A DYNAMIC SYSTEM-LEVEL SIMULATION TOOL FOR
UMTS FDD
Nuno Cota
Instituto Superior de Engenharia de Lisboa, Rua Conselheiro Emídio Navarro-1, 1950-062 Lisboa, Portugal
António Rodrigues
Instituto Superior Técnico, Av. Rovisco Pais, 1049-001 Lisboa, Portugal
Keywords: Simulation, Modelling, UMTS, WCDMA
Abstract: This paper presents a tool for dynamic simulation of radio resources in UMTS FDD. The developed tool
provides an integrated environment for users scenario definition, environment edition, base stations
planning, simulation and analysis. This tool can also be used to evaluate the impact of new services, different
system settings or new subscriber profiles, in the system’s global performance. For implementing the
different mechanisms of WCDMA radio technology, the simulation tool kernel was supported by a system-
level, dynamic, stochastic and event driven simulation model. Both uplink and downlink directions are
considered, including soft handover connections. The final results confirm the developed tool validity and its
good functionality in relation to simulation scenario definition, simulation and analysis phases.
1 INTRODUCTION
The new third generation (3G) services and user
profiles, as well the radio access technology itself
puts important challenges when planning and design
a wideband code division multiple access
(WCDMA) based network. The multi-service
characteristic of these networks lead to different bit
rates in combination with different profile mobility
settings, different link level gains, margins and E
b
/N
0
requirements. Furthermore, WCDMA radio
resources management mechanisms like transmit
power control, admission control, load control or
soft-handover, associated with the characteristic of
implicit interference limited capacity in WCDMA,
imply simulation tools usage in the planning
process.
Static simulation has been extensively used on
second generation (2G) networks planning.
However, the use of this type of simulators carries
some limitations for 3G network performance
analysis. The multi-service aspect, combined with
the radio resource mechanisms, constitutes the main
limitation of static simulation. Another aspect that
will not be able to be included in a static simulation
is the subscriber mobility. This aspect has a
significant impact in the system performance,
mainly in WCDMA where, due to the systems
characteristics, handover has a considerable
influence in the cell resources usage. Therefore, the
use of dynamic simulation tools has a crucial
importance in the radio resources performance
evaluation on WCDMA networks. The main purpose
of this paper is to present a tool for system level and
dynamic simulation of radio resources in UMTS
FDD. This tool considers the main aspects of 3G
network usage, like multi-service, multi-profile and
non uniform traffic distribution. Moreover the
mobility characterization of user’s profiles is
considered, which gives a more realistic approach.
2 SIMULATOR
CHARACTERIZATION
The developed software application provides an
integrated environment for simulation, where a
single application can be used in the different phases
of the simulation process, including the scenario
definition, environment morphology edition, system
simulation and system performance analysis. The
17
Cota N. and Rodrigues A. (2004).
A DYNAMIC SYSTEM-LEVEL SIMULATION TOOL FOR UMTS FDD.
In Proceedings of the First International Conference on E-Business and Telecommunication Networks, pages 17-23
DOI: 10.5220/0001389700170023
Copyright
c
 SciTePress

application runs in Windows™ operating system,
providing a user friendly interface for parameters
input, simulation and results analysis. In this section,
the main characteristics of the simulator tool will be
explained.
2.1 Application Structure
The simulation tool usage can be structured on three
consecutive phases, as is presented in Figure 1. One
of the main aspects considered on the developed tool
is its flexibility. To obtain this, we use a flexible
simulation scenario, allowing new scenarios creation
for a system planning or system behaviour study
purposes. Thus, the application provides three types
of windows based interfaces to allow simulation
scenario edition. First, a scenario layout is displayed,
in which environment and base stations settings are
presented that can be edited. On the second window
interface, all morphologic information from
simulated environment, including roads and
buildings data, can be edited. The third configuration
interface provides a visual tool for subscriber’s
density areas creation and edition. Beyond these
three main types of interfaces, a collection of general
parameters, which configures several aspects of the
simulation scenario and tool, can be set by several
dialog boxes.
Configuration
Users
scenario
Simulation
Radio
coverage
prediction
System
level
simulation
Modelo de
Propagação
Modelo de
Propagação
Propagation
models
Modelo de
Propagação
Modelo de
Propagação
Mobility
models
Modelo de
Propagação
Modelo de
Propagação
Traffic
models
Analysis
Base
Stations
configuration
Radio
coverage
and losses
Demand
characte-
rization
Load and
Interferences
System
performance
Users
generation
General
parameters
General
parameters
Environment
Figure 1: Application structure.
