Winprop Ebook 31 031-2021

EBOOK
Learn Basics of Wave Propagation

with Altair WinProp
© Altair Engineering, Inc. All Rights Reserved. / altair.com / Nasdaq: ALTR / Contact Us
eBook / Learn Basics of Wave Propagation with Altair WinProp
Contents
Acknowledgement ............................................................................................................................................. 4
1 INTRODUCTION TO WAVE PROPAGATION ANALYSIS .................................................. 5

1.1 History of wireless communication and its applications .................................................................... 5
1.2 Mobile Radio Channel ............................................................................................................................. 6
1.3 Wave Propagation Principles ................................................................................................................. 7
1.3.1 Path loss .......................................................................................................................................... 10
1.4 Propagation Path-Loss Models ........................................................................................................... 10
2 INTRODUCTION TO WINPROP ........................................................................................ 12

2.1 WinProp Overview................................................................................................................................. 12
2.2 WinProp Applications ........................................................................................................................... 13
2.2.1 Automotive ...................................................................................................................................... 14
2.2.2 Aerospace ....................................................................................................................................... 14
2.2.3 Telecommunication ......................................................................................................................... 15
2.2.4 Wireless Sensors ............................................................................................................................ 16
2.2.5 Railway ............................................................................................................................................ 16
2.3 WinProp Components........................................................................................................................... 17
2.3.1 ProMan ............................................................................................................................................ 17
2.3.2 WallMan .......................................................................................................................................... 17
2.4 TuMan ..................................................................................................................................................... 18
2.4.1 AMan ............................................................................................................................................... 18
2.4.2 CoMan ............................................................................................................................................. 19
2.4.3 CompoMan ...................................................................................................................................... 19
3 WAVE PROPAGATION MODELS FOR INDOOR SCENARIOS ....................................... 20

3.1 Indoor Models ........................................................................................................................................ 20
3.2 Types of Propagation Models .............................................................................................................. 20
3.3 Required Database ................................................................................................................................ 21
3.4 Modified Free Space Model (MF): ........................................................................................................ 21
3.5 Motley-Keenan Model (MK): ................................................................................................................. 22
3.6 COST- Multi-Wall Model (MW) .............................................................................................................. 23
3.7 Standard Ray Tracing (SRT) ................................................................................................................ 24
3.7.1 Propagation Paths ........................................................................................................................... 24
3.7.2 Computation of Each Ray’s Contribution ........................................................................................ 25

© Altair Engineering, Inc. All Rights Reserved. / altair.com / Nasdaq: ALTR / Contact Us 1
3.8 Intelligent Ray Tracing (IRT) ................................................................................................................ 25
3.8.1 Propagation Paths ........................................................................................................................... 26
3.8.3 Pre-processing ................................................................................................................................ 26
3.9 Dominant Paths (DP)............................................................................................................................. 27
3.9.1 Prediction ........................................................................................................................................ 27
3.9.2 Multipath Propagation in Indoor Environments ............................................................................... 28
3.9.3 Principle of the Dominant Paths ...................................................................................................... 28
3.9.4 Determination of the Dominant Paths ............................................................................................. 29
3.9.5 Path selection criteria ...................................................................................................................... 31
3.9.6 Share of the empirical and neural model ........................................................................................ 32
3.10 Indoor Scenario Example ..................................................................................................................... 32
3.11 Indoor Advanced Scenario Model Example ....................................................................................... 49
4 WAVE PROPAGATION MODEL FOR URBAN SCENARIOS ........................................... 70

4.1 Motivation .............................................................................................................................................. 70
4.2 Required Databases .............................................................................................................................. 70
4.3 Empirical COST-Walfisch-Ikegami Model ........................................................................................... 71
4.4 Urban Scenario Example ...................................................................................................................... 73
4.5 Basics of Ray-Optical Models .............................................................................................................. 87
4.5.1 Specular Reflection ......................................................................................................................... 87
4.5.2 Diffraction ........................................................................................................................................ 88
4.5.3 Multiple Diffraction ........................................................................................................................... 88
4.5.4 Scattering ........................................................................................................................................ 88
4.5.5 Penetration and Absorption ............................................................................................................. 88
4.5.6 Ray Tracing Versus Ray Launching................................................................................................ 89
4.6 Ray Tracing ............................................................................................................................................ 89
4.6.1 Acceleration ..................................................................................................................................... 89
4.6.3 Breakpoint ....................................................................................................................................... 91
4.7 Description of the Individual Ray-Optical Models ............................................................................. 91
4.7.1 Standard 3D Ray Tracing (SRT) ..................................................................................................... 92
4.7.2 Intelligent Ray Tracing (IRT) ........................................................................................................... 92
4.8 Combined (Urban + Indoor) Propagation-Hybrid Prediction ............................................................ 97
4.9 Urban + Indoor Scenario Example ...................................................................................................... 98

5 WAVE PROPAGATION MODEL FOR RURAL/SUBURBAN SCENARIOS .................... 114

5.1 Motivation ............................................................................................................................................ 114
5.2 Required Databases ............................................................................................................................ 114
5.3 Hata-Okumura Model .......................................................................................................................... 114
5.4 Parabolic Equation Method ................................................................................................................ 116
5.4.1 Basic Principles ............................................................................................................................. 116
5.4.2 Absorbing Media ........................................................................................................................... 117
5.4.3 Impedance Boundary at the Ground ............................................................................................. 117
5.4.4 Wide Angle Parabolic Equation..................................................................................................... 119
5.4.5 Comparison between the Hata-Okumura and Parabolic Equation Models .................................. 119
5.5 Rural/Suburban Scenario Example ................................................................................................... 121
Acknowledgement
A very special Thank You goes to all the many colleagues who contributed in different ways:
Shannon Harshkant Mistry for adapting the main content from Reiner Hoppe`s original work and for authoring content.
Smitha Ramaiah for reviewing, testing, editing & updating the tutorials and creating videos for the tutorials.
Reiner Hoppe and Jordi Soler for the detailed technical review, For sure, your feedback and suggestions had a significant
impact on the “shape” and content of this book.
Priyanka Nagraj for the formatting and editing, Rahul Ponginan, for reviewing and editing the book
The entire Altair WinProp Documentation Team for putting together 1000’s of pages of documentation.
Lastly, the entire Altair WinProp team deserves huge credit for their passion & dedication! It is so exciting to see how WinProp
has evolved throughout the last couple of years.
Every effort has been made to keep the book free from technical as well as other mistakes. However, publishers and authors will not
be responsible for loss, damage in any form and consequences arising directly or indirectly from the use of this book.
© 2021 Altair Engineering, Inc. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, or
translated to another language without the written permission of Altair Engineering, Inc. To obtain this permission, write to the attention
Altair Engineering legal department at:
1820 Big Beaver, Troy, Michigan, USA, or call +1-248-614-2400.
1 Introduction to Wave Propagation Analysis
1.1 History of wireless communication and its applications
Wireless communication has now become the most dominant mode of communication. The history of wireless communication can
be traced back to 1864 when James Clerk Maxwell formulated the electromagnetic theory. He predicted the existence of
electromagnetic waves and framed a series of differential equations that showed the movement of EM waves. These were known as
Maxwell’s equations. Before Maxwell’s discovery, it was believed that energy can only be transported over short distances. Maxwell’s
prediction of transporting energy at the speed of light with no barriers of distances was a great contribution to the development of
wireless technology. This discovery paved a way for Heinrich Hertz who then showed the physical existence of radio waves in 1887
using a spark transmitter and a receiver. Guglielmo Marconi’s approach was different from other scientists. In the year 1895, he used
radio waves to transmit a message a few miles near his home. He wanted to improve the wireless system so that it could be
commercially used. Simultaneously, Sir Jagdish Chandra Bose generated radio waves and constructed devices like horn antennas
and microwave reflectors. In 1897 a wireless telegraph was sent across the English Channel by Marconi. This was the first time a
communication link was established for longer distances. This marked the beginning of wireless technology. On 23 December 1900,
the Canadian inventor Reginald A. Fessenden became the first person to send audio (wireless telephony) utilizing electromagnetic
waves, successfully transmitting over about 1.6 kilometers, and six years later Christmas Eve 1906, he became the first person to
make a public radio broadcast. [Wikipedia]. Eventually, due to advances in the field of wireless technology, the first bidirectional
communication link was established in the year 1902. A lot of progress in the field of wave modulation took place around the early
1900s. Soon audio and video data were transmitted through wireless signals over any distance. By 1927 the first long-distance
television program was broadcasted in the United States. The word “Telecommunication” was first coined in 1932 and the
International Telecommunications Union (ITU) was formed. In 1946, the first public mobile telephone systems were introduced in 5
cities in the US. The launch of the first satellite Sputnik, into space, marked a new era of wireless systems. The wireless technology
has seen great advancements over the years that have caused a paradigm shift from wired to wireless technology.
There has been significant growth in commercial wireless technologies like mobile telephony, pagers, wireless computer networks,
cellular networks, wireless Internet, and laptop and handheld computers with a wireless connection.
The wireless communications have transitioned from first-generation systems that were only focused on voice communications to
well-designed fourth-generation systems that provide mobile web access, gaming services, HD mobile television, video conferencing,
etc. Today, wireless devices are almost everywhere. We can see how common cell phones have become for day-to-day use. The
television transmissions are broadcasted through satellites, directly to our homes. Wireless networks have started to replace Ethernet.
Mobility.
Wireless technology is used in a wide range of applications in automotive, commercial, telecommunication, railway and aerospace
industries.
Some of the applications are as follows:
• Digital audio and video broadcasting like radio and television.
• Mobile telecommunication systems for personal use.
• Wireless power transfer
• Home automation systems
• Global positioning system (GPS) that allows users to ascertain their location
• Cordless computer peripherals like mouse, keyboards, headphones

• Cordless phones with limited range
• Radio navigation used by air traffic control
• Wireless medical equipment to measure the heart rate, oxygen level, etc.
• Low power wireless communication networks like Wi-Fi
• Mobile cellular networks (4G, 5G) for personal use.
• Local area networks like Wi-Fi
• Internet of thinks (IoT) to connect every device also in home & industry, e.g. for automation.
• Wireless medical equipment to measure the heart rate, oxygen level, etc.
• Cordless phones and computer peripherals like mouse, keyboards, headphones
• Digital audio and video broadcasting like radio and television.
• Global positioning system (GPS) that allows users to ascertain their location
• Radio navigation for air traffic control
1.2 Mobile Radio Channel
The performance of wireless communication systems depends in a fundamental way on the mobile radio channel. In contrast to wired
channels that are stationary and easy to design, radio channels show a time-variant behavior which complicates their analysis. The
transmission path between transmitter and receiver can vary from simple line-of-sight to one severely obstructed by buildings, walls
and furniture. Therefore, predicting the propagation characteristics between two antennas still belongs to the most important tasks of
the design and installation of wireless networks, ranging from low bit-rate IoT to high bit-rate cellular and local area networks.
The mobile radio channel concerning transmission outdoors but also within buildings is characterized by a multipath scenario as
shown in Figure 1.1.
Figure 1.1 : Multipath propagation within buildings
The signal from the transmitting antenna - usually only the downlink is considered as the principle of reciprocity applies - propagates
along different paths to the antenna of the (mobile) receiver. In many cases there is no direct line-of-sight and only paths connecting
transmitter and receiver penetrate several walls and are reflected, diffracted, and scattered at several different obstacles. Since the
phases of the waves are randomly distributed, the superposition of these contributions causes constructive and destructive
interference (i.e. small-scale fading) which leads to rapidly fluctuating signal levels over very small distances.
Figure 1.2 : Large scale and small-scale fading
Figure 1.2 illustrates this small-scale fading and the slower large-scale signal variation for an indoor radio communication system.
While the small-scale fading is random, the large-scale variations occur due to fundamental changes of the propagation paths (e.g.
larger distances, different obstacles). Typically, the local average of the received power is computed by averaging signal
measurements over an interval of 10 to 20 wavelengths, which corresponds to movements of the receiver of 1.5 m to 3 m at a
frequency of 2 GHz.
1.3 Wave Propagation Principles
With decreasing wavelength, i.e. increasing frequency, the wave propagation becomes more and more like the propagation of light.
A radio ray is assumed to propagate essentially along a straight line and is influenced only by the given obstacles. As a criterion for
this type of modeling to be successful, the wavelength should be much smaller than the extensions of considered objects, e.g. the
building structures. At the frequencies used for wireless communications this criterion is typically sufficiently fulfilled.
The phenomena which influence radio wave propagation can generally be described by four basic mechanisms:
Reflection – The impinging of an electromagnetic wave on an object which has a larger dimension than the wavelength of the
propagating wave causes it to reflect. This phenomenon is known as reflection.
Reflection loss depends on
• Angle of incidence
• Properties of reflecting material: permittivity, conductance, permeability
• Polarization of incident wave
• Fresnel coefficients for modelling the reflection
Ei
br
Ei bi Er
Er
a i a r Material 1
m 1 , e 1, s 1
n
QR Material 2
m 2 , e 2, s
Et 2
a t
Et bt
Figure 1.3 : Reflection
Transmission – Transmissions are relevant for penetration of obstacles (as e.g. walls) Transmission loss depends on:
• Angle of incidence
• Properties of material: permittivity, conductance, permeability
• Polarization of incident wave
• Fresnel coefficients for modelling the transmission
Ei
br
Ei bi Er
Er
a i a r Material 1
m 1 , e 1, s 1
n
QR Material 2
m 2 , e 2, s
Et 2
a t
Et bt
Figure 1.4 : Transmission
Penetration loss includes two parts:
• Loss at border between materials
• Loss for penetration of plate
Figure 1.5 : Penetration loss
Diffraction- When the path between a transmitter and the receiver is blocked by a sharp surface that has edges, the wave bends and
deviates around the obstacle surface, this phenomenon is called diffraction. This bent secondary wave travels behind the obstacle,
giving rise to a shadow region that does not exist in the line of the sight of a transmitter and receiver.
Diffraction loss depends on:
• Angle of incidence & angle of diffraction
• Properties of material: epsilon, µ and sigma
• Polarisation of incident wave
• UTD coefficients with Luebbers extension for modelling the diffraction
k Ei
Ei Ed
F Ed
QD F
a
(2-n) p
i
Figure 1.6 : Diffraction
Scattering - When the signal hits an obstacle having a rough surface whose size is smaller or on the order of the signal wavelength,
scattering of the signal takes place. Scattering causes the signal to spread out in all directions.
• Scattering occurs on rough surfaces
• Subdivision of scattering object into numerous scattering elements
• Consideration of the relevant part to obtain acceptable computation effort
• Example: Ground properties
Figure 1.7 : Scattering
1.3.1 Path loss
The signal that is received by the receiver is a mixture of these modes of propagation as it hits a lot of obstacles on its way to the
receiver. Due to reflection, diffractions, and scattering, several signal paths are created, and the signals arrive at the receiver with
different time delays and over several different paths. This is referred to as multipath propagation.
These multiple signal paths can interfere with each other in a constructive or destructive manner, causing the signal energy to vary.
There are fluctuations in the amplitude and phase of the signal which leads to the fading of the signal energy. Superposition of
multiple paths leads to fading channel
• Fast fading – due to random phase variations
• Slow fading – due to principle changes in propagation channels
The above-mentioned factors contribute to Path loss. Path loss is the reduction in signal energy as it propagates through space.
Propagation losses are included in path loss. The factors like the terrain and environment (indoor, rural, urban, etc.), the medium of
propagation, the distance between source antenna and receiver antenna, and dimensions and locations of antennas influence the
path loss. Path loss is technically stated as ratio of transmitted power to the received power.
1.4 Propagation Path-Loss Models

