Wave Propagation inside and around Vehicles

in Dynamic Time Variant Scenarios

René Wahl, Gerd Wölfle
AWE Communications GmbH, Otto-Lilienthal-Str. 36, 71034 Boeblingen, Germany, www.awe-communications.com

Abstract–With the growing demand for wireless The frequency range available and suitable for wireless
communication systems, a wireless concept for the data links in and around vehicles will be above 450 MHz and
exchange between different system components inside a vehicle might go up to several GHz. At those frequencies, an
and also between vehicles becomes interesting. In order to approach with full wave methods is not possible anymore due
describe and thus assess the mobile radio channel in and around
the vehicle, simulations are required. For low frequencies, full
to the required computational efforts and the memory
wave simulations using the Method of Moments are adequate requirements. However, with increasing frequency ray optical
means. For higher frequencies, those models get inadequate approaches can be used alternatively, if the dimensions of the
because of their computational effort. Systems employing high objects are much larger than the wavelength [5].
frequencies are also used for car-to-car communication systems.
The main aspect in such applications is the time variance of the
B. Description of the new Approach
scenario, as the vehicles move continuously and thus the channel Ray optical propagation models are well known in the
impulse response is time-variant. This paper discusses the domain of urban radio network planning [6]. In the last years,
demands for models to describe the radio channel inside and these models have become more and more important also for
around vehicles at high frequencies and presents an appropriate indoor scenarios because of their high accuracy [7]-[9].
ray tracing approach. Therefore, a 3D Ray Tracing model was used as a basis for
the described approach. Some improvments (e.g. higher
Index Terms–Automotive, Vehicles, Multimedia in Vehicles,
Time Variant Scenarios, Ray Tracing, Car-to-Car resolution of the databases, handling of bent surfaces,
Communication consideration of a very high number of polygons in the
prediction) were made to the Ray Tracing model, so that the
I. INTRODUCTION algorithm can deal with vehicle databases and fits to the
After the great success of wireless communications used needs of such scenarios.
in land mobile radio systems, wireless communications inside As the 3D Ray Tracing propagation model allows the
vehicles and between vehicles becomes more and more vector-oriented definition of an arbitrary number of polygons
interesting. with different material properties, a vehicle with all details
The steadily increasing demand for multimedia can be modeled easily. The vehicles in fig. 1 are built up of
applications in modern cars requires looking for new wireless more than 50,000 polygons with individual material
concepts for the planning of wireless in-car systems. properties, such as permittivity, permeability and
To increase driving safety, assistance systems exchanging conductance.
data using wireless channels between two or more vehicles
will be introduced during the next years. Such a wireless
system is the Dedicated Short-Range Communications
Standard (DSRC) which is a short to medium range wireless
protocol specifically designed for automotive use. It offers
communication between vehicles and roadside equipment.
This technology is working in the 5.9 GHz band. The main
aspect in such applications is the time variance of these
scenarios, as the positions of vehicles change continuously.
II. MODELING INSIDE VEHICLES Figure 1: Vector oriented databases of vehicles: Wire frame
(left) and with filled polygons (right).
A. Current Status The determination of interaction losses along each ray is
In the automotive sector, simulations by using full wave based on the Uniform Theory of Diffraction (UTD) for
approaches (e.g. Method of Moments [1]) have been diffractions and on the Fresnel coefficients for reflections and
established to e.g. optimize antenna positions in the vehicle transmissions [10]. Due to the material properties of objects
under consideration of EMC problems and electromagnetic used in vehicles, the consideration of diffractions is much less
field exposure of the user. Therefore several software important than the consideration of reflections and
packages are available, such as FEKO [2], Microwave Studio transmissions/penetrations, (as diffractions lead to higher
[3] or XFDTD [4]. attenuations). Thus, propagation paths with diffractions are
highly attenuated and nearly irrelevant if many reflected ray approaches are nearly similar (fig. 3 and fig. 4). The
paths for one pixel are superposed. deviations between the results can be traced back to the
In urban and indoor environments, the position of walls differences in the treatment of the material properties.
and buildings are usually rather inaccurate. Therefore the
(deterministically determined) phase of the signal
contributions of each propagation path is not considered for
these environments, as the geometrical inaccuracies could be
in the range of several wavelengths. So radio network
planning tools only accumulate the power contributions (i.e.
incoherent superposition) [11]. For vehicles, very accurate
CAD data is available. This makes it possible to compute the
individual ray path contributions with consideration of the
phases. This is very important, as a large number of
interactions must be taken into account due to the low
reflection losses of the materials utilized in vehicles [12]. The
extreme multipath propagation is shown in figure 2.

