Content uploaded by Aniruddha Chandra
Author content
All content in this area was uploaded by Aniruddha Chandra on Mar 24, 2020
Content may be subject to copyright.
Multipath Propagation Analysis for Vehicle-to-
Infrastructure Communication at 60 GHz
Ales Prokes*, Jiri Blumenstein*, Josef Vychodil*, Tomas Mikulasek*, Roman Marsalek*,
Erich Zӧchmann†‡*, Herbert Groll‡, Christoph F. Mecklenbräuker‡,
Thomas Zemenx, Aniruddha Chandra⸸, Hussein Hammoud††, Andreas F. Molisch††
*Department of Radio Electronics, Brno University of Technology, Czech Republic, prokes@vutbr.cz
† Christian Doppler Laboratory for Dependable Wireless Connectivity for the Society in Motion
‡ Institute of Telecommunications, TU Wien, Austria
x Safety and Security Department, Austrian Institute of Technology, Austria
⸸National Institute of Technology, Department of Electronics and Communication Engineering, India
†† Wireless Devices and Systems Group, University of Southern California, USA
Abstract— The paper deals with an analysis of multipath
propagation environment in the 60 GHz band using a pseudo-
random binary sequence-based time-domain channel sounder
with 8 GHz bandwidth. The main goal of this work is to analyze
the multipath components (MPCs) propagation between a moving
car carrying a transmitter with an omnidirectional antenna and a
fixed receiver situated in a building equipped with a manually
steered directional horn antenna. The paper briefly presents the
time dependence of the dominant MPC magnitudes, shows the
effect of the surrounding vegetation on the RMS delay spread and
signal attenuation, and statistically evaluates the reflective
properties of the road which creates the dominant reflected
component. To understand how the MPCs propagate through the
channel we measured and analyzed the power and the RMS delay
spread distributions in the static environment surrounding the car
using an automated measuring system with a controlled receiver
antenna tracking system. We give some examples of how the MPC
magnitudes change during the antenna tracking and demonstrate
that a building and a few cars parked close to the measuring car
create a lot of MPCs detectable by the setup with a dynamic range
of about 50 dB.
Keywords—millimeter wave, channel measurement, channel
sounder, channel impulse response, delay spread.
I. INTRODUCTION
Automated driving systems of future vehicles will be much
more reliable and safer when employing inter-vehicular
communication [1]. Being able to disseminate information on
safety-critical events detected or caused by individual
participants in the traffic with other relevant traffic members is
for sure beneficial not only in terms of safety, but also in terms
of traffic flow and overall smoothness of the transportation
process. The vehicle-to-vehicle communication (V2V) had been
previously studied for various sub-6 GHz frequency bands [2];
however, it is tempting to utilize specific parts of a much higher
frequency band designated as the millimeter wave (MMW)
band, which spans the frequency range 30-300 GHz. Many
studies consider especially the frequency band around 60 GHz
because the International Telecommunication Union (ITU)
assigns the 60 GHz band to the Industrial, scientific and medical
(ISM) bands allowing license-free operations and offers several
GHz of bandwidth, which is unheard of in the mentioned sub-
6 GHz band [3].
It needs to be mentioned that the 60 GHz band for vehicular
connectivity had been studied several decades earlier, see e.g.
[4], however, in a very narrowband setup leading to the fact that
the extensive bandwidth, available in the already mentioned
60 GHz band, stayed unused. At present, as the advances in the
monolithic microwave integrated circuitry (MMIC) have
brought the attention of both the industrial and the scientific
community again to the 60 GHz band, the V2V MMW
communication is revisited, for example in [5]. The path-loss of
the links at 60 GHz is of course still high due to the oxygen
absorption in the atmosphere which reaches 15 dB/km; therefore
it is deemed suitable for short-range communication rather than
for long-range backhauls [6]. The broadband V2V MMW
communication is studied in [7] demonstrating exemplary
power-delay profiles (PDPs), path-loss and root-mean-square
(RMS) delay spreads of the radio channel between two
oncoming vehicles. In [7], the information exchange is direct,
without any intermediary. In this paper, we show another
possible scenario termed vehicle-to-infrastructure (V2I), where
the information from the individual vehicles is gathered by the
(usually elevated) roadside infrastructure. Subsequently, the
information might be backhauled to a distant area (e.g. in
difficult non-line-of-sight situations (NLOS)) or transmitted
back to other vehicles even in the vicinity of the original
transmitter car. This relaying scenario might be beneficial due
to the elevated installation of the roadside infrastructure, thus
easing the issues of shadowing by other vehicles on the road [8].
