiving antenna after
multiple reflections and scattering. In the LoS scenario, it includes the LoS signal from
the transmitting antenna to the receiving antenna. In order to study the doppler spread of
T2T communication in the tunnel environment, it is assumed that the receiving and the
transmitting antennas are omnidirectional uniform antennas.
 signal consists of the NLoS signal that reaches the receiving antenna after multiple re
tions and scattering. In the LoS scenario, it includes the LoS signal from the transmi
antenna to the receiving antenna. In order to study the doppler spread of T2T comm
Sensors 2022, 22, 4289 cation in the tunnel environment, it is assumed that the receiving and 4the of 13transmi
antennas are omnidirectional uniform antennas.

𝑣𝑟 z
z Rx
𝑣𝑡
Tx y
Multipath
H Link o x
y
W
o x

Figure 1. Three-dimensional map of T2T communication in tunnel scenario.

Figure 1. Three-dimensional map of T2T communication in tunnel scenario.
Figure 2 shows the schematic diagram of the signal leaving the transmitting antenna
Figureat2the
and arriving shows the antenna.
receiving schematic diagram
θ ZOD,i of the
and φ AOD,i signalthe
represent leaving
zenith the angle transmitting
of depar- ant
and arriving at the receiving antenna. 𝜃𝑍𝑂𝐷,𝑖 and 𝜙𝐴𝑂𝐷,𝑖 represent the zenith
ture (ZOD) and the azimuth angle of departure (AOD) of the i-th transmitted signal. θ ZOA,m angle o
and
parture
φ AOA,m represent
(ZOD) andthe thezenith angle of
azimuth arrival
angle of(ZOA) and azimuth
departure (AOD) angle
of of arrival
the i-th (AOA)
transmitted si
of the m-th received signal, where θ ZOD,i , θ ZOA,m ∈ [0, 180◦ ], φ AOD,i , φ AOA,m ∈ [0◦ , 360◦ ),
𝜃𝑍𝑂𝐴,𝑚 and 𝜙𝐴𝑂𝐴,𝑚 represent the zenith angle of arrival (ZOA) and azimuth angle o
1≤ i ≤ I, 1 ≤ m ≤ M, and I = M, I and M represent the number of multipaths of trans-
rival (AOA)
mitted of the
and receiving m-th Let
signals. received signal,
Pt,i denote where power
the transmit 𝜃𝑍𝑂𝐷,𝑖 ,of𝜃𝑍𝑂𝐴,𝑚
the i-th∈path
[0,180° ], and
signal, 𝜙𝐴𝑂𝐷,𝑖 , 𝜙𝐴𝑂𝐴
[0°, 360°), 1 ≤ 𝑖 ≤ 𝐼, 1 ≤ 𝑚 ≤ 𝑀, and 𝐼 = 𝑀, 𝐼 and 𝑀 represent
Pr,m denote the receive power of the m-th path signal. Then the spherical coordinate form the number of
tipaths
of of transmitted
the transmitter signal and andthereceiving signals.
receiver signal canLet 𝑃𝑡,𝑖 denote
be expressed the
as ( Pt,i , θtransmit power
ZOD,i , φ AOD,i ) of t
and ( P
th path ,
r,m signal,
θ ,
ZOA,m and
φ AOA,m𝑃𝑟,𝑚 denote the receive power of the m-th path signal. Then the sp
) . Due to the movement of the transmitting antenna, the Doppler
i
ical fcoordinate
shift d,t of the signal
form( Pt,iof
, θ ZOD,i , φ AOD,i ) can be
the transmitter expressed
signal and the as receiver signal can be expresse
(𝑃𝑡,𝑖 , 𝜃𝑍𝑂𝐷,𝑖 , 𝜙𝐴𝑂𝐷,𝑖 ) and (𝑃𝑟,𝑚 , 𝜃𝑍𝑂𝐴,𝑚
i
,v𝜙
t 𝐴𝑂𝐴,𝑚
). Due to the movement of the transmi
𝑖 f d,t = cos(ψt,i ) (1)
antenna, the Doppler shift 𝑓𝑑,𝑡 of the λ0signal (𝑃𝑡,𝑖 , 𝜃𝑍𝑂𝐷,𝑖 , 𝜙𝐴𝑂𝐷,𝑖 ) can be expressed as
T 𝑡 𝑣
) 𝑖= r̂=
cos(ψt,i 𝑓 ·v̂t cos(𝜓 ) (2)
𝑑,𝑡 t,i 𝑡,𝑖
 𝜆0 
sin(θ ZOD,i ) cos(φ AOD,i )
𝑇
r̂t,i cos(𝜓
=  sin(θ ZOD,i (φ)AOD,i
) sin𝑡,𝑖 = 𝑟̂𝑡,𝑖
) ∙ 𝑣̂𝑡 (3)
cos(θ ZOD,i )
sin(𝜃𝑍𝑂𝐷,𝑖 ) cos(𝜙𝐴𝑂𝐷,𝑖 )
v̂t = [1, 0, 0] T (4)
𝑟̂𝑡,𝑖 = [ sin(𝜃𝑍𝑂𝐷,𝑖 ) sin(𝜙𝐴𝑂𝐷,𝑖 ) ]
where ψt,i represents the angle between the signalcos(𝜃 ( Pt,i , θ ZOD,i , φ AOD,i ) and the moving
𝑍𝑂𝐷,𝑖 )
direction of Tx. r̂t,i and v̂t represent the unit vector in the direction of the transmitted signal
and the unit vector in the direction of Tx movement. λ0 is ]the𝑇 signal carrier wavelength,
𝑣̂𝑡 = [1,0,0
T
and (·) represents the transpose of the matrix. Due to the movement of the receiving
where the𝜓𝑡,𝑖Doppler
represents m angle between the signal (𝑃 at, 𝜃
shift fthe ) and
, 𝜙𝐴𝑂𝐷,𝑖end can the mo
antenna, d,r of the signal ( Pr,m , θ ZOA,m , φ AOA,m )𝑡,𝑖 the receiving
𝑍𝑂𝐷,𝑖
expressedofasTx. 𝑟̂𝑡,𝑖 and 𝑣̂𝑡 represent the unit vector in the direction of the transm
direction
be
vr
signal and the unit vector in the m direction
f d,r = cos(of Tx
ψr,m ) movement. 𝜆0 is the signal(5) carrier w
𝑇 λ0
length, and (∙) represents the transpose of the matrix. Due to the movement of th
T
ceiving antenna, the Doppler shift 𝑚) = r̂r,m
cos(ψ𝑓r,m ·v̂signal (6)the recei
(𝑃𝑟,𝑚 , 𝜃𝑍𝑂𝐴,𝑚 , 𝜙𝐴𝑂𝐴,𝑚 ) at
𝑑,𝑟 of the
r

end can be expressed as

 
sin(θ ZOA,m ) cos(φ AOA,m )
r̂r,m =  sin(θ ZOA,m ) sin𝑣(φ AOA,m )  (7)
𝑚 𝑟
𝑓𝑑,𝑟
cos = )cos(𝜓𝑟,𝑚 )
(θ ZOA,m
𝜆0
T
v̂r = [1, 0, 0] 𝑇 (8)
cos(𝜓𝑟,𝑚 ) = 𝑟̂𝑟,𝑚 ∙ 𝑣̂𝑟
where ψr,m represents the angle between the signal ( Pr,m , θ ZOA,m , φ AOA,m ) and the moving
direction of Rx. r̂r,m and v̂r represent the unitsin(𝜃 of the) cos(𝜙
vector𝑍𝑂𝐴,𝑚 received ) direction and the
signal
𝐴𝑂𝐴,𝑚
unit vector of the Rx moving direction.𝑟̂ = [ sin(𝜃 ) sin(𝜙 )]
𝑟,𝑚 𝑍𝑂𝐴,𝑚 𝐴𝑂𝐴,𝑚
cos(𝜃𝑍𝑂𝐴,𝑚 )

