99Akihito KATO et al.

3-4 ITS Wireless Transmission Technology

3-4-1 Technologies of Millimeter-Wave Inter-vehicleCommunications– Propagation Characteristics –

Akihito KATO, Katsuyoshi SATO, and Masayuki FUJISE

In this paper, we introduce developed technologies for millimeter-wave inter-vehiclecommunication (IVC) system in intelligent transport systems (ITS), especially propagationcharacteristics of 60 GHz band for the system design of IVC. First we introduce the outlineof an IVC system using millimeter wave and its research subjects. Next we show experi-mental results of propagation characteristics of radio wave at 60 GHz between running vehi-cles. The propagation model and mechanism of fading propagation are argued. The jointresearch activity of IVC system in Yokosuka Research Park (YRP) is also introduced.

Keywords Inter-vehicle communication, Millimeter wave, Propagation, ITS

1 Introduction

The number of cell-phone service sub-scribers reached 60 million at the end ofMarch 2001 and has continued to increase,exceeding the number of fixed-phone sub-scribers. In addition to the growth in wirelessmobile communications systems, the range oftheir applications is further expanding intonew fields such as Intelligent Transport Sys-tems (ITS), while moving toward themicrowave and even millimeter wavelengthregions. With such technological innovationsin its scope, the Mobile CommunicationGroup of the Yokosuka Radio Communica-tions Research Center has been investigatingradio communications technologies in theuntapped millimeter wave band since FY1998,in anticipation of new developments in theinfo-communications area of ITS. In Yokosu-ka Research Park (YRP) where this researchcenter is located, the world’s top mobile-device manufacturers and research instituteshave gathered to push ahead with leading-edge mobile communications research on

mobile multimedia systems represented by thenext-generation cell-phone and ITS radiocommunications system, in cooperation withindustry, academia, and government[1].

This paper describes our research on on-road radio propagation during millimeter-wave Inter-Vehicle communications, a subjectthat is among the research activities conductedat YRP. We first explain the features of themillimeter-wave inter-vehicle communicationand its applications, clarifying the relevanttechnological issues that remain to beresolved. Research on ITS millimeter-waveinter-vehicle communication conducted inYRP is also introduced. Next, we considerpropagation models and fading mechanismsbased on the experimental results of our fieldpropagation tests. The paper concludes with adiscussion of future research activities.

2 Features of Millimeter-WaveInter-Vehicle Communicationsand Problems to Resolved

The millimeter-wave inter-vehicle com-

100

munications system is a wireless communica-tions system that can improve safety and con-venience during driving by enabling directcommunications between vehicles, withoutrelying on roadside infrastructure, by using amillimeter-wave band such as the 60-GHzband. Fig.1 shows some example applicationsof the millimeter-wave inter-vehicle commu-nications system.

When the millimeter-wave band isemployed for such inter-vehicle communica-tions, the signal is less attenuated by fog andsnow than communications systems based oninfrared light, and is less affected by interfer-ence from sunlight. In particular, the inter-vehicle communications system, as a dedicat-ed short-range communications system(DSRC) using the 60-GHz band in the mil-limeter-wave region, provides a number ofadvantages. For example, the reuse efficiencyof channels is high due to the large atmospher-ic absorption, capacity for high-speed broad-band communication, and the ability to pro-vide linkage operations and share circuits withonboard radar, which has already been com-mercialized.

As with communications for transmittingcontrol signals for unmanned operations suchas platooning, an inter-vehicle communica-tions system using millimeter waves basically

assumes communication among vehicles mov-ing in the same direction within line of radiosight. Although the vehicles move quickly onthe ground, the relative speed between thecommunicating vehicles is low. Thus, directcommunications between such mobile stationsis similar to that of conventional fixed sta-tions. In contrast, millimeter-wave-basedcommunication may be seriously affected byeven small amounts of relative movementbetween vehicles, due to the short wavelengthof the millimeter wave. The influence of themultipath effect from the surrounding terres-trial features near the road may be relativelysmall, as the antenna will have a relativelynarrow beam width. However, the signalreflection from the road surface is not negligi-ble, and very strong fading occurs due to inter-ference between the direct waves and thereflected waves from the road surface. In thisway, the radio-wave propagation phenomenonobserved during millimeter-wave inter-vehiclecommunication will differ significantly fromthat of conventional mobile communications.Thus, it is very important to clarify the propa-gation mechanism and establish a propagationmodel[2][3] for practical use of this system.

