Content uploaded by Sinem Coleri
Author content
All content in this area was uploaded by Sinem Coleri on Nov 16, 2023
Content may be subject to copyright.
Visible Light and mmWave Propagation Channel
Comparison for Vehicular Communications
Ali Uyrus, Bugra Turan, Ertugrul Basar and Sinem Coleri
Department of Electrical and Electronics Engineering, Koc University, Sariyer, Istanbul, Turkey, 34450
E-mail: [auyrus18, bturan14, ebasar, scoleri] @ku.edu.tr
Abstract—Future connected vehicles are expected to require
fast and reliable exchange of road information to increase
safety and enable cooperative driving. Currently, standardized
vehicular communication technologies aim to enable basic safety
message exchanges with limited bandwidth. Recently, alternative
technologies, based on millimeter-wave (mmWave) and visi-
ble light spectrum are proposed as complementary vehicle-to-
everything (V2X) communication schemes, provisioned to sup-
port future connected vehicles with high bandwidth and increased
security. However, the understanding of channel propagation
characteristics is the key to achieve reliability, due to higher
path loss compared to 5.8 GHz band. In this work, we compare
channel path loss characteristics of mmWave and vehicular visi-
ble light communication (VVLC) schemes to provide an overview
regarding technology selection in an indoor parking garage. Path
loss measurements are conducted with respect to various inter-
vehicular distances, receiver angles, nearby vehicle existence, and
lane occupation scenarios. Measurement results indicated path
loss of 21.47 dB for VVLC from 3 m to 20 m distances. Moreover,
path loss for mmWave 26.5 GHz and 38.5 GHz channels increased
12.5 dB, and 12.7 dB, respectively. Nearby vehicles are shown
to decrease path loss of 26.5 GHz and 38.5 GHz signals up to
9.78 dB, and 9.56 dB, respectively, whereas VVLC channel path
loss decreases 0.4 dB at the same scenario. Channel frequency
response (CFR) measurements indicated frequency flat behavior
of VVLC channels while mmWave channel exhibits frequency
selectivity induced dispersion due to parking garage structure.
Obstructed line-of-sight (OLoS) measurements further reveal
that blocking vehicle interrupts VVLC signals while selecting
a favorable antenna location leads up to 30 dB less path loss for
mmWave signals.
Index Terms—Vehicular communication, visible light commu-
nication, vehicle-to-vehicle communication, mmWave Vehicular
Communication, vehicle-to-infrastructure communication
I. INTRODUCTION
Next generation connected and autonomous vehicles
(CAVs) are expected to be equipped with vehicle-to-vehicle
(V2V) and vehicle-to-everything (V2X) communication sys-
tems providing fast and reliable data exchange with nearby
vehicles and infrastructure to increase road safety. Current
V2X communication technologies aim to support exchange
of basic safety messages including vehicle size, location,
heading, acceleration, speed, and weather data with limited
bandwidth and latency. Therefore, applications requiring high
bandwidth and low latency, such as raw sensor data sharing
and see through, are not readily supported with standardized
IEEE 802.11p and Cellular Vehicular to Everything Commu-
nication (C-V2X) V2V communication technologies.