𝑣̂𝑟 = [1,0,0]𝑇
 Sensors 2022, 22, x FOR PEER REVIEW 5 of 13
where 𝜓𝑟,𝑚 represents the angle between the signal (𝑃𝑟,𝑚 , 𝜃𝑍𝑂𝐴,𝑚 , 𝜙𝐴𝑂𝐴,𝑚 ) and the mov-
ing direction of Rx. 𝑟̂𝑟,𝑚 and 𝑣̂𝑟 represent the unit vector of the received signal direction
and the unit vector of the Rx moving direction.
where 𝜓𝑟,𝑚 represents the angle between the signal (𝑃𝑟,𝑚 , 𝜃𝑍𝑂𝐴,𝑚 , 𝜙𝐴𝑂𝐴,𝑚 ) and the mov-
Sensors 2022, 22, 4289 ing direction of Rx. 𝑟̂𝑟,𝑚 and 𝑣̂𝑟 represent the unit vector of the received signal direction
5 of 13
z z of the Rx moving direction.
and the unit vector z z
𝜃𝑍𝑂𝐷,𝑖 𝜃𝑍𝑂𝐷,𝑖 𝜃𝑍𝑂𝐴,𝑚 𝜃𝑍𝑂𝐴,𝑚
z Tx Tx z z Rx Rx z
o x y o x y
𝜃𝑍𝑂𝐷,𝑖 𝜃𝑍𝑂𝐷,𝑖 𝜃𝑍𝑂𝐴,𝑚 𝜃𝑍𝑂𝐴,𝑚
Tx
o o
Rx view of Rx Rx
o Front Tx
view of Tx x y Left view of Tx
o
Front
x y Left view of Rx
y o y o
Front view of Tx Left view of Tx Front view of Rx Left view of Rx
𝜙𝐴𝑂𝐴 ,𝑚
y 𝜙𝐴𝑂𝐷 ,𝑖 y x
o Tx x o 𝜙𝐴𝑂𝐴 ,𝑚
Rx
𝜙𝐴𝑂𝐷 ,𝑖 x
o Tx x o Rx
Top view of Tx Top view of Rx
(a) Top view of Rx (b)
Top view of Tx
(a)Figure 2. Three views of T2T communication scenario: (a) transmitter,
(b) (b) receiver.

Figure
2.2. 2. Three of
Matching views of T2T communication
Receiving scenario:
and Transmitting Rays(a)(a) transmitter, (b) receiver.
Figure 2. Three views of T2T communication scenario: transmitter, (b) receiver.
In the proposed
2.2.Matching
Matching system, the transmitted
Rays and the received signals have a one-to-one
2.2. ofofReceiving
Receiving andTransmitting
and Transmitting Rays
mapping relationship, which means that for each received signal, a unique corresponding
InInthe
theproposed
proposedsystem,
system,thethetransmitted
transmittedand andthe
thereceived
receivedsignals
signalshave
haveaaone-to-one
one-to-one
transmitted signal can always be located, completing a full signal chain from Tx to Rx.
mappingrelationship,
mapping relationship,whichwhichmeans
meansthatthatfor
foreach
eachreceived
receivedsignal,
signal,aaunique
uniquecorresponding
corresponding
According to the roughness of the sidewall of the tunnel, the mapping relationship be-
transmitted signal
transmitted signal cancanalways
always be be located,
located, completing
completing aa full
full signal
signal chain
chain from
from TxTx to
to Rx.
Rx.
tween
According
According thetotransmitted
tothethe and
roughness
roughness of receiving
of the
the signals
sidewall
sidewall of can
of the be the
solved
the tunnel,
tunnel, theby the
mapping
mapping spatial mirror
relationship
relationship betweenmethod
be-
and
the
tween thetherandom
transmitted and
transmittedscatterer
and distribution
receiving signals can
receiving method.
be solved
signals The
can be by space mirror
the spatial
solved by method
themirror method
spatial is shown
mirror and in Fig-
the
method
ure
and 3.
random the scatterer distribution
random scatterer method. The
distribution spaceThe
method. mirror
spacemethod
mirrorismethod
shown is inshown
Figure in
3. Fig-
ure 3.

Tx
Tx

Tx Rx
Tx LoS Rx
NLoS LoS
NLoS

Figure3.3.Channel
Figure Channelmodel
model based
based onon spatial
spatial mirror
mirror method.
method.
Figure 3. Channel model based on spatial mirror method.
For
ForthetheNLoS
NLoSpropagation
propagation path
pathin in
thethe
tunnel,
tunnel,thisthis
method
method solely considers
solely reflection,
considers reflection,
and For
the the NLoS
mirror space propagation
is formed path
with theinreflection
the tunnel, this method
surface as the solely
axis. For considers
the signal reflection,
reflected
and the mirror space is formed with the reflection surface as the axis. For the signal re-
and
for the
k (k =formirror
0, 1, 2, .=space
. 0, K)2,
. , 1, is formed
times, the with
mirror the reflection
image point surface
Tx’ point as the axis. antenna
of transmitting For the Tx
signal
needsre-
flected
flected k (k
for k (kthrough
= 0, 1, 2, k…, …, K) times,
K) times, the mirror
the mirror image Tx’ of transmitting antenna Tx
to be obtained times of mirror image.image point Tx’ of transmitting antenna Tx
needs
needs to be
to be obtained
obtainedthrough throughk ktimes
times ofof mirror
mirror image.
image.
The scatterers’ location at the tunnel’s size wall follows a uniform random distribution,
The scatterers’
The scatterers’ location at the tunnel’s size wall follows a uniform random distribu-
as shown in Figure 4.location at the tunnel’s
In this model, scattererssizeare
wall follows
randomly a uniform
distributed random
along the distribu-
inner
tion, as
tion,ofastheshown
shown in
in WhenFigure
Figurethe 4.
4. InIn this model,
thissignal
model, scatterers
scatterers are
are the randomly
randomly distributed
distributed along
the the
wall tunnel. NLoS passes through scatterer, the rough along
scatterer
inner
surface wall
inner wallleads ofto
of the
the tunnel.randomness
tunnel.
a certain Whenthe
When theNLoS
NLoS
in the signal
signal passes
passes
direction through
ofthrough
the thethe
secondary scatterer,
scatterer, the
the wave.
radiation roughrough
When the signal is repeatedly scattered, the mapping between the transmitted signal and
the received signal can be considered as random mapping.
Sensors 2022, 22, x FOR PEER REVIEW 6 of 13

scatterer surface leads to a certain randomness in the direction of the secondary radiation
Sensors 2022, 22, 4289 wave. When the signal is repeatedly scattered, the mapping between the transmitted6 of 13 sig-
nal and the received signal can be considered as random mapping.

scatterer

NLoS
Tx LoS Rx

Figure4.4.Channel
Figure Channelmodel
modelbased
basedonon geometric
geometric space
space random
random scatter.
scatter.

Themapping
The mappingrelationship
relationship between
between thethe transmitter
transmitter multipath
multipath signal
signal andand the receiver
the receiver
multipath
multipathsignal
signalcan
canbebeexpressed
expressedas as

a11 𝑎 𝑎·12 ⋯
a1M 𝑎1𝑀 1 1
 Γ
Γ1 11 ··
   
1 a12
ΓΓ2= (Γ2 ) = 𝐴𝑁𝑇a= 𝑎a 21 𝑎·22·· ⋯ 𝑎  2 
a2M 2𝑀 2
21 ( 22 ) ( ) (9)
Γ =  .  = ⋮AN T =  ⋮. ⋮ ⋱ ⋮ . ⋮ (9)
   
. . ..
 ..  .. 𝑎 .. 𝑎. .  .. 
  
Γ𝐼 
𝐼1 𝐼2 . ⋯ 𝑎𝐼𝑀 𝑀
ΓI 𝑎11a I1 𝑎a I2 ⋯· · ·𝑎 a I M M
12 1𝑀
𝑎21 𝑎22 ⋯ 𝑎 2𝑀
= (a12 · · · a1M )

𝐴a11 (10)
 a21 a22 ⋮ · ·⋮· ⋱ a 2M
⋮
𝑎 𝑎 ⋯ 𝑎

A= .. ..𝐼1 . . 𝐼2 .. 𝐼𝑀 𝐼×𝑀 (10)


 . . . . 
a I2(1 · 2
a I1𝑁 = a I M 𝑀 )1×𝑀
·· ⋯ I×M
(11)
where A is the mapping matrix with I rows· and
N= 1 2 ·· MM columns.

