EuCAP 2006 - European Conference on Antennas & Propagation

Session: Session 2P3P - Terrain and Building Diffraction (06)
Type: Oral Propagation
Date: Tuesday, November 07, 2006
Time: 14:00 - 17:40
Room: Hermes
Chair: Tjelta & Liniger

Seq   Time   Title   Abs No
1   14:00   Deterministic Propagation Prediction Method Developed by the EBU
Grosskopf, R.


A deterministic propagation prediction method was developed by the European Broadcasting Union (EBU) in order to provide a prediction model which is acceptable to all parties concerned, and which has demonstrated accuracy and reliability. This method is also submitted and proposed to ITU-R WP 3K as a deterministic site-specific prediction method.

Path general methods such as Recommendation ITU-R P.1546 are good for general planning and coordination. They can be used with the minimum of information about the propagation path and can be agreed easily between countries. However, as they take little account of terrain there can be significant prediction errors on individual paths. Thus, for detailed planning and coordination of specific transmitter locations there are considerable benefits in using deterministic prediction methods that take account of the terrain.

Prediction method

This prediction procedure, based on ITU-R Recommendations P.452, P.526, P.1546, is appropriate to broadcasting stations operating in the frequency range of about 100 MHz to 3000 MHz. The method includes a complementary set of propagation models, which ensure that the predictions embrace all the significant propagation mechanisms that can arise.

The procedure calculates 6 propagation losses:

- line-of-sight,

- diffraction,

- tropospheric scatter;

- anomalous propagation (ducting),

- height-variation in ground cover,

- building penetration.

Depending on the type of path, as determined by a path profile analysis, one or more of these propagation mechanisms are used.

The prediction procedure employs three radio-meteorological parameters to describe the variability of anomalous propagation conditions at the different locations around the world:

- delta-N (N-units/km), the average radio-refractive index lapse-rate through the lowest 1 km of the atmosphere, provides the data upon which the appropriate effective Earth radius can be calculated.

- β0 (%), the time percentage for which refractive index lapse-rates exceeding 100 N units/km can be expected in the first 100 m of the lower atmosphere, is used to estimate the relative incidence of anomalous propagation.
- N0 (N-units), the sea-level surface refractivity, is used by the troposcatter model.


The method was tested against the EBU database containing more than 3500 profiles with measurements for a variety of paths, antenna heights and frequencies. In general, the results of the EBU prediction method are very encouraging. The standard deviation of the prediction error is much smaller than in case of predictions with Recommendation ITU-R P.1546 and the correlation coefficient is much greater. However, the mean prediction error of the EBU method is approximately 4 dB, i.e. the predictions are too optimistic. More thought will be needed to determine the cause of this 'optimistic' situation. For one thing, more terrain profiles and corresponding field strength measurements are needed to ensure a proper basis for further testing of the method.

The software should be made available to EBU members to aid in planning exercises and coordination, in particular in preparation of RRC-06.

2   14:20   Digital Elevation Data and Their Use for Improved Broadcast Coverage and Frequency Utilization
da Silva, M.A.N.1; Costa, E.1; Liniger, M.2

Collective effort is in progress within ITU-R Working Party 3K to develop, test, and finally recommend path-specific propagation models for improved broadcast coverage and interference predictions. It is to be expected that, to be effective, increasingly sophisticated prediction models demand more detailed and accurate information on terrain, particularly on its topography. The combination of the horizontal resolution and vertical accuracy of digital elevation models with the methods selected to extract path profiles and to estimate their parameters, as well as with the associated propagation models, will affect predictions. However, increased spatial resolution and vertical accuracy of a digital elevation model are also directly related with the cost and size of the database and with the computer time required to run different applications. Additionally, the used of the same propagation model coupled with topographic data from different sources may lead to also different prediction results, even when all data are based on a common geodetic datum and coordinate system.