Before any wireless system is implemented, propagation characteristics of the environment should be known. A series of propagation
prediction models are used for designing and analyzing the propagation characteristics and path loss. Predictions are required for
proper coverage planning, the determination of multipath effects as well as for interference and cell calculations, which are the basis
for the high-level network planning process. Without these predictions, statistical data of the environments can be obtained only by
field measurements. This makes it expensive and very time-consuming.
Propagation models are broadly classified in to three types
1. Empirical models
2. Semi-deterministic models
3. Deterministic models
Besides the prediction of the path loss, the delay and angular spread can be computed as well as LOS / NLOS, directional channel
impulse response, angular profile, and propagation.
2 Introduction to WinProp
WinProp is a complete suite of tools in the domain of wave propagation and radio network planning. With applications ranging from
satellite to terrestrial, from rural via urban to indoor radio links, WinProp’ s innovative wave propagation models combine accuracy
with short computation time.
2.1 WinProp Overview

WinProp is a comprehensive and leading simulation tool in the domain of wave propagation and radio network planning
Propagation Models
WinProp’ s powerful simulation methods are all high-frequency methods, valid when the geometrical features are much larger than
the wavelength of the signal. Several methods are fully three-dimensional, consider 3D object data and compute all rays in 3D. Others
focus on the vertical plane only.
Databases
WinProp supports the following databases to describe geometry:
Pixel database
Typically used for rural/suburban scenarios. Elevations and optionally land usage are stored on a pixel-by-pixel basis.
2.5D Vector database
Typically used for urban scenarios. Each building or vegetation object is described by a polygonal ground plane and an individual
height above street level using polygonal cylinders. If the urban area is not flat, the topography can be included.
3D Vector database
Typically used for indoor scenarios or small outdoor scenarios. Each building or other object is described by flat polygons with arbitrary
orientations. This allows maximum geometrical freedom.
Computation and Simulation
WinProp allows the planning of coverage and capacity as well as network simulations (for example, the performance of algorithms,
analysis of delays). Depending on the application, WinProp offers static, Monte-Carlo, and dynamic network simulators. You can
define the (location-dependent) traffic for circuit and packet-switched services (for example, statistical distributions and mobility).
Coverage
Different transmission modes can be defined (with parameters including bandwidth, coding schemes, required signal-to-noise-and-
interference ratio, signal threshold, transmit power) upon which the coverage maps (for example, cell assignment, best server, active
set, channel quality, received power in downlink and uplink, signal-to-noise-and-interference ratio) is computed individually for each
transmission mode. Link adaptation is considered and depends on the channel quality predicted with the propagation models.
Maximum received power as well as maximum achievable data rates are predicted accurately for each location in the coverage area.
Capacity
WinProp calculates the capacity (for example, throughput, maximum data rates, packet delays, QoS) of the different radio links and
cells in the network based on the coverage analysis and the traffic assumptions. Capacity limitations and overloaded cells are
detected easily, networks are optimized to provide both high capacity and throughput. Capacity improvements due to MIMO and / or
beamforming are modeled accurately due to the sophisticated deterministic propagation models.
Arbitrary antenna configurations (linear, circular) are possible and their impact on the radio channel is determined during the
propagation analysis.
WinProp API
The WinProp API enables flexible integration into other software tools.
A description of the API interface is available in the installation directory, together with C/C++ sample project for MS Visual Studio.
WinProp and Feko Interaction
Design the antenna and calculate the radiation characteristic in terms of a 3D antenna pattern using Feko. Then superimpose the far
field on the 3D radio channels calculated with WinProp.
WinProp supports the import of antenna patterns from several sources, including Feko. You can design the antennas and calculate
the radiation patterns in Feko and use those patterns to represent transmitting and receiving antennas in the radio channel analysis
in WinProp.
WinProp also supports the import and processing of radar cross-section patterns calculated by Feko. To accelerate certain
simulations in WinProp, for example, for automotive RADAR, complex vehicles can be substituted by their radar cross-sections.
2.2 WinProp Applications
Altair WinProp will show key applications in Automotive, Railway, and Telecom industries like radar sensors and installed antenna
performance analysis by virtual-drive tests, the wireless network design for railway scenarios including tunnels and metro stations,
and radio coverage planning for heterogeneous wireless networks including 5G.
WinProp is used in, but not limited to, the following applications:
Automotive- antenna evaluation
Analysis of car-antenna performance along a trajectory using a virtual drive test or a simulation with the explicit time variance
Aerospace - airport radio coverage
Analysis of radio coverage at airports between the terminals and airplanes. An example includes the coverage of an airplane as it
moves away from the terminal to the runway.
Telecommunication - radio network planning for an urban scenario
Analysis of a radio network in an urban scenario. For example, you can optimize the network coverage of a cellular network while
considering the effects of buildings, vegetation and topography.
Telecommunication - indoor radio network planning
Analysis of a radio network indoor. An example includes optimizing a WIFI network to give the best coverage inside an office building.
Wireless Sensors - planning of sensor network
Connectivity analysis of wireless sensor networks that monitors physical or environmental conditions. An example includes the
placement of a sensor network inside an industrial environment to automate assembly lines.
Railway - in-cabin radio coverage
Connectivity analysis of the transmit antennas in a train. An example includes optimizing the WIFI network in-cabin and considering
the seats and a separate room.
Let us know look at the applications in detail
2.2.1 Automotive
Evaluate antenna performance using a virtual-drive test in WinProp. Evaluate different car antenna configurations regarding receiving
power and MIMO throughput.
Define the topography, building, cars, test route and air interface (for example, LTE). Superimpose three-dimensional car antenna
patterns (calculated with Feko) and analyze the radio waves impinging on the car antenna along the test route as well as possible
antenna configurations based on performance indicators.
The following configurations are supported:
Car-to-car communication
Car-to-infrastructure communication
Radar channel (collision avoidance)
Altair WinProp considers the full environment including buildings, cars, street objects to get accurate representations of the radio
waves impinging on the installed car antennas and the multipath radar channels including reflections, diffractions and scattered
contributions. For the efficient analysis, the car objects can be also replaced by their corresponding radar cross-sections (pre-
calculated in Feko). Thus, allowing the realistic and fully reproducible evaluations of different options for the antennas and sensors
including their integration and configuration.
Cars have started being developed to save lives by talking to each other and the infrastructure around them using communication
antennas in a connected car concept. The automotive evolution towards autonomous driving includes also ADAS radar sensors to
be aware of surrounding obstacles. Both radar sensors and installed car antennas can be efficiently evaluated using virtual-drive
tests in WinProp.
Figure 2.1 : Automotive Virtual Drive test

2.2.2 Aerospace
WinProp is well suited to the analysis of airport radio coverage.
At airports, communication between airplanes and terminals is paramount to ensure the safety of people. WinProp is the leading tool
to analyze the radio coverage within terminals and airplanes. It allows for the evaluation of different transmitter options and the
resulting coverage situation as well as considering airplanes as obstacles when determining radio coverage.
Figure 2.2 : Ray tracing between communication tower (on the left) and airplane and the radio coverage at the airport terminal (to
the right).
2.2.3 Telecommunication
WinProp’s technologies make the tool applicable to the planning of both urban and indoor radio networks.
Indoor Radio Network Planning
Process 3D building vector data that includes outer and inner walls, subdivisions like doors and windows. Specify the construction
materials and their electrical properties or import from a database containing industry-standard materials. Analyze the receiving
coverage and throughput (data rate) for WIFI planning or an indoor site for a cellular network.
Figure 2.3 : An example of an indoor radio network planning.
Urban Scenario Radio Network Planning
Process 3D building vector data superimposed to terrain layer using highly accurate wave propagation models for macro-cells and
micro-cells.
Figure 2.4 : Radio Coverage Planning for Heterogeneous Wireless Networks, including 5G
WinProp offers highly accurate and fast wave propagation models as well as radio network planning modules for almost every
standard (cellular incl. 5G, LTE, WIFI, broadcasting, TETRA, etc.). Based on this WinProp is ideally suited for the radio coverage
planning in arbitrary scenarios, the network roll-out and optimization, the comparison of deployment concepts, the evaluation of
installed antenna performance, the co-existence analysis of radio networks and for IoT deployments.
Radio Coverage Planning for Heterogeneous Wireless Networks, including 5G Heterogeneous wireless networks combine outdoor
macro- and micro-cells as well as indoor pico- and femto-cells. By exploiting cellular and WIFI standards the growing demand for
enhanced mobile broadband in various environments shall be served. New use cases, the usage of higher frequency bands and new
antenna techniques have significantly increased the need for radio propagation modelling and network planning simulation.
2.2.4 Wireless Sensors

Wireless sensor networks consist of spatially distributed autonomous sensors to cooperatively monitor physical or environmental
conditions (for example, temperature, sound, vibration, pressure, motion or pollutants).
Applications for wireless sensor networks are varied, typically involving monitoring, tracking, or controlling using, for example,
Wireless Hart, ZigBee and LoRa. Specific applications include habitat monitoring, object tracking, nuclear reactor control, fire
detection, and traffic monitoring.
WinProp covers the following simulation aspects:
• Connectivity between the sensor nodes of the wireless network.
• Channel assignment in a self-organizing sensor network.
• Determination of optimum paths between sensor and gateway nodes considering different criteria, such as best SNIR,
minimum path delay.
2.2.5 Railway
WinProp can be used to evaluate the in-cabin radio coverage.
In the world of ubiquitous networking, people want to have continuous connectivity even when travelling. Using WinProp you can
evaluate and plan different transmitting antennas and leaky feeders to ensure strong in-cabin radio coverage on trains. Altair WinProp
is used for the wireless network design and deployment in various railway scenarios including tunnels and metro stations, inside train
wagons, as well as along railway tracks. Key challenges for train/metro operators are increasing traffic volumes, ensuring passengers
safety and security during their journey, as well as providing real-time multimedia information and access to social networks in stations
and tunnels. To meet these requirements various broadband telecommunication networks based on WIFI, GSM-R, LTE need to be
put in place.
Both antennas and leaky feeder cables can be deployed in the 3D environment of the station/tunnel scenario including the train
Figure 2.5 : Wireless Network Design for Railway Scenarios, Including Tunnels and Metro Stations
2.3 WinProp Components
2.3.1 ProMan
The central tool in the WinProp suite is ProMan (propagation manager). ProMan is where you define and edit the simulation project
settings, the simulator for all scenarios and where you display and evaluate results.
2.3.2 WallMan
Figure 2.6 : A detailed multi-floor indoor building in an urban surrounding.
WallMan is the graphical editor for vector building databases. It is used to prepare a geometry database that can become part of the
ProMan project. Preparing such geometries0 in WallMan, or even just converting them from other sources in WallMan, is a common
step in indoor and urban scenarios.
In rural scenarios, WallMan is often not needed, since databases such as elevation maps and land-usage maps can be imported
directly into ProMan.
WallMan is also instrumental in preparing hybrid urban/indoor scenarios, where some buildings are described with interior detail while
others are not.
Figure 2.7 : The graphical database editor for WallMan.
2.4 TuMan
TuMan (tunnel manager) is a graphical editor to prepare tunnel and stadium geometries efficiently. Geometries from TuMan can be
exported to WallMan to add further detail if needed.
Figure 2.8 : The tunnel propagation models included in ProMan’ s tunnel module can predict the propagation inside the tunnel.
2.4.1 AMan
AMan (antenna manager) is the graphical editor for antenna patterns that are used to prepare antenna patterns for use in ProMan. It
is not an antenna simulator but can convert antenna patterns from other sources to the correct format. It can also generate an
approximate 3D antenna pattern in cases where only two 2D pattern cuts are available. Furthermore, when antennas are combined
on masts, AMan will generate the combined pattern, taking all necessary effects into account.
Figure 2.9 : A 3D antenna pattern.
2.4.2 CoMan
CoMan (connectivity manager) is the connectivity simulator for sensor and MESH networks used to simulate wireless sensor
networks.
Figure 2.10 : Point to multipoint prediction on multiple height levels
2.4.3 CompoMan
CompoMan is the editor for components used in wireless indoor network installations.
Figure 2.11 : The CompoMan editor
3 Wave Propagation Models for Indoor Scenarios
3.1 Indoor Models

The propagation of a signal in an indoor environment is prone to interference. Indoor radio channels are different from outdoor models
in that the:
• Distances are shorter between transmitter and receiver
• Environment variability is greater for a shorter distance
Unlike the outdoor environment, these propagations are not affected by natural factors like rain, snow, temperature changes, etc.
However, there are several elements in an indoor environment like furniture, walls, windows, etc. that contribute to reflection,
refraction and scattering of the original signal.
• Propagation can be influenced by
• The layout of the buildings
• Construction materials
The type of building is an important contributor - fixed walls (hard partitions), movable walls (soft partitions), sports arenas, residential
buildings, etc.
This results in multipath fading where the transmitted signal is received by the receiver through more than one path with different
time delays. Very little can be done to avoid such losses in an indoor environment. However, if we characterize the multipath medium,
the placement of transmitters can be done such that the signal undergoes minimum losses. As discussed in earlier sections, we now
know that site measurements can become a tedious task, so propagation models are a low-cost alternative.
As it is the case in outdoor systems, there are several important propagation parameters to be predicted. The path loss and the
statistical characteristics of the received signal envelope are most important for coverage planning applications. The wide-band and
time variation characteristics are essential for the evaluation of the system performance.
3.2 Types of Propagation Models