In [9], the V2I scenario, where the transmitter and the
receiver are not moving but the surrounding traffic on the two-
lane road causes the time-varying nature of the channel, is
analyzed. It is shown in [9] that the presence of moving vehicles,
although making the channel non-stationary, may reduce the
RMS delay spread. This is due to the presence of cars, which
effectively serve as artificial reflecting surfaces and thus
increase the total received power.
In [10], the outage probability of MMW V2I propagation
channels is researched with the conclusion that the two-wave
Rice model is not necessary and “the road-reflected wave can be
neglected in the numerical computations”. The conclusions in
[10] are based on a ray-tracing analytical channel model. On the
other hand, however, based on the real-world channel sounding
campaign in [5], a two-way diffuse power (TWDP) channel
model is derived and validated via Akaike’s information
criterion. In [11], a small-scale fading model is derived for a V2I
scenario similar to that in [5] and again, the TWDP statistics are
confirmed. The two-way nature of the V2I MMW channel is
also observed in this paper; however, here we show different
statistical properties of the reflected path as compared with the
line of sight (LOS).
With this background, the detailed contributions of this
paper are as follows:
• Analysis of multipath components (MPCs) propagating
between a moving car and a receiver situated in a
building in terms of time dependence of direct and
reflected MPCs magnitudes, RMS delay spread of the
received MMW signal, and statistical evaluation of the
strongest reflected component.
• Examination of 2D power and the RMS delay spread
distributions in a static environment surrounding the car.
The rest of the paper is organized as follows. Section 2
briefly describes the channel sounder. Section 3 informs about
the measurement scenarios used for time-varying and static
environments. Section 4 deals with the real word measurements.
Then in Section 5, the analyses of both the time-varying and the
static channels are presented. A summary of the paper is given
in the conclusion.
II. MEASUREMENT SETUP
The channel measurement was carried out using the 60 GHz
time-domain channel sounder described in detail in [12]. It is
composed of a pseudo-random binary sequence (PRBS)
transmitter (TX) and a correlation receiver (RX). It employs
Golay complementary sequences as the excitation PRBS signal,
because of their very good correlation properties, minimal
leakage effects caused by FFT [13], and a great ability to
mitigate unwanted nonlinearity effects produced by channel
sounder analog circuits [14]. The transmitter is based on an
Anritsu MP1800A Signal Quality Analyzer working as a PRBS
generator. The receiver is created using a Tektronix
MSO72004C (20 GHz, 50 GS/s) Mixed Signal Oscilloscope
working as a very fast analog-to-digital converter. The baseband
PRBS signal is converted into the MMW band and back using
SiversIma FC1000V series V-band up/down converters [15].
The channel sounder bandwidth is 8 GHz, the number of
samples per measured channel impulse response (CIR) is set to
NSa = 8092 and the number of saved CIRs per measurement is
NCIR = 932. Due to the correlation gain the sounder dynamic
range is about 45 dB. In the case of static channel measurement,
it can be increased by another 5-10 dB using an averaging
technique. Downloading data from the oscilloscope to a PC, and
their basic processing are controlled by LabView.
The transmitter was equipped with an omnidirectional SIW
slot antenna described in [16], whose radiation pattern related to
the car is shown in Fig. 1. An angle of 0° in both the E- and the
H-plane corresponds to the left side of the car (as seen by the
driver) and the car roof in the H‐plane is situated at an angle of
270°. The MMW signal was received using a directional horn
antenna with a dielectric lens. The antenna gain dependence on
the angle is shown in Fig. 2.
All measured values mentioned below are related to the low
noise MMW preamplifier output. The following down converter
SiversIma then increases them by a gain of about 15-20 dB.
Fig. 1. E‐plane (left) and H‐plane (right) measured radiation pattern of
double-sided SIW slot antenna at 55 GHz, 60 GHz, and 65 GHz.
