Volcano_Reference_Guide 3.0.pdf

Volcano 3 Reference Guide

Siradel copyright 2010. All rights reserved. Page 2

Copyright

© 2001-2010 SIRADEL SAS. All rights reserved. Volcano 3.0 Reference Guide This manual, as well as the software described in it, is furnished under license and may be used or copied only in accordance with the terms of such license. The content of this manual is furnished for informational use only, is subject to change without notice, and should not be construed as a commitment by SIRADEL SAS. SIRADEL SAS assumes no responsibility or liability for any errors or inaccuracies that may appear in this documentation. Except as permitted by such license, no part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, recording, or otherwise, without the prior written permission of SIRADEL SAS.

France Headquarters SIRADEL S.A.S. 3, allée Adolphe Bobierre CS 24343 35043 RENNES Cedex FRANCE Tel: +33 (0) 2 23 480 500 Fax:+33 (0) 2 23 480 599

Toronto, Canada, American Office Radio Expertise and Research Centre Suite 200 1920 Yonge Street Toronto – ON M4S3E2 - CANADA

Hong Kong, China, Asian Office SIRADEL Asia Limited 1806-08, Tower Two, Lippo Center, 89 Queensway, Hong Kong


Content

Copyright ............................................................................................................................................... 2 Content .................................................................................................................................................. 3 1. Introduction ................................................................................................................................... 5 2. Explanation of Symbols ................................................................................................................... 5 3. Cells definition ................................................................................................................................ 6

3.1. Macro-cells ............................................................................................................................... 6 3.2. Micro-cells ................................................................................................................................ 6 3.3. Mini-cells ................................................................................................................................. 7

4. Volcano Rural model ...................................................................................................................... 8 4.1. Introduction ............................................................................................................................... 8 4.2. Map Data Requirements.............................................................................................................. 8 4.3. Volcano Rural Model Parameters ................................................................................................. 9 4.4. Concept .................................................................................................................................. 10

4.4.1. Introduction...................................................................................................................... 10 4.4.2. Processing the Vertical Profile ............................................................................................ 10 4.4.3. Path Loss Estimate ............................................................................................................ 12

4.5. Volcano Rural Functionalities ................................................................................................... 13 4.5.1. Antenna Masking for Colocalized Transmitters .................................................................... 13 4.5.2. Coverage Prediction Mode ................................................................................................. 13 4.5.3. Raster Resolution Strategy ................................................................................................. 14

4.6. Volcano Rural Model Tuning .................................................................................................... 14 4.6.1. Description ...................................................................................................................... 14

4.7. Advanced Tuning for Low-Resolution Map Data ......................................................................... 15 4.7.1. Description ...................................................................................................................... 15 4.7.2. Tuning Clutter Heights ...................................................................................................... 16 4.7.3. Tuning Clutter Attenuations ............................................................................................... 19

5. Volcano Urban Model ................................................................................................................... 22 5.1. Introduction ............................................................................................................................. 22 5.2. Map data requirements ............................................................................................................. 22 5.3. Concept .................................................................................................................................. 23

5.3.1. Introduction...................................................................................................................... 23 5.3.2. Direct Path Component ...................................................................................................... 24 5.3.3. Multiple Path Component .................................................................................................. 25 5.3.4. Model parameters ............................................................................................................. 27

5.4. Path loss estimate ..................................................................................................................... 28 5.5. Volcano Urban tuning .............................................................................................................. 29 5.6. Advanced tuning for low resolution map data .............................................................................. 29 5.7. Volcano Urban functionalities ................................................................................................... 29

5.7.1. Antenna Correction ........................................................................................................... 29 5.7.2. Spatial averaging .............................................................................................................. 30 5.7.3. Street level mode .............................................................................................................. 31 5.7.4. Coverage prediction mode .................................................................................................. 31 5.7.5. Indoor penetration ............................................................................................................. 33 5.7.6. Indoor transmitter ............................................................................................................. 35 5.7.7. Antenna Masking/Unmasking............................................................................................. 36 5.7.8. Wideband additional outputs .............................................................................................. 36

6. Volcano Indoor Model .................................................................................................................. 39 6.1. Description.............................................................................................................................. 39 6.2. Model Overview ...................................................................................................................... 39 6.3. Supported geographical data ...................................................................................................... 39

7. Antenna modelling ........................................................................................................................ 41 7.1. Description.............................................................................................................................. 41


7.2. Concept .................................................................................................................................. 41 8. References .................................................................................................................................... 43 9. Technical Support ........................................................................................................................ 44


1. Introduction

SIRADEL's Volcano is a high performance software component giving an implementation of deterministic radio propagation models based on the description of the environment. It is delivered as a library that can easily be integrated in products such as radio planning and optimization tools thanks to a well designed application programmer interface. Volcano models are able to simulate macro-cells, mini-cells and micro-cells in rural, suburban and dense urban environments. Volcano models also offer possibility to compute predictions in indoor environment for pico-cells and femto-cells. This document aims at describing the principles of SIRADEL's Volcano propagation models. Three models are available in the Volcano propagation model suite, namely the Volcano Rural Model, the Volcano Urban Model and the Volcano Indoor Model. Those three models will be described in this document, giving the necessary information for the user to understand Volcano models functionalities, the role of the different model parameters and the meaning of the results.

2. Explanation of Symbols

Reference Note Tip Warning


3. Cells definition

In this document, a “cell” refers to the coverage area around one transmitter antenna. Different types of cells may be distinguished according to the transmitter configuration as well as to its surrounding environment. For each type of cell, the way the radio energy is propagated is different leading to several approaches in terms of propagation modeling. This section aims at describing the kinds of cells that may be encountered during radio planning processes.

3.1. Macro-cells

The Macro-cell configuration refers to transmitter antennas located at a height that provides a clear view over the surrounding building or vegetation. Such configurations can be retrieved in rural and urban environments with transmitters located on pylon or on a rooftop higher than the surrounding roof level (Figure 1). The radio energy with a macro-cell antenna is therefore mainly propagated through diffraction over top of the obstacles.

Figure 1: Macro-cell transmitters

3.2. Micro-cells

The Micro-cell configuration refers to transmitter antennas located at very low level (below the surrounding obstacles). Micro-cells are present in urban and suburban environments with transmitters located on street lamps or against a wall at street level for example (Figure 2). The radio energy is mainly propagated by canyon effect along the streets.


3.3. Mini-cells

The mini-cell configuration refers to intermediate configurations between Macro-cells and Micro-cells in urban and suburban environments:

The emitting antenna is located on a rooftop but not higher than the surrounding roof level

The emitting antenna is against a building façade but not lower than the surrounding roof level In these configurations, the radio energy is partly propagated above obstacles and partly along the streets.

Figure 2: Street-level transmitter

Figure 3: Mini-cell transmitters


4. Volcano Rural model

4.1. Introduction

The Volcano Rural model is a proprietary deterministic model that computes the attenuation from the transmitter to the receiver using the vertical terrain profile between those two points. The model is based on the multiple knife-edge diffraction method [1], which is specially adapted to Macro-cell transmitters (refer to section 3 for further details about Macro-cell configurations). Volcano Rural model is suited to predictions in rural and suburban environments based on low-resolution geographical map data (i.e. raster resolution greater or equal to 20m).

4.2. Map Data Requirements

This section aims at describing the required Map data to use with the Volcano Rural Model. Table 1 summarizes the different possibilities whether Map data is based on Raster or on Vector.