The propagation models can be divided into three types:
• Empirical narrow-band models
• Empirical wide-band models
• Deterministic models
Figure 3.1 : Types of Propagation Models
Empirical narrow-band models are expressed in a form of simple mathematical equations which give the path loss as output. The
equations are obtained by fitting the model to measurement results. The empirical wide-band models (Dominant Paths) allow also
the prediction of the wide-band characteristics of the channel (e.g. delay spread).
Deterministic models are calculation methods that physically simulate the propagation of radio waves. These models yield both
narrow-band and wide-band information of the mobile radio channel inside buildings.
The calibration of these propagation models is done through measurements. These measurements have mostly been carried out at
1800 MHz which is most appropriate considering the common indoor systems. The usage of the models at other frequency bands is
possible when the material parameters of the propagation environment are known.
3.3 Required Database

The basis for any propagation model is a database that describes the propagation environment. Considering indoor propagation
suggests, for propagation modelling, that each building element should be categorized into classes (wall, floor, door, window,) and
specified by its coordinates and finally its material properties (thickness, permittivity, conductivity).
To get a more accurate description of wave propagation, the building data are stored in a 3D-vector format including all walls, doors,
and windows. All elements inside the building are described in terms of plane elements. Every wall is for instance represented by a
plane and its extent and location are defined by its corners Additionally, for each element individual material properties can be
considered. For efficient use, it is also possible to import .dxf-files, a very common data format in architecture.
3.4 Modified Free Space Model (MF):

The modified-free-space model computes the path loss by analyzing the distances between walls and penetration losses of the walls.
This model is like the Free-Space loss model but with an adjustable exponent and offset. This model does not consider the walls of
the building. Also, it does not take into consideration the individual positions of the walls and their material properties. Hence, no
database is required.
Figure 3.2 : Modified Free Space Model
The excess path loss (in dB) is linearly dependent on the distance. The path loss is given by
Where,
n= attenuation constant
With constant values for n and lC the prediction leads to field strength values decreasing in concentric circles around the transmitter.
Since these prediction results do not consider the position and materials of the walls, the results are inaccurate and only suited for a
rough estimation. The attenuation constant and the offset can be varied by using the Settings button to calibrate any model with
measurement results.
3.5 Motley-Keenan Model (MK):

The Motley-Keenan method computes the path loss based on the direct path between transmitter and receiver. In contrary to the
modified free space model this model considers the exact locations of the walls, floors and ceilings of the building considered.
Additional factors for absorption of the direct ray path by walls consider these shadowing effects.
Figure 3.3 : Motley-Keenan Model
lMK = l FS + lC + k wlw
Where,
Kw= number of walls intersected by the direct path between transmitter and receiver
Lw= transmission loss (material properties of individual walls are not taken into consideration)
This uniform transmission loss can be specified by using the Settings button.
3.6 COST- Multi-Wall Model (MW)

The multi-wall model gives the path loss as the free space loss including losses introduced by the walls and floors in the direct path
between transmitter and receiver. It has been observed that the total floor loss is a function of the number of penetrated floors. This
characteristic is considered by introducing an additional empirical correction factor.
Figure 3.4 : COST multiwall model (MW)
The individual penetration losses for the walls (depending on their material parameters) are considered for the prediction of the path
loss. Therefore, the multi-wall model can be expressed as follows:
N
lMW = l FS + lC + ∑ k wi lwi + k f l f
i =1
Where,
l FS = free space loss between transmitter and receiver,
lC = constant loss,
k wi = number of penetrated walls of type i,
kf = number of penetrated floors,
l wi = loss of wall type i,
lf = loss between adjacent floors,
N = number of different wall types.
The constant loss in the equation above is a term that results when wall losses are determined from measurement results by using
multiple linear regression. Normally it is close to zero. The third summand in the equation represents the total wall loss as a sum of
the walls between transmitter and receiver. For practical reasons in ProMan the individual wall loss of the intersected walls is
considered.
It is important to note that the loss factors in the formula are not physical wall losses but model coefficients which are optimized with
the measured path loss data. Consequently, the loss factors implicitly include the effect of furniture. This model has a low dependency
on database accuracy and because of this simple approach, it also has a very short computation time. However, waveguiding effects
are not considered with this model leading to a moderate accuracy level.
No pre-processing of the building data is needed for the computation of the prediction, and no settings must be changed for this
prediction model.
3.7 Standard Ray Tracing (SRT)

The standard ray-tracing model (SRT) performs a rigorous 3D ray-tracing prediction which results in very high accuracy but at the
cost of a large computational effort.
Deterministic models are used to model the propagation of radio waves physically. Therefore, the effect of the environment on the
propagation parameters can be considered more accurate. Another advantage is that deterministic models make it possible to predict
several propagation parameters. For example, the path loss, impulse response, and angle-of-arrival can be predicted at the same
time (Figure 1.1).
As smaller wavelengths (higher frequencies) are considered, the wave propagation becomes like the propagation of light. Therefore,
a radio ray is assumed to propagate along a straight line influenced only by refraction, reflection, diffraction or scattering which is the
concept of geometrical optics (GO). The criterion considered for this modelling approach is that the wavelength should be much
smaller in comparison to the extension of the considered obstacles, which are typically the walls of a building. At the frequencies
used for mobile communication networks, this criterion is also sufficiently fulfilled inside buildings.
There are two different basic approaches for the determination of the ray paths between transmitter and receiver in the geometrical
optics technique:
The so-called standard ray-tracing model (SRT) performs a rigorous 3D ray-tracing prediction which results in very high accuracy.
Due to the determination of the individual paths, the computational effort is very large. Therefore, several acceleration techniques
both with and without loss of accuracy are developed and integrated into this rigorous 3D approach.
This model has a long computation time because only small parts of the prediction are pre-processed, and every propagation path is
analytically determined. The initialization data is stored in a .idw file. This file is created at the first prediction run. The computation
time of subsequent predictions that are based on the new. idw database is thus reduced. For this purpose, a new project based on
the. idw database must be created.
3.7.1 Propagation Paths
The computed propagation paths can either be selected by using a selection file (*.sel) and choosing the option Only user defined
propagation paths. But normally, the option Number of interactions is used. Each transmission through a wall, each reflection at a
wall and each diffraction at an edge counts as an interaction.
The value Max. defines the maximum number of interactions that are allowed for each propagation path. The appropriate value
depends on the building structure. If, for example, the building has a corridor that runs around a corner three times, then it would be
better to compute more interactions, because multiple diffractions are needed to reach all prediction pixels. The same occurs if a
building has a structure where the rays must pass many walls to reach every point of the building, because then more transmissions
are needed.
The only constraint is that the computation time naturally increases if more interactions are computed. On the other hand, if too less
interactions are computed, the accuracy decreases. In this case additionally more prediction pixels might not be reached by the SRT
prediction, which leads to the need to compute those pixels with an empirical model which will decrease the accuracy even more.
As a basic rule, an appropriate setting for the Max. value would be 2 – 4, depending on the building structure.
3.7.2 Computation of Each Ray’s Contribution
For the computation of the rays, not only the free space loss has to be considered but also the loss due to the transmissions,
reflections and (multiple) diffractions.
This is either done using a physical deterministic model or using an empirical model.
Note: This does only affect the determination of the transmission, reflection and diffraction coefficients. The prediction itself always
remains a deterministic one, i.e. the same rays are considered.
The deterministic model uses Fresnel Equations for the determination of the reflection and transmission loss and the GTD/UTD for
the determination of the diffraction loss. This model has a slightly longer computation time and uses three physical material
parameters (permittivity, permeability and conductivity)
The empirical model uses five empirical material parameters (min. loss of incident ray, max. loss of incident ray, loss of diffracted ray,
reflection loss, transmission loss). For correction purposes or the adaptation to measurements, an offset to those material parameters
can be specified.
Herewith the empirical model has the advantage that the needed material properties are easier to obtain than the physical parameters
required for the deterministic model. Also, the parameters of the empirical model can more easily be calibrated with measurements.
It is therefore easier to achieve high accuracy with the empirical model.
The effect of bookshelves and cupboards covering considerable parts of walls is considered by including an additional loss of the
wall’s penetration (transmission) loss. An additional loss of 3 dB was observed to be appropriate. This additional loss is introduced
in the context of walls covered by bookshelves, cupboards or other large pieces of furniture. Furthermore, it was found necessary to
set an empirical limit for the wall transmission loss which otherwise becomes very high when the angle of incidence is large.
3.8 Intelligent Ray Tracing (IRT)

Ray optical models are very time consuming because all possible rays must be determined. A new approach to the acceleration of
ray optical models reduces the computation time to that of empirical models. This new method combines the advantages of both ray
optical models and neglects their disadvantages. It is based on a single reprocessing of the building database. All walls of the buildings
are subdivided into tiles and all wedges are subdivided into segments. The visibility relations between all tiles, segments and receiving
points in the database are computed in the pre-processing because they are independent of the transmitter location.
3.8.1 Propagation Paths
This deterministic model allows a very accurate rigorous 3D Ray Tracing prediction because many interactions can be considered.
The selection of propagation paths is similar to the method for the standard Ray Tracing (SRT) mentioned in section 3.7 That means
rays can either be selected by using a selection file (*.sel) and choosing the option Only user defined propagation paths or the option
Number of interactions is used. Therefore, each transmission through a wall, each reflection at a wall and each diffraction at an edge
counts as an interaction. Due to a preprocessing of the database, the IRT model has a very short computation time.
For the computation of the rays, not only the free space loss has to be considered but also the loss due to the transmissions,
reflections and (multiple) diffraction.
This is either done using a physical deterministic model or using an empirical model similar to the method for the standard Ray
Tracing (SRT) mentioned in section 3.7
Note: This does only affect the determination of the transmission, reflection, and diffraction coefficients. The prediction itself always
remains a deterministic one, i.e. the same rays are considered.
3.8.3 Pre-processing
In the preprocessing, each wall is divided into tiles and each edge is divided into segments (see Figure 4.7) These tiles and segments
are then used to determine the visibility relations between these elements which are needed to determine the propagation paths in
the prediction. For the determination of visibility relationships each tile and segment is represented by its center, while the
transmission, reflection or diffraction point can be an arbitrary point on the tile respective segment.
The size of the tiles and segments has an important impact on the preprocessing and the prediction. The following happens, if smaller
tiles respective segments are used:
• The accuracy of the prediction will increase because the rays are determined more precisely.
• The computation time (of the preprocessing and the prediction) will increase because more visibility relations must be
computed (at preprocessing time) and checked (at prediction time).
• The size of the database file after preprocessing will increase because more visibility relations must be stored.
The maximum size of the tiles and the segments can be specified. Appropriate values would be 1 – 3 meters for the tiles and 2.5 –
3.5 meters for the segments. The appropriate max. sizes depend on the structure of the building. The smaller the rooms, the smaller
the max. sizes should be chosen. The size of the segments is less critical than the size of the tiles. If the max. size of the segments
is chosen too large, the effects on the prediction are smaller. Therefore the max. size of the segments can be chosen higher than the
one for the tiles.
If a wall is smaller than the chosen maximum tile size or if an edge is smaller than the chosen maximum segment size then the wall
respective segment is treated as a whole part, but it is not combined with other walls respective edges.
Resolution Pre-processing Size of the pre-processed Prediction time

for the discretization time file
1m 7212 s 51.6 MB 110 s
1.5 m 1835 s 24.9 MB 100 s
2m 1095 s 12.2 MB 67 s
2.5 m 680 s 7.3 MB 16 s
3m 457 s 3.2 MB 1s
Table 1 : Dependency of preprocessing time, file size, and prediction time
Table 1 represents the dependencies of the pre-processing time, size of the pre-processing file and the prediction time for different
resolution values used for the discretization of the building database i.e. extension of the tiles and segments while the resolution of
the receiving points remains constant at 1 m (on a PC Pentium II 350 MHz with 128 MB RAM).
3.9 Dominant Paths (DP)