The digital elevation model provided by the Shuttle Radar Topography Mission (SRTM) may be suited for the above application. It can be downloaded for all land areas of the Earth between 60° north and 56° south latitude, with data points located at every 3-arc second (approximately 90 meters) on a latitude/longitude grid. Its performance requirements (absolute vertical accuracy of 16 meters and absolute geolocation error of 20 m for 90% of the data) were met and even exceeded for selected test samples, often by a factor of two.

In the present contribution, thousands of path profiles incorporated into the ITU-R Study Group 3 (DBSG5) database and those estimated from the SRTM digital elevation model will be initially compared. It should be observed that the DBSG5 data base receives contributions from several sources and that its path profiles have been estimated using different survey techniques. Virtually all of these path profiles have been obtained independently from the SRTM digital elevation model. Next, diffraction effects on the propagation of radio waves over irregular terrain in the VHF and UHF bands will be estimated using: (i) classical prediction models (Bullington, Epstein-Peterson, and Deygout); (ii) the ones described by the most recent versions of Recommendations ITU-R P.526 and ITU-R P.1546, which may be key elements in the path-specific propagation models being developed; and (iii) computer-intensive models such as the Advanced Propagation Model (APM) and ray tracing. Finally, the predicted field strengths will be compared with those from the corresponding measurements. The mean value and the standard deviation of the difference between measurements and predictions will be presented for each model and for both the DBSG5 and the SRTM terrain profiles, as functions of their number of main obstacles. In addition, the results from SRTM profiles will also be presented for different horizontal resolutions.

3   14:40   Time Variation Characteristics of Wireless Broadband Channel in Urban Area
Suzuki, H.1; Wilson, C.D.1; Ziri-Castro, K.2
2University of Southern Queensland, AUSTRALIA

The use of wireless broadband technologies has been gaining popularity in recent years. Here, wireless broadband refers to a system where a fixed (possibly portable) user device using radio frequencies communicates with base stations (BSs), providing high data rate Internet connection. One of the important differences between emerging wireless broadband technologies (e.g. IEEE 802.16a/e) and traditional services, such as wireless local loop (WLL) and broadband wireless access (BWA), is that the former are designed to operate in a non line-of-sight (NLoS) environment with an omnidirectional user antenna while the latter in LoS with a directional user antenna. Typically the wireless broadband user device is installed by an unskilled person without taking into account an optimal siting of the antenna, while a skilled technician installs the WLL/BWA user device, ensuring that the LoS path or a dominant path exists between the user and BS antennas. Because of this, more variation of signal of signal strength and/or delay spread is experienced with the wireless broadband channel, due, for example, to nearby foliage blown by wind, pedestrian/car traffic, or the movement of the user. The time variation of signal quality due to movement of objects (not of antennas) in multipath environments in the frequency range under interest has been investigated for indoor and outdoor environments, but very rarely for outdoor-to-indoor environments. We have previously investigated the temporal variation of signal quality for a near range (130 m) outdoor-to-indoor link at 2.4 GHz, showing severe variation that approached Rayleigh distribution when nearby foliage is blown by wind at speeds greater than 2 m/s [1]. We have also shown that temporal variations of signal strength due to pedestrian traffic in 5.2 GHz indoor channels approached Rice distribution with K factor decreasing proportionally to the pedestrian traffic in the vicinity of the link [2]. Such effects cause a deteriorated or lost link, yet very few investigations have been reported in the literature that characterise the time variation of wireless broadband channels. This paper investigates the time variation characteristics of outdoor-to-indoor wireless broadband channels in an urban area by recording signal quality measured by commercially available wireless broadband devices for a duration of several weeks. The effects of utilising spatial diversity antennas to mitigate the detrimental effects of temporal fading on the downlink are investigated by measuring common constant pilot signals simultaneously by multiple receivers. Due to increasing interests in utilising VHF and lower UHF bands for their possibly favourable channel characteristics for wireless broadband services, the measurement of digital video broadcasting pilot tones are conducted at the same site and the results are compared with those at higher frequencies.

