Propagation Modeling For ITS Applications in The Roadway Environment

ITS Journal - Intelligent Transportation Systems Journal
ISSN: 1024-8072 (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/gits19
Propagation Modeling for ITS Applications in the

Roadway Environment
NICHOLAS DE MINCO & CHRISTOPHER L. HOLLOWAY
To cite this article: NICHOLAS DE MINCO & CHRISTOPHER L. HOLLOWAY (1997) Propagation
Modeling for ITS Applications in the Roadway Environment, ITS Journal - Intelligent
Transportation Systems Journal, 3:4, 287-331, DOI: 10.1080/10248079708903728
To link to this article: http://dx.doi.org/10.1080/10248079708903728
Published online: 24 Oct 2007.
Submit your article to this journal
Article views: 35
View related articles
Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=gits20
Download by: [Australian National University] Date: 13 June 2016, At: 08:45
mJourwr
1997. VOI.
i(4). pp. 287-111 0 1997 OPA (Ovcrrca~PublisherrAssaeiaion)
Reprints available dtmedy from the publisher Amrlerdam B.V. Publirhd inThc Nclhnlanb
Pholocopying p m i n r d by liccnx only under lircnw by Gordon and Breach Science Publishers
P n n l d in Malaysia
Propagation Modeling for ITS

Applications in the Roadway
ITS Journal - Intelligent Transportation Systems Journal 1997.3:287-331.
Environment
NICHOLAS DE MlNCO and CHRISTOPHER L. HOLLOWAY
Institute for Telecommunication Scrences

National Telecommunications and Information Administration
U.S. Department of Commerce
(Received 5 February 1996; In final form 3 October 1996)
It is expected that the Intelligent Transportation Systems (ITS) of the future will involve communica-
tion between vehicles and the roadside. In designing these communication systems it is imperative that
one has a sound and detailed understanding of radio-wave propagation in both short-distance roadside
and complicated urban environments. In this paper we discuss the type of propagation models that are
needed for ITS designers to predict system performance. This paper has approached this problem by
examining the radio-wave propagation in the roadway environment over the range of distances and at
frequencies that will have the potential to meet the needs of ITS. The requirements of the individual
functions of ITS were taken into consideration in addition to what poilions of the radio frequency
spectrum will be available.
Keywords: ITS; IVHS; ATMS; ATS; traffic control; trafic management; route guidance; electronic
toll and trafic management; radio-wave propagation; roadway environment; radio frequencies
INTRODUCTION
The functions of ITS will probably be implemented using radio communication

links and collision avoidance radars ranging from meters to hundreds or even
thousands of meters. Figure 1 shows some of the applications that would require
communications and radar systems to operate over these distances in the road-
way environment. Radio-wave propagation for ITS applications on short paths
in the roadway environment for wideband systems and digital modulation has
had little attention. This type of environment would most likely contain severe
multipath and its effect on different modulation methods would have to be
determined. Sufficient knowledge of radio-wave propagation in such an environ-
ment is not readily available.
288 N. DE MlNCO AND C.L. HOLLOWAY
Radio-wave propagation prediction models, that can be used for engineering

analyses of communication systems, determine the basic transmission loss
between a communications transmitter and receiver. Prediction of the behavior
of separate functional communication subsystems of ITS is difficult at best
without this knowledge. This knowledge is necessary for planning, architectural
evaluation, system performance prediction, system design and testing, standards
development, and standards conformance testing. It is necessary that this knowl-
edge be developed early in the ITS program. This would allow system perfor-
mance evaluation of these separate systems within ITS at proposed frequencies.
FIGURE I Intelligent Transportation Systems applications.
If reliable performance predictions cannot be made for subsystems of ITS,

operability problems could arise. Reliable predictions are necessary to evaluate
link margins, so that transmitter power requirements can be minimized for both
energy consumption and radiation hazard considerations. At the same time the
communications engineer needs to design with a link margin that will result in
adequate received signal that is not masked by noise. Two-way communications
systems operating between vehicle-to-roadside could not be designed efficiently.
The design of electronic toll and traffic management systems requires the proper
radio-wave propagation models and performance prediction algorithms to design
for system operation with minimized error occurrences. Large losses of revenue
PROPAGATION MODEL IN ROADWAY ENVIRONMENT 289
and dissatisfied customers could result from toll collection subsystem malfunc-
tion. False alarms and missed detections for ITS collision avoidance radars
could be minimized with paper analysis using these improved models. Malfunc-
tion of subsystems in ITS could cause user acceptance to decrease and could
also involve personnel safety considerations.
In this paper we review the radio-wave propagation concepts that need to
be incorporated into models that will allow the ITS propagation channels to be
analyzed. Specifically, we discussed propagation in both the general roadway
and an urban canyon environment. A discussion of the present capabilities of

existing propagation models is included with a discussion of the improvements
that are needed to ensure reliable radio-wave propagation loss predictions.
THE ITS ROADWAY ENVIRONMENT
The ITS roadway environment involves considering both the general roadway
environment and an urban canyon environment. The urban canyon environment
refers to downtown streets and buildings typical of cities.
When radio waves propagate down the streets of urban canyons or the
short-range roadway environment, they will encounter buildings, doors,
windows, automotive vehicles, ground, and other obstacles with both comers
and rounded surfaces that can be smooth or rough. These obstacles must be
represented as real surfaces with finite conductivity and permitivity. They can
be poor or good conductors or insulators. The reflection off the surface is
directly related to the composition of the material as well as whether it is
hollow or solid. This affects the amount of energy reflected in the specular
direction. The reflecting surfaces and diffraction geometries include streets,
ground and building walls that are not perfectly conducting, but have surfaces
with finite conductivity and are not smooth. The reflecting surfaces can be
concrete, asphalt, soil, or many other natural and man-made materials. Special
treatment is necessary for concrete block and other composite walls of build-
ings. Rough and smooth surfaces behave vastly different in how much energy
they reflect from these different surfaces. The polarization of the radio waves
also has a major impact on the radio-wave propagation behavior. For these
propagation paths in the ITS roadway environment, correctly representing
reflections of electromagnetic energy from surfaces is critical for accurate
prediction of radio-wave propagation.
charactenzing radio-wave propagation for various environments has been the
object of much attention. The references at the end of this paper summarize
much of the research in this area. The main difficulty with much of this research
290 N. DE MINCOAND C.L. HOLLOWAY
is that only line-of-sight (LOS) and limited diffraction propagation channels

have been modeled. Only very recently have models of complex out-of-sight
(00.5) propagation been invesiigated such as propagation down side streets and
over roof tops (Bergljung and Olsson, 1991; Maciel et al., 1993; Erceg et al.,
1992; Yim and Siang, 1992; Tan and Tan, 1995). One problem with these current
LOS and OOS models is that simple assumptions about the reflecting surfaces
and diffraction geometries are made. For example, both ground and building
walls are assumed to be perfectly conducting smooth surfaces, but a large class
of ITS situations do not resemble this scenario. Real surfaces have a finite
conductivity and they are not smooth and, as a result, behave in a significantly
different manner.
Diffraction around obstacle comers depends heavily on this permittivity and
conductivity. Modeling these obstacles as perfectly conducting is not adequate.
If structures have multiple edges and comers that are closely spaced, such as
with narrow comers and narrow streets, then coupling between the comers will
occur. This has an important effect on diffraction computations.
The coupling of energy into the side,streets in an urban environment involves
both reflected rays into the side streets as well as diffracted rays from the comer
of the building at the main and side street intersection. This environment also
includes propagation over rooftops on city blocks lined with both two-story and
many-story buildings. The diffracted and reflected rays should not only include
those rays due to the buildings but also those due to the multiple ground and
groundlbuilding reflection and diffraction combinations.
When the roadway environment contains buildings, rough terrain, trees,
forests, or vegetation, these obstacles must be modeled differently at different
frequencies, because the analysis techniques and radio-wave propagation
phenomena are different at different frequencies. The signal may have to propa-
gate through and around these obstacles.
The roadway environment contains congested vehicle traffic moving down
urban canyons. The road surface becomes a rough surface where the dimensions
of the roughness (automobiles) are comparable to a wavelength, and therefore
haye a significant effect on radio-wave propagation. The vehicles are in motion
and also have an effect on the reflection, diffraction, and scattering of the radio
waves.
ITS subsystems may need to function in areas where the roadway is lined
with thick forest or in residential neighborhoods, where streets are heavily lined
with trees and shrubs. In order to accurately model propagation in this environ-
ment it is important to accurately characterize the reflection coefficient from the
forest boundary. In the ITS environment there may be a need to propagate
through this foliage.
PROPAGATION MODEL IN ROADWAY ENVIRONMENT 29 1
The roadway environment in urban settings contains bridges composed of

complex metal structures. The environment also contains many different types
of tunnels. The metal structure of the bridges and the material content of the
tunnels results in very complicated propagation channels. Attenuation in these
types of environments can be severe.
When vehicles are in close proximity to one another, mutual coupling occurs
in both the near-field and far-field regions of an antenna under consideration.
This can affect the performance of the subsystems of ITS. Situations such as toll
collection and vehicle identification in the roadway environment can occur in

the near field of the antennas involved. These must be treated separately from
the far-field cases.
One system for disseminating information to vehicles in a rural environment
uses the AM broadcast band where both a sky wave and ground wave must be
considered. In the open rural ITS roadway environment there is at least a
conducting ground to consider, and in many cases the analysis would involve
00.5 radio-wave propagation. This is far from a free-space condition. An FM
broadcast system for disseminating information to vehicles would be affected
more than an AM broadcast system by the buildings and trees in the propagation
environment.
Thus, radio-wave propagation modeling in the ITS roadway environment will
require some special considerations to account for all of these phenomena.
POTENTIAL FREQUENCY BANDS FOR ITS
Many frequency bands (ranging from below the AM broadcast band to millimeter
wave frequencies and all frequencies in between these limits) have been proposed
for all different applicatiois of ITS including: toll collection, vehicle location and
identification, traffic information dissemination, and collision avoidance. In addi-
tion, the actual frequencies in these bands at which manufacturers are planning to
build the subsystems of ITS for the Federal Highway Administration (FHWA)
operational tests have also been taken into consideration.
The frequencies being proposed for use by ITS are a result of: what fre-
quency bands are available in the radio spectrum, what frequencies are eco-
nomically supported by the technology, what frequencies and systems are pro-
posed for the FHWA operational tests, the beamwidths and antenna pattern
requirements, and the propagation effects at these frequencies and bandwidths.
That area of the frequency spectrum would also have to be available or reas-
signed for use. The physics of radio-wave propagation would suggest potential
bands of the frequency spectrum where specific functions of ITS could be
292 N. DE MlNCO AND C.L.HOLLOWAY
performed. Certain areas of the radio frequency spectrum would be suitable for
collision avoidance and other areas would be more appropriate for position
location,, dissemination of information to vehicles; communication between
vehicles and the roadside, emergency beacon location, toll collection, and vehicle
identification (see Figure I). In-vehicle information and route guidance could
be provided using vehicle-to-roadside communication, or it could be broadcast
over subcarriers from commercial FM and AM broadcast stations. These broad-
casts would provide traffic information. Position location for these functions
could be obtained from the Global Positioning System (GPS). Automatic

vehicle identification and toll collection could be provided using vehicle-to-
roadside communications with interrogators and transponders in the vehicle
and at the roadside terminals. Two-way communication in vehicles would permit
vehicles to act as traffic probes to the Traffic Management Center (TMC) for
indicating traffic conditions. This two-way communication could be via
vehicle-to-roadside, roadside-to-TMC, or directly between vehicles and TMC.
The information on frequencies planned for use in the subsystems of ITS is
based on the activities of the ITS Communications Committee. A frequency
requirements matrix was generated from four activities of this committee: the
Transportation Research Board (TRB) Workshop in San Francisco (14-16 July,
1993); the Communications Spectrum Task Force; the 'A' Team from the Com-
munications Committee; and the Group of Experts Meeting (March 16-17.1994).
Many of the Federal Highway Administration's operational tests are using sys-
tems at the frequencies indicated in this matrix. The matrix will not be repeated .
here, because of its size and complexity, but the information taken from it will
be discussed to indicate the frequency ranges that are important to ITS and the
appropriate propagation models for analysis and performance prediction. The
applications and functions in the ITS Communications Committee matrix are
discussed by frequency in the following paragraphs.*The activities of the ITS
Communications Committee has decided that these bands would be useful for
ITS applications.
Applications at Frequencies of 40 GHz and Above