In the simulation phase several modules are
considered including the traffic, mobility and
propagation models. Furthermore, three simulation
modules are implemented, of which the most
important is the system level simulation. This
module uses the mobility and traffic models with the
users generation module output to simulate system
subscriber’s behaviour, and conjugates it with the
output of radio coverage prediction module.
The purpose of the users generation module is to
populate the system by setting the initial location,
services and profiles from all system subscribers. In
the radio coverage prediction module all static radio
link calculations will be made, based on propagation
models, which in turn depends on base stations
configuration and environment settings. All outputs
from simulation modules can be analyzed separately
or can be saved on a file for later analysis.
Finally, based on the results taken from
simulation, the software user is able to continue the
analysis phase. Different kinds of formats can be
used to display all outputs taken from simulation
modules.
2.2 Simulation Scenario
The structure adopted to store all simulation related
data was designed to give a great flexibility on new
scenarios definition and on independency
preservation between scenario components.
General Parameters
SimulationSystem VisualizationPropagation
Base Stations
Data
Users Scenario Environment
Users
profiles
Services
Users density
areas
Link
Performance
Data
Mobility
Models
Base
Stations
Link
performance
measures
Antennas
Environment
Digital image
Geographical
information
Urban
morphology
Simulation
Scenario
Figure 2: Simulation scenario components.
On Figure 2 we depict the information set that,
defines the simulation scenario, which is structured
and collected on five different groups:
• General parameters – auxiliary information set to
several purposes that allow the definition of the
working conditions of simulation model and
application;
• Environment – this information group defines the
geographic context where the system is installed,
and contains geographic and morphologic
information;
• Base stations data – this information set defines all
system radio equipment. The site location, radio
and system operating parameters and antennas
types and placement information constitute this
database. To define the antennas physical
parameters we use an equipment database, which
includes the horizontal and vertical radiation
pattern information and gains;
• Users scenario – contains all subscriber’s
behaviour, services and profiles information that
ensures the characterization of simulation scenario.
All this information will be described on Section
3;
ICETE 2004 - WIRELESS COMMUNICATION SYSTEMS AND NETWORKS
18

• Link performance data – information obtained
from link level simulations, allowing the radio link
modelling at the system level simulation model.
2.3 Environment Data
The geographical information consists on a raster
database, where each entry has the ground height
and needed classification information to predict the
link path losses. Besides the geographical
information, a vector database is essential to know
the building limits and roads definition. This
information is compiled in a database denominated
by urban morphology. Buildings are represented by
two dimensional polygons which describes the
boundary of the area containing the building. Roads
are described by lines, which represent a subscriber
trajectory. The positions resulting from lines
overlapping constitutes possible crossroads. Each
road structure also includes information about
allowed speed.
The developed tool has the particularity to
include a visual edition environment to make it
possible to create and change buildings and roads on
the studied environment.
2.4 Radio Link Modelling
On the system-level simulation model that
constitutes the tool kernel, the modelling of all
aspects, concerning radio link, has a crucial
importance. This modelling gives us an abstraction
level relatively to the link physical layer. Thus,
several tables could be imported from link-level
simulators or radio-link measures, changed and used
on simulation model. These tables contain:
• Uplink (UL) and downlink (DL) E
b
/N
0
requirements, for different channel rates and
speeds;
• Downlink orthogonality factor;
• Link gains due the soft handover;
• Fast fading margins.
The link path loss is predicted by a propagation
model, which is chosen according to the
environment type. In large macro-cellular areas the
COST 231 Hata model (Damoso, 1999) is used. For
a micro-cellular urban environments the tool
considers the model COST 231 Walfish Ikegami.
Finally, for the particular case where the
environment is a Manhattan model like, a specific
model is used.
Regarding to slow fading, a method defined in
(Viterbi, 1995) was used. On this method, it is
considered that the link resulting slow fading is
modelled by a Gaussian random variable, and its
value depends on the sector and site location, and
the subscriber location. Fast fading effects are
considered by a link margin (Cota, 2004).
3 DEMAND
CHARACTERIZATION
In 3G wireless multimedia systems, each subscriber
can use one or more simultaneous services. Besides,
performance requirements within a service may
differ depending on the user’s mobility and
propagation environment. On the other hand,
subscriber’s location may not be uniform, which
means that several density areas and hotspots
existence has to be considered.
Thus, for an accurate demand characterization, a
flexible and scalable demand modelling must be
included in the tool, making a new demand scenarios
possible and a system resources utilization growth.
In this section we propose a flexible and modular
user scenario for a complete demand
characterization. Figure 3 presents the user scenario
structure.
User Scenario
Services
Traffic
Model
Mobility
Model
User
Profiles
User density
areas
Demand
Characteri-
zation
Environment
Geographical
Data
Urban
Morphology
Figure 3: User scenario structure.