[1] H. Suzuki, "Diurnal signal fading characteristics of IEEE 802.11b outdoor fixed link," In Proc. of ClimDiff, Nov. 2003.
[2] K. I. Ziri-Castro, N. E. Evans, and W. G. Scanlon, "Propagation Modelling and Measurements in a Populated Indoor Environment at 5.2 GHz," In Proc. of AusWireless, March 2006.

4   15:00   A Hybrid Site-Specific Radio Propagation Model

McKenna, P.
Institute for Telecommunication Sciences, UNITED STATES

Site-general radio propagation models were developed before there was widespread availability of digital terrain elevation databases and low cost computation platforms. As such, these models assumed that only very little information was available to the model user about the details of the radio propagation path between the transmitting and receiving sites. When more detailed information about the propagation path is available, in particular, the terrain elevation as a function of distance along the path, it may be desirable to utilize this information to refine the accuracy of the predictive model by accounting for diffraction over irregular terrain in a path-specific sense. By definition, the use of this more detailed information will make the model site-specific. That is, the resulting prediction will only be applicable to the path in question. A different path can yield a different prediction, even though it shares the same frequency of operation, terminal heights and distance, if its terrain profile is different from the first path. By using a site-specific model in point-to-multipoint mode, one can obtain an area coverage at specific locations for which a site-general model might otherwise be employed. This can improve accuracy and spectral efficiency when designing and planning radio systems.

This site-specific model presented here is intended to be generally applicable to the ranges of antenna heights, frequencies and distances found in Rec. ITU-R P.1546, but also to provide a method for more accurately predicting the effects of irregular terrain diffraction on both line-of-sight and transhorizon land paths when detailed terrain elevation information may be utilized. The fundamental assumption that underpins the model is that, for the cold sea curves of Rec. ITU-R P.1546 (50% time) and for path lengths less than 200 km, the excess losses are dominated by the ground wave (at short to modest distances) blending to the diffraction of the radio wave propagating through a well mixed, temperate atmosphere above a smooth spherical earth (at longer distances), albeit with the electrical material properties of the sea. Therefore, it should be possible to derive a new irregular terrain diffraction loss term or, equivalently, to "correct" the Rec. ITU-R P.1546 land curves by using the methods of Rec. ITU-R P.452 (which includes the methods of Rec. ITU-R P.526, by reference.).

This paper provides a detailed implementation of the model that follows a step-by-step procedure. Following that, comparisons of the models predictions to radio propagation measurements are also provided.

5   15:20   Analysis of the Over-Rooftop Multiple-Building Diffraction in Urban Areas with Shadowing Caused by Terrain Effects
Juan Llacer, L.; Rodríguez, J. V.; Molina García Pardo, J. M.
Universidad Politécnica de Cartagena (UPCT), SPAIN

In this work, a hybrid Uniform Theory of Diffraction (UTD) - Physical Optics (PO) formulation is used for the analysis of the multiple-building diffraction which takes place in urban areas where a bare hill causes shadowing of the transmitter from the receiver (a non-line-of-sight situation).
Cases where the hill is modelled as a cylindrical section (Fig. 1) and where the hill is represented by a wedge (Fig. 2) are analyzed, thus furthering the study presented in [1]. The mentioned UTD-PO solution, validated by numerical results from the technical literature (Fig. 3), has a good computational efficiency, while at the same time increasing the versatility over the method given by Bertoni in [2] (Fig. 4) for the analysis of the same scenarios, since neither a plane-wave incidence over the city nor an elevated number of buildings have to be assumed. Hence, the considered hybrid formulation allows for the study of the above-mentioned environments when the urban area consists of a small number of buildings as well as when a spherical-wave incidence over the latter is considered.


[1] J-V. Rodriguez, J-M. Molina-Garcia-Pardo y L. Juan-Llacer, "A Solution Expressed in Terms of UTD Coefficients for the Analysis of the Multiple Diffraction in Urban Areas with Shadowing Caused by Terrain Effects", Microwave and Optical Technology Letters, vol. 47, no. 6, pp. 523-525, Dec. 2005.
[2] H. L. Bertoni, "Radio Propagation for Modern Wireless Systems", Prentice Hall, pp. 189-192, 2000.