Table 1: Map data supported by Volcano Rural Model (Based on Raster)

Map data elements Based on Raster

Comments Recommendations

Height raster data or Digital Terrain Model

(DTM) Optional

Strongly recommended, especially in hilly regions

Clutter raster data or Digital Land Usage

(DLU) Mandatory

25-metre-resolution (low-resolution) recommended

5-metre-resolution (high-resolution) can be used in urban areas

Clutter height raster data or Digital Height

Model (DHM) Optional

Recommended in areas with high-resolution clutter data

Average clutter height

Association between an average clutter height and

one clutter type Used in areas where clutter

height raster data is not available

As mentioned in the introduction, Volcano Rural model is dedicated to predictions in rural environments where low-geographical databases are available. Although the rural model is not dedicated to be used in vector areas, Volcano allows computations on these areas. This feature is mainly used when simulating coverage from rural sites to urban areas.

Volcano Rural model allows a mixed use of raster and vector data. Volcano Rural Model offers the possibility to define priorities between raster and vector data, when both are available:

Using the raster favorite mode, the terrain profile between the transmitter and the receiver is preferably extracted from the raster layers: DTM, DLU and DHM when available

Using the vector favorite mode, the terrain profile between the transmitter and the receiver is preferably extracted from the vector layer and the raster DTM layer


More information about Map data is available in Map data requirement [SIR2].

4.3. Volcano Rural Model Parameters

Table 2 lists and gives a rough description of the parameters involved in the Volcano Rural model. The role of each parameter will then be explained in detail in the following sections.

Table 2: Volcano Rural model parameters

Parameters

Description

User-defined or Tunable

Default value Refer to section

ALOS (dB) Correction of free space loss predicted at 1 meter for Line Of Sight (LOS) links

Tunable 0 4.4.3

BLOS (dB/dec) Correction of the distance dependence

for Line Of Sight (LOS) links Tunable 24 4.4.3

ANLOS (dB) Correction of free space loss predicted

at 1 meter for Non Line Of Sight (NLOS) links

Tunable 0 4.4.3

BNLOS (dB/dec)

Correction of the distance dependence for Non Line Of Sight (NLOS) links

Tunable 24 4.4.3

α Weight of the Volcano deterministic

attenuation Tunable 0.7 4.4.3


4.4. Concept

4.4.1. Introduction

This section describes in detail the concept of the Volcano Rural model and the role of each parameter. Along the radio link from the transmitter to the receiver, various kinds of obstacles, such as hills, trees or buildings, give rise to diffraction phenomena. The main obstacles are replaced by knife-edges and the diffraction attenuation due to any single knife-edge is calculated by means of an approximate expression of the Huygens principle. The description of the vertical terrain profile between a transmitter and a receiver is extracted from the map raster data and/or 3D vector data according to the available map data and the user preferences (raster/vector favourite). Figure 4 illustrates the vertical terrain profile extraction based on raster data.

(a) (b) (c)

Figure 4: Extraction of the vertical terrain profile between the transmitter and the receiver from the (a) height raster

data, (b) clutter raster data and (c) clutter heights raster data; the vertical terrain profile is made up of the terrain

height (brown), built-up areas (grey) and vegetation (green)

4.4.2. Processing the Vertical Profile

4.4.2.1. Earth Bulge

In order to take into consideration the curvature of the earth’s surface and the bending of the radio path in the atmosphere, the heights are corrected for long radio links. That correction is carried out by use of the effective earth’s radius. Actually the gradient of the refractivity index in the atmosphere causes bending of the radio propagation trajectories and the substitution of the real earth’s radius Re by the effective earth’s radius Re’= kxRe allows to consider that radio waves are travelling along straight lines. The multiplication factor k is directly related to the refractivity gradient; it is roughly equal to 4/3 for a typical gradient of -39 N/km.

150

200

250

300

350

400

450Altitude (m)


4.4.2.2. Visibility Conditions

The Volcano Rural model detects optical visibility conditions: Line Of Sight (LOS) or Non Line Of Sight (NLOS)

4.4.2.3. Fresnel Analysis and Knife Edge Definition

Volcano Rural model detects the main obstacles (buildings, vegetation or ground) along the radio link and approximate them as knife-edges, as illustrated in Figure 5. In this example, the hill placed in the middle of the radio link is the main diffracting obstacle. It is replaced by the first order knife-edge (red). The woods on the left-hand side and the building on the right-hand side are replaced by second order knife-edges (blue).

The main obstacles are detected from their Fresnel-Kirchoff parameter , which represents the portion

of the first-Fresnel ellipsoid that is obstructed. This parameter is calculated as shown in Figure 6. The total diffraction loss LDET is obtained from a combination of the individual knife-edge diffraction

losses, computed from as described in Figure 7 [1]. Note that in the case of hills with rounded tops, correction is introduced in the calculation to properly consider diffraction by rounded obstacles.

Figure 6: Knife-edge diffraction (Fresnel-Kirchoff parameter )

Figure 5: Vertical profile between the transmitter and the receiver (green: wood; grey: building) and resulting knife-

edges


4.4.3. Path Loss Estimate

The path-loss is calculated from the total diffraction loss (or deterministic loss) LDET as follows:

𝑷𝑳 𝒅𝑩 = 𝑨 + 𝑩. 𝒍𝒐𝒈 𝒅 + 𝜶. 𝑳𝑫𝑬𝑻 + 𝑳𝑭𝑺 + 𝑳𝑪 + 𝑳𝑨𝑵𝑻

Eq. 1

A is the correction of free space loss at 1 meter (dB)

B is the correction of distance dependence (dB.dec-1)

d is the path length (m)

α is the weight of the deterministic loss

LDET is the deterministic loss based on multiple knife-edge diffraction (dB)

LFS is the free space loss at distance 1m (dB)

f is the frequency in MHz

LC is the clutter attenuation at the receiver location (dB)

LANT is the antenna loss, including antenna correction (dB)

The values of both parameters A and B involved in the calculation of the free space correction are given from parameters ALOS and BLOS for Line Of Sight (LOS) radio paths and given from parameters ANLOS and BNLOS for Non Line Of Sight (NLOS) radio paths. These values, as well as the value of

parameter , are generally defined by means of the model tuning process (see section 4.6). The free space loss at distance 1m is given by

𝑳𝑭𝑺 𝒅𝑩 = −𝟐𝟕.𝟓𝟔 + 𝟐𝟎𝐥𝐨𝐠(𝒇) Eq. 2

Specific clutter attenuation LC is associated to any clutter type of the map data. The objective of this clutter attenuation is to correct the path loss depending on the receiver location and the receiver clutter type. The following expression is used to determine the received power at the receiver location

𝑹𝒆𝒄𝒆𝒊𝒗𝒆𝒅 𝑷𝒐𝒘𝒆𝒓 𝒅𝑩𝒎 = 𝑬𝑰𝑹𝑷−𝑷𝑳 Eq. 3

EIRP is the Equivalent Isotropically Radiated Power (dBm)

PL is the Path loss (dB)

Please refer to Map data requirements for Volcano [SIR2] for further information on clutter attenuations. As explained in this document, the clutter attenuation assigned to a clutter type is not

Figure 7: Diffraction loss vs. Fresnel-Kirchoff parameter


added to the propagation loss in a systematic way. It depends on the clutter type (land, vegetation, building, built-up area, bridge, water) and the clutter location (inside the clutter or above the clutter).

4.5. Volcano Rural Functionalities

4.5.1. Antenna Masking for Colocalized Transmitters

Volcano Rural model integrates an automatic feature to speed up the computation process of co-localized transmitters. When using co-localized transmitters, the path loss matrix is only computed for the first transmitter. The received power matrix of every transmitter is computed from the same path loss matrix, but changing the impact of antenna pattern, antenna gain and EIRP.