These applications include both longitudinal and lateral automatic collision avoid-
ance radar. These radars would provide warnings to the driver in the event that
another vehicle is too close (on either side or in the front or rear). These radars
would be operated for close ranges of meters to hundreds of meters. Collision
warning systems.are currently being developed by commercial companies using
millimeter-wave technology at 77 GHz that will be installed in the front of
vehicles (Schumacher et al., 1996).
Applications in the 5800-MHz Band
The 5800-MHz band currently used for Industrial, Scientific, and Medical (ISM)
is a likely candidate for many ITS applications, since the frequencies are
available for ISM and Part 15 (National Archives and Records Administration,
1994) uses. The 5800 MHz band has the highest equipment cost, but at present
there is very little ISM or Part 15 usage. Two-way communications will be
needed for such applications as: automatic vehicle identification and location,
driverlvehicle status reporting, electronic credentials (commercial), electronic

toll collection, and hazardous material access control and warning beacons.
One-way communications will be needed for such applications as: driverlvehicle
status reporting, electronic siren, in-vehicle signing, parking information. This
could be covered by cellular or Personal Communications Services (PCS), and
signal priority (emergency and transit vehicles). The distances involved are less
than 100 meters. The FHWA is anticipating the use of the 5850- to 5925-MHz
band for dedicated short range communication systems that include vehicle-to-
roadside communication and automated toll collection. Wireless local area
networks (WLAN) are also being considered in the ISM band (5800 MHz f 75
MHz) for a variety of ITS applications. Electronic Toll and Traffic Management
(ETTM) systems are currently being tested in Europe (Elliot and Dailey, 1995).
These systems currently identify cars passing through toll plazas and charge
driver's toll accounts.
Applications in the 2450-MHz Band

The 2450-MHz band currently used for ISM is a likely candidate for many ITS
applications, since the frequencies are available'for ISM and Part 15 uses. The
2450 MHz band has moderate cost equipment, but there are many microwave
ovens near the center of this band. The Government is giving up part of the
lower end of this band. A specialized data link will be needed to coordinate
evasive actions for automatic collision avoidance. Traffic probing would require
an outbound poll to the vehicle and an inbound response from the vehicle. The
2450-MHz band is to be used for WLAN applications for ITS. E'ITM applica-
tions for ITS will also use this band.
Applications at Frequencies Greater than 1 GHz (Private Microwave

Bands)
Closed-circuit television communications and control functions will require two-
way communications as used in traffic management centers and other areas of
ITS. The PCS band (1.850 to 2.200 GHz) with cell sizes from 1,000 ft. to 2.5
miles could provide an inexpensive and convenient means of transmitting data
for users demanding immediate access or mobility in the roadway environment.
This type of function would be used by field crews relaying information to a
central Pepartnient of Transportation (DOT) office or from commuters to high-
way information centers.
Applications in the 902- to 928-MHz Band

The 902-.to 928-MHz band currently used for ISM applications is a likely
candidate for many ITS applications, since the frequencies are available for ISM
and Part 15 uses. The 902- to 928-MHz band bas the most inexpensive equip-
ment implementation, however, because it contains the greatest number of users,
it will probably be too crowded in the near future with ISM, Government radars,
and Part 15 devices. Manufacturers prefer this band due to the low cost of
equipment in this band compared to other bands. Automatic vehicle identifica-
tion will require both one-way and two-way communications. Automatic vehicle
location will require only one-way communications for the application in this
band, since it is a beacon implementation. Many ETTM systems in the United
States are currently using this band, but crowding problems in this band could
make the higher bands more suitable in the future.
Applications in the 150-,450-,800-, and 900-MHz Bands

These frequency bands are selected as the existing land mobile radio bands.
Two-way communications will be needed for such applications as: automatic
vehicle location, closed-circuit television communications and control, dispatch-
ing (commercial, emergency, transit, transit-handicap), driverlvehicle status
reporting, and centralized route guidance. The automatic vehicle location service
utilizing two-way communication is a different application than that mentioned
above for the 902- to 928-MHz band. One-way communication applications
include: changeable message sign control and driverlvehicle status reporting.
Both wireline and radio communications will be used for in-road and road-
side detector reports, as well as traffic signal control. These bands will also
include transit schedule status reporting from transit vehicle to dispatch center.
Cellular telephone currently uses the 800-MHz band along with Cellular Digital
Packet Data (CDPD) systems (824- to 849-MHz and 869- to 894 MHz bands)
and voice communications. Radio data networks such as the Advanced Radio
Data Information Service (ARDIS) in the 800-MHz band show promise for
making economical data transmission available with tested technology
(Elliot and Dailey, 1995).
Applications in 220-MHz Band
One-way communications will be needed for such applications as: in-vehicle

signing, terrainhazard warnings, and variable message sign control. An auto-
matic air-bag sensor MAYDAY application planned for this band will require
two-way communications. A traffic probing application will use an outbound
poll to and an inbound response from the vehicle. ITS has been allocated 5
channel pairs in the 220-MHz band for one- and two-way applications.
Applications in the 88- to 108-MHz Band (FM Broadcast Band)

A one-way communications function for changeable message signs control will
be used in this band. An FM broadcast application to provide an outbound
channel of information is planned to provide incident reporting to vehicles.
Traffic probing will also be performed with an outbound poll to the vehicles.
Several FM subcanier systems are currently being tested in the United States.
These include the Radio Broadcast Data System (RBDS) and the Subcarrier
Traffic Information Channel (STIC) (Mastrangelo and Rust, 1996). An FM
subcanier application that is currently being used extensively in Europe is the
Radio Data System (RDS) (Whelan, 1995).
Other Systems, Applications, or Functions not lncluded Above

One implementation of the Advanced Traveler Information System (ATIS) is
being tested but is not in the matrix. This is the Herald AM subcarrier system
for Rural ATIS applications that is under development and testing by the Enter-
prise Consortium that includes the Colorado Department of Transportation and
the Iowa Department of Transportation. The ranges involved will be fractions of
a kilometer to many kilometers. This system can be analyzed using a combina-
tion of ground-wave and sky-wave models, because it is in the AM broadcast
band (535- 1605 kHz).
Other systems under test and or development by manufacturers are listed in
a report prepared by the National Institute of Standards and Technology, Robot
Systems Division (Scott et al., 1994). This report discusses 15 proposed systems
to be used for vehicle-to-roadside communications. Some of these frequencies
are mentioned in the matrix and some are not. Eleven of these systems are
in the 902- to 928-MHz band. One of these systems requires operation at
49.86 MHz in addition to the 902- to 928-MHz band. Another of these systems
operates at 845-850 MHz. Of the remaining three systems one specifies no
frequency and the other two operate at 199 and 1.8 MHz. These systems operate
. at ranges from 1 to 1610 meters. Under certain conditions they would require
special consideration for propagation analysis, because they are not in the
far field of the antenna systems. Many of these situations must be treated as
coupling problems in the near field of the antennas.
BASIC CONCEPTS IN WAVE PROPAGATION

In all of the radio-wave propagation concepts that are discussed, it is assumed that
the propagation is taking place in the far field of the source and receptor antennas.
The far field of an antenna operating at a frequency corresponding to a wave-
length (h) having an effective antenna diameter (D) is governed by the following:
Locations at distances greater than (d) are in the far field of the antenna. An
antenna has a static field, induction field, and a radiation field. All three of these
fields exist in the near field. Only the radiation field exists in the far field. In the
near-field of the antenna constructive and destructive interference phenomena
predominate and as a result the power density does not decrease monotonically
as the distance from the antenna increases. The power density oscillates with
distance from the antenna in the near field.
Free Space Propagation

If a point source radiates power (p,) with a transmitting antenna that has a
directional gain (g,), then the received power for a receiver antenna of gain (g,)
at a distance (r) away is given by the following (Friis, 1946, 1971):
where (h) is the wavelength in meters of the electromagnetic wave in the

medium corresponding to the frequency (f) in Hertz, and (c) is the speed light
in the medium in meters per second. This equation is usually expressed in dB,
and is given by (ITU, 1986; Norton, 1959):
where (P,) and (P,) are the received and transmitted powers respectively in dB,
and (G,) and (G,) are the receiver and transmitter antenna gains respectively in
dB. (Lfsb) is the free-space basic transmission loss in dB and is given by the
following (Rice er al., 1967):
LBb= 32.45 + 20 IogCf,,,) + 20 log(r) (dB) (4)

where (f& is the frequency of the radio wave in GHz and (r) is the distance
between the transmitter and receiver in meters. Free-space loss is a theoretical
reference or limit, against which the actual basic transmission loss (Lb) can be
compared or expressed as a ratio.
Equation (3) can be used to calculate the received power for a transmitter and
receiver located in free space. Free space is a highly idealized environment and
is not the situation for ITS propagation channels. In the ITS environment there
is at least a conducting ground to consider and in many situations the wave may
have to propagate through a very complicated environment. For example, the
energy may have to propagate: through or reflect from precipitation, through an
atmosphere with oxygen and water vapor, through buildings, around comers,
down side streets, into and around wooded areas, and over rough terrain. In this
type of environment the received signal can be expressed by the simple equation
given in (3) if the free-space basic transmission loss (Lfsb) were replaced by the
basic transmission loss (Lb) computed by propagation models that take all of
these propagation effects into account. Basic transmission loss is the transmis-
sion loss that would occur if the antennas were replaced by isotropic loss-free
antennas with the same polarization as the real antennas (ITU, 1986; Norton,
1959). The remaining part of this discussion describes the different types of
radio-wave propagation phenomenon that need to be considered for general
radio-wave propagation models in the roadway environment. Radio-wave propa-
gation prediction models generally compute basic transmission loss (Lb) which
is combined with antenna gains and transmitter power to perform engineering
analyses of communication systems.
Attenuation Due to Precipitation