Journal of the Communications Research Laboratory Vol.48 No.4 2001

Application examples of millimeter-wave inter-vehicle communicationsFig.1

101

3 ITS Joint Research in YokosukaResearch Park[1]

Yokosuka Research Park (YRP) wasestablished in October 1998 to contribute tothe development of domestic and internationalinfo-communications technologies and to helpestablish an affluent society with a high quali-ty of life for the 21st century by strengtheningthe research base for info-communicationstechnologies and by proceeding with researchon both basic and new technologies. Activi-ties in this area are currently developing on aglobally unprecedented scale, as a researchpark dedicated to mobile-communicationstechnology. Indeed, research institutes and theworld’s top manufacturers of mobile commu-nications devices have gathered at YRP, andare resolutely forging ahead with research oncutting-edge mobile-communications tech-nologies, such as mobile systems representedby next-generation cell phones. A wide rangeof research is underway involving industry,academia, and government to realize the mul-timedia information society of the 21st centu-ry. As part of such research activities, ITScommunications technology has been understudy since FY1998 by private companies(joined by 19 additional companies inFY2000), led by the CommunicationsResearch Laboratory. They have been jointlyconducting research on ITS wireless-commu-nications technologies (for inter-vehicle com-munication, road-vehicle communication)using millimeter waves. Fig.2 illustrates the

cooperative research organization for FY2000.The Inter-Vehicle Communications Work-

ing Group launched its activities in 1998 tocreate the ITS inter-vehicle wireless-commu-nications system based on millimeter waves,and to provide insight into its technologicalbasis and standardization issues. In FY 2000,joined by some 12 organizations, it conductedresearch and development of specific applica-tions and worked toward the widespread dis-tribution of useful systems. Currently empha-sizing activities that target commercialization,it is divided into three sub-working groups forthe undertaking of specific studies.

4 60-GHz-Band Outdoor RadioPropagation Measurement Sys-tem

We have been investigating a 60-GHz-band outdoor radio propagation measurementsystem since 1998 to determine its on-roadradio propagation characteristics during mil-limeter-wave inter-vehicle communication andhave conducted a variety of measurements.The system configuration and its specifica-tions are shown in Fig.3 and Table 1, respec-tively. Fig.4 shows an external view of the RFunit of the system, along with views showinginstallation in test vehicles. This system con-figuration is such that the RF units and meas-urement devices are aboard two test vehicles.The configuration assumes a situation inwhich the experiment is conducted using apair of vehicles traveling on roads and

Akihito KATO et al.

Joint research organizations in YRP (for FY 2000)Fig.2

102

expressways. In principle, the RF unit iscapable of two-way communication based onfrequency division duplex. During our propa-gation experiments, however, the experimentwas performed on a single-direction basis,with the front vehicle serving as the transmit-ting side and the rear vehicle serving as thereceiving side. The rear vehicle is equippedwith two series of wireless systems to examinethe effects of space diversity by changing theheights of the systems. This diversity is basedon selective signal synthesis in which signal-

receiving paths are switched with time lagdetermined at the moment conditions are met,such that the receiving powers and thresholdvalues are compared and matched in the tworeceiving systems. Thus, the timing of switch-ing is not synchronized with the demodulationsymbols. The data on received power isobtained from the control voltage of AGC inthe IF unit, and data acquisition is possible inthe 150-kHz band at maximum in both sys-tems. Our system temporarily includes aDFSK-type modem that can provide BER atintervals of 1 s or 10 ms during data transmis-sion at transmission rates of 1 Mbps, 5 Mbps,

Journal of the Communications Research Laboratory Vol.48 No.4 2001

Configuration of the 60-GHz outdoor radio-wave propagation measurement systemFig.3

Center frequency

Maximum transmission power

Data transmission rate

Modulation-demodulation method

Detection mode

Measurement of received power

Diversity

59.1GHz

10dBm

1Mbps～10Mbps

DFSK(Manchester code)