Fig. 2. E‐plane (top) and H‐plane (bottom) measured gain of horn antenna
with dielectric lens at 55 GHz, 60 GHz, and 65 GHz.
TX TX
55 GH z
60 GH z
65 GH z
55 GH z
60 GH z
65 GH z
Angle
Angle
III. MEASUREMENT SCENARIOS
All V2I measurements were performed in the Brno
University of Technology campus between the building at
Technicka 12 and a VW CC car driving on the road in front of
the building as shown in Fig. 3. The car was moving in both
directions at different speeds. In the text below, we will use the
designation "Scenario 1a" for the direction of moving shown in
Fig. 3 and "Scenario 1b" for the opposite direction of the car
movement. Since the measured results contained several
multipath components whose origin was ambiguous, we made
further stationary measurements in order to be able to correctly
interpret the measured data. In "Scenario 2" the car was situated
in front of the building and the channel was scanned by the horn
antenna directed to uniformly distributed points as shown in Fig.
4. To determine the direction of propagation of certain MPCs,
the TX antenna was in a few measurements shielded towards the
receiver by an MMW absorber as shown in the upper left corner.
In all the scenarios the transmitter was situated in the car
and its antenna together with the power amplifier and cooler
were placed on the car roof as shown in Fig. 5 (left). The receiver
was situated in a room on the 6th floor of the university building.
In Scenarios 1a and 1b the RX antenna was mounted on a
photographic gimbal head and tripod and directed out of an open
window (see Fig. 5 (middle)). The antenna was manually
directed to the moving car using a riflescope. This manual
tracking of the directional antenna simulates electronic beam
steering of an antenna array assumed for 5G wireless networks
[17]. The correctness of the horn antenna alignment was
checked using a video recorded by a camera mechanically
coupled to the antenna. In Scenario 2 the RX antenna was
directed using a motorized Sky-Watcher AllView mount (see
Fig. 5 (right)) controlled by PC and LabView.
For Scenarios 1a and 1b the car speed was chosen
v = 30 km/h to 50 km/h. As shown in Fig. 3, the height of the
TX and RX antennas above the ground is =1.55 m and =
14.5 m respectively. It is obvious that for Scenario 1a, where the
distance between the TX antenna and the building is D = 28.5
m, the propagation distance between the antennas is =
()+ =31.3 m. Similarly, for Scenario 1b, where D
is only 25 m because the car goes on the roadside closer to the
building, we can get =28 m.
IV. MEASUREMENT
For all the car speeds and directions, we had performed three
measurements and then we selected the measurement where the
car tracking was the most accurate. Note that the tracking of a
moving car is not easy especially at higher speeds. The
magnification of the telescope must not be too high, as it must
allow the car to be quickly found. In such a case, the transmitter
is too small for watching. Thus, in most cases the RX antenna
was directed to the car center (to the bottom part of the doors).
Examples of measured CIRs for Scenario 1a, are shown in
Fig. 6, where for a clearer interpretation of the results we used
the propagation distance at the vertical axis instead of the more
common time or delay.
Fig. 3. Measuring workplace in the university campus at
Technická 12, Brno
Fig. 4. Distribution of the measurement points for CIR evaluation.
First, we chose a long period of CIRs measurement
TCM = 5 ms and the car speed v = 50 km/h to be able to observe
changes in the MPCs and the influence of surrounding trees on
signal propagation. The corresponding total measurement time
TTM =TCM×NCIR was 4.675 s. Further, to analyze the influence of
the road roughness on the MMW signal reflection during the car
movement we set the measurement period to TCM = 200 µs
which gives the total measurement time TTM = 0.186 s. For
v = 50 km/h (v = 14 m/s) the CIRs are taken every v ×TCM =
2.8 millimeters of the car movement.