6   16:00   Analysis of the Dominant Propagation Mechanisms for Radio Coverage and Interference Prediction in Urban Scenarios at 2.4, 5.8 & 28.0 GHz
Thomas, N; Willis, M; Craig, K
Rutherford Appleton Laboratory, UNITED KINGDOM

Radio coverage and interference in communications systems is provided by various propagation mechanisms depending on the frequency of the radiowaves and the propagation environment. Key propagation mechanisms are line-of-sight (LOS), diffraction, transmission (or penetration), incoherent scattering, and specular reflections. In urban environments with many tall buildings and antenna locations below rooftop height, the latter mechanisms become increasingly significant. Understanding which propagation mechanisms dominate for coverage and interference calculations is important for cell planning issues.

In this paper we present simulations results determined under the EU Framework 6 Broadwan programme. A novel surface model unifying scatter and Fresnel models has been implemented and used to evaluate the relative importance of the different propagation mechanisms at three frequencies of interest through simulation using an accurate database of a typical UK urban landscape. The results obtained show that a significant coverage fraction may be provided by reflected paths at low frequencies. At higher frequencies incoherent scattering becomes an increasingly important mechanism and may significantly affect interference predictions.

7   16:20   Indoor Coverage Maps Over Large Urban Areas: An Enhanced Ray-Tracing Method
Lostanlen, Y.; Corre, Y.

A majority of mobile communication take place inside buildings (office, home, shopping mall, train station, airport). Consequently radio networks (3G, TV on mobile), must be carefully planned today considering receivers inside buildings. This enlarges the requirements on precise field strength predictions, especially in urban environments. New reliable propagation prediction methods must be set up.

The empirical method commonly used today to get indoor coverage maps consists in the addition of a loss (depending on the elevation of the floor) to the field strength simulated in the street. But this approach does not consider the specificities of the propagation environment, e.g. the position of the transmitter relative to the building or the height of the surrounding obstacles.
3D ray-tracing techniques have the capabilities to simulate the multiple contributions that penetrate inside the buildings. But most of these techniques are limited to small calculation areas.
The paper presents an outdoor-indoor propagation model based on a ray-tracing technique that is specifically optimised for calculations over large urban areas. The model was elaborated to match the requirements of radio planning (fast, robust, precise, tuneable). It provides large multi-floor in-building predictions that may be used to generate precise interference matrices and QoS estimates. It predicts also the channel estimations parameters (delay spread, angular spread, orthogonality factor for CDMA networks).

The outdoor-indoor model is based on a fast ray-tracing technique that computes the predominant 3D ray contributions in any type of urban configuration [1].The ray paths are prolonged along straight lines inside the building polygons. Attenuation is added to the path loss when crossing the building wall (typically 10dB at 900MHz) and propagating inside the building (typically 0.5dB/m at 900MHz).

The path loss computed for the whole ray (from the outdoor transmitter to the indoor receiver) takes precisely into account the elevation of all the interacting elements. Thus the model has the ability to predict realistic coverage maps at different floor levels.

A measurement campaign was carried recently out inside buildings in Paris based on test DVB-T broadcasting signals [2]. Different distances from the transmitter to the building are considered, from few cents of metres to more than 10 km. Measurements were collected along straight lines in the corridors and over regular grids inside the rooms.
Propagation predictions are compared to these measurements. The paper presents some statistical results to illustrate the accuracy of the model (mean error 1dB, error standard deviation 5dB).

The paper will present a validated solution used for predicting the urban outdoor-indoor propagation. The approach is based on an optimised ray-launching method and the used techniques will be clearly illustrated. Besides the simulation of multiple contributions will be also helpful in future studies to evaluate the reception quality, especially associated with diversity techniques.