3.9.1 Prediction
Figure 3.4 illustrates the visibility relations computed in the pre-processing in the shape of a “visibility tree”. Only the relations in the
first layer of the tree must be computed in the prediction which can be done very fast, all other relations are determined in the pre-
processing and can be read from a file. The stored visibility relations (except the first layer) can be used for all predictions with the
same database.
Figure 3.5 : Field strength prediction with IRT (transmitting power 0.1 W at 1800 MHz)
Figure 3.5 shows a field strength prediction with the new approach after pre-processing of the building database. This prediction
indicates a typical behavior of ray optical modelling including waveguiding along the corridor. The difference between measurement
and prediction for a part of the building leads to good values for the mean error (1.9 dB) and the standard deviation (2.8 dB).
While for the prediction shown in Figure 3.5 a resolution of 1 m for the discretization of the building database has been selected it is
possible to increase this resolution to some extent without losing accuracy.
3.9.2 Multipath Propagation in Indoor Environments
In indoor scenarios, there are many possible rays between a transmitter and a receiver, as shown in Figure 3.6. On the one side, the
computation of all these rays is very time consuming, on the other, these rays are not time-invariant and depend on the accuracy of
the database.
If e.g. a door is opened, the determined rays (for the situation with a closed door) are no longer valid and so the predicted field
strength is wrong. Also, people and moved furniture are time—variant obstacles and cannot be included in the database of the
building. The second problem is if the positions of the walls and edges are not known exactly (they may vary inside a specified area)
some of the determined rays do not reach the receiver. In most databases the positions of the walls are not known as precisely as
they should be for a ray-tracing algorithm. So, a new approach for the
Figure 3.6 : Multipath propagation in an indoor scenario and dominant paths.
prediction of the field strength should not rely on the knowledge of all possible rays between a transmitter and a receiver. Only the
representative paths should be determined, and they should be independent of the accuracy of the database and of the time--variant
effects. These effects influence wave propagation, but there are many rays passing the same rooms between the transmitter and the
receiver with different reflection and diffraction points, so changes in these points of interaction do not influence the received power
It is only necessary to know which rooms are passed by the rays and which walls are penetrated. The waveguiding by multiple
reflections must not be determined by the computation of all possible reflection and diffraction points. It should be determined
independently of the different points of interaction. All these aspects are included in the approach with dominant paths which is
described in the following.
3.9.3 Principle of the Dominant Paths
If the rays in the left part of Figure 3.6 are analyzed, it seems to be obvious that most of them are nearly similar to one another. So,
the rays can be subdivided into different groups, each group defined by the following criteria:
• Similar sequence of rooms passed
• Transmissions through the same walls
It must be pointed out that in contrast to the deterministic models the number of interactions is not used for the classification of the
rays. It is not important how many reflections or diffractions occur along the ray -- more important is the sequence of rooms passed
and walls transmitted.
All rays, passing the same rooms and transmitting the same walls can be described by a representative dominant path. So, each
receiver point is reached by different dominant paths, passing the rooms and transmitting the walls in a different sequence.
This effect is pointed out with the example given in Figure 3.6. All rays reaching receiver R2 in the left part of Figure 3.6 can be
described by their corresponding dominant path in the right part of Figure 3.6. As these dominant paths have no reflection or diffraction
points but only points of changing directions, it is impossible to compute the field strength at their ends by using GTD/UTD. One
possibility for computing the field strength is empirical models, based on the regression of measurements. Another approach is
artificial neural networks.
One of the basic ideas of the new prediction model is the fast determination of the dominant paths. While in principle it is possible to
determine all rays with a ray tracing algorithm and then combining the different groups of rays to dominant paths, a new and very fast
algorithm is developed for the determination of the dominant paths. This algorithm is nearly as fast as the prediction with empirical
models.
3.9.4 Determination of the Dominant Paths
In a first step, the sequence of the transmitted walls and rooms passed by the dominant path must be determined. Therefore, an
analysis of the database is mandatory. In the database only information on walls and the material of the walls is given, no information
about rooms is available. So, rooms must be determined in a first step. After this initial step, a tree of the room.
Figure 3.7 : Tree of the room-structure of the building presented in the left part
The root of the tree corresponds to the room in which the transmitter is placed. The first layer contains all neighboring rooms. If there
are different walls between the transmitter room and the neighboring rooms (e.g. wall E and F between room 1 and 5), the neighboring
room is placed in this layer of the tree as many times as there are coupling walls between the rooms. After this first layer of the tree,
the second layer is determined similarly, i.e. all neighboring rooms (and coupling walls) are branches of the corresponding rooms of
the first layer. The tree contains as many layers as necessary for completeness, i.e. each room of the building must occur in the tree
at least once.
After the determination of the tree, the dominant paths between the transmitter and the receiver can be computed very easily, because
if the receiver is in room i, the tree must only be examined for room i. If room i is found in the tree, the corresponding dominant path
can be determined by following all branches back to the root of the tree. For example, the path to room 5 through wall E is highlighted
in Figure 3.7 to show the determination of the paths.
Figure 3.8 : Combination of rooms
The coordinates of the path are always computed with the same algorithm, independent of the number of transmitted walls and
passed rooms. This is only possible by combining all passed rooms into one room for the determination of the path. This is shown in
Figure 3.8 for a given dominant path. Rooms 1 and 5 are combined, erasing wall E (Names of the walls and rooms are given in Figure
3.7). Now the situation is like the situation where the receiver and the transmitter are in the same room. The solution for the
determination of the path inside a single room is described in the following.
There are two different cases for the determination of the dominant path between a transmitter and a receiver located in the same
room: Line of sight and obstructed line of sight (No line of sight is not possible because the rooms of the transmitter and the receiver
are combined to a single room as described above.
In the first case (Line of sight) the determination of the dominant path is very easy because it corresponds to the direct ray between
the two points. The second case (Obstructed line of sight) is a bit more complicated. In the first step, all the corners of the room get
a number and are arranged in two lists, one containing all convex corners and the second one all concave corners.
The concave corners are not used for determining the path and so the path between the two points must pass different convex
corners (at least one convex corner). For the corners of the room two trees are generated, as shown in Figure 3.8.
Figure 3.9 : Determination of dominant paths in a single room.
Two trees are necessary, one for the transmitter and one for the receiver. If there is a line of sight between the transmitter (receiver)
and a convex corner, this corner is put in the first layer of the tree. The second layer consists of all convex corners which are visible
(by the line of sight) from the corners in the first layer and so on.
The determination of the path is very easy now because it is only necessary to compare the trees for the transmitter and the receiver.
If there is the same corner number on the first layer in both trees, the path leads via this corner. If there are different numbers, the
second layer in both trees must be compared to the first layer in the other tree and if there occurs the same number, the path leads
via this corner and the corresponding corner in the first layer. If no path is found at this stage, the corners in the second layers of both
trees are compared. This is done till the same number in both trees is found.
In the last step the dominant path must be modified to be independent of the exact location of the corners, as mentioned above.
Therefore, the path is moved inside the room, depending on the angle of the corner and the distance to the next wall.
The option Layers deeper than optimal path layer determines how many layers deeper than the optimal layer of the room tree should
be used for the path determination. A minimum and a maximum value can be specified. Appropriate values for min are 1 – 3 (default
2), for max 2 – 4 (default 3).
The option Threshold for consideration of paths determines, how much dB less than the best path a path can have to be considered.
Appropriate values are 2 – 6 dB, the default value is 3 dB. If a path does not meet the criteria, it is neglected.
If the box Consider 3D path determination is checked (default), then a rigorous 3D path search is performed. Otherwise only a 2D
search is performed, leading to reduced accuracy.
If the box Consider outdoor paths is checked, additionally paths that go through the outside of the building and “reenter” the building
is also considered. Checking the box might increase the accuracy in some building structures where the propagation through the
outdoor area is an important part of the coverage, but it also increases the computation time.
3.9.5 Path selection criteria
The different path selection criteria parameters are described in Table2:
LP Allows to change the weighting of the path loss (free space loss) that was computed for
the rays. The default value is 1. A larger value increases the weight of the path loss.
LT Allows to change the weighting of the transmission loss (due to the walls) that was
computed for the rays. The default value is 1. A larger value increases the weight of the
transmission loss.
LI Allows to change the weighting of the interaction loss (due to diffraction) that was
computed for the rays. The default value is 1,5. A larger value increases the weight of the
interaction loss.
LW Allows to change the weighting of the computed waveguiding gain (due to long parallel
walls, e.g. in corridors) that was computed for the rays. The default value is 2. A larger
value increases the weight of the waveguiding gain.
nI This is an additional correction factor (> 0) which allows to consider an additional loss
depending on the number of interactions. Appropriate values are between 0 and 2.
n min This is an additional correction factor (> 0) which allows to consider an additional gain for
the pixels that are reached in the first layer of the room tree. This is suggestive if these
pixels are predicted with a too pessimistic value due to the building structure. Appropriate
values are between 0 and 2.
3.9.6 Share of the empirical and neural model
In the current version, predictions with a neural part > 0 % are not possible.
This empirical model allows an accurate prediction with moderate computation time. The dependency on the database accuracy is
low but complex room detection is required, which determines all existing rooms and corridors. The computation time is moderate,
because only small parts of the prediction are preprocessed, and every propagation path is analytically determined. Still, it is much
faster than the Standard Ray Tracing, as only the dominant paths are determined.
3.10 Indoor Scenario Example

Let’s see an example of Indoor propagation. The antenna pattern is imported using Aman. The structure of the environment is
constructed using WallMan. The model is then run using ProMan with different propagation models.
Main Exercise Start
Start Aman
1. Launch AMan interface
The main window is grey at this stage.
2. File / Open Antenna Pattern
3. Select router.aha
(aha = antenna pattern horizontal ASCII)
5. Select router.ava
(ava = antenna vertical ASCII)
6. Take a moment to look at the contents of the aha and ava files with a text editor. Note the simplicity: always easy to
generate.
Interpretation
7. Theta graph shows the elevation pattern,
i.e. a vertical cut.
θ=0 points vertically upward to the ceiling;
θ=180 points down to the floor.
8. Phi graph shows the azimuth pattern,
i.e. a horizontal cut.
Minimum gain > -1 dBi: good horizontal coverage in all directions.
Intermediate step: convert to the binary file
9. Select the window with the horizontal pattern
10. File / Save as…
11. Save in. ahb format (Antenna pattern Horizontal Binary)
12. Select the window with the vertical pattern
13. File / Save as…
Save in. avb format (Antenna pattern Vertical Binary)
Convert two 2D Patterns to one 3D Pattern
14. File / Convert 2x2D to 3D
15. Accept the default conversion algorithm.
16. When prompted, select the binary files you just produced, one at a time.
17. When prompted, specify the name of the 3D file.
apb = antenna pattern binary
Visualize and Exit
Explore the visualization options.

Hint: to rotate the pattern, depress both mouse buttons simultaneously and drag the mouse.
To zoom/unzoom, depress the right mouse button and drag (slowly).
File / Exit.
Main Exercise Start
Start WallMan
1. Launch WallMan interface
2. The main window is gray at this stage.
3. File / New Database
4. Indoor Database
Draw Manually
Browse to GlobalMaterialCatalogue
5. Click OK
Set Default Values
1. Define default wall height 3 m
2. Select default wall material Brick 10 cm
3. Select default sub-division material Wood 5 cm
The latter is for doors; we will modify it later for windows.
4. OK
5. After this window has closed, you can find it under
Edit / Default Values.
Grid
1. Settings / Local Settings
2. Specify Grid Size 0.25 m
3. Note several other optional settings:
Fill Objects
Display Thickness of Walls
Adjust Zoom Level
1. Activate Mouse Pan and Zoom Tool
2. Use mouse wheel to zoom in the 2D windows
until the grid is visible.
3. Note the coordinates at the bottom of the interface
as you move the mouse around in a window.
4. At the end, leave the XY window active.
Draw Outer Walls
1. Click with the left mouse button in the X/Y window to make it active.
2. Use the button “Add polygonal objects with orthogonal walls and a ceiling.” (Can also be found under Objects / Enter Basic
Objects)
3. With single left clicks in the X/Y window, enter
(-10, 6), (-5, 6), (-5, 0), (7, 0), (7, -5), (-10, -5)
With a single right-click, close the polygon at (-10, 6).
Draw Interior Walls
1. Ensure that the 3rd coordinate (the Z value) is at 0.0 m
2. Use the Add Objects Orthogonal icon (or the menu Objects / Enter Basic Objects / Orthogonal) to add interior walls.
Create your floor plan.
Explore Visualization
1. Temporarily activate the XZ or YZ window.
2. Access the 3rd button to move the “cut plane”
in the third coordinate.
3. Observe how the cross-section that is displayed changes.
4. Observe how the purple cut plane in the 3D window moves as you move the slider.
Add Doors
1. Edit / Default Values
2. Set upper and lower coordinate as shown, appropriate
for doors.
3. Verify that material for new subdivisions is set to Wood.
4. Using the same orthogonal drawing feature as before,
add several doors: interior doors, front doors.
Each door may require two single left clicks at the start point
and one single left click at the endpoint.
Add Windows
1. Edit / Default Values: adjust upper and lower coordinate
and adjust the material of the new subdivision.
2. Create several windows.
Add Glass Patio Doors
1. Set default lower coordinate to 0.
2. Use the same orthogonal drawing feature to add patio doors,
e.g. from (-2.5, 0) to (-0.5, 0) unless you already had a
window there.
Add Floor
1. Activate XY Window if not still active
2. Verify that the 3rd coordinate is set to zero.
3. Edit / Default Values: Choose Concrete 20 cm
4. Draw a rectangle under the house and the patio.
5. Filled images were obtained through Settings / Local Settings.
Save and Exit
6. File / Save As
Name rooms_1.idb
idb = indoor database binary
7. File / Exit
Main Exercise Start
Start ProMan
1. Launch ProMan interface
2. File / New Project
Indoor; browse to rooms_1.idb, the model you just created.
3. When prompted, set Height 1.0 m
(Height of a desk or a table where receiving computer maybe).
4. Model is Displayed
5. Cross-section at z=1.0 m.
6. The red rectangle is the suggested extent of the computational domain.
Setup Simulation Parameters
1. File / Save As
Browse to suitable directory; save as Indoor_Router.net
2. Project / Edit Parameter
This brings up a key panel with several tabs. We will visit the tabs one by one and set up the simulation. This will include
importing the antenna pattern.
Edit Parameters / Simulation Tab
1. Enter Prediction Resolution and Prediction Height as shown.
The Rectangle Area refers to the red rectangle around the model, i.e. the extent of the computational domain.
Edit Parameters / Propagation Tab
1. Specify an output directory. Here we call it prop01_MW as we plan to use the Multi-Wall model first.
2. Specify desired output quantities.
Edit Parameters / Sites Tab
1. Click Site to define initial
properties.
2. Specify location (4, -1, 2)
3. Under Antennas, click Add.
Edit Parameters / Sites Tab (ctd)

1. Specify the location of the antenna
(height z was adjusted)
2. Specify frequency 2400 MHz
3. Specify power 0.5 W
4. Under Antenna Pattern, change the selection from Omnidirectional to Directional / Sector Antenna
5. Browse to the file router.apb, produced earlier, and click OK.
Edit Parameters / Sites Tab (ctd)
Back at the Edit Parameters window / Sites tab, click Add.