Delay detection

Use of AGC voltage

Selective synthesis basedon condition fulfillment

Specifications of the 60-GHz out-door radio-wave propagationmeasurement system

Table 1

External view of the RF unit and instal-lation in the test vehicle

Fig.4

103Akihito KATO et al.

and 10 Mbps. To precisely monitor the ever-changing measurement conditions, the meas-urement system also features a distancebetween vehicles measurement system usinglaser radar, a three-axial measurement systemusing optical gyroscopes, and a recording sys-tem using a video camera that records the sur-rounding views. All data provided by thesesystems is synchronized with the GPS signal,and off-line analysis can be performed basedon the synchronized data after measurementshave been taken.

5 Two-Ray Model for On-RoadRadio Propagation

It has been pointed out[2]～[5] that millime-ter-wave propagation on a road is multipathpropagation combining the direct wave andother waves reflected by the road surface andsurrounding structures. In the current experi-ment, we conducted measurements in an openarea with relatively few surrounding struc-tures, such as buildings. Thus, we considereda two-ray model that limited the waves incom-ing to the receiver to direct waves and wavesreflected from the road surface, and then com-pared the measurement results with the calcu-lations for received power provided by themodel. As shown in Fig.5, received power Pr

is given by the following equation[6], consid-ering only the direct wave and the reflectedwave when a transmitter held at a height ht anda receiver at a height hr , with each rd apartfrom one another in the horizontal direction:

where Gt and Gr are the gains of the boresightof the transmitter and receiver, respectively, rd

and rr are the optical path lengths of the directwave and the wave reflected from the roadsurface, L(rd) is the absorption factor[7] in the60-GHz band, λis the wavelength of the car-rier wave, 2π/λ,φis the phase rotation duringroad reflection (assuming π in this study), Dd

and Dr are the coefficients of antenna directiv-ity for the direct wave and the reflective wave

in the arrival direction, and Γ is the reflectioncoefficient of the road surface.

Next, considering the current experimentalconditions, if antenna heights ht and hr are50cm when the distance between vehicles rd is50-100 m, the incident angle of the reflectedwave onto the road surface falls in the rangeof 88.8-89.4 degrees. Assuming that therefractive index of asphalt is 2.0-j 0.05[8], theabsolute value of the reflection coefficient ofasphalt is estimated to be 0.90-0.99 at 60 GHzwithin the above incident angle range.Although various polarization antennas maybe used in actual measurements, the reflectioncoefficient is expected to fall in the aboverange. Thus, in the present study, we haveassumed the absolute value of the on-roadreflection coefficient to be 1. In addition, thetwo test vehicles travel in the same lane nearlythroughout the experiment. Thus, both directand reflected waves strike the antenna at inci-dent angles much smaller than the antennahalf-width, making it unnecessary to considerantenna directivity. Therefore, Equation 1 canbe approximated using a simpler equation thatdoes not account for antenna directivity, asfollows[9]:

We have conducted a theoretical investigationusing the above equation for the two-raymodel of the on-road propagation phenome-non.

6 On-Road Millimeter-Wave Prop-agation Characteristics

With respect to radio-wave propagationbetween moving vehicles, a number of factorsinfluence radio-wave propagation characteris-

Two-ray model for on-road propagationFig.5

(1)(2)

104

tics, such as vehicle motions and changes insurrounding conditions. However, we haveconducted measurement under conditions thatmay be affected as little as possible by themotion of vehicles and the surrounding condi-tions, so as to gain an understanding of thefundamental characteristics of on-road radio-wave propagation. In the experiment, twovehicles nearly at a standstill transceived mil-limeter-wave signals on a straight road withclear sight, and measured the received powerand bit error rate (BER) at each distance (dis-tance characteristics), gradually changing thedistance between the transmitter and receiverin the 10-200 m range.(1) Distance characteristics with antennaheight as a parameter[10]