For radio link systems, precipitation attenuation may be significant depending
on the radio frequency and the rate of precipitation. For general ITS propagation
models it is important to have the capability to account for this added loss if
necessary. The attenuation due to precipitation can be determined by subtracting
the additional loss term (Lp). the loss due to precipitation from equation (3).
Values for (Lp) have been measured for different precipitation ntes and for
different radio wave frequencies (Boithias, 1987; Doviac and Zrnic, 1993;
Shibuya, 1987). Figure 2 illustrates the dependance of Lp on precipitation rate,
relative humidity, and frequency. It shows the attenuation due to rain at different
rainfall rates, relative humidity, and frequencies. Figure 3 shows the attenuation
due to clouds at three different temperatures.
100
m
5 lo
i
0
53
z
[o.lo
MOIST AIR + RAIN
r-*o1.3 "P. T.O t
25
0 50 100 150 200 250 300 350
FREQUENCY. GHz
FIGURE 2 Attenuation due to different rainfall rates for GHz frequencies (from page M 8 (Liebe,
1989)).
FREOUENCY - GHz
FIGURE 3 Attenuation due to cluds for 1 dm3 cloud water content at 20, 2, and -20 Celsius (from
p. 5 (Allen. 1983)).
Attenuation Due to Water Vapor and Oxygen Absorption

Attenuation due to clear-air absorption by oxygen and water vapor can have
a significant effect on communication systems above approximately 4 GHz.
This attenuation can range from less than one d B h at 4 GHz up to about 15
dBkm at 100 GHz as shown in Figure 4. Note that the absorption is highly
dependent on frequency, with maximum absorption occurring near certain
molecular resonances. Many ITS applications could use the 5.8 GHz band
where the attenuation due to water vapor absorption is negligible and the
attenuation due to oxygen absorption is less than 0.01 d B h . The distances
for these applications are also less than 100 meters, so water vapor and oxygen
absorption will be negligible for this band. Oxygen absorption occurs at a
series of frequencies in the 57- to 63- GHz band, where the attenuation can
be as high as 15 dBkm. The attenuation due to oxygen absorption is still
appreciable in the 70- to 80- Ghz band where ITS collision avoidance radar
is expected to be deployed at 77 GHz. The distances are short but the attenu-
ation is high (as much as 6 dB/km). Another possible band for ITS collision
avoidance radar is the 24-GHz band which is near the 22.3 -GHz water vapor
resonance with .1 dBkm attenuation, but the distances for ITS applications in
this band are short.
0 50 1W 150 200 250

FREQUENCY - - GHZ
FIGURE 4 Attenuation due to clear-air absorption by oxygen and water vapor (from p. 38 (Liebe.
1983)).
The phase and amplitude response of an electromagnetic wave interacting

with the atmospheric medium can be determined by calculating the complex
refractive index of the medium. The complex refractive index (n) of the
N. DE MlNCO AND C.L.HOLLOWAY
atmosphere can be used to compute the phase and amplitude of a plane electro-
magnetic wave propagating a distance (r) at frequency (f) as (Liebe, 1989):
where (EO)is the initial wave amplitude, (c) is the speed of light in a vacuum, and
j = 0. Lc,(dB) is the attenuation due to clear-air absorption and is given by:
(-&)
LC, = 20 log (dB)'
This loss is also subtracted from equation (3)
Radio Waves in The Vicinity of Planar Surfaces

When radio waves propagate down urban canyons or the short-range roadway
environment, they inevitably will be incident onto buildings and onto the ground.
Depending upon the nature of these surfaces the reflected wave will behave in
different ways.
Reflections from smooth perfectly conducting surfaces

Waves incident onto a smooth surface will be reflected in the specular direction.
The angle of reflection equals the angle of incidence. The reflection coefficient
( r ) for a smooth perfectly conducting surface is: rE= -1 for an E-field perpen-
dicular to the plane of incidence, rH= +I for an E-field parallel to the plane of
incidence.
Reflections from smooth surfaces

If the surface is not a perfect conductor, then the amount of energy reflected in
the specular direction is related to the composition of the material and the
polarization of the incident wave. The reflection coefficient (TEof an E-field
perpendicular to the plane of incidence (E-polarization) and incident at an angle
Y measured from the surface is given by (Brown er al., 1973):
sin Y - (c, - i60 o h - cos2 ' # ' ) I n

r - (7)
- sin Y + ( 6 , - i60 o h - cos2 ' # ' ) I n
The reflection coefficient (rH) of an E-field parallel to the plane of incidence

(H-polarization) incident at an angle Y measured from the surface is given by
(Brown et al., 1973):
rH= ( Q - i60 o h) sin Y - (6, - i60 o h - cos2 Y ) ' ~

(6, - i60 o h ) sin Y - (.E, - i60 o h - cos2 y)ln
where (6,) is the relative permittivity of the surface, (o) is the conductivity of
the surface in Slm, and (h) is the wavelength in meters corresponding to the
frequency (0.
Some of the reflecting surfaces encountered in ITS channels will be
composed of concrete. Common examples of this are buildings and road

surfaces. Figure 5 illustrates the behavior of the reflection coefficient for two
polarizations for a infinitely thick slab of concrete at 1500 MHz (see (Achatz,
draft report)). The results shown in this figure were calculated assuming that for
dry concrete f , = 6.57 and o = 1.95 * Slm (Westphal and Sils, 1972).
Incident Angle (degrees)
FIGURE 5 Reflection coeficients for both an E-polarized (rE)and H-polarized (rH)

field incident
onto an infinite thick slab of concrete.
Reflections from rough surfaces

There are three different scenarios that are encountered with reflections from
rough surfaces. The limits of these are governed by the ratio of the dimension
of the roughness of the surface to the wavelength (1) of the radio wave, and are
discussed below.
1. Roughness dimension very small compared to a wavelength
When the roughness dimensions of the surface are very small compared to a
wavelength (R c .OIL), the surface can be treated as a smooth surface. For this
particular case, the reflection coefficient is given in equations (7) and (8).
2. Roughness dimension small compared to a wavelength (R c 1 A) .

When the roughness dimensions are small compared to a wavelength (R < .lh),
the rough surfaces can be analyzed with an effective boundary condition. The
effective boundary condition takes into account the roughness effects.
The effective boundary condition for a perfectly conducting rough surface is
treated in (Sivov, 1964; Vainshtein, 1963; Boit, 1973; Verbitsby, 1976; Holloway,
1992; Holloway and Kuester, 1994a; 1994b). For this problem it is assumed that
all the energy is reflected in the specular direction. Therefore, the magnitude of
the reflection coefficient is one. Incident energy is temporarily stored in the
roughness profile, causing the phase of the reflected energy to be different than
,180 degrees (as for a smooth perfectly conducting surface). The phase of this
reflection coefficient is a function of the dimensions of the roughness profile as
well as the incident angle. Figure 6 shows the variation of the phase of the
reflection coefficient for a perfectly rectangular rough profile (Holloway, 1992;
Holloway and Kuester, 1994a). Figure 6 compares this work with that of other
researchers (Sivov, 1964; Verbitsky, 1976).
The effective boundary condition for a highly conducting rough interface
is treated in (Holloway, 1992) and (Holloway and Kuester, 1994b). For this
problem it is assumed that some fraction of energy penetrates the rough interface
and the remaining energy is reflected in the specular direction. The phase of the
reflection coefficient is a function of the roughness profile, the angle of
incidence and the conductivity of the surface. Figure 7 shows results for a
highly conductive interface (Holloway, 1992; Holloway and Kuester, 1994b).
The parameter p/& in Figure 7 is the ratio of the horizontal roughness dimen-
sion to the skin depth of the surface.
3. Roughness dimension large compared to a wavelength
When the fields are incident onto a rough surface whose roughness dimensions
are large compared to a wavelength (R > A) and the horizontal dimensions of
the roughness profile of that surface are small compared to a wavelength, then
the fields inside the rough troughs cannot develop into higher order propagation
modes, therefore, the energy is reflected in the specular direction. On the other
hand, when the horizontal dimensions of the roughness profile are large
compared to a wavelength, an incident field can cause higher order modes to
propagate. When this occurs, the energy is not only carried in the specular
direction, but propagates in other directions. Consequently, the magnitude of the
specular reflection coefficient for a rough perfect conductor is no longer one.
-
&160.00
W
u
- - HK THEORY
FIGURE 6 The phase of the reflection coefficient for a perfectly conducting rectangular profile,
where the dimensions of the profile are small compared to a wavelength (from p. 169 (Holloway,
1992)).
0.00 20.00 40.00 60.00 80.08

Angle of Incidence (degrees)
FIGURE 7 The phase of the reflection coefiicient for a highly conducting rectangular profile,
where the dimensions of the profile are small compared to a wavelength (from (Holloway and
Kuester, 1994b)).
A plane wave is incident onto the surface shown in Figure 8. The phase
difference between two rays of the field is (Brown et al., 1973):
where
A1 = the difference in total distance traversed by the two rays as a result of
reflection from the rough surface,
p = 2x/h.
If PA1 is small, the surface can be approximated as smooth. However, if

PA1 = x, no energy is reflected in the specular direction. When PA1 is large, this
corresponds to the condition where roughness dimension is large compared to a
wavelength (R > h). For this scenario, the energy is reflected into non-specular
directions. The behavior of the magnitude of the specular reflection coefficient
as a function of surface dimensions is described as periodic as a function of
surface roughness depth and the period of the horizontal roughness dimensions
(Beckmann and Spizzichino, 1987).
FIGURE 8 Illustration of the phase difference for a wave reflection off of a rough interface
(from p. 409 (Brown er a/., 1973)).
There have been a number of theoretical approaches analyzing electromag-

netic (and the mathematically similar acoustic) field near rough surfaces. The
first approach was probably by Rayleigh, which was later refined and extended
by many scientists (good accounts of this research may be found in (Bass and
Fuks, 1979; Beckmann and Spizzichino, 1987; Wait, Wenzel, 1974; Rosich and
Wait, 1977, 1985)). For good accounts of periodic rough surfaces see (Beckmann
and Spizzichino, 1987; Ishimaru, 1991; Kong, 1986).
Diffraction from Objects

Radio waves propagating down streets will encounter buildings, doors,
windows, automobiles and other obstacles with corners. The theory of geometric
optics and diffraction are used together to analyze and account for the total
electromagnetic field in these situations. The theory of geometric optics assumes
that radio waves travel in straight lines except for the effects of reflection and
refraction. Geometric optics alone can not accurately account for the electro-
magnetic field propagation around these comers. The effects of diffraction must
be considered. Diffraction effects are those effects that cause the propagation of
radio waves to deviate from straight lines due to the influence of effects other
than reflection and refraction. Diffraction effects usually occur with the passage
of radio waves by the edges of object surfaces that are opaque to radio waves.
Diffraction is the bending of radio waves around the edges of these opaque
objects or around the edges of apertures existing in these opaque objects. Opaque
objects include: certain types of buildings, tollbooth structures, bridges, auto-
motive vehicles, and other obstructions occurring in the roadway environment.
For example, assume that a field is incident onto the comer shown in Figure 9.
This could be the edge of a building in the urban canyon environment. The space
in this figure is divided into three regions. The fields in region I consist of the
incident field, the reflected field, and fields due to the comer effects (the
diffracted field). Region U is the reflected ray shadow boundary (RSB). The
fields in this region consist of the incident field and the fields due to the comer
effects (the diffracted fields). Region 111 is the incident ray shadow boundary
(ISB). The fields in this region consist of the diffracted fields only.
Tall building edge (90 degree wedge)
t' 1 ngle of incidence

i n g l e of observation
IS6 = Incident shadow boundary
RSB = Reflection shadow boundary
Region I = LOS, Reflection. Diffraction rays
Region I1 = LOS, Diflmction rays
Region Ill = Diffraction rays
FIGURE 9 Diffraction o f a ray by a comer