The antenna we just defined is added to the simulation.
Edit Parameters / Database Tab
• Skip the Components Tab
• No changes at the Database Tab
• Edit Parameters / Computation Tab
The main choices are
- 3D Ray Tracing
Accurate but can be computationally expensive if many interactions are considered
- Dominant Path Model

Recommended; the right balance between accuracy and simulation time
- Multi-Wall Model
Fast but maybe too pessimistic when a ray traverses multiple walls. Rays that diffract around corners are not considered.
Choose Multi-Wall Model first (optionally inspect settings)

and click OK.
Verify Router Position
Display / 3D Display
This is a toggle button; choosing it again brings you back to 2D
Floor and Roof Visibility (for your information only)
Settings / Local Settings / Layout Tab
Indoor Database / Options
“Draw horizontal walls/objects”

checkbox governs visibility.
Launch Simulation
Computation / Propagation: Compute All or use the RUNPRO button on the left.
The simulation completes in a matter of seconds.
Result with Multi-Wall Model
• Straightforward: straight rays from the router; distance and number of walls determine field strength.
Try Dominant Path Model
Return to Project / Edit Parameters
Specify the new output directory
Specify Dominant Path Model
Computation / Propagation: Compute All
Result with Dominant Path Model
• Path Loss
• Note the effect of diffraction around a corner!
Result with Dominant Path Model, ctd.
• Field Strength with additional ray information loaded.
• Right-click and choose Rays: Display, then left-click points of interest.
Line of Sight Plot
• Light blue: “obstructed line of sight”. There is no wall blocking the ray,
but it’s not a pure line of sight either.
6. Display the Difference between two Sets of Results
1. With Field Strength of Dominant Path Model displayed and with that window active (click in the white space if needed),
choose Edit / Subtract Data / Value (File, linear).
2. Browse to the directory you defined for the Multi-Wall simulation (prop01_MW)
3. Select the .fpf file you find there and click Open
4. The field plot will now show the difference
Difference
Main differences: when dominant path involves diffraction around a corner, instead of crossing straight through walls.
This is the case in the red and purple areas: Dominant Path Model is reliable
while Multi-Wall Model is too pessimistic.
Simulate with 3D Ray Tracing
1. Create a new sub-directory for results
2. Select 3D Ray Tracing with default settings
3. Computation / Propagation: Compute All
Result with 3D Ray Tracing
Display the Field Strength.
Note that reflections are visible, especially in the room farthest from the router.
Accuracy
Whether the Dominant Path Model (DPM) or the ray-tracing model is more accurate depends on the situation.
Altair recommends using the DPM for pure coverage studies (without multipath effects) and using the ray-tracing model in case the
temporal or angular properties of the radio channel are of interest.
DPM has advantages for regions far away. See next slide.
Propagation Models: Comparison for a larger building
End
File / Save Project
File / Exit
https://altair-2.wistia.com/medias/u9ng9k47f5
3.11 Indoor Advanced Scenario Model Example
Main Exercise Start
Start AMan
1. Open router.pfs with POSTFEKO
2. Inspect the antenna pattern – does it make sense?
3. Run the script APA_Export2.lua
(Home tab; open with Script editor; click the Play button)
Script
1. Export .apa file for the 3D pattern
(specify the extension as well)
2. Close the script editor and exit POSTFEKO. The file is available on disk.
Start Aman
1. Launch AMan interface Main window is gray at this stage.
3. Select router_3Dpattern.apa (apa = antenna pattern ASCII)
Produce Binary File
1. File / Save As
Save in. apb format
(antenna pattern binary)
2. File / Exit
Main Exercise Start
Start WallMan
1. Launch WallMan interface Main window is grey at this stage.
3. Indoor Database
Draw with bitmap in background
Browse to GlobalMaterialCatalogue.mcb (Verify all selections before clicking OK)
OK
Default Values
1. Specify wall height 3 m
2. Specify wall material Concrete; 20 cm
(for outer walls)
3. Specify subdivision material Wood; 5 cm
(e.g. for doors)
4. OK
Note: this can later be modified through
Edit / Default Values.
Specify image
1. Floorplan.jpg
2. Width 60 m
3. Lock ratio
4. Apply and Close
5. The image appears in the XY window
and in the 3D window
Define Grid
1. Zoom in on the Z-shaped part of the building.
2. Settings / Local Settings:
- Grid Size 0.5 m
- Display Thickness of Walls
- Fill Objects
Start Drawing
1. Activate the X/Y window with a left mouse click.
2. Use the symbol Add Objects Orthogonal \ to enter this mode.
3. Define the outer walls one by one, usually with a single left-click at the start and endpoint.
(The first one with two single-left clicks at the start point)
4. File / Save Database As my_office.idb
Add Interior Walls
1. Edit / Default Values
Set material properties of new walls to Brick; 10 cm
2. You are still in “Add Objects Orthogonal” mode. If not, choose Objects / Enter Basic Objects
3. Enter Orthogonal Object.
4. Draw interior walls, one by one, in the XY view, as you did before for the outer walls.
Trace over doors; doors will be created next as subdivisions in walls.
3D image is shown for clarity.
Add Doors
1. Edit / Default Values. Set upper and lower coordinate as shown.
2. You are still in “Add Objects Orthogonal” mode. If not, choose Objects / Enter Basic Objects /
Enter Orthogonal Object.
3. Draw the first door.
4. Answer the question about Subdivision. OK
5. Repeat for the other doors.
Add Windows
1. Edit / Default Values Upper and lower coordinates as shown for windows.
Subdivision material Glass; 5 mm.
2. You are still in “Add Objects Orthogonal” mode. If not, choose Objects / Enter Basic Objects /
Enter Orthogonal Object.
3. Draw the first window (in the XY view).
4. Answer the question about Subdivision. OK
5. Repeat for the other windows.
6. File / Save
Add Floor
1. Make sure the 3rd coordinate in the XY view is at 0.00 m.
2. Make the view as large as possible for easy drawing.
3. Edit / Default Values: set material to Concrete: 20 cm.
4. Select Add Polygonal Objects and trace the shape of the floor surface. Single left clicks; close with a single right click.
XY View is filled (not shown here), because it shows the view at 0.00 m.
Increase the 3rd coordinate setting for the XY View.
5. File / Save Database As my_office_2.idb.
Add Next Level
1. File / Import / WinProp Binary Database.
2. Navigate to the most recent .idb file. Use material properties from the existing file.
3. Specify 3 m
4. Group walls? Yes
5. Move the 3rd coordinate a bit in the XZ and YZ windows to see the cross sections. Note we don’t have a roof yet.
Add Roof
1. XY Window: set 3rd coordinate to 6 m.
2. XZ or YZ window: select the floor with a left click.
It is highlighted.
3. Edit / Copy (or CTRL+C)
4. Edit / Paste (or CTRL+V)
5. Move the mouse in the XY window to position the roof, click when satisfied.
Save
1. Saving is not possible yet, because there are objects with duplicate numbers.
Edit / Renumber all Objects.
2. File / Save Database As my_office_3.idb.
Create Stairwell
1. XY View; 3rd coordinate 3 m
2. For a better visibility, disable “fill objects” under Settings
OR
disable the filling option for the floor material under Edit / Materials.
3. Draw a rectangle where the stairwell must be.
4. Hole in the surrounding wall – OK. An opening is created.
5. Exit rectangle mode by clicking on select-mode icon.
Create Staircase (optional)
1. Set the 3rd coordinate for XY View to 0.0.
2. Use LEGO tools to create stairs.
3.
4. After placing the object, it is selected. Use Objects / Rotate Selected Objects
and Objects / Move Selected Objects to place the staircase as desired.
5. Set the 3rd coordinate for the XY View at 1.5 m.
6. Create a second flight of stairs and rotate/move it.
7. File / Save Database As my_office_4.idb
8. File / Close.
Pre-Processing
1. The goal of pre-processing is to investigate and store visibility relations between many parts of the model, so propagation
simulations can later be done more efficiently.
2. This is only necessary for the Intelligent Ray Tracing (IRT) method.
4. Open the most recent office database you just saved.
5. Pre-Processing / Enter Pre-processing Parameters
6. Specify output file name, different from the office database you just loaded.
7. Specify IRT mode
8. Specify Prediction Area: Total
9. Continue to the next tab in this window
(don’t click OK yet)
10. Wedges tab  0.1 m Wedges shorter than this can be neglected.
IRT General tab  Specify as shown here.
1. Reduced Resolution Factor 2 is recommended.
2. OK.
3. File / Save Project As office_prep_IRT.pin
4. Pre-processing / Compute Current Project Duration a couple of minutes.
Exit
File / Exit
Main Exercise Start
Start ProMan
1. Launch ProMan interface. Main window is grey at this stage.
Indoor Scenarios;
3. check “Use only pre-processed databases” browse to office_prep_IRT.idi, the pre-processed database you just created.
4. When prompted, accept
Height 1.5 m.
Model is Displayed
• Cross section at z=1.5 m.
• Red rectangle is the suggested extent of the computational domain.
3D View
3D Toggle button near the top
Display Settings
1. Settings / Local Settings (Display)  Layout tab  Indoor Database – Options button
2. Verify that floors and roof are there
3. Toggle back to the 2D View afterward.
Save; then set up Simulation

1. File / Save Project As… my_office.net
2. Project / Edit Parameter Simulation tab
Set up Simulation, ctd.
1. Propagation tab:
- Create directory for IRT results
- Specify desired results
2. Sites tab:
- Add
- Site with sectors; one Cell/Sector; OK
Setup Simulation, ctd.
1. Specify location (X, Y, Z) = (16, 32, 0)
2. Click Edit button below Site 1 Ant 1
3. Specify height, frequency, power
4. Specify antenna pattern from earlier exercise.
5. OK OK Yes
6. Click Add to place another router
7. Name the second site “upstairs”
8. (X, Y, Z) = (26, 14, 3)
Z=3 is the height of the ceiling; 2.5 m will be added next.
9. Select “Site 2 Ant 1”; click Edit
10. Specify height (2.5 m), frequency, power
11. Select the router’s antenna pattern again (the .apb file)
12. OK Yes
Don’t click OK at the bottom of this window yet.
13. Skip the Components tab
14. Look under the Database tab
15. Move on to the Computation tab
Setup Simulation, Computation tab
1. Note “Compute always direct ray”
Every pixel will always be reached by at least one ray, even if it takes many transmissions.
2. Number of Interactions: try fewer than shown here and compare your results with results.
3. Have a look under Selection of Paths. These settings avoid wasted efforts.
Close this sub-window.
.
4. OK (confirm IRT settings and close this window)
5. Done with all tabs, OK to close the window.
Computation
1. File / Save
2. Computation / Propagation: Compute All or the RUNPRO button on the left.
Try Dominant Path Model

Return to Project / Edit Parameters.
1. Under the aPropagation tab, create a new directory for DPM results (DPM = Dominant Path Model)
2. Under the Computation tab, inspect settings for Dominant Path Model. Consult the manual for recommendations (the
manual contains a LOT of good information).
3. OK to close Edit Parameters window.
4. Computation / Propagation: Compute All
Accuracy
• Whether the Dominant Path Model (DPM) or the ray tracing model is more accurate depends on the situation.
• Altair recommends using the DPM for pure coverage studies (without multipath effects) and using the ray tracing
model in case the temporal or angular properties of the radio channel are of interest.
• DPM has advantages for locations far from the transmitter.
End
1. File / Save Project
2. File / Exit
https://altair-2.wistia.com/medias/115z7bfhv1
4 Wave Propagation Model for Urban Scenarios
4.1 Motivation
In urban environments, the propagated signal is subjected to a variety of obstacles that lead to multipath propagation. Obstacles like
tall buildings, signposts, traffic signals, etc after the transmitted signal. To characterize these multipath signals, we use empirical
methods.
In Urban propagation scenarios, the propagation path of a signal depends on the base station antenna height for the heights of
surrounding buildings. Various propagation models that work with the assumption of antenna placement over the rooftop. Methods
like Walfisch-Ikegami or COST 231 assume rooftop propagation.
Site-specific propagation methods can also be used for base station heights below as well as above rooftop levels. Such methods
include three-dimensional descriptions for the propagation environments.
4.2 Required Databases
An appropriate database for the propagation model considered is important for accurate prediction data.
The basis for any propagation model is a database that describes the propagation environment. Microcell and indoor propagation
modeling rely strongly on high-resolution geographical information. Most organizations are now using high-resolution databases of
the building structures in the urban area with an accuracy in the range of 1-2m derived from aerial photography measurements. For
urban propagation, it is essential to have accurate information at least about the average height of individual buildings, especially
when base stations are operating close to rooftop height.
Figure 4.1 : Building a database in vector format to describe the urban environment.
As microcells are planned to increase the network capacity in urban areas, it is obvious to use building-focused databases. To get a
more accurate description of wave propagation, the building data are stored in a vector format. Every building is modeled as a vertical
cylinder with a polygonal ground plane and a uniform height above street level. With this approach, only the propagation environment
including building vertical walls and horizontal flat roofs are considered. Also, the material properties (thickness, permittivity,
conductivity) of the building surfaces can be considered, which is important for the calculation of the reflection and diffraction
coefficients and the penetration into buildings.
Considering the influence of database information on prediction accuracy, it is noted in Damosso that prediction errors in microcells
of up to 15 dB were attributed to database inaccuracies arising from the poor resolution of the building data. Given the sensitivity to
the terrain surface, the terrain profile should be considered for propagation modeling if the considered area is not flat. Therefore,
terrain databases in pixel format are required with resolutions of about 20-30 m. This resolution is higher compared to terrain models.
4.3 Empirical COST-Walfisch-Ikegami Model
The empirical COST-Walfisch-Ikegami model considers only vertical plane propagation for faster prediction. Accuracy is reduced but
is acceptable in specific scenarios described.
The so-called empirical models (for example, the model according to Walfisch/Ikegami) consider only the propagation in a vertical
plane that contains the transmitter and receiver. For the field strength prediction, significant parameters must be extracted from this
vertical section (for example, average building height).
Equations containing these parameters must be optimized and fitted to numerous measurements to get a prediction model which is
applicable in different propagation environments. The main advantage of empirical models is their short computation time.
However, their prediction accuracy is limited because only a small number of parameters is considered and the influence of the
distance from the transmitter is over-emphasized. Additionally, waveguiding effects in streets cannot be considered with an empirical
approach.
The empirical model implemented in ProMan was developed in the course of the European COST 231 project by a combination of
the Walfisch and Ikegami models. The model allows for improved path loss estimation by consideration of more data to describe the
character of the urban environment, namely:
• height of the transmitter
• height of the receiver
• mean value of building heights
• mean value of widths of road w
• mean value of building separation b
• road orientation for the direct radio path
Figure 4.2 : Typical propagation situation in urban areas and the definition of the parameters used in the COST-Walfisch-Ikegami
model
However, this model is still statistical and not deterministic because only characteristic values are considered for the prediction. The
model distinguishes between line-of-sight (LOS) and non-line-of-sight (NLOS) situations. In the LOS case – between the base station
and mobile antenna within a street canyon – a simple propagation loss formula different from free space loss is applied. The calibration
of this formula is done by measurements performed in European cities.
LOS case:
where the first constant is determined in such a way that
is composed of the terms,
free space loss
multiple screen diffraction loss
rooftop-to-street diffraction loss
NLOS case:
The free space loss is given by:
The term describes the coupling of the wave propagation along a multiple screen path into the street where the mobile station is
located. The determination of is mainly based on Ikegami’s model. It considers the width of the street and its orientation. COST
231, however, has applied another street-orientation function to improve on the Ikegami model:
The orientation loss is an empirical correction factor gained from measurements:
Scalar electromagnetic formulation of multi-screen diffraction results in an integral for which Walfisch and Bertoni published an
approximate solution for the case of base station antennas located above rooftops. This model is extended by COST 231 for base
station antenna heights below the rooftop levels using an empirical function based on measurements. The heights of buildings and
their spatial separations along the direct radio path are modeled by absorbing screens for the determination of :
where:
4.4 Urban Scenario Example
Main Exercise Start
Start Aman
1. Launch AMan interface
2. File / Open 2D Antenna Pattern
Import Example01.msi
It contains both horizontal and vertical data.
Convert two 2D Patterns to one 3D Pattern
1. File / Convert MSI to 3D
(for imported 2D patterns in other formats,
use File / Convert 2x2D to 3D)
2. Accept the default conversion algorithm.
3. Specify the original .msi file once more
(this is because you can have multipy loaded)
4. Specify the destination file name
.apb = antenna pattern binary (3D)
Visualize, Exit
1. Explore the visualization options.
Hint: to rotate the pattern, depress both mouse buttons simultaneously and drag the mouse.
2. File / Exit.
(You had already saved the .apb file).
Main Exercise Start
Start WallMan
1. Launch WallMan interface Main window is grey at this stage.
2. File / Convert Urban Database / Vector Database.
Convert
1. Set filter to AutoCAD
2. Accept coordinate options as shown
3. OK
4. Browse to Frankfurt.dxf and select it.
5. Save as Frankfurt.odb when prompted.
6. Select the desired layers
7. Click Convert to Walls
The layer gets a wall symbol in front of it.
8. Start Conversion
9. Ignore 2D objects, OK
10. OK when prompted
Note Settings
1. Settings / Local Settings
2. Fill Objects
3. Display Building Heights in Status Bar
Note you can zoom / unzoom with the mouse wheel.
4. Select and Modify
5. Select a building by left-clicking on an edge of the building in the 2D display.
6. Access the properties through right-click  Properties
7. Changing the z coordinate of one corner is enough
to change the height of the entire building.
8. Note the ability to change building type
(leave it Standard Building)
9. Note the ability to import selected materials
from the file GlobalMaterialCatalogue.mcb.
Note on Materials
• Note the ability to change materials for many buildings simultaneously.
• Note ability to left-click-and-drag a building corner to change the building shape.
1. Add a Courtyard or a Tower
2. Use the Polygon icon (on the left edge) to draw a polygon within an existing building (single left-clicks ending with one
right-click).
3. Courtyard or Tower  Yes
4. Change from Polygon mode to Select-Object mode
5. Select the new object and access its properties.

