Mss Sp 72 Pdf Free Download

This deliverable describes WINNER II channel models for link and system level simulations. Both generic and clustered delay line models are defined for selected propagation scenarios. Disclaimer: The channel models described in this deliverable are based on a literature survey and measurements performed during this project. The authors are not responsible for any loss, damage or expenses caused by potential errors or inaccuracies in the models or in the deliverable.

Figures - uploaded by Pekka Kyösti

Author content

All figure content in this area was uploaded by Pekka Kyösti

Content may be subject to copyright.

Discover the world's research

20+ million members
135+ million publications
700k+ research projects

Join for free

WINNER II D1.1.2 V1.2

Page 1 (82)

IST-4-027756 WINNER II

D1.1.2 V1.2

WINNER II Channel Models

Part I Channel Models

Contractual Date of Delivery to the CEC: 30/09/2007

Actual Date of Delivery to the CEC: 30/09/2007 (updated 04/02/2008)

Author(s): Pekka Kyösti, Juha Meinilä, Lassi Hentilä, Xiongwen Zhao, Tommi

Jämsä, Christian Schneider, Milan Narandzić, Marko Milojević, Aihua

Hong, Juha Ylitalo, Veli-Matti Holappa, Mikko Alatossava, Robert

Bultitude, Yvo de Jong, Terhi Rautiainen

Participant(s): EBITG, TUI, UOULU, CU/CRC, NOKIA

Workpackage: WP1 Channel Model

Estimated person months: 62

Security: PU

Nature: R

Version: 1.1

Total number of pages: 82

Abstract:

This deliverable describes WINNER II channel models for link and system level simulations. Both

generic and clustered delay line models are defined for selected propagation scenarios.

Keyword list: Channel modelling, radio channel, propagation scenario, channel sounder, cluster,

polarisation, measurements, delay spread, angle spread, MIMO, fading

Disclaimer: The channel models described in this deliverable are based on a literature survey and

measurements performed during this project. The authors are not responsible for any loss, damage or

expenses caused by potential errors or inaccuracies in the models or in the deliverable.

WINNER II D1.1.2 V1.2

Page 2 (82)

Executive Summary

This deliverable presents WINNER II channel models for link level and system level simulations of local

area, metropolitan area, and wide area wireless communication systems. The models have been evolved

from the WINNER I channel models described in WINNER I deliverable D5.4 and WINNER II interim

channel models described in deliverable D1.1.1. The covered propagation scenarios are indoor office,

large indoor hall, indoor-to-outdoor, urban micro-cell, bad urban micro-cell, outdoor-to-indoor, stationary

feeder, suburban macro-cell, urban macro-cell, rural macro-cell, and rural moving networks.

The generic WINNER II channel model follows a geometry-based stochastic channel modelling

approach, which allows creating of an arbitrary double directional radio channel model. The channel

models are antenna independent, i.e., different antenna configurations and different element patterns can

be inserted. The channel parameters are determined stochastically, based on statistical distributions

extracted from channel measurement. The distributions are defined for, e.g., delay spread, delay values,

angle spread, shadow fading, and cross-polarisation ratio. For each channel snapshot the channel

parameters are calculated from the distributions. Channel realisations are generated by summing

contributions of rays with specific channel parameters like delay, power, angle-of-arrival and angle-of-

departure. Different scenarios are modelled by using the same approach, but different parameters. The

parameter tables for each scenario are included in this deliverable.

Clustered delay line (CDL) models with fixed large-scale and small-scale parameters have also been

created for calibration and comparison of different simulations. The parameters of the CDL models are

based on expectation values of the generic models.

Several measurement campaigns provide the background for the parameterisation of the propagation

scenarios for both line-of-sight (LOS) and non-LOS (NLOS) conditions. These measurements were

conducted by seven partners with different devices. The developed models are based on both literature

and extensive measurement campaigns that have been carried out within the WINNER I and WINNER II

projects.

The novel features of the WINNER models are its parameterisation, using of the same modelling

approach for both indoor and outdoor environments, new scenarios like outdoor-to-indoor and indoor-to-

outdoor, elevation in indoor scenarios, smooth time (and space) evolution of large-scale and small-scale

channel parameters (including cross-correlations), and scenario-dependent polarisation modelling. The

models are scalable from a single single-input-single-output (SISO) or multiple-input-multiple-output

(MIMO) link to a multi-link MIMO scenario including polarisation among other radio channel

dimensions.

WINNER II channel models can be used in link level and system level performance evaluation of

wireless systems, as well as comparison of different algorithms, technologies and products. The models

can be applied not only to WINNER II system, but also any other wireless system operating in 2 – 6 GHz

frequency range with up to 100 MHz RF bandwidth. The models supports multi-antenna technologies,

polarisation, multi-user, multi-cell, and multi-hop networks.

This report is divided into two parts. The first part defines the channel model structure and parameters.

The second part (separate volume) contains more detailed information about channel measurements and

analysis.

WINNER II D1.1.2 V1.2

Page 3 (82)

Authors

Partner Name Phone / Fax / e-mail

EBITG Pekka Kyösti Phone: +358 40 344 2000

Fax: +358 8 551 4344

e-mail: firstname.lastname@elektrobit.com

EBITG Juha Meinilä Phone: +358 40 344 2000

Fax: +

e-mail: firstname.lastname@elektrobit.com

EBITG Tommi Jämsä Phone: +358 40 344 2000

Fax: +358 8 551 4344

e-mail: firstname.lastname@elektrobit.com

EBITG Xiongwen Zhao Phone: +358 40 344 2000

Fax: +358 9 2561014

e-mail: firstname.lastname@elektrobit.com

EBITG Lassi Hentilä Phone: +358 40 344 2000

Fax: +358 8 551 4344

e-mail: firstname.lastname@elektrobit.com

UOULU/EBITG Juha Ylitalo Phone: +358 40 344 3352

Fax: +358 8 551 4344

e-mail: firstname.lastname@elektrobit.com

UOULU Mikko Alatossava Phone: +358 8 814 7638

Fax: +358 8 553 2845

e-mail: mikko.alatossava@ee.oulu.fi

WINNER II D1.1.2 V1.2

Page 4 (82)

UOULU Veli-Matti Holappa Phone: +358 8 814 2890

Fax: +358 8 553 2845

e-mail: crimson@ee.oulu.fi

TUI Milan Narandžić Phone: + 49 3677 69 3722

Fax: + 49 3677 69 1113

e-mail: milan.narandzic@tu-ilmenau.de

TUI Aihua Hong Phone: + 49 3677 69 1157

Fax: + 49 3677 69 1113

e-mail: aihua.hong@tu-ilmenau.de

TUI Marko Milojević Phone: + 49 3677 69 2673

Fax: + 49 3677 69 1195

e-mail: marko.milojevic@tu-ilmenau.de

TUI Christian Schneider Phone: + 49 3677 69 1157

Fax: + 49 3677 69 1113

e-mail: christian.schneider@tu-ilmenau.de

TUI Gerd Sommerkorn Phone: + 49 3677 69 1115

Fax: + 49 3677 69 1113

e-mail: gerd.sommerkorn@tu-ilmenau.de

CRC Robert Bultitude Phone: 1-613-98-2775

Fax: 1-613-990-7987

e-mail: robert.bultitude@crc.ca

CRC Yvo de Jong Phone: 1-603-990-9235

Fax: 1-613-990-6339

e-mail: yvo.dejong@crc.ca

NOK Terhi Rautiainen Phone: +358 50 4837218

Fax: + 358 7180 36857

e-mail: terhi.rautiainen@nokia.com

WINNER II D1.1.2 V1.2

Page 5 (82)

Table of Contents

1. Introduction................................................................................................. 7

2. Definitions ................................................................................................... 9

2.1

Terminology................................................................................................................................ 9

2.2

List of Symbols .........................................................................................................................12

2.3

Propagation Scenarios...............................................................................................................14

2.3.1

A1 – Indoor office.............................................................................................................16

2.3.2

A2 – Indoor to outdoor......................................................................................................16

2.3.3

B1 – Urban micro-cell.......................................................................................................17

2.3.4

B2 – Bad Urban micro-cell............................................................................................... 17

2.3.5

B3 – Indoor hotspot...........................................................................................................17

2.3.6

B4 – Outdoor to indoor..................................................................................................... 17

2.3.7

B5 – Stationary Feeder......................................................................................................17

2.3.8

C1 – Suburban macro-cell.................................................................................................19

2.3.9

C2 – Urban macro-cell......................................................................................................19

2.3.10

C3 – Bad urban macro-cell ...............................................................................................19

2.3.11

C4 – Urban macro outdoor to indoor................................................................................ 19

2.3.12

D1 – Rural macro-cell....................................................................................................... 20

2.3.13

D2 – Moving networks......................................................................................................20

2.4

Measurement Tools...................................................................................................................20

2.4.1

Propsound (EBITG, UOULU, Nokia)...............................................................................21

2.4.2

TUI sounder......................................................................................................................22

2.4.3

CRC sounder.....................................................................................................................24

3. Channel Modelling Approach ..................................................................26

3.1

WINNER Generic Channel Model............................................................................................ 27

3.1.1

Modelled parameters......................................................................................................... 27

3.2

Modelling process .....................................................................................................................27

3.3

Network layout..........................................................................................................................28

3.3.1

Correlations between large scale parameters....................................................................30

3.4

Concept of channel segments, drops and time evolution...........................................................33

3.4.1

Basic method for time-evolution.......................................................................................33

3.4.2

Markov process based method of time evolution..............................................................34

3.5

Nomadic channel condition.......................................................................................................34

3.6

Reduced complexity models......................................................................................................35

3.6.1

Cluster Delay Line models for mobile and portable scenarios.......................................... 36

3.6.2

Cluster Delay Line models for fixed feeder links.............................................................36

3.6.3

Complexity comparison of modelling methods................................................................36

4. Channel Models and Parameters............................................................. 37

4.1

Applicability..............................................................................................................................37

4.1.1

Environment dependence..................................................................................................37

4.1.2

Frequency dependence......................................................................................................37

4.2

Generation of Channel Coefficients..........................................................................................37

4.2.1

Generation of bad urban channels (B2, C3)...................................................................... 42

4.3

Path loss models........................................................................................................................ 43

WINNER II D1.1.2 V1.2

Page 6 (82)

4.3.1

Transitions between LOS/NLOS ......................................................................................46

4.4

Parameter tables for generic models..........................................................................................46

4.4.1

Reference output values....................................................................................................48

4.5

CDL Models..............................................................................................................................49

5. Channel Model Usage............................................................................... 50

5.1

System level description............................................................................................................ 50

5.1.1

Coordinate system.............................................................................................................50

5.1.2

Multi-cell simulations.......................................................................................................51

5.1.3

Multihop and relaying.......................................................................................................53

5.1.4

Interference .......................................................................................................................54

5.2

Space-time concept in simulations............................................................................................55

5.2.1

Time sampling and interpolation.......................................................................................55

5.3

Radio-environment settings....................................................................................................... 55

5.3.1

Scenario transitions...........................................................................................................55

5.3.2

LOS\NLOS transitions......................................................................................................55

5.4

Bandwidth/Frequency dependence............................................................................................ 55

5.4.1

Frequency sampling.......................................................................................................... 55

5.4.2

Bandwidth down scaling...................................................................................................55

5.4.3

FDD modeling...................................................................................................................56

5.5

Comparison tables of WINNER channel model versions..........................................................56

5.6

Approximation of Channel Models........................................................................................... 60

6. Parameter Tables for CDL Models........................................................... 61

6.1

A1 – Indoor small office............................................................................................................61

6.2

A2/B4 – Indoor to outdoor / outdoor to indoor.........................................................................62

6.3

B1 – Urban micro-cell...............................................................................................................63

6.4

B2 – Bad Urban micro-cell........................................................................................................64

6.5

B3 – Indoor hotspot...................................................................................................................64

6.6

C1 – Urban macro-cell ..............................................................................................................66

6.7

C2 – Urban macro-cell ..............................................................................................................67

6.8

C3 – Bad urban macro-cell........................................................................................................ 68

6.9

C4 – Outdoor to indoor (urban) macro-cell............................................................................... 69

6.10

D1 – Rural macro-cell...............................................................................................................70

6.11

D2a – Moving networks............................................................................................................ 71

6.12

Fixed feeder links - Scenario B5...............................................................................................72

6.12.1

Scenario B5a..................................................................................................................... 72

6.12.2

Scenario B5b.....................................................................................................................73

6.12.3

Scenario B5c..................................................................................................................... 75

6.12.4

Scenario B5f......................................................................................................................75

7. References................................................................................................. 77

WINNER II D1.1.2 V1.2

Page 7 (82)

1. Introduction

The goal of WINNER is to develop a single ubiquitous radio access system adaptable to a comprehensive

range of mobile communication scenarios from short range to wide area. This will be based on a single

radio access technology with enhanced capabilities compared to existing systems or their evolutions.