Figure 3: Comparison of prediction results of Method of

Transmitter
Moments (left) to 3D Ray Tracing (right).

Figure 4: Database used for computation (left) and difference

between prediction with Method of Moments and 3D Ray
Tracing (right).
The horizontal area predicted for this comparison was
about 1.3m x 1.7m. The resolution used was 3 cm. The
computation time for the 3D RT is much shorter than the time
needed by FEKO, using the MoM approach. Even more, the
3D RT needs less memory for the computation. See table Ι for
the values. In the computation four reflections, four
diffractions and four transmissions are computed.

Table I: Comparison between FEKO and the 3D RT

Figure 2: Propagation paths (top) and channel impulse response
approach (3 GHz CPU, 1.5 GB RAM)
(bottom) for one receiver pixel.
Approach Computation Time Memory needed
Another point is the movement of objects inside the
vehicle. Objects like the engine, the wheel suspension or parts MoM (FEKO) some hours about 1 GB
of the steering are not stationary and move or vibrate. High 3D Ray Tracing some seconds about 100 MB
frequent vibrations lead to Doppler shift of the propagation
paths and must be considered in the computation.
D. Example Application
C. Comparison of Approaches
Figure 5 shows an example of a prediction of a wireless
For a comparison between two different approaches, communication system inside a vehicle operating at 433
namely the Method of Moments (MoM) and the 3D Ray MHz. The transmitter (40 dBm) is located at the mirror
Tracing (3D RT), a scenario with an engine compartment was beneath the windscreen and the prediction layer is above the
chosen. The MoM is included in the software package FEKO dashboard. Only the relevant parts of the vehicle (no wheels,
[2]. rims, lights) are used for the computation.
In the left picture of figure 4 the database used for the The database of the vehicle is preprocessed once, before
computations is shown. The transmitter (40 dBm) was placed the prediction. This takes several hours of time on a standard
near the battery. The frequency used was 433 MHz. At this PC. The computation time of the prediction itself for arbitrary
frequency the conditions for ray optical models are not transmitter locations is below than one minute, even if two
completely fulfilled, so this is a worst-case scenario for the diffractions, two reflections and an arbitrary number of
ray tracing model. Nevertheless, the results of both simulation transmissions for each propagation path are allowed.
 T = to

Figure 5: Prediction of path loss for an in-car communication

scenario.

III. MODELING AROUND VEHICLES

A. Time-variant Scenarios
In real traffic situations, many objects influence the
propagation of waves around vehicles. One example is shown
in fig. 6, where a car-to-car communication scenario is
depicted. Buildings, vegetation, guard rails and other vehicles
have a significant influence on the received channel impulse
response (CIR).
T = to+ t1