306 N. DE MINCO AND C.L.HOLLOWAY
Using geometric optics one would conclude that region I consists of only
incident and reflected fields and region I1 consists of only incident fields, and
region 111 would be null of fields. The Uniform Theory of Diffraction (UTD)
(Kouyoumjian and Pathak, 1974; McNamara et al., 1990) must be used to
accurately model the effects of the comers. The effects of comers modeled by
UTD are then combined with the geometric optics field to model the total radio-
wave propagation in the roadway environment.
The diffraction fields are expressed as:
where (E') is the incident field at the diffraction point, (s) is the distance from
the diffraction point to the observation point, (k) is the wavenumber, and (D) is
the dyadic diffraction coefficient. (D) is a function of the incident angle, the
observation angle, the wave number (k), the distance to the observation point
(s), the geometry of the edge, and the material makeup of the edge.
Expressions for (D) for perfectly conducting wedges and comers for both
two-dimensional and three-dimensional situations are available (McNamara
et al., 1990). Perfectly conducting materials do not represent the properties of real
wedges and comers with finite conductivity. Impedance wedges or comers have
finite conductivity representing real materials. Two different expressions for (D)
for impedance wedges or comers have been derived (Maliuzhinets, 1958; Leubbers,
1984). The first is a rigorous solution for the impedance wedge (Maliuzhinets,
1958). while the second is based on a heuristic solution (Leubbers, 1984). Results
of a comparison of these expressions are reproduced in Figures 10 and 11. The
vertical axes in these figures represent relative diffracted field in decibels. The
geometry for the angle of incidence ( @ I ) is defined in Figure 9. These comparisons
show that depending on the location of the source with respect to the building
comer, the difference between the two expressions can be significant.
Another consideration is building materials. Comers of buildings may behave
more like an impedance comer as opposed to a perfectly conducting comer,
depending on the building materials utilized. We have made comparisons
for the diffracted field at a comer of a building assuming both a perfectly
conducting comer (McNamara expression (McNamara et al., 1990) for D) and
an impedance wedge (Maliuzhinets rigorous expression (Maliuzhinets, 1958)
for D and Luebbers heuristic expression (Leubbers, 1984) for D), the results are
shown in Figures 10 and I1 for two different source locations. These figures
demonstrate that correctly representing the comers of a building, either as a
perfectly conducting corner or as an impedance corner, is important to
accurately predicting diffraction around buildings.
I
-- -._ , ;,'
... , , . I
I
: , ;
,, ., -
. I .
-80.00'
,'I , _ _ _ PI ms rpi eecdtalny cCe oWn de du gc tei n(gr i gWoer do gu se)
i ; -----.
, I Impedance Wedge (heuristic)
-100.00
8.00 50.00 100.00 150.00 200.00 250.00
0 (degrees )
FIGURE 10 A comparison of the diffracted fields from a perfectly conducting comer to the
diffracted field assuming an impedance comer using both the rigorous expression (Maliuzhinets.
1958) and the heuristic expression (Luebbers. 1984). These results are for a source location of
@' = 5 degrees.
-Perfectly
- - - I m p e d a n c Ce oWn de du gc tei n(r8gorous)
g Wedge
.-----
I m p e d a n c e W e d g e lheurlstoc)
-180.00
0.00
0
50.00 100.00 150.00 200.00 250.00
,
0 ( degrees )
FIGURE I1 A comparison of the diffracted fields from a perfectly conducting comer to the
diffracted field assuming a impedance comer using both the rigorous expression (Maliuzhinets.
1958) and the heuristic expression (Luebbers. 1984). These results are for a source location of
0' = 17 degrees.
Previously we discussed single diffraction, that is diffraction of an isolated

comer or edge. If structures have multiple edges or comers, such as narrow
buildings or narrow streets, then coupling between the comers will occur. This
type of diffraction is called multiple or higher-order diffraction and represents
an effect that is of lesser magnitude than standard single diffraction. However,
it is more important when the separation between the comers is small, and then
should be considered. For a detailed account of multiple diffraction see (Balanis,
1989; Pathak and Marhefka, 1991).
EXISTING PROPAGATION MODELS
The radio-wave propagation models developed to date include reflecting smooth

surfaces with finite conductivity and permittivity. These models also consider
the terrain contours and atmospheric refraction between a transmitter and a
receiver and perform diffraction loss computations for a propagation path. Models
have been developed for the following frequency ranges: 100 kHz to 2 MHz,
2-30 MHz, 30 MHz to 20 GHz, and 20-1000 GHz.
General Radio-wave Propagation Models

A large number of propagation models are used to analyze radio communication
systems. Some are specialized and others are more general, but all require sound
engineering judgement on which is the better model to use for a particular
analysis. The models described in this section can be used to analyze and predict
ITS performance for medium to long range distances without modification.
Short distance LOS radio-wave propagation models can be used after certain
modifications are made for near-field coupling.
Models for 100 kHz to 30 MHz

For the frequency bands of 100 KHz to 2 MHz and 2-30 MHz both the ground
wave and the sky wave must be considered. Both the sky-wave and the ground-
wave model can interface with many different antenna models. The AM broadcast
band (0.535- to 1.605-MHz) is used by the ITS ATIS application called the Herald
AM subcarrier. Another ATIS application similar to this has been proposed to
operate at an HF frequency ( s 2 MHz). These models can be used for both
designing and making performance predictions for these ITS applications.
The sky wave is the radio wave that reflects from and refracts through
different layers. of the ionosphere and returns back to Earth. Medium frequency
(MF) and high frequency (HF) communication via the ionosphere depend upon
the ability of the ionosphere to return these radio signals back to Earth. The
degree of ionization of the atmosphere determines the amount of signal that is
reflected or refracted. Ionospheric propagation models predict the ionization
levels in the various regions of the ionosphere to determine sky-wave circuit
performance. There are two different sky-wave propagation models, one for use
above 2 MHz and one below 2 MHz. The nature of sky-wave propagation is
different for each of these bands and thus two separate models must be used.
The sky wave can contribute to a desired signal and also generate undesir-
able interference to the desired signal depending on the distance between the
transmitter and receiver. Close and far distances between the transmitter and
the receiver depend on the frequency and ground conductivity. At 0.5 MHz
over average ground, the ground wave predominates over the sky wave from
the transmitter site out to distances of about 150 km, where the two signals are
equal. The signals add as vectors and destructive and constructive interference
can occur. At distances beyond 150 km the sky wave is the predominant
signal. At a signal frequency of 1.5 MHz the distance where the two signals
are equal reduces to 45 km, because of the increased loss at the higher
frequency. At even higher frequencies the attenuation of both the ground-wave
and the sky-wave signal is greater, so the distance becomes smaller where the
signal levels are equal in amplitude. If the transmitter and receiver are close
to each other, the ground-wave signal is usually predominant over the
sky-wave signal, and the ground-wave signal is normally considered the
desired signal. For large separations, the sky-wave signal predominates and is
the desired signal. In the region between close and far transmitter to receiver
separations, either the sky-wave or ground-wave signal can cause interference
to the other dominant signal. The desired signal in many cases is the ground-
wave signal. One sky-wave model (Teters et a[., 1983) calculates a myriad of
HF sky-wave propagation parameters to enable a complete system analysis to
be performed. Typical parameters calculated include: median values of
maximum useable frequency (MUF), optimum frequency of transmission (FOT),
lowest useable frequency (LUF), field strength, reliability, antenna gain, noise
power, system loss, and signal-to-noise ratio. This sky-wave model contains
the latest International Telecommunication Union Radiocommunication Study
Group (ITU-R) approved atmospheric noise coefficients, improved manmade
noise calculations, and realistic specifications/predictions for the global
variations of the f-region critical frequencies. Another model, the MF
sky-wave model (Haakinson er al., 1988) predicts sky-wave propagation
parameters using empirical ionospheric reflection coefficients for an
undisturbed ionosphere.
The ground-wave signal can be determined using one of several models that
specifically addresses the propagation phenomena at frequencies below 30 MHz.
One general ground-wave model (DeMinco, 1986) computes propagation loss,
electric field strength, received power, noise, received signal-to-noise power
ratio, and antenna factors over lossy Earth. The smooth-Earth and irregular-
Earth (terrain dependent) propagation loss prediction methods within this model
can be used over either homogeneous or mixed paths. This computer program
combines three propagation loss prediction methods for both smooth and
irregular Earth, a system interface algorithm, a reliability algorithm, and an

antenna algorithm into a single analysis tool. The propagation loss prediction
methods, system interface algorithm, and noise prediction algorithm are valid
from 10 kHz to 30 MHz.
The ground wave includes the direct line-of-sight space wave, the ground-
reflected wave, and the surface wave that diffracts around the curved Earth. The
following five computation techniques are used in this model to account for all
possible propagation geometries: flat-Earth attenuation function, flat-Earth
attenuation function with a curvature correction, the residue series calculation,
geometric optics, and numerical integration of the full-wave theory. The appro-
priate technique is selected according to the relative geometry of the transmitter
and receiver antenna heights and the Earth. Diffraction is considered where
necessary in the last four of these models. The first model, the flat-Earth
attenuation function, is used for line-of-sight propagation.
If the terrain contour is 'smooth' or the terrain irregularities are much smaller
than a wavelength, then the smooth-Earth method is mathematically and numeri-
cally accurate for the ground-wave predictions for frequencies from 10 kHz to
100 MHz, however above 30 MHz the irregularities of the atmosphere make
statistical methods more appropriate. Irregularities in the terrain have a greater
effect at higher frequencies, so an irregular terrain model is more appropriate
when terrain irregularities become appreciable in size with respect to a wave-
length. The smooth-Earth model is much more computationally efficient and
many orders of magnitude faster than the irregular-Earth model, so in cases
where the terrain is smooth enough, the smooth-Earth model can be used
with minimal sacrifice in accuracy. Specific comparisons of smooth-Earth and
irregular-Earth predictions with actual measurements have been made for
different terrain irregularities (DeMinco, 1986). When the terrain contour is
snlooth, the smooth-Earth model is valid for all combinations of antenna heights,
frequency, and dielectric constants by virtue of the computation techniques
contained within its structure. It should be used only out to the maximum
distances considered useful for ground-wave propagation at each frequency, since
the sky wave will become significant from those distances to points beyond.
PROPAGATION MODEL 1N ROADWAY ENVIRONMENT 311
The irregular-Earth, mixed-path method in this ground-wave model uses

an integral equation to compute the propagation loss of a vertically polarized
electromagnetic wave over irregular terrain that is covered with forests,
buildings, or snow. The terrain cover is modeled as a slab of user-specified
thickness, length, conductivity, and dielectric constant. Antenna heights of the
transmitter and receiver antennas without a slab are included in the irregular-
Earth propagation loss computations using the same height-gain functions as the
smooth-Earth model. When a slab is included, a special height-gain function is
used for the antennas within or above the slab. This approach is a point-to-point
prediction method valid for frequencies between I0 kHz and 30 MHz. At higher
Frequencies, other techniques must be used to account for the large variability of
the ground wave in time and space.
Models for 20 MHz to 20 GHz