Fig. 5. From left to right: placement of TX antenna on the car roof, RX
antenna on the gimbal head and on the motorized mount.
d = 2.7m
D = 28.5m
H = 14.5 m
h = 1.55 m
Direct
component
Reflected
component
Road level
α
β
γ
TX
RX
β
’
TX
1 5
6
3
810
11 13
16 18 20
12 14
17 19
2 4
7 9
15
21 22 23 24 25
It is obvious that in both cases there are two dominant
components: a direct component (the stronger and nearer one)
and a reflected component. To find a reflective spot, we parked
the car at the position shown in Fig. 4 and placed the MMW
absorber on the road, sidewalk and various parts of the car. We
observed that this component was reflected off the road at about
2.7 meters from the car as shown in Fig. 3. The difference
between the direct and reflected path lengths =+ =
++()+ is then 1.4 m, which
corresponds to the measured difference shown in Fig. 6. The
angle between the direct component and the reflected
component is =4.95°. It is given by = =
arctg[()
]arctg[()
]. Note that the angles
and slightly differ because the road is not perfectly even. As
the maximum measurable distance of the channel sounder is
49 m [12], the multipath components propagating over a longer
distance (multiply reflected from the building behind the car) are
aliased as shown in Fig. 6. The white dashed curves approximate
the maximum MPCs position (see below). The measurements
were made so that the center of records corresponded to the
moment when the transmitter and the receiver were both in a
vertical plane perpendicular to the building wall. Different
distances at the beginning and the end of record shown in Fig. 6
(top) are caused by an imperfect parallelism of the road and the
building.
In Scenario 1b, when the car was moving in the opposite
direction, the data records are very similar to those shown in Fig.
6 (top), but there is no component reflected from the road since
the road and the sidewalk were shadowed by the car roof.
Scenario 2 measurements were performed for several car
positions. The most interesting measurement (the richest in
MPC components) is shown in Fig. 4. To get information about
the MPCs power distribution in 2D plane we used the motorized
mount and scanned the space around the transmitter in
rectangular coordinates with horizontal and vertical steps of 5°.
Since the measurement setup was working in the same way as
in Scenario 1 but it recorded only 100 CIRs for any measurement
point, we averaged them to increase the signal to noise ratio
(SNR). We evaluated the power at all points by summing the
CIR peaks above the noise floor, which was set to -60 dBV. A
smooth representation of the distribution, shown in Fig. 7, was
achieved by interpolating the values by a ratio of 32. The
elongated shape in the vertical direction is caused by strong
MPCs reflected by the car and the building behind the
transmitter and by the road. Of non-negligible influence are also
the different beam-widths of the RX antenna in the vertical and
the horizontal axes as shown in Fig. 2.
V. CHANNEL CHARACTERIZATION
As mentioned above the aim of this work is to analyze the
time varying MPC characteristics caused by a moving car such
as MPC magnitude variation in time, its statistical properties,
and the time variation of the delay spread. Another research goal
is to analyze the power and RMS delay spread distributions in a
static environment surrounding the car and to find any
dependence between them. So, we can divide the channel
characterization into two subsections devoted to the time-
varying and the static channels.
Fig. 6. The CIR magnitude in dBV measured in Scenario 1a with
sampling periods TS C = 5 ms (top) and and TCM = 200 µs
(bottom).
Fig. 7. Power distribution in the vinicity of the car.
0.5 11. 5 22.5 33. 5 44. 5
Measur emen t time [ s]
5
10
15
20
25
30
35
40
45
Propaga tion dista nce [m ]
-60
-55
-50
-45
-40
-35
-30
Measur emen t time [ s]
5
10
15
20
25
30
35
40
45
-60
-55
-50
-45
-40
-35
-30
Propaga tion dista nce [m ]
Component re flected from road
00.02 0. 04 0.06 0. 08 0.1 0.12 0. 14 0. 16 0.18
Direct component
Component re flected from road
Aliased components
TSC
= 5 ms
Direct component
Aliased components
TSC
= 200 µs
12 3 4 5
678910
11 12 13 14 15
16 17 18 19 20
21 22 23 24 25
-10 -5 0 5 10
Azimu th [deg ]
-10
-5
0
5
10
Ele vatio n [d eg ]
-60
-55
-50
-45
-40
-35
-30
A. Analysis of time-varying channel
To analyze the MPCs magnitude variation we approximated
the coordinates (indices) of direct and reflected components by
a smooth curve as shown in Fig. 6. It was accomplished by the
following three steps:
1. Detecting the strongest and the second strongest
component (i.e. the direct component and the component
reflected from the road) in each CIR record and
determining their coordinates.
2. Filtering the coordinates by a Hampel filter to remove
outliers (coordinates of the other higher components not
belonging to the two above).