[1] Y. Corre, Y. Lostanlen, "A 2.5D model for predicting the in-building propagation in urban environments", COST 273, Espoo, Finland, May 02. TD(02)052.
[2] A. Fluerasu, A. Sibille, Y. Corre, Y. Lostanlen, L. Houel and E. Hamman, "A measurement campaign of spatial, angular, and polarization diversity reception of DVB-T", COST 273, Bologna, Italy, Jan. 05.

8   16:40   Full-3D MR-FDPF Method for the Simulation of Indoor Radio Propagation


Several softwares have been developed for computer-aided design of radio networks. The first constitutive element of such a tool is the propagation modeling algorithm. Empirical approaches have difficulties to predict accurately the coverage of access points and ray tracing like models lead to a difficult trade-off between computational load and efficiency when the number of walls is too high. We proposed a new method based on a multi-resolution extension of the TLM approach. Its main property holds in the fact that all reflections and diffractions are taken into account, with no impact on the computational load.

Standard MR-FDPF Model

The 2D MR-FDPF Method has been described in [1]. In this approach the environment is discretized in nodes having inward and outward flows associated with. Gorce et al. proposed to formulate a multi-resolution decomposition of the problem leading to a very efficient computational approach. A coverage calculation is divided into two stages:
-a preprocessing stage that exploits the environment characteristics to compute a binary tree and its associated propagation matrices. This stage has to be computed once and does not depend on the emitters location
-a propagation stage that computes a coverage map once the source location is known. The propagation of flows is made in each nodes and the process either ends at pixels, or stops when a predefined size of nodes is reached
A 2.5D MR-FDPF method can be implemented to evaluate coverage areas in different floors of standard buildings. We obtained in [2] a mean square error of about 5 dB between simulation and measurement in a standard multi-floored building. But the 2.5D approach fail for a full 3D environment such as a wide hall having high free-spaces and connecting multiple floors. In this case only a full 3D approach can succeed in.

Full 3D MR-FDPF Model

Implementation of the full 3D MR-FDPF is not an easy problem. Instead of 2D nodes with 4 vectors for the flows coresponding to the 4 cardinal directions, we now work with 3D nodes with 6 bi-dimensional vectors for the flows on each face of the cube.So the size of matrices and the memory needs increase a lot. A pecular implementation has been proposed to compensate for. In our approach flows are not stored in each node, but each nodes' flows are stored in 6 global environment matrices containing all the flows corresponding to the 6 directions.
The method is implemented in a java developed software with JNI interfaces and C Blas optimized libraries. In this article we will give results concerning performances and precision of the new method.


[1] JM.Gorce, E.Jullo, and K.Runser, 'An adaptative multi-resolution algorithm for 2D simulations of indoor propagation' in Proceedings of the 12th International Conference on Antennas and Propagation. Exeter, UK, April 2003
[2] la Roche, X.Gallon, JM.Gorce, and G.Villemaud, A 2.5d extension of frequency domain parflow method for 802.11b/g propagation simulation in multifloored buildings submitted in:VTC 2006 Spring,Montreal, Canada, September 2006

9   17:00   Combined Urban and Indoor Network Planning Focusing on the Dominant Propagation Paths
Wahl, R; Woelfle, G
AWE Communications, GERMANY

With the growing interest for broadband mobile services in mobile communication networks, the investigation of radio transmission in and into buildings is getting more important. Popular empirical and deterministic models for the propagation inside buildings compute the field strength based on the inner structure of the buildings (walls, furniture). But for current and future wireless networks (3G, B3G, W-LAN, WiMax,..), the neighboring buildings must also be considered to avoid interference problems in these buildings.

Additionally the indoor coverage of outdoor transmitters must be analyzed to guarantee a high QoS even inside buildings. A new concept for the prediction of the field strength in such hybrid scenarios (urban and indoor) is presented in this paper. The new model focuses on the most dominant path(s) between transmitter and receiver. This model allows also the computation of the transition from an urban to an indoor scenario and vice versa, thus allowing an accurate computation of the received power inside and around buildings. For the validation of the new model, measurements were accomplished.