Recently, complementary V2X communication technolo-
gies, utilizing millimeter-wave (mmWave) and visible light
communication (VLC) have been proposed to fulfill the re-
quirements of next generation CAVs. When compared to cur-
rent IEEE 802.11p and cellular based communications utiliz-
ing frequencies between 5.850 −5.925 GHz, mmWave based
schemes, enable utilization of higher bandwidths, whereas,
VLC provides secure, line-of-sight (LoS), and radio frequency
(RF) interference free communications. Therefore, both tech-
nologies can be considered as complementary V2V schemes
providing the additional advantages of high bandwidth and
secure communications for vehicular applications such as
platooning [1], [2]. However, considering the high propagation
path loss of both technologies, large scale fading character-
ization at practical V2X communication usage scenarios is
key for reliable system designs. Currently, channel models for
V2X communications for mmWave frequencies, 30 GHz and
63 GHz, targeting vehicle-to-base station (V2B), pedestrian-to-
base station (P2B), and base station-to-road side unit (B2R)
including urban and highway scenarios are detailed in [3],
whereas, mmWave channel models for V2V and vehicle-to-
pedestrian (V2P) communication applications are not spec-
ified. The mmWave vehicular communication schemes are
explored to date for their characteristics at different frequen-
cies [4]–[7]. In [4], the path loss characteristics of 60 GHz
radio waves are investigated on both metropolitan highways
and regular roads; the authors of [5] studied V2V mmWave
channel with time domain characteristics for antennas located
near a vehicle’s headlight. Ben-Dor et.al [6] explored 60
GHz outdoor channel characteristics considering in-vehicle
antennas, whereas [7] evaluated 38 GHz and 60 GHz V2V
channel characteristics at a university campus. Moreover, [8]
characterized the effects of vehicle blockage through measure-
ments at 6.75 GHz, 30 GHz, 60 GHz, and 73 GHz in both
urban and highway scenarios. The authors in [9] characterized
28 GHz channel with respect to vehicle blockage in an open
parking lot.
Vehicular visible light communication (VVLC) channel
characteristics are addressed only in a few studies. Specif-
ically, [10] conducted VVLC Channel Frequency Response
(CFR) measurement-based channel modeling, [11] proposed
geometry-based stochastic model for VVLC, [12] investigated
VVLC received signal power variations with respect to vehicle
movements, and [13] provided an analytical expression to
compute received signal strength (RSS) of VVLC channel.
Figure 1: Experimental Setup
In this work, channel path loss comparisons between
mmWave and VVLC bands targeting direct V2V or V2P com-
munications in an indoor environment are performed. To the
best of our knowledge, this is the first study in which mmWave
and VVLC channel comparisons are performed to provide an
overview on the vehicular communication technology selection
in an environment comprising a parking garage. We utilize
26.5 GHz and 38.5 GHz frequencies for mmWave bands
considering vehicle user equipment using 5G-New Radio
Frequency Range 2, n257 and n260 bands [14]. Time domain
channel path loss measurements are conducted in the −1level
of a 2 story underground parking garage with up to 28 m inter-
vehicular distances, to investigate nearby vehicles, building
columns, floor slabs, and wall effects on signal propagation.
Moreover, CFR measurements of the technologies under con-
sideration are conducted to further evaluate the received power
variations with respect to the building structure and nearby
vehicles. Finally, vehicle blockage effects for mmWave signals
in the target underground parking garage are characterized
through different cross sections of a production vehicle at 11
different antenna locations.
The rest of the paper is organized as follows. In Section II,
we briefly describe the experimental setups utilized to mea-
sure mmWave channel path loss, VVLC RSS, mmWave and
VVLC channel frequency responses. In Section III, we pro-
vide measurement results and compare mmWave and VVLC
channel path loss characteristics including vehicle blockage
loss for mmWave channel targeting indoor scenarios. Finally,
we conclude the paper in Section IV.
II. EX PE RI ME NTAL SE TU P
Experiments are conducted to understand the channel loss
propagation differences between mmWave and VVLC chan-
nels in an indoor environment. Transmitter and receiver
front ends of both technologies (i.e., Light Emitting Diode
(LED)-avalanche photodiode (APD) for VVLC, antennas for
mmWave bands) are located at the same height of 70 cm, to
provide fair comparison. For direct LoS measurements, both
antenna boresights and optical transmitter-receiver front-ends
are aligned. Our experimental setup is depicted in Fig. 1 and
detailed in following subsections.
A. mmWave Channel Measurement Setup
MmWave signals at 26.5 GHz and 38.5 GHz are generated
by Keysight E8267D Vector Signal Generator for channel path
loss characterization. Generated signals are transmitted and
received through QPar QMS-00361 wideband horn antennas,
with 22◦to 41◦half power beam width, 14.3 dBi and 14.7 dBi
gain at 26.5 and 38.5 GHz, respectively. Received signals
are fed into Keysight UXA N0940B Vector Signal Analyzer,
where received signal power levels are averaged and recorded
for analysis.