The element 𝑎𝑖𝑚 in the map-
(11)
1× M
ping matrix represents the mapping relationship between the signal (𝑃𝑟,𝑚 , 𝜃𝑍𝑂𝐴,𝑚 , 𝜙𝐴𝑂𝐴,𝑚 )
at theAreceiving
where is the mapping
end andmatrixthewith I rows
signal and
(𝑃𝑡,𝑖 M columns.
, 𝜃𝑍𝑂𝐷,𝑖 , 𝜙𝐴𝑂𝐷,𝑖 )The
at element aim in the mapping
the transmitting end. Let the
matrix represents the mapping relationship between the signal ( P
transmitter signal (𝑃𝑡,𝑥 , 𝜃𝑍𝑂𝐷,𝑥 , 𝜙𝐴𝑂𝐷,𝑥 ) and the receiver signal (𝑃𝑟,𝑦 , 𝜃𝑍𝑂𝐴,𝑦
r,m , θ ZOA,m , φ AOA,m ) at)the
, 𝜙𝐴𝑂𝐴,𝑦 be the
receiving end and the signal ( Pt,i , θ ZOD,i , φ AOD,i ) at the transmitting end.
Let the transmitter
same signal, where 1 ≤ 𝑥 ≤ 𝑀, 1 ≤ 𝑦 ≤ 𝑀, then the element 𝑎𝑖𝑚 is
signal ( Pt,x , θ ZOD,x , φ AOD,x ) and the receiver signal Pr,y , θ ZOA,y , φ AOA,y be the same signal,
where 1 ≤ x ≤ M, 1 ≤ y ≤ M, then the element 1, 𝑖 a=
im 𝑥,
is 𝑚 = 𝑦
𝑎𝑖𝑚 = { (12)
 0, 𝑖 = 𝑥, 𝑚 ≠ 𝑦
1, i = x, m = y
The matrix N is the subscript =
aimmatrix of the multipath signal at the receiving end.(12)After
0, i = x, m 6= y
rearranging the mapping matrix, the subscript matrix Γ corresponding to the signal at the
transmitting end
The matrix obtained.
N is the subscript matrix of the multipath signal at the receiving end. After
rearranging the mapping matrix, the subscript matrix Γ corresponding to the signal at the
2.3. Doppler Effect
transmitting end isat the Transmitter and Receiver
obtained.
Rearrange the Doppler shifts of each ray at the receiver so that they are in the same
2.3. Doppler Effect at the Transmitter and Receiver
order as the corresponding ray at the transmitter
Rearrange the Doppler shifts of each ray at the receiver so that they are in the same
𝑖 Γ
𝑓𝑓𝑑,𝑟
order as the corresponding ray at the transmitter= 𝑓𝑑,𝑟𝑖 (13)
When both the receiving and transmitting
i Γi ends move, the Doppler shift 𝑓 𝑖 of the
f f d,r = f d,r (13)
𝑑
communication signal can be expressed as
i
When both the receiving and transmitting 𝑖 ends 𝑖move, the Doppler shift f d of the
𝑓𝑑𝑖 = 𝑓𝑑,𝑡 + 𝑓𝑓𝑑,𝑟 (14)
communication signal can be expressed as
Assuming that the transmitted signal is 𝑥𝑝 (𝑡 ) = cos(2𝜋𝑓𝑐 𝑡), the bandpass form 𝑅𝑝𝑖 (𝑡 )
f di =
of the i-th path received signal can be
i
f d,t i
+ f f d,r
expressed as (14)

Assuming that the transmitted 𝑅𝑝𝑖 (𝑡 )signal

= 𝑐𝑖 cos[2𝜋(𝑓 𝑐+
is x p (t) = 𝑓𝑑𝑖()(𝑡
cos c t )𝜏,𝑖 )]
2π f− the bandpass form Rip (t) (15)
of the i-th path received signal can be expressed as
where 𝑐𝑖 and 𝜏𝑖 represent the power normalized amplitude and time delay of the i-th
path signal, so the baseband form 𝑅𝑏𝑖 (𝑡 )2πof the i-th path received signal is expressed as
h   i
Rip (t) = ci cos f c + f di (t − τi ) (15)
 Sensors 2022, 22, 4289 7 of 13
Sensors 2022, 22, x FOR PEER REVIEW 7 of 13

where ci and τi represent the power normalized amplitude and time delay of the i-th path
i 𝑖
𝑖 Rb ( t ) of𝑗2𝜋(𝑓
signal, so the baseband form 𝑡−𝑓𝑑𝑖path
the 𝑑i-th 𝜏𝑖−𝑓𝑐 𝜏received
𝑖) signal
−𝑗𝜑 is 𝑖expressed as
𝑖 𝑗2𝜋𝑓𝑑 𝑡 (16)
𝑅𝑏 (𝑡 ) = 𝑐𝑖 𝑒 = 𝑐𝑖 𝑒 𝑒
where 𝜑𝑖 = 2𝜋(𝑓𝑑𝑖 𝜏𝑖 + 𝑓𝑐R𝜏ib𝑖 ), = ci e j2π
(t)when
i 𝑖i
d t −𝑓f𝑑d ,τi −
𝑓𝑐( f≫ 𝜑f𝑖c τ≈ ci𝑐e𝜏−𝑖jϕ
2𝜋𝑓
i) =
i
i e j2π f d tthe baseband form (16)
; thus, of the
multipath received signal can be expressed as
where ϕi = 2π f di τi + f c τi , when f c  f di , 𝐼ϕi ≈ 2π f c τi ; thus, the baseband form of the

multipath received signal can be expressed as 𝑖

𝑅𝑏 (𝑡 ) = ∑ 𝑐𝑖 𝑒 −𝑗𝜑𝑖 𝑒 𝑗2𝜋𝑓𝑑𝑡 (17)
I 𝑖=1

𝐼 → ∞, the∑received
i
Rb (t) = ci e− jϕi e j2π f d t (17)
When the number of rays i =1
signal can be expressed as the integral
function of all frequency components from the minimum Doppler frequency 𝑓𝑑,𝑚𝑖𝑛 to the
WhenDoppler
maximum the number I → ∞ , the received signal can be expressed as the integral
of rays 𝑓
frequency 𝑑,𝑚𝑎𝑥
function of all frequency components from the minimum Doppler frequency f d,min to the
𝑓
maximum Doppler frequency f d,max 𝑑,𝑚𝑎𝑥 −𝑗𝜑𝑖 𝑗2𝜋𝑓𝑑 𝑡
𝑅𝑏 (𝑡 ) = ∫ √𝑃(𝑓𝑑 )𝑒 𝑒 d𝑓𝑑 (18)
Z f𝑓𝑑,𝑚𝑖𝑛q
d,max
Rb (t) = P( f d )e− jϕi e j2π f d t d f d (18)
where 𝑃(𝑓𝑑 ) represents the continuous
f d,min Doppler spectral function.

3.where P( f dSpread
Doppler ) represents the continuous
Simulation Doppler spectral function.
and Results
3.1. Simulation
3. Doppler ModelSimulation
Spread and Parameter
andSettings
Results of RT
3.1. In order toModel
Simulation verifyandthe Doppler
Parameter model
Settings proposed in this paper, the RT simulation
of RT
method In order to verify the Doppler model proposed in this software
is used in the Wireless Insite (WI) simulation paper, the to RTobtain the angle
simulation methodand
amplitude characteristics of the transmitted and received signals, and
is used in the Wireless Insite (WI) simulation software to obtain the angle and amplitude the mapping rela-
tionship of the of
characteristics transmitted and received
the transmitted and receivedsignals is extracted.
signals, The 3D relationship
and the mapping model of the oftunnel
the
istransmitted
built using and the 3D modeling
received software
signals Inventor.
is extracted. TheThe tunnel of
3D model is athe
rectangular straight
tunnel is built usingtun-
nel
thewith a length ofsoftware
3D modeling 300 m, aInventor.
width of 5Them, and
tunnela height of 5 m. In the
is a rectangular simulation,
straight tunnel the
withsignal
a
length of 300 m, a width of 5 m, and a height of 5 m. In the simulation,
carrier frequency is 28 GHz, and both the transmitting antenna and the receiving antenna the signal carrier
frequency
are is 28 GHz,
omnidirectional and bothlocated
antennas, the transmitting
in the center antenna
of theand the receiving
tunnel 100 m and antenna
200 m areaway
omnidirectional antennas, located in the center of the tunnel 100 m
from the tunnel entrance. The distance between the antenna and the tunnel ground is 2 m,and 200 m away from
asthe tunnel
shown inentrance.
Figure 5.The Thedistance between the
tunnel structure andantenna
transmit andandthereceive
tunnel antenna
ground is 2 m, as
parameters
shown in Figure
are listed in Table 1. 5. The tunnel structure and transmit and receive antenna parameters are
listed in Table 1.