The left picture in figure 1 shows the problem of empirical propagation models: They rely on the direct ray between transmitter and receiver. In urban scenarios this ray includes the over-rooftop diffractions (e.g. Walfisch-Ikegami model) while in indoor scenarios the ray is always the direct line between Tx and Rx. In both scenarios, this direct ray is not always dominant as it is highly attenuated. A model based on this path must lead to errors in scenarios where the direct ray is contributing only a very small part to the total received signal power. In the center part of fig. 1 the principle of ray-optical propagation models is shown. Up to hundreds of rays can be computed for each receiver, which leads to long computation times. The contributions of all rays are superposed (in most cases incoherent superposition) to obtain the received power. But in most cases only 2 or 3 rays are contributing more than 98% of the energy, i.e. by focusing on these dominant rays the accuracy for a pth loss prediction would be sufficient.

Based on the restrictions mentioned above, the Dominant Path Model (DPM) was developed. The algorithm is based on the dominant paths from transmitter to receiver and was adapted to handle objects with arbitrarily located and rotated planes (typical for indoor) as well as very large areas (urban cells with cell radii of several kilometers, see fig. 2 and 3). For the prediction of the path loss along a ray path, the distance from Tx to Rx is considered, as well as the frequency, the attenuations due to interactions, the waveguiding and the Tx power.

An empirical waveguiding factor is introduced to take reflections and scatterings in long straight corridors (indoor) or street canyons (urban) into account. A transition interface for the computation of the indoor penetration (coverage) as well as for the outdoor interference will be described in the final version of this paper. The final paper will also contain further examples and will show several measurement campaigns confirming the accuracy of the model.

10   17:20   Electromagnetic wave propagation using 3D vectorial Parabolic Wave Equation coupled with Leontovich boundary condition

Radar systems and communication links rely on a suitable knowledge of the spatial distribution of the electromagnetic field radiated by the source antenna in urban or natural medium. The two dimensional approximation of the environment ignores lateral reflections or diffractions caused by terrain obstacles and depolarization of the field. To take into account these effects, three dimensional (3D) vector Parabolic Wave Equation (PWE) resolutions have been developed. Assuming that the backscattered field can be neglected, the propagation can be modelled by forwardly marching the electric field, which, in the atmosphere, is governed by the equation:

x is the main propagation axis, y and z are transversal axes, k0 is the free space wave-number, m is the modified refractive index and E is the vector electric field. Refractive index allows us to take into account atmospheric refraction, which cannot be neglected in long range applications.
Two resolutions have been considered: Split-Step Fourier (SSF) and Finite Differences (FD) algorithms. Both are coupled with Leontovich boundary condition to take into account the ground, the relief and the obstacles:
n is the outward normal unit vector to the terrain, E is the vector electric field, Z is the terrain impedance and H is the vector magnetic field.
SSF is based on a plane wave spectral decomposition using FFT transforms on only two components of the electric field [1]. This method is very efficient under grazing angle and applicable on relief and obstacles presenting sligth slopes compared to the main propagation axis. However, on complex cases, its coupling with boundary condition is more delicate and requires small propagation step.
This observation leads us to consider resolution based on FD algorithm which also requires small steps in range but which is more accurate for complex cases. Alternating Direction Iteration (ADI) technique has been implemented to optimize computational time and required memory space. The resolution is performed component per component and the coupling between components is added next. This approximation combined with ADI technique allows us to deal only with tridiagonal matrixes.

Both methods have been implemented and a comparison is realized in terms of accuracy, validity domain, computational time and required memory space. A realistic three dimensional case is represented on the figure below: the propagation inside a valley. This configuration permits to focus on terrain effects and to put in evidence the necessity of a three dimensional resolution.

Fig.: valley configuration

[1] Ramakrishna Janaswamy, Path Loss Predictions in the Presence of Buildings on Flat Terrain: A 3D Vector Parabolic Equation Approach, IEEE Trans Antennas Propagation, 51(8), August 2003