WINNER II is a continuation of the WINNER I project, which developed the overall system concept.

WINNER II has developed and optimised this concept towards a detailed system definition. [WINNERII]

The radio interface supports the challenging requirements of systems beyond 3G. It is scalable in terms of

carrier bandwidth and carrier frequency range. The system concept supports a wide range of radio

environments providing a significant improvement in performance and Quality of Service (QoS). The

radio interface optimises the use of spectral resources, e.g. through the exploitation of actual channel

conditions and multiple antenna technology. New networking topologies (e.g. relaying) supports cost-

effective deployments. Support of advanced resource management and handover eases the deployment of

the WINNER system concept enabling seamless service provision and global roaming. [WINNERII]

It has been widely understood that radio propagation has a significant impact on the performance of

wireless communication systems. The impact on future broadband systems is even more important due to

increased data rate, bandwidth, mobility, adaptivity, QoS, etc. Because of the major influence on the

system performance and complexity, radio channel models and simulations have to be more versatile and

accurate than in earlier systems.

WINNER I work package 5 (WP5) focused on wideband multiple-input multiple-output (MIMO) channel

modelling at 5 GHz frequency range. Totally six partners were involved in WP5 during 2004 – 2005,

namely Elektrobit, Helsinki University of Technology, Nokia, Royal Institute of Technology (KTH) in

Stockholm, Swiss Federal Institute of Technology (ETH) in Zurich, and Technical University of Ilmenau.

In the beginning of Phase I, existing channel models were explored to find out channel models for the

initial use in the WINNER I project. Based on the literature survey, two standardised models were

selected, namely 3GPP/3GPP2 Spatial Channel Model [3GPPSCM] and IEEE 802.11n. The former is

used in outdoor simulations and the latter in indoor simulations. Because the bandwidth of the SCM

model is only 5 MHz, wideband extension (SCME) was developed in WINNER I. However, in spite of

the modification, the initial models were not adequate for the advanced WINNER I simulations.

Therefore, new measurement-based models were developed. WINNER I generic model was created in

Phase I. It allows creating of arbitrary geometry-based radio channel model. The generic model is ray-

based double-directional multi-link model that is antenna independent, scalable and capable of modelling

channels for MIMO connections. Statistical distributions and channel parameters extracted by

measurements at any propagation scenarios can be fitted to the generic model. WINNER I channel

models were based on channel measurements performed at 2 and 5 GHz bands during the project. The

models covered the following propagation scenarios specified in WINNER I: indoor, typical urban

micro-cell, typical urban macro-cell, sub-urban macro-cell, rural macro-cell and stationary feeder link.

In the WINNER II project work package 1 (WP1) continued the channel modelling work of WINNER I

and extended the model features, frequency range (2 to 6 GHz), and the number of scenarios. Five

partners were involved, namely Elektrobit, University of Oulu / Centre for Wireless Communications

(CWC), Technical University of Ilmenau, Nokia, and Communication Research Centre (CRC) Canada.

WINNER I models were updated, and a new set of multidimensional channel models were developed.

They cover wide scope of propagation scenarios and environments, including indoor-to-outdoor, outdoor-

to-indoor, bad urban micro-cell, bad urban macro-cell, feeder link base station (BS) to fixed relay station

(FRS), and moving networks BS to mobile relay station (MRS), MRS to mobile station (MS). They are

based on generic channel modelling approach, which means the possibility to vary number of antennas,

the antenna configurations, geometry and the antenna beam pattern without changing the basic

propagation model. This method enables the use of the same channel data in different link level and

system level simulations and it is well suited for evaluation of adaptive radio links, equalisation

techniques, coding, modulation, and other transceiver techniques. Models have been developed in two

steps, WINNER II Interim Channel Models [WIN2D111] and the final WINNER II Channel Models (this

deliverable, D1.1.2).

This deliverable describes the (final) WINNER II Channel Models. The models are based on WINNER I

models [WIN1D54] and WINNER II interim models [WIN2D111]. This deliverable covers new features

and new scenarios, such as outdoor-to-indoor urban macro-cell and line-of-sight (LOS) urban macro-cell.

Some scenarios have been updated. The indoor part of the moving network scenario has been determined

WINNER II D1.1.2 V1.2

Page 8 (82)

and whole the scenario has been updated considerably, as well as the model for indoor hot-spot. Bad

urban scenarios have also been updated. New features of the WINNER II Channel Models include

modelling of the elevation of rays, treating the LOS component of the channel model as a random

variable, and moving scatterers in fixed connections. The differences in the scenarios Indoor-to-Outdoor

and Outdoor-to-Indoor were noticed to be negligible. Therefore these two scenarios have been merged.

Model parameters have been revised in the cases, where new results have pointed this necessary.

Valuable comments have been received also via standardisation work in various standardisation bodies,

especially in IEEE802.16m and ITU-R/8F. We have taken into account several such change proposals.

Probably most important of them is the tuning of our path-loss models.

During the projects WINNER I and WINNER II the models have been evolved, mainly by adding new

scenarios in the models, but also by including new features. In this process we have tried to conserve the

model parameters from changes as much as possible. However, some changes have been inevitable.

Therefore the models are not exactly the same in this and the earlier deliverables. The propagation

scenarios from WINNER Phase I have been included in this document, partly updated. In WINNER

Phase II the following new propagation scenarios have been created and documented in this document:

indoor-to-outdoor, outdoor-to-indoor, bad urban micro-cell, bad urban macro-cell and moving network

scenario. All the propagation scenarios have been listed and introduced in section 2.3. WINNER I,

WINNER II interim, and WINNER II final models are compared in section 5.5.

The deliverable is divided into two major parts. This first part is the main part and defines the channel

model structure and parameters. The second part contains more detailed information about channel

measurements and analysis performed during projects WINNER I and II. The two parts are published in

separate volumes to keep the size of each part reasonable.

SCM, SCME, and WINNER I channel models have been implemented in Matlab, and are available via

WINNER web site. WINNER II channel model implementation is planned to be available by the end of

the year 2007.

Sections 1 - 7 cover the following topics. Section 1 introduces this deliverable. Section 2 expresses some

definitions, like the propagation scenarios and introduces the used measurement tools. Section 3 defines

the channel modelling approach. Section 4 explains the generation of channel coefficients and describes

path loss models as well as parameters for generic models. Section 5 discusses how the channel models

are used in system level (multi-link) simulations, sampling, transition scenarios, bandwidth/frequency

dependence of the models. Parameter tables for reduced variability (CDL) models can be found from

Section 6. Reference list is in Section 7.

WINNER II D1.1.2 V1.2

Page 9 (82)

2. Definitions

2.1 Terminology

3GPP 3rd Generation Partnership Project

3GPP2 3rd Generation Partnership Project 2

ACF Auto-Correlation Function

ADC Analog-to-Digital Converter

AN Antenna Array

AoA Angle of Arrival

AoD Angle of Departure

AP Access Point (BS)

APP A Posteriori Probability

APS Angle Power Spectrum

AS Azimuth Spread

ASA Azimuth Spread at Arrival

ASD Azimuth Spread at Departure

AWGN Additive White Gaussian Noise

B3G Beyond 3G

BER Bit Error Rate

BRAN Broadband Radio Access Networks

BS Base Station

C/I Carrier to Interference ratio

CDF Cumulative Distribution Function

CDL Clustered Delay Line

CG Concept Group

CIR Channel Impulse Response

CRC Communications Research Centre Canada

CW Continuous Wave

DoA Direction of Arrival

DoD Direction of Departure

DS/DES Delay Spread

EBITG Elektrobit

ECDF Experimentally determined cumulative probability distribution function

ESA Elevation Spread at Arrival

ESD Elevation Spread at Departure

ESPRIT Estimation of Signal Parameters via Rotational Invariance Techniques

ETHZ Eidgenössische Technische Hochschule Zürich (Swiss Federal Institute of Technology

Zurich)

ETSI European Telecommunications Standards Institute

FDD Frequency Division Duplex

FIR Finite Impulse Response

FL Floor Loss, loss between different floors

FRS Fixed Relay Station

FS Fixed Station

GPS Global Positioning System

HIPERLAN High Performance Local Area Network

HUT Helsinki University of Technology (TKK)

IR Impulse Response

WINNER II D1.1.2 V1.2

Page 10 (82)

ISIS Initialization and Search Improved SAGE

KTH Kungliga Tekniska Högskolan (Royal Institute of Technology in Stockholm)

LA Local Area

LNS Log-Normal Shadowing

LOS Line-of-Sight

LS Large Scale

MA Metropolitan Area

MCSSS Multi-Carrier Spread Spectrum Signal

METRA Multi-Element Transmit and Receive Antennas (European IST project)

MIMO Multiple-Input Multiple-Output

MPC Multi-Path Component

MRS Mobile Relay Station

MS Mobile Station

MUSIC Multiple Signal Classification

NLOS Non Line-of-Sight

NOK Nokia

OFDM Orthogonal Frequency-Division Multiplexing

OLOS Obstructed Line-of-Sight

PAS Power Azimuth Spectrum

PDF Probability Distribution Function

PDP Power-Delay Profile

PL Path Loss

PLO Phase-locked oscillator

PN Pseudo Noise

RIMAX Maximum likelihood parameter estimation framework for joint superresolution estimation

of both specular and dense multipath components

RF Radio Frequency

RMS Root Mean Square

RT Roof-top

RX Receiver

SAGE Space-Alternating Generalized Expectation-maximization

SCM Spatial Channel Model

SCME Spatial Channel Model Extended

SF/SHF Shadow Fading

SIMO Single-Input Multiple-Output

SISO Single-Input Single-Output

SoS Sum of Sinusoids

std Standard deviation

SW Software

TDD Time Division Duplex

TDL Tapped Delay-Line

TUI Technische Universität Ilmenau

TX Transmitter

UE User Equipment (MS)

UOULU University of Oulu

UT User Terminal (MS)

WA Wide Area

WINNER Wireless World Initiative New Radio

WPx Work Package x of WINNER project

WINNER II D1.1.2 V1.2

Page 11 (82)

XPR Cross-Polarisation power Ratio

XPRH Horizontal Polarisation XPR

XPRV Vertical Polarisation XPR

WINNER II D1.1.2 V1.2

Page 12 (82)

2.2 List of Symbols

∆(• ) Change in parameter value

(• )

Transpose

(• )

Hermitian transpose

(• )* Complex conjugate

A Pairing matrix

C Correlation matrix

Tx antenna array response matrix

Rx antenna array response matrix

H MIMO channel transfer matrix

N Normal distribution

U Uniform Distribution

Azimuth arrival angle AoA

Azimuth departure angle AoD

Elevation arrival angle (EAoA)

Elevation departure angle (EAoD)

Delay

RMS delay spread

RMS angle spread of AoA

RMS angle spread of AoD

AoA

cluster-wise RMS angle spread of AoA

AoD

cluster-wise RMS angle spread of AoA

Shadow fading standard deviation

Variance

Per cluster shadowing standard deviation

Wavelength

Wave number

Vertical-to-horizontal XPR

Horizontal-to-vertical XPR

Doppler frequency

Complex gain of a propagation path

c Speed of light

Central frequency

BS antenna height

bs'

Effective BS antenna height

MS antenna height

ms'

Effective MS antenna height

Ricean K-factor

n Index to cluster

P Power

AoA distribution proportionality factor

AoD distribution proportionality factor

Break point distance

Delay distribution proportionality factor

WINNER II D1.1.2 V1.2

Page 13 (82)

s Index to Tx antenna element

t Time

u Index to Rx antenna element

WINNER II D1.1.2 V1.2

Page 14 (82)

2.3 Propagation Scenarios

The propagation scenarios modelled in WINNER are shown in Table 2-1. The propagation scenarios are

explained in more detail in the following paragraphs. In WINNER II the work was divided between

Concept Groups (CG) according to the environment they were working at. There were CG:s Local Area

(LA), Metropolitan Area (MA) and Wide Area (WA), Mapping of scenarios to Concept Groups is shown

in the table Table 2-1 in column CG.

Table 2-1. Propagation scenarios specified in WINNER.