CFR measurements for mmWave channel are conducted
with Rohde Schwarz ZVA-67 vector network analyzer (VNA),
where antennas are connected to two ports of VNA with
Pasternack PE361 RF cables.
B. VVLC Channel Measurement Setup
A single tone sinusoidal signal at 1 MHz is generated
through Rigol DG4000 Arbitrary Waveform Generator, and fed
into LED driver block. LED driver block includes a cascaded
set of Mini-Circuits ZHL-6A+ and ZFL-500LN amplifiers
with Mini-Circuits ZFBT-4R2GW+ bias tee. Amplified and
DC-Biased signals are fed into a 2017 Ford Mondeo multi-
beam LED headlight (see Fig.3) from the output of the driver
block.
At the receiver side, optical signals were captured through
Hamamatsu C5331-03 APD and 20 dB amplified with Mini-
Circuits ZFL-1000LN+ amplifier. Received signal amplitudes
are measured with Keysight DSOX3034A Oscilloscope for
further analysis.
(a) (b)
(c) (d)
(e)
Figure 2: (a) Transmitter Setup, (b) Receiver Setup (c) and (d)
obstructed line-of-sight (OLoS) scenarios for longitudinally
and laterally blocking vehicles (e) OLoS Antenna Locations
Table I: VVLC Front End Specifications
Parameter Value
Transmitter
Headlight 3-dB Bandwidth 2 MHz
DC Bias Voltage 24 V
Driver Block Input Signal Amplitude 63 mVpp
Driver Block Total Gain 47 dB
Driver Block Output Signal Amplitude 14.1 Vpp
LED Input Signal Amplitude 5.6 Vpp
LED Optical Transmitted Power -6.72 dBm
Transmitter Height 0.7 cm
Receiver
Avalanche Photodiode Module Hammamatsu C5331-03
APD Active Area 1 mm
APD 3 dB Frequency Bandwidth 4kHz to 100 MHz
APD Spectral Response Range 400 to 1000 nm
APD Peak Sensitivity Wavelength 800 nm
Amplifier Mini-Circuits ZFL-1000LN+
Amplifier Gain 20 dB
Amplifier Frequency Range 0.1 to 1000 MHz
Receiver Height 0.7 m
Headlight Radiation Pattern
5 10 15 20
Longitidunal Distance (m)
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
Lateral Distance (m)
-40 dBm
-30 dBm
-20 dBm
-10 dBm
Figure 3: Headlight Radiation Pattern at 70cm Height
12345678910
Frequency (MHz)
-5
0
5
10
15
S21 (dB)
3 dB Bandwidth @ 2 MHz
Figure 4: Headlight Frequency Response
Anritsu MS-2026C VNA is utilized for VVLC CFR mea-
surements. LED Headlight is connected to the Port 1 of VNA
through the driver block, where the amplified output of the
APD through ZFL-1000LN+ amplifier is connected to port
2 of the VNA. VVLC CFR measurements are conducted
between 200 kHz and 10 MHz due to limited modulation
bandwidth of vehicle headlight power LEDs (see Fig.4). Table
I depicts VVLC measurement setup specifications.
Table II: CFR Measurement Parameters
Parameters VVLC mmWave
Frequency Range 200 kHz - 10 MHz 25 GHz- 40 GHz
Number of Points 4001 4001
Bandwidth 100 kHz 100 kHz
Cable Length 7 m 7 m
Tx Power -25 dBm 8 dBm
C. Measurement Scenarios
1) Path Loss Measurements: For path loss measurements,
one vehicle is equipped as mmWave and VVLC transmitter
(Fiat Tipo, See Fig. 2(a)), while the other vehicle, Ford
Courier, depicted in Fig. 2(b), is utilized as mmWave and
VVLC receiver vehicle. Measurement scenarios are further
detailed as follows.
a) One Lane Scenario: Vehicles are moved back-to-back
at single lane in the parking garage to distinguish building
structure effects such as columns (see Fig. 2(c) and Fig. 2(d)).