3. Filtering the coordinates using a Savitzky-Golay FIR
filter to obtain a smooth approximation.
The Hampel filter is a configurable-width sliding window
filter, calculating for each window the median and the standard
deviation . If any point in the window is more than n out of
the median, where n is a user-definable value (default n = 3),
then the Hampel filter identifies this point as an outlier and
replaces it with the median. Compared with a simpler median
filter the Hampel filter is defined by another parameter (n),
which improves the filter configurability and offers better
fitting. Both the window length and n were set empirically.
The Savitzky-Golay filter is a frequently used smoothing
filter based on local least-squares polynomial approximation.
The filter is defined by the polynomial order and by the
approximation interval. Optional parameters implemented in
MATLAB, which was used for simulations, are positive-valued
weights used during the least-squares minimization. Thanks to
many configurable parameters, it can be very well optimized for
the desired approximation. The filter is necessary to
approximate the signal tendency when its magnitude is strongly
attenuated (for example when the signal penetrates trees) and
when the Hampel filter does not give a satisfactory result. In
fact, both filters represent the momentum of the car and smooth
the curves representing the distance change.
The time dependence of the direct and reflected component
magnitudes depicted in Fig. 6 (top) plotted for the fitted
coordinates is shown in Fig. 8 (grey waveforms). The time
dependence of the direct component obtained in Scenario 1b is
then shown in Fig. 9. The trends (black and red dashed
waveforms) were obtained by lowpass zero-phase bidirectional
filtering, where the data is processed in both the forward and the
reverse directions, which does not shift the filtered signal in
time.
The effect of trees in Figs 8 and 9 is obvious. The
attenuation varies in a wide range, depending on which part of
treetops the MMW signal penetrates. Its maximum is slightly
above 30 dB. Note that the difference in the magnitudes
between the direct and the reflected component in Scenario 1a
is relatively small. In the “unshadowed” record segments the
trends differ between 6-12 dB. As mentioned above it is caused
by pointing the traced RX antenna at the car doors.
Fig. 8. The time dependence of the direct and reflected component
magnitudes measured in Scenario 1a for TVM = 5 ms.
The ripple of the direct component magnitude is due to
imperfect manual tracking and angularly dependent TX antenna
irradiation (see Fig. 1). The ripple of the reflected component
magnitude is additionally affected by road irregularities. Note
that the bottom parts of component magnitudes are also affected
by the measurement setup noise.
The effect of the road on a reflected component is also
obvious in Fig. 10, where the component magnitudes depicted
in Fig. 6 (bottom) are plotted. While the small changes in the
direct component magnitude are caused by the receiver noise,
downconverter IQ imbalance, and possibly by the phase noise
of the reference rubidium oscillators, the larger variation of the
reflected component is caused predominantly by the rough road
surface (flat asphalt). To confirm this, one measurement was
performed with the stationary vehicle and compared with all the
time-varying channel measurements (made at different speeds
between 40 and 50 km/h) with respect to the reflected
component magnitude variance. The reflected component for
the stationary car is shown in Fig. 10 (blue waveform). It can be
concluded that the influence of receiver noise is small as the
variance (normalized power in W) of the reflected component
measured for a stationary car is 2.08e-07 while the variance of
the reflected component measured for a moving car is in the
interval from 2.30e-05 to 7.85e-06. Note that the direct
component variance in the static measurement is 2.10e-07.
Fig. 9. The time dependence of the direct component magnitudes
measured in Scenario 1b for TCM = 5 ms.
00.5 11.5 22.5 33 .5 44.5
Measurement time [s]
-70
-65
-60
-55
-50
-45
-40
-35
-30
Direct component
MPC Magnitude [dBV]
Component reflected
from road
00.5 11.5 22. 5 33.5 44.5
Measur ement time [s]
-70
-60
-50
-40
-30
Direct component
MPC Magnit ude [dBV]
Fig. 10. The time dependence of the LOS and reflected component
magnitudes measured in Scenario 1a for TCM = 200 µs.