4.5.2. Coverage Prediction Mode

Volcano Rural model proposes two different modes for prediction of received power matrices (coverage prediction): Regular mode (higher accuracy) and Boost mode (faster). In Regular mode, the received power is calculated from a precise analysis of the vertical terrain profile between the transmitter and the receiver positions. In particular, the local spatial variations of the received power are directly related to the precise position of obstacles and receivers. The boost mode in the Volcano Rural model permits to strongly reduce the computation time, using approximates of the terrain vertical profiles, thus slightly degrading the received power prediction by removing the deterministic local variations. Degradation on the received power accuracy is controlled: mean error is close to 0 and standard deviation is lower than 1dB. Computation gain varies according to the resolution of the map data that is used for the calculation and the prediction radius. However, for one given map data resolution, the larger the prediction radius, the greater the computation time gains.

The choice of the coverage mode (Regular or Boost) has no impact on the tuning of the model. A model that has been tuned without the boost option can be reused for large scale predictions with the boost option switched on (no need to realize a specific tuning).

Figure 8: Volcano Rural prediction (40km²) with regular mode (left) and boost mode (right)


4.5.3. Raster Resolution Strategy

When having several resolutions for raster data, this option lets the user define a strategy regarding the map data resolution to be used for the prediction. It will set the priority for the resolutions to be used by Volcano according to the prediction step that is set for the prediction within the radio planning tool. Closest to prediction step Highest resolution In Figure 9 and Table 3: Raster resolution strategy, a concrete example describes the impact of the raster resolution strategy on the map data used by Volcano to construct the vertical profile between both black dots.

Table 3: Raster resolution strategy

Closest to prediction step Highest resolution

Prediction step Path 1 Path 2 Path 1 Path 2

5m 5m 25m 5m 25m

25m 25m 25m 5m 25m

Remark that Volcano permits to define a reference prediction step for tuning and predictions along routes. The resolution strategy will consider this reference prediction step for selection of the appropriate resolution strategy. When tuning Volcano Rural model for a specific usage (e.g. for coverage with 50m resolution step), we recommend setting the reference prediction step equal to the future usage prediction step.

4.6. Volcano Rural Model Tuning

4.6.1. Description

Volcano Urban model has a unique auto-tune function that provides all adjusted parameters in one run, hence facilitating the tuning process and reducing tuning time. It provides a set of parameter values that gives a zero mean error and low RMS error. The purpose of the tuning is to adjust Volcano parameters to the radio channel characteristics and to reduce prediction errors. Tuning process is based on comparisons between measurements and Volcano predictions. The tuning process provides two kinds of results:

Figure 9: Raster resolution strategy illustration


The error statistics between the measurements and the predictions before and after tuning

The adjusted parameter values of the calibrated model The objective of the tuning process is to improve the Volcano Rural predictions by introducing statistical correction (free space corrections A/B and weight of deterministic attenuation alpha). This is performed by a balanced technique: reducing the weight of the deterministic attenuation calculated from the terrain profile and adding a statistical correction to free space loss.

The model tuning is recommended in order to get robust operational propagation models. Table 4 gives the list of tuned parameters and typical tuning ranges.

Table 4: Calibrated parameters (Description, Typical values and range after tuning)

Parameters Description Typical Value

Typical Tuning ranges

ALOS (dB) Correction of free space loss predicted at 1 meter for

Line Of Sight (LOS) links 0 [-2;2]

BLOS (dB/dec) Correction of the distance dependence for Line Of Sight

(LOS) links 24 [20;26]

ANLOS (dB) Correction of free space loss predicted at 1 meter for

Non Line Of Sight (NLOS) links 0 [-2;2]

BNLOS (dB/dec) Correction of the distance dependence for Non Line Of

Sight (NLOS) links 24 [20;26]

α Weight of the deterministic attenuation 0.7 [0.5;1]

To get further details about recommendations, please refer to the Volcano Tuning guide [SIR3]. Tuning recommendations concern:

The distribution of the measurement bins around the transmitter

The accuracy on the transmitter location

4.7. Advanced Tuning for Low-Resolution Map Data

4.7.1. Description

This feature allows the adjustment of two further types of statistical parameters in the Volcano Rural model tuning:

The clutter attenuations

The average clutter heights The clutter attenuation represents the attenuation added to the propagation loss when the receiver is located in the clutter. The default clutter attenuations that are proposed by Volcano are 0dB for open areas, 2dB for vegetation clutter, 10dB for building clutter, 0dB for built-up areas and –2dB for water. The average heights are associated to the low-resolution clutter types (i.e. clutter types used in the low-resolution clutter raster layers). The average clutter heights are typically 0 meter for open areas, a few tens of meters for vegetation areas, from 3 meters to many tens of meters in urban areas, etc. The initial value of both these parameters, the clutter attenuations and the average heights, are generally defined either from Volcano default values (only clutter attenuations), from the user experience or from the geo map data provider (only average heights). However those parameters can be advantageously adjusted for the propagation predictions (adjusted for one specific region, one specific frequency band, etc.). The Volcano advanced tuning permits an automatic adjustment by comparison between measurements and predictions.


The correction of all clutter attenuations and heights is not enforced when the Volcano advanced tuning is carried out. The user has the possibility to select the list of clutter attenuations he wants to adjust. He selects as well the list of clutter average heights he wants to adjust. The Volcano advanced tuning offers automatic feature to find, on a per clutter basis, both the clutter attenuations and clutter average heights optimal to reduce the prediction error.

4.7.2. Tuning Clutter Heights

4.7.2.1. Case Study

Consider the map data proposed in Figure 10. This is a multi-resolution map data, with low-resolution raster data in the whole area and high-resolution raster data in the urban part. The description of the clutter height is available only in the urban part.

The user has to associate an average height to each low-resolution clutter class. This is detailed in Table 5.

Table 5: List of average clutter heights

Clutter class Type Volcano class Average height (m)

Open Low-resolution Land 0

Inland water Low-resolution Water 0

Woodland Low-resolution Vegetation 20

Villages Low-resolution Built-up 10

Agricultural field Low-resolution Vegetation 0

Industrial Low-resolution Built-up 10

Figure 10: Clutter description of a digital map composed of high-resolution data superimposed to the low-resolution

data in the dense urban areas.


Urban Low-resolution Built-up 15

Open High-resolution Land 0 (HR clutter)

Tree High-resolution Vegetation 0 (HR clutter)

Habitation High-resolution Building 0 (HR clutter)

Industry High-resolution Building 0 (HR clutter)

Road bridge High-resolution Bridge 0 (HR clutter)

No Average height is defined for High-resolution clutter since this information is already included in the clutter height layer (DHM). The average heights given to the low-resolution clutter types in the table above are arbitrary. They may be standard heights automatically given to any kind of forest, any kind of village, etc… The role of the advanced tuning process is to determine the optimal height values, i.e. the ones that give the most realistic Volcano radio simulations in the area under study based on the survey data. A typical result of the tuning process is presented in Table 6.

Table 6: Average clutter heights after tuning

Clutter class Type Average height after

tuning (m)

Open Low-resolution 0

Inland water Low-resolution 0

Woodland Low-resolution 24

Villages Low-resolution 8

Agricultural field Low-resolution 0

Industrial Low-resolution 7

Urban Low-resolution 11

Open High-resolution 0

Tree High-resolution 0

Habitation High-resolution 0

Industry High-resolution 0

Road Bridge High-resolution 0

If an average height is specified to a high-resolution clutter and raster clutter height is available, then only the height extracted from the raster map data is used during the prediction; the specified average height has no impact on the prediction and no adjusted value can be returned by the tuning process.


If the measurement bins used in the tuning are all located in the high-resolution area, and highest resolution is preferred (see model parameters in section 4.5.3), then no adjustment can be done for the average clutter heights.

4.7.2.2. Tuning Rules

The tuning of the average clutter heights complies with the rules defined in Table 7.

Table 7: Description of tuning rules

No tuned clutter heights

The tuning of average heights associated to clutters of type land, water or bridge is not allowed. The average heights equal to zero before tuning (e.g. for agricultural fields or parks) are not changed by the tuning process

Tuned clutter heights Only the non-zero average heights associated to clutters of type building, vegetation or built-up may be adjusted by the tuning process.