6. Addition becomes a tower by choosing a large z-coordinate
for any of the corners.
7. Addition becomes a courtyard by choosing a small z-coordinate
for any of the corners.
8. In the latter case, turn off Fill Objects (Settings / Local Settings), otherwise the courtyard won’t be displayed, even though
it will be recognized and locally take priority.
9. Choose whether to create a tower or a courtyard.
10. File / Save as Frankfurt_modified.odb.
Add Vegetation
1. Objects / Enter Vegetation
2. With the polygon drawing tool, add a polygon where there is room for vegetation.
3. Switch to Select-Object mode and select the new polygon by single-left-clicking on an edge (you may need to try more
than once).
4. Right-click and access Properties.
5. Verify its indeed vegetation.
6. Inspect the Default Vegetation material properties.
Save and Exit
1. File / Save
2. File / Exit
Main Exercise Start
Start ProMan
Urban; browse to Frankfurt_modified.odb, the model you just created.
3. When prompted, accept Height 1.5 m
Save
• File / Save Project as Frankfurt.net
1. Add Site
2. Use the Set Site icon on the left to define a new antenna site.
3. Choose 3-Sector Site.
4. OK
5. With the cursor, click on the tall building shown below on the map and left-click.
Add Site, ctd.
1. Specify z=135.84 m Edit antennas one by one to adjust frequency, pattern, tilt
2. Frequency 1800 MHz
3. Down tilt 4 degrees
4. Antenna pattern: Select .apb file from AMan exercise
5. OK
6. Repeat for antennas 3 and 4 of this site.
Add Second Site
1. Still in Add-Site mode, click on other building,
2. This building has height 30.11 m
3. Specify sector antennas as before
4. Once the site has been completely defined,
close the windows with OK
5. Press Esc to get out of Add-Site Mode.
Set up Simulation
A window with several tabs comes up; we will visit them one by one.
2. Under the Simulation tab, accept most defaults.
3. Optionally, set the Resolution to 2 m.
Note: you can choose the Prediction Height; it doesn’t have to be the same as the Display Height.
Propagation tab
1. Specify the output directory
2. Specify desired output
Sites tab
• Sites had already been defined.
We can disable sites if desired.
Building Data tab
• Accept the defaults
Computation Tab
1. Dominant Path model is recommended.
2. Optionally, add Indoor Coverage.
3. OK
Run Simulation
1. Computation / Propagation: Compute All or the RUNPRO button on the left.
2. Inspect Results, ctd.
3. To view fields inside the buildings, use Settings / Local Settings (Display of Data)
4. Choose the Layout tab
5. Choose Urban Database Options
6. Un-check Fill Buildings
Display / 3D Display
• Buildings filled
• Note the 3D View confirms that antennas are at the desired altitudes (blue symbols).
Other Method
1. Try the Extended Walfisch- Ikegami Model.
• This is valid when the transmitter is high and the dominant propagation mechanism is over rooftops, followed by
diffraction down into the street.
• A wave that is guided along a street between buildings would not be included.
• Simulations tend to be fast.
End
2. File / Exit
https://altair-2.wistia.com/medias/u4o5adz4ee
4.5 Basics of Ray-Optical Models

Knowing the terminology and concepts of ray-optical methods aids the selection of an appropriate ray-optical technique regarding
accuracy and prediction time.
Primary Criterion for Ray-Optical Models
The mobile radio channel in urban environments is characterized by strong multipath propagation. Dominant propagation
mechanisms in these scenarios are reflection, diffraction, shadowing by discrete obstacles and waveguiding effects in street canyons.
With a ray-optical approach, it is possible to consider these effects in a propagation model.
As smaller wavelengths (higher frequencies) are considered, the wave propagation becomes like the propagation of light. A radio ray
is assumed to propagate along a straight line influenced only by refraction, reflection, diffraction or scattering. These are the concepts
of geometrical optics (GO). The criterion considered for this modeling approach is that the wavelength should be much smaller in
comparison to the extent of the considered obstacles (buildings for urban environments). At the frequencies used for mobile
communication networks, this criterion is sufficiently satisfied.
4.5.1 Specular Reflection
Specular reflection phenomena are the mechanism by which a ray is reflected at an angle equal to the incident angle. The reflected
wavefields are related to the incident wavefields through a reflection coefficient which can be expressed as a matrix when the full
polarimetric description of the wavefield is considered. The most common expression for the reflection is the Fresnel reflection
coefficient which is valid for an infinite boundary between two mediums, for example, between air and concrete.
The Fresnel reflection coefficient depends on the incident wavefield and upon the permittivity and conductivity of each medium. The
application of the Fresnel reflection coefficient formulas is popular, and these equations are also applied in ProMan.
To calibrate the prediction model with measurements some ray-optical software tools consider an empirical reflection coefficient is
varying with the incidence angle to simplify the calculations. Such an empirical approach is also available in ProMan.
4.5.2 Diffraction
The diffraction process in ray theory is the propagation phenomena that explain the transition from the illuminated region to the
shadow regions behind a corner or over rooftops. Diffraction by a single wedge can be solved in various ways: empirical formulas,
perfectly absorbing wedge, geometrical theory of diffraction (GTD) or uniform theory of diffraction (UTD). The advantages and
disadvantages of using either formulation are difficult to address since it may not be independent of the environments under
investigation. However reasonable results are possible with either formulation. The various expressions differ mainly in the
approximations being made on the surface boundaries of the wedge under consideration. One major difficulty is to express and use
the proper boundaries in the derivation of the diffraction formulas. Another problem is the existence of wedges in real environments
- the complexity of a real building corner or the building’s roof illustrates the modeling difficulties.
Despite these difficulties, however, diffraction around a corner or over rooftops is commonly modeled using the heuristic UTD formulas
since they are well behaved in the illuminated/shadow transition region, and account for the polarization as well as for the wedge
material. Therefore, these formulas are also used in ProMan to calculate the diffraction coefficient.
4.5.3 Multiple Diffraction
In the case of multiple diffractions, the complexity increases dramatically. In the case of propagation over rooftops, the result of
Walfisch and Ikegami has been used to produce the COST-Walfisch-Ikegami model. One method frequently applied to multiple
diffraction problems is the UTD. The main problem with straightforward applications of the UTD is in many cases one edge is in the
transition zones of the previous edges. Strictly speaking, this forbids the application of ray techniques, but in the spirit of the UTD,
the principle of local illumination of an edge should be valid. At least to some approximate degree, a solution can be obtained which
is quite accurate in most cases of practical interest.
The key point in the theory is to include slope diffraction, which is usually neglected as a higher order term in an asymptotic expansion,
but in the transition zone diffraction, the term is of the same order as the ordinary amplitude diffraction terms. ProMan utilizes the
UTD including slope diffraction for over rooftop propagation for the calculation of multiple diffractions. Also, there is an empirical
diffraction model available which can easily be calibrated with measurements.
4.5.4 Scattering
Rough surfaces and finite surfaces (thus surface with small dimensions regarding the wavelength) scatter the incident energy in all
directions with a radiation diagram which depends on the roughness and size of the surface or volume. The dispersion of energy
through scattering means a decrease of the energy reflected in the specular direction.
One can account for the scattering process by simply decreasing the reflection coefficient. You can do this by multiplying it by a factor
smaller than one which depends exponentially on the standard deviation of the surface roughness according to the Rayleigh theory.
To consider the true dispersion of radio energy in various directions, one can also specify the additional empirical scattering loss or
a polarimetric scattering matrix. For ease of use, this specification is for 30 degrees off-specular only, and values for other angles will
be derived from it. This will spawn extra rays in non-specular directions, thus increasing the computational cost
4.5.5 Penetration and Absorption
Penetration loss due to building walls has been investigated and found very dependent on the system. Absorption due to trees or
absorption by the human body is also propagation mechanisms difficult to quantify with precision. Therefore, in the radio network
planning process, adequate margins should be considered to ensure overall coverage.
Another absorption mechanism is the one due to atmospheric effects. These effects are usually neglected in propagation models for
mobile communication applications at radio frequencies but are important when higher frequencies (for example, 60 GHz) are used.
4.5.6 Ray Tracing Versus Ray Launching
Ray-optical propagation models are often used for the prediction of field strength in urban scenarios. They are very accurate because
they consider waveguiding effects in street canyons and they include diffraction at wedges. There are two different approaches to
determine the ray-optical propagation paths between transmitter and receiver:
• Ray tracing
• Ray launching
Ray tracing computes all rays for each receiver point individually and guarantees the consideration of each ray wall as well as a
constant resolution. In the ray launching approach, the rays are launched from the transmitter with a constant angular increment.
Thus, ray launching might neglect a wall because it is very small and located between two rays. Different approaches to a better
resolution with ray launching were proposed recently, but Ray Launching has still kept the disadvantage of a variable resolution
depending on the distance to the transmitter. The only advantage of ray launching is the shorter computation time compared to
standard ray-tracing algorithms (This is not valid for intelligent ray tracing, which is much faster than ray launching).
• Regardless of the model that is used for the determination of the rays between the transmitter and the receiver, the
received power must be calculated by the superposition of all contributions. WinProp offers two options, uncorrelated
(a statistical summation of contributions based on power considerations) and correlated (taking the phases into
account). The latter is important in some
• multipath situations but may often be undesired 0in urban scenarios because it will include the small- scale fading.
Figure 4.3 : Deterministic propagation models.
4.6 Ray Tracing