Another aim of the channel characterization was to fit the
distribution of the reflected component magnitudes by a proper
probability density function (PDF). It was found that the best
fitting can be obtained by a generalized extreme value (GEV)
distribution given by the location parameter µ, scale parameter
σ, and shape parameter k ≠ 0. For all four unfiltered datasets we
defined intervals for the above parameters: µ∈〈4.4 mV,
5.8 mV〉, σ ∈〈2.3 mV, 4.0 mV〉, and k ∈〈-0.055, -0.108〉. No
relationship between the amplitude distribution and the speed of
the vehicle was found. The PDF calculated as a histogram, and
the cumulative distribution function (CDF) of the reflected
component depicted in Fig. 10 (gray waveform), fitted by the
GEV distribution are shown in Fig. 11. Note that when the
filtered trend (red curve) is subtracted from an unfiltered record
(gray waveform) the result can be successfully fitted with the
normal distribution. In other words, the effect of small particles
in asphalt can be fitted by a normal distribution.
Fig. 11. PDF (left) and CDF (right) calculated and fitted by
generalized extreme value distribution for the
component reflected from the road.
Fig. 12. The time dependence of the instantaneous RMS delay spread
for TCM= 5 ms.
Finally, we calculated the instantaneous RMS delay spread
for the components depicted in Fig. 6 (top) according to [18]
,=|(,)|
|(,)|
|(,)|
|(,)|
, (1)
where |h(τi, t)| are the magnitudes of CIR taps at the delay τi, t is
the measurement time (capturing time of CIRs), and L is the
number of taps. The number of taps was chosen experimentally
as the number of the CIR peaks exceeding the receiver noise
threshold (-60 dBV). The RMS delay spread plotted in the
logarithmic vertical scale is shown in Fig. 12. It corresponds
well to the magnitudes in Fig. 8. The range of averaged values
(black dashed waveform) between 2 ns (LOS propagation) and
60 ns (shadowing by trees) are typical of outdoor V2X scenarios
[19]. By applying FFT to the complex values of the component
reflected from the road we found that at speeds of up to 50 km/h,
the significant components do not exceed the 600 Hz as evident
from Fig. 13.
B. Analysis of static channel
The aim of the static channel analysis is to find how some
MPCs propagate through the channel in order to correctly
interpret the data measured and to find some dependence
between the received signal power and the RMS delay spread.
For this purpose, we first calculated the power distribution in the
car vicinity as mentioned above and shown in Fig. 7, and then
evaluated the RMS delay spread for all the measured points
according to (1). A map of the RMS delay spread distribution is
shown in Fig. 14. It is evident that the minimum value occurs
very close to the transmitter (point 13), due to the dominant
direct component. The relatively low delay spread at points 8,
12, 14 and 18 is caused by the presence of direct components
and either strong MPCs traveling close to the direct component
(reflected from the road) or weak MPCs traveling far from the
direct component (e.g. reflected from the building behind the
transmitter).
Fig. 13. Spectrum of reflected component measured in Scenario 1a for TSC =
200 µs and the car speed v = 40 km/h (black) and v = 50 km/h (red).
00.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18
Measurement time [s]
-60
-55
-50
-45
-40
-35
-30
MPC Magnitude [dBV]
Direct component
Component reflected from the road
for st ationa ry car
Component reflected
from road
0 5 10 15
0
Probabi lity density function [-]
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
Magn itude [m V] Magn itude [m V]
0 5 10 15
Cumula tive disstrib ution func tion [-]
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
00.5 11. 5 22.5 33. 5 44. 5
Measur emen t time [ s]
10-8
10-9
RMS delay spread [s]
00.1 0.2 0. 3 0.4 0.5 0. 6 0. 7 0.8 0.9 1.0
Frequen cy [kHz]
-30
-20
-10
0
Normali zed magnitud e [dB]
Fig. 14. Map of RMS delay spread in nanoseconds calculated for a
set of CIRs measured in the vinicity of the car.
The CIRs corresponding to points 13 and 18 are shown in
Fig. 15. The black waveforms were measured without the
MMW absorber and the red waveforms were obtained with the
absorber shielding the TX antenna. The blue circles indicate the
values used for the calculation. It is obvious that the component
C was reflected from the building behind the transmitter,
because it was also received with the absorber, while the
component D propagated towards the receiver and it was
probably multiply reflected. The difference between the points
13 and 18 is 5° in the vertical direction. This difference causes a
noticeable changes in the direct component A (13 dB) and in the
component reflected from the road (7 dB). This unequal
difference in the changes is due to the angle depended RX
antenna gain and the deviation of the position of central point 13
and the transmitter as shown in Fig. 4.