The main components of user scenario are the
services and user profiles. They summarize the main
system resources demand aspects. The traffic models
will be responsible for the event generation that will
characterize the traffic and resources usage. Several
models had been considered according to service
type. Another aspect that needs to be included on
user scenario is the subscriber mobility, since it will
go to influence all system performance. Each
subscriber mobility is described by the mobility
model, which in turn is defined on their own profile.
Besides profile settings, mobility models also need
information about urban morphology to characterize
the subscriber mobility behaviour. The user density
areas database defines the subscriber initial
placement and movement limits.
3.1 Services
The information set that constitutes a service will
define the network resources usage context by
A DYNAMIC SYSTEM-LEVEL SIMULATION TOOL FOR UMTS FDD
19

subscribers. Service structure contains information
about traffic behaviour, resources usage relative
information and the maximum values of supported
packet delay. All service parameters defined will
ensure the tool versatility to create new services and
to change the existing services parameters. These
parameters are:
• Source bit rate – average UL and DL source bit
rate;
• Channel bit rate – average UL and DL physical
source rate. For this value estimation the signalling
channel has been taken into account, as well as
CRC and tail bits attachment, coding and final rate
matching into UMTS standard channel bit rates;
• Switching mode – packet or circuit switched
mode;
• Maximum transfer delay – a quality of service
parameter. If the delay experimented by a packet
exceed this value, the session will be dropped.
• Source traffic model – one of three available
models. Choice depends on switching mode and
service characteristics;
• Source traffic model parameters – values that
define traffic behaviour.
3.2 Traffic Models
In order to generate the UMTS multi-service traffic,
three source models had been implemented. Traffic
models choice not only considers resources usage
behaviour, but also the flexibility to create new
services and implementation viability on a
reasonable computational platform. On this basis,
the models that had been implemented are:
• Poisson – model indicated to circuit switched
services. In this model calls are generated
according to a Poisson process and call duration
may be described by several probability
distributions, which depends on service
applications type.
• ON-OFF – a simple packet services model, that
includes inactivity periods and activity periods into
a packet session. For both UL and DL direction
packets generation configuration, several
parameters are provided. A complete model
description can be found in (Cota, 2004).
• ETSI – the implementation of ETSI TR 101-112
(UMTS, 1998) model for packet-switched traffic.
In this model it is considered that a session
consists of a sequence of packet calls. During a
packet call, several packets may be generated with
different sizes. This model considers only traffic in
DL direction.
3.3 Mobility
In mobile systems, the terminal mobility has an
important influence on service performance. Due to
the UMTS characteristics, this influence must be
considered and quantified on simulation tools, in
order to achieve an accurate performance evaluation.
On the other hand, mobility inclusion brings an
increased complexity of simulation model. Based on
these facts, a simple but flexible mobility model was
included.
The simulation tool considers five classes that
describe the existing mobility types in an urban
scenario:
• Static;
• Pedestrian;
• Low speed vehicular;
• Medium speed vehicular;
• High speed vehicular.
The model adopted to describe subscriber’s
motion direction and speed was the Gauss-Markov
Mobility Model (Liang, 1999) that was designed to
adapt to different levels of randomness via two
tuning parameters. At fixed intervals of time,
movement occurs by updating the speed and
direction of each terminal. Specifically, the value of
speed and direction at one instance is calculated
based upon the value of speed and direction at the
previous
instance and a random variable using
equations (Cota, 2004).
However, this mobility model has a limitation
due the fact that he does not consider the urban
morphology. In order to become a more realistic
model, an improvement was made, considering the
vector data of streets and crossroads. Thus, it is
considered that vehicles moves along streets and
may turn at cross streets with a given probability.
However, streets change only occurs if new street
support terminal speed. Speed limitation can be
configured to each road of the database.
3.4 User Profiles
To evaluate the demand of WCDMA resources, it is
necessary to identify different types of customers,
since their calling patterns, usage and mobility
behaviour may differ. Thus, system subscribers
population would be classified according to their
user profiles, which includes information about
service usage, mobility and placement.
User profile contains a service list with the busy
hour call attempts (BHCA) respected values.
Placement information refers to location limits
during simulation, e.g., a high-speed vehicular
subscriber cannot transit to an indoor environment.
ICETE 2004 - WIRELESS COMMUNICATION SYSTEMS AND NETWORKS
20

Finally, mobility information classifies user profiles
concerning their mobility class.
3.5 User Density Areas
Due to the fact that real spatial distribution of
subscribers is not uniform, the determination of
initial position and mobility boundaries of all users
is an important task of demand characterization.
This information is determined based on user density
areas database contained on user scenario.