Scenario Definition LOS/

NLOS Mob.

km/h Frequ

ency

(GHz)

Note

In building Indoor office /

residential

NLOS 0–5 2 - 6 LA

Indoor to outdoor

NLOS 0–5 2 - 6 LA AP inside UT

outside. Outdoor

environment urban

Hotspot

Typical urban micro-

cell LOS

NLOS

0–70 2 - 6 LA,

Bad Urban

micro-cell NLOS 0–70 2 - 6 MA

Same as B1 +

long delays

Hotspot Large indoor hall

LOS/

NLOS 0–5 2 - 6 LA

Outdoor to indoor.

micro-cell NLOS

0–5 2 - 6 MA

-Outdoor typical

urban B1.

-Indoor A1

B5a

Hotspot

Metropol

LOS stat. feeder,

rooftop to rooftop LOS 0 2 - 6

Same channel

model for hot spot

and metropol.

B5b

Hotspot

Metropol

LOS stat. feeder,

street-level to street-

level

LOS 0 2 - 6

B5c

Hotspot

Metropol

LOS stat. feeder,

below- rooftop to

street-level

LOS 0 2 - 6 MA

Extended B1

B5d

Hotspot

Metropol

NLOS stat. feeder,

above rooftop to

street-level

NLOS

0 2 - 6 MA

Extended C2

B5f

Feeder link BS ->

FRS. Approximately

RT to RT level.

LOS/

OLOS/

NLOS

0 2 - 6 WA

Desired link: LOS

or OLOS,

Interfering links:

LOS/(OLOS)

/NLOS

FRS -> MS = B1*

WINNER II D1.1.2 V1.2

Page 15 (82)

Table 2-1 (continued).

The propagation scenarios listed above have been specified according to the requirements agreed

commonly in the WINNER project [WIN1D72]. These are the environments and conditions, where all the

WINNER simulations have been carried out. There are a couple of facts that need to be understood about

the scenarios and channel models adapted to them:

1. The scenarios cover some typical cases. They are not intended to cover all possible environments

and conditions: e.g. the mountaineous or even hilly rural environments have not been covered.

Similarly the antenna heights do not cover all values that could be seen reasonable. Generally

speaking, the environments are such that are found in urban areas of European and North-

American countries.

2. The environments are described in two levels of details: firstly, most of the scenarios use the

ordinary way placing the transmitters and receivers, so that the only location parameter is the

distance between transmitter and receiver, called non-grid-based models. Secondly, the other

group of the scenarios is grid-based. This means that there is a grid of streets or a building lay-

out or both, where the transmitters and receivers can be located e.g. by Cartesian coordinates.

This latter group of scenarios include A1, A2, B1, B2 and B4, see 2.3.1 to 2.3.13. Other

scenarios belong to the first group.

With these selections we have been able to restrict the number of scenarios reasonable, and still

presumably covered representatively the conditions encountered by radio equipment in the field. We have

also been able to run some simulations in grid-based scenarios with higher precision than is possible in

conventional scenarios.

Scenario Definition LOS/

NLOS Mob.

km/h Frequ

ency

(GHz)

Note

Metropol

Suburban LOS/

NLOS 0–120 2 - 6

Metropol Typical urban

macro-cell LOS/

NLOS 0–120

2 - 6 MA

C3 Bad Urban macro-

cell NLOS 0–70 2 - 6 - Same as C2 + long

delays

C4 Outdoor to indoor

macro-cell NLOS 0-5 2 - 6 MA

-Outdoor typical

urban C2.

-Indoor A1

Rural

Rural macro-cell LOS/

NLOS 0–200

2 - 6 WA

a) Moving

networks:

BS – MRS, rural

LOS 0 –350

2 - 6 WA

Very large Doppler

variability.

b) Moving

networks:

MRS – MS, rural

LOS /

OLOS/

NLOS

0 – 5 2 - 6 LA Same as A1 NLOS

WINNER II D1.1.2 V1.2

Page 16 (82)

2.3.1 A1 – Indoor office

The scenario A1 has been modelled in D5.4. The layout of the scenario is shown in Figure 2-1. Base

stations (Access Points) are assumed to be in corridor, thus LOS case is corridor-to-corridor and NLOS

case is corridor-to-room. In the NLOS case the basic path-loss is calculated into the rooms adjacent to the

corridor where the AP is situated. For rooms farther away from the corridor wall-losses must be applied

for the walls parallel to the corridors. E.g. for the UE at the bottom wall of the lay-out in the Figure 2-1

there are three walls to be taken into account. Finally, we have to model the Floor Loss (FL) for

propagation from floor to floor. It is assumed that all the floors are identical. The Floor Loss is constant

for the same distance between floors, but increases with the floor separation and has to be added to the

path-loss calculated for the same floor.

Rooms: 10 x 10 x 3 m

Corridors: 5 x 100 x 3 m

Figure 2-1. Layout of the A1 indoor scenario.

2.3.2 A2 – Indoor to outdoor

In indoor-to-outdoor scenario (Figure 2-2) the MS antenna height is assumed to be at 1 – 2 m, and BS

antenna height at 2 – 2.5 m + floor height. The corresponding outdoor and indoor environments are B1 an

A1, respectively. It is assumed that the floors 1 to 3 are used in simulations, floor 1 meaning the ground

floor. The parameters of this scenario have been merged with B4 and C4 in table 4-7. We explain the

merging in detail in Part II of the deliverable. The comparison of Outdoor-to-Indoor and Indoor-to-

Outdoor scenario characteristics is presented in [AHHM07] and in [HACK07].

LOS/NLOS

Figure 2-2. Indoor to outdoor scenario.

WINNER II D1.1.2 V1.2

Page 17 (82)

2.3.3 B1 – Urban micro-cell

In urban micro-cell scenarios the height of both the antenna at the BS and at the MS is assumed to be well

below the tops of surrounding buildings. Both antennas are assumed to be outdoors in an area where

streets are laid out in a Manhattan-like grid. The streets in the coverage area are classified as "the main

street", where there is the LOS from all locations to the BS, with the possible exception in cases where

the LOS is temporarily blocked by traffic (e.g. trucks and busses) on the street. Streets that intersect the

main street are referred to as perpendicular streets, and those that run parallel to it are referred to as

parallel streets. This scenario is defined for both the LOS and the NLOS cases. Cell shapes are defined

by the surrounding buildings, and energy reaches NLOS streets as a result of the propagation around

corners, through buildings, and between them.

2.3.4 B2 – Bad Urban micro-cell

Bad urban micro-cell scenarios are identical in layout to Urban Micro-cell scenarios, as described above.

However, propagation characteristics are such that multipath energy from distant objects can be received

at some locations. This energy can be clustered or distinct, has significant power (up to within a few dB

of the earliest received energy), and exhibits long excess delays. Such situations typically occur when

there are clear radio paths across open areas, such as large squares, parks or bodies of water.

2.3.5 B3 – Indoor hotspot

Scenario B3 represents the propagation conditions pertinent to operation in a typical indoor hotspot, with

wide, but non-ubiquitous coverage and low mobility (0-5 km/h). Traffic of high density would be

expected in such scenarios, as for example, in conference halls, factories, train stations and airports,

where the indoor environment is characterised by larger open spaces, where ranges between a BS and a

MS or between two MS can be significant. Typical dimensions of such areas could range from 20 m × 20

m up to more than 100m in length and width and up to 20 m in height. Both LOS and NLOS propagation

conditions could exist.

2.3.6 B4 – Outdoor to indoor

In outdoor-to-indoor urban microcell scenario the MS antenna height is assumed to be at 1 – 2 m (plus the

floor height), and the BS antenna height below roof-top, at 5 - 15 m depending on the height of

surrounding buildings (typically over four floors high). Outdoor environment is metropolitan area B1,

typical urban microcell where the user density is typically high, and thus the requirements for system

throughput and spectral efficiency are high. The corresponding indoor environment is A1, typical indoor

small office. It is assumed that the floors 1 to 3 are used in simulations, floor 1 meaning the ground floor.

The parameters of this scenario have been merged with A2 and C4 in table 4-7. We explain the merging

in detail in Part II of the deliverable. The comparison of Outdoor-to-Indoor and Indoor-to-Outdoor

scenario characteristics is presented in [AHHM07] and in [HACK07].

2.3.7 B5 – Stationary Feeder

Fixed feeder links scenario is described in [WIN1D54] and defined as propagation scenario B5. This

scenario has also been partly modelled in [WIN1D54]. In B5, both terminals are fixed. Based on this, the

scenario is divided in four categories or sub-scenarios in [WIN1D54]. These are B5a (LOS stationary

feeder: rooftop to rooftop), B5b (LOS stationary feeder: street level to street level), B5c (LOS stationary

feeder: below rooftop to street level) and B5d (NLOS stationary feeder: rooftop to street level). Height of

street level terminal antenna is assumed to be 3-5 meters. To cover the needs of CG WA one modified

sub-scenario is needed in phase 2, scenario B5f: LOS/NLOS stationary feeder: rooftop-to-below/above

rooftop. All the sub-scenarios will be described below.

In stationary scenarios, the Doppler shifts of the rays are not a function of the AoAs. Instead, they are

obtained from the movement of the scatterers. In B5 we let one scatterer per cluster be in motion while

the others are stationary. In [TPE02] a theoretical model is built where the change of phase of scattered

waves between time t and t+∆t is given by

( ) ( )

αγπ

coscos4 ∆

(2.1)

where

is the angle between the direction of scatterer movement and

the direction orthogonal to

the reflecting surface and the reflection angle. By proper selection of these angles different Doppler

spectrums may be achieved. For B5d also an additional term in the path-loss model has to be included.

WINNER II D1.1.2 V1.2

Page 18 (82)

The feeder scenarios are specified here in connection of the micro-cellular environment. Actually the

feeders can be used also in the macro-cellular cases. In this document it is assumed that the useful macro-

cellular feeder link, C5, is identical with the feeder model B5c.

2.3.7.1 B5a

The signal in B5a can be assumed to consist of a strong LOS signal and single bounce reflection. Also far

away reflections can occur. The connection is almost like in free space, so that the path-loss does not

depend noticeably on the antenna heights. For this scenario fixed angle spread, delay spread and XPR

values are applied. Directive antennas are very effective in reducing the delay spread and other multi-path

impacts as explained in [PT00]. However, the model is applicable for omni-directional antennas for up to

300 meters in distance. By using directive antennas the range can be extended approximately to 8 km.

A static (non-fading) channel component is added to the impulse response. We select its power to be 10

dB. The power-delay profile (of all paths except the direct) is set as exponential, based on the results in

[OBL+02] and [SCK05]. The shadow fading is Gaussian with mean zero and standard deviation of 3.4 dB

based on [PT00]. B5a sub-scenario was specified and modelled in [WIN1D54]. The same channel model

is used also in Phase II.

2.3.7.2 B5b

In B5b it is assumed that both the transmitter and receiver have many scatterers in their close vicinity

similar as theorized in [Sva02]. In addition there can also be long echoes from the ends of the street.

There is a LoS ray between the transmitter and receiver and when this path is strong, the contribution

from all the scatters is small. However, beyond the breakpoint distance the scatterers start to play an

important role.

In papers e.g. [Bul02], [SBA+02] the results for different carrier frequencies are very similar. Therefore,

in B5b model the frequency is disregarded. The principle adopted for the WINNER phase 1 model allows

for various correlations between different parameters such as angle-spread, shadow-fading and delay-

spread. In this case, dependency between path loss and delay-spread [MKA02] is applied. This

dependence is handled by selecting one of three different CDL models given in [WIN1D54]. Based on the

delay-spread formula in [MAS02] we select the delay spread to be 30 ns when the path loss is less than 85

dB, 110 ns when the path loss is between 85 dB and 110 dB, and finally 380 ns when the path loss is

greater than 110 dB. With these settings the delay-spread used here is a factor 40%-156% of the delay-

spread formula of [MAS02] for path losses up to 137 dB. We call these path-loss intervals range1, range2

and range3 and different clustered-delay line models will be provided for the three cases.

In terms of path loss, the break point distance calculated as

0b0b

4hhhh

r−−

(2.2)

becomes important leading to so called two slope -model. The power delay profile (of all paths except the

direct) is set as exponential, based on the results in [SMI+00]. A per-path shadow fading of 3 dB is used

to obtain some variation in the impulse responses. A static (non-fading) channel component is added to

the impulse response. Based on [FDS+94] we select this parameter to be 10 in range1, 2 in range2, and 1

in range3. Also K-factor changes according to range. B5b sub-scenario was specified and modelled in

[WIN1D54]. The same channel model is used also in Phase II.

