Professional Documents
Culture Documents
User Manual
Version 2.8.0
AS280_UM_E0
Aster 2.8.0 User Manual
Release: AS280_UM_E0 (October 2019)
© Copyright 1997-2019 Forsk. All Rights Reserved.
Published by:
Forsk
7 rue des Briquetiers
31700 Blagnac, France
Tel: +33 562 747 210
Fax: +33 562 747 211
The software described in this document is provided under a license agreement. The software may only be used or copied under the terms and
conditions of the license agreement. No part of the contents of this document may be reproduced or transmitted in any form or by any means
without written permission from the publisher.
The software shall be used by a human user in interactive mode, or in autonomous batch mode via scripts as described in the software
documentation. Licensee shall not drive or activate the software through a robot or through an external application that is not part of the
software.
The product or brand names mentioned in this document are trademarks or registered trademarks of their respective registering parties.
Third party services that are not part of Aster are governed by the terms and conditions of their respective providers, which are subject to change
without notice.
The publisher has taken care in the preparation of this document, but makes no expressed or implied warranty of any kind and assumes no
responsibility for errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of the
use of the information contained herein.
Aster 2.8.0 User Manual
AS280_UM_E0 Table of Contents
Table of Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
About Aster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
About Atoll . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
About Forsk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Getting Help . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Printing Help Topics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
About Aster Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Contacting Technical Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Introduction
Aster is a ray-tracing propagation model based on a new concept of implementation for ray tracing and is
particularly designed for dense urban environments with high resolution geo raster data.
The purpose of this User Manual is to provide general information about using the Aster propagation model with
your Atoll documents. It contains the following chapters:
◼ "Installation and Configuration" on page 9
◼ "Aster Model Overview" on page 17
◼ "Using the Aster Model" on page 23
◼ "Calibrating the Aster Model" on page 39
In addition to this manual, the Aster documentation package also includes the following items:
◼ Aster Technical Reference Guide
You can also download manuals from the Forsk web site at:
http://www.forsk.com/MyForskAccount/
About Aster
Aster is a fast, accurate, and easy to use radio propagation model and is fully integrated with Atoll and Naos. It is
provided with pre-calibrated parameters, can auto-calibrate, and supports all wireless technologies.
Aster can be used with Atoll version 3.2.1 and later, as well as the Atoll distributed calculation server.
Aster can also be partially used with Atoll 3.1.2. However, only matrices and profiles
can be calculated with the Aster propagation model. For information, contact Forsk
Support.
Starting with 2.7.0, Aster can be used with Naos version 1.0.0 and later.
The Aster propagation model has the following features:
◼ Ray tracing: Aster is based on two major components:
◼ Vertical diffraction over rooftops based on Walfisch-Ikegami model and multiple knife-edge Deygout
method.
◼ Horizontal diffraction based on ray tracing.
Aster can use geographical data such as vectors for ray tracing, but it can also perform ray tracing with raster
data only.
◼ mmWave support: Aster supports extremely high frequencies also known as millimetre wave or mmWave,
which makes Aster compliant with all wireless technologies and frequencies from 150 MHz up to 60 GHz,
including 5G networks.
◼ Extremely fast: Aster uses a unique high-speed ray-tracing technique, based on the raster sampling of building
angles. For example, 5 seconds are enough to calculate a 1,500 m radius cell with a 5m grid on an ordinary
laptop.
◼ Highly accurate: Aster can take all the main radio propagation effects into account, leading to highly accurate
coverage prediction results. Its accuracy applies to antennas above rooftops (where signal levels are mainly
due to vertical diffractions) as well as antennas under rooftops (where signal levels are mainly due to
horizontal diffractions).
◼ Ready & Easy-to-use: Aster is fully integrated in the Aster environment and there is no need for a special
database or for any type of data pre-processing.
◼ Compatible with high resolution (less than 25 m) raster data and with all types of clutter data. It is also
compatible with vector data (ESRI Shapefiles SHP and MapInfo TAB formats are currently supported).
◼ Supplied with pre-calibrated parameters using more than 1.5 million measurement points. The standard
deviation from measurements is typically less than 6.5 dB. Model configurations are intuitive and easy to
access.
◼ Auto-calibration: Aster supports measurement-based auto-calibration. The standard deviation can drop to
less than 6.5 dB in scenarios and environments with high-resolution geo data and good-quality
measurements.
About Atoll
Atoll is a multi-technology wireless network design and optimisation platform that supports wireless operators
throughout the network lifecycle, from initial design to densification and optimisation. Atoll offers unique
capabilities of using both predictions and live network data throughout the network planning and optimisation
process.
Atoll includes integrated single RAN–multiple RAT network design capabilities for both 3GPP and 3GPP2 radio
access technologies including 5G NR, LTE, NB-IoT, UMTS, GSM, and CDMA. It provides operators and vendors with
a powerful framework for designing and optimising current and future integrated multi-technology networks.
Atoll supports the latest technology advances such as massive MIMO, 3D beamforming, and mmWave propagation
for the design and roll-out of 5G networks.
Atoll’s integration and customisation features help operators smoothly streamline planning and optimisation
processes. Aster supports a wide range of implementation scenarios, from standalone to enterprise-wide server-
based configurations. Atoll has become the industry standard for radio network planning and optimisation.
If you are interested in learning more about Atoll, please contact your Forsk representative to inquire about our
training solutions.
About Forsk
Forsk is an independent software company providing operators and vendors with wireless network design and
optimisation products. Atoll, Forsk’s flagship product, is the market-leading wireless network planning and
optimisation software on the market; it allows operators to streamline planning and optimisation activities by
combining predictions and live network data.
With more than 9000 active licenses installed with 500+ customers in 140 countries, Atoll has become the industry
standard for wireless network design and optimisation.
Forsk distributes and supports Atoll directly from offices and technical support centres in France, USA, and China
as well as through a worldwide network of distributors and partners.
Getting Help
The online help system that is installed with Aster is designed to give you quick access to the information you need
to use the product effectively. It contains the same material as the Aster 2.8.0 User Manual.
You can browse the online help from the Contents view, the Index view, or you can use the built-in Search feature.
You can also download manuals from the Forsk web site at:
http://downloads.forsk.com
You can print individual topics or chapters from the online help.
To print help topics or chapters:
1. In Aster, click Help > Help Topics.
2. In the Contents tab, expand the table of contents.
3. Right-click the section or topic that you want to print and click Print. The Print Topics dialog box appears.
4. In the Print Topics dialog box, select what you want to print:
◼ If you want to print a single topic, select Print the selected topic.
◼ If you want to print an entire section, including all topics and sections in that section, select Print the
selected heading and all subtopics.