Each user density area includes a user profiles
list where each entry has a profile and its
probability. This probability refers to percentage
from the total users contained on area. Each area
boundary is defined by a polygon that can be defined
or edited on the tool. In this way any spatial
distribution of users may be defined according to
real distributions. Users are uniformly distributed
inside each area. This distribution also respects the
profile settings concerning mobility limits. An
example showing several user density areas in
Lisbon urban area is depicted in Figure 4.
Figure 4: User density areas definition example.
4 SIMULATION MODEL
The simulation model is the most critical component
of the simulator and it must reflect UMTS system
behaviour. Thus, several radio resources
management mechanisms has to be considered on
simulation model design. In this section, we explain
the main characteristics of system-level simulation
model.
The main system related aspects considered in
the simulation model, are:
• UL and DL directions;
• Open and closed loop power control;
• Admission and load control;
• Handover and soft handover;
• Simultaneous services usage by a single
terminal;
• User mobility;
• Circuit and packet switching modes;
In order to implement these aspects keeping a
reasonable time response from simulation, an
efficient system-level model has been designed,
which main features are:
• Dynamic and Stochastic;
• Discrete events with a continuous time
reference;
• Event driven and process oriented simulation.
The model events set definition and
implementation details are given in (Cota, 2004).
5 SYSTEM PERFORMANCE
ANALYSIS
In this section we describe the analysis features
included on simulator. The analysis is the post-
simulation phase in which all data collected along
the simulation can be displayed to the tool user. The
analysis information can be divided in coverage,
users, base station and services analysis.
Figure 5: Downlink best server plot example.
The coverage analysis contains several
information about radio coverage which is plotted as
maps. To this analysis belongs the plots of link
losses, CPICH channel receive power and E
c
/I
0
, best
server, active set size and service coverage
prediction. In Figure 5, a plot of downlink best
server of an example scenario is depicted.
A DYNAMIC SYSTEM-LEVEL SIMULATION TOOL FOR UMTS FDD
21

Figure 6: Users distribution plot.
The users analysis includes all user and related
mobile equipment informatio. This information is
often plotted on maps, allowing that users services
performance could be related with their locations.
Also belonging to this group are the mobile
distribution, average transmission power,
satisfaction and outage, circuit and packet services
performance. In Figure 6 an example of users
distribution is displayed. In this scenario, six user
profiles had been defined (Cota, 2004). In the
presented plot we can see that users hotspot
locations correspond to users density area definition.
Another example of users analysis is the UL
transmit power. This analysis can be done for a
single service only or to all services. Transmit power
distribution can be analysed through its distribution
histogram or a CDF graph. An example is shown in
Figure 7.
Figure 7: Terminal transmit power histogram and CDF.
Base stations analysis allows the identification of
each base station concerning its performance. The
information can be plotted as a histogram or a XY
graph. Examples of this analysis type are the base
station load, interference, transmission power and
circuit services satisfaction, outage and performance.
Figure 8 presents the downlink average packet delay
of each cell, for the scenario in Figure 5.
Figure 8: Terminal transmit power histogram and CDF.
Finally, on services analysis group, we can
visualize all information about global performance.
This includes transmission power, satisfaction and
outage level, circuit and packet services
performance. Figure 9 shows the downlink transmit
power for an example scenario.
Figure 9: Services downlink transmit power.
6 CONCLUSIONS
This work concentrates in presenting a dynamic and
system-level simulation tool for UMTS FDD. It
integrates several simulation phases in one single
Windows™ based application. A flexible simulation
scenario was adopted allowing the creation of new
services and users profiles, and all simulation
parameters are fully interchangeable. The obtained
results allow the service, base stations and system
performance analysis. The final results confirm the
developed tool validity and its good functionality in
relation to simulation scenario definition, simulation
and analysis phase.
ICETE 2004 - WIRELESS COMMUNICATION SYSTEMS AND NETWORKS
22

REFERENCES
Cota, N., 2004. Radio Resource Planning Simulation in
UMTS FDD. (In portuguese), M.Sc. Thesis, IST,
Technical University of Lisbon
Damoso, E., Correia, L., 1999. Digital mobile radio
towards future generations systems – COST 231 Final
Report. COST Office, European Commission,
Brussels, Belgium.
Viterbi, A., 1995. Principles of Spread Spectrum
Communication. Addison-Wesley, 1995.
UMTS, 1998. Selection procedures for the choice of radio
transmission technologies of the UMTS (UMTS 30.03
version 3.2.0). TR 101 112 v3.2.0.
Liang, B. and Haas, Z., 1999. Predictive distance-based
mobility management for PCS networks. In
proceedings of the Joint Conference of the IEEE
Computer and Communications Societies, INFOCOM
A DYNAMIC SYSTEM-LEVEL SIMULATION TOOL FOR UMTS FDD
23