Direct LoS and 30◦tilted receiver cases are considered for
empty parking garage and full parking garage with nearby
vehicles. Considering common 30◦tilt angle of production
vehicle rear view cameras, optical receiver is oriented at
30◦angle to observe the inclination angle effects on VVLC
link. On the other hand, nearby vehicle scenario (NBV) is
explored to investigate tunneling effect of nearby vehicles in
an indoor environment.
b) Two lane Scenario: Two lane scenarios are considered
to compare both distance and angle dependent path loss of
mmWave and VVLC channels for vehicles moving in opposite
traffic directions.Two lane scenarios emulate weak received
signal regions, as mmWave receiver antennas were out of
boresight and VVLC transmitter LED emits light with less
intensity due to its radiation pattern.
c) Blockage Scenario: Vehicle blockage induced path
loss, OLoS, is measured for mmWave channel, where a third
vehicle (Fiat Linea) is located between transmitter and receiver
vehicles apart from 15m. For OLoS scenarios, 2 different
cases, a blocking vehicle moving in the same direction with
transmitter −receiver vehicles, longitudinal OLoS (Fig. 2(c))
and a vehicle located between transmitter −receiver vehicles,
emulating a vehicle manoeuvring to leave from parking spot,
lateral OLoS (Fig. 2(d)) are considered. Transmitter antenna
is located on the rear bumper in 11 different locations at same
height as shown in Fig. 2(e). All antenna locations are selected
to be in the main beam of the receiver antenna, with negligible
antenna gain differences. Thereby, mmWave channel path loss
dependency on antenna locations are explored.
2) CFR Measurements: VVLC and mmWave CFR mea-
surements are conducted to evaluate channel frequency char-
acteristics with respect to varying locations covering different
regions of the parking garage with pure walls and building
columns. CFR measurements of mmWave and VVLC are exe-
cuted simultaneously, where mmWave and VVLC transmitter
−receiver front ends are located on two rolling platforms,
as shown in Fig. 2(a). Both rolling platforms are moved
in the parking structure, encapsulating one lane direct LoS
path loss scenarios. Measured CFR (S21) parameters are
recorded at static inter-vehicular distances from 1 m to 7 m
with 1 m intervals for each measurement run due to cable
limitations. Four measurement runs are executed for a total
distance of 28 m, covering all measurement points from direct
LoS scenarios of path loss measurements. CFR measurement
parameters are listed in Table II.
Table III: mmWave Channel Path Loss Parameters
Frequency Angle NBV n µ σs
26.5 GHz
0No 1.49 0.156 3.76
Yes 1.31 -0.22 3.48
30 No 1.51 0.179 4.611
Yes 1.477 0.040 2.20
38.5 GHz
0No 1.581 0.075 4.23
Yes 1.06 -0.35 2.18
30 No 1.82 -0.16 3.76
Yes 1.78 -0.060 2.89
III. RES ULTS A ND DISCUSSION
A. Path Loss and Received Power
For 26.5 GHz and 38.5 GHz mmWave measurements, path
loss is extracted, whereas received electrical signal power is
recorded for VVLC, as exact transmit power, APD gains,
optical-to-electrical and electrical-to-optical conversion effi-
ciencies are not readily known. Path loss related parameters
for mmWave channel, PL(d), are extracted from,
PL(d) = PL(d0) + 10nlog10 d
d0+Xσ(1)
where, PL(d0)is the path loss at the reference distance d0
of 1 m, nis the path loss exponent, dis the inter-vehicular
distance at each measurement point, and Xσis a zero mean
random variable which has Gaussian distribution with standard
deviation σs. Linear regression with minimized mean squared
error (MMSE) is utilized in order to obtain nand σs[15]
for mmWave channel, and parameters are listed in Table III.
Both nand σsvalues are observed to be consistent with
the reported indoor mmWave channel path loss studies [16],
[17], indicating wave guiding effects of indoor environment.