2.3.7.3 B5c and B5d

Scenarios B5c and B5d can be considered as LOS of B1 and NLOS of C2 respectively. Only support for

Doppler spectrum of stationary cases has to be introduced. B5c is probably the most important feeder link

scenario, because it will be used in urban micro-cell relay scenario. B5c is almost identical to the B1

micro-cellular LOS scenario. The only difference in environment is the assumed antenna height of the

mobile/relay. Same channel model will cover both of the cases, except the difference in Doppler spectrum

(mobility). Feeder link ends are stationary and the Doppler frequency results from motion of the

environment. In scenario B5c some clusters represent vehicles with speed of ~50 km/h and the rest of the

clusters represent stationary objects like walls and building corners.

Actually B5d seems less useful for a feeder link scenario. Therefore it is not discussed here further.

WINNER II D1.1.2 V1.2

Page 19 (82)

2.3.7.4 B5f

The sub-scenario is shown in the figure below.

Feeder-link

B5f

Master-

station

Relay

Relay to MS:

Feeder-link

B5f Master-

station

Relay

Relay to MS:

Interfering

Feeder-link

Relay

MS: B1

Relay

Desired

Feeder-link

a b c

Figure 2-3 B5f scenario for three cases: a) NLOS (OLOS) b) LOS c) Combined interference case.

B5f scenario consists of the cases with relay antennas some meters over the roof-top or some meters

below the roof-top. Critical information is, if the link is LOS or NLOS: It is possible to create LOS links

with the antennas below roof-tops. As well it is possible to implement NLOS links with antennas above

the average roof-top level. Our approach is that the desired BS to FRS links can be planned to be LOS or

OLOS, or at least "good" links. It is assumed that the interfering links from undesired BS to FRS can be

LOS or NLOS. (Although in practice this can be also affected by careful planning.) It should be pointed

out that the link FRS to MS is covered by the model B1. Interference to undesired feeder link may occur.

In B5f it is assumed that the relay station is shadowed due to some obstacle. The proposed model is based

on literature and formed from the B5a LOS fixed relay model by attenuating artificially its direct

component by 15 dB in average and summing to it a normally distributed random decibel number with

standard deviation 8 dB. The path loss formula is based on the references [ZEA99] and [GEA03]. The

other model parameters are the same as in B5a. The model B5f can also be understood as NLOS part of

the model B5a.

2.3.8 C1 – Suburban macro-cell

In suburban macro-cells base stations are located well above the rooftops to allow wide area coverage,

and mobile stations are outdoors at street level. Buildings are typically low residential detached houses

with one or two floors, or blocks of flats with a few floors. Occasional open areas such as parks or

playgrounds between the houses make the environment rather open. Streets do not form urban-like

regular strict grid structure. Vegetation is modest.

2.3.9 C2 – Urban macro-cell

In typical urban macro-cell mobile station is located outdoors at street level and fixed base station clearly

above surrounding building heights. As for propagation conditions, non- or obstructed line-of-sight is a

common case, since street level is often reached by a single diffraction over the rooftop. The building

blocks can form either a regular Manhattan type of grid, or have more irregular locations. Typical

building heights in urban environments are over four floors. Buildings height and density in typical urban

macro-cell are mostly homogenous.

2.3.10 C3 – Bad urban macro-cell

Bad urban environment describes cities with buildings with distinctly inhomogeneous heights or

densities, and results to a clearly dispersive propagation environment in delay and angular domain. The

inhomogeneties in city structure can be e.g. due to large water areas separating the built-up areas, or the

high-rise skyscrapers in otherwise typical urban environment. Increased delay and angular dispersion can

also be caused by mountains surrounding the city. Base station is typically located above the average

rooftop level, but within its coverage range there can also be several high-rise buildings exceeding the

base station height. From modelling point of view this differs from typical urban macro-cell by an

additional far scatterer cluster.

2.3.11 C4 – Urban macro outdoor to indoor

The Outdoor-to-Indoor scenario is specified here as follows: The outdoor environment is the same as in

urban macrocellular case, C2, and the indoor environment is the same as in indoor case, A1. The base

station antenna is clearly above the mean building height. This means that there will be quite long LOS

paths to the walls penetrated by the signals, mainly in the higher floors of the buildings. On the other hand

there is often quite a severe shadowing, especially in the lower floors. The propagation in the

macrocellular outdoor scenario is different from the corresponding microcellular case in that the outdoor

WINNER II D1.1.2 V1.2

Page 20 (82)

environment produces higher delay spreads and higher path-losses than the indoor environment.

Propagation through building walls and inside the building is assumed to be quite similar in both cases.

The parameters of this scenario have been merged with A2 and B4 in table 4-7.

2.3.12 D1 – Rural macro-cell

Propagation scenario D1 represents radio propagation in large areas (radii up to 10 km) with low building

density. The height of the AP antenna is typically in the range from 20 to 70 m, which is much higher

than the average building height. Consequently, LOS conditions can be expected to exist in most of the

coverage area. In case the UE is located inside a building or vehicle, an additional penetration loss is

experienced which can possibly be modelled as a (frequency-dependent) constant value. The AP antenna

location is fixed in this propagation scenario, and the UE antenna velocity is in the range from 0 to 200

km/h.

In WINNER Phase I, measurements were conducted in a flat rural environment near Oulu in Finland, at

both 2.45 and 5.25 GHz, and with an AP antenna height of 18 - 25 m. A channel model derived from

these measurements is available and has been reported in [WIND54]. The channel model from Phase I for

propagation scenario D1 is generalised for the frequency range 2 – 6 GHz and different BS and MS

antenna heights.

2.3.13 D2 – Moving networks

Propagation scenario D2 ("Rural Moving Network") represents radio propagation in environments where

both the AP and the UE are moving, possibly at very high speed, in a rural area. A typical example of this

scenario occurs in carriages of high-speed trains where wireless coverage is provided by so-called moving

relay stations (MRSs) which can be mounted, for example, to the roof. The link between the fixed

network and the moving network (train) is typically a LOS type. Later we call this link as D2a. In

addition there is a link from the MRS to the UE. It is assumed that the indoor part of the MRS is mounted

in the ceiling in the middle of the carriage. Later on we call this link D2b.

2.3.13.1 D2a

The scenario for D2a is specified as follows:

- There is a track accompanied with base stations in the intervals of 1000 - 2000 m.

- The base stations are

 50 m away from the tracks and the antenna heights are 30 m, or alternatively

 2 m away from the tracks and the antenna heights are 5 m.

- Height of the train (and MRS) is 2.5 m

- Speed of the train is nominally 350 km/h.

No tunnels are assumed in the route, but the lower BS antenna height can be used to simulate situations

compatible with the ones encountered in tunnels as regards high change rate in Doppler frequencies.

2.3.13.2 D2b

D2b model is applicable in an environment inside the fast train carriage. The carriage is assumed to

consist of one floor, but this should not make big difference, because one floor of a double floor carriage

is quite similar as a single floor carriage. The MRS indoor part is assumed to be located in the ceiling of

the carriage. It is assumed that there are chairs and tables densely as usual in train carriages. This makes

that typically there is NLOS connection between the MRS and UE. Finally, it is assumed that the

windows of the carriage are made of heat protective glass. This is important, because then we can assume

that the relatively very fast moving scatteres do not affect considerably to the propagation. The reason is

that such heat protective glass attenuates radio waves about 20 dB in both directions giving a total

attenuation of 40 dB to the signals transmitted out from the carriage, scattered in the outside environment

and penetrated back to the interior of the carriage.

2.4 Measurement Tools

Five different radio channel measurement systems have been used in the propagation measurements

during Phase I and II. Main characteristics of the channel sounders used in Phase II are summarised in

this section. Measuring equipment used in Phase I have been described in [WIN1D54].

WINNER II D1.1.2 V1.2

Page 21 (82)

2.4.1 Propsound (EBITG, UOULU, Nokia)

The Propsound™ multi-dimensional radio channel sounder is a product of Elektrobit, Finland [PSound].

Propsound has been designed to enable realistic radio channel measurements in both the time and spatial

domains. It is based on the spread spectrum sounding method in the delay domain. The other domains,

including polarization, FDD frequency and the spatial domain, are covered using an advanced time-

domain switching technique. Together with optional super-resolution techniques (based on the SAGE

algorithm), this allows accurate measurements of SISO, SIMO, MIMO, geolocation and multi-user

propagation channels. Some key features of Propsound are presented in Table 2-2.

Table 2-2 Propsound

characteristics

Propsound Property Range of values

RF bands 1.7 - 2.1, 2.0 - 2.7, 3.2 - 4.0, 5.1 - 5.9 GHz

Sustained measurement rate Up to 30,000 CIR/s (code length: 255 chips)

Maximum cycle (snapshot) rate 1500 Hz

Chip frequency up to 100 Mchips/s

Available code lengths 31 - 4095 chips (M-sequences)

Number of measurement channels up to 8448

Measurement modes SISO, SIMO, MIMO

Receiver noise figure better than 3 dB

Baseband sampling rate up to 2 GSamples/s

Spurious IR free dynamic range: 35 dB

Transmitter output up to 26 dBm (400 mW), adjustable in 2 dB steps

Control Windows notebook PC via Ethernet

Post processing MATLAB package

Synchronisation rubidium clock with stability of 10

-11

Table 2-3 Propsound

terminals.

Trasmitter with a trolley.

Receiver with a trolley.

WINNER II D1.1.2 V1.2

Page 22 (82)

Table 2-4 Propsound

antennas.

Name ODA_5G25 PLA_5G25 UCA_5G25

Owner Elektrobit Elektrobit Elektrobit

Array structure omnidirectional array rectancular array uniform circular array

Polarization dual (+/- 45°) dual (+/- 45 °) vertical

Center frequency

[GHz]

5.25 5.25 5.25

Number of elements 50 (25 dual) 32 (16 dual) 8

Element type patch patch monopole

Picture

Name SPH_5 PLA_5 Mockup

Owner Radio Laboratory / Helsinki

Univ. of Technology Radio Laboratory / Helsinki

Univ. of Technology Nokia Research Center

Array structure Semi-spherical array Planar array Terminal mockup

Polarization dual (H/V) dual (+/- 45°) -

Center frequency

[GHz]

5 5 5

Number of elements 42 (21 dual) 32 (16 dual) 4

Element type patch patch -

Picture

2.4.2 TUI sounder

The RUSK TUI-FAU channel sounder used at TU Ilmenau for MIMO measurements was designed by

Medav, Germany [Medav]. RUSK is a real-time radio channel impulse response measurement system that

supports multiple transmit and receive antenna element configurations.

The RUSK MIMO channel sounder measures the channel response matrix between all transmitting and

receiving antenna elements sequentially by switching between different (Tx,Rx) antenna element pairs.

WINNER II D1.1.2 V1.2

Page 23 (82)

This means that the sounder uses only one physical transmitter and receiver channel, which reduces

sensitivity to channel imbalance. The switched-antenna approach offers a simple way of changing the

effective number of antenna elements in the array. Additionally, since antennas are not transmitting at the

same time, separation of transmitted signals at the receiver side is straightforward. To accomplish

synchronous switching, rubidium reference oscillators are used at both the transmitter and the receiver.

Timing and switching frame synchronization is established during an initial synchronization process prior

to measurement data recording and must be maintained during the entire measurement.

For channel excitation RUSK uses a multi-carrier spread spectrum signal (MCSSS) with an almost

rectangular shape in the frequency domain. This approach allows precise concentration of the transmitted

signal energy in the band of interest. Simultaneous sounding of multiple bands (e.g., separated up- and

down-link bands in FDD) is supported by setting some spectral magnitudes to zero.

Table 2-5 summarizes the key features of the RUSK TUI-FAU channel sounder.

Table 2-5 Key features of the Medav RUSK TUI-FAU channel sounder.

RUSK TUI-FAU Sounder Property Range of values

RF bands 5…6 GHz

Max. meas. data storage rate (2x)*160 Mbyte/s

Test signal Multi Carrier Spread Spectrum Signal (MCSSS)

Sequence length

(defines maximum excess delay)

256 – 8192 spectral lines, depending on IR length

Number of measurement channels up to 65536 (2

)

Measurement modes SISO, SIMO, MIMO

Sampling frequency 640 MHz at Tx and Rx

Spurious free IR dynamic range 48 dB

Transmitter output up to 33 dBm (2 W),

Propagation delay resolution 4.17 ns (1/bandwidth)

Impulse response length 0.8 µs – 25.6 µs

RF sensitivity -88 dBm

Control Windows PC

Post processing MATLAB package

Synchronisation rubidium clock with stability of 10

-10

Rate is doubled with additional disk storage. Second storage enables shorter time gap between Tx-Rx sub-channels.