Tuning range If the average height before tuning is h, then the adjusted height after tuning is in the range from MAX{ h/2, (h – 10m)} to (h + 10m).

Integer values The tuned average heights are integer values.

4.7.2.3. Recommendations

The advanced tuning is suitable only for low-resolution clutters without any accurate height information. In the case of clutter types such as building_8m_10m, building_10m_12m, vegetation_14m_16m, vegetation_16m_18m, etc., the average height associated to the clutter is very accurate. Then it is not required to apply the tuning on those clutter heights. Volcano allows the user to select the list of clutter average heights to be tuned. Table 8 gives the default selection that is made in the Volcano interfaces depending on the clutter classes.

Table 8: Default selection of the tuned average heights

Volcano clutter class

Tuning the average height Default selection

Comments

Land No Not allowed.

Vegetation Yes

This clutter type may be either high-resolution clutter or low-resolution clutter. Volcano suggests by default to tune the associated average height.

However, to be changed for vegetation clutters that already have accurate height description.

Building No

This clutter type is generally high-resolution clutter. That is why Volcano proposes by default to

not tune their associated height.

However, to be changed for building clutters that do not have accurate height description available.

Built-up area Yes

That clutter type is found only in low-resolution map data. Then the default selection suggests

tuning the associated height.

However, to be changed for built-up clutters that already have accurate height description.


Water No Height is necessarily zero; no need for an

adjustment.

Bridge No Not allowed

4.7.3. Tuning Clutter Attenuations

4.7.3.1. Study Case

Consider the map data and the clutter types presented previously. Volcano users must associate a clutter attenuation to any clutter type (from low-resolution data or high-resolution data). Table 9 gives an example of such an association.

Table 9: List of clutter attenuations

Clutter class Type Volcano class Attenuation (dB)

Open Low-resolution Land 0

Inland water Low-resolution Water -2

Woodland Low-resolution Vegetation 2

Villages Low-resolution Built-up 0

Agricultural field Low-resolution Vegetation 0

Industrial Low-resolution Built-up 0

Urban Low-resolution Built-up 0

Open High-resolution Land 0

Tree High-resolution Vegetation 2

Habitation High-resolution Building 10

Industry High-resolution Building 10

Road bridge High-resolution Bridge 0

The clutter attenuations in Table 9 were defined according to the Volcano default values. However those values are not necessary the most suitable for the area under study, the emitting frequency band, the kind of trees met in the woodland areas, the material used in the urban areas, the density of habitations in the villages, etc… The role of the advanced tuning process is to determine the optimal attenuation values, i.e. the ones that give the most realistic Volcano radio simulations in the area under study. A typical result of the tuning process is presented in Table 10.

Table 10: Clutter attenuations after tuning

Clutter class Type Clutter attenuation after

tuning (dB)

Open Low-resolution 0.0

Inland water Low-resolution -2.0


Woodland Low-resolution 3.2

Villages Low-resolution 1.4

Agricultural field Low-resolution 0.6

Industrial Low-resolution - 1.8

Urban Low-resolution -1.5

Open High-resolution 0.0

Tree High-resolution 1.0

Habitation High-resolution 10.0

Industry High-resolution 10.0

Road bridge High-resolution 0.0

Measurement bins are generally not present on water; for this reason, in most of the cases, the tuning process does not adjust the attenuation for water clutter classes.

Measurement bins are generally not present inside buildings; for this reason, in most of the cases, the tuning does not adjust the attenuation for building clutter classes.

4.7.3.2. Tuning Rules

The tuning of the clutter attenuations complies with the rules given in Table 11 and Table 12.

Table 11: Description of tuning rules for Clutter attenuation

Land and Bridge No tuning is allowed

Vegetation and Building

Only positive or 0 dB attenuations are returned for vegetation clutter types and for building clutter types. Indeed that kind of clutter is expected to provide additional propagation loss.

Water Only negative or 0 dB attenuations are returned for water clutter types. Indeed that kind of clutter is expected to increase the signal strength by reflection (conductive material)

Built-up The returned attenuation for built-up clutter types may be positive or negative (≥ -2dB).

Table 12: Minimum requirements for measurement bins per clutter type

Measurement bins A minimum number of bins located on a particular clutter type is required to adjust the associated attenuation.


The tuning must be used to adjust the attenuation of one particular clutter only if measurements were really carried out in this clutter. Usually the measurements are done in the streets or any other kind of open areas. Then, in most of the cases, the measurement bins that are located on clutter types such as water or building are not in the real investigated clutters. They are misplaced because of the database inaccuracy, the database resolution or the positioning inaccuracy. In the case where no measurement


bins are expected to be in building clutter or in water clutter, we recommend not to activate the tuning process for those clutter types. The resulting clutter attenuations could not be properly adjusted. Volcano allows the user to select the list of clutter attenuations to be calibrated. Table 13 gives the default selection that is made in the Volcano interfaces depending on the clutter classes.

Table 13: Default selection of the calibrated clutter attenuations

Volcano clutter class

Calibrating the attenuation Default selection

Comments

Land No Not Allowed

Vegetation Yes To change only if no measurement bin is expected

to be in vegetation.

Building No To change only if measurements were really carried

out inside building.

Built-up area Yes To change only if no measurement bin is expected

to be in built-up areas.

Water No To change only if measurements were really carried

out above water.

Bridge No Not Allowed


5. Volcano Urban Model

5.1. Introduction

Volcano Urban model is dedicated to predictions in dense urban, urban and suburban environments based on 2D low-resolution (i.e. raster resolution greater or equal to 20m) or 3D high-resolution geographical map data (i.e. raster resolution less or equal to 10m and/or 3D vector representation). Volcano Urban model is able to carry out predictions with urban transmitters in Macro-cells, Mini-cells and Micro-cells (refer to section 3 for the description of cell categories). Depending on the transmitter configuration and user settings, Volcano Urban model takes advantage of direct path only or direct and multi-path contributions together.

5.2. Map data requirements

This section aims at describing the required Map data to use with the Volcano Urban Model. Table 14 and Table 15 summarize the different possibilities whether Map data is based on Raster or on Vector.

Table 14: Map data supported by Volcano Urban Model (Based on Raster)

Map data elements Based on Raster


Height raster data (DTM)

Optional Strongly recommended, especially in hilly

regions

Clutter raster data (DLU)

Mandatory 5-metre-resolution (high-resolution) is

recommended for urban areas, 25-metre-resolution (low-resolution) elsewhere

Clutter height raster data (DHM)

Optional Recommended in clutter high-resolution

areas

Figure 11: 3D View contributions from a mini-cellular site to street-level receivers


Average clutter height

Association between an average clutter height and one clutter

type Used in areas where clutter

height raster data is not available

Building vector data Mandatory when Multipath

component is activated

3D building vector data is strongly recommended (in opposition to 2D

building contours)

Table 15: Map data supported by Volcano Urban Model (Based on Vector)

Map data elements Based on Vector


Height raster data (DTM) Optional Strongly recommended, especially in hilly

regions

Vector layer Mandatory 3D vector layer is strongly recommended

The description of all clutter types is strongly recommended as well

Average clutter height Only if the elevation is not

available in the vector layer Not recommended

Volcano Urban model allows a mixed use of raster and vector data:

Using the raster favorite mode, the terrain profile between the transmitter and the receiver is preferably extracted from the raster layers: DTM, DLU and DHM when available. The building vector data are necessary to predict multi-paths contributions

Using the vector favorite mode, the terrain profile between the transmitter and the receiver is preferably extracted from the vector layer and the raster DTM

More information about Map data is available in the Map data requirements [SIR2].