One widely used propagation model (Hufford et al., 1982) is a computer method
for predicting long-term median radio transmission loss over irregular terrain
and is valid over a frequency range of 20 MHz to 20 GHz. The method is based
on well-established propagation theory and has been tested against a large num-
ber of propagation measurements. Predictions have been tested against data for
wide ranges of frequency, antenna height and distance, and for all types of
terrain from very smooth plains to extremely rugged mountains. This method
can be used with either detailed terrain profiles for actual paths or with profiles
that are representative of median terrain characteristics for a given area.
The program requires the following input parameters: frequency, antenna
heights, and either an estimate of terrain irregularity or the detailed terrain
profile. The program computes median reference values of attenuation relative
to the transmission loss in free space as a function of distance. This median
reference value is determined from a variety of times and paths when the atmo-
sphere is in a quiet state, well-mixed, and conforming to a standard atmospheric
model. The calculated reference is based on two-ray geometric optics theory for
radio line-of-sight paths with an extrapolated value of diffraction attenuation.
The reference value for transhorizon paths is the lesser of either diffraction
attenuation or forward scatter attenuation.
Radio-wave propagation in the terrestrial environment at VHF, UHF, and
SHF can be difficult to predict in the physical world, because received signal
levels vary. The received signal power varies in time because of changing
atmospheric conditions and in space because of changes in terrain. This model
is statistical and describes this variability in the median value of received signal
level and the magnitude of expected deviations from the median value.
Models for 1-1000 GHz
Several propagation models are used in this frequency range. Each has its own
specific applications with respect to geometry. Some of these models are
specific to terrestrial applications where both the receiver and transmitter are on
the ground, and others have one or both the transmitter or receiver airborne.
One terrestrial model (Allen, 1986) that is used for performing engineering
calculations on links is a complete line-of-sight model that considers Fresnel
zone clearances and antenna heights. The program predicts cumulative

distribution of propagation losses and converts this to cumulative distribution
of received signal level. A climatological database is used to interpolate the
appropriate parameters using the midpoint of the path under study. The
computations include clear-air absorption attenuation due to oxygen and water
vapor, and attenuation due to rain.
There are also specialized models for computing attenuation due to atmo-
spheric effects (Liebe, 1983; 1989). These models are also line-of-sight models,
but treat more rigorously clear-air absorption attenuation due to oxygen and
water vapor and the attenuation due to precipitation. The attenuation and propa-
gation delay effects are predicted from meteorological data sets containing
pressure, temperature, humidity, suspended particle concentration, and rain rate.
Atmospheric propagation limitations critically affect the development of milli-
meter-wave applications. Reduced radio signals are caused by adverse weather
conditions such as rain, wet snow, suspended particles, and water vapor. Radio
signal levels can also be elevated by atmospheric ducts.
A more general and flexible model than either of these two is available for
analyzing communication systems at frequencies of 1 to I00 GHz (Allen, 1990).
This model provides an integrated procedure for modeling propagation effects
on telecommunication links of various scenarios and is used for predicting
system performance. The scenarios include earth-to-satellite, ground-to-aircraft,
aircraft-to-satellite, and terrestrial. The results obtained include the effects of
rain, clear air, and multipath attenuation. Received signal levels and link margin
calculations can be performed. Terrain profiles are used for terrestrial links to
calculate attenuation due to diffraction and troposcatter.
A more recent model analyzes communication systems in a jamming or
interference environment in the 2-MHz to 20-GHz frequency range (Geikas and
DeMinco, 1994; DeMinco, 1995). This highly structured model allows for greater
flexibility and expandibility. It includes a user-created catalog of equipment,
ground stations, and aircraft and satellite platforms; the software for creating and
maintaining this catalog; a climatological database for much of the world; and
the analysis software. The analysis software includes subroutines for calculating
clear-air attenuation, rain attenuation, multipath attenuation, diffraction, and

troposcatter. Data entry for this model is simplified by user-friendly menus and
options.
Databases are created as a result of this data entry and saved as scenario
descriptions. These scenario descriptions completely characterize the communi-
cation link or jamming situation. The scenario description includes: ground or
airborne station location; equipment characteristics; and other pertinent physical
factors such as climate and terrain. Each analysis program within a scenario
analyzes the case represented by the scenario description data.

This model is organized into six scenarios. A scenario type represents either
a communication path geometry description or a jamming geometry description.
The four scenario types in the communication geometry description are: ground-
to-ground, ground-to-satellite, ground-to-aircraft, and aircraft-to-satellite. The
two scenario types in the jamming geometry description are: jamming and
jammer versus network. The jamming scenario analyzes: received jammer power
versus distance, received transmitter power versus distance, jammer footprint,
and isopower contours. The jammer versus network scenario analyzes the effects
of a jammer or interference source on up to five communication nodes. For the
jamming geometry description, the receiver, transmitter, and jammer platforms
can be on the ground or airborne. The jamming and the jammer versus network
scenarios are the major features of this model for electronic warfare and
intersystem interference analysis. The jamming scenarios of this model can be
used for analysis in the frequency range of 2MHz to 20 GHz. The 2- to 30-MHz
range includes propagation models for both the ground wave and the sky wave.
An irregular terrain model for the 20-MHz to 20-GHz frequency range has also
been integrated into the model. The other four scenario types are used as an aid
in evaluating and designing of microwave communication systems operating
between 1 and 300 GHz. They allow the user to simulate a wide variety of
propagation effects on a system.
SPECIFIC ITEMS THAT NEED TO BE ADDRESSED IN

PROPAGATION MODELS
Many of these radio-wave propagation models described above could be used
for system performance or interference calculations for the different subsystems
of ITS that use one-way or two-way communication systems in the 220-MHz,
900-MHz, 2.4-GHz, and 5.8-GHz bands. The models can also be used in the FM
broadcast band for ATIS applications.
The existing propagation models described can analyze and predict ITS
subsystem performance at medium to long range distances without modifica-
tion. A typical ITS roadway environment includes LOS and non-LOS (also
referred to as 0 0 s ) communications. Models that predict communications
performance beyond LOS around buildings and obstacles other than Earth are
needed to supplement the existing models. These models would include mul-
tiple diffractions and diffraction-reflection phenomena to predict communica-
tions performance in an urban environment. Below is a discussion of specific
items that must be addressed in developing short-path propagation models for
ITS analysis and performance predictions. Many frequency bands (ranging
from below the AM broadcast band to millimeter-wave frequencies and all

frequencies in between these limits) have been proposed for all different
applications of ITS functions including: toll collection, vehicle location and
identification, traffic information dissemination, and collision avoidance, etc.
Of particular interest here are the two-way communication applications at 900
MHz and 5.8 GHz where radio waves would be propagating in the urban
canyon environment of the downtown areas of a city.
Modeling Out-Of-Sight Propagation
Cross Streets Propagation
With this scenario the goal is to predict the signal level of a receiver placed at
location B (see Figure 12) when the transmitter is located at A. Recent work
investigated propagation down these side streets, and as one might expect, they
found that if the receiver is located close to the intersection, the reflected rays
into the side street are the dominant contribution to the sign& levels (Bergljung
and Olsson, 1991; Maciel et al., 1993; Erceg et al., 1992; Yim and Siang, 1992).
As the receiver is moved away from the intersection, the diffracted fields from
the comers of the building become the dominant contribution to the signal
levels. The total fields that were calculated by combining the reflected and
diffracted fields correspond closely to the results of measurements (Yim and
Siang, 1992).
Propagation Down Multiple Side Streets
It will be necessary to incorporate the ability to predict the multiple side street
OOS propagation. That is, to be able to predict the signal level of a receiver
located at either point C, D or E with a transmitter located at point A (see
Figure 12). The majority of signal strengths are from diffracted fields around
the comers of the buildings at the various intersections, and therefore it is not
FIGURE 12 Illustration of side street receiver locations. A represents the location of the transmitter.
B,C,D,E represent receiver locations.
necessary to include the multiple reflections off the walls. The model should
include ground reflections of the diffracted fields in the model.
Impedance edge and comer diffraction

Depending upon the building materials used, the comers of a building do not
appear as perfectly conducting corners, but more like impedance corners.
Previously, we compared the results of the diffracted fields for both a 90 degree
perfectly conducting comer, and a 90 degree impedance comer. These compari-
sons (see Figures 10 and 11) illustrate that as much as a 40-dB difference in the
diffracted fields exist for the two cases. This suggests that large errors in the
estimated diffracted fields can be achieved if the wrong type of edge or comer
condition is used. It is therefore necessary to incorporate the option of diffrac-
tion from an impedance comer into the existing propagation model.
Diffracted fields from corners to street to receiver

In current propagation models the diffraction effects are handled as follows: The
field at the receiver due to diffraction is obtained by only considering a wave
from the transmitter hitting the comers of a building and diffracting straight to
the receiver. The model does not consider a wave diffracting from the comers,
316 N. DE MlNCOAND C.L. HOLLOWAY
then striking the street, and then hitting the receiver. The amplitude of this
diffracted field is approximately the same as the diffracted fields from the comer
hitting the.receiver directly. The difference will be found in the phase. This is
similar to the effect of the LOS propagation plus reflection from the ground; that
is, the difference in the phase of the two rays causes an interference pattern in
the total received signal. Therefore, if both the direct diffracted fields and the
reflected diffracted fields are considered, an interference pattern in the total
diffracted fields would be seen depending on the antenna heights and the radio
frequency. We have investigated this effect, and the results are illustrated in
Figures 13 and 14. In these figures, the total diffracted field (the direct diffracted
field plus the ground reflections) is plotted as a function of the receiver distance,
measured from the comer of the building for various frequencies and antenna
heights. Also shown on these figures is the LOS diffraction. Depending upon the
frequency and antenna height, the total diffracted fields can differ from the LOS
diffracted field.
-LOS w i t h Ground Reflection

- - -- - - LOS
FIGURE 13 Comparison of the diffracted field signal levels obtained from just the LOS diffracted
field and from the combination of both the LOS diffracted field and the ground reflected diffracted
field. These results are for f = 5 GHz, transmitter height of IOm, and a receiver height of 2 m.
- LO5 with Ground Reflecfion

--.--- LO5
FIGURE 14 Comparison of the diffracted field signal levels obtained from just the LOS diffracted
field and from the'combination of both the LOS diffracted field and the ground reflected diffracted
field. These results are for f = I CHz, transmitter height of ]Om, and a receiver height of 2 m.
The contribution of the street reflected diffracted fields to the total signal
levels may help explain the differences in the predicted to measured signal levels
obtained by (Yim and Siang, 1992). In this model, predictions of the signal
levels down a side street were calculated using the reflections from walls and
only the LOS diffracted fields. For large distances down a side street the majority
of the signal level is from the total diffracted fields only. The results of these
computations indicate that the predicted signal levels match reasonably well with
measurements. However, the predicted results appear to have a different slope
than the measured data at large distances down the side street. Figures 13 and 14
show that by considering both the LOS diffraded field and the reflected and
diffracted field the slope of the signal levels change over that of only the LOS
diffracted fields. Not considering the total diffracted fields could very well
explain the difference between the results obtained by (Yim and Siang, 1992).
Multiple diffractions from closely spaced corners and objects

It was shown previously that ifcomers ofbuildings are spaced close together, then
a second-order diffraction known as 'multiple diffraction' will result. In general,
this type of diffraction has less effect than standard single diffraction. However, it
plays a bigger role when the separation between the comers is small, and it should
then be considered. Multiple diffraction may be more important when predictions
down narrow side streets are needed.
Propagation over Rough Streets

All of the propagation models assume that the street is a smooth perfectly
conducting plane. In most ITS applications this assumption is not valid. Propa-
gation of radio waves down urban canyons will occur when the streets are
congested with automobiles. The road surface no longer appears as a smooth
surface, but rather as a rough surface due to the presence of automobiles as
shown in Figure 15. Given the automobile dimensions and the radio-wave fre-
quencies to be used for ITS (0.0517 m at 5.8 GHz, 0.33 m at 900 MHz, 0.67 m
at 450 MHz, 1.36 m at 220 MHz, 2.00 m at 150 MHz, and 3.00 m at 100 MHz),
the roughness dimensions of this rough surface are large compared to a wave-
length. It was shown previously that a wave reflecting in the specular direction
from a rough surface with large dimensions compared to a wave length has a
reflection coefficient less than one and a phase not equal to 180 degrees. It was
also shown that if the dimensions of the rough surface present the appropriate
geometry, then very little of the energy is reflected in the specul& direction, but
is reflected in the nonspecular direction. If these vehicles are moving along the
street, then the rough moving surface would cause significant multipath effects.
Approximately
FIGURE 15 Illustration of a realistic street surface. Due to the automobiles, the street appears as
a rough surface.
In Figures 16 and 17, we show results of a signal level at a receiver for a LOS
ray and a reflection off the ground, assuming a phase of the reflection coefficient
from the ground obtained from Figure 6. The phase of the reflection coefficient
given in Figure 6 is for the roughness dimensions that are small compared to a
wavelength, and these do not correlate to the realistic roughness dimensions for
roadway propagation. However, using these results gives an indication of the
roughness effects on signal levels. From Figures 16 and 17 we see that the
roughness effects not only shift the null of the signal level, but also cause the
null and maximum to be less pronounced.