Fig. 15. The CIRs obtained at points 13 and 18 without (black) and
with (red) absorber.
Fig. 16. The CIRs obtained at the edge points 6 and 21.
The averaging effect implemented in this measurement
increases the dynamic range and allows detecting many weak
MPCs. Thanks to it we can evaluate the RMS delay spread also
for the edge points (far from the transmitter). An example of the
CIRs measured for the edge points with the highest and the
lowest RMS delay spread is shown in Fig. 16. The highest RMS
delay spread at point 21 is caused by a two very distant peaks
with similar power.
Comparing the power distribution depicted in Fig. 7 and the
RMS delay spread shown in Fig. 14 we can conclude that there
is no easily describable dependence between power and delay
spread of received signal except point 13. Note that in the case
of the RX antenna with a broader beam-width the RMS delay
spread would be probably higher.
VI. CONCLUSION
We analyzed the multipath components propagation
between a moving car on a road and a fixed receiver situated in
a campus building. We showed very good reflective properties
of the asphalt surface and demonstrated them by the small
difference between the direct and the reflected component
magnitudes varying between 6-17 dB (see Fig. 8 and Fig. 10).
Similar small difference lower than 12 dB can also be observed
between these components in “static scenario” (Fig. 15). Then
we examined the attenuation caused by trees. We discovered
that it varies in a wide range and its maximum for all the
measurements slightly exceeds the value 30 dB. The time
variation of the reflected component magnitudes was fitted by a
generalized extreme value distribution and the marginal values
of the distribution parameters were found for all the
measurements. For “moving scenario” we then evaluated the
time dependence of the RMS delay spread. We found that its
averaged values (trends) vary in the range from 2 ns (obtained
for obstructed LOS propagation) to 60 ns (obtained for the case
the signal shadowed by trees). Finally, we calculated the
spectrum of the component reflected from the road and
1 2 3 4 5
678910
11 12 13 14 15
16 17 18 19 20
21 22 23 24 25
-10 -5 0 510
Azimut h [deg]
-10
-5
0
5
10
Elevation [deg ]
5
10
15
20
25
30
35
40
45
50
0 5 10 15 20 25
-80
-60
-40
0510 15 20 25
-80
-60
-40
Point 13
Point 18 Dis tan ce [m]
MPC Magnit ude [dBV]MPC Magnit ude [dBV]
Dis tance [m]
B
C
A
A
B
D
0 5 10 15 20 25 30 35 40 45 50
-80
-60
-40
0 5 10 15 20 25 30 35 40 45 50
-80
-60
-40
Dis tance [m]
Point 6
Point 21
Dis tance [m]
MPC Mag nitude [dBV]
MPC Mag nitude [dBV]
discovered that at speeds of up to 50 km/h, there are no
significant components above 600 Hz.
To understand how the MPCs propagate through the channel
we measured and analyzed the power and the RMS delay spread
distributions in the static environment surrounding the car. We
obtained very similar values of the RMS delay spread between
3 and 50 ns as in the case of the previous time-varying channel
analysis. We justified these values by analyzing the multipath
component in the corresponding CIRs. Finally, by analyzing the
2D power and the RMS delay spread distribution we came to the
conclusion that there is no provable relationship between them.
By the results presented we generally demonstrate that a
building and a few cars parked close to the measuring car create
an environment relatively rich in multipath components.
ACKNOWLEDGMENT
The research described in this paper was financed by the
Czech Science Foundation, Project No. 17-27068S, by the
National Sustainability Program under grant LO1401 and by the
California Transportation Authority under the METRANS
program. For the research, the infrastructure of the SIX Center
was used.
REFERENCES
[1] C. Sommer and F. Dressler, Vehicular networking, Cambridge University
Press, 2015.
[2] C. F. Mecklenbrauker et al., "Vehicular Channel Characterization and Its
Implications for Wireless System Design and Performance, " Proceedings
of the IEEE, July 2011.
[3] ITU, “Radio Regulations, Section IV. Radio Stations and Systems –
Article 1.15, definition: Industrial, scientific and medical (ISM)
applications (of radio frequency energy)”.