5m 100 m 100 m 100 m

Tx/Rx
5m 5m Tx Rx
2m 2m

Left view Front view

Schematicdiagram
Figure5.5.Schematic
Figure diagram of WI simulation
simulationtunnel.
tunnel.

Parametric structure
Table1.1.Parametric
Table structure of
of transmitted
transmittedand
andreceiving
receivingantennas in in
antennas tunnel.
tunnel.
Parameter
Parameter Value
Value
fcfc(GHz)
(GHz) 2828
(L, W, H) (m) (300,5,5)
(L, W, H) (m)
(xt , yt , zt ) (m)
(300,5,5)
(100,0,2)
(x(xt,r ,yytr, ,zztr)) (m)
(m) (100,0,2)
(200,0,2)
(vrt,, y
(x vr,) z(km/h)
r) (m) (160,80), (160,160)
(200,0,2)
(vt, vr) (km/h) (160,80), (160,160)
The tunnel material and simulation ray parameter settings are shown in Table 2. The
tunnel
Thematerial is concrete,
tunnel material and and the parameters
simulation such as the
ray parameter permittivity
settings and conductivity
are shown in Table 2. The
tunnel material is concrete, and the parameters such as the permittivity and conductivity
of the tunnel material are calibrated according to the measured data of Shanghai Metro
Sensors 2022, 22, x FOR PEER REVIEW 8 of 13
Sensors 2022, 22, 4289 8 of 13
Sensors 2022, 22, x FOR PEER REVIEW 8 of 13

Line 7 [10].
of the Moreover,
tunnel material the signal’s maximum
are calibrated accordingreflection time indata
to the measured the tunnel
of Shanghaiis set Metro
to 10, its
maximum
Line scattering
Line 77[10].
[10]. time
Moreover,
Moreover, theissignal’s
the set to 2,
signal’s and its transmission
maximum
maximum reflection
reflection timetime
time in is
in theset
the to 0. isis set
tunnel
tunnel set to
to10,
10, its
its
maximum
maximumscattering
scatteringtime
timeisisset
setto
to2,2,and
andits
itstransmission
transmissiontime
timeisisset
settoto0.0.
Table 2. WI simulation parameter settings.
Table 2. WI simulation parameter settings.
Table 2. WI simulation parameter settings.
Material Permittivity Conductivity (S/m) Thickness(m) Roughness(m)
Tunnel Parameters Material Permittivity Conductivity
Tunnel Parameters
Material
Concrete Permittivity
5.31 0.48(S/m)
Conductivity (S/m) Thickness 0.5(m)
Thickness(m) Roughness
Roughness(m)
0.005 (m)
Tunnel Parameters
Concrete
Concrete
Reflections times Transmission 5.31 times Scattering
5.31 0.48 times Interval
0.48 0.5 of rays Number
0.5 0.005 of rays
0.005
Ray Parameters
Reflections
10 Times
Reflections times Transmission
Transmission
0 Times times Scattering
Scattering times Interval
2Times Interval of of
0.25° rays Number
Rays Numberof of
250 rays
Rays
Ray Parameters
Ray Parameters ◦
10
10 00 22 0.25
0.25° 250
250
3.2. Simulation Results and Analysis
3.2. Simulation
3.2. Simulation Resultsand andAnalysis
Analysis
The anglesResults
and powers of 250 multipath signals are obtained through the RT simula-
tion, and Thethe
The angles
angles and powers
propagation
and paths
powers of
of250
of multipath
250the multipath
multipath signals are
signals
signals areobtained
inobtained through
the tunnel thethe
RTRT
are shown
through simula-
in sim-
Figure
tion,
ulation,and the
and propagation
the propagationpaths of
paths the
of multipath
the multipathsignals in the
signals intunnel
the
6. The polar coordinate form of the multipath signal angle is shown in Figure 7, in which are
tunnelshown
are in
shown Figurein
6. The
Figure
the polar coordinate
pitch6.departure form
The polar coordinate of the
form of
angle and pitch multipath signal
the multipath
arrival angle
angle aresignal is shown in
angle is shown
concentrated Figure
around 7,
in 90°,in
Figure which
while7, inthe
the pitch
which departure
the departure angleangle
pitch departure and pitchpitch
arrival angleangle
are concentrated around 90°,90
while the
◦ , while
horizontal angle and and horizontal arrival
arrival are are
angle concentrated
around 0° around
and 180°. The an-
horizontal
the horizontal departure
departureangle and horizontal arrival angle are around 0° ◦
and 180°. ◦
The an-
gular range of the arrival angle
angleand horizontal
is greater thanarrival angle are
the angular rangearound
of the0 departure
and 180 . The angle.
gular range
angular rangeofofthe
thearrival
arrivalangle
angleisisgreater
greaterthan
thanthetheangular
angular range
range of of the departure angle.
The angles and powers of the five largest energy paths in the simulation results are shown
Theangles
The anglesand
andpowers
powersof ofthe
thefive
fivelargest
largestenergy
energypaths
pathsin inthe
the simulation
simulationresults
resultsare
areshown
shown
in Table 3.
inTable
in Table3.3.

Figure 6. 6.Multipath
Figure MultipathSignal
SignalPropagation
Propagation Path.
Path.
Figure 6. Multipath Signal Propagation Path.

(a) (b) (c) (d)

(a) (b) (c) (d)
Figure7.7.Angle
Figure Angleand
andpower
powerof
ofmultipath
multipathsignals
signals(a)
(a)ZOD,
ZOD,(b)
(b)AOD,
AOD,(c)
(c)ZOA,
ZOA,and
and(d)
(d)AOA.
AOA.
Figure 7. Angle and power of multipath signals (a) ZOD, (b) AOD, (c) ZOA, and (d) AOA.
Table3.3.Angle
Table Angleand
andpower
powerofof55maximum
maximumenergy
energypaths.
paths.
Table 3. Angle and power of 5 maximum energy paths.
Number ofof φAOD (◦ )𝜽 θZOD (◦ ) 𝝓 φAOA (◦ ) ◦
Number
Number of
Multipath
𝝓𝑨𝑶𝑫 (°) 𝒁𝑶𝑫 (°) 𝑨𝑶𝑨 (°) 𝜽𝒁𝑶𝑨
θZOA(°)( ) 𝑷P𝒓r (𝐝𝐁𝐦)
(dBm)
Multipath 𝝓 (°) 𝜽 (°) 𝝓 (°) 𝜽 (°) 𝑷
1 𝑨𝑶𝑫 0 𝒁𝑶𝑫 90 𝑨𝑶𝑨−180 𝒁𝑶𝑨 90 𝒓 (𝐝𝐁𝐦)
−39.8109
Multipath
1 2 0 −2.86241 90 90 −180
−177.138 9090 −39.8109
−40.3604
12 3 0
−2.86241 2.86241 90 90 −180
−177.138
177.138 9090
90 −−39.8109
−40.3604
40.3604
23 4 2.86241−5.71059 90 90
−2.86241 174.289
−177.138
177.138 90
9090 −−40.3604
41.6511
−40.3604
5 5.71059 90 −174.289 90 −41.6511
34 2.86241
−5.71059 90 177.138
174.289 9090 −40.3604
−41.6511
45 5.71059
−5.71059 90 −174.289
174.289 9090 −41.6511
−41.6511
5 5.71059 90 −174.289 90 −41.6511
The transceiver signal mapping matrix extracted from the RT simulation results can
be represented as ansignal
The transceiver identity matrix with
mapping matrix 250extracted
rows andfrom
250 columns
the RT simulation results can
be represented as an identity matrix with 250 rows and 250 columns
Sensors 2022, 22, x FOR PEER REVIEW 9 of 1
Sensors 2022, 22, 4289 9 of 13

The transceiver signal mapping matrix1extracted

0 ⋯ from0 the RT simulation results can
be represented as an identity matrix 𝐴
with 0 1 ⋯ 0) columns
250 rows and 250
=( (19
⋮ ⋮ ⋱ ⋮
0· · · 0 0 ⋯ 1

1 0 250×250
0 1 ··· 0 
A= . . (19)
 
 .. .. . . .. 
In order to verify the spatial . . method
mirroring 
proposed in this paper and the ran
· · · 1after250
0 0 signals,
dom matching method for scattered randomly
×250 arranging the transmitted sig
nals,Infirst determine the mapping relationship of the reflected signals between the trans
order to verify the spatial mirroring method proposed in this paper and the random
ceivers according
matching method fortoscattered
the spatial mirroring
signals, principle,
after randomly and then
arranging theperform random
transmitted matching
signals,
first determine the mapping relationship of the reflected signals between the transceiversshown in
on the unmatched signals. The complete transceiver signal mapping matrix is
Figure 8. to the spatial mirroring principle, and then perform random matching on the
according
unmatched signals. The complete transceiver signal mapping matrix is shown in Figure 8.