4.6.1 Acceleration
ProMan offers the possibility to accelerate the prediction by canceling the search for rays as soon as the accumulated prediction
value reaches the free space loss. The consequent prediction error is small because the received power is rarely above the received
power in the free space case. This might be only the case very near to the transmitter in a LOS situation with additional waveguiding
effects.
Mathematical background and equations show how WinProp calculates the contribution to the prediction for each ray.
For the computation of the rays, not only the free space loss has to be considered but also the loss due to the reflections and (multiple)
diffractions. This is either done using a physical deterministic model or using an empirical model
The deterministic model uses Fresnel equations for the determination of the reflection and transmission loss and the GTD/UTD for
the determination of the diffraction loss. This model has a slightly longer computation time and uses three physical material
parameters (permittivity, permeability and conductivity).
The empirical model uses five empirical material parameters (minimum loss of incident ray, maximum loss of incident ray, loss of
diffracted ray, reflection loss, transmission loss). For correction purposes or the adaptation to measurements, an offset to those
material parameters can be specified.
The empirical model has the advantage that the needed material properties are easier to obtain than the physical parameters required
for the deterministic model. Also, the parameters of the empirical model can more easily be calibrated with measurements. It is,
therefore, easier to achieve high accuracy with the empirical model.
Both diffraction models are based on the angles shown in Figure 4.4
Figure 4.4 : Angles for diffraction on a wedge
For the empirical diffraction model, the loss LBLB of the diffracted rays is computed depending on the angles φ and φ' using the
following equations
Φ−Φ'
Figure 4.5 : Empirical variation of the diffraction loss depending on the angle
The angle dependencies are derived from the uniform diffraction theory (UTD) by the evaluation of measurements with different
materials (brick, concrete) in an anechoic chamber and can be varied with the parameters abmin, abmax and akaus within appropriate
limits. With these three parameters, the model can be calibrated with measurements.
4.6.3 Breakpoint
In free space, there is a reverse proportional relation between the square of the distance from the transmitter to the
receiver and the power at the receiver ( is the propagation factor):
1
A0 = 2
This dependency of the received power is not valid for longer distances at the line of sight in urban scenarios. Starting
d0
from a certain distance, the relation gets (approximately) reverse proportional to the distance with exponent 4. The position of this
so-called breakpoint can theoretically be deduced from a two-path model, but different authors report that measurements contradicted
these results.
To account for different propagation scenarios and to allow the user to manipulate the free space loss computation, the above
conditions were considered by an extension of the equation:
1
A0 = p1
d0
With this approach, a smooth transition of the free space loss is ensured. In general, the breakpoint distance depends on the
transmitter height, the height of the antenna at the mobile station and the frequency. The parameter BP is the breakpoint distance
that is set to a default value according to the following
E.g. for transmitter heights of approx. 5 m, receiver heights in the range of 1.5 m and a frequency of 900 MHz, a breakpoint distance
of about 280 m can be expected. The default value derived from equation (30) can be corrected with the option (additional) breakpoint
offset within a ProMan dialog.
The parameters p1 (exponent before breakpoint) and p2 (exponent after breakpoint) can also be set by the user. The default values
are p1 = 2.0 and p2 = 4.0.
4.7 Description of the Individual Ray-Optical Models
4.7.1 Standard 3D Ray Tracing (SRT)
This model performs a rigorous 3D ray-tracing prediction which results in very high accuracy, but computation time is increased.
Due to the determination of the individual paths, the computational effort is very large. Therefore, several acceleration techniques
both with and without loss of accuracy are developed and integrated into this rigorous 3D approach.
4.7.2 Intelligent Ray Tracing (IRT)
This unique model performs a rigorous 3D ray-tracing prediction which results in very high accuracy and due to the preprocessing of
the database in a very short computation time.
To cope with the excessive computation times for the deterministic approach with ray tracing the following aspects have been
investigated:
• The deterministic modeling of pathfinding generates many rays, but only a few of them deliver the main part of the
received electromagnetic energy.
• The degree to which visibility relations between walls and edges are independent of the position of the base station.
• The number of cases where neighboring receiving points are hit by rays for which the paths differ only slightly. For
example, every receiving point in a street orthogonal to the street in which the transmitter is located is hit by a double
reflected, single vertically diffracted ray and the points of interaction of these rays with walls and edges are
independent of the receiver position.
Based on these considerations it is possible to accelerate the time-consuming process of ray pathfinding by a single intelligent
preprocessing of the database for buildings.
Figure 4.6 : Rays between transmitter and receiver determined by ray-tracing
Pre-processing
In the first step, as indicated in Figure 4.7, the walls of the buildings are divided into tiles and the edges into horizontal and vertical
segments.
After this discretization of the database, the visibility relationships of the different elements are determined and stored in a file. For
this process, all elements are represented and stored using their center positions. This leads to a simplification of the problem of ray
pathfinding since possible interaction points are only determined by the center positions of the tiles and segments.
The visibility relations between each tile (segment) and all other tiles (segments) are computed in preprocessing because they are
independent of the transmitter and receiver locations.
For the decision about the visibility relations, the line of sight criterion between the center positions of the tiles (or segments) is
evaluated. If there is a line of sight between the center positions, the rays from the center positions of the first tile to the corners of
the second tile are determined. Then the projection of the angles of the rays on the first and second tile is stored together with the
visibility relation.
A similar computation for the visibility relations between tiles and segments and between segments and other segments is performed
and stored in the pre-processing file
Figure 4.7 : Subdivision of the building database into small elements.
The angles of the projection are very important because they define a range of possible reflection (or diffraction) angles for the
illuminated tile (or segment). The angle also continues onto the neighboring tile resulting in a very accurate prediction of the rays
even if the tiles or segments are large - up to 50 or 100 meters for urban databases
A further improvement is possible if the grid of the prediction points is also used in the preprocessing because the prediction plane
can be subdivided into tiles and the visibility relations between the tiles of the prediction grid and the tiles (and segments) of the walls
represent the last part of the ray in the direction to the receiver.
If the receiver visibility relations are determined in the preprocessing, the only remaining visibility relations to be computed in the
prediction are the ones from the transmitter to the tiles (of walls and prediction grid) and segments. Figure 4.8 depicts the visibility
relations between a tile and a receiving point. For the calculation of the angles, the connecting straight lines between the receiving
point and the four edges of the tile are considered.
By projecting these four lines onto the XY plane and additionally into a plane perpendicular for the inspected wall, four angles are
determined which give an adequate description of this visibility relation.
Memory Requirement and Computation Time Table 3 shows the memory requirements for the preprocessed database file and
computation times for different urban scenarios. They were all computed with a maximum extension of the tiles and segments of 50
meters and a prediction resolution of 10 meters.
Example Area [km2] Number of Buildings File Size [MB] Relative

Computation
Time
Munich 8 2000 50 100%
Stuttgart 4 300 18 4%
Lille 1 86 6 0.60%
Table 3: Memory requirements and relative computation times for the pre-processing.
These computation times are shorter than the computation times of a single prediction for the same area using the standard ray-
tracing because each visibility relation is only computed once in the preprocessing while in the prediction with standard ray tracing
similar visibility relations might be considered and computed for many prediction points.
Prediction
The result of the preprocessing of the building database is a tree structure containing tiles, segments and receiving points of the
prediction area. In this tree, every branch symbolizes a visibility relationship between two elements.
For the prediction only the tiles, segments and receiving points, which are visible from the base station, must be determined. Also,
the angles of incidence for the visible tiles and segments must be calculated. Subsequently, pathfinding can be done like the ray
launching algorithm by recursively processing all visible elements and checking if the specific conditions for reflection or diffraction
are fulfilled. The ray search is stopped if a receiving point or a given maximum number of interactions is reached. Finally, the field
strength is summed up at all potential receiving points.
The preprocessing of the building database reduces the time-consuming process of pathfinding to the search in a tree structure. A
comparison between the number of branches in the first layer (determined in the prediction) with the number of branches in the
remaining layers (determined in the preprocessing) in the tree structure given in Figure 4.8 indicates the relation between the
computational effort in the prediction and the computational effort in the preprocessing.
Figure 4.8 : Determination of ray paths by searching in a tree structure.
Due to the small number of visibility relations in the first layer of the tree, the computation times are very short. Most of the time is
spent reading the visibility data from a file. If more than one the transmitter is considered at the same time; the preprocessed data
must only be read once and thus the prediction of the second transmitter is even faster than the prediction of the first transmitter
because the visibility tree is already loaded into memory.
Low Dependency of Prediction Area
The computation time for the prediction is nearly independent of the size of the prediction area because the whole tree is computed
once for each prediction and all receiver points are included in the prediction. Only the time for computing the field strength is
necessary for the evaluation of a prediction point and this time depends on the location of the point (inside or outside the prediction
area). The rays to all preprocessed receiver points are always determined in each prediction.
Therefore, if the size of the prediction area is reduced, a smaller number of prediction points must be considered (more preprocessed
prediction points are neglected) and the time for computing the field strength is reduced. But this part of the computation time is very
short compared to the time for determining the rays and therefore the total computation time is nearly independent of the size of the
prediction area. The number of interactions influences the computation time because each new interaction corresponds to a further
layer in the visibility tree.
Results
Very good results are achieved with a maximum number of three interactions (reflections and multiple diffractions in different
combinations with a maximum number of two diffractions in each ray).
Table 4 shows the computation times for the different urban scenarios using intelligent raytracing. The computation times are
compared to the ones of an accelerated 3D standard ray-tracing model.
Table 4: Relative computation times for the two different ray-tracing algorithms.
Number of Relative Relative

Example Area km2
Buildings Time, IRT Time, SRT
Munich 8 2000 0.30% -
Stuttgart 5 300 0.05% 100%
Lille 1 86 0.01% 20.80%
Figure 4.9 : Prediction of the received power with intelligent Ray tracing
A comparison of this prediction with preprocessing of the database on the one hand with a prediction calculated with the standard
3D raytracing, on the other hand, leads to very small differences.
2 x 2D Modelling
From the many transmitter-to-receiver propagation paths, the most dominant ones must be selected to obtain the total received power
with moderate computation time. A useful acceleration of the process of ray pathfinding under the consideration of the main
propagation mechanisms is the limitation to two orthogonal planes (double 2D). Rooftop diffracted paths are included in the vertical
plane approach, while for buildings the diffracted paths are modeled with the transverse plane approach. The propagation in both the
vertical and the transverse plane is considered in two dimensions. However, the determination of the building corners in the transverse
plane is not necessarily performed in a horizontal plane. This principle can also be considered for the intelligent ray- tracing approach
which leads to the following two 2 x 2D models included in ProMan:
2x2D (2D-H IRT + 2D-V IRT)
The preprocessing and the determination of propagation paths are done in two perpendicular planes. One horizontal plane (for the
waveguiding, including the vertical wedges) and one vertical plane (for the over rooftop propagation including the horizontal edges).
In both planes, the propagation paths are determined similarly to the 3D-IRT by using ray optical methods. This approach neglects
the contributions from reflections at the building walls which are in most cases only relevant for the streets with LOS to the transmitter.
2x2D (2D-H IRT + COST231-W-I)
This model treats the propagation in the horizontal plane in the same way as the previously described model, thus by using ray-
optical methods (for the waveguiding, including the vertical wedges). The over rooftop propagation (vertical plane) is considered by
evaluating the COST 231-Walfisch-Ikegami model. By using this model, only the propagation in the horizontal plane is determined
by ray-optical methods considering the vertical wedges of the buildings, while the over rooftop propagation is modeled by an empirical
approach.
Due to the restriction to two orthogonal planes, this approach has better computational performance but at the cost of a slightly
reduced accuracy - not all buildings are considered for the determination of ray paths.
Figure 4.10 : Ray tracing within a vertical and a transverse transmitter-receiver plane
4.8 Combined (Urban + Indoor) Propagation-Hybrid Prediction

In a straightforward way, you achieve the best of both worlds: the speed of an urban simulation combined with the detail of an
indoor simulation. A typical application is cellular coverage inside a building of interest, while several other buildings of lesser
interest are in the vicinity, possibly partly obstructing the line of sight to the base station.
4.9 Urban + Indoor Scenario Example
Main Exercise Start
Start Aman
1. Launch AMan interface. The main window is grey at this stage.
2. File / New 2D Antenna Pattern Empty polar graph appears.
Import Bitmap and Position, it
1. Bitmap / Import Bitmap
2. Browse to Azimuth.bmp and import it.
3. Bitmap / Move Bitmap
Left-click-and-drag center of bitmap to center of AMan polar graph.
Adjust Bitmap Further
1. Bitmap / Zoom factor for bitmap…
2. Adjust zoom factor until the outer circle of bitmap coincides with the outer circle of AMan graph.
Access this menu multiple times if needed.
Adjust Scale
1. Click Settings
2. Specify the bitmap’s scale:
maximum, minimum gain; step size
Even if bitmap is in dB, no need to check “Logarithmic scale”
Define Points
1. Right-click in the main window
“Select type of antenna pattern” window pops up
2. Specify Horizontal in case of azimuth pattern, click OK
3. Through many single right mouse clicks, trace the pattern accurately.
The cursor should look like a + sign. If not, just move it a little.
4. Pattern always automatically forms a closed polygon; just stop right-clicking when done.
5. File / Save As For horizontal pattern, the file extension is. ahb (Antenna Horizontal Binary).
6. If desired, correct existing points with left-click-and-drag;
insert extra points with one extra right click at the desired location;
delete existing point by right-click on it and confirm Delete.
7. Save.
Next 2D Pattern
1. Repeat the process: File / New 2D Antenna Pattern
2. Bitmap / Import Bitmap  Elevation.bmp
3. Bitmap / Move
4. Bitmap / Zoom factor
5. Settings
Define Points
1. Right-click  select Vertical pattern  OK
2. With many single right clicks, trace the pattern approximately.
3. File / Save As For vertical pattern, *.avb
Produce 3D Pattern
1. File / Convert 2x2D to 3D
2. Accept the default conversion algorithm (HPI)
3. When prompted, select the .avb file and select the .ahb file you just saved to disk.
4. When prompted, specify a name for the 3D pattern (*.apb) and press Save.
This pattern can be used by ProMan.
Visualize and Exit
1. Explore the visualization options
Hint: to rotate the pattern, depress both mouse buttons simultaneously
and drag the mouse.
2. File / Exit
Main Exercise Start
Start WallMan
1. Launch WallMan interface
The main window is gray at this stage.
3. Urban Building Database
Draw with bitmap in background
Browse to GlobalMaterialCatalogue
(Verify all selections before clicking OK)
OK
Intermediate Steps
1. Coordinate System not important in this case.
It is only required for geo-referencing, which is not needed here. Type a zone number and click OK.
2. Specify default materials for new buildings and vegetation.
Select Image
1. When prompted, browse to the image file.
2. Specify estimated actual width. Note the “lock ratio” button.
Image can also be scaled later.
3. Apply.
4. Close.
Image appears in WallMan
We will use the left image to draw a few buildings and vegetation.
Option to Scale After Import
1. Edit / Scale
You can draw a line with a known length on the image to adjust all sizes.
2. Try this: a certain building wall is known to be 20 m wide (line in WallMan shows up without
arrows).
Draw Buildings
1. Zoom in if desired
2. Use “Add Polygonal Objects” to trace several building outlines.
One left clicks per corner, one right click when closing the polygon.
3. After drawing, leave Add-Polygon mode by going to Select mode.
4. Adjust building heights by selecting them, then right-click and choose Properties.
5. File / Save As Office_Park.odb odb = outdoor database binary
Add Vegetation
1. Objects / Enter Vegetation
ensures that the next object(s) will be vegetation, not buildings.
2. Draw polygons around rows of trees.
3. When done, return to the
Select mode.
4. Adjust heights of vegetation.
5. File/Save.
The model so far
• Database is ready for a standard urban simulation.
• However, we want to add an indoor database near the bottom of the image.
• Hybrid Urban/Indoor:
high accuracy where needed, high simulation speed elsewhere.
Import Indoor Database
1. Objects / Import Indoor Database
2. Browse to the *.idb file you produced in the first exercise.
3. Follow prompts and insert it in the location.
The windows and doors, although not visible here, are included.
Finalize
1. With Objects / Rotate and Objects / Move, place the indoor database nicely between the trees.
No need to use Objects / Scale.
2. File / Save
3. File / Exit.
Main Exercise Start
Start ProMan
Urban; browse to OfficePark.odb, the model you just created.
3. When prompted, accept Height 1.5 m
Save
• File / Save Project as OfficePark.net
Add Site
1. Use the Set Site icon on the left to define a new antenna site.
2. Choose 3-Sector Site.
3. OK
4. With the cursor, click on the building shown below on the map and left-click.
5. Specify height 5 m (the roof is 5 m high; Tx height is always relative to the ground, so this specification gets the Tx on the
roof).
6. Edit the antennas one by one:
- 4 m high
- 1800 MHz
- For Antenna Pattern, select the file sector. Apb from the AMan exercise.
Setup Simulation
2. Simulation Tab
3. The computational red rectangle, which you saw earlier, will be enlarged (at the end of the set-up).
4. Propagation Tab
- Request Field Strength
- Create a sub-directory for the results.
5. Sites Tab
- Nothing to modify. There is only one site and will be active.
6. Building Data Tab
- Inspect the settings.
- No need to modify them
7. Computation Tab
8. Dominant Path Model for Urban
(recommended) - Dominant Path Model for Indoor as well (recommended).
9. OK
Run Simulation
1. Computation / Propagation: Compute All or the RUNPRO button on the left
Result with Dominant Path Model, ctd.
• Note the effects of windows and doors.
• Scale 65-85 dB µV/m
Result with COST 231: Extended Walfisch-Ikegami Model
• Propagation is along straight lines only (always in the vertical plane between transmitter and point of interest).
• Fields are lower than with DPM, as expected.
• Scale 65-85 dB µV/m
• Effects of windows and doors visible.
Save and Exit
2. File / Exit
https://altair-2.wistia.com/medias/tf3i7q01vs
5 Wave Propagation Model for Rural/Suburban Scenarios