Fig. 5 depicts that nearby vehicles, decrease path loss, while
increasing tunneling effect. Furthermore, path loss is observed
to decrease for one lane 30◦tilted receiver antenna scenario,
likely due to reflections from parking garage floor slabs.
Path loss for mmWave channel and RSS for VVLC channels
with respect to nearby vehicle and receiver angle variations
for 5 m and 20 m are summarized in Table IV. Nearby
vehicles together with 30◦receiver angle are demonstrated
to decrease path loss up to 10.69 dB for mmWave channels,
clearly indicating waveguide effect sourced from building floor
and vehicles. VVLC RSS is observed to consistently decrease
with tilted receiver, and increase with nearby vehicles.
Fig. 6 shows the cumulative density function (CDF) of
mmWave channel path loss, ranged from -118 dB to -70.7
dB fitting to Gaussian distribution with -96.5 dB mean and
8.66 dB standard deviation.
Fig. 7 depicts the VVLC RSS for one lane scenarios, where
the mean received power decreases by 2 dB with 30◦tilted
receiver, and nearby vehicles are observed to have very subtle
incremental effect (0.4 dB) on VVLC RSS. Inter-vehicular
distance (div), dependent path loss equations are derived for
all scenarios, where relevant coefficients (a1, a2, a3, a4) are
listed in Table V.
H0(div) = a1ea2div +a3ea4div (2)
0 5 10 15 20 25
Inter-vehicular Distance (m)
60
70
80
90
100
110
120
Path Loss (dB)
26.5 GHz, Direct LOS
26.5 GHz, Direct LOS, NBV
26.5 GHz, 30°
26.5 GHz, 30°, NBV
38.5 GHz, Direct LOS
38.5 GHz, Direct LOS, NBV
38.5 GHz, 30°
38.5 GHz, 30°, NBV
Figure 5: mmWave Channel Path Loss for 1-lane Scenario
70 75 80 85 90 95 100 105 110 115 120
x (dB)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
F(x) = P(PL<x)
26.5 GHz, Direct LOS
26.5 GHz, Direct LOS, NBV
26.5 GHz, 30°
26.5 GHz, 30°, NBV
38.5 GHz, Direct LOS
38.5 GHz, Direct LOS, NBV
38.5 GHz, 30°
38.5 GHz, 30°, NBV
N( = -96.5, = 8.663)
Figure 6: mmWave Channel Path Loss Empirical Cumula-
tive Distribution Function Compared to Normal Distribution
(N(µ, σ))
0 5 10 15 20 25
Inter-vehicular Distance (m)
-90
-80
-70
-60
-50
-40
-30
Received Signal Strength (dBm)
Direct LOS
Direct LOS, NBV
30°
30°, NBV
Figure 7: VLC Received Signal Strength for 1-lane Scenario
Table IV: Fitted Path Loss Comparison at 5 m and 20 m (dB)
Frequency Distance NBV 0◦30◦
26.5 GHz
5 m No 92.41 88.50
Yes 82.88 78.01
20 m No 101.9 97.59
Yes 94 86.90
38.5 GHz
5 m No 95.97 90.37
Yes 86.19 80.29
20 m No 104.9 101.3
Yes 96.09 91.93
VVLC (RSS (dBm))
5 m No -58.50 -60.89
Yes -58.06 -59.72
20 m No -79.97 -82.12
Yes -79.79 -82.06
Table V: VVLC Path Loss Coefficients
Scenario a1a2a3a4
Direct LoS -82.73 0.0023 39.33 -0.0892
Direct LoS w/NBV -73.62 0.0056 33.9 -0.1307
30◦-71.13 0.0077 32.84 -0.1817
30◦w/NBV -69.92 0.0084 36.5 -0.2036
2-lane Direct LOS, d > 10m-27.97 0.0097 -0.0021 0.3130
2-lane 30◦,d > 10m-28.11 0.0126 -0.00052 0.3430
Measurement results from opposite traffic two lane scenar-
ios are demonstrated for mmWave bands in Fig. 8(a) and
VVLC in Fig. 8(b). VVLC two lane RSS measurements
indicate that headlight radiation pattern has substantial effect
on VVLC angular dependency, as same received signal power
is obtained at 1 m and 25 m distances. On the other hand, 26.5
GHz and 38.5 GHz path loss measurement results demonstrate
that, at the out of boresight gain of antennas, nearby and sur-
face reflections become dominant, as path loss decreases with
30◦tilted receivers when compared to direct LoS receivers.