5. Click OK.
The following PDF manuals are available to customers with a valid maintenance contract for Aster and Aster
Microwave and can be downloaded from the Forsk web site at:
http://downloads.forsk.com/
To read PDF manuals, download Adobe Reader from the Adobe web site at:
http://get.adobe.com/reader/
Hardcopy manuals are also available. For more information, contact to your Forsk representative.
Forsk provides global technical support for its products and services. To contact the Forsk support team, visit the
Forsk web site at:
http://downloads.forsk.com
Alternatively, depending on your geographic location, contact one of the following support teams:
Forsk US
For North and Central America, contact the Forsk US support team:
◼ Tel.: 1-888-GO-ATOLL (1-888-462-8655)
◼ Fax: 1-312-674-4822
◼ Email: support_us@forsk.com
Opening Hours: Monday to Friday 8.00 am to 8.00 pm (Eastern Standard Time)
Forsk China
For Asia (except Japan), contact the Forsk China support team:
◼ Tel: +86 20 8557 0016
◼ Fax: +86 20 8553 8285
◼ Email: atollsupport@forsk.com.cn
Opening Hours: Monday to Friday 9.00am to 5.30pm (GMT+08:00) Beijing, Chongqing, Hong Kong, Urumqi.
◼ Email: support@forsk.com
Opening Hours: Monday to Friday 9.00 am to 6.00 pm (GMT +1:00)
If only one token is available and the number of threads to be used per calculation
is defined to be larger than 4, then the token will be used by the first 4 threads and
the remaining threads will use that token afterwards.
Aster can be used with fixed and floating licenses for Aster. However only floating licenses for Aster are currently
supported:
Floating Net HASP for current Aster customers (1) Served by SafeNet Sentinel LDK Run-time (2)
floating protection key, even if more than one Aster session is run on the same computer. The token is released and
made available for other users when all the Aster sessions on the computer are closed.
Depending on the number of Aster licenses you purchased, you may find it useful to be able to authorize specific
users to use an Aster license and block others from using one. To that end, the following lines must added in the
Atoll.ini file:
[License]
ASTER = 0
The way the Sentinel LDK Run-time is installed depends on how Aster, Aster, and the Aster protection key are
deployed:
Aster protection key on computer, and - Sentinel LDK Run-time None (LDK Run-time is
Aster and Aster on same computer installed automatically by
Aster setup)
Aster protection key on license server con- With Windows update Sentinel LDK Run-time None (LDK Run-time is
nected to the internet, and installed automatically by
Aster and Aster on other computers Windows Update)
http://<server>:1947
where <server> is the name or IP address of the individual computer or license server you want to monitor.
To manage Aster licenses from the Sentinel ACC, open the Sentinel ACC as shown in "Opening the Sentinel ACC" on
page 11.
Only some of the columns displayed in Sentinel ACC pages are described in this
section. Click Help in the left-hand pane of Sentinel ACC for information on the
other columns.
◼ Display denotes that access is permitted from remote clients, via the display on a terminal server.
◼ Counting: indicates the basis on which logins to the feature are counted:
◼ Station denotes that all login requests for a single machine count as one login.
◼ Login denotes that each login request is counted.
◼ Logins: indicates the number of users currently logged in to the feature on the protection key.
◼ Limit: indicates the maximum number of users that may concurrently log into the feature on the protection
key.
To display custom vendor, product, and features’ names in the Sentinel ACC:
1. Log in with administrator privileges to the machine running the Sentinel ACC.
2. Stop the Sentinel LDK Licenses Manager service.
3. Create the 62733.xml file with the following content:
errorlog = 1
requestlog = 1
broadcastsearch = 0
serveraddr = IPADDR_OR_HOSTNAME
Where IPADDR_OR_HOSTNAME should be replaced by the IP address or the hostname of a computer running
an instance of Sentinel LDK License Manager service.
4. Copy the hasp_62733.ini file into the Atoll installation directory.
Alternatively, if you want the hasp_62733.ini to be specific to the current user, you
can copy the file into %LocalAppData%\SafeNet Sentinel\Sentinel LDK. This
location takes precedence over the Atoll installation directory.
Where:
◼ <vendor id> corresponds to <VendorID>
◼ <hasp id> corresponds to <KeyID>
◼ Statistical propagation classes : Aster uses statistical propagation when the ASODM description is rough and
not well geo-referenced. They are determined by correspondence with the clutter class layer or from a vector
file.
For a list of deterministic and statistical propagation classes in Aster, see "Propagation Classes" on page 27.
For example, if the clutter file has a 25m resolution and defines "Residential" or "DenseUrban" clutter classes , you
should use the Sparse Buildup or Dense Buildup statistical propagation class. For parks and mixtures of trees and
open areas, you should use the Mixed Vegetation statistical propagation class. For areas that are mainly open, you
can still use the Open and Water deterministic propagation classes.
These classes should be used when the clutter class description is rough, for example with low resolution clutter
that cannot accurately describe the terrain.
2.3.1 ASODM Model with both Clutter Heights and Clutter Classes
In this case, the Clutter Heights folder in the Geo explorer contains a clutter heights layer and the Clutter Classes
folder contains a clutter class layer. This is the recommended setup and the best accuracy in propagation modeling
is reached when the maximum pixel size in either layer is 5m or less. When both layer types are available, the
following scenarios are be considered:
◼ "Clutter Heights with Higher or Same Resolution as Clutter Classes" on page 18
◼ "Clutter Classes with Higher Resolution than Clutter Heights" on page 19
◼ "Multiple Clutter Class Layers with Different Resolutions" on page 19
◼ "Partial Clutter Heights Information" on page 19
However, in order to prevent mismatches when the clutter class layer and the clutter heights layer have
different resolutions:
◼ If that propagation class is "Open" or "Water", the ASODM type is set to the corresponding deterministic
propagation class defined under Clutter heights on the Geo tab of Aster properties.m
◼ If that propagation class is one of the statistical propagation classes and the clutter heights file has a high
resolution (less than or equal to 10m), the ASODM type is reset to the corresponding deterministic
propagation class defined under Clutter heights on the Geo tab of Aster properties.
A proprietary algorithm insures that the transition zones are handled properly from
areas where clutter height and clutter class information is available to areas where
only clutter class information is available.
When the Clutter Classes folder in the Geo explorer contains a clutter class layer and the Clutter Heights folder is
empty, Aster will uses the clutter class layer to model the ASODM types as well as the ASODM heights.
For best ASODM modeling accuracy in urban areas, the lowest clutter class resolution (defined by Pixel Size on the
Geocoding tab of the clutter class file properties) should be:
◼ 5m when no vector layer is available
◼ 25m when vector layers are available
The classes in the clutter class layer must be mapped to deterministic or statistical propagation classes. The choice
depends on the resolution of the clutter class layer:
◼ High resolution clutter class layer: the classes should be mapped to deterministic propagation classes
◼ Low resolution clutter class layer: the clutter classes should be mapped to statistical propagation classes
The ASODM heights appear under Default Clutter Height (m) on the Clutter tab of Aster properties:
◼ By default, ASODM heights are inherited from the Description tab of the clutter class file properties and are
read-only.