An overview of measurement-relevant technical data for the antenna arrays used in the TU-Ilmenau

campaigns is given in Table 2-6.

Table 2-6 Overview of TU-Ilmenau antenna arrays.

Name PULA8

(PULA8@10W) UCA16 PUCPA24 SPUCPA4x24

Vendor IRK Dresden TU Ilmenau IRK Dresden IRK Dresden

Array structure uniform linear array

uniform circular

array uniform circular

array stacked uniform

circular array

Polarization dual (vertical+

horizontal) vertical dual (vertical+

horizontal) dual (vertical+

horizontal)

WINNER II D1.1.2 V1.2

Page 24 (82)

Name PULA8

(PULA8@10W) UCA16 PUCPA24 SPUCPA4x24

Center

frequency

[GHz]

5.2 5.2 5.2 5.2

Bandwidth

[MHz]

120 120 120 120

Max. Power

[dBm]

27 (40) 27 25 24

Number of elements 8 16 24 96

Element type patch disk cone patch patch

Dimensioning

element spacing

0.4943 λ diameter

10.85 cm diameter

19.5 cm diameter 19.5 cm

ring spacing 0.4943 λ

Element orientation

Picture

The monopole antenna that is mounted on the ICE roof was manufactured by Huber&Suhner, and is of

type SWA 0859 – 360/4/0/DFRX30. The disc-conical antenna used for the ICE SISO measurements was

2.4.3 CRC sounder

The sounder used for the CRC measurements is the fourth generation of a PN sounder design that was

first implemented with 20 MHz bandwidth at CRC in 1981. Its construction is bread-board style, with

semi-rigid cables connecting various commercially-available modules, such as phase-locked oscillators,

power splitters, mixers, filter modules, and amplifiers. The bread-board style construction is maintained

so as to allow easy reconfiguration and recalibration for different measurement tasks, with different

operating frequencies and different bandwidths, as required. Its PN sequence generator is a CRC

implementation that can generate sequences of length between 127 and 1021 chips, and it can be clocked

at rates up to 65 mchips/s. Both CRC-Chanprobe's transmitter and its receiver have two RF sections with

operating bandwidths centred on 2.25 GHz and 5.8 GHz. The transmitter transmits continuously in both

bands. Operation at other frequencies is made possible by substituting different up-converter PLOs and

bandpass filters.

The receiver front ends are connected sequentially, using an RF switch, to its IF section. Operation at

other centre frequencies is accomplished via an extra, external RF section, with frequency translation to

either 2.25 or 5.8 GHz. Final downconversion is from IF to baseband via quadrature downconversion

circuitry. The in-phase (I) and quadrature (Q) baseband outputs can each be sampled at rates up to 100

MSamples/s. CRC-Chanprobe's operating characteristics are summarized in Table 2-7.

WINNER II D1.1.2 V1.2

Page 25 (82)

Table 2-7 CRC-Chanprobe operating characteristics

CRC-Chanprobe Property Range of values

RF bands 0.95, 2.25, (4.9), 5.8, 30, 40, 60 GHz

[1]

Sustained measurement rate 10,000 snapshots/s

[2]

Maximum cycle (snapshot) rate 40,000 snapshots/s

[3]

Chip rate up to 50 Mchips/s

Useable code lengths 127 – 1021 chips (M-sequences)

Number of measurement channels 32 Switched Rx antennas, 1 Tx antenna

Measurement modes SISO, SIMO

Receiver noise figure < 2 dB

Baseband sampling rate 100 MSamples/s

Spurious IR free dynamic range: 40 dB

Transmitter output up to 42 dBm at 2.25 GHz, up to 30 dBm at other

frequencies

Control Windows PC

Post processing MATLAB package

Synchronisation rubidium clock with stability of 10

-11

Minimum Received Power level (20 dB

MPSR) -89 dBm

Linear Dynamic Range without pre-

attenuation -69 dBm to -89 dBm with 20 dB MPSR

Transmit Antenna Vertical Quarter-Wavelength Monopole, with drooping

radials

Receive Antenna 32 Element UCA of Vertical Quarter-Wavelength

Monopoles with drooping radials

Note: Transfer rate specs are quoted assuming a single Rx channel, 50 Mchps m-sequence, sequence length 255

chips, 2 samples per chip, 1 sequence length per snapshot.

1) 0.95, 4.9 & 5.8 GHz characteristics are SISO

2) Based on a verified average data acquisition rate of ~20 Mbytes/S when logging data to hard disk in real

time (needed for long measurement runs).

3) Based on a verified average data acquisition rate of ~80 Mbytes/S when not logging data to hard disk in

real time (valid for short measurement runs).

CRC-Chanprobe can be operated in SISO or SIMO modes. A 32-element switched uniform circular array

and a 32-element 3D cross array have been implemented for use at the receiver. Both arrays employ

quarter-wavelength monopole antennas for the reception of vertically polarized waves.

WINNER II D1.1.2 V1.2

Page 26 (82)

3. Channel Modelling Approach

WINNER channel model is a geometry based stochastic model. Geometry based modelling of the radio

channel enables separation of propagation parameters and antennas. The channel parameters for

individual snapshots are determined stochastically, based on statistical distributions extracted from

channel measurement. Antenna geometries and field patterns can be defined properly by the user of the

model. Channel realisations are generated with geometrical principle by summing contributions of rays

(plane waves) with specific small scale parameters like delay, power, AoA and AoD. Superposition

results to correlation between antenna elements and temporal fading with geometry dependent Doppler

spectrum [Cal+07].

A number of rays constitute a cluster. In the terminology of this document we equate the cluster with a

propagation path diffused in space, either or both in delay and angle domains. Elements of the MIMO

channel, i.e. antenna arrays at both link ends and propagation paths, are illustrated in Figure 3-1.

Path N

Array 1

(S Tx elements)

Array 2

(U Rx elements)

1, rx

Urx,

Stx,

1, tx

Path 1

Figure 3-1 The MIMO channel

Transfer matrix of the MIMO channel is

( ) ( )

∑

nn tt 1 ;;

ττ

(3.1)

It is composed of antenna array response matrices F

for the transmitter, F

for the receiver and the

propagation channel response matrix h

for cluster n as follows

ϕφφϕφτϕτ

ddtt

txnrxn

,,;; FhFH

(3.2)

The channel from Tx antenna element s to Rx element u for cluster n is

( )

( ) ( )

mnmn

stxmnurxmn

mnHstx

mnVstx

HHmnHVmn

VHmnVVmn

mnHurx

mnVurx

nsu

rjrj

0,,

,,,

,,,,

,,,

2exp

2exp2exp

ττδπυ φπλϕπλ

ααα

−× ⋅⋅×





































−−

∑

(3.3)

where F

rx,u,V

and F

rx,u,H

are the antenna element u field patterns for vertical and horizontal polarisations

respectively,

n,m,VV

and

n,m,VH

are the complex gains of vertical-to-vertical and horizontal-to-vertical

polarisations of ray n,m respectively. Further

is the wave length of carrier frequency,

is AoD unit

vector,

is AoA unit vector,

stx

and

urx

are the location vectors of element s and u respectively,

and

n,m

is the Doppler frequency component of ray n,m. If the radio channel is modelled as dynamic, all

the above mentioned small scale parameters are time variant, i.e. function of t. [SMB01]

For interested reader, the more detailed description of the modelling framework can be found in

WINNER Phase I deliverable [WIN1D54].

WINNER II D1.1.2 V1.2

Page 27 (82)

3.1

WINNER Generic Channel Model

WINNER generic model is a system level model, which can describe arbitrary number of propagation

environment realisations for single or multiple radio links for all the defined scenarios for desired antenna

configurations, with one mathematical framework by different parameter sets. Generic model is a

stochastic model with two (or three) levels of randomness. At first, large scale (LS) parameters like

shadow fading, delay and angular spreads are drawn randomly from tabulated distribution functions.

Next, the small scale parameters like delays, powers and directions arrival and departure are drawn

randomly according to tabulated distribution functions and random LS parameters (second moments). At

this stage geometric setup is fixed and only free variables are the random initial phases of the scatterers.

By picking (randomly) different initial phases, an unlimited number of different realisations of the model

can be generated. When also the initial phases are fixed, the model is fully deterministic.

3.1.1 Modelled parameters

Parameters used in the WINNER II Channel Models have been listed and shortly explained below.

Parameter values are given in a later section, see sub-section 4.4.

The first set of parameters is called large scale (LS) parameters, because they are considered as an

average over a typical channel segment i.e. distance of some tens of wave-lengths. First three of the large

scale parameters are used to control the distributions of delay and angular parameters.

Large Scale Parameters

• Delay spread and distribution

• Angle of Departure spread and distribution

• Angle of Arrival Spread and distribution

• Shadow Fading standard deviation

• Ricean K-factor

Support Parameters

• Scaling parameter for Delay distribution

• Cross-polarisation power ratios

• Number of clusters

• Cluster Angle Spread of Departure

• Cluster Angle Spread of Arrival

• Per Cluster Shadowing

• Auto-correlations of the LS parameters

• Cross-correlations of the LS parameters

• Number of rays per cluster

All of these parameters have been specified from the measurement results or, in some cases, found from

literature. Number of rays per cluster has been selected to be 20 as in [3GPPSCM]. Analysis of the

measurement data for the different parameters has been described in the Part II document of this

deliverable. In the WINNER Channel Models the parameters are assumed not to depend on distance.

Although this assumption is probably not strictly valid, it is used for simplicity of the model. The

parameter values are given in paragraph 4.4 and represent expected values over the applicability range.

In the basic case the Angles of Arrival and Departure are specified as two-dimensional, i.e only azimuth

angles are considered. For the indoor and outdoor-to-indoor cases the angles can also be understood as

solid angles, azimuth and elevation, and the modelling can be performed also as three-dimensional.

3.2

Modelling process

The WINNER Channel Modelling Process is depicted in Figure 3-2. The process is divided into three

phases. The first phase starts from definition of propagation scenarios, which means selection of

environments to be measured, antenna heights, mobility, and other general requirements. Generic model

is needed to know what parameters have to be measured. Planning of measurement campaign can be

started when scenarios and generic model exist. Campaign planning has to be done carefully to take into

WINNER II D1.1.2 V1.2

Page 28 (82)

account several aspects – e.g. channel sounder setup, measurement route, link budget. Channel

measurements are done according to the campaign planning and documented accurately. Measurement

data is stored onto a mass memory (e.g. magnetic tape or hard disk).

The second phase of the channel modelling process concentrates on data analysis. Depending on the

required parameters, different analysis methods and items are applied. Output of data post-processing

could be, e.g., a set of impulse responses, path-loss data, or extracted multidimensional propagation

parameters. For the post-processed data, statistical analysis is done to obtain parameter PDFs.

The third phase of the channel modelling process covers the items required in simulation. Parameters are

generated according to the PDFs, by using random number generators and suitable filters. MIMO transfer

matrix is obtained by using the generated parameters, and information about the antennas. In our

approach MIMO transfer matrices are generated by using the sum-of-rays method. Generated impulse

responses are called channel realisations, which are then used in simulations.

generic

model

measurement

data

measurement

data

measurement

data

Campaign

planning Channel

measurements

parameter

PDFs

parameter

PDFs

parameter

PDFs

data post-

processing /

analysis

parameter

generation MIMO transfer

matrix

generation

parameter

PDFs

parameter

PDFs

channel

realisations

array

responses

Simulations

generic

model

measurement

data

measurement

data

measurement

data

Campaign

planning Channel

measurements

measurement

data

measurement

data

measurement

data

Campaign

planning Channel

measurements

parameter

PDFs

parameter

PDFs

parameter

PDFs

data post-

processing /

analysis

parameter

generation MIMO transfer

matrix

generation

parameter

PDFs

parameter

PDFs

channel

realisations

array

responses

Simulations

Figure 3-2 WINNER channel modelling process

3.3

Network layout

WINNER MIMO radio channel model enables system level simulations and testing. This means that

multiple links are to be simulated (evolved) simultaneously. System level simulation may include

multiple base stations, multiple relay stations, and multiple mobile terminals as in Figure 3-3. Link level

simulation is done for one link, which is shown by blue dashed ellipse. The short blue lines represent

channel segments where large scale parameters are fixed. System level simulation consists of multiple

links. Both link level and system level simulations can be done by modelling multiple segments, or by

only one (CDL model).