5.3. Concept

5.3.1. Introduction

Volcano Urban model is able to deal with transmitters in urban Macro-cells, Mini-cells and Micro-cells. Depending on the transmitter configuration and user settings, Volcano Urban model takes advantage of direct path only or direct and multi-paths contributions together. The direct-path component is mandatory, resulting from the analysis of the vertical terrain profile along the straight path between both radio terminals. The multi-path component relies on ray-tracing. It is optional. Its activation permits to get accurate results in areas where signals are propagating along the streets (even after propagation above rooftops). This multi-path component is recommended in configurations where canyoning is significant: transmitter located in dense urban area, Mini-cell transmitter or Micro-cell transmitter. It is necessary to predict wideband channel properties (delay spread, departure angles, arrival angles, etc.). When the multi-path component is activated, the computation area is divided into three reception regions where the categories of rays allowed to catch the receivers differ. This division permits the computation of large coverage maps with competitive times, whereas the multi-paths having a preponderant impact on the narrowband power are still constructed. The division of the computation area is illustrated in Figure 12 and described in Table 16. In the NEAR-reception region, all types of ray contributions are computed. Thus the narrowband and wideband


channel characteristics may be predicted there. In the FAR-reception region, only the direct-path is computed, allowing only for the computation of the narrowband received power.

Table 16: Description of the computation regions, when multi-path component is activated

NEAR region All types of contributions are calculated.

FAR region Only the direct path is calculated

TRANSITION region A linear transition is carried out from the field strength as calculated in the

NEAR region and the field strength as calculated in the FAR region.

The user specifies the width of the NEAR region. The recommended widths (when dealing with a mini-cellular site) are from 500 meters to 1000 meters. The default width is 800 meters.

The width of the transition region is 100 meters. The user cannot modify it.

Figure 12: Calculation area around the transmitter is divided in three regions: the circular NEAR region (orange

area), the circular TRANSITION region 100 meter wide (yellow area) and the FAR region

5.3.2. Direct Path Component

Direct-component is always present in Volcano Urban predictions. It is based on the on the same basic principles than the multiple knife-edge (MKE) technique used in the Volcano Rural model (refer to section 4.4).


The terrain vertical profile is extracted from the map data. The main obstacles (buildings, vegetation or hills) are identified according to their Fresnel- the diffraction loss along the direct path.

The Volcano Urban model implements a multiple knife-edge diffraction technique that has been adapted to urban and dense urban environment in order to strongly enhance the result accuracy. In urban radio configurations, the radio terminals are usually located close to significant building obstacles (rooftops or building façades). In this situation, the methods based on the multiple knife-edge diffraction generally provide a path loss that is overestimated. Actually, in reality, various propagation paths generated around the direct path provide contributions passing above rooftops or resulting from lateral interactions. The diffraction loss calculated by the Volcano Urban model is corrected to avoid this overestimation. Figure 13 is a typical example of a transmitter located in dense urban environment. In such a case, the adaptation of the Volcano Urban model permits to stick with reality.

5.3.3. Multiple Path Component

Multi-path component is optional. It may be activated to compute accurate received power in dense urban areas, especially for Mini-cell and Micro-cell transmitters. It also enables the user to get Wideband channel properties (delay spread, departure angles, arrival angles, etc). When multiple paths option is activated, Volcano Urban model combines the direct-path contribution as well as multi-paths that undergo reflections on the vertical building façades and diffraction on the vertical building edges. These multi-paths are predicted in the NEAR and TRANSITION regions only, although the direct-path is predicted over the whole reception area (refer to Figure 12). The multi-path calculation is based on a ray launching technique that computes 3D contributions from interactions with the building facades, building rooftops, hills and vegetation [2]. An optimised combination of multiple knife-edge technique and UTD (Uniform Theory of Diffraction [3]) permits the construction of those 3D contributions and the calculation of the field strength. A specific antenna gain is calculated for each contribution according to its emitting and arrival angles. A list of 3D contributions is calculated at each receiver location. Wideband channel properties like the delay spread, departure and arrival angle spreads may be derived. And the propagation loss (called deterministic loss in section 0) is obtained from the sum of field strength squares, assuming uncorrelated random field phases. The user sets the maximum number of reflections (NR) and maximum number of diffractions (ND) allowed on vertical building façades along one single ray path.

Figure 13: Direct path illustration


There is no maximum number of diffractions in the ray vertical profile (resulting from interaction with rooftops, hills or vegetation), as this is handled by the multiple knife-edge technique.

5.3.3.1. Forward Ray Contributions (Reflection/Diffraction after

Propagation Above Rooftops)

The multi-path ray contributions described in this section result from lateral interactions (reflections and/or diffraction) taking place at street level after possible propagation above rooftops (Figure 14). Reflections occur on the vertical building facades and diffractions occur at the vertical building edges. At most NR successive reflections and ND diffraction are allowed along one single ray path.

The first reflection of the multi-path contributions is allowed to take place above the rooftop of a podium building (represented as a polygon included in another polygon), as illustrated in Figure 15.

5.3.3.2. Backward ray contributions (Reflections/Diffractions before

propagation above rooftops)

The multi-path contributions described in this section result from lateral interactions (reflections and/or diffraction) taking place in the transmitter vicinity before propagation above rooftops. The first interaction is obligatorily in Line-Of-Sight (LOS) from the transmitter, either a reflection on the vertical building façade or diffraction at an horizontal building edge. It may be followed by (NR-1) reflections and ND diffraction on building façades before propagation above rooftops.

(a) (b) (c)

Figure 14: Top view of (a) reflected rays, (b) diffracted rays, (c) reflected and diffracted rays after

propagation above rooftops.

Figure 15: Reflection above rooftops (podium)


5.3.4. Model parameters

Table 17 and Table 18 list and give a rough description of the parameters involved in the Volcano Urban model. The role of each parameter will then be explained in detail in the following sections.

Table 17: Volcano Urban general parameters

Parameters

Description


Recommended or default value

Refer to section

AFAR (dB) Correction of free space loss predicted at 1 meter for receivers in the FAR region,

]-∞;+∞[ Tunable 0 5.4

BFAR (dB/dec) Correction of the distance dependence for receivers in the FAR region, [0;+∞[

Tunable 24 5.4

α Weight of the Volcano deterministic

attenuation, [0;1] Tunable 0.7 5.4

Antenna Correction

Correction of the antenna diagram: Yes/No

User-defined Yes 5.7.1

Table 18: Volcano Urban models related to multiple paths components.

Parameters

Description


Recommended or default value

Refer to section

ANEAR (dB)

Correction of free space loss predicted at 1 meter for the receivers in the NEAR

region, ]-∞;+∞[

Tunable 0 5.4

BNEAR (dB/dec)

Correction of the distance dependence for the receivers in the NEAR region,

[0;+∞[ Tunable 24 5.4

Multi-path activation

Activation of the multi-path component: Yes/No

User-defined - 5.3.3

NR Maximum number of reflections, [0;4] User-defined 2 5.3.3

ND Maximum number of diffractions, [0;1] User-defined 1 5.3.3

Figure 16: Backward ray contributions resulting from a reflection (blue) or diffraction (green) in the transmitter LOS

vicinity


WNEAR Width of the near region, [0;+∞[ User-defined 800 meters 5.3.1

Street level mode

Street level mode, No/Flat/Hilly User-defined No 5.7.3

WR Weighting of reflection, [0;1] Tunable 1 5.3.3

WD Weighting of diffraction, [0;1] Tunable 1 5.3.3

WB Weighting of first backward ray

interaction, [0;1] Tunable 1 5.3.3

When the multiple-paths component is deactivated, all receivers are considered being located in the FAR region. The NEAR region does not exist.