◼ You can remove the read-only attribute by setting isHeightReadOnly to 0 in the [Clutter] section of the Aster.ini
or AsterMmWave.ini file.
+ Another advantage of using the isHeightReadOnly option is that you can define
several Aster instances where a given clutter class will have a different default
height from one instance to another.
In the following example, 5 clutter classes are defined in Aster with a high resolution clutter class layer. Since the
clutter class layer has a high resolution, Aster maps each clutter class to one of the 8 deterministic propagation
classes described in "Propagation Classes" on page 27:
When the Clutter Heights folder in the Geo explorer contains a clutter heights layer and the Clutter Classes folder
is empty, the ASODM heights are retrieved from the clutter heights layer only. This usage of Aster is not
recommended because clutter classes provide a much better model of the various ASODM types found in the field.
For best ASODM modeling accuracy in urban areas, the lowest resolution of the clutter heights layer should be 5m
(defined by Pixel Size on the Geocoding tab of the clutter heights file properties).
When no clutter classes are available for the definition of propagation classes, the following applies:
◼ For the pixels where the clutter height is 0, Aster will use a deterministic propagation class on these pixels:
◼ The deterministic propagation class defined next to "No Data", or
◼ Open if the class defined next to "No Data" is a statistical class ("Mixed Vegetation", "Sparse Buildup", or
"Dense Buildup").
◼ For the pixels where the clutter height is strictly positive:
◼ When the clutter heights resolution is low (i.e. pixel size greater than 10m), Aster uses a statistical
propagation class on these pixels:
Both ESRI Shapefiles (SHP) and MapInfo (TAB) vector formats are supported.
The corresponding data files are the dBF file and the DAT file, respectively.
The TAB format provides better performance with Aster.
The ASODM model is first initialised from the geo raster layers as shown in the previous sections. Vector layers are
not mandatory to build the ASODM model but they improve its accuracy.
If vector data is available, then it is used to refine the accuracy of the model. Vector data takes priority over raster
data when defining the propagation class of a pixel in the polygon. A proprietary algorithm insures that the transition
zones are handled correctly from areas where vector layers are available to areas where only raster layers are
available.
Vector usage is enabled when the Use vectors check box is selected on the Vectors tab of Aster properties. You can
define one or more vector files and configure the properties of each vector file.
For each polygon defined by a vector layer, an "above-surface object" is built with height and type attributes, which
are found in the corresponding data file:
◼ Height is directly recovered from the data file. The height is considered as above ground level by default and,
if needed, it can also be defined as above sea level by selecting the Height defined above mean sea level
(AMSL) check box in the corresponding Polygon Definition dialog box.
◼ Type is obtained from a mapping between the data file attributes and the propagation classes.
A specific propagation class is assigned to each pixel falling within a polygon of the vector layer and the pixel is
given the height defined by the polygon.
Propagation classes that are applied to pixels that are not in a polygon of the vector layer are either:
◼ The propagation classes set by the raster layers when these pixels are located outside the vector support area.
In this case, the vector support area is automatically detected.
◼ Reset to the "Open" deterministic propagation class when the pixels are located inside the vector support area,
because vectors are assumed to have a properly defined support area without any holes. This propagation
class takes priority over the clutter-level propagation class.
When a vector layer is available, only the pixels located inside the zone defined by the vector layer can be reset to
Open. All other pixels keep the propagation class set by the clutter classes raster layer.
You can specify main matrix and extended matrix calculation parameters for each individual transmitter. These
parameters include the propagation model, resolution, and calculation radius. The propagation model assigned to
an individual transmitter always has precedence over any other assigned propagation model.
Aster enables you to calculate two levels of path loss matrices:
◼ A main matrix, which covers a smaller radius with a higher resolution
◼ An extended matrix, which covers a wider radius with a lower resolution
The extended matrix allows you to reduce calculation time and storage size by calculating lower resolution matrices
in areas that are far from the transmitter. Aster supports extended matrix configurations and can perform smooth
dual-grid calculations. However, a slight level of discontinuity can appear between two matrices when the difference
between radius settings and/or the difference between the resolution settings is too big
To specify the main and extended propagation parameters for a transmitter:
1. In the Network explorer, expand the Transmitters folder, right-click the transmitter to which you want to assign
a main and extended propagation model, and select Properties from the context menu. The Properties dialog
box appears.
You can also access the Properties dialog box by right-clicking the transmitter in the map window and
selecting Properties from the context menu.
2. To use Aster as the main matrix propagation model, in the Propagation tab, under Main matrix:
◼ Propagation model: Select "Aster Propagation Model"
◼ Radius: Enter an effective radius for which the signal level will be significant. If the radius it is too big, the
accuracy will not be degraded but the calculation time and the memory usage will be significantly
impacted.
◼ Resolution: Enter a pixel size that will account properly for the terrain features. Knowing that Aster is
designed for all types of environments, a minimum resolution of 5m is particularly recommended in urban
environments. If it is lower (above 5m), the resulting ASODM model will not be accurate enough to translate
urban characteristics such as buildings and streets.
Some of Aster’s internal features are automatically disabled when the resolution too low. For example, ray
tracing is disabled if the path loss matrix resolution value is larger than 10m.
Aster is also able to work on non-urban environments for which geo raster data is more statistical than
deterministic, typically at a low resolution. It provides a set of statistical propagation classes that can be
used for such environments and resolutions as low as 50m can be used. When no vector layers are
available, you can use the same resolution as the one defined for the raster layers.
3. If necessary, specify the Propagation model, Radius and Resolution for the Extended matrix.
The radius and resolution of the extended matrix must be larger than the radius and resolution of the main
matrix.
4. Click OK. The selected calculation parameters will be used for the selected transmitter.
When transmitters and/or cells have been configured to use the Aster propagation model, all of Aster’s calculation
features (such as calculating path losses, calculating predictions, and performing point-to-point analysis) remain
available.
Analysis and automatic calibration based on measurement files can be run directly from the context menu of the
Aster Propagation Model folder in the Parameters explorer of Aster. For more information about calibrating
propagation models, see the Model Calibration Guide.
For more information about assigning propagation models, see the chapter Radio Calculations and Models of the
Atoll User Manual.
Penetration loss and indoor loss, used for in-building interpolation in the
deterministic "Building" propagation classes 4 to 6, are defined by the Indoor
Attenuation set of parameters. They are also used for propagation through
obstacles in all classes except "Open" and "Water". You can also define different
losses for in-building interpolation and propagation in the Aster.ini or
AsterMmWave.ini file.