WINNER II D1.1.2 V1.2

Page 29 (82)

FRS

segments

link

Figure 3-3. System level approach, several drops.

A single link model is shown in Figure 3-4. The parameters used in the models are also shown in the

figure. Each circle with several dots represents scattering region causing one cluster. The number of

clusters varies from scenario to another.

Figure 3-4. Single link.

In spatial channel model the performance of the single link is defined by small-scale parameters of all

MPCs between two spatial positions of radio-stations. According to this, if only one station is mobile

(MS), its position in space-time is defining a single link. The more complex network topology also

includes multihop links [Bap04] and cooperative relaying [Lan02], however more complex peer-to-peer

connections could be easily described as collections of direct radio-links.

Large-Scale Parameters (LSP) are used as control parameters , when generating the small-scale channel

parameters. If we are analyzing multiple positions of MS (many MSs or multiple positions of the single

MS) we have a multiple-link model for system level simulations. It can be noted that different MSs being

at the same spatial position will experience same LSP parameters.

For multi-link simulations some reference coordinate system has to be established in which positions and

movement of radio-stations can be described. A term network layout is designating complete description

WINNER II D1.1.2 V1.2

Page 30 (82)

of the relative positions of the system elements, as well as vectored description of their movements

(speeds). In general, positions (coordinates) of scatterers are unknown. Only exceptions are related to far

cluster scatterers (FCS) that are actually positioned in the same coordinate system as radio-stations. In

multi-link simulations spatial correlations of channel parameters are important. In order to establish

correlations between links at system level the LSPs have been generated with the desired correlation

properties. This has been described in the following subsection.

3.3.1 Correlations between large scale parameters

For single position of radio-stations (one link) we can describe inter-dependence of multiple control

parameters (LSP) with correlation coefficient matrix. Correlations of LSPs that are observed in measured

data are not reflected in joint power or probability distributions. Instead LSPs are estimated from

marginal power distributions (independently for angles and delays), and necessary dependence is re-

established through cross-correlation measure:

yyxx

, (3.4)

where

is the cross-covariance of LS parameters x and y.

At system level two types of correlations could be defined: a) between MSs being connected to the same

BS and b) correlations of links from the same MS to multiple BSs (Figure 3-5). These correlations are

mostly caused by some scatterers contributing to different links (similarity of the environment).

a) b)

Figure 3-5 Links toward common station will exibit inter-correlations: a) fixed common station, b)

mobile common station

In the first case WINNER models are using exponential correlation functions to describe dependence of

LSP changes over distance. In other words LSPs of two MSs links toward same BS would experience

correlations that are proportional to their relative distance d

. As a consequence correlation coefficient

matrices for neighbouring links (for MSs at certain distance) are not independent and they also have to

reflect observed correlations over the distance dimension:

yyxx

MSxy

)(

, (3.5)

For this reason elements of link cross-correlations coefficient matrix should reflect exponential decay

with distance, as shown in Figure 3-6

WINNER II D1.1.2 V1.2

Page 31 (82)

)0( ρ













∝∆

−MS

ed )( ρ

Figure 3-6 Dependence of cross-correlation coefficient matrix over distance.

In 3GPP SCM, shadowing fading for links from one MS to different BSs exhibits constant correlation

coefficient equal to 0.5. This correlation does not depend on distances between BSs or their relative

angular positions as seen from MS and therefore it is not layout dependent. Additionally, this property is

estimated from few measurements and therefore it is not considered as being fully representative for

different WINNER scenarios. This phenomenon has also been investigated in WINNER project in some

extent.

Correlation properties of links from the same MS to multiple BSs (inter-site) were investigated in Phase I

of WINNER project [WIN1D54]. The results showed rather high correlation for one measurement route

and quite low for another. The amount of measurement data was limited, so that we could not specify

correlation other than zero.

The inter-site correlation of shadowing fading is also investigated in the literature for the outdoor macro-

cell scenarios: in [Gud91],[PCH01] and [WL02], the authors proposed that the inter-site correlation is a

function of the angle between BSs directions being seen from the MS (

); while in [Sau99] author

studied the dependence of the inter-site correlation on the distance between BSs,

. Although some

correlations could be found in the references afore, the results could not show clear correlation behaviour

between different BS:s. Although we also believe that such correlation most probably exists in many

scenarios, at least between Base Stations near each other, at this point we decided to let the correlation be

modelled as zero.

Inter-correlations between links of one MS to multiple sectors of the same BS could be analyzed in a

similar way, by treating different sectors of the BS as independent one-sector BSs. As a matter of fact, the

links from two different sectors to an MS are correlated so that the LS parameters for the links are the

same.

Correlation of large-scale parameters (LSPs) is achieved by using wighed sums of independent Gaussian

random processes (IGRP). If i-th LSP,

, have distribution that differs from Normal (Gauss), required

distribution is generated by applying mapping from random variable

having Gaussian distribution.

Random variable

will be referred as transformed LSP (TLSP). Prior of mapping

correlated with TLSPs

, belonging to other LSPs or different links (being at certain distance - for

system level correlations). Process applied to introduce or to calculate correlations (from measured data)

is illustrated in Figure 3-7.

WINNER II D1.1.2 V1.2

Page 32 (82)

)(⋅

)(

⋅

−

)(⋅

)(

⋅

−

Figure 3-7 Correlations of LSP are introduced in transformed domain.

In cases when mapping

sgs ii

1 −

is unknown, necessary relations between LSP and transformed

domain can be established using knowledge about cumulative density distribution (cdf) of

)(

. In

such cases

can be generated from

using expression:

(

sFFsgs

ssi −−

(3.6)

where

)

(

is cdf of Normally distributed process that can be calculated using Q-function (or erf/erfc).

In simpler cases, e.g. when LSP is log-normally distributed it is possible to use known mappings:

sgs

−

(3.7)

ssgs

i10

log

(3.8)

As a correlation measure cross-correlation coefficient is used, expression (3.4). Above is explained that

for one link (single position of MS) inter-dependence of multiple control parameters can be described

with correlation coefficient matrix. Additionally if parameters of intra-site links are correlated according

to distance between MS positions, then correlation matrix gets additional dimension that describes

changes in correlations over distance, Figure 3-6. This means that for each pair of TLSP we can define

cross-correlation coefficient dependence over distance, as in expression (3.5):

llkk

ssrr

lksr

~~~~

)(

)( =

(3.9)

Cross-variances

)(

~~ lksr

are calculated from measurement data using knowledge about positions of

MS during measurement, and in general have exponential decay over distance.

If each link is controlled by M TLSPs, and we have K links corresponding to MS locations at positions

Kk ..1

, then it is necessary to correlate values for N= M

K variables.

Generation of N Normally distributed and correlated TLSPs can be based on scaling and summation of N

independent zero-mean and unit variance Gaussian random variables,

NNNN

yxyxyx ,,,,),(

111

ξξ

K = ξ

. Using matrix notation that can be expressed:

),(),(

yxyx

NNxN

(3.10)

This will ensure that final distribution is also Gaussian. Scaling coefficients have to be determined in such

way that cross-variances

)(

~~ lksr

dC lk

)()(

lklklk

yyxxd −+−=

are corresponding to measured

values. If element

of matrix

NxN

represents cross-variance between TLSPs

and

, then

scaling matrix can be calculated as:

NxNNxN

(3.11)

This approach is not appropriate for correlation of large number of parameters, since dimensions of

scaling matrix are increasing proportionally to the total number of TLSPs in all links (squared dependence

in number of elements). For that reason it is more convenient to generate separately the influence of LSP

cross-correlation and exponential auto-correlation.

WINNER II D1.1.2 V1.2

Page 33 (82)

Let us assume we have M LSPs per link and K correlated links, i.e. K MSs linked to the same BS site at

locations (x

), where k = 1,…,K. Auto-correlation is generated to the LSPs the following way. At first

we generate a uniform grid of locations based on co-ordinates of the K MSs. Size of the grid is

DyyDxx

kkkk

2)min()max(2)min()max(

. To each grid node we assign M Gaussian iid

~N(0,1) random numbers, one for each LSP. Then the grid of random numbers is filtered with a two

dimensional FIR filter to generate exponential auto-correlation. Impulse response of the filter for the mth

LSP is

( )













∆

−=

dh exp

, (3.12)

where d is distance and ∆

is the correlation distance both in meters (see Table 4-5). Each of the M

random numbers in nodes of the grid, representing M LSPs, is filtered with a specific filter, because the

correlation distances may be different in Table 4-5. After filtering the correlated random numbers

),(

kkM

at K grid nodes (K MS locations) are saved and the redundant grid nodes are discarded.

Cross-correlation is generated independently to the LSPs of K links by linear transformation

),()0(),(

kkMMxMkk

yxyx

, (3.13)

where elements of correlation matrix













=)0()0(

)0()0(

)0(

~~~~

111

MMM

ssss

MxM

MOM

(3.14)

are defined in Table 4-5.

3.4 Concept of channel segments, drops and time evolution

Channel segment represents a period of quasi-stationarity during which probability distributions of low-

level parameters are not changed noticeably. During this period all large-scale parameters, as well as

velocity and direction-of-travel for mobile station (MS), are practically constant. To be physically

feasible, the channel segment must be relatively confined in distance. The size depends on the

environment, but it can be at maximum few meters. Correlation distances of different parameters describe

roughly the proper size of the channel segment, see the paragraph 4.4.

Allowing the channel segment length go to zero, we specify a drop: In a drop all parameters are fixed,

except the phases of the rays. Motion within a drop is only virtual and causes fast fading and the Doppler

effect by superposition of rotating phasors, rays. It can be said, that a drop is an abstract representation of

a channel segment, where the inaccuracies caused by the change of the terminal location have been

removed. In a simulation, the duration of a drop can be selected as desired. It is a common practice to use

drops in the simulations. The main advantage is the simplicity of the simulation, because successive

simulation runs do not need to be correlated. The drawback is that it is not possible to simulate cases,

where variable channel conditions are needed. However, the drop-based simulation is the main method of

simulations in WINNER projects I and II. In the final WINNER II Channel Models there is also an

alternative for the drop-based simulation, i.e. simulation with time evolution., where correlated drops are

used

In the WINNER II models the propagation parameters may vary over time between the channel segments.

In the multi segment modelling two options are available, either drops (stationary channel segments like

in WINNER I) or continuous channel evolution with smooth transitions between segments. There are two

approaches for time evolution modelling discussed below. First is the one that is proposed to be

implemented, due to the simplicity of the method. Second is a method using Markov process that can be

regarded as a more advanced method and it requires parameters that have not been determined yet.

3.4.1 Basic method for time-evolution

In this report time evolution of propagation parameters is modelled like depicted in Figure 3-8. The route

to be modelled is covered by adjacent channel segments. The distance between segments is equal to the

stationarity interval. Transition from segment to segment is carried out by replacing clusters of the "old"

segment by the clusters of the "new" segment, one by one. The route between adjacent channel segments

WINNER II D1.1.2 V1.2

Page 34 (82)

is divided to number of sub-intervals equal to maximum number of clusters within the channel segments.

During each sub-interval the power of one old cluster ramps down and one new cluster ramps up. Power

ramps are linear. Clusters from the old and new segments are coupled based on their power. If number of

clusters is different in the channel segments, the weakest clusters are ramped up or down without a pair

from other cluster.

time

delay

amplitude

Figure 3-8 Smooth transition between channel segments by power ramp-up and ramp-down of

clusters.

3.4.2 Markov process based method of time evolution

In [ZTL+05] the authors propose a dynamic channel model, where paths are arised and disappeared

according to a Markov process. The birth and death probabilities are specified from measurements. This

approach leads to a more realistic behaviour of the channel. However, to apply this approach, the birth

and death parameters are needed for all the channels, which are not available at the moment. Another

disadvantage is the variable number of instantaneous paths.

In spite of the drawbacks listed above this approach seems quite promising, and should be investigated

and adopted in a later stage, if the benefits are deemed more important than the disadvantages. One way

would be to use only the N strongest paths in the model based on the Markov process, where N is a

constant.

3.5 Nomadic channel condition

Propagation environment is called nomadic, if the transmitter and receiver locations are normally fixed

during the communication, but may have moved between different uses of the network [OVC06]. In such

conditions we have to assume that some of the scatterers may move. Actually this is quite typical in many

cases, like when there are people working in the vicinity of the transceiver. For the nomadic environment

it is also typical that an access point and especially user terminals can change place, e.g. in the room and

even go out from the room. However, the most important feature to be taken into account in channel

modelling is the moving scatterers. Nomadic channels can be regarded as a special case of the WINNER

generic model shown in eq. (3.3). In principle, nomadic channels can exist in all the WINNER

deployment scenarios, both in indoor and outdoor. For feeder links we assume that the LOS component is

strong enough, so that the reflections from moving objects can be neglected. Therefore we use nomadic

modelling only for the scenarios A1 Indoor and B4 Outdoor-to-indoor.

