Radiation Protection Dosimetry (2004), Vol. 111, No. 4, pp. 339–342
doi:10.1093/rpd/nch050
RAY-TRACING TECHNIQUES TO ASSESS THE
ELECTROMAGNETIC FIELD RADIATED BY RADIO BASE
STATIONS: APPLICATION AND EXPERIMENTAL
VALIDATION IN AN URBAN ENVIRONMENT
S. Adda, L. Anglesio1, G. d’Amore1, M. Mantovan1 and M. Menegolli2
1
ARPA Piemonte, Dipartimento di Ivrea, Via Jervis 30, 10015 Ivrea (To), Italy
2
Provincia di Torino, Area Ambiente, Via Valeggio 5, 10128 Torino, Italy
INTRODUCTION
A careful assessment of the exposure of urban populations to electromagnetic fields, generated by radio
base stations for mobile telephones, requires the use
of deterministic models that take into account the
interferences caused by the buildings in the propagation of the field.
One of these models is based on the ray-tracing
method, which allows the simulation of the propagation of electromagnetic waves between the transmitting antenna and the reception point by taking into
account the influence of all reflection and diffraction
phenomena due to the presence of obstacles in the
propagation area, on the basis of the theory of
geometrical optics and other related developments
[Geometrical Theory of Diffraction (GTD)/Uniform
Theory of Diffraction (UTD)]. Although it is possible to model very complex urban environments using
such a method, it is often impossible to get all the
input data needed to take into account, for instance,
the geometrical and electrical features of the buildings within a certain urban area. Moreover, the
knowledge of such features always suffers from
uncertainties that influence the accuracy of the calculated results. The aim of this paper is to estimate
the uncertainties in the application of the ray-tracing
method in different urban environments through
several measurements planned in order to experimentally validate the adopted code (the VIGILA
3.0 software). This is necessary because of the lack
of comparisons in the literature between calculations
with ray-tracing techniques and measurements of
significant environment field levels, carried out in
Corresponding author: s.adda@arpa.piemonte.it
order to assess the exposure of the population and
compare it to the legal limits. Such a validation has
been carried out through narrow band measurements in both near- and far-field situations.
MATERIALS AND METHODS
Ray-tracing techniques and the adopted model
Ray-tracing techniques allow the estimation of the
propagation of electromagnetic waves in complex
environments by taking into account the presence
of obstacles and on the basis of the laws of modern
geometrical optics. The interference caused by the
obstacles is estimated by first calculating the transmission and reflection coefficients, which depend
on the angle of incidence of the rays and on the
electrical features of the materials [Luneberg–Kline’s
theory concerning asymptotic developments(1,2)].
Similarly, GTD and UTD require the evaluation of
diffraction coefficients, which also allow a ‘geometrical’ treatment of this phenomenon. Ray-tracing
methods can be used at high frequencies, namely
when the obstacle dimensions are large in comparison with the wavelength, and therefore it is correct
to describe electromagnetic wave propagation by
means of optical rays.
In order to assess the propagation path, and calculate the electric field along each path, it is necessary
to model the ground and buildings by taking into
account the geometrical and electrical descriptions
of the propagation environment.
The tested software (VIGILA 3.0), which was
developed by TiLab laboratories, implements a
backward ray-tracing technique, taking into account
the first- and second-order contributions, for six
339
Radiation Protection Dosimetry Vol. 111, No. 4 ª Oxford University Press 2004; all rights reserved
Downloaded from http://rpd.oxfordjournals.org/ at Freie Universitaet Berlin on June 25, 2015
This paper aims to validate a ray-tracing model for electromagnetic field calculation, which is used in urban environments
to predict irradiation from radio base stations for population exposure evaluation. Validation was carried out through a
measurement campaign by choosing measurement points in order to test different propagation environments and analysing
broadcast control channels through narrow band measurements. Comparison of the calculated and measured fields indicates
that the ray-tracing model used calculates electric field with good accuracy, in spite of the fact that the propagation
environment is not described in detail, because of difficulties in modelling the geometrical and electrical characteristics of
urban areas. Differences between the calculated and measured results remain below 1.5 dB, with a mean value of 1 dB.