For propagation through obstacles in all classes except "Open" and "Water", the
LoS Attenuation set of parameters is used.
◼ Indoor calculations only: When this option is set to "Yes", Aster performs indoor-only calculation above
minimal receiver heights. This is useful in 3D coverage prediction calculations performed by Aster’s Multi-
storey add-in as it performs successive 2D layer calculations at different heights.
In the following example, the signals will be computed indoor and outdoor at ground level (receiver height
< 2m) and only indoor in storeys above 2m:
◼ Indoor antennas:
◼ Enable outdoor relocation: Select this check box to relocate indoor antennas to outdoor. Antennas are
often interpreted as being indoors because of error and mismatches between geo information and
transmitter data. You can either correct the geo data or automatically relocate indoor antennas to an
outdoor location, such as on a building facade or a rooftop.
Strategy: Click this button to open the Indoor Antenna Relocation Strategy dialog box. In this dialog box,
you can modify the automatic relocation algorithm. For more information, see "Configuring Antenna
Relocation" on page 36.
◼ Indoor calculation within antenna building: Select how Aster calculates path loss inside the building where
the indoor antenna is located:
◼ At antenna height: No interpolation is made and the signals will be computed from indoor losses per
metre, assuming that the receiver is at the same height as the antenna.
◼ At receiver height: No interpolation is made and the signals will be computed from indoor losses per
metre, assuming that the receiver is at the defined height.
◼ Receiver on top of building: Select this option to consider that the receiver is placed on top of a building
instead of inside. If this option is selected, building penetration is ignored and the default receiver height is no
longer used.
◼ LOS calculation only: Select this option to restrict Aster calculations to line-of-sight vertical attenuation only.
When this option is enabled, propagation is mostly blocked by buildings, vegetation, terrain. A potential
penetration loss is calculated within certain limits. For more information on how line-of-sight propagation is
calculated, see the Aster Technical Reference Guide.
The bottom line on the Settings tab contains the current version number of Aster.
The configuration parameters correspond to a series predefined model types that are available with Aster.
For the standard Aster propagation model, the following configurations are predefined:
◼ Standard: A general purpose configuration that works in all environments. Ray tracing is enabled.
◼ Macro: A configuration optimised for Macro cells. Ray tracing is disabled.
◼ Micro: A configuration optimised for Micro cells. Ray tracing is enabled.
◼ Rural: A configuration optimised for rural environments. Ray tracing is disabled.
For the mmWave Aster propagation model, the following configurations are predefined:
◼ Default_Model: A general purpose configuration that works in all environments with extremely high
frequencies. Ray tracing is disabled by default but can be enabled.
◼ LOS_Model: A configuration optimised for line-of-sight communication with extremely high frequencies. Ray
tracing is disabled.
◼ Micro_28Ghz: A general purpose configuration optimised for micro or small cells in urban or suburban
environments. Ray tracing is enabled.
◼ Macro_28Ghz: A configuration optimised for macro cells in urban or suburban environments. Ray tracing is
disabled.
If necessary, you can edit and add predefined configurations by editing the [Configuration] section of the Aster.ini
or AsterMmWave.ini file. For more information, see the Aster Technical Reference Guide.
You can also save and load Aster configurations in the Configuration window.
The Calibration menu can be disabled in the Aster.ini or AsterMmWave.ini file. For
more information, see the Aster Technical Reference Guide.
For Aster to work properly, terrain categories must be mapped to the appropriate Aster deterministic or statistical
propagation classes. This mapping can be done on the Propagation Classes tab of the Aster Propagation Model
Properties window.
Open All open areas, streets, squares, flat parks, etc. Open space, radio signals propagate freely.
Water Sea, rivers, lakes, etc.
Elevated Open Elevated areas. Filled obstruction area for radio signals.
Reception done on top.
Building type 1 Three types of buildings are defined to model dif- Deterministic buildings that create an obstruction
Building type 2 ferent propagation effects. on which radio signals reflect and diffract. Radio
Building type 3 signal levels are not calculated in a direct way but
interpolated from signal levels in surrounding open
areas.
You can configure up to 32 propagation classes of each type based on different propagation effects. For example,
you can create Building classes to model multiple types of buildings based on construction materials such as
concrete, wood, mostly glass covered, and so on.
+ Initially, it is recommended to use a single type for each propagation class. You can
improve the model later by introducing different types for more precise modeling.
Bridge and Elevated Open classes allow you to model objects for which reception is expected on top. The difference
between the two is that Bridge allows the radio waves to propagate underneath while Elevated Open creates a full
obstruction and is more suited for elevated areas. In all other propagation classes, reception is expected near
ground level.
The Trees class is used when vegetation is defined with accuracy. Otherwise, the Mixed Vegetation statistical class
is used. To use the Trees class:
◼ Specify a vegetation loss model: Linear, ITU-R, Cost 235 (in leaf or out of leaf), or FITU-R (in leaf or out of leaf).
◼ Click the button ( ) in the "..." column to specify the Under foliage ratio. When the value has been changed
from the default, the button in the "..." column turns green ( ).
The Building classes enable you specify the outdoor to indoor and indoor to outdoor attenuation:
◼ Specify the Penetration loss (dB) that is the applied to the radio signals that go through the external walls
of the building.
◼ Specify the Linear loss (dB/m) that is applied to signals as they propagate throughout the building.
◼ Click the button ( ) in the "..." column to specify the Roof penetration loss, which is applied to radio
signals that penetrate the roof of the building instead of the walls. When the value has been changed from
the default, the button in the "..." column turns orange ( )
Mixed Vegetation Parks or areas with a mixture of trees and open Radio signals can go through with some loss
areas. and diffract vertically.
Sparse Buildup Sparse building areas with a mixture of build- The clearance area is considered at the receiver
location and reception is done at ground level.
ings and open areas.
Mixed Vegetation, Sparse Buildup, and Dense Buildup classes are characterised by an assumed clearance zone
near the reception point or near the transmitting point.
When a low resolution umbrella clutter class layer is used, the propagation classes can be any of the statistical
classes or either the "Open" or "Water" deterministic classes for the clutter classes which are mostly open areas.
For example:
Grass Open
When a high resolution clutter class layer is used, choose the corresponding propagation classes among the
deterministic classes. For example:
Park Open
Mixed Vegetation
Vegetation Trees
Roads Bridge
+ You can modify the default height if you remove the read-only attribute by setting
the isHeightReadOnly option to 0 in the [Clutter] section of the in Aster.ini or
AsterMmWave.ini file. Another advantage of setting the isHeightReadOnly option is
that several Aster model instances can be defined with different default heights per
clutter class.
◼ Advanced Clutter Parameters: When the clutter class is mapped to a statistical class, a button ( ) is
available in the "..." column. Click this button to open the Advanced Clutter Parameters window.