[4] H. Meinel and A. Plattner, "Millimetre-wave propagation along railway
lines," IEE Proceedings F (Communications, Radar and Signal
Processing), no. 130, p. 688–694, 1983.
[5] E. Zöchmann et al., "Position-Specific Statistics of 60 GHz Vehicular
Channels During Overtaking," IEEE Access, vol. 7, pp. 14216–14232,
2019.
[6] M. Khatun, H. Mehrpouyan and D. Matolak, "60-GHz Millimeter-Wave
Pathloss Measurements in Boise Airport," in IEEE Global Conference
on Signal and Information Processing (GlobalSIP), pp. 1276–1280, 2018.
[7] J. Blumenstein, et al., "Vehicle-to-Vehicle Millimeter-Wave Channel
Measurements at 56-64 GHz," in IEEE Vehicular Technology Fall
Conference (VTC-Fall) (accepted for publication), Honolulu, pp. 1–4,
2019.
[8] M. G. Nilsson, D. Vlastaras, T. Abbas, B. Bergqvist and F. Tufvesson,
"On multilink shadowing effects in measured V2V channels on highway,"
in 9th European Conference on Antennas and Propagation (EuCAP),
Lisbon, 2015.
[9] J. Blumenstein et al., "Measured High-Resolution Power-Delay Profiles
of Nonstationary Vehicular Millimeter Wave Channels," in IEEE 29th
Annual International Symposium on Personal, Indoor and Mobile Radio
Communications (PIMRC), Bologna, 2018.
[10] R. Verdone, "Outage probability analysis for short-range communication
systems at 60 GHz in ATT urban environments," IEEE Transactions on
Vehicular Technology, vol. 46, no. 4, pp. 1027–1039, 1997.
[11] E. Zöchmann, J. Blumenstein, R. Marsalek, M. Rupp, K. Guan:
"Parsimonious Channel Models for Millimeter Wave Railway
Communications", in IEEE Wireless Communications and Networking
Conference (WCNC), 2019, pp. 1 - 6.
[12] A. Prokes et al., "Time-Domain Broadband 60 GHz Channel Sounder for
Vehicle-to-Vehicle Channel Measurement," in 2018 IEEE Vehicular
Networking Conference (VNC), Taipei, Taiwan, 2018, pp. 1–7.
[13] R. Zetik, M. Kmec, J. Sachs, and R. S. Thomä, "Real-Time MIMO
Channel Sounder for Emulation of Distributed Ultrawideband Systems,"
Int. J. Antennas Propag., vol. 2014, Article ID 317683, pp. 1–16, 2014.
[14] J. Vychodil, M. Pospisil, A. Prokes and J. Blumenstein, "Millimetre wave
band time domain channel sounder," IET Communications, vol. 13, no. 3,
pp. 331–338, 19 2 2019.
[15] FC1005V/00 V-band Converter with LO. [Online]. Available:
https://www.siversima.com/wp-content/uploads/FC1005V00-Data-
Sheet.pdf. [Accessed: 15- Jun- 2018].
[16] T. Mikulasek, J. Lacik, and Z. Raida, "SIW slot antennas utilized for 60-
GHz channel characterization," Microw Opt. Tech. Lett., vol. 57, no. 6,
pp. 1365–1370, 2015.
[17] S. Sun, T. S. Rappaport, R. W. Heath, A. Nix and S. Rangan, "Mimo for
millimeter-wave wireless communications: beamforming, spatial
multiplexing, or both?," IEEE Communications Magazine , vol. 52, no. 12,
pp. 110–121, December 2014.
[18] A. F. Molisch, "Statistical properties of the RMS delay-spread of mobile
radio channels with independent Rayleigh-fading paths," in IEEE
Transactions on Vehicular Technology, vol. 45, no. 1, pp. 201-204, Feb.
1996.
[19] E. Ben-Dor, T. S. Rappaport, Y. Qiao and S. J. Lauffenburger,
"Millimeter-Wave 60 GHz Outdoor and Vehicle AOA Propagation
Measurements Using a Broadband Channel Sounder," in IEEE Global
Telecommunications Conference (GLOBECOM), Kathmandu, 2011, pp.
1–6.