S. ADDA ET AL.
possible configurations (direct path, single reflection,
double reflection, single diffraction, diffraction–
reflection and reflection–diffraction). The software
uses a vector database containing three-dimensional
cartographic information as well as the building
material used for each building as input. Based on
these data, the software calculates the visibility array
and the possible optical paths between the source
and the reception points.
Experimental validation
(1) a receiving biconical antenna (Australian
Research PBA10200) with frequency response
in the 100 MHz–2.1 GHz range;
(2) a Rohde & Schwarz FSP-3 spectrum analyser,
with frequency response in the 9 kHz–3 GHz
range;
(3) a Suhner 20 m coaxial cable.
For both types of areas examined, the evaluation of
the fields was carried out using the VIGILA software
and by creating the appropriate geometrical and
electrical models for the urban area, in order to
characterise the propagation environment. Cartographic information for Site 1 (Figure 2a); on a 1 :
1000 scale was extracted from a database for the
urban area of the city of Turin, whereas for Site 2
the reconstruction of the propagation environment
required a greater accuracy because the measurements were carried out in near-field conditions
(Figure 2b). For this purpose, a building planimetry
on a 1 : 50 scale was used, digitizing every structure
(dividing walls, brick chimneys, etc.) using an
AUTOCAD application.
In order to evaluate reflection and transmission
coefficients for each building, the prevailing type of
building material was assessed by inspection and the
corresponding dielectric constants were chosen from
among the list of standard walls supplied by the
VIGILA software. The buildings were placed on a
DTM obtained by interpolation using the spline
method(4) and the topographical points of the cartographic information (on a scale of 1 : 1000) of the
city of Turin.
RESULTS
Figure 1. Sampling grid.
The ray-tracing model was used to compare the
measured and calculated electric field levels by
taking an average for the 12 points on the grid.
The four sites considered are different with respect
to the propagation conditions: Site 1A is an uncovered car park located on the third floor with the
340
Downloaded from http://rpd.oxfordjournals.org/ at Freie Universitaet Berlin on June 25, 2015
The ray-tracing model was validated through several
measurements that were carried out in certain areas
of the city of Turin. These areas were chosen according to the following criteria: the availability of particularly detailed and up-to-date vector cartography,
the type of urban environment and the presence
of significant electric field levels with respect to the
urban background(3). To select the areas with higher
field levels, the distribution of the electromagnetic
field generated by radio base stations was assessed,
for the entire municipal territory, by means of a
simplified far-field calculation model. The field was
calculated at a height of 1.5 m from the ground,
using the digital terrain model (DTM) on a grid of
points with a 50 m step.
The data obtained allowed the determination of
areas with theoretical levels of electric fields between
1.5 and 3 V/m, often characterised by the presence of
several radio base stations coexisting at the same
site. Two types of sites were thus selected: the first
site (referred to as Site 1) was located in a far-field
zone, where measurements were carried out at locations with different visibility situations; the second
site (referred to as Site 2) was a terrace, where three
radio base stations were set up and where measurements were carried out under near-field conditions.
For each measurement site, electric field sampling
was carried out at points located at the corners of
a 1 m box, at three different heights from the ground
level for 12 points overall (Figure 1).
Then the average of the field levels measured
at the 12 grid points was calculated in order to
compare the calculated and measured values. For
each radio base station, a narrow band measurement at each grid point was carried out, detecting
only the control channel, which is always active at
constant power (the Broadcast Control Channel
of Global System for Mobile communication).
The assessment of the electric field level by control
channel detection made the measurement independent of the working conditions of the radio base
station.
The instrumental apparatus, which includes an
electrically insulated van to ensure electromagnetic
immunity, comprises the following parts:
EXPERIMENTAL VALIDATION OF RAY-TRACING TECHNIQUES
(a)
(b)
Table 1. Comparison between the average values of the
measured and calculated electric field levels obtained using
the ray-tracing technique at the different sites.