Aster applies a loss when a ray goes through an obstacle and can also apply a loss for the interpolated
pixels inside buildings (indoor loss). This loss is based on the sum of the two types of losses defined for
each of Aster’s propagation classes: Penetration Loss (dB) and Linear Loss (dB/m).
In the Advanced Clutter Parameters window, select Specific propagation class parameters if you want to
modify the Penetration loss, Linear loss, and Clearance for this specific clutter class.
The values specified in this window override the corresponding values in the
Statistical propagation class default parameters table, for the current clutter class
only (any other clutter class that is mapped to a statistical class still uses the values
of the Statistical propagation class default parameters table or its own Specific
propagation class parameters).
When settings in the Advanced Clutter Parameters window have been changed from the default, the
corresponding active cell in the "..." column turns green ( ).
3. Click OK.
When a vector layer defines a unique clutter type, a Fixed propagation class is applied to all the polygons in the
vector layer. For example: if Fixed propagation class is set to "Building type 2", this class is defined for all the
polygons in the data file.
To define a new vector layer:
1. In the Parameters explorer, expand the Propagation Models folder, right-click Aster Propagation Model, and
select Properties from the context menu. The Aster Propagation Model Properties dialog box appears.
2. Select the Vectors tab and select Use vectors.
3. Under Path, click Browse ( ). The Open dialog box appears.
4. Select a vector file and click Open. The Polygon Definition dialog box appears.
5. In the Polygon Definition dialog box, define the following parameters:
◼ Height definition:
◼ Fixed height (m): Enter a default height to use for the polygon when no valid value is found in the file
attribute. It is strongly recommended to provide a valid height for each record to avoid using this value.
◼ Use ’height’ from file attribute: Select this option to define the height from a file attribute then select the
file attribute in the drop-down list on the right, for example: "HEIGHT". If no valid value is found in the file
attribute, the Fixed height (m) value is used for the polygon.
◼ Height defined above mean sea level (AMSL): By default, heights are defined as "above ground level"
(AGL). Select this option to specify that the Fixed height (m) is defined above mean sea level.
◼ Propagation class definition:
◼ Fixed propagation class: Select the default deterministic or statistical propagation class to use for the
polygon when no valid value is found in the file attribute.
◼ Use ’propagation class’ from file attribute: Select this option to define a file attribute then select the file
attribute in the drop-down list on the right, for example: "DESCRIPTION". Aster parses the data file,
detects all the string attributes, and makes them available in the drop-down list on the right. When this
option is selected, the table below it displays the relevant information under Field Values and
Propagation Class.
◼ Use this definition as default: Select this option to use this vector definition as default.
6. Click OK. The Aster Propagation Model Properties dialog box reappears with the path to the vector layer that
you just defined under Vectors.
7. If you want to reopen the Polygon Definition window, under Definition, click Browse ( ).
Aster enables you to consider indoor areas when you calculate a coverage prediction.
When indoor calculation is disabled, the output calculation grid has undefined values for all the pixels inside
buildings and coverage predictions will not be available for indoor locations, (Figure 3.3).
Aster does not apply specific indoor losses when it performs interpolation inside buildings. Considering indoor
losses or not depends on the workflow used in Aster when the signal level coverage predictions are calculated.
This section covers the following topics:
◼ "Enabling Indoor calculation without Indoor Losses" on page 33
◼ "Calculating a Service Level Map" on page 33
◼ "Calculating Clutter-based Indoor Loss from Atoll" on page 33
◼ "Calculating User-defined Indoor Loss from Aster" on page 34
Please bear in mind that you also activate user-defined indoor loss from Aster,
indoor loss may end up being considered twice.
If you activate clutter-based indoor loss both in Aster and in Aster, indoor loss may
end up being considered twice.
The Aster propagation model uses ray tracing techniques to incorporate horizontal diffraction and reflection, which
allows for efficient calculation. Ray tracing mode can be enable in the Aster Propagation Model Properties dialog
box.
Ray tracing is strongly recommended for micro cells where the antenna height is below the average building height
in the vicinity of the antenna. In this case, the horizontal diffraction and reflection component of the radio signal is
more significant than the vertical diffraction component.
However, ray tracing is not necessary when considering macro cells where the vertical diffraction component of the
radio signal is dominant. In this case, enabling ray tracing can have an adverse effect on calculation time with little
or no gain in model accuracy.
When in doubt, or when your network uses a mix of micro and macro cells managed by a single instance of Aster,
ray tracing should be enabled.
Aster generally disables ray tracing calculation in the following situations:
◼ When path loss matrix resolution is above 10 meters.
◼ When most propagation classes are statistical classes.
◼ When no deterministic propagation class is available in the ray tracing radius.
Radius (m) 800 N/A N/A The recommended value should be kept for
best compromise between accuracy and calcu-
lation time
5. Click OK.
Radio waves in extremely high frequency bands (mmWave) are prone to high atmospheric attenuation from being
absorbed by the gases in the atmosphere. Absorption by humidity in the atmosphere is significant except in desert
environments, and attenuation by rain (rain fade) is a serious problem even over short distances. The Aster
mmWave propagation model can include these factors in its calculations.
To enable atmospheric and rain attenuation:
1. In the Parameters explorer, expand the Propagation Models folder, right-click mmWave Aster Propagation
Model, and select Properties from the context menu. The Aster mmWave Propagation Model Properties
dialog box appears.
2. Select the Ray Tracing tab.
3. Select Atmospheric absorption to take atmospheric attenuation into account.
4. Select Rain attenuation and specify an average attenuation value (in dB/km) for rain fade.
5. Click OK.
You can switch the Aster Propagation Model Properties dialog box to advanced mode by setting the
hideAdvancedParamUI option to 0 in the [Core] section of the Aster.ini or AsterMmWave.ini file. In the advanced
mode, the following additional items appear:
◼ On the Clutter tab: The Calibration Correction (dB) and Building Calibration Correction (dB) columns appear.
◼ On the Ray Tracing tab: The Reflection/Diffraction Parameters frame appears.
◼ Parameters tab: new tab on which you can modify the low-level configuration parameters.
For more information on the advanced mode, see the Technical Reference Guide.
The results of the antenna relocation audit are based on the relocation strategy that is defined in the properties of
the Aster propagation model. under Conditions and Where to locate.
When Facade or Facade or roof are selected, the relocation algorithm takes into
account antenna azimuths so that final locations remain realistic with respect to
antenna azimuths (and hence avoid configurations in which antennas point
towards facades).
5. Click OK.
You can perform an antenna relocation audit for a single instance of multiple instances of the Aster propagation
model. For more information on using multiple instances of the propagation model, see "Creating a New Instance
of Aster" on page 24.