B. Channel Frequency Response
CFR measurement results for mmWave bands are depicted
on Fig. 9, where parking structure columns and nearby ve-
hicle effects at various inter-vehicular distances are clearly
demonstrated. However, VVLC CFR measurements, denoted
in Fig. 10, imply frequency flat characteristics of VVLC
channel, mainly shaped through LED frequency response,
and channel gain mainly depends on inter-vehicular distances.
Nearby vehicles are shown to increase channel gain while no
substantial effects of parking structure columns are observed
0 5 10 15 20 25 30
Inter-vehicular Distance (m)
80
85
90
95
100
105
110
115
120
125
130
Path Loss (dB)
26.5 GHz, Direct LOS
26.5 GHz, 30°
38.5 GHz, Direct LOS
38.5 GHz, 30°
(a)
0 5 10 15 20 25 30
Inter-vehicular Distance (m)
-65
-60
-55
-50
-45
-40
-35
-30
-25
Received Signal Strength (dBm)
VLC, Direct LOS
VLC, 30°
(b)
Figure 8: (a) mmWave path loss and (b) VLC received signal
strength for 2-lane Scenario
25 30 35 40
Frequency (GHz)
-70
-60
-50
-40
-30
S21 (dB)
7.3 m
25 30 35 40
Frequency (GHz)
-70
-60
-50
-40
-30
S21 (dB)
6.3 m
25 30 35 40
Frequency (GHz)
-70
-60
-50
-40
-30
S21 (dB)
5.3 m
25 30 35 40
Frequency (GHz)
-70
-60
-50
-40
-30
S21 (dB)
4.3 m
25 30 35 40
Frequency (GHz)
-70
-60
-50
-40
-30
S21 (dB)
3.3 m
25 30 35 40
Frequency (GHz)
-70
-60
-50
-40
-30
S21 (dB)
2.3 m
1.3 m Free Space
Nearby Vehicle
No Vehicle, No Column
Column
Figure 9: mmWave Channel Frequency Response
Frequency (MHz)
1 2 3 4 5 6 7 8 9 10
S21 (dB)
-60
-55
-50
-45
-40
-35
Nearby Vehicle
7 m
6 m
5 m
4 m
3 m
2 m
Frequency (MHz)
1 2 3 4 5 6 7 8 9 10
S21 (dB)
-60
-50
-40
No Vehicle, No Column
Frequency (MHz)
1 2 3 4 5 6 7 8 9 10
S21 (dB)
-60
-50
-40
Nearby Column
Figure 10: VVLC Channel Frequency Response
with VVLC CFR. Therefore, continuous channel estimation
is important for mmWave channels, considering the building
and nearby vehicle effects on the channel frequency domain
response. However, as VVLC channels exhibit similar fre-
quency response characteristics for various building structure
and nearby vehicle existence, channel estimation periods can
be extended, yielding less overhead, hence, lower latency.
C. OLoS with Vehicle Blockage
Vehicle blockage effects on V2V communication signal
loss is investigated, where a third vehicle is located between
transmitter and receiver vehicles. VVLC transmissions were
not captured at the receiver with the blocking vehicle despite
the high sensitivity of the receiver. Therefore, vehicle blockage
can be regarded to interrupt VVLC transmissions.