Figure8.8.Receive
Figure Receiveand transmit
and signal
transmit mapping
signal matrix.
mapping matrix.
After obtaining the angle, power information, and mapping relationship of the trans-
mitting After
and obtaining the angle,
receiving signals, power information,
the normalized Doppler power and spectrum
mappingofrelationship of the trans
the transmitting
mitting
and and antennas
receiving receivingat signals,
different the normalized
moving speeds is Doppler power spectrum
obtained through the Dopplerofmodel,
the transmit
as shown in Figure 9. The normalized amplitude in the graph is defined as the ratio ofthe
ting and receiving antennas at different moving speeds is obtained through the Dopple
model,ofas
power shown
each singleinpath
Figure
to the9.total
Thepower
normalized amplitudeWhen
of the multipath. in the thegraph
movingis speeds
defined as the
of the of
ratio receiving
the powerand transmitting
of each single antennas
path toare the160 km/h
total and 80
power of km/h, the Doppler
the multipath. Whenshiftthe mov
of
ing speeds of the receiving and transmitting antennas are 160 km/h and 80 km/h,isthe Dop
the LoS signal in the RT simulation results is 2.074 KHz, and the Doppler spread
2.516 KHz. The Doppler shift of the LoS signal is 2.074 KHz and the Doppler spread is
pler shift of the LoS signal in the RT simulation results is 2.074 KHz, and the Dopple
2.542 KHz in the simulation results of the image space method and the random matching
spread is 2.516 KHz. The Doppler shift of the LoS signal is 2.074 KHz and the Dopple
method of scattered signals. When the receiving and transmitting antennas travel in the
spread
same is 2.542
direction at KHz in the
the same simulation
speed, results
the Doppler shiftofofthe
the image space
LoS signal method
in the and the random
RT simulation
is 0 Hz, and the Doppler spread is 4.702 KHz. The Doppler shift of the LoS signal is 0 Hzantenna
matching method of scattered signals. When the receiving and transmitting
travel
and the in the same
Doppler direction
spread is 4.783atKHz theinsame speed, the
the simulation Doppler
results of theshift
image of space
the LoS signal in the
method
RT simulation
and is 0 Hz, and
the random matching the Doppler
method of scatteredspread
signals. is The
4.702 KHz. The
Doppler Doppler
frequency shift of the LoS
component
to the right
signal is 0 of
Hztheand
LoSthe
path is the scattered
Doppler spreadsignal
is 4.783component.
KHz in the Thesimulation
simulation results
resultsofofthe
the imag
Doppler spread simulation model in this paper are consistent
space method and the random matching method of scattered signals. The with the simulation results of
Doppler fre
RT, which verifies the availability of the Doppler spread model. The simulation results also
quency component to the right of the LoS path is the scattered signal component. Th
show that in the T2T communication scenario, the Doppler spread of the signal is not only
simulation results of the Doppler spread simulation model in this paper are consisten
related to the antenna moving speed, but also to the departure and arrival angles. Due to
with
the the simulation
inconsistent angularresults
ranges ofof the
RT,arrival
which andverifies
departuretheangles
availability of the Doppler
of the scattered signals, spread
the extent of the Doppler frequency spreading to the right of the LoS path is increased. the Dop
model. The simulation results also show that in the T2T communication scenario,
pler spread of the signal is not only related to the antenna moving speed, but also to the
departure and arrival angles. Due to the inconsistent angular ranges of the arrival and
departure angles of the scattered signals, the extent of the Doppler frequency spreading
to the right of the LoS path is increased.
 Sensors 2022, 22, x FOR PEER REVIEW 10 of 13
Sensors 2022, 22, 4289 10 of 13

(a) (b)
Figure
Figure 9. Doppler
9. Doppler spread
spread simulation
simulation results
results of T2T
of T2T communication:
communication: (a) (a)
vt =v160
t=160 km/h, vr=80 km/h, (b)
km/h, vr = 80 km/h,
vt=160 km/h, vr=160 km/h.
(b) vt = 160 km/h, vr = 160 km/h.

4. 4. Physical
Physical Simulation
Simulation Model
Model ofof
T2TT2T Communication
Communication Channel
Channel in in Tunnel
Tunnel
In In a complex
a complex propagation
propagation environment,
environment, two two
kindskinds of fading
of fading channelschannels
will bewill be gener-
generated
dueated duedelay
to the to thespreading
delay spreadingeffect ofeffect of multipath
multipath channels,channels,
namely, the namely, the frequency
frequency flat fadingflat
fadingand
channel channel and the selective
the frequency frequency selective
fading channel.fading channel.effects
Multipath Multipath
causeeffects cause the
the amplitude
of amplitude
the received ofsignal to shift over
the received signal timeto when
shift overthe signal
time bandwidth
when the signal Bs is smaller
bandwidth 𝐵𝑠 is
than the
coherence bandwidth
smaller than Bc , but the
the coherence bandwidth 𝐵𝑐 , but does
signal spectrum not. Inspectrum
the signal this case,doesthe duration of the
not. In this case,
symbol Ts is greater
the duration of the than the maximum
symbol 𝑇𝑠 is greatertime delaythan the of multipaths,
τmaxmaximum time and
delay this𝜏𝑚𝑎𝑥
channel is
of mul-
called the flat fading channel. In a flat fading channel, the influence
tipaths, and this channel is called the flat fading channel. In a flat fading channel, the in- of the time delay on
thefluence
communication
of the timesystem
delay can be ignored.
on the communication Existingsystem
test results
can be show that the
ignored. multipath
Existing test re-
delay
sultsinshow
tunnelthat
scenarios is tens ofdelay
the multipath nanoseconds
in tunnel [11–14],
scenariosthat is
is,tens
τmax of < nanoseconds
Ts , so the bandpass
[11–14],
form
thatofis,Equation
𝜏𝑚𝑎𝑥 < (17) can
𝑇𝑠 , so thebebandpass
expressedform as the ofproduct
Equation of(17)
the bandpass transmitas
can be expressed signal x p (t)
the product
and the multiplicative spreading
of the bandpass transmit signal 𝑥𝑝 factor(𝑡) H ( t )
and the multiplicative spreading factor 𝐻(𝑡)
I 𝐼
( )
∑∑
 it
j2π f𝑗2𝜋𝑓 𝑖
R(t) 𝑅(𝑡)
= Re=xRe{𝑥p (t) H𝑝 (𝑡)𝐻(𝑡)}
(t) = Re = Re (t𝑝)(𝑡)
x p{𝑥 ci e 𝑐𝑖 𝑒 d 𝑑𝑡 } (20)(20)
i =1 𝑖=1
𝑗𝜑(𝑡)
(𝑡), 𝐻(𝑡) = |𝐻(𝑡)jϕ|𝑒(t) = H
where
where x p (𝑥t𝑝) (𝑡) t)(𝑡)
= x=p,I𝑥(𝑝,𝐼 + 𝑗𝑥𝑝,𝑄
+ jx p,Q ( t ), H ( t ) = | H ( t )| e
= (𝐻t𝐼)(𝑡)
I
+ 𝑗𝐻(𝑄t)(𝑡),
+ jH Q
then
, then Equation
Equation (20)
(20)
can be expressed
can be expressed as as
𝑅(𝑡) n= Re{𝑥𝑝 (𝑡)|𝐻(𝑡)|𝑒o𝑗𝜑(𝑡) } = 𝑥𝑝,𝐼 (𝑡)𝐻𝐼 (𝑡) − 𝑥𝑝,𝑄 (𝑡)𝐻𝑄 (𝑡) (21)
R(t) = Re x p (t)| H (t)|e jϕ(t) = x p,I (t) H I (t) − x p,Q (t) HQ (t) (21)
where 𝜑(𝑡) = tan−1 (𝐻𝐼 (𝑡)⁄𝐻𝑄 (𝑡)) is the phase shift caused by the Doppler shift of mul-
tipaths.
where ϕ(t) = tan−1 H I (t)/HQ (t) is the phase shift caused by the Doppler shift of multipaths.