Traditionally these scenarios have been modelled using very low speed for the User Equipment. By

applying an approach using fixed links with moving scatterers, we can certainly get more accurate

channel model and parameters for the generation of the channel coefficient.

The idea of modelling nomadic (or fixed) environments has been introduced in some open literature. Here

we follow the approach introduced in [OP04, OC07, Erc+01, ESB+04]. Based on measurements, we can

define a temporal K-factor, for both LOS and NLOS connections. Based on the temporal K-factor,

pathloss model including shadow fading, cross polariztion discrimination etc., the channel coefficients

can be generated [ESB+04]. In [ESB+04], 2x2 MIMO was discussed from theory, measurements,

generation of channel coefficients, and validation of the channels, but without information of angular

domain.

The overall procedure is roughly as follows. Assume that we have generated initial channel parameters

(delays, powers, AoA/AoD etc.) for the nomadic situation. Then we draw the clusters that are moving.

Next we draw the Doppler frequencies for all moving rays in all the clusters containing movement. (Note

that all or only part of the rays are moving in those clusters.) Next we can simply generate the channel

WINNER II D1.1.2 V1.2

Page 35 (82)

coefficients for whole the channel segment. In addition it is possible to define an extra attenuation or

cases, where a moving object (e.g a person) is shadowing paths from other scatterers. However, we

neglect this phenomenon for simplicity. The reasoning is as follows: The shadowing situation in the

indoor environment is assumed to be statistically the same, irrespective of the position of the scatterers.

Therefore we conclude that the measurements and literature results already contain this shadowed

situation, precisely enough for our modelling needs.

In indoors the moving objects (called clusters) are assumed to be humans. Reflection is the main

interaction with human body at WINNER target frequency range, as analysed in [VES00] and [GTD+04].

In our model only a cluster can be in linear motion for longer times, and this is modelled by an

accompanying mean cluster Doppler shift. A cluster is composed of 20 rays. If the scatterer described by

the cluster is assumed rigid, the relative movements come from the geometry and the movement of the

cluster, and can be directly calculated from the geometric model plus the known motion. In addition, there

are moving scatterers within a cluster (e.g. limbs), the parts of which are moving relatively. This

phenomenon can be governed e.g. through a Doppler spectrum assigned to a cluster.

Assumptions:

A cluster can be either moving or static.

A moving cluster has a random velocity that can be zero.

Static cluster, contains no movement at all, moving cluster can have a random fluctuation on top

of its mean movement (random velocity).

A moving cluster can shadow signals from other clusters. (Neglected here, as discussed afore.)

To create a model for the situation described afore, we have to fix the probabilities of static and moving

clusters and the accompanying distributions of the directions of the rays and the Doppler spectra of the

moving rays. The distributions for the directions of the rays, power levels etc. are all given by the

ordinary random process (i.e. non-nomadic) for the creating of the channel coefficients. All that remains

are the Doppler frequencies of the rays based on the virtual movement of the clusters. This means that, in

addition to the ordinary process, we have to specify:

the number of static clusters (e.g. 80% of all clusters),

mean velocity and direction for all moving clusters, with some velocities being possibly

zero (e.g. 50% zero velocity, 50% 3km/h, direction ~Uni(360°) (uniformly distributed

over 360°)),

additional Doppler frequency for each of the moving scatterers (e.g. calculated by ray

AoA/AoD, velocity 3km/h, direction of motion ~Uni(360°)),

The number of moving scatterer in a cluster is determined by targeted cluster-wise temporal K-factor. The

temporal K-factor will be K

= F/S, where F is the number of fixed rays and S is the total number of rays

per cluster.

3.6 Reduced complexity models

A need has been identified for reduced-complexity channel models that can be used in rapid simulations

having the objective of making comparisons between systems alternatives at link-level (e.g. modulation

and coding choices). In this report, such models are referred to as reduced-complexity models, and have

the character of the well-known tapped delay line class of fading channel models. However, to address

the needs of MIMO channel modelling, temporal variations at the taps are determined by more detailed

information than that required for the specification of relative powers, envelope fading distributions, and

fading rates, which are typical inputs to traditional tapped delay line models.

Specifically, multipath AoD and AoA information is inherent in the determination of tap fading

characteristics. For these reasons, the reduced complexity models reported herein are referred to as

Cluster Delay Line (CDL) models. A cluster is centred at each tap. In general, each cluster is comprised

of the vector sum of equal-powered MPCs (sinusoids), all of which have the same or close to same delay.

Each MPC has a varying phase, but has fixed AoA and AoD offsets. The latter depend on the angular

spreads at the MS and the BS, respectively, as shown in Table 4-1. The values in this table were chosen to

realise a specified Laplacian PAS for each cluster, appropriate to the scenario being modelled. In cases

where there is a desire to simulate Ricean-like fading, an extra MPC is added, which is given a power

appropriate to the desired Rice factor, and zero angular offset. The powers and delays of the clusters can

be non-uniform, and can be chosen to realise the desired overall channel rms delay spread. Parameters of

all CDL models reflect the expected values of those used in the more complex models described in other

sections of this report.

WINNER II D1.1.2 V1.2

Page 36 (82)

Doppler information is not specified explicitly for CDL models. This is because Doppler is determined by

the AoAs of the MPCs, MS speed and direction, and the specified antenna patterns at the MS and BS,

upon which there are no restrictions, except in fixed feeder link scenarios, as discussed in the section of

feeder link models.

3.6.1 Cluster Delay Line models for mobile and portable scenarios

Cluster delay line (CLD) models for all mobile scenarios have been generated from the corresponding

generic models by selecting typical values from a set of random channel realisations. The CLD models

consist of the average power, mean AoA, mean AoD, and angle spreads at the BS and MS associated with

each cluster within the cluster delay line models. Tables of CDL parameters for the above-cited scenarios

can be found in Section 6. Although AoA and AoD values are fixed, it is recommended to have

directional variation for e.g. beamforming simulations by adding network layout related angle parameter

Ω

and

Ω

to all tabulated angles (see Figure 5-2).

3.6.2 Cluster Delay Line models for fixed feeder links

Only CDL models have been created for fixed feeder links (B5 scenarios). Model parameters have mostly

been derived from the literature as described in [WIN1D54], but some of them have been created by

applying models generated in WINNER. CDL models for B5 scenarios are given in the tables of Section

6. As for the mobile and portable scenarios, any desired antenna patterns can be chosen. However, for

scenarios B5a and B5b, at distances greater than 300 metres, the 3 dB beamwidth of the antenna at one

end of the link should be less than 10 degrees, while that at the other end of the link should be less than

53 degrees. Different parameters are specified in the cited tables for scenarios B5a, b, c, and d.

For fixed link scenarios B5a, B5b, B5d and B5f, Doppler shifts are independent of AOAs. Instead, they

are derived from considerations concerning the movement of interacting objects. One interacting object

per cluster is modelled as having motion, while the others are fixed. Associated Doppler frequencies are

specified in CDL tables. For the scenario B5c, two whole cluster are moving with random velocity.

3.6.3 Complexity comparison of modelling methods

Computational complexity of simulation of channel models is an important issue in system performance

evaluations. Complexity comparison of WINNER modelling approach with the popular correlation matrix

based method is studied in [KJ07]. A common supposition is that the correlation method is simpler and

computationally more effective than the geometric method. Conclusion of [KJ07] is that complexity of

both methods is about the same order of magnitude. With a high number of MIMO antenna pairs (>16)

correlation based method is clearly more complex.

The computation complexity is compared in terms of the number of "real operations". With the term "real

operations" is equated complexity of real multiplication, division, addition and table lookup. In Figure

3-9

the number of real operations per delay tap per MIMO channel time sample (matrix impulse response),

with different MxN MIMO antenna numbers, is depicted assuming 10 or 20 rays (M in eq. 3.3) and 8

order IIR filter in correlation matrix method. It was also noted that complexity of channel realisation

generation is several order of magnitudes lower than computational complexity of simulation of channel

convolution.

Figure 3-9. Computational complexity comparison.

WINNER II D1.1.2 V1.2

Page 37 (82)

4. Channel Models and Parameters

In this section, we summarize all the channel models and parameters. The path loss models are mainly

based on 5 GHz and 2 GHz measurements. However, the frequency bands are extended for 2 − 6 GHz

range.

It should be noted that the scenarios Indoor-to-Outdoor and Outdoor-to-Indoor have been combined and

represented by a single channel model in this deliverable. This combining has been discussed in Part II

document of this report.

4.1 Applicability

4.1.1 Environment dependence

Different radio-propagation environment would cause different radio-channel characteristics. Instead of

attempt to parameterize environment directly (e.g. street widths, average building height etc.) WINNER

models are using (temporal and spatial) propagation parameters obtained from channel measurements in

different environments. In this context, environments in which measurements are conducted to observe

radio-channel characteristics are called propagation scenarios. For each scenario measured data is

analyzed and complemented with results from literature to obtain scenario-specific parameters. After this

point, same generic channel is used to model all scenarios, just by using different values of channel

parameters.

Usually, even for the same scenario, existence of LOS component substantially influences values of

channel parameters. Regarding to this property, most WINNER scenarios are differentiating between

LOS and NLOS conditions. To enable appropriate scenario modelling, transition between LOS and

NLOS cases have to be described. For this purpose distance dependent probability of LOS is used in the

model.

4.1.2 Frequency dependence

Dependence on carrier frequency in WINNER model is found in path-loss models. All the scenarios

defined by WINNER support frequency dependent path-loss models valid for the ranges of 2 – 6 GHz.

The path-loss models are based on measurements that are mainly conducted in 2 and 5 GHz frequency

range. In addition the path-loss models are based on results from literature, like Okumura-Hata and other

well-known models [OOK+68], [OTT+01], which have been extended to the desired frequency range.

Path-loss frequency dependence has been considered in more detail in the paragraph 4.3.

From WINNER measurement results and literature survey it was found that model parameters DS, AS

and Ricean K-factor do not show significant frequency dependence [BHS05]. For that reason these

parameters show only dependence on environment (scenario).

For modelling of systems with time-division-duplex (TDD) all models are using same parameters for both

uplink and downlink. If system is using different carriers for duplexing (FDD), then (additionally to path

loss) random phases of scatterer contributions between UL and DL are modelled as independent.

For the WINNER purposes it is required that channel model supports bandwidths up to 100 MHz.

Following the approach described in [SV87] (for indoor propagation modelling) and further with SCME

[BHS05] WINNER II model introduces intra-cluster delay spread as a mean to support 100 MHz

bandwidth and to suppress frequency correlation. Instead of zero-delay-spread-cluster approach of Phase I

model, the two strongest clusters with 20 multipath components (MPCs) are subdivided into 3 zero-delay

sub-clusters. Thus we keep the total number of MPCs constant, but introduce four additional delay taps

per scenario.

4.2 Generation of Channel Coefficients

This section gives general description of the channel coefficient generation procedure, depicted also in

Figure 4-1. Steps of the procedure refer to parameter and model tables of Sections 4.2 to 4.4 give the

minimum description of the system level channel model.

WINNER II D1.1.2 V1.2

Page 38 (82)

Figure 4-1 Channel coefficient generation procedure

It has to be noted, that the geometric description covers arrival angles from the last bounce scatterers and

respectively departure angles to the first scatterers interacted from the transmitting side. The propagation

between the first and the last interaction is not defined. Thus this approach can model also multiple

interactions with the scattering media. This indicates also that e.g. the delay of a multipath component can

not be determined by the geometry.

General parameters:

Step 1: Set the environment, network layout and antenna array parameters

Choose one of the scenarios (A1, A2, B1,…)

Give number of BS and MS

Give locations of BS and MS, or equally distances of each BS and MS and relative

directions

LOS

and

LOS

of each BS and MS

Give BS and MS antenna field patterns F

and F

, and array geometries

Give BS and MS array orientations with respect to north (reference) direction

Give speed and direction of motion of MS

Give system centre frequency

Large scale parameters:

Step 2: Assign the propagation condition (LOS/NLOS) according to the probability described in Table

4-7.

Step 3: Calculate the path loss with formulas of Table 4-4 for each BS-MS link to be modelled.