5.4. Path loss estimate

The path loss predicted by the Volcano Urban model at a receiver point is calculated from only the direct-path component in the FAR region (or when multi-path component is not activated) and from the combination of all the direct and multi-path contributions in the NEAR region (when multi-path component is activated). See section 5.3.1 for details about these NEAR and FAR reception regions, as well as TRANSITION region.

A is the correction of free space loss at 1 meter (dB)

B is the correction of free space loss dependence on the distance logarithm (dB.dec-1)


α is the weight of the deterministic loss

LDET is the deterministic loss from only direct path or from direct path and multi-path combination (dB)

LFS is the free space loss at distance 1m (dB)

LANT is the antenna loss from only direct-path or from direct-path and multi-path combination, including antenna correction (dB)

LC is the clutter attenuation at the receiver location (dB)

The following expression is used to determine the received power at the receiver

location.𝑹𝒆𝒄𝒆𝒊𝒗𝒆𝒅 𝑷𝒐𝒘𝒆𝒓 𝒅𝑩𝒎 = 𝑬𝑰𝑹𝑷−𝑷𝑳 Eq. 5

The following expression is used to determine the received power at the receiver location


PL is the Path loss (dB) In the NEAR region, the path loss (or more precisely, the deterministic loss LDET) results from the scalar sum of the ray field strength squares, then not from a vectorial sum including ray field phases. This is a common way to get the averaged path loss in the close vicinity of the receiver location (i.e. around a surface one-wavelength wide), provided that the phases of the multiple field contributions may be assumed uncorrelated. Actually, in the case of urban radio predictions, where the accuracy for the location of the building walls is in the order of one meter (at most), the phase of the multi-path contributions cannot be estimated. Indeed at the 1GHz frequency, an uncertainty of only 30cm on the path length gives rise to an uncertainty of 2π on the phase. That is the reason why the vectorial sum of urban radio contributions

𝑷𝑳(𝒅𝑩) = 𝑨 + 𝑩. 𝐥𝐨𝐠 𝒅 + 𝜶. 𝑳𝑫𝑬𝑻 + 𝑳𝑭𝑺 + 𝑳𝑨𝑵𝑻 + 𝑳𝑪 Eq. 4


cannot provide real insight into the nominal power at the receiver point. But the scalar sum is perfectly appropriate to give average power in the spatial vicinity of the receiver. The parameter α weights the impact of the deterministic loss LDET. When α = 1, the computed deterministic loss is totally included in the path loss calculation. On the contrary, when α = 0, only free space loss, antenna loss and clutter attenuation are considered in the path loss calculation. Values in the range from 0.5 to 1 are typically returned by the tuning process.

5.5. Volcano Urban tuning

The purpose of the tuning is to adjust Volcano parameters to the radio channel characteristics and to reduce prediction errors. Tuning process is based on comparisons between measurements and Volcano predictions. The tuning determines the set of model parameters (including free space correction parameters, interaction weighting, deterministic loss weighting) that best fit radio channel conditions. Volcano Urban model has a unique auto-tune function that provides all adjusted parameters in one run, hence facilitating the tuning process and reducing tuning time. It provides a set of parameter values that gives a zero mean error and low RMS error. The tuning process provides different kinds of results:

The error statistics between the measurements and the predictions, after calibration

The error per measurement bin, after tuning

The adjusted parameter values of the calibrated model All these results are returned in a tuning report. The list of tunable model parameters is given in Table 17 and Table 18

To get further details about recommendations, please refer to the Volcano Tuning Guidelines [SIR3].

The model tuning is recommended in order to get robust operational propagation models. Tuning recommendations concern:

The distribution of the measurement bins around the emitting site

The accuracy on the emitting site location

The accuracy on the measurement locations

5.6. Advanced tuning for low resolution map data

Please refer to section 4.7 for the description of this feature.

5.7. Volcano Urban functionalities

5.7.1. Antenna Correction

This section aims at presenting the antenna correction that is implemented in the Volcano Urban model. Antenna correction consists in modifying the departure antenna gain G that is calculated from extrapolation of the manufacturer antenna diagrams (see section 7). This correction is optional in the Volcano Urban model. It is activated by default.


The need for this antenna correction was demonstrated from many empirical observations, and proved to be a very simple solution for taking into account the influence of the complex outdoor propagation environment on the observed antenna mask. SIRADEL recommends enabling antenna correction when using directional antennas, as this significantly improves the model accuracy. However the correction may be disabled, either when:

Predicting omni-directional antennas or antennas having a large horizontal beamwidth

Predicting antennas located far above the surrounding urban environment (typical for broadcast transmitters); in that configuration, the emitted signal experiences a smaller departure angular spread, then correction is not required

When the choice between enabling/disabling the antenna correction is not straightforward, we recommend testing both approaches, then selecting the one that gives the lowest prediction errors.

This feature is complementary to the antenna modeling detailed in section 0.

More information is available in the technical note Antenna Correction in Volcano Urban model.

5.7.2. Spatial averaging

The spatial-averaging reduces the small-scale variations of the predicted received-power, by averaging the usual received-power prediction. Prediction accuracy is improved, as the averaged results get advantage of site-specific calculations, whereas the small-scale variations are filtered out. Indeed those simulated variations are likely to be uncorrelated with actually observed in-the-field variations.

SIRADEL recommends enabling the spatial-averaging in the following situation:

Coverage resolution is between 5m and 20m

Reduction of the prediction error is an important concern

There is no need for prediction of the received-power small-scale fluctuations (i.e. fluctuations over intervals less than 20m wide)

More information is available in the technical note Spatial averaging in Volcano Urban model.

Figure 17: Illustration of coverage map obtained without spatial averaging (left) and with spatial averaging (right)


5.7.3. Street level mode

The optional street-level mode significantly reduces the time for computing the multi-path component. It restricts the constructed rays to the rays propagating only at the street level. No propagation above rooftops is simulated (except for direct-path). This optional mode has to be activated only when the radio terminals are located under the rooftop level, otherwise the coverage map will be erroneous. When appropriately used, the street-level mode permits to strongly reduce the computation times without any significant loss in accuracy. Three optional values can be associated to the street-level mode as described in Table 19.

Table 19: Street level option

No All types of contributions are calculated. The street level mode is deactivated

Flat

The street-level mode is enabled. Only the rays propagating at the street level are computed. The variations of the terrain along the ray path are not investigated.

This option is suitable only for street-level configurations in flat and vegetation-free areas. It is far less time consuming than the usual mode (street level mode = no).

Hilly

The street-level mode is enabled. Only the rays propagating at the street level are computed. The variations of the terrain along the ray path are investigated and taken

into account for the path loss computation.

This option is suitable for all street-level configurations, especially in hilly areas or in presence of vegetation. It is far less time consuming than the usual mode (street level

mode = no).

Keep in mind that when street level mode is activated, reflected and diffracted ray paths only propagate in the street (outside building contours). As a consequence, interactions cannot be simulated above podium buildings. Actually, when the interactions above podium buildings are likely to provide significant contributions, we recommend disabling the street-level mode.

5.7.4. Coverage prediction mode

Volcano Urban model proposes three different coverage prediction modes:

Regular mode (default one)

Boost mode (approximate of the regular coverage, obtained with a shorter computation time);

Along linears (prediction is realized only at receivers located along linear vectors)

5.7.4.1. Boost Mode

The Boost mode impacts on the coverage prediction is illustrated in Figure 18. As illustrated in the right picture (boost mode activated), the global aspect of the coverage is perfectly preserved. Only the prediction of local signal fluctuations is slightly degraded. The coverage in open areas is calculated as if the receiver was located in the middle of the street. The coverage in vegetation, built-up or building areas is obtained from adding the attenuation of the receiver land usage to the path loss calculated in the middle of the street. Remark that the Boost mode does not use the building linear attenuation, contrary to the Regular mode.