Measurement Emeas (V/m)
site
Site
Site
Site
Site
1A
1B
1C
2
0.810
0.025
0.100
1.623
0.158
0.005
0.021
0.358
Ecalc (V/m) SEcalc (%)
0.952
0.023
0.127
1.895
0.032
0.001
0.006
0.022
17.5
8.0
27.0
16.8
obstacles between the source and the evaluation
point and the field values are very low, the calculated
level is slightly lower than the measured level. The
same data reported in Figure 3 show a good agreement between the calculated and measured values
as the difference between these values is, for each
site, within the uncertainty of the measured level.
CONCLUSIONS
radio base station within its line of sight, Site 1B is a
green area located at about 300 m from the radio
base station installation site with several buildings in
between, Site 1C is a green area under shadow conditions due to one single row of buildings between
the source and the measurement point and Site 2 is
the building terrace where the radio base stations
are located (the calculation is made by using
the near-field algorithm supplied by the VIGILA
software).
The results for the different measurement sites are
indicated in Table 1, which lists the measured electric
field levels and the field levels calculated using the
model. The uncertainty associated with each measured value was obtained by adding the instrumental
uncertainty and the standard deviation of the distribution of the values measured on the considered
volume, which is related to the uncertainty due
to antenna positioning. The standard deviation of
the calculated values [s/H(n 1)], in the volume
considered for each measurement site, was assessed
to estimate the uncertainty for the average calculated
field level (SEcalc). The last column in Table 1 lists the
percentage difference () between the calculated and
measured levels.
It is noticeable that the measured and calculated
data differ by 8–27%. In three of the four sites the
calculated level is higher than the measured level;
only in the case of Site 1B, where there are several
This research shows that there is a good agreement
between the results of ray-tracing calculations and
the measurements made in urban environments
under different propagation conditions. It is highly
significant that the differences between the calculated and measured data are always within the measurement uncertainty even if approximations such
as morphologic data of the urban propagation
area, topographical data of the transmitting antennae
and characterisation of the surfaces (considered
perfectly flat and smooth), and the temperature and
humidity conditions when carrying out the measurements were not taken into account in the uncertainty
estimation.
In particular we emphasise the agreement
obtained in those sites where there are buildings in
between (Sites 1B and 1C) and where a simple calculation under free-space and far-field conditions led to
an overestimate of the field by 2–12 times the
measured value.
It has to be pointed out that under visibility conditions also the results of the ray-tracing model come
closer to the measured levels compared with the
simulations under far-field conditions: in Site 1A
the far-field model gives an overestimate of 59%
compared with the measured level, whereas in Site 2
we find an overestimate of 30% (in both cases the
overestimate is higher than the uncertainty in the
experimental value).
This validation demonstrates that a ray-tracing
model is an useful instrument for population
341
Downloaded from http://rpd.oxfordjournals.org/ at Freie Universitaet Berlin on June 25, 2015
Figure 2. Three-dimensional model of the measurements and calculation sites.
Electric field (V/m)
S. ADDA ET AL.
2
1.5
Emeas
E ray-tracing
1
E far field
0.5
0
SITE1A
SITE1B
SITE1C
SITE2
Figure 3. Comparison between the measured and calculated electric field levels obtained using the ray-tracing model at
the different sites. An uncertainty is associated with each value.
exposure assessment in an urban propagation environment because the results obtained are much
closer to real exposure conditions than the results
obtained using a simplified model (such as a freespace model), notwithstanding the introduced
approximations.
REFERENCES
1. Luneberg, R. M. Mathematical Theory of Optics.
(Boston: Artech House) (1990).
342
2. Kline, M. An asymptotic solution of Maxwell’s
equations. Comm. Pure Appl. Math. 4, 225–262 (1951).
3. Anglesio, L., Benedetto, A., Bonino, A., Colla, D.,
Martire, F., Saudino Fusette, S. and d’Amore, G.
Population exposure to electromagnetic fields generated
by radio base stations: evaluation of the urban background by using previsional model and instrumental
measurements. Radiat. Prot. Dosim. 97(4), 355–358
(2001).
4. Dierckz, P. Curve and Surface Fitting With
Spline. (Amsterdam: Oxford Science Publications)
(1996).
Downloaded from http://rpd.oxfordjournals.org/ at Freie Universitaet Berlin on June 25, 2015
site