1234567891011
Antenna Location Number
90
100
110
120
130
140
150
Path Loss (dB)
26.5 GHz, No Vehicle
38.5 GHz, No Vehicle
26.5 GHz, Lateral Vehicle
38.5 GHz, Lateral Vehicle
26.5 GHz, Longitudinal Vehicle
38.5 GHz, Longitudinal Vehicle
Figure 11: OLoS Path Loss
Table VI: OLoS Measurement Statistics
Frequency Scenario µ σ
26.5 GHz
No Vehicle 100.67 5.63
Lateral 106.81 6.95
Longitidunal 109.95 4.62
38.5 GHz
No Vehicle 103.81 6.53
Lateral 120.09 8.33
Longitidunal 115.21 1.30
Fig. 11 displays the path loss for one lane scenarios with
a blocking vehicle, where the transmitter antenna location
is varied at same height. Mean path loss µ, for each OLoS
scenario is depicted with solid lines in Fig. 11 and values are
listed in Table VI along with standard deviations σ. Results
indicate that path loss fluctuates up to 30 dB (38.5 GHz) with
laterally located blocking vehicle, even though longitudinal
vehicle blockage leads maximum 15.5 dB path loss fluctuation.
IV. CONCLUSION
In this paper, we have empirically compared propaga-
tion characteristics of mmWave and VLC technologies for
V2V communication scenarios in an indoor parking garage
environment with empty and full cases. Our measurement
results indicate that both mmWave and VVLC signals take
advantage of full parking scenario case with decreased path
loss, due to nearby vehicles. Moreover, 30◦tilted receiver
has been demonstrated to decrease path loss for mmWave
signals, while decreasing RSS for VVLC. CFR measurements
have indicated that, VVLC channel exhibits frequency flat
behavior, while mmWave CFR varies with respect to building
columns. Considering 12.7 dB direct LoS path loss increase
for mmWave channel at 38.5 GHz, and 21.47 dB path loss
for VVLC from 3 m to 20 m inter-vehicular distances, with
penetration capability through blocking vehicles, mmWave
communications can be regarded more reliable for indoor
parking garages, with symmetrical walls. On the other hand,
VVLC can be utilized more reliably for more complex parking
garages with additional building columns. Regarding LoS
optical propagation characteristics of VVLC, and directional
antenna usage for mmWave bands, two lane scenario mea-
surements depicted that angular pattern of vehicle LED is key
for the performance of VVLC, whereas mmWave signals take
advantage of floor slab reflections at 30◦tilted receiver.
CFR measurement results imply that VVLC channel ex-
hibits frequency flat characteristics, despite the building struc-
ture with columns and walls, whereas frequency selective
characteristics of mmWave signals play an important role for
path loss. OLoS measurement results indicate that VVLC
signals are completely shadowed through blocking vehicles,
whereas lateral and longitudinal blocking vehicles lead differ-
ent path loss for mmWave V2V communications. Furthermore,
different antenna locations at same height is demonstrated to
have substantial effect on mmWave channel direct and OLoS
channel path loss, which can be further considered for diversity
schemes.
V. ACKNOW LE DG EM EN T
The authors acknowledge the support of Ford Otosan, 5G
Valley Open Test Site (5G VATS) and Spark Measurement
Technologies. E.Basar acknowledges the support of Turkish
Academy of Sciences GEBIP programme.
REFERENCES
[1] J. Choi, V. Va, N. Gonzalez-Prelcic, R. Daniels, C. R. Bhat, and R. W.
Heath, “Millimeter-wave vehicular communication to support massive
automotive sensing,” IEEE Commun. Mag., vol. 54, no. 12, pp. 160–
167, 2016.
[2] S. Ucar, S. C. Ergen, and O. Ozkasap, “IEEE 802.11 p and visible light
hybrid communication based secure autonomous platoon,” IEEE Trans.
Veh. Tech., vol. 67, no. 9, pp. 8667–8681, 2018.
[3] “3GPP TR 37.885 v15.3.0 Study on evaluation methodology of new
Vehicle-to-Everything V2X use cases for LTE and NR,” Jun. 2019.
[4] S. Takahashi, A. Kato, K. Sato, and M. Fujise, “Distance dependence of
path loss for millimeter wave inter-vehicle communications,” in Proc.