Figure 6: A car-to-car communication scenario [13]. T = to+ t2

A new software module was developed to create and edit

time-variant scenarios. The basis of the databases are planar Figure 7: Example of a dynamic database with several cars for
polygons, like in ordinary indoor databases. Each element in different timestamps. The time is increasing from the upper to
the lower picture.
the database can be either stationary (not moving) or non-
stationary (dynamic). Translation and rotation vectors, as well
B. Example Prediction
as a scalar value for the velocity are assigned to dynamic
objects for the definition of their behavior in the scenario, Figure 8 presents a result of a car-to-car communication
depending on the time. Predictions are then accomplished for scenario with the most important propagation paths for one
arbitrary defined timestamps. snapshot. The computation time is below one minute on a
An example for a modeling of a time-variant scenario is standard PC. The transmitting antenna (ideal omni antenna) is
shown in the pictures of fig. 7. The scenario contains a located at the mirror of a vehicle as well as the receiving
highway, guardrails, buildings and several cars. For each car antenna. For the prediction a maximum of two reflections,
a route was defined. The cars are built up of very simple one diffraction and zero transmissions for one propagation
databases in contrast to the vehicle databases in fig. 1, as for path are computed. More interactions do not improve the
computations around vehicles highly accurate CAD databases accuracy significantly. Therefore they have been neglected.
are not required to achieve accurate results. All elements in the database with their individual material
For the prediction of the scenarios, a 3D Ray Tracing properties (buildings, streets, guardrails, vehicles) are
model was used. The algorithm supports an arbitrary number considered for the computation.
of reflections, diffractions and transmissions/penetrations for
each ray path. It was adapted to the needs of dynamic
scenarios, e.g. the consideration of the Doppler Shift was
added.
 Transmitter
IV. CONCLUSIONS
In this paper the requirements for modeling the radio
channel inside vehicles as well as for large time variant
scenarios are described. A ray tracing approach for the
propagation modeling in such scenarios is presented and a
concept for the modeling of large scenarios is shown.
Until now costly measurement campaigns are needed to
determine the CIR in time variant scenarios. Simulations now
offer the possibility to modify a time variant scenario and re-
simulate the CIR within short time intervals instead of
carrying out time-consuming measurement campaigns.
V. REFERENCES
[1] U. JAKOBUS, N. BERGER, F. M. LANDSTORFER: Efficient
Receiver
techniques for modelling integrated windscreen antennas
within the method of moments, Millennium Conference on
Figure 8: Propagation paths in a dynamic car-to-car-
Antennas and Propagation, AP2000, Davos, Apr. 2000
communication scenario.
[2] ELECTROMAGNETIC SOFTWARE AND SYSTEMS: FEKO –
Comprehensive EM Solutions, www.feko.info
C. Using RCS for the Modeling of complex Objects [3] COMPUTER SIMULATION TECHNOLOGY: Microwave Studio,
As already mentioned above, the vector-oriented www.cst.de
databases of the vehicles consist of a lot of polygons. Hence, [4] REMCOM: XFDTD, www.remcom.com
the usage of ray optical models for scenarios shown in figure [5] D. DIDASCALOU, M. YOUNIS and W. WIESBECK: Millimeter
wave propagation mechanisms for mobile intervehicle
6 is no longer possible, as memory restrictions exist and
communications, Proc. EUNARV'98 Vehicle Navigation -
computation time should be limited. New methods to define ITS ready for market?, Hannover, Germany, pp. 81-90, June
complex objects (like cars) must be found. 1998.
One method is to use simpler databases of vehicles, which [6] R. GAHLEITNER: Radio Wave Propagation in and into Urban
do not have so many polygons and are a rough approximation Buildings, PhD Thesis, Technical University of Vienna,
of the cars as described above in the previous section. Institut für Nachrichtentechnik und Hochfrequenztechnik,
Another method is to substitute complex objects with their 1994.
bistatic radar cross sections (RCS). A relation between each [7] V. DEGLI-ESPOSTI, G. LOMBARDI, C. PASSERINI: Millimeter
incident and each scattered ray/angle to an object is defined, Measurement and ray-tracing prediction of indoor channel
parameters, Electronics Letters, Volume 34, Issue 22, 29
usually with attenuation and phase shift. RCS data has to be
Oct. 1998 Page(s):2167 - 2168
measured or computed [14]. [8] G. WÖLFLE: Adaptive Modelle zur Funknetzplanung und zur
The Ray Tracing algorithm does not consider polygons of Berechnung der Empfangsqualität in Gebäuden,
objects defined by RCS data for reflection and diffraction. Dissertation, Universität Stuttgart, Shaker Verlag, Aachen,
The polygons are only considered for transmission/ 1999.
penetration losses in the scenario, as reflections and [9] R. HOPPE, P. WERTZ, G. WÖLFLE, AND F. M. LANDSTORFER:
diffractions are defined by the RCS data assigned to an Wideband Propagation Modeling for Indoor Environments
object. and for Radio Transmission into Buildings, PIMRC 2000,
London (UK), 18-21 September 2000
[10] R. HOPPE: Effiziente Modellierung der Wellenausbreitung für
die Funknetzplanung in urbanen Szenarien und innerhalb
Transmitter von Gebäuden, Dissertation, Universität Stuttgart, Shaker
Receiver Verlag, 2002.
[11] AWE COMMUNICATIONS: Software Tool WinProp for the
Planning of Mobile Communication Networks (incl. demo-
version), www.awe-communications.com, 2005
[12] J. P. DEWECK: Real Time Characterization of Wideband
Mobile Radio Channels, PhD thesis, Technical University of
Vienna, May 1992.
[13] CAR 2 CAR COMMUNICATION CONSORTIUM: Mission and
Objectives, www.car-2-car.org
[14] F. WEINMANN, T. F. EIBERT: Radar Cross Section Modelling
of Large Complex Objects, Kleinheubacher Tagung, 2005
Figure 9: Propagation paths in a dynamic scenario where all
vehicles are defined by RCS data.
An example for a prediction including RCS data is shown
in fig. 9. As shown in the figure, all propagation rays are
concentrated in the center of the vehicles, as the RCS data is
defined for this point.