It will be necessary to develop a set of curves for the magnitude and phase
of the reflection coefficient for typical or realistic roughness dimensions that
would be encountered in an ITS setting. These reflection coefficients for a
nonsmooth street could then be incorporated into the existing propagation model.
Optimal Antenna Height

Based on a brief study (see Figures 16 and 17), fading of the total signal levels
due to the ground reflection can be minimize'd, depending upon frequency and
antenna height. Further investigation is necessary to determine the optimal
transmitter and receiver height for an ITS system.
-SFree Space
m o o t h Ground
------ Rough Ground R = 0 . B
--_
-200.00
0.00
s 50.00 100.00 150.00 200.00
Down the Street Im)
FIGURE 16 Illustration of the fading of a signal due to the LOS and ground reflection. These
results are for I GHz and the transmitter and receiver height equal to I m.
N. DE MINCO AND C.L. HOLLOWAY
Free S p a c e
- - Smooth Ground
rn
------ Rough Ground R=0.8
2 -50.00:
L
;
0
a
-100.00-
C
.-vl0
-
0.00 0.50 100.00 150.00 200.00

Down the S t r e e t Iml
FIGURE 17 lllustntion of the fading of a signal due to the LOS and ground reflection. These
rcsults are for a I GHz and the transmitter and receiver height equal to 2 m.
Reflections from C o m p o s i t e Walls for S h o r t P a t h Propagation

Modeling
For short propagation paths, such as those found in ITS environments, correctly
representing reflections of electromagnetic energy from surfaces is critical for
accurate prediction of signal characteristics. In this section models to calculate
the reflection and transmission coefficients for block walls and other types of
materials are presented. The importance of accurately representing the reflec-
tions for signal level prediction models is also investigated here. It is shown that
the 5 to 10 dB error in received power can occur if the composite walls are not
handled appropriately.
When the wavelength is on the same order of magnitude or larger than the
dimension of the composite structure, the electromagnetic field sees the blocks as
an equivalent effective media. The effective material properties for the block wall
of Figure 18 (Note: the direction of y is out of the page) are given by the follow-
ing (Holloway et al., 1995a; Holloway et al., 1995b; Holloway er al., 1996):
where g = alp is the relative volume of space occupied by the material, f, and
pa and the complex permittivity and permeability parameters, respectively of the
bulk material, and f o and p, are the free space values for permittivity and per-
meability, respectively.
FIGURE 18 Illustration of a concrete block wall
For the two-dimensional structure in Figure 19 the effective material

properties' are given by the following:
where f = a2/p2is the volume fraction of space occupied by air.
FIGURE 19 lllustration of a two-dimensional periodic block wall

N. DE MINCOAND C.L.HOLLOWAY
/
x-axis o r i e n t a t i o n
/
-- Solld wall /
/
/
FIGURE 20 Reflectiiity versus angle of incidence for a perpendicular polarized wave. These
results are for a 6 inch block wall with slabs oriented along both the y-axis and x-axisivith f = 900
MHz and for I2 = 2.25 cm, p = h,E, = 3.0 and o = 1.95 * lo-', a = 2.6 cm, d = 14.3 cm, and
I, = 10.0 cm. The large dashed curve represents the results for a single layered slab of thickness
equal to 212 + I,, the solid curve represents the actual concrete block wall with the slabs oriented
along the y-axis, and the small dashed curve represents the results for the actual concrete block wall
with the slabs oriented along the x-axis.
Figure 20 shows results for the reflectivity (defined as the magnitude

'
squared of the reflection coefficient) of a 6-inch block wall for a perpendicu-
larly polarized E field as a function of incidence angle, for a frequency of
900 MHz. The results are for an E-field aligned along either the vane of the
block or perpendicular to the vane. Also shown in this figure are results for
a solid layer of concrete of thickness equal to 12.25 cm. From this figure one
can see that the resonant behavior of the reflectivity of the wall cannot be
achieved if a composite wall (block wall) is approximated by a solid layer.
In Figure 21 we show results for the total received power as a function
'
of antenna spacing for an antenna placed above a perfectly conducting ground
and between two walls. The received power was calculated by assuming that
four different rays contribute to the total power; the direct path, the ground
reflection, and one reflection off each of the two walls. For distances be-
tween 5 and 30 m the results for the composite wall show a received signal
that is 5-7 dB less than the results for the other two types of walls; as much
as 10 dB less signal has been seen in some examples. Results for different
size concrete block walls and for two-dimensional structures show similar
effects.
20.00
- y-axis orientation
0.00 -
- - x-axis orientation
- - - - Solid wall
Perfect conducting wall
z -20.00
-
-0
n-
-40.00
-
al
.->
-
0
M
0, -60.00
-80.00
FIGURE 21 Received power versus antenna separation for the four ray model. These results are
for a 6 inch block wall with slabs oriented along the y-axis, where f = 900 MHz. The antennas are
1 rn off the ground and are spaced 1 rn from each of the two walls.
From these results it is demonstrated that for short path propagation

(< 500 m), differences of 5-10 dB in received power can be achieved by
modelling composite walls as either a single layer structure or as a perfectly
conducting wall. This illustrates the importance of properly representing the wall
reflections for short propagation paths.
Propagation Down Roads In Thickly Forested Areas

ITS subsystems may need to function in areas where the roadway is lined with
thick forest or in residential neighborhoods, where streets are heavily lined with
trees and shrubs. In order to accurately model propagation in this environment
it is important to accurately characterize the reflection coefficient from the forest
boundary. This characterization can be accomplished either by measuring
reflection coefficients or by theoretically modelling the reflection from a free
spaceltree interface.
Individuals in the past have theoretically modeled this situation (Brown and
Jackson, 1955; Lewin, 1947; Waterman and Pedersen, 1986; Brown, 1953; Came
324 N. DE MINCO AND C.L. HOLLOWAY
and Brown, 1959). In these works, r (the reflection coefficient) is determined

in the same way it is determined when two regions are homogeneous and
isotropic. The reflection coefficient was calculated by:
where the (qeff)is the effective material properties seen by a wave propagating
into the forest.
However, evidence suggests see (Holloway and Keuster, 1994b; Holloway,

draft report) that this approach does not accurately predict the reflection
coefficient. A correction term is needed in equation (1 3) that considers the higher-
order modes that are present at the air-tree interface (Holloway, draft report).
In the ITS environment there may also be a need to propagate through
foliage. Empirical formulas for propagation through foliage have been introduced
(Yacoub, 1993). The attenuation for a vertically polarized wave is given by:
and for a horizontally polarized wave the attenuation is given by:
where (o) is the foliage conductivity in Slm and (f) is the frequency in MHz.
Values for (o) can range from 0.5*10" to 50*10-~Slm depending on the types
of foliage and whether the foliage is dense, dry, or wet (Yacoub, 1993). These
equations are valid from 300 MHz to 10 GHz. At frequencies below 300 MHz
the attenuation is negligible. Two-way communications for ITS where the
attenuation would be significant include applications at 900 MHz, 2.4 GHz, and
5.8 GHz.
These expressions can be incorporated into existing propagation models. It is
also necessary to theoretically investigate the error that occurs in calculating the
reflection coefficient by not considering the correction as mentioned above.
PropagationThrough Tunnels and Bridges

In many urban settings, automobiles drive on bridges composed of complex
metal structures and drive through different types of tunnels. The metal structure
of the bridges and the material content of the tunnels results in very complicated
propagation channels for ITS which may in turn compromise its efficiency.
Attenuation of signal levels in these types of environments can be severe. The
effects of frequency and antenna spacing on the attenuation of a signal propa-

gating through a tunnel have been analyzed (Jakes, 1974; Yacoub, 1993).
Attenuation along the length of a tunnel is only 2 4 dB per 1,000 feet in the
frequency range of 2.4 to 11 GHz (Jakes, 1974). At 900 M H z the received
signal power level is inversely proportional to the fourth power of distance, and
at 2.4 GHz the received signal power is inversely proportional to the square of
the distance (Jakes, 1974). Signal attenuation in tunnels is much more severe
(> 10 dB per 1,000 feet) at frequencies below 600 MHz (Jakes, 1974). The
references contain more detailed information on signal attenuation in tunnels.
Propagation Through and Over Buildings

Depending upon the building material used, the straight line propagation path
through and over the buildings (point A to C, A to D, or A to E, in Figure 12).
may be a significant contribution to the total signal. The propagation through a
building may be the same order of magnitude as the diffracted fields down two
to three city blocks. In earlier work, (Alexander, 1983; Seidel and Rappaport,
1992) performed measurements of signal propagation inside buildings. They
found that free space loss did not vary as the typical I/?, but varied as I/$.
Where n is a number ranging from 1.2 to 6.5, depending on the type of building
material (Alexander, 1983). An experimental effort similar to that done by
(Alexander, 1983; Seidel and Rappaport, 1992) for the indoor channel will need
to be performed on various urban canyon settings, in order to accurately char-
acterize the 11'1' propagation loss for different types of buildings. In this effort
the straight line propagation through a building model for a multiple side street
scenario may be incorporated into the model to be used under certain conditions.
Compensation for Near-field and Close-in Distances

In the models described previously, compensation must be made to allow for
near-field coupling when the distances between the transmitter and receiver of
an ITS subsystem are so close that a far-field analysis is an invalid approach for
performance prediction. The electromagnetic waves arriving at a receiver from
a transmitter at such short distances are not true plane-electromagnetic waves
and the antennas are not effectively characterized by their far-field gains. At
these close-in distances the field strength or received power from the transmitter
does not depend on the far-field gains of the transmitter and receiver antennas.
A near-field analysis approach is the correct way to analyze the different sub-
systems of ITS operating under these conditions. One analysis approach would
examine the near field of both the specific transmitter and receiver antennas
326 N . DE MlNCO AND C.L. HOLLOWAY
separately and combine the results to compute either signal field intensity or
power at the receiver. Another approach'would be to perform a near-field
coupling calculation using the specific electrical characteristics and physical
geometry of the configuration under study.
ANTENNA ISSUES FOR ITS