Fig.6(a) and (b) show examples of themeasurement results for received power andthe bit error rate plotted for two antennaheights. The solid line represents the meas-urement, while the dotted line represents themodel calculation provided by the propagationmodel considering both direct waves androad-surface reflected waves. In both graphs,the measurement and calculation approximate-ly agree, implying that the two-ray model con-sidering both direct waves and road-surfacereflected waves can be used under open meas-urement conditions such as those employed inthe experiment. In addition, we can expecteffects of space diversity in the vertical direc-tion, as the distance that provides the nullpoint resulting from the interference of thetwo waves changes with antenna height. Fig.7shows the measurement results obtained whenspace diversity was utilized. In comparisonwith the previous figures, this figure indicatesthat vertical space diversity eliminates thedeep null point and improves BER.(2) Distance characteristics with anten-na-beam width as a parameter[7][11]

Fig.8 shows the measurement results forreceived power in relation to the transmitter-receiver distance when the beam width of thetransceiver antenna was set at 4 or 10 degrees.The solid line represents the measurement,while the broken line represents the simulation

results provided by a two-ray model consider-ing the directivity of the rectangular apertureantenna. This figure indicates that the influ-ence of reflected waves is reduced as directivi-

Journal of the Communications Research Laboratory Vol.48 No.4 2001

Distance characteristics with antennaheight as a parameter

Fig.6

Distance characteristics with diversityFig.7

105

ty is augmented, and that the decrease at thenull point then becomes small. Thus, the fad-ing caused by road-surface reflected waveswill be reduced by narrowing antenna direc-tivity.(3) Signal blocking by intermediate vehi-cles[12][13]

We measured the change in receivedpower by employing two vehicles equippedwith transceivers on a straight road 46 m apartfrom each other, and slowly moving an inter-mediate vehicle (sedan) from the receiver sideto transmitter side so that it intervened in thepropagation path between the transmitter andreceiver. Fig.9 shows the measurementresults, with the horizontal axis representingthe distance from the receiver to the interme-diate vehicle. The received power with nointermediate vehicle was – 46 dBm. Thegraph indicates that the received powerpeaked when the intermediate vehicle reachedthe midpoint between the transmitter and thereceiver. This is probably due to the fact thatthe direct waves were blocked by the sedan,but that signals reached the receiver by pass-ing under the intermediate vehicle, beingreflected off the road surface. The regionbetween the two vertical lines in the figurerepresents the distance when the reflectedwave enters the line of sight by passing thevehicle, based on the calculation of clearance.The figure indicates that the signal decaybecomes relatively small when the propaga-

tion path enters the line of sight. Indeed, thereare cases in which the decay due to blocking ismerely a few dB. In other words, even if thereexists an intermediate vehicle, blocking lossmay be relatively small, depending on specificconditions. Communications may then becontinued with no shadowing. On the otherhand, this implies that interference willbecome a serious problem in a system thatuses a co-channel among several platooningvehicles in anticipation of a large blockingloss.

7 Received-Power Fading duringHigh-Speed Travel [10]

The preceding section examined the valid-ity of adopting the two-ray model of the inter-vehicle radio propagation phenomenon for

Akihito KATO et al.

Distance characteristics with antenna-beam width as a parameterFig.8

Beam width of 10 degrees Beam width of 4 degrees

Shadowing characteristics of interme-diate vehicle

Fig.9

periods sufficiently short to disregard thevehicle motion and the dependence ofreceived power upon the antenna-to-antennadistance and antenna height. However, theheight of a traveling vehicle and the distancebetween vehicles change from time to time, asdo the surrounding terrestrial features. There-fore, the simple two-ray model may not accu-rately simulate the propagation phenomenonof actual travel under complex conditions. Wetherefore operated two vehicles on a commonexpressway to measure the inter-vehicle prop-agation characteristics and data-transmissioncharacteristics under such realistic conditions.

In the experiment, we operated two vehi-cles in the same lane, maintaining a distanceof approximately 80 m at a speed of 80 km/hon the Yokohama-Yokosuka Expressway(between Sawara and Namiki), and measuredthe instantaneous received power and bit errorrate. Table 2 lists the measurement condi-tions, while Fig.10 shows a view of the meas-urement site.