To perform an antenna relocation audit:
1. In the Parameters explorer, expand the Propagation Models folder, right-click Aster Propagation Model, and
select Audit > Antenna Outdoor Relocation from the context menu. The Outdoor Relocation window appears.
2. In the Outdoor Relocation window select one or several instances of the Aster propagation model or select
Only current propagation model.
3. Click Start. The audit starts and a progress bar is displayed. You can click Cancel to stop the calculation.
4. When the audit is finished, click the Results tab. You can copy and paste the results of the audit into Excel or
you can use the results to update the antenna locations in Atoll.
4.1 Preparation
The first step of a calibration process consists of creating an Aster project with all of the network and geographical
data necessary to recreate the CW measurement survey area. When the Aster project has been created with all the
necessary data, the CW measurement data can be imported and filtered in order to ensure that only meaningful data
is used for calibration.
In the Network explorer, the CW measurements are assigned to each transmitter. In the example below, the Aster
project contains three transmitters and a CW measurement file is imported for each transmitter.
For more information, see the Setting up the Calibration Project section in the Atoll Model Calibration Guide.
The Aster Propagation Model is optimised for urban areas, where the radio propagation is calculated with two major
components:
◼ Vertical diffraction over the rooftops (Deygout and Ikegami-Walfisch)
◼ Horizontal reflection/diffraction (ray launching)
Aster is fully integrated with Aster. Therefore, its main inputs are directly retrieved from the Aster project:
◼ Model parameters are stored in the Aster model properties
◼ Transmitter and Antenna parameters are retrieved from the Aster tables
◼ Geo inputs are retrieved from Aster geo layers
◼ Continuous wave measurements are retrieved from the CW Measurements folder in the Network explorer of
Aster.
Automatic calibration is a straightforward process in Aster. For data collection and preparation, it is strongly
recommended to follow the procedure in the Atoll Model Calibration Guide.
It is recommended to use CW measurements for the Aster calibration: CW (Continuous Wave) measurements are
collected in the field for a single transmitter using a single frequency, and provides the highest accuracy in the
calibration process.
Drive test data can in theory be used (and is supported by Aster), but the resulting calibration process is impaired
due to the various limitations inherent to this kind of data:
◼ Presence of interference
◼ Impossibility to collect data through a large distribution of distances, thus limiting the validity of model
calibration
◼ Directive antenna decreasing the accuracy of the calibrated model
◼ Low sampling rate (compatibility with the "Lee" criteria)
◼ Etc.
For recommendations on the measurement process, see the Atoll Model Calibration Guide.
Accurate geographic data is needed for realistic coverage predictions by Aster, but also for proper CW
measurement analysis and preparation in Aster.
The recommended types of geographic data for Aster are the following:
◼ Raster geographic data: a DTM (Digital Terrain Model) layer and either (or both) the Clutter Classes and
Clutter Heights layer. These files must have a sufficiently high resolution to obtain high-quality and accurate
results for the prediction and calibration process.
In urban areas, the recommended geo data resolution for Aster is:
◼ 50 m or less for DTM
◼ 5 m or less for height defining raster, i.e. usually the Clutter Heights raster (but it can be the Clutter Class
raster with default heights for clutter classes).
◼ 25m or less for the Clutter Class file when a Clutter Height raster is provided
In rural areas, the recommended geo data resolution for Aster is:
◼ 100m or less for DTM
◼ 50m or less for the Clutter Class file (rural settings are usually not based on a Clutter Heights layer; the
default heights of the Clutter Class layer are used instead).
◼ Vector data (optional): two types of vector data are useful:
◼ Building vectors, for refining the precision of the Aster height model
◼ Road vectors, representing at least major roads, useful for verifying measurement survey routes.
◼ Scanned maps: scanned maps are useful for verifying measurement survey routes in urban areas.
For more information on the preparation of the geographic data setup, see the Aster User Manual.
For more information on the following preparation steps for CW measurement data, see the Atoll Model Calibration
Guide:
◼ Data analysis
◼ Data filtering
◼ Data smoothing
Once measurement data is acquired and prepared, the Aster calibration process can start.
A preliminary task is to determine the number of different Aster calibrated instances you want to calibrate. It is
possible to:
◼ Calibrate a single model instance with all the measurements data, based on the semi-deterministic principle
of Aster. A single calibrated Aster instance can accurately model the following:
◼ Different frequencies
◼ Different environments: micro-urban, macro-urban, macro-suburban, etc.
◼ Create an instance of Aster for a subset of measurements representing different environment types, and/or
different frequency bands: the prediction performance of each calibrated model may be better for transmitters
in this specific type of environment (or frequency).
Choosing between these two strategies can be done by trying both, then analyzing the quality performance provided
by the calibrated models. If the two strategies provide almost similar prediction quality on measurement surveys, it
is recommended to use the first one as it is simpler to manage (i.e. assigning to each transmitter its own calibrated
model).
This chapter describes the calibration of a single Aster instance from a set of related measurements.
Some configuration parameters cannot be calibrated and must be defined before calibration in the Aster
Propagation Model Properties dialog box. The Parameters tab and items on other tabs will only appear if the
hideAdvancedParamUI option is set to 0 in the [Core] section of the Aster.ini or AsterMmWave.ini file (advanced
mode).
◼ General: Aster model instance name, view the register signature, and enter comments.
◼ Configuration: indoor calculation and indoor antennas parameters.
◼ Clutter: clutter parameters.
The Calibration Correction (dB) and Building Calibration Correction (dB) columns appear only in advanced
mode.
◼ Geo: geo data configuration.
◼ Ray Tracing: ray launching algorithm parameters.
The Reflection/Diffraction Parameters frame appears only in advanced mode.
◼ Parameters: configuration parameters. This tab appears only in advanced mode.
The Calibration menu can be disabled in the Aster.ini or AsterMmWave.ini file. For
more information, see the Aster Technical Reference Guide.
If you force the calibration module to use indoor points while these measurements
points are actually outdoor, then the Apply Indoor losses option should always be
deactivated. Since Aster performs interpolation of surrounding outdoor points,
these points will provide useful information as long as no additional indoor losses
are applied.
Regarding the indoor antennas relocation mode, it is recommended to keep it
activated (Relocate outdoor set to "Always" in Indoor Antenna Relocation Settings
dialog box) when dealing with scanned data from real transmitters, and deactivate
it when dealing with CW measurements.
In any case, it is unwise to rely on such automatic transmitter reposition features in
a calibration project. It is preferable to review all Aster warning messages and
manually relocate all the transmitters which were wrongly detected by Aster as
indoor.
As a rule of thumb, make sure that most measurement points are correctly geolocated as outdoor by:
◼ Verifying that the geo data is accurate enough to model the buildings with high precision
◼ Correcting the measurement points geolocation manually or via some pre-processing tools
◼ Filtering indoor points either from Aster or from Aster's CW measurements filtering feature
For each propagation class, the auto-calibration module will not calibrate some parameters that are related to:
◼ Penetration loss and linear loss when the signal runs into an obstacle.