Mutual Coupling of Antennas in Adjacent Vehicles

The electromagnetic environment in which ITS subsystems are to operate is
very con~plex.This is especially true when vehicles are in close proximity to one
another. Mutual coupling occurs in both the near-field and far-field regions of
an antenna under consideration. For this scenario, mutual coupling of electro-
magnetic fields from the antennas can affect the performance of the subsystems
of ITS.
The effect of mutual coupling can be explained by investigating two dipole
antennas spaced a distance (d) apart. The induced voltage (the received or trans-
mitted signal) at the antenna terminals can be expressed as the following
(Stutzman and Thiele, 1981; Lee, 1984; Kraus, 1988):
where V I and V2 are the voltages on antennas 1 and 2, 11 and 12 are the currents
on the antennas, Z l l and Z 2 2 are the self-impedance of the two antennas, and
Z I 2 and ZZ1are the mutual impedances between the two antennas. This equation
states that energy in antenna 2 is coupled into antenna 1 through the mutual
impedance Z12.
If the antennas are far apart (i.e., the distance between the antennas is large
compared to a wavelength) then the mutual impedance term can be neglected
and the induced voltages on the antennas are governed only by the self-
impedance. For a dipole of length h/2, the self-impedance is:
Z,, = 73 + j42.5 R (17)
and for a dipole of length 3U2, the self impedance is:

The mutual impedance of the antennas is a function of antenna lengths,

antenna spacing, and the orientation of the antennas. The mutual impedance for
a dipole of approximate length )J2 for different antenna spacings and orienta-
tions varies over a wide range. If the antennas are within approximately one
wavelength of each other, the value of the mutual impedance becomes compa-
rable to the value of the self-impedance (Stutzman and Thiele, 1981). Thus, the
signal on antenna 1 will couple or interfere with the signal on antenna 2. Some
of the frequencies proposed for two-way communication for ITS applications
include: 902 MHz (h = .33m), 450 MHz (h = .67m), 220 MHz (1= 1.36m), and
150 MHz (h = 2.00m). Antennas of adjacent vehicles may be spaced at distances
on the order of a wavelength. Under these conditions the mutual coupling effects
could be important.
The antennas used for applications in ITS will not behave exactly like the
dipole examples above. Therefore, research is needed to determine the order of
magnitude of the mutual impedance for the antenna systems used.
The mutual impedance behavior discussed above can be extended to 'N'
number of antennas, and the induced voltages on the 'N' antennas are given by:
v, = Z , , I, + Z , , I,+ ... + Z,, IN
V,=Z,,I,+ z,, I , + ... + z , I,
The coupling of the signal from one antenna to another has two effects. First,
the energy can actually alter the antenna radiation pattern of the antennas.
Secondly, the coupling can distort and disrupt received and transmitted signals
from the antennas. The order of magnitude of the mutual impedance and its
effect on different subsystems of ITS need to be investigated to ensure the
reliability of the systems.
CONCLUSIONS AND RECOMMENDATIONS
In this paper we have summarized the parts of radio-wave propagation

prediction models that are needed to simulate the ITS environment. It will be
necessary to modify the existing models to account for propagation down side
streets and combinations of side streets. This will include: cross streets propa-
gation, propagation down multiple side streets, impedance edge and comer
diffraction, diffracted fields from comers to street to receiver, and multiple
diffractions from closely spaced comers and objects. For propagation over rough
328 N. DE MlNCO AND C.L. HOLLOWAY.
streets a set of curves must be developed for the magnitude and phase of the
reflection coefficient for typical or realistic roughness dimensions that would be
encountered in an ITS setting. These reflection coefficients for a non grnooth
street could then be incorporated into existing propagation models. For propa-
gation down roads in thickly forested areasthe expression listed in the paper for
propagation through foliage must be incorporated into the propagation model.
Another area of investigation would involve determining the best technique to
model propagation through tunnels and bridges.
Development of the propagation model components that do not already'exist

will allow both performance and interference prediction of ITS subsystems and
equipment. It is necessary that this knowledge be developed early in the ITS
program to ensure reliable system performance.
References
Achatz, R.J.. "Performance Prediction of Direct Sequence Modems in the Indoor Channel," NTlA
draft rcpon.
Alexander, S.E., (1983) "Characterising Buildings For Propagation at 900 MHz," Electmnics
Letters. VOI. 19, NO. 20, p. 860.
Allen, KC., (1983) LLAllenuationof Millimeter Waves on Earth-Space Paths by Rain Clouds," NTlA
Report 83-132.
Allen, K.C.. (1986) "EHF Telccommunication System Engineering Model," NTlA Report 86-192.
Allen, K.C., (1990) "Software For The Analysis Of Microwave Operational Scenarios, Reprinted
from Proceedings of the IEEE Military Communications Conference," Monterey, CA, June
1990, pp. 17.5.1-17.5.5.
Allen. K.C.. (1991) "A Model of Millimeter-Wave Propagation for Personal Communication
Networks in Urban Settings," NTlA Report 91-275.
Balnnis, C.A., (1989) Advanced Engineering Electmn~agnetics.Chapter 13, pp. 2-31, pp. 743-835,
New York: John Wiley & Sons.
Barrick, D.E., (1971) "Theory of HF and VHF Propagation Across the Rough Sea, 1, The Effective
Surface Impedance For a Slightly Rough Highly Conducting Medium At Grazing Incidence,"
Radio Sci., Vol. 6, pp. 517-526.
Bass, F.G. and Fuks, I.M., (1979) Wave Scattering Fmm Statistically Rough Surfaces, Chapter 3 , pp.
72-90, Oxford: Pergamon Press.
Bcckmann, P. and Spizzichino, A., (1987) The Scattering of Electmmagnetic Waves from Rough
Surfaces. pp. 34-1 18, Norwood, MA: Anech House.
Bergbung, C, and Olsson, L.G., (1991) "Rigorous Diffraction Theory Applied to Street Microcell
Propagation," lEEE Global Telecommunication Conference, Phoenix, AZ, pp. 1292-1296.
Bcnoni, H.L., Honcharenko, W., Maciel, L.R. and Xia, H.H., (1994) "UHF Propagation Prediction
for Wireless Personal Com'munications," Proc. of the IEEE, Vol. 82, No. 9, pp. 1333-1359.
Boithias, L., (1987) Radio Wove Pmpagotion, pp. 118-143, New York: McGraw-Hill.
Boit. MA., (1973) "Lagrangian Analysis of Multiple Scatter in Acoustic and Electromagnetic
Rcflexion." Bull. Acad Rayale CI. Sci., Vol. 59, pp. 153-169.
Brown, G.S., (1985) "A Comparison of Approximate Theories for Scattering From Rough
Surfaces," Wave Motion. Vol. 7, pp. 195-205.
Brown, J., (1953) "Artificial Dielectrics Having Refractive Indices Less Than Unity," The
Institution of Electrical Engineers, No. 62, May.
Brown, J. and Jackson, W., (1955)"The Properties of Artificial Dielectrics At Centimetre
Wavelengths," Pmc. IEE (London), Vol. lOZB, pp. 11-16.
Brown, R.G., Sharpe, R.A., Hughes, W.L. and Post, R.E., (1973) Lines Waves and Antennas,
pp. 392-394, pp. 408-409, New York: The Ronald Press Company.
Carne, A. and Brown, J., (1959)"Theory of Reflections From The Rodded-Type Artificial
Dielectric," Proc. IEE (London), Vol. 106, pt. B, pp. 107-1 15.
DeMinco, N., (1986) "Automated Performance Analysis Model for Ground-Wave Communication
Systems," NTlA Report 86-209.
DeMinco, N., (1995) "Engineering Manual for the Jammer Effectiveness Model," NTIA Repon
95-322, September.
Doviak, R.J. and Zmic, D.S., (1993) Doppler Radar and Weather Observations, Second Edition,
pp. 38-43, San Diego, CA: Academic Press.
Elliot, S.D. and Dailey, D.J., (1995) Wireless Communications for lnrelligent Transportalion
Systems, pp. 5 6 6 7 , 121-137. 141-186, 223-224, Boston, MA: Artech House.
Erceg, V., Ghassemzadeh, S., Taylor, M., Li, D. and Schilling, D.L., (1992) "UrbanISuburban
Out-of-Sight Propagation Modeling," IEEE Communications Magazine, pp. 5 M 1 , June.