TheThe circuitstructure
circuit structureofofthe theIQ IQ vector
vector phase
phase shifter
shifter[15,16]
[15,16]isisshown
showninin Figure
Figure10.10.
The
phase shifter circuit consists of a quadrature splitter (QS), a variable
The phase shifter circuit consists of a quadrature splitter (QS), a variable gain amplifier gain amplifier (VGA),
and aandquadrature combiner (QC). The input RF signal 𝑅𝐹𝑖𝑛 generates thethe in−phase
in−phasecom-
(VGA), a quadrature
⁄ √combiner (QC). The input RF signal RFin generates √
⁄
ponent 𝑉
component 𝐼VI = RF = 𝑅𝐹 √2 and the quadrature component
𝑖𝑛 in / 2 and the quadrature component VQ𝑄= jRFin𝑖𝑛 𝑉 = 𝑗 𝑅𝐹 √2 after passing
/ 2 after passing
through
through thethe quadrature
quadrature splitter.
splitter. AfterAfter 𝑉𝐼 and
VI and VQ pass 𝑉𝑄 pass
through through independent
independent VGAs,VGAs, they
they are
are summed
summed in thein the combiner,
combiner, and the andoutput
the output RFout 𝑅𝐹
signalsignal is a function
is a𝑜𝑢𝑡function of theof VGAthe VGA
gains gains
AI
and𝐴𝐼 Aand
Q 𝐴𝑄
RF𝑅𝐹 RFin
RFout = A I VI + AQ VQ = A I √ in +𝑖𝑛 jA Q √𝑅𝐹𝑖𝑛 (22)
𝑅𝐹𝑜𝑢𝑡 = 𝐴𝐼 𝑉𝐼 + 𝐴𝑄 𝑉𝑄 = 𝐴𝐼 2 + 𝑗𝐴𝑄 2 (22)
√2 √2
where the gain range of VGA is {−1,1}, the amplitude of the output signal RFout and
where the gain range of VGA is {−1,1}, the amplitude of the output signal − 𝑅𝐹 and
the
1 𝑜𝑢𝑡 /A
the input signal RFin remain unchanged, and phase difference ∆ϕ = tan −1 (𝐴 ⁄ A I ). Q .
input signal 𝑅𝐹𝑖𝑛 remain unchanged, and phase difference ∆𝜑 = tan √ 𝐼 𝐴𝑄 √Com-
Compared to (21) and (22), make x p = RFin , then x p,I = RFin / 2, x p,Q = jRFin / 2,
pared to (21) and (22), make 𝑥𝑝 = 𝑅𝐹𝑖𝑛 , then 𝑥𝑝,𝐼 = 𝑅𝐹𝑖𝑛 ⁄√2, 𝑥𝑝,𝑄 = 𝑗 𝑅𝐹𝑖𝑛 ⁄√2, 𝐴𝐼 = 𝐻𝐼 ,
A I = H I , AQ = − HQ . Therefore, the physical simulation of the channel can be theoretically
𝐴𝑄 = −𝐻𝑄. Therefore, the physical simulation of the channel can be theoretically realized
realized by using the program-controlled IQ vector phase shifter. The physical simulation
 Sensors
Sensors 2022,
2022, 22,22, x FORPEER
x FOR PEERREVIEW
REVIEW 11 of1113of 13

Sensors 2022, 22, 4289 11 of 13

by using
by using the
the program-controlled
program-controlled IQ IQvector
vectorphase shifter.
phase The
shifter. Thephysical simulation
physical model
simulation model
is shown in Figure 11. In Figure 11, the IQ vector phase shifter is a program-controlled
is shown in Figure 11. In Figure 11, the IQ vector phase shifter is a program-controlled
phase shifter
modelshifter
phase that
that can
is shown be
becontrolled
in Figure
can 11. In in
controlled inreal
realtime.
Figure 11, the IQ vector phase shifter is a program-
time.
controlled phase shifter that can be controlled in real time.
AI
AI
VI
VI VGA
VGA
RFin RFout
RFin QS QC RFout
QS QC
VQ
VQ VGA
VGA

AQ
AQ
Figure 10. The circuit structure of IQ vector phase shifter.
Figure10.
Figure Thecircuit
10.The circuit structure
structure of
of IQ
IQ vector
vectorphase
phaseshifter.
shifter.
𝑅(𝑡)
1 𝑐1 𝑅(𝑡)
𝑒 𝑗 2𝜋𝑓𝑑 𝑡
1 𝑐1 QC
𝑒 𝑗 2𝜋𝑓𝑑 𝑡
2
𝑐2 QC
𝑒 𝑗 2𝜋𝑓𝑑 𝑡 𝐻𝐼 (𝑡)
2
𝑐2
𝑒 𝑗 2𝜋𝑓𝑑 𝑡
VGA VGA
+ 𝐻 (𝑡)
𝑐3 𝐻𝑄𝐼(𝑡)
3
𝑒 𝑗 2𝜋𝑓𝑑 𝑡 +
VGA VGA

3
𝑐3 𝐻𝑄 (𝑡) 𝑥𝑝,𝐼 (𝑡) 𝑥𝑝,𝑄 (𝑡)
𝑒 𝑗 2𝜋𝑓
⋮ 𝑑𝑡
𝑐𝐼
𝐼 QS(𝑡 )
𝑥𝑝,𝐼 𝑥𝑝,𝑄 (𝑡)
𝑒 𝑗 2𝜋𝑓
⋮
𝑑𝑡
𝑐𝐼
𝐼
𝑒 𝑗 2𝜋𝑓𝑑 𝑡 QSPhase Shifter
(𝑡) Shifter
𝑥𝑝 Phase
Figure11.
Figure Physicalsimulation
11.Physical simulationmodel
modelof
ofthe
the (𝑡 )
𝑥channel.
channel.
𝑝