IEEE Veh. Tech. Conf. (VTC-Fall) (IEEE Cat. No. 03CH37484), vol. 1,
pp. 26–30, IEEE, 2003.
[5] E. Kampert, P. A. Jennings, and M. D. Higgins, “Investigating the V2V
millimeter-wave channel near a vehicular headlight in an engine bay,”
IEEE Commun. Lett., vol. 22, no. 7, pp. 1506–1509, 2018.
[6] E. Ben-Dor, T. S. Rappaport, Y. Qiao, and S. J. Lauffenburger,
“Millimeter-wave 60 GHz outdoor and vehicle AOA propagation mea-
surements using a broadband channel sounder,” in Proc. IEEE Glob.
Telecommun. Conf. (GLOBECOM, pp. 1–6, 2011.
[7] M. G. Snchez, M. P. Tboas, and E. L. Cid, “Millimeter wave radio
channel characterization for 5G vehicle-to-vehicle communications,”
Measurement, vol. 95, pp. 223 −− 229, 2017.
[8] M. Boban, D. Dupleich, N. Iqbal, J. Luo, C. Schneider, R. M¨
uller, Z. Yu,
D. Steer, T. J ¨
ams¨
a, J. Li, et al., “Multi-band vehicle-to-vehicle channel
characterization in the presence of vehicle blockage,” IEEE Access,
vol. 7, pp. 9724–9735, 2019.
[9] J.-J. Park, J. Lee, J. Liang, K.-W. Kim, K.-C. Lee, and M.-D. Kim,
“Millimeter wave vehicular blockage characteristics based on 28 GHz
measurements,” in Proc.IEEE Veh. Tech. Conf. (VTC-Fall), pp. 1–5,
IEEE, 2017.
[10] B. Turan, G. Gurbilek, A. Uyrus, and S. C. Ergen, “Vehicular VLC
frequency domain channel sounding and characterization,” in Proc. IEEE
Veh.Netw.Conf. (VNC), pp. 1–8, Dec. 2018.
[11] A. Al-Kinani, J. Sun, C.-X. Wang, W. Zhang, X. Ge, and H. Haas,
“A 2-D non-stationary GBSM for vehicular visible light communication
channels,” IEEE Trans. Wireless Commun., vol. 17, no. 12, pp. 7981–
7992, 2018.
[12] A. Chen, H. Wu, Y. Wei, and H. Tsai, “Time variation in vehicle-
to-vehicle visible light communication channels,” in Proc. IEEE
Veh.Netw.Conf. (VNC), pp. 1–8, Dec. 2016.
[13] A. Memedi, H.-M. Tsai, and F. Dressler, “Impact of realistic light radia-
tion pattern on vehicular visible light communication,” in GLOBECOM
2017-2017 IEEE Global Communications Conference, pp. 1–6, IEEE,
2017.
[14] “3GPP TS 38.101-2 User Equipment (UE) radio transmission and
reception; Part 2: Range 2 Standalone Release 16,” Jun. 2019.
[15] T. Rappaport, Wireless Communications: Principles and Practice. Upper
Saddle River, NJ, USA: Prentice Hall PTR, 2nd ed., 2001.
[16] G. R. Maccartney, T. S. Rappaport, S. Sun, and S. Deng, “Indoor
office wideband millimeter-wave propagation measurements and channel
models at 28 and 73 GHz for ultra-dense 5G wireless networks,” IEEE
Access, vol. 3, pp. 2388–2424, 2015.
[17] A. M. Al-Samman, T. Abd Rahman, and M. H. Azmi, “Indoor corridor
wideband radio propagation measurements and channel models for 5G
millimeter wave wireless communications at 19 GHz, 28 GHz, and 38
GHz Bands,” Wireless Commun. and Mobile Comput., vol. 2018, 2018.
[18] T. S. Rappaport, G. R. MacCartney, M. K. Samimi, and S. Sun, “Wide-
band millimeter-wave propagation measurements and channel models for
future wireless communication system design,” IEEE Trans. Commun.,
vol. 63, pp. 3029–3056, Sep. 2015.