The Boost mode is not compatible with:

Multi-path component (this limitation will be overcome in future versions)

Spatial-averaging (refer to 5.7.2)

Receiver height defined above clutter We recommend using the Boost mode in the following situation:

Coverage maps to be computed are medium or large (i.e. more than 10,000 pixels)

Coverage resolution is lower or equal to 25m (e.g. 5m)

Times are a critical concern in the computation process

Small degradation of the received-power accuracy is acceptable

There is no need for realistic prediction of the received-power local spatial fluctuations (i.e. fluctuations over about 100m large intervals)

Generally, the user requirements match all these criteria when performing the radio network dimensioning.

More information is available in the technical note Boost Mode in Volcano Urban model.

5.7.4.2. Along linear mode

The “Along linears” mode aims at predicting only on a set of selected linear vectors, for specific applications (received signal in the streets only, on railways…). Prediction results are provided as a raster grid (as in regular and boost coverage modes), but pixels that do not cross selected linear vectors have undefined values. This option strongly reduces the computation time.

Figure 18: coverage map with radius 1km in dense urban area obtained from Regular mode (left) and Boost mode

(right)


Figure 19: Prediction along linears

5.7.5. Indoor penetration

The clutter exclusion feature lets the user exclude coverage prediction for receivers located inside specific clutter types. As illustrated in Figure 20, when building clutter is excluded, a default value (undefined) is returned for the in-building receiver locations (left picture). When the building clutter is not excluded, it provides an estimate of the in-building received power (right picture).

The indoor penetration is realized in two different ways depending on the receiver location and Volcano Urban parameters:

NEAR region, when multi-path is activated: rays resulting from outdoor interactions are prolonged inside the buildings and combined to compute the total received power

FAR region: the received-power is predicted from the direct-path prolonged inside the building, plus contributions coming from the streets surrounding the building

These algorithms provide compatible results in respectively NEAR and FAR regions. However only the prediction in the NEAR region results from multiple ray paths related to the real outdoor geometry. A smooth transition between both algorithms is done in the TRANSITION region to get a seamless indoor coverage. Please refer to section 5.1 for more information about different reception regions. Indoor penetration relies on two user-defined key parameters, which are associated to each building clutter:

Figure 20: Coverage prediction with indoor penetration (right) and without (left)


Clutter attenuation (dB): represents the attenuation when a ray penetrates inside the building (resulting from the exterior wall or rooftop)

Building linear attenuation (dB/m): represents the attenuation per meter when a ray propagates inside the building (resulting from internal partitions, furniture, etc.)

5.7.5.1. Indoor penetration in NEAR region (multi-path component is

activated)

The penetrating rays (associated to the multi-path component) are assumed to propagate inside the receiver floor along a straight trajectory from the penetration point to the receiver location, as illustrated in Figure 21. The direct-path contribution is processed in a quite different way. It can either be simulated as propagating inside the reception floor or propagating through different floors separating the penetration point (that may be on the rooftop) to the receiver location. In the latter case, a floor loss is calculated assuming that the loss due to the crossing of one 3-meter floor is equal to the clutter attenuation. Volcano Urban model selects the most favourable propagation mechanism.

5.7.5.2. Indoor penetration in FAR region (or when multi-path

component is not activated)

The basic principle is quite similar to the NEAR region, except that the multiple penetrating rays do not correspond to physical propagation paths, but are street-contribution rays coming from different building façades and whose field strength depends on the façade orientation and in-the-street field strength level.


The number of predicted pixels when indoor penetration is activated is generally about 50% greater than the number of predicted pixels without indoor penetration (i.e. 33% of pixels are indoors). Thus computation times when indoor penetration is activated are about 25% - 35% greater than computation times without indoor penetration. The indoor penetration algorithm does not apply for:

Figure 21: Penetration of multiple rays, from a rooftop transmitter to a ground-floor in-building receiver


Receivers located into built-up areas represented by clutters like “urban”, “dense urban”, “industrial”, etc; in these built-up reception areas, Volcano Urban model simply adds the clutter attenuation to the propagation path loss

Receivers located above a building clutter; in this situation, Volcano Urban model does not consider any clutter attenuation

More information is available in the technical note Indoor penetration in Volcano Urban model.

5.7.6. Indoor transmitter

The Volcano Urban model can simulate the propagation around indoor (i.e. in-building) transmitters. That feature is suitable for interference studies to take into account the impact of some in-building transmitters on the exterior network. The propagation from the in-building transmitter to the exterior building façade is computed in a statistical way, based on a indoor-to-outdoor loss. It is calculated from two user parameters that are already available, i.e. the clutter attenuation [dB] (i.e. interface loss) and the in-building linear attenuation [dB/m] while taking into account the length of the path inside the building. When the multi-path component is activated, the indoor-to-outdoor loss is calculated individually for each ray path. Remark that no interaction is constructed inside the building. All interactions take place in the outdoor environment, as illustrated in Figure 23.

Figure 22: Coverage predicted from an in-building transmitter. The transmitter is located 19 meters above the

ground inside a 22-meter high building. The antenna is omni-directional emitting 0 dBm at the frequency 900 MHz.

The coverage is computed at the street level

Volcano Urban model can automatically detect whether a transmitter is indoor or outdoor, by checking its coordinates against the map data. But it is up to the user to decide the strategy Volcano has to follow:

Really Indoor: Volcano will consider the transmitter as really indoor and therefore apply a loss due to the indoor-to-outdoor interface

Actually outdoor: Volcano will consider the transmitter as actually outdoor (error in the transmitter location or map data inaccuracies) and therefore will not apply a loss due to the


indoor-to-outdoor interface. Clearance is applied around the transmitter to remove any close building obstacle before computing the propagation loss

In the case where the direct path from the transmitter to the receiver is entirely inside the same building, the in-building loss includes the loss due to the crossing of separating floors. It is assumed that the distance between two successive floors is about 3 meters. It is assumed also that the loss due to the crossing of one floor is similar to the clutter attenuation (i.e. interface loss). When the vertical distance is less than 2 meters, the transmitter and the receiver are assumed to be in the same floor, the floor loss is null.

5.7.7. Antenna Masking/Unmasking

When using multiple path contributions, it is possible to take advantage of the antenna masking/unmasking feature. Indeed, for co-located sectors (margin of 1m around the site), the whole calculation is performed for the first sector only. For the following sectors, the prediction is unmasked (removing impact of first antenna and masked with the new antenna impact). This option enables to dramatically decrease the computation time. Indeed, for 3 co-located sectors (typical configuration), the computation time is divided by 2.

5.7.8. Wideband additional outputs

The Volcano Urban model lets the user the possibility to get further information regarding the predictions they performed. These additional outputs are suitable for advanced analysis. Indeed, the radio channel impulse response can be estimated from the multiple predicted ray paths.

The impulse response includes information about the time delay i , the field strength Pi, the angle of

departure i and the angle of arrival i of every path of index i. Some other wideband parameters can be deduced from the estimated channel response that may be helpful in the analysis of the radio systems performance: delay spread, spread of arrival angles and spread of departure angles. Additional outputs can be divided into two main families:

Single value: a single value is returned for each receiver

Array of values: an array of values is returned for each receiver (also called spectrum outputs) The full list of additional outputs is given in Table 20 (single-value additional outputs) and Table 21 (array-of-values additional outputs).