5. Conclusions
Figure 11. Physical simulation model of the channel.
5. Conclusions
In this paper, the Doppler shift and Doppler spread of T2T communication in a tunnel
In this paper, the Doppler shift and Doppler spread of T2T communication in a tun-
environment
5.nel
Conclusions are studied. Independent coordinate systems are established at the receiving
environment are studied. Independent coordinate systems are established at the re-
and transmitting antennas. According to the angle and amplitude characteristics of the
ceiving In this
andpaper, the Doppler
transmitting antennas.shift and Doppler
According to thespread
angle andof T2T communication
amplitude in a tun-
characteristics
receiving and transmitting signals, the Doppler spread caused by the movement of the
nel
of theenvironment
receiving and are studied.
transmitting Independent coordinate
signals, the Doppler systems
spread caused
receiving and transmitting antennas is analyzed. The use of the mapping matrix approach
are byestablished
the movement at the
of re-
the receiving
ceiving and and transmitting
transmitting antennas
antennas. is analyzed.
According to theThe
angleuseandof the mappingcharacteristics
amplitude matrix ap-
to solve the matching problem of the transmitting and receiving signals is presented as
proach
of anthe to solvesolution,
receiving
innovative the matching
and transmitting
and twoproblem
signals,
methods of for
theobtaining
transmitting
Doppler spread
the andcaused
mapping receivingby signals
matrix the is pre- of
aremovement
described
sented
theand as an innovative
receiving
verified. and
In order tosolution,
transmitting
verify theand
antennas
T2T two methods
is analyzed.
Doppler for obtaining
spread The use of
simulation the mapping
the
model mapping matrix
proposed in are
matrix
thisap-
described
proach
paper, the and
to solve verified.
the matching
RT method In order to verify
is used to problem
simulate the the
of T2TT2T Doppler
the transmitting
communication spread
and simulation
receiving
channel model
in thesignals pro-
tunnel, is
and pre-
posed
sented
the angle in
as this
an
and paper,
innovativethe RT
amplitude method
solution,
informationandisofused
two to simulate
themethods
transmittedforthe T2T
obtaining
and communication
received the mapping
signals channel
matrix are
are obtained.
inBythe
described tunnel,
comparing andand the angle
verified.
the Doppler and amplitude
In order
results to
ofverify information
the
the simulationT2T model
Dopplerofwith
the transmitted
spread
those and received
simulation modelthe
of WI simulation, pro-
signals
correctnessare obtained.
of the T2TBy comparing
Doppler the
spread Doppler
simulation results
modelof the
is
posed in this paper, the RT method is used to simulate the T2T communication channel simulation
proved. model
Based on with
the those
Doppler
of WI simulation, the correctness
in spread
the model,and
tunnel, a physical
the angle andof
channel the T2T Doppler
simulation
amplitude methodspread
informationusing ofsimulation
anthe model
IQtransmitted
vector phaseis proved.
shifter
and to
received
Based
complete on the Doppler spread model, a physical channel simulation method using an IQ
signals are T2T communication
obtained. By comparing in a the
tunnel environment
Doppler results ofis the
proposed,
simulation which can provide
model with those
vector
a phasefor
reference shifter
the to complete
physical T2T communication
channel simulation of 5Gin mmWave
a tunnel environment
T2T is proposed,
communication in a
of WI simulation, the correctness of the T2T Doppler spread simulation model is proved.
which
tunnel can provide a in
environment reference
the for the physical channel simulation of 5G mmWave T2T
future.
Based on the Doppler spread model, a physical channel simulation method using an IQ
communication in a tunnel environment in the future.
vector phase shifter to complete T2T communication in a tunnel environment is proposed,
Author Contributions: Conceptualization, P.Z. and G.Z.; data curation, P.Z. and X.W.; formal analysis,
which
Author can provide
Contributions:
P.Z.; funding aConceptualization,
reference
acquisition, G.Z. andfor
Y.J.;the physical
P.Z. channel
and G.Z.;
investigation, data
P.Z. simulation
curation,
and X.W.; ofX.W.;
P.Z. and 5G P.Z.
methodology, mmWave
formal
andanal-T2T
G.Z.;
communication
ysis, P.Z.; fundingin a tunnel
acquisition, environment
G.Z. and Y.J.; in the future.
investigation, P.Z. and X.W.; methodology, P.Z. and
project administration, P.Z. and K.Z.; resources, P.Z. and G.Z.; software, P.Z.; validation, P.Z.; writing—
G.Z.; project
original administration,
draft, P.Z. and and
P.Z.; writing—review K.Z.;editing,
resources,
P.Z.,P.Z. and
K.Z., G.Z.;
G.Z. andsoftware, P.Z.; validation,
Y.J. All authors have readP.Z.;
and
Author Contributions: Conceptualization, P.Z.
agreed to the published version of the manuscript. and G.Z.; data curation, P.Z. and X.W.; formal anal-
ysis, P.Z.; funding acquisition, G.Z. and Y.J.; investigation, P.Z. and X.W.; methodology, P.Z. and
G.Z.; project administration, P.Z. and K.Z.; resources, P.Z. and G.Z.; software, P.Z.; validation, P.Z.;
Sensors 2022, 22, 4289 12 of 13

Funding: This work was supported by National Natural Science Foundation of China (61871261) and
Natural Science Foundation of Shanghai (22ZR1422200).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Noor-A-Rahim, M.; Liu, Z.; Lee, H.; Khyam, M.; He, J.; Pesch, D.; Moessner, K.; Saad, W.; Poor, H. 6G for Vehicle-to-Everything
(V2X) Communications: Enabling Technologies, Challenges, and Opportunities. Proc. IEEE Inst. Electr. Electron. Eng. 2022, 1–23.
[CrossRef]
2. Alnasser, A.; Sun, H.; Jiang, J. Recommendation-Based Trust Model for Vehicle-to-Everything (V2X). IEEE Internet Things J. 2020,
7, 440–450. [CrossRef]
3. Naik, G.; Choudhury, B.; Park, J. IEEE 802.11bd & 5G NR V2X: Evolution of Radio Access Technologies for V2X Communications.
IEEE Access 2019, 7, 70169–70184.
4. Gyawali, S.; Xu, S.; Qian, Y.; Hu, R. Challenges and Solutions for Cellular Based V2X Communications. IEEE Commun. Surv. Tutor.
2021, 23, 222–255. [CrossRef]
5. Sehla, K.; Nguyen, T.; Pujolle, G.; Velloso, P. Resource Allocation Modes in C-V2X: From LTE-V2X to 5G-V2X. IEEE Internet Things
J. 2022, 9, 8291–8314. [CrossRef]
6. Tariq, F.; Khandaker, M.; Wong, K.; Imran, M.; Bennis, M.; Debbah, M. A Speculative Study on 6G. IEEE Wirel. Commun. 2020,
27, 118–125. [CrossRef]
7. Yang, P.; Xiao, Y.; Xiao, M.; Li, S. 6G Wireless Communications: Vision and Potential Techniques. IEEE Netw. 2019, 33, 70–75.
[CrossRef]
8. Akyildiz, I.; Kak, A.; Nie, S. 6G and Beyond: The Future of Wireless Communications Systems. IEEE Access 2020, 8, 133995–134030.
[CrossRef]
9. Saad, W.; Bennis, M.; Chen, M. A Vision of 6G Wireless Systems: Applications, Trends, Technologies, and Open Research Problems.
IEEE Netw. 2020, 34, 134–142. [CrossRef]
10. Khan, W.; Ihsan, A.; Nguyen, T.; Javed, M.; Ali, Z. NOMA-enabled Backscatter Communications for Green Transportation in
Automotive-Industry 5.0. IEEE Trans. Industr. Inform. 2022, in press. [CrossRef]
11. Khan, W.; Jameel, F.; Kumar, N.; Jäntti, R.; Guizani, M. Backscatter-Enabled Efficient V2X Communication with Non-Orthogonal
Multiple Access. IEEE Trans. Veh. Technol. 2021, 70, 1724–1735. [CrossRef]
12. ElHalawany, B.; Jameel, F.; Costa, D.; Dias, U.; Wu, K. Performance Analysis of Downlink NOMA Systems Over κ-µ Shadowed
Fading Channels. IEEE Trans. Veh. Technol. 2020, 69, 1046–1050. [CrossRef]
13. Zhang, F.; Wang, M.; Bao, X.; Liu, W. Centralized Resource Allocation and Distributed Power Control for NOMA-Integrated NR
V2X. IEEE Internet Things J. 2021, 8, 16522–16534. [CrossRef]
14. Zhang, F.; Wang, M.; Chen, S.; Shan, L.; Xiao, M.; Xu, M. Applying NOMA to NR V2X: A graph-based matching and cooperative
game approach. In Proceedings of the 2021 IEEE 93rd Vehicular Technology Conference (VTC2021-Spring), Helsinki, Finland,
25–28 April 2021; pp. 1–5.
15. Ihsan, A.; Chen, W.; Zhang, S.; Xu, S. Energy-Efficient NOMA Multicasting System for Beyond 5G Cellular V2X Communications
with Imperfect CSI. IEEE Trans. Intell. Transp. Syst. 2021. [CrossRef]
16. Do, D.; Anh Le, T.; Nguyen, N.; Li, X.; Rabie, K.M. Joint Impacts of Imperfect CSI and Imperfect SIC in Cognitive Radio-Assisted
NOMA-V2X Communications. IEEE Access 2020, 8, 128629–128645. [CrossRef]
17. Liao, Y.; Yang, G.; Liang, Y. Resource Allocation in NOMA-Enhanced Full-Duplex Symbiotic Radio Networks. IEEE Access 2020,
8, 22709–22720. [CrossRef]
18. Khan, W.; Jamshed, M.; Lagunas, E.; Chatzinotas, S.; Li, X.; Wang, H. Energy Efficiency Optimization for Backscatter Enhanced
NOMA Cooperative V2X Communications under Imperfect CSI. arXiv 2022, arXiv:2202.01592.
19. Liu, J.; Ahmed, M.; Mirza, M.; Khan, W.; Xu, D.; Li, J.; Aziz, A.; Han, Z. RL/DRL Meets Vehicular Task Offloading Using Edge
and Vehicular Cloudlet: A Survey. IEEE Internet Things J. 2022, 9, 8315–8338. [CrossRef]
20. Huang, C.; He, R.; Ai, B.; Molisch, A.; Lau, B.; Haneda, K.; Liu, B.; Wang, C.; Yang, M.; Oestges, C.; et al. Artificial Intelligence
Enabled Radio Propagation for Communications—Part I: Channel Characterization and Antenna-Channel Optimization. IEEE
Trans. Antennas Propag. 2022, in press. [CrossRef]
21. Huang, C.; He, R.; Ai, B.; Molisch, A.; Lau, B.; Haneda, K.; Liu, B.; Wang, C.; Yang, M.; Oestges, C.; et al. Artificial Intelligence
Enabled Radio Propagation for Communications—Part II: Scenario Identification and Channel Modeling. IEEE Trans. Antennas
Propag. 2022. early access. [CrossRef]
22. Guo, S.; Zhou, X. Robust Resource Allocation with Imperfect Channel Estimation in NOMA-Based Heterogeneous Vehicular
Networks. IEEE Trans. Commun. 2019, 67, 2321–2332. [CrossRef]
Sensors 2022, 22, 4289 13 of 13