◼ Clearance area for statistical classes.
It is recommended to leave these parameters to their default values.
Calibration Correction (dB) and Building Calibration Correction (dB) columns appear only in the advanced Aster interface.
For more information, see the Technical Reference Guide.
+ It is recommended to keep the default values in the selected configuration type. For
example:
◼ Macro: Ray tracing disabled.
◼ Micro and Standard: Ray tracing enabled the following settings:
◼ Radius: 800m
◼ Max number of diffractions/reflections: 4
◼ Rural: Ray tracing disabled.
Only the following parameters in the advanced Aster interface can be calibrated:
◼ Reflection/Diffraction parameters: for each propagation class, two reflection and diffraction losses are
defined; one for the first obstacle encountered during the ray-launching process, and one for the other
obstacles:
◼ Reflection Loss (dB) / First: reflection loss due to the first obstacle encountered during ray launching.
◼ Reflection Loss (dB) / Others: reflection loss due to other obstacles encountered during ray launching.
◼ Diffraction Loss (dB) First: diffraction loss due to the first obstacle encountered during ray launching.
◼ Reflection Loss (dB) / Others: diffraction loss due to other obstacles encountered during ray launching.
To analyse the performance of Aster on a set of CW measurements from Aster, the predicted signal levels should
first be added to the CW measurements in Aster.
To analyse Aster’s performance from Aster, you need to:
◼ Add predicted signal levels to CW measurements in Aster.
◼ Calculate the signal levels predicted by Aster.
◼ Verify that the signal levels predicted by Aster were added in the corresponding measurement path table.
◼ Display the statistics of CW measurement paths in Aster.
To add predicted signal levels to CW measurements in Aster:
1. In the Network explorer, right-click the CW Measurements folder and select Properties from the context menu.
The CW Measurements Properties dialog box appears.
2. In the CW Measurements Properties dialog box, select the Propagation tab.
3. On the Propagation tab, set Propagation Model to "Aster Propagation Model".
4. Click OK to close the CW Measurements Properties dialog box.
To calculate the signal levels predicted by Aster for all the CW measurement points:
1. In the Network explorer, right-click the CW Measurements folder and select Calculation > Calculate signal
levels from the context menu. Aster calculates the signal levels for all CW measurement paths and adds the
signal levels it predicted to all CW measurement points.
To verify that the signal levels predicted by Aster were added in the corresponding measurement path table:
1. In the Network explorer, expand the CW Measurements folder, right-click a CW measurement path, and select
Open Table and verify that some values were added in the corresponding P (dBm) column.
To display the statistics of CW measurement paths in Aster:
1. In the Network tab in the explorer window, right-click the CW Measurements folder, and select Display
Statistics from the context menu. The Measurement Path Selection dialog box appears.
2. In the Measurement Path Selection dialog box, select All measurement paths.
3. Click OK. The CW Measurements dialog box appears with the initial global Aster performance under Global
Statistics and the average and standard deviations per clutter class, per transmitter, and per measurement
path. For more information, see the Aster User Manual.
The performance of Aster on a set of CW measurements can also be analysed from Aster by using its own analysis
mode.
To analyse Aster performance:
1. In the Parameters explorer, right-click Aster Propagation Model and select Analysis from the context menu.
The Analysis wizard appears.
2. Select the Settings tab.
3. Under CW measurement path(s) to be used, select the check boxes of the measurement paths you want to
consider:
◼ You can click Select All to select all measurement paths or Unselect All to unselect them all.
◼ You can select contiguous rows by clicking the first row, pressing SHIFT and clicking the last row.
◼ You can select non-contiguous rows by pressing CTRL and clicking each row separately.
Under Analysis method, the settings of Resolution (m) and Indoor points are read-only. The values are
inherited from the Calibration wizard. If you want to modify them, do it from the Calibration wizard.
4. Click Start analysis. A progress bar appears and, when analysis is finished, the results appear on the Result
tab:
◼ Global Performance: provides overall statistics on the model fit to measurements for the initial default
model. The values are similar to the ones calculated by Aster.
◼ No points/Total: Number of measurement points used in analysis / total number of points in all
measurement paths used in analysis.
◼ Mean: Mean error value between predictions and measurements.
◼ Std Dev: Standard deviation of the differences between predictions and measurements.
◼ RMS: Root mean square of the differences between predictions and measurements.
◼ Corr: Correlation coefficient between measurements and predictions.
◼ 95% Error Range: 95% percentile values of the distribution function of the difference between
predictions and measurements. This means that 95% of the points are in the [Min,Max] range.
◼ Spread of Means: Standard deviation of the mean errors of the different measurement paths.
◼ Detailed Performance: gives similar statistics individually on each measurement path.
◼ Measurement path: measurement path name.
◼ No points/Total: Number of points used in analysis per measurement path / total number of points in
the same measurement path.
◼ No points: Number of points per measurement path.
◼ Mean: mean error between predictions and measurements.
◼ Std Dev: Standard deviation between predictions and measurements.
◼ RMS: Root mean square of the differences between predictions and measurements.
◼ Corr: Correlation coefficient between measurements and predictions..
◼ 95% Error Range: 95% percentile values of the distribution function of the difference between
predictions and measurements. This means that 95% of the points are in the [Min,Max] range.
◼ Input Parameters: lists the following additional information:
◼ Used CW measurement path: names of the measurement paths used in this analysis.
◼ Resolution (m): resolution used in this analysis, defined in the Analysis wizard.
◼ Indoor points: "Use" or "Ignore" indoor points, as defined in the Analysis wizard.
◼ Context:
◼ Duration: elapsed time from beginning to end of analysis.
For more information on Aster analysis indicators, see "Calibration Results Indicators" on page 51.
4.4 Calibration
The goal of the calibration process is to reduce the mean error and standard deviation of measured values versus
calculated values. For each parameter, the calibration process tries to find the best value that gives the best
standard deviation and the best mean error. The acceptable data ranges and step for each parameter have to be
defined.
The following sections describe the steps to follow to calibrate an Aster instance:
◼ "Creating a Calibration Copy of Aster" on page 46.
◼ "Calibrating an Aster Instance" on page 46.
◼ "Verifying Calibration Results" on page 50.
◼ "Committing Calibrated Parameters" on page 52.
Before starting a calibration, it is recommended to work on a new instance of the Aster model. You can later
calibrate the original model if needed.
To create a new instance of the Aster model:
1. In the Parameters explorer, expand the Propagation Models folder, right-click Aster Propagation Model, and
select Duplicate from the context menu. A new Aster instance, Copy of Aster Propagation Model, is added in
the Propagation Models folder.