Figure 23: Multi-path contributions from an in-building transmitter


Table 20: Single-value additional outputs

Output Description Unit Range value

Mean delay Average of the ray path delays ns [0 ; +∞[

Delay spread Standard deviation of the ray path delays ns [0 ; +∞[

Azimuth mean arrival angle Azimuth of the mean arrival direction (North:

0°, West: -90°) deg [0 ; 360[

Azimuth arrival angle spread Standard deviation of the ray path arrival

azimuths deg [0 ; 180[

Azimuth mean departure angle

Azimuth of the mean departure direction (North: 0°, West: -90°)

deg [0 ; 360[

Azimuth departure angle spread

Standard deviation of the ray path departure azimuths

deg [0 ; 180[

Elevation mean arrival angle Elevation of the mean arrival direction (Top:

-90°, Bottom: 90°) deg [–90 ; 90[

Elevation arrival angle spread Standard deviation of the ray path arrival

elevations deg [0 ; 90[

Elevation mean departure angle

Elevation of the mean emitting direction (Top: -90°, Bottom: 90°)

deg [-90 ; 90[

Elevation departure angle spread

Standard deviation of the ray path departure elevations

deg [0 ; 90[

Ray number Number of predicted paths (incl. the direct-

path) - [0; +∞[

Table 21: Array-of-values additional outputs

Output Description Unit Range value

Received power spectrum Ray power levels at the receiver end dBm ]-∞ ; +∞[

Delay spectrum Ray path delays between the transmitter and

the receiver ns [0 ; +∞[

Azimuth departure angle spectrum

Ray horizontal emitting angles at the receiver (North: 0°, West: -90°)

deg [0 ; 360[

Elevation departure angle spectrum

Ray vertical emitting angles at the receiver (horizontal: 0°, up: +90°, down: -90°)

deg [0 ; 360[

Azimuth arrival angle spectrum

Ray horizontal arrival angles at the receiver (North: 0°, West: -90°)

deg [0 ; 360[


Elevation arrival angle spectrum

Ray vertical arrival angles at the receiver (horizontal: 0°, up: +90°, down: -90°)

deg [0 ; 360[


6. Volcano Indoor Model

6.1. Description

The Volcano Indoor Model is a proprietary model based on the COST 231 multi-wall model that calculates the path loss between in-building transmitter and receiver terminals. An accurate description of the indoor environment permits to include losses from building partitions (internal walls, doors, external walls, windows, etc) placed on the direct path between both terminals. Besides, the floor loss (added when terminals are not located in the same floor) is a non-linear function of the number of floors that separate both terminals, as generally observed from measurements.

6.2. Model Overview

The Volcano Indoor Model is a proprietary model based on the COST 231 multiwall model for indoor-to-indoor radio links. The general equation for the Volcano Rural Model is given hereafter.

𝑷𝑳(𝒅𝑩) = 𝑨 + 𝑩. 𝐥𝐨𝐠 𝒅 + 𝑳𝑻𝑹𝑨𝑵𝑺 + 𝑳𝑭𝑺 + 𝑳𝑨𝑵𝑻 + 𝑳𝑭.𝒏𝑭.𝒆𝒏𝑭+𝟐𝒏𝑭+𝟏

−𝒄

Eq. 6


A is the additional loss (dB)

B is the additional loss per distance decade (dB.dec-1)

LTRANS is the transmission loss through the building partitions found along the direct-path (dB)

LFS is the propagation loss in free-space at distance 1m (dB)

LANT is the antenna loss in direct-path direction (dB)

LF is the transmission loss through a floor

nF is number of floors crossed by the direct path

c is the correction parameter for transmission through multi-floors The following expression is used to determine the received power at the receiver location

𝑹𝒆𝒄𝒆𝒊𝒗𝒆𝒅 𝑷𝒐𝒘𝒆𝒓 𝒅𝑩𝒎 = 𝑬𝑰𝑹𝑷−𝑷𝑳 Eq. 7


PL is the Path Loss (dB)

6.3. Supported geographical data

The Volcano Indoor Model requires an accurate description of the indoor environment. This environment must described a DBM (Digital Building Model) file that is volcano specific XML file format including description of building layers (similar to menu file in outdoor geo map data), a list of DXF file references (similar to index file in outdoor geo map data) and DXF files that represent the building geometry.


More information about DBM is available in the technical note Creation of a Digital Building Model [SIR1].

Figure 24: Digital building model (DBM)


7. Antenna modelling

7.1. Description

Volcano models provide “masked” predictions, i.e. the antenna masking at the transmitter and receiver ends is included in the calculation. The antenna gain is given from a 3D extrapolation based on the horizontal diagram and the vertical diagram provided by the antenna manufacturer. Among others, one advantage of this 2D to 3D extrapolation is to take into account the dissymmetry occurring between the forward and backward antenna patterns.

7.2. Concept

The antenna radiation pattern for one particular linear polarization is described by means of two bi-dimensional and orthogonal diagrams. On the one hand, the horizontal diagram is the cut of the 3D pattern in the horizontal plane at elevation 0°. As illustrated in Figure 25 (left), the gain is given as a function of the horizontal angle or azimuth, which is in the range from –180° to +180°. On the other hand, the vertical diagram is the cut of the 3D pattern in the vertical plane at azimuth 0°. As illustrated in Figure 25 (right), the gain is given as a function of the vertical angle or elevation which is in the range from –180° to +180°.

Figure 25: Antenna diagrams examples.

The 3D extrapolation implemented in Volcano models to compute the antenna gain is a proprietary technique based on the following relation:

The gain G is a function of:

The maximum gain Gmax of the antenna, in dBi

The elevation angle θ in the range from –90° to +90°, which is the angle between the propagation direction and its projection in the horizontal plane

The azimuth angle φ in the range from –180° to +180°, which is the angle between the antenna axis and the projection of the propagation direction in the horizontal plane

The gain Gv (θ ) given by the forward part of the vertical diagram, in dB

The gain Gv(180°-θ) given by the backward part of the vertical diagram, in dB

The gain Gh (φ ) given by the horizontal diagram, in dB

Antenna diagram in the horizontal plane given as a

function of the azimuth angle. Gain in dB is relative

to the maximum antenna gain

Antenna diagram in the vertical plane given as a

function of the elevation angle. Gain in dB is

relative to the maximum antenna gain

𝑮 = 𝒇(𝑮𝒎𝒂𝒙,𝑮𝒗 𝛉 ,𝑮𝒗 𝟏𝟖𝟎° − 𝛉 ,𝑮𝒉 𝝋 ) Eq. 8


One main advantage of the 3D extrapolation implemented in Volcano models is that results are consistent with the description of the backward vertical diagram, as illustrated in Figure 26. This kind of consistency is not obtained from typical extrapolation methods only based on the description of the forward vertical diagram (i.e. diagram described in the elevation angles range [-90°; +90°]).

Basic method Volcano method

Figure 26: Extrapolated 3D antenna diagrams


8. References

[1] Saunders R. Simon, Antennas and Propagation for wireless communication systems. John Wiley & Sons Ltd, 1999. [2] Y. Corre and Y. Lostanlen, “Three-dimensional urban EM wave propagation model for radio network planning and optimization over large areas”, IEEE Trans. on Vehicular Technology, vol. 58, pp. 3112-3123, Sept. 2009. [3] R. Luebbers, “A heuristic UTD slope diffraction coefficient for rough lossy wedges”, IEEE Trans. on Antennas and Propagation, vol. 37, pp. 206-211, Feb. 1989. [4] COST Action 231, Digital mobile radio towards future generation systems, COST 231 Final report, 1989. [SIR1] Creation of a Digital Building Model: description of the indoor geographical data (representation of walls and partitions, format, coordinate system, etc.) needed for the Volcano Indoor Model [SIR2] Map Data Requirements for Volcano: description of the outdoor geographical map data (type of data, format, resolution, etc.) needed for the Volcano Rural and Urban Models [SIR3] Volcano Model Tuning Guidelines: description of the steps and instructions to perform Volcano model tuning


9. Technical Support

For technical support when you are using Volcano, please contact the SIRADEL support team:

E-mail: [email protected]

Telephone: +33 (0) 223 480 500

Fax: +33 (0) 223 480 599

mailto:[email protected]

Volcano_Reference_Guide 3.0.pdf

Documents