7.1 Received power and BER fluctua-tions over time

Fig.11 shows the median value of theshort-term median values of received powermeasured at intervals of 10 ms in each receiv-er, the bit error rate (BER), and the distance

between the vehicles over time. The measure-ment was begun immediately before theSawara Interchange and ended immediately

106 Journal of the Communications Research Laboratory Vol.48 No.4 2001

Short-term median value of received power, bit error rate, and horizontal distance over timeFig.11

View of the experimental site on anexpressway

Fig.10

Transmission power

Modulation-demodulation method

Data transmission rate

Antenna

Antenna gain

Beam width (horizontal)

Beam width (vertical)

Polarized wave

Height of transmitter

Diversity threshold (level)

Diversity threshold (difference in level)

Diversity timing delay

10dBm

DFSK

1Mbps

Patch array

20dBi

30degrees

7.2degrees

45degrees tilt

47cm

－80dBm

20dB

10ns

Measurement conditionsTable 2

107

after the Namiki Exit. This figure indicatesthat there was serious fading of more than 20dB in received power over the entire measure-ment period, and that BER degraded as thereceived power declined. Fig.12(a) and (b)demonstrate the relationship between theshort-term median value of received powerand the distance between vehicles. The curvein each figure represents the received powerpredicted by a model using Equation (2) forthe individual inter-vehicle distances. Thecalculated curve implies that the level ofreceived power changes along with the dis-tance between vehicles due to interference

between the direct waves and reflected waves.The measurement results also indicate that thelevel of received power fluctuates significant-ly, even when the distance between vehiclesremains constant. This is probably due to thefact that the actual height of the vehiclechanges as it traveled on the road, althoughthe theoretical study assumed a constantheight for each transceiver. Under actual con-ditions, the received power fluctuates even ifthe distance between vehicles remainsunchanged.

7.2 Cumulative distributionFig.13 shows the cumulative distribution

of the short-term median value of receivedpower at each receiver. This figure includesthe theoretical predictions of the Rayleigh dis-tribution and the log-normal distribution(assuming a standard deviation of 1). Thecumulative distribution obtained throughmeasurement is not dependent on the height ofthe receiver, which is located halfway betweenthe Rayleigh distribution and the log-normaldistribution in the figure. Specifically, theheight of the transceiver is not constant, butchanges over time in actual driving on a road.Measurement using the optical gyroscope hasshown that the distribution of the vertical fluc-tuations of a transceiver is close to the normaldistribution. In addition, calculations usingthe two-ray model based on Equation (2) pre-

Akihito KATO et al.

Cumulative distribution of the short-term median value of received power

Fig.13

Measurement results for the short-term median value of receivedpower in relation to the horizontal dis-tance for each receiver height

Fig.12

108

dict that the distance characteristics willchange considerably if the height of the trans-ceiver changes by even a few centimeters[10].Specifically, the vertical movement of thevehicle is among the crucial factors affectingthe fading mechanism for millimeter-waveinter-vehicle communication.

We then estimated the vehicle height ateach moment based on the pitch of the vehiclemeasured during a drive on an expressway.Referring to the obtained results and the meas-urement results for distance between vehicles,we estimated the instantaneous value ofreceived power based on a theoretical studyusing Equation(2). We then calculated thecumulative distribution for these fluctuationsover time. Fig.14(a) and (b) show the cumula-

tive distribution obtained from the previousmeasurement of received power and the theo-retical prediction based on the current analy-sis, respectively. As they show good agree-ment overall, we can assume that the verticalmotion of the vehicle is an important factor indetermining the fading mechanism.

Fig.15 shows the cumulative distributionof BER during 1 Mbps data transmission andselective-synthetic space diversity. In this fig-ure, the horizontal axis represents the expo-nent of BER (log10), while the vertical axisrepresents the time ratio as a percentage indi-cating the length of time for which BER isequal to or smaller than the correspondingvalue on the horizontal axis. Note that theposition denoted -8 indicates that there is noerror. The figure indicates that there were noerrors for 70% of the entire span of time, andthat the figure rose to 87% when diversity wasemployed. The results of this study imply thatvertical space diversity will be effective inmillimeter-wave inter-vehicle communication.