2. You can rename the new Aster instance on the General tab of its properties dialog box, for example: Aster
Copy.
The settings defined in this dialog box for Resolution (m) and Indoor points are
inherited as read-only settings in the Analysis wizard. You can only modify them
from the Calibration wizard.
◼ Click Reset to defaults to reset the parameters in this dialog box to their default values.
◼ Click Save Parameters to save the current parameters for later use in the propagation model.
◼ Advanced Parameters: Click this button (available only in advanced mode) to open the Advanced
Calibration Parameters dialog box. For more information, see "Advanced Calibration Parameters (Inputs)"
on page 47.
5. Optionally, if you are using the mmWave Aster propagation model, you can tune the linear loss through
vegetation by precalibrating the deterministic propagation classes associated with vegetation:
a. Click Precalibration. The Precalibration Results window opens.
The Vegetation Propagation Classes list displays the linear loss (in dB/m) for each vegetation type.
b. Click Start Precalibration to run the precalibration calculation.
c. When the precalibration is finished, information about the number of points is displayed. You can choose
to Commit the result.
d. Click Close.
6. Click Start Calibration. A progress bar appears.
When calibration is finished, the results appear on the Result tab. For more information, see "Verifying
Calibration Results" on page 50.
◼ For the parameters used to minimise the standard deviation (e.g. "K3 - los"), a minimum value, a maximum
value, and a step must be defined. For each parameter, the calibration module scans all the possible values
between the defined minimum and maximum, with the defined step. It then finds the value that best
minimises the standard deviation.
◼ For the parameters used to minimise the mean error (e.g. "K1 - los"), these columns are not needed. The
calibration module finds automatically the value that makes the mean error equal to zero.
In general, no changes need to be made in the Advanced Calibration Parameters dialog box and the default values
can be kept. If necessary, the parameter range can be Reset to defaults on the Parameters tab of the Calibration
wizard.
For indoor points, the calculation module can be instructed to either "Ignore" or
"Use" the points detected as indoor.
m is the standard deviation of the mean errors of the different measurement paths.
During a standard calibration, the calibration engine looks for the best parameters to minimise the standard
deviation , while keeping the overall average close to zero. Hence, the minimised cost function will be equal to the
standard deviation:
When the spread of means optimisation is enabled, the cost function becomes:
where the weighting m can be configured from the Aster interface (default value = 1).
In this case, the calibration engine will look for the best set of parameters to minimise both the standard deviation
and the spread of means.
Where:
This means that when one of the two items (standard deviation or spread of means) is above the target, the penalty
for that part above this target is 10 times the standard value. This cost function will tend to make both the spread
of means and standard deviation converge to the target values.
Example:
[Calibration]
spreadOfMeansWeight = 1.0
spreadOfMeansTarget = 2
standardDeviationWeight = 1.0
standardDeviationTarget = 6.75
In this case, the module will try to have the spread of means below 2 dB and the standard deviation below 6.75 dB,
with equal weight between the two targets. When the targets are set to zero, this is equivalent to the standard default
cost function
When calibration is finished, the results are displayed on the Result tab of the Calibration wizard:
◼ Global Performance: this section provides overall statistics on the model fit to measurements for the initial
default model and the final calibrated model. The initial values are similar to the values calculated by Aster.
For more information, see "Calibration Results Indicators" on page 51.
◼ Detailed Performance: this section gives similar statistics individually on each measurement path.
For more information, see "Calibration Results Indicators" on page 51.
◼ Input Parameters: lists the following additional information:
◼ Used CW measurement path: names of the measurement paths used in this calibration.
◼ Resolution (m): resolution used in this calibration, defined in the Calibration wizard.
◼ Indoor points: "Use" or "Ignore" indoor points, as defined in the Calibration wizard.
◼ Context:
◼ Duration: elapsed time from the beginning to the end of this calibration.
◼ Date: end date (<YYYY-MM-DD HH:MM:SS>) of this calibration.
The following buttons are also available on the Result tab:
◼ Advanced Parameters: click this button (available only in advanced mode) to display the Advanced
Calibration Parameters dialog where initial and final values are indicated for each calibrated parameter.
◼ Export: to export the results to a TXT, HTML, XML or XLS file.
◼ Commit: to apply the calibrated values directly in this Aster model instance.
◼ Std Dev: Pearson-type standard deviation ( ) of the differences between predictions and measurements.
◼ RMS: Root mean square ( rms ) of the differences between predictions and measurements.
◼ Corr: Correlation coefficient between measurement (x) and prediction (y) variables.
N
1
----
N x y – x y i i
p (x,y) = 1
-------------------------------------
-
x y
Where:
N
1
x = ----
N x i
i=0
N
1
y = ----
N y i
i=0
◼ 95% Error Range: The 95% percentile values of the distribution function of the difference between predictions
and measurements. This means that 95% of the points are in the range [Min,Max]. This percentile ratio can be
modified from the Aster.ini or AsterMmWave.ini file:
[Calibration]
percentileMinMax = 95
◼ Spread of Means: Standard Deviation of the mean errors ( m ) of the different measurement paths.
Where:
M is the number of measurement paths
m j is the mean error between predictions and measurements for path " j ".
◼ Advanced Performance Results: You can display additional performance results at the end of the calibration
results. The Advanced Results section displays the initial and calibrated values for the following indicators:
◼ No points
◼ Mean (dB)
◼ Std Dev (dB)
◼ RMS (dB)
Results are provided for All, LOS, NLOS, and RT radio components.
You can enable the Advanced Results section from the Aster.ini or AsterMmWave.ini file:
[Calibration]
CalibAdvancedResults = 0
When calibration is finished, and before committing the new calibrated model parameters, advanced users can
make manual adjustments to compensate for values that could not be calibrated due to missing or incomplete data.
For example, the missing values can be extrapolated from existing data or from standard values.
To make manual calibration adjustments:
1. Click Start calibration in the Calibration wizard. When calibration is finished, the results appear on the Result
tab.
2. On the Result tab, click the Advanced Parameters button. The Advanced Calibration Parameters dialog box
appears.
◼ Parameters: for each parameter in this column, there is an Initial Value and a Final Value.
◼ Initial Value: initial value before calibration. This value is read-only.
◼ Final Value: final value after calibration. This value can be modified by advanced users.
In order for the calibrated parameters to be taken into account by the model, they need to be committed by clicking
OK in the Advanced Calibration Parameters dialog box, then Commit and Close on the Result tab of the previous
window. The new parameter values will then appear in the Aster properties.
4.5 Validation
After calibration has been performed and committed, you can verify the calibration accuracy directly in Aster by
performing the same steps as in the initial analysis of the Aster model.
The Global Statistics section provides the global performance of the model. These values are usually the same as
the ones displayed on the Result tab of the Calibration wizard.