5.1 Motivation
In Rural scenarios, the diffractions over terrain and buildings are computed. In this case scattering and reflections coming from hills,
mountains, etc. are neglected since the base station antenna is usually placed above such obstacles.
The predictions are based on the knowledge of topography, land usage and in some models also the building height information.
This section describes the different approaches to propagation modeling in terrain scenarios.
Figure 5.1 : Typical scenario for a terrain prediction
5.2 Required Databases

Topographical databases (digital elevation model or DEM) are required for the prediction of propagation.
Topographical databases consist of binary stored pixel data with an arbitrary resolution, for example, 50m x 50 m.
Additionally, morphological data can be considered by empirical correction values to improve the accuracy of the model. This data is
also stored as binary data. The different morphological properties are coded, for example:
• 1 for urban
• 2 for suburban
• 3 for forest
• 4 for waters
• 5 for acre
5.3 Hata-Okumura Model

The Hata-Okumura model is a simple empirical approach for macro-cellular areas with vertical polarization. This model features a
very short computation time.
The Hata-Okumura model is based on the evaluation of intensive measurements at frequencies between 200 MHz and 2 GHz with
vertical polarization. The equations derived from the measurement data require only the following four parameters:
frequency ,
the distance between transmitter and receiver
antenna height of the transmitter ,
antenna height of the receiver .
Figure 5.2 : Definition of the effective antenna height
Because of the calibration with measurement data the Hata-Okumura model is restricted to the following ranges for the different
parameters:
The basic transmission loss in urban areas is then computed according to the following formulas. There is an equation for the basic
loss and different correction terms according to different propagation environments (dense urban, suburban and open):
where for suburban areas
For urban areas
1.1
For urban areas
In addition to the formulas for the urban case, there are some modifications for rural (village) and open areas (acre). This leads to
the following equations:
These formulas describe the wave propagation assuming a flat terrain because the terrain profile between the transmitter and the
receiver is not considered. If there is, for example, a hill between the transmitter and receiver, the results will not be affected. Also,
local effects around the receiver are neglected (for example, reflection or shadowing).
COST 231 has extended the Hata-Okumura model to the frequency band between 1500 MHz and 2000 MHz by analyzing Okumura's
propagation curves in the upper-frequency band. This combination isdenoted as the “COST-Hata-Model”
5.4 Parabolic Equation Method

The parabolic equation method efficiently describes diffraction and forward-scattering processes in inhomogeneous terrains.
Due to the limited accuracy of the empirical models, there is a need for methods that describe diffraction and forward-scattering
processes in inhomogeneous terrain, thus the effect of the terrain on the large-scale variation. Therefore, different full wave
approaches have been investigated, and one efficient possibility is the parabolic equation method.
This method employs a numerical evaluation of the parabolic equation (PE) to compute the field strength in a macro-cellular area
based on terrain data. Because the different propagation mechanisms (free space propagation, reflection and diffraction) are implicitly
considered this model is very accurate. Due to the sophisticated algorithm, the computation time is quite long in comparison to the
empirical models.
5.4.1 Basic Principles
Figure 5.3 shows the vertical terrain section and the computation domain of the PE model
Figure 5.3 : Vertical terrain section and computation domain
Neglecting backward propagation and assuming a rotationally symmetric problem, a partial differential equation of parabolic type out
of the Maxwell equations can be derived:
2
∂Ψ j ∂2 Ψ k − k 02
+ + j Ψ=0
∂ r 2k 0 ∂ z 2 2k 0
Ψ is the field strength related to the field of a linear source. In the far-field the vertical component of the electrical field is
Ψ − jk0 r
E z = ZF0 ⋅ H Φ = ZF0 ⋅ e
r
and ko denotes wave number in free space. k is the complex wave number in an in-homogenous lossy atmosphere
This so-called standard parabolic equation (SPE) leads to valid results if the propagation angle for the horizon is in the range, –15°
to 15°. Therefore, the PE must not be used in the vicinity of the transmitter.
A finite difference technique solves the PE in the computation domain using an iterative method. In WinProp the required start field
is computed using free space propagation and a single ground reflection ray.
5.4.2 Absorbing Media
As the implementation of a transparent boundary condition at the top of the computation domain yielded to high computation times,
an absorbing medium is used instead to avoid reflections at the upper boundary. Having a width of about dA = 150 wavelengths, the
absorbing medium works well for all frequencies if an adequate distance hA to the highest terrain point is assumed.
The prediction plot in Figure 5.4 shows the field in the artificial absorbing medium above the dotted line
5.4.3 Impedance Boundary at the Ground
Figure 5.4 : Sample prediction (945 MHz, vertical terrain section).
For all parabolic equation’s calculations, the conductivity and dielectric permittivity of the ground surface is considered. Three
variations how to model the impedance boundary condition are presented here.
1. Discrete terrain approximation
A step-like height profile is assumed in Figure 5.5 , which leads to the simplified Leontovich boundary condition.
Figure 5.5 : Step-like height profile approximation.

Due to the numerical instabilities of this method, the radial step size of the computation grid is reduced at down-sloping ground
automatically.
Continuous terrain approximation
By setting appropriate values for the complex wave number k , the ground is modeled as if it is a part of the atmosphere. Together
with a variable step size this approach is very robust but does not match the boundary condition exactly because abrupt
alterations of the atmospheric medium are not allowed when using the parabolic approximation.
Terrain profile approximation with coordinate transformation
Another possibility to prevent numerical instabilities is the transformation of the irregular computation domain to a rectangular area
as shown in Figure 5.6 for down-sloping ground
Figure 5.6 : Transformation of the computation domain
Therefore, the coordinate transformation
is applied at each radius r and the modified Leontovich boundary condition
is used, which ensures the uniqueness of the solution. Θ denotes the sloping angle for the horizon.
Because of the distortion of the computation grid, the transformation should only be used for slight hilly terrains.
5.4.4 Wide Angle Parabolic Equation
The disadvantage of the small valid propagation angle of the standard parabolic equation can be remedied if an extended parabolic
approximation of the Maxwell equations is used.
 2
k − k 02 1 ∂2  ∂Ψ j ∂2 Ψ
2
k − k 02
1 + 
 4k 02
+
4k 02 ∂ z 2  ∂ r + 2k ∂ z 2 + j 2k Ψ = 0
  0 0
Therewith propagation angles up to about 40° are acceptable. The additional terms do hardly prolong the computation time and thus
the wide-angle PE should be preferred.
5.4.5 Comparison between the Hata-Okumura and Parabolic Equation Models
Differences between the Hata-Okumura and parabolic equation models are exemplified in the prediction of shadowing.
Figure 5.7 : Hata-Okumura model prediction
Figure 5.8 : Parabolic equation model prediction.
Figure 5.9 : Height Profile for the Prediction
Figure 5.10 : Difference plot between the two models
When looking at the difference plot with consideration of the height profile, the parabolic equation model accounts much better for
shadowing effects as they occur in valleys. On the other hand, the Hata-Okumura model is too pessimistic in cases where both the
transmitter and the receiver are located on hills.
5.5 Rural/Suburban Scenario Example

Main Exercise Start
POSTFEKO
1. Open Basestation_Sector.pfs with POSTFEKO
2. Inspect the antenna pattern – does it make sense?
3. Run the script APA_Export2.lua
(Home tab; open with Script editor;
Click Play (Run) button)
Script
1. Export .apa file for the 3D pattern
(specify the extension as well)
2. Close the script editor and exit POSTFEKO.
The file is available on disk.
Start Aman
1. Launch AMan interface Main window is gray at this stage.
3. Select Basestation_Sector.apa (apa = antenna pattern ASCII)
4. Explore the visualization options.
Left-mouse and drag
Right-mouse and drag
Both mouse buttons and drag
Menu options on the left
Note: linear view still works with dB values.
The view distinction is just about the display of those dB values.
Produce Binary File
1. File / Save As
Save in. apb format (antennapattern binary)
2. File / Exit That’s all there is to it
Main Exercise Start
Start ProMan
1. Launch ProMan interface Main window is gray at this stage.
Rural and Suburban; browse to the Hilly Terrain databases.
.tdb = topo data binary
.mdb = clutter (land usage)
.mct = morpho clutter table
3. OK
Topography and Clutter Databases
• Usually supplied by external vendors.
• Can be converted by ProMan under the Data menu.
Setup Simulation Parameters
1. File / Save Project as HillyTerrain01.net
3. Under Simulation tab, accept values as shown.
4. Click the Change button and create a sub-directory “DPM” for results.
We will first analyse with the Dominant Path Model (DPM) (to be specified later).
5. Request Field Strength and Path Loss.
6. Add a site with three sectors as shown.
7. First location: (X, Y) = (513628, 5400187) This is on the highest hill.(*)
8. You can leave Z=0 at this stage.
9. Edit properties as shown; do this for every sector.
Note we select the antenna pattern we obtained from FEKO earlier.
Note we now enter Height 30 min the Antenna/Cell dialogue.
10. Place another three-sector base station at (X,Y) = (516000, 5402142) This is on a lower hill closer to town.
11. Edit the three sector antennas as before; use the pattern from FEKO.
Setup Simulation Parameters, ctd.: Computation tab
12. Dominant Path Model with default settings
13. OK
After Base Station Placement
• Sectors with azimuth 45 point to the city center.
• File / Save Project
Run Simulation
Computation / Propagation: Compute All or the RUNPRO button on the left

The simulation takes only seconds to complete.
End
Dominant Path Model is the best choice in this situation: cellular base stations in hilly terrain.
Other methods have value in other scenarios, e.g. not-very-hilly terrain.
2. File / Exit
https://altair-2.wistia.com/medias/tohxozd287

Winprop Ebook 31 031-2021

Uploaded by

Copyright:

Available Formats

Winprop Ebook 31 031-2021

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Winprop Ebook 31 031-2021

Uploaded by

Copyright:

Available Formats

EBOOK

Learn Basics of Wave Propagation

1 INTRODUCTION TO WAVE PROPAGATION ANALYSIS .................................................. 5

1.2 Mobile Radio Channel ............................................................................................................................. 6

1.3 Wave Propagation Principles ................................................................................................................. 7

1.3.1 Path loss .......................................................................................................................................... 10

1.4 Propagation Path-Loss Models ........................................................................................................... 10

2 INTRODUCTION TO WINPROP ........................................................................................ 12

2.2 WinProp Applications ........................................................................................................................... 13

2.2.1 Automotive ...................................................................................................................................... 14

2.2.2 Aerospace ....................................................................................................................................... 14

2.2.3 Telecommunication ......................................................................................................................... 15

2.2.4 Wireless Sensors ............................................................................................................................ 16

2.2.5 Railway ............................................................................................................................................ 16

2.3 WinProp Components........................................................................................................................... 17

2.3.1 ProMan ............................................................................................................................................ 17

2.3.2 WallMan .......................................................................................................................................... 17

2.4 TuMan ..................................................................................................................................................... 18

2.4.1 AMan ............................................................................................................................................... 18

2.4.2 CoMan ............................................................................................................................................. 19

2.4.3 CompoMan ...................................................................................................................................... 19

3 WAVE PROPAGATION MODELS FOR INDOOR SCENARIOS ....................................... 20

3.2 Types of Propagation Models .............................................................................................................. 20

3.3 Required Database ................................................................................................................................ 21

3.4 Modified Free Space Model (MF): ........................................................................................................ 21

3.5 Motley-Keenan Model (MK): ................................................................................................................. 22

3.6 COST- Multi-Wall Model (MW) .............................................................................................................. 23

3.7 Standard Ray Tracing (SRT) ................................................................................................................ 24

3.7.1 Propagation Paths ........................................................................................................................... 24

3.7.2 Computation of Each Ray’s Contribution ........................................................................................ 25

3.8 Intelligent Ray Tracing (IRT) ................................................................................................................ 25

3.8.1 Propagation Paths ........................................................................................................................... 26

3.8.2 Computation of Each Ray’s Contribution ........................................................................................ 26

3.8.3 Pre-processing ................................................................................................................................ 26

3.9 Dominant Paths (DP)............................................................................................................................. 27

3.9.1 Prediction ........................................................................................................................................ 27

3.9.2 Multipath Propagation in Indoor Environments ............................................................................... 28

3.9.3 Principle of the Dominant Paths ...................................................................................................... 28

3.9.4 Determination of the Dominant Paths ............................................................................................. 29

3.9.5 Path selection criteria ...................................................................................................................... 31

3.9.6 Share of the empirical and neural model ........................................................................................ 32

3.10 Indoor Scenario Example ..................................................................................................................... 32

3.11 Indoor Advanced Scenario Model Example ....................................................................................... 49

4 WAVE PROPAGATION MODEL FOR URBAN SCENARIOS ........................................... 70

4.2 Required Databases .............................................................................................................................. 70

4.3 Empirical COST-Walfisch-Ikegami Model ........................................................................................... 71

4.4 Urban Scenario Example ...................................................................................................................... 73

4.5 Basics of Ray-Optical Models .............................................................................................................. 87

4.5.1 Specular Reflection ......................................................................................................................... 87

4.5.2 Diffraction ........................................................................................................................................ 88

4.5.3 Multiple Diffraction ........................................................................................................................... 88

4.5.4 Scattering ........................................................................................................................................ 88

4.5.5 Penetration and Absorption ............................................................................................................. 88

4.5.6 Ray Tracing Versus Ray Launching................................................................................................ 89

4.6 Ray Tracing ............................................................................................................................................ 89

4.6.1 Acceleration ..................................................................................................................................... 89

4.6.2 Computation of Each Ray’s Contribution ........................................................................................ 90

4.6.3 Breakpoint ....................................................................................................................................... 91

4.7 Description of the Individual Ray-Optical Models ............................................................................. 91