8 Conclusions

We have reported on the on-road radiopropagation characteristics that must be under-stood in the design of inter-vehicle communi-cations systems using millimeter waves. Ourresearch activities entered a new phase in FY2001, in anticipation of the implementation of

Journal of the Communications Research Laboratory Vol.48 No.4 2001

BER cumulative distribution for eachreceiver height for diversity reception

Fig.15

Comparison between measurementand calculation of the cumulative dis-tribution of received power

Fig.14

109Akihito KATO et al.

References1 A. Kato, et.al., "Research Activity on 60 GHz Band Inter-Vehicle Communication at Yokosuka Research

Park," Technical Report of IEICE ITS2001-6, 2001.

2 Yoshio Karasawa, "Maltipath Fading due to Road Surface Reflection and Fading Reduction by means of

Space Diversity in ITS Vehicle-toVehicle Communications at 60 GHz," Trans. on IEICE Vol.J83-B, No.4,

pp.518-524 2000.

3 K. Tokuda, et.al., "Analysis of Millimeter-Wave Band Road Surface Reflection Fading(RSRF) in Vehicle-to-

Vehicle Communications," Technical Report of IEICE AP98-134, 1999.

4 A. Kato, et.al., "Propagation Characteristics at 60 GHz on the Road for ITS Inter-Vehicle Communuications,"

Technical report of IEICE SST99-105, 2000.

5 T. Wada, et.al., "Theoretical Analysis of Propagation Characteriostics in Millimeter Waves Inter-Vehicle Com-

munication System," Trans. on IEICE J81-B-II, No.12, pp.1116-1125, 1998.

6 N. Taguchi, et. al., "Propagation Characteristics of 60 GHz Millimeter Wave for ITS Inter-Vehicle Communi-

cations (3)," Proc. ITST2000 S9-3, pp.259-262, 2000.

7 "Attenuation by atmospheric gases," Rec. ITU-R, P676-3, pp.244-260, 1997.

8 K. Sato, el.al., "Reflection Characteristics of Asphalt Surfaces in Millimeter-Wave Band," Proc. of IEICE B-1-

10, 1998.

9 Y. Hosoya, et.al., Radiowave Propagation Handbook, Realize Inc., pp.125-126, 1999.

10 A. Kato, et. al., "Propagation characteristics of 60-GHz millimeter waves for ITS inter-vehicle communica-

tions", IEICE Trans. Commun., Vol. E84-B, No.9, pp.2530-2539, 2001.

11 N. Taguchi, et.al., "Propagation Characteristics on the Road at 60 GHz for ITS Inter-Vehicle Communica-

tions," Proc. of IEICE A-17-32, 2000.

12 S. Noda, et. al., "Propagation characteristics of 60 GHz millimeter wave for ITS inter-vehicle communications

(4) -shadowing effect by interrupting vehicle-," ITST2000, pp.263-266, 2000.

13 A. Yamamoto, et.al., "Propagation Characteristics for ITS Inter-Vehicle Communications at 60 GHz (2),"

Proc. of IEICE A-17-26, 2001.

the millimeter-wave inter-vehicle communica-tions system in around FY 2003. Targeting afew applications that are expected to be readi-ly accepted in our society, we have directedour efforts toward commercialization, focus-ing on individual system designs and fieldtests. In addition, we are concentrating ourefforts on realizing this service by carrying outfield tests to verify its effectiveness, and on

realizing technical standardization as quicklyas possible by providing standardizationorganizations with a YRP draft that can serveas the basis of official technological standards.

We would like to express our deep appre-ciation to the members of the Inter-VehicleCommunication Working Group for numerousexciting discussions concerning the experi-ment and analysis.

110 Journal of the Communications Research Laboratory Vol.48 No.4 2001

Akihito KATO, Ph. D.

Senior Researcher, Mobile Communi-cations Group, Yokosuka Radio Com-munications Research Center

Propagation, Mobile Wireless Commu-nication

Masayuki FUJISE, Dr. Eng.

Leader, Mobile CommunicationsGroup, Yokosuka Radio Communica-tions Research Center

Optical Fiber Communication, Wire-less Communication

Katsuyoshi SATO

Senior Researcher, Mobile Communi-cations Group, Yokosuka Radio Com-munications Research Center

Propagation