23. Wang, J.; Jiang, C.; Zhang, H.; Ren, Y.; Chen, K.; Hanzo, L. Thirty Years of Machine Learning: The Road to Pareto-Optimal
Wireless Networks. IEEE Commun. Surv. Tutor. 2020, 22, 1472–1514. [CrossRef]
24. Wu, T.; Yin, X.; Lee, J. A novel power spectrum-based sequential tracker for time-variant radio propagation channel. IEEE Access
2020, 8, 151267–151278. [CrossRef]
25. Jiang, W.; Schotten, H. Deep Learning for Fading Channel Prediction. IEEE Open J. Commun. Soc. 2020, 1, 320–332. [CrossRef]
26. Ratnam, V.; Chen, H.; Pawar, S.; Zhang, B.; Zhang, C.; Kim, Y.; Lee, S.; Cho, M.; Yoon, S. FadeNet: Deep Learning-Based mm-Wave
Large-Scale Channel Fading Prediction and its Applications. IEEE Access 2021, 9, 3278–3290. [CrossRef]
27. Bai, L.; Xu, Q.; Wu, S.; Ventouras, S.; Goussetis, G. A Novel Atmosphere-Informed Data-Driven Predictive Channel Modeling
for B5G/6G Satellite-Terrestrial Wireless Communication Systems at Q-Band. IEEE Trans. Veh. Technol. 2020, 69, 14225–14237.
[CrossRef]
28. Zhao, X.; Du, F.; Geng, S.; Fu, Z.; Wang, Z.; Zhang, Y.; Zhou, Z.; Zhang, L.; Yang, L. Playback of 5G and beyond measured MIMO
channels by an ANN-based modeling and simulation framework. IEEE J. Sel. Areas Commun. 2020, 38, 1945–1954. [CrossRef]
29. Levie, R.; Yapar, Ç.; Kutyniok, G.; Caire, G. Radiounet: Fast radio map estimation with convolutional neural networks. IEEE
Trans. Wireless Commun. 2021, 20, 4001–4015. [CrossRef]
30. Zhang, J.; Shen, C.; Wang, S.; Bao, Y.; Yin, J. The design and performance test of train control data communication systems
based on train-to-train communication. In Proceedings of the 2020 Chinese Automation Congress (CAC), Shanghai, China, 6–8
November 2020; pp. 588–592.
31. Song, Z.; Ding, J.; Wang, Y.; Sun, X.; Sun, B.; Zhong, Z. Measurement-based doppler analysis in viaduct scenarios for railway. In
Proceedings of the 2021 13th International Symposium on Antennas, Propagation and EM Theory (ISAPE), Zhuhai, China, 1–4
December 2021; pp. 1–3.
32. Unterhuber, P.; Walter, M.; Fiebig, U.-C.; Kürner, T. Stochastic Channel Parameters for Train-to-Train Communications. IEEE
OJAP 2021, 2, 778–792. [CrossRef]
33. Kim, C.; Kim, J.; Hong, J.; Lim, J.; Chong, Y. Propagation characteristics of Urban and Highway Vehicle-to-Everything(V2X)
Channels at 5.9 GHz. In In Proceedings of the 2021 International Conference on Information and Communication Technology
Convergence (ICTC), Jeju Island, Korea, 20–22 October 2021; pp. 872–876.
34. Abd, H.; Al-Heety, A.; Mahmoud, M. LTE-V2I system performance under different stochastic channel model for highway
scenarios. In Proceedings of the 2021 International Conference on Communication & Information Technology (ICICT), Basrah,
Iraq, 5–6 June 2021; pp. 140–145.
35. Xiong, B.; Zhang, Z.; Zhang, J.; Jiang, H.; Dang, J.; Wu, L. Novel Multi-Mobility V2X Channel Model in the Presence of Randomly
Moving Clusters. IEEE Trans. Wirel. Commun. 2021, 20, 3180–3195. [CrossRef]
36. Gong, Z.; Li, C.; Jiang, F.; Win, M. Data-Aided Doppler Compensation for High-Speed Railway Communications Over mmWave
Bands. IEEE Trans. Wirel. Commun. 2021, 20, 520–534. [CrossRef]
37. Jiang, H.; Chen, Z.; Zhou, J.; Dang, J.; Wu, L. A General 3D Non-Stationary Wideband Twin-Cluster Channel Model for 5G V2V
Tunnel Communication Environments. IEEE Access 2019, 7, 137744–137751. [CrossRef]
38. Wang, S.; Zhao, H.; Zheng, G. Channel characteristics analysis of ultra-high mobility MIMO wireless system in tunnel scenario.
In Proceedings of the 2022 16th European Conference on Antennas and Propagation (EuCAP), Madrid, Spain, 27 March–1 April
2022; pp. 1–5.
39. Saleem, A.; He, Y. Investigation of massive MIMO channel spatial characteristics for indoor subway tunnel environment. In
Proceedings of the 2021 Computing, Communications and IoT Applications (ComComAp), Shenzhen, China, 26–28 November
2021; pp. 162–167.
40. Liu, Y.; Wang, C.-X.; Lopez, C.F.; Goussetis, G.; Yang, Y.; Karagiannidis, G.K. 3D Non-Stationary Wideband Tunnel Channel
Models for 5G High-Speed Train Wireless Communications. IEEE Trans. Intell. Transp. Syst. 2020, 21, 259–272. [CrossRef]
41. Wang, C.; Ji, W.; Zheng, G.; Saleem, A. Analysis of Propagation Characteristics for Various Subway Tunnel Scenarios at 28 GHz.
Int. J. Antennas Propag. 2021, 2021, 7666624. [CrossRef]
42. Wei, Y.; Sun, K.; Xu, D.; Gao, X.; Huang, S. Wideband RF photonic phase shifter with a full 360◦ phase tunable range. In
Proceedings of the 2017 International Topical Meeting on Microwave Photonics (MWP), Beijing, China, 23–26 December 2017;
pp. 1–3.
43. Xia, J.; Slim, B. Digitally Assisted 28 GHz Active Phase Shifter With 0.1 dB/0.5◦ RMS Magnitude/Phase Errors and Enhanced
Linearity. IEEE Trans. Circuits Syst. II Express Briefs 2019, 66, 914–918. [CrossRef]
44. Hu, Z.; Ji, W.; Zhao, H.; Zhai, X.; Saleem, A.; Zheng, G. Channel Measurement for Multiple Frequency Bands in Subway Tunnel
Scenario. Int. J. Antennas Propag. 2021, 2021, 9991758. [CrossRef]
45. Saleem, A.; Cui, H.; He, Y.; Boag, A. Channel Propagation Characteristics for Massive MIMO Systems in Tunnel Environment.
IEEE Antennas Propag. Mag. 2022, accepted.
46. Yin, X.; Liu, X.; Zheng, G.; Saleem, A.; Zhai, X. 28-GHz Channel Characterization for a Short Tunnel. IEEE Microw. Wirel. Compon.
Lett. 2018, 28, 1146–1148. [CrossRef]