Friis, H.T., (1946) "A Note on a Simple Transmission Formula,' Proc. of the IRE. Vol. 41,
pp. 254-256, May.
Friis, H.T., (1971) "Introduction to Radio and Radio Antennas." IEEE Spectrum, pp. 55-61, April.
Geikas, J. and DeMinco, N., (1994) "Jammer Effectiveness Model," NTlA Report 95-317.
Haakinson, E., Rothschild, S. and Bedford, B., (1988) "MF Broadcasting System Performance
Model," NTlA Report 88-237.
Holloway, C.L., (1992) "Edge and Surface Effects on Conductor Loss Associated With Planar
Circuits," MIMICAD Technical Report no. 12, Department of Electrical and Computer
Engineering, University Colorado, Boulder, Colorado, Chapter 3.
Holloway, C.L. and Kuester, E.F., (1994a) "The Use of Homogenization in Deriving Effective
Boundary Conditions for Perfectly Conducting Periodic Surfaces," draft report to he submitted
to Wave Motion Journal.
Holloway, C.L. and Kuester, E.F., (1994b) "Impedance Type Boundary Conditions for Both
2-Dimensional and 3-Dimensional Periodic Interfaces," draft report to be submitted to
J. Electromagnetic Waves and Appl.
Holloway, C.L., Perini, P.L., DeLyser, R.R. and Allen, KC., (1995a) "A Study of the Electromag-
netic Properties of Concrete Block Walls for Short Path Propagation Modeling," NTlA Repon
96-326.
Holloway, C.L., Allen, K.C. and Laflin, M.G., (1995b) "Analysis of Composite Walls For Shon Path
Propagation Modeling," in Proc. 1995 IEEE Antennas and Propagation Society International
Symposium, Newport Beach, CA, 18-23, pp. 526-529.
Holloway, C.L., "Reflections From An Array of Periodically Spaced Conducting Scatterers," draR
report.
Holloway, C.L., Perini, P.L., De Lyser, R.R., and Allen, K.C., (1997) "Analysis of Composite Walls
and Their Effects on Shon Path Propagation Modeling," accepted for publication in IEEE
Trans. on Vehicular Tech. to appear early 1997.
Hufford, G.A., Longley, A.G. and Kissick, W.A., (1982) "A Guide to the Use of the ITS Irregular
Terrain Model in the Area Prediction Model," NTlA Repon 82-100.
International Telecommunications Union (ITU-R), (1986) "The Concept of Transmission Loss,"
Vol. V, ITU-R Recommendation 341-2, Dubrovnik. Yugoslavia, pp.9-14.
Ishimaru, A,, Electromagnetic Wave Propagation, Radiation, and Scattering, Chapter 7, pp.178-203,
New Jersey: Prentice Hall, 1991.
Jakes, W.C., (1974) Microwave Mobile Communications, pp. 110-113, New York: John Wiley &
Sons.
Kong, J.A., (1986) Elecrmmagnetic Wave Theory, Chapter 6, pp. 495-507, New York: John Wiley &
Sons.
Kouyoumjian, R.G. and Pathak, P.H., (1974) "A Uniform Geometrical Theory of Dimaction for an
Edge in a Perfectly Conducting Surface," Proceedings of the IEEE, Vol. 62, No. I I ,
pp. 1448-1461.
Kraus, J.D., (1988) Antennas, Chapter 10, pp. 4 0 9 4 3 3 , New York: McGraw-Hill Book Company.
Lee, K.F., (1984) Principles ofdntenna Theory, Chapter 5, pp. 85-104, Chichester: John Wiley &
Sons.
Lewin, L. (1947) "The Electrical Constants of a Material Loaded With Spherical Particles," The
Institution of Electrical Engineers. Vol. 94, pt. 111, No. 27,.
Liebe, H.J., (1983) "An Atmospheric Millimeter Wave Propagation Model," NTlA report 83-137.
Liebc, H.J., (1989) "MPM- an Atmospheric Millimeter Wave Propagation Model," lnlernational
Journal ufl~ljrored and Millimerer Waves. Vol. 10, No. 6, pp. 6 3 1 6 5 0 .
Longley, A.G. and Rice. P.L., (1968) "Prediction of Tropospheric Radio Transmission Loss Over
lreggular Ternin," ESSA Technical Report ERL 79-ITS 67.
Luebbers. R.J., (1984) "Finite Conductivity Uniform GTD Versus Knife Edge Diffraction in
Prediction of Propagation Path Loss," IEEE Trans. on Antenna and Propagation, Vol. 32,
No. I, pp. 70-76.
Maciel, L.R., Bertoni, H.L. and Xia. H.H., (1993) "Unified Approach to Prediction of Propagation
Over Buildings for All Ranges of Base Station Antenna Height," IEEE Trans. on Vehicular
Technology. Vol. 47. No. I, pp. 4 1-45.
Maliuzhinets. G.D., (1958) "Excitation, Reflection and Emission of Surface Waves From a Wedge
With Given Face Impedances," Sou Phys. Dokl.. Vol. 3 , No. 4, pp. 752-755.
Mastrangelo, J.R and Rust, W.R., (1 996) "Subcarrier Traffic Information Channel Testing and Evalu-
ation," NTlA Report 9 6 3 3 3 , October.
Mnystre. D., Mendez, O.M. and Roger, A,, (1983) "A New Electromagnetic Theory for Scattering
From Shallow Rough Surfaces," Oplica Acra. Vol. 30, pp. 1707-1723.
McNamara. D.A., Pistorius, C.W.I. and Malherbe, J.A.G., (1990) lnrroducrion lo the Uni/orm
Geomerrical Theory of Drffraction, pp. 159-261. Boston: Artech House.
National Archives and Records Administration, (1994) Code of Federal Regulations-Communica-
riuns, Vol. 47, pp. 5 6 8 4 2 5 , October.
Norton. K.A., (1953) "Tmnsmission Loss in Radio Propagation," Proc. of the IRE. Vol. 41,
No. I, pp. 146-152, January.
Norton. K.A... 11959)
. , "Svstem, ...
Loss in Radio Wave Prooaeation.". J. Res. NBS. Vol. 63D
(dadio Propagation), No. 1, pp. 53-73, JulyIAugust.
Pathnk, P.H. and Marhelka, R.J., (1991) Asymproric High Frequency Methods for E M Anrenna and
Scarrering Anal)ais, course notes for an IEEE AP-S short course, June.
Rice, P.L., Longley, A.G., Norton, K.A. and Barsis. A.P., (1967) "Tropospheric Loss Predictions for
Tropospheric Communication Circuits," NBS Technical Note 101, May.
.Rosich, R.K. and Wait, J.R., (1977) "A General Perturbation Solution For Reflection From
Two-Dimensional Periodic Surfaces," Radio Sci.. Vol. 12, pp. 719-729.
Rustako, A.J., Amitay, N., Owens, G.J. and Roman, R.S., (1991) "Radio Propagation at Microwave
Frequencies for Line-Of-Sight Microcellular Mobile and Personal Communications," IEEE
Trans. on Vehicular Technology. Vol. 40. No. I , pp. 203-210.
Schumaker, R.W., Olney, R.D., Wragg, R., Landau, F.H., (1996) "Collision Warning System
Technology," Trnflic Technology International '96, pp. 70-73, United Kingdom and Interna-
tional Press, United Kingdom.
Scott. H., Stouffer, K., Rowe, P., "Recommendations on Selection of Vehicle-to-Roadside
Communications Standards for Commercial Vehicle Operations." NlST Report NIST-IR-5444.
Seidel, S.Y. and Rappaport, T.S., (1996) "914 Mhz Path Loss Prediction Models For Indoor Wireless
Communications In Multifloored Buildings." IEEE Trans. on Anrennas and Propagarion.
Vol. 40, No. 2, pp. 207-217.
. -
Shibuva. S.. (1987) A Basic Atlas ofRadio-Wave Pmpa~arion, . 332-368, New York: John Wilev
.DD.
& Sons, 1987, pp. 332-368. -
Srvov, A.N., (1964) "Reflection of Electromagnetic Waves From Cormgated Surface With Small
-
Period." Radioreth. Elekrmn.. Vol. 9. DD. 1821-1827. 1964 fin Russian: Enel. transl. in Radio
Eng. Electron., Phys., Vol. 9, 1509-'l>i5].
Stutzman, W.L. and Thiele, G.A., (1981) Anrenna Theory and Design, pp. 154-160, New York:
John Wiley & Sons.
Takeuchi, T., Sako, M. and Yoshida, S., (1991) "Multipath Delay Prediction on a Workstation for
Urban Mobile Radio Environment," IEEE Global Telecommunication Conference Proceedings,
Phoenix, AZ, pp. 308-1312.
Tan. S.Y. and Tan, H.S., (1995) "A Theory for Propagation Path-Loss Characteristics in a
City-Street Grid," IEEE Trans. on Elccrromagneric Comparibilify, Vol. 37, No. 3, pp. 333-342.
Teters, L.R., Lloyd, J.L., Haydon, G.W. and Lucas, D.L., (1983) "Estimating the Performance of
Telecommunication Systems Using the Ionospheric Transmission Channel." NTlA Report
83-127.
Vainshtein (Wainstein) L.A.. (1966) "On the Electrodynamic Theory of Grids," in Elektronika
Bol'shiW1 Moshchnostei, Vol. 2 (P. L. Kapitsa and L. A. Vainshtein, eds.). Moscow: Nauka,
1963, pp. 2 6 7 4 [in Russian; Engl. transl. in High-Power Electronics. Oxford: Pergamon Press,
pp. 14-48].
Verhitskiy, I.L., (1976) "The Diffraction of a Plane Wave By a Dense Comb," Radio Eng. Electron.
Phys, No. 3, pp. 41-46.
Violette, E., Espeland, R. and Allen, K.C. (1988) "Millimeter-Wave Propagation Characteristics and
Channel Performance for Urban-Suburban Environments," NTIA Report 88-239,.
Wait, J.R., (1971) "Perturbation Analysis for Reflection From Two-Dimensional Periodic Sea Waves,"
Radio Sci., Vol. 6 , pp. 387-391.
Waterman, P.C. and Pedersen, N.E., (1986) "Electromagnetic Scattering By Periodic Array of Par-
ticles," J. Appl. Phvs.. Vol. 59, No. 8, pp. 2609-2618.

Wenzel. A.R.. (1974) "Smoothed Boundaw Conditions For Randomlv. Rough - Surfaces." J. Math.
P ~ Y S . voi.
, 1 5 , ' ~317-323.
~.
Westphal, W.B. and Sils. A,, (1972) "Dielectric Constant and Loss Data," Massachusetts Institute of
Technoloev. -.. Technical Reoon AFML-TR-72-39.
Whelan, R., (1995) Smarl Highways-Smart Cars, Chapter 11, pp. 188-190, Boston, MA: Artech
House.
Yacoub, M.D., (1993) Foundoriom ofMobile Radio Engineering, pp. 61-123. Boca Raton, FL: CRC
Press.
Yamaguchi, Y. and Takeo, T., (1988) "Streets Crowded With Pedestrians," IEEE Transactions on
Eleclromagnelic Comporibility. Vol. 30, No. 2, pp. 130-136.
Yim, T.S. and Siang, T.H., (1992) "A New Propagation Model for a Street Scene for Microcellular
Communications," International Electromagnetic Compatibility Symposium, Singapore,
pp. 200-207.

Propagation Modeling For ITS Applications in The Roadway Environment

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Propagation Modeling For ITS Applications in The Roadway Environment

Uploaded by

Copyright:

Available Formats

ITS Journal - Intelligent Transportation Systems Journal

ISSN: 1024-8072 (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/gits19

Propagation Modeling for ITS Applications in the

NICHOLAS DE MINCO & CHRISTOPHER L. HOLLOWAY

To link to this article: http://dx.doi.org/10.1080/10248079708903728

Published online: 24 Oct 2007.

Submit your article to this journal

View related articles

Full Terms & Conditions of access and use can be found at

Propagation Modeling for ITS

Institute for Telecommunication Scrences

(Received 5 February 1996; In final form 3 October 1996)

The functions of ITS will probably be implemented using radio communication

Radio-wave propagation prediction models, that can be used for engineering

mance evaluation of these separate systems within ITS at proposed frequencies.

FIGURE I Intelligent Transportation Systems applications.

If reliable performance predictions cannot be made for subsystems of ITS,

and an urban canyon environment. A discussion of the present capabilities of

THE ITS ROADWAY ENVIRONMENT

is that only line-of-sight (LOS) and limited diffraction propagation channels

The roadway environment in urban settings contains bridges composed of

collection and vehicle identification in the roadway environment can occur in

POTENTIAL FREQUENCY BANDS FOR ITS

could be obtained from the Global Positioning System (GPS). Automatic

Applications at Frequencies of 40 GHz and Above

Applications in the 5800-MHz Band

driverlvehicle status reporting, electronic credentials (commercial), electronic

Applications in the 2450-MHz Band

Applications at Frequencies Greater than 1 GHz (Private Microwave

Applications in the 902- to 928-MHz Band

Applications in the 150-,450-,800-, and 900-MHz Bands

Applications in 220-MHz Band

One-way communications will be needed for such applications as: in-vehicle

Applications in the 88- to 108-MHz Band (FM Broadcast Band)

Other Systems, Applications, or Functions not lncluded Above

BASIC CONCEPTS IN WAVE PROPAGATION

Free Space Propagation

where (h) is the wavelength in meters of the electromagnetic wave in the

LBb= 32.45 + 20 IogCf,,,) + 20 log(r) (dB) (4)

Attenuation Due to Precipitation

Attenuation Due to Water Vapor and Oxygen Absorption

0 50 1W 150 200 250

The phase and amplitude response of an electromagnetic wave interacting

LC, = 20 log (dB)'

This loss is also subtracted from equation (3)

Radio Waves in The Vicinity of Planar Surfaces

Reflections from smooth perfectly conducting surfaces

Reflections from smooth surfaces

sin Y - (c, - i60 o h - cos2 ' # ' ) I n

The reflection coefficient (rH) of an E-field parallel to the plane of incidence

rH= ( Q - i60 o h) sin Y - (6, - i60 o h - cos2 Y ) ' ~

composed of concrete. Common examples of this are buildings and road

Incident Angle (degrees)

FIGURE 5 Reflection coeficients for both an E-polarized (rE)and H-polarized (rH)

Reflections from rough surfaces

1. Roughness dimension very small compared to a wavelength

2. Roughness dimension small compared to a wavelength (R c 1 A) .

3. Roughness dimension large compared to a wavelength

0.00 20.00 40.00 60.00 80.08

If PA1 is small, the surface can be approximated as smooth. However, if

There have been a number of theoretical approaches analyzing electromag-

Diffraction from Objects