Step 4: Generate the correlated large scale parameters, i.e. delay spread, angular spreads, Ricean K-factor

and shadow fading term like explained in section 3.2.1 (Correlations between large scale parameters).

Small scale parameters:

Step 5: Generate the delays

τ.

Delays are drawn randomly from the delay distribution defined in Table 4-5. With exponential delay

distribution calculate

Xr ln'

ττ

στ

−=

, (4.1)

where r

is the delay distribution proportionality factor,

is delay spread, X

~ Uni(0,1) and cluster

index n = 1,…,N. With uniform delay distribution the delay values

are drawn from the

corresponding range. Normalise the delays by subtracting with minimum delay and sort the

normalised delays to descending order.

WINNER II D1.1.2 V1.2

Page 39 (82)

'min'sort

nnn

τττ

−=

. (4.2)

In the case of LOS condition additional scaling of delays is required to compensate the effect of LOS

peak addition to the delay spread. Heuristically determined Ricean K-factor dependent scaling

constant is

000017.00002.00433.07705.0 KKKD ++−=

, (4.3)

where K [dB] is the Ricean K-factor defined in Table 4-5. Scaled delays are

LOS

ττ

, (4.4)

they are

not

to be used in cluster power generation.

Step 6: Generate the cluster powers P.

The cluster powers are calculated assuming a single slope exponential power delay profile. Power

assignment depends on the delay distribution defined in Table 4-5. With exponential delay distribution

the cluster powers are determined by

exp n

Ζ−

⋅











−

−=

ττ

(4.5)

and with uniform delay distribution they are determined by

10exp n

Ζ−

⋅











−

, (4.6)

where

~ N(0,

) is the per cluster shadowing term in [dB]. Average the power so that sum power of

all clusters is equal to one

∑

(4.7)

Assign the power of each ray within a cluster as P

/ M, where M is the number of rays per cluster.

Step 7: Generate the azimuth arrival angles

and azimuth departure angles

If the composite PAS of all clusters is modelled as wrapped Gaussian (see Table 4-5) the AoA are

determined by applying inverse Gaussian function with input parameters P

and RMS angle spread

maxln2

AoA

−

. (4.8)

On equation above

4.1

AoA

σσ

is the standard deviation of arrival angles (factor 1.4 is the ratio

of Gaussian std and corresponding "RMS spread"). Constant C

is a scaling factor related to total

number of clusters and is given in the table below:

# clusters 4 5 8 10 11 12 14 15 16 20

C 0.779

0.860

1.018

1.090

1.123

1.146

1.190

1.211

1.226

1.289

In the LOS case constant C is dependent also on Ricean K-factor. Constant C in eq. (4.10) is

substituted by C

LOS

. Additional scaling of angles is required to compensate the effect of LOS peak

addition to the angle spread. Heuristically determined Ricean K-factor dependent scaling constant is

0001.0002.0028.01035.1

KKKCC

LOS

+−−⋅=

, (4.9)

where K [dB] is the Ricean K-factor defined in Table 4-5.

Assign a positive or negative sign to the angles by multiplying with a random variable X

with

uniform distribution to discrete set of {1,–1}, add component

5,0N~

AoAn

to introduce

random variation

LOSnnnn

ϕϕϕ

++=

, (4.10)

where

LOS

is the LOS direction defined in the network layout description Step1.c.

WINNER II D1.1.2 V1.2

Page 40 (82)

In the LOS case substitute (4.12) by (4.13) to enforce the first cluster to the LOS direction

LOS

LOSnnnnn

YXYX

ϕϕϕϕ

−+−+=

. (4.11)

Finally add the offset angles

from Table 4-1 to cluster angles

mAoAnmn

αϕϕ

, (4.12)

where c

AoA

is the cluster-wise rms azimuth spread of arrival angles (cluster ASA) in the Table 4-5.

Table 4-1 Ray offset angles within a cluster, given for 1

°°

rms angle spread.

Ray number m Basis vector of offset angles

1,2 ± 0.0447

3,4 ± 0.1413

5,6 ± 0.2492

7,8 ± 0.3715

9,10 ± 0.5129

11,12 ± 0.6797

13,14 ± 0.8844

15,16 ± 1.1481

17,18 ± 1.5195

19,20 ± 2.1551

For departure angles

the procedure is analogous.

Step 7b If the elevation angles are supported: Generate elevation arrival angles

and elevation

departure angles

Draw elevation angles with the same procedure as azimuth angles on Step 7. Azimuth rms angle

spread values and cluster-wise azimuth spread values are replaced by corresponding elevation

parameters from Table 4-6.

Step 8: Random coupling of rays within clusters.

Couple randomly the departure ray angles

n,m

to the arrival ray angles

n,m

within a cluster n, or

within a sub-cluster in the case of two strongest clusters (see step 11 and Table 4-2).

If the elevation angles are supported they are coupled with the same procedure.

Step 9: Generate the cross polarisation power ratios (XPR)

for each ray m of each cluster n.

XPR is log-Normal distributed. Draw XPR values as

, (4.13)

where ray index m = 1,…,M , X ~ N(

) is Gaussian distributed with

and

from Table 4-5 for

XPR.

Coefficient generation:

Step 10: Draw the random initial phase

,,,

mnmnmnmn

ΦΦΦΦ

for each ray m of each cluster n and

for four different polarisation combinations (vv,vh,hv,hh). Distribution for the initial phases is uniform,

Uni(-

π,π

In the LOS case draw also random initial phases

hhvv

LOSLOS

ΦΦ

for both VV and HH polarisations.

Step 11: Generate the channel coefficients for each cluster n and each receiver and transmitter element

pair u,s.

For the N – 2 weakest clusters, say n = 3,4,…, N , and uniform linear arrays (ULA), the channel

coefficient are given by:

WINNER II D1.1.2 V1.2

Page 41 (82)

( )

( ) ( )

( )

tjjdjd

mnmnumns

mnHurx

mnVurx

hh mn

hvmnmn

vhmnmn

vvmn

mnHstx

mnVstx

nnsu

,,,

2expsin2expsin2exp

expexp

πυϕπλφπλ

κκ

−−

⋅

























ΦΦ ΦΦ













∑

(4.14)

where F

rx,u,V

and F

rx,u,H

are the antenna element u field patterns for vertical and horizontal polarisations

respectively, d

and d

are the uniform distances [m] between transmitter elements and receiver

elements respectively, and

is the wave length on carrier frequency. If polarisation is not

considered, 2x2 polarisation matrix can be replaced by scalar

exp Φ

and only vertically

polarised field patterns applied.

With the fixed feeder link models (B5 scenarios) the Doppler frequency component

n,m

is tabulated

for the first ray of each cluster. For the other rays

n,m

= 0. With all other models the Doppler

frequency component is calculated from angle of arrival (downlink), MS speed v and direction of

travel

cos

λθϕ

vmn

v−

, (4.15)

For the two strongest clusters, say n = 1 and 2, rays are spread in delay to three sub-clusters (per

cluster), with fixed delay offset {0,5,10 ns} (see Table 4-2). Delays of sub-clusters are

ns10

ns5

ns0

+= +=

ττ ττ

(4.16)

Twenty rays of a cluster are mapped to sub-clusters like presented in Table 4-2 below. Corresponding

offset angles are taken from Table 4-1 with mapping of Table 4-2.

Table 4-2 Sub-cluster information for intra cluster delay spread clusters.

sub-cluster # mapping to rays power delay offset

1 1,2,3,4,5,6,7,8,19,20 10/20

0 ns

2 9,10,11,12,17,18 6/20

5 ns

3 13,14,15,16 4/20

10 ns

In the LOS case define

nsunsu ,,,,

'HH =

and determine the channel coefficients by adding single line-

of-sight ray and scaling down the other channel coefficient generated by (4.14). The channel

coefficients are given by:

( ) ( )

( )

( ) ( )

( )

tjjdjd

LOSLOSuLOSs

LOSHurx

LOSVurx

LOS

LOSHstx

LOSVstx

nsu

πυϕπλφπλ

2expsin2expsin2exp

exp0

0exp

,,,,

−−

⋅





































−+

=HH

(4.17)

where

(

) is the Dirac's delta function and K

is the Ricean K-factor defined in Table 4-5 converted to

linear scale.

If non-ULA arrays are used the equations must be modified. For arbitrary array configurations on

horizontal plane, see Figure 4-2, the distance term d

in equations (4.14) and (4.17) is replaced by

uuu

mnu xyyx

',, sin

arctancos

ϕϕ

−+

, (4.18)

WINNER II D1.1.2 V1.2

Page 42 (82)

where (x

) are co-ordinates of uth element A

and A

is the reference element.

n,m

',, mnu

Figure 4-2 Modified distance of antenna element u with non-ULA array.

If the elevation is included (4.16) will be written as

( )

( ) ( )

( )

( ) ( )

( )

tjrjdrj

mnmnumns

mnHstx

mnVstx

hh mn

hvmnmn

vhmnmn

vvmn

mnHurx

mnVurx

nnsu

,,,

2exp2exp2exp

expexp

πυπλπλ

κκ

Ψ⋅Φ⋅⋅

























ΦΦ ΦΦ













−−

∑

(4.19)

where scalar product

mnsmnmnsmnmnsmns zyxr ,,,,,, sinsincoscoscos

γφγφγ

++=Φ⋅

, (4.20)

is location vector of Tx array element s,

mn,

is departure angle unit vector of ray n,m and x

, y

and z

are components of

to x,y and z-axis respectively,

mn,

is ray n,m arrival azimuth angle and

mn,

is ray n,m arrival elevation angle.

mnu

Ψ⋅

is a scalar product of Rx antenna element u and

arrival angle n,m.

Further on in the case of elevation assuming horizontal only motion, eq. (4.15) will be written as

,,,,

sincossincoscoscos

λφγθφγθ

mnmnvmnmnvmn

vvv +

Ψ⋅

. (4.21)

Step 12: Apply the path loss and shadowing for the channel coefficients.

4.2.1 Generation of bad urban channels (B2, C3)

Bad urban channel realizations can be created as modified B1 and C2 NLOS procedures as follows:

Step 1:

Drop five far scatterers within a hexagonal cell, within radius [FSmin, FSmax]. For FSmin and FSmax

values see Table 4-3. For each mobile user determine the closest two far scatteres, which are then used for

calculating far scatterer cluster parameters.

WINNER II D1.1.2 V1.2

Page 43 (82)

Table 4-3 Far scatterer radii and attenuations for B2 and C3.

Scenario FS

min

max

loss

B2 150 m 500 m 4 dB/µs

C3 300 m 1500 m 2 dB/µs

Step 2:

For C3 create 20 delays as described for C2 model in section 4.2. step 5. For the shortest 18 delays create

a typical urban C2 channel profile (powers and angles) as in section 4.2.

Similarly, create 16 delays for B1 NLOS, and for the shortest 14 delays create a typical B1 NLOS

channel profile as in section 4.2.

The last two delays in B2 and C3 are assigned for far scatterer clusters.

Step 3:

Create typical urban channel powers P

for FS clusters substituting equation (4.5) of section 4.2 step 6

with

Ζ−

, where

~ N(0,

) is the per cluster shadowing term in [dB].

Step 4:

Next create excess delays due to far scatterer clusters as

LOSMSFSBS

excess

−

>−>−

(4.22)

Attenuate FS clusters as FS

loss,

given in Table 4-3.

Step 5:

Select directions of departure and arrival for each FS cluster according to far scatterer locations. i.e.

corresponding to a single reflection from far scatterer.

It is worth noticing that depending on the location of the mobile user within the cell the FS clusters may

appear also at shorter delays than the maximum C2 or B1 NLOS cluster. In such cases the far scatterers

do not necessarily result to increased angular or delay dispersion. Also the actual channel statistics of the

bad urban users depend somewhat on the cell size.

4.3 Path loss models

Path loss models for the various WINNER scenarios have been developed based on results of

measurements carried out within WINNER, as well as results from the open literature. These path loss

models are typically of the form of (4.23), where d is the distance between the transmitter and the receiver

in [m], f

is the system frequency in [GHz], the fitting parameter A includes the path-loss exponent,

parameter B is the intercept, parameter C describes the path loss frequency dependence, and X is an

optional, environment-specific term (e.g., wall attenuation in the A1 NLOS scenario).

[ ]

CBdAPL













++= 0.5

GHz

log)m(log

1010

(4.23)

The models can be applied in the frequency range from 2 – 6 GHz and for different antenna heights. The

path-loss models have been summarized in Table 4-4, which either defines the variables of (4.23), or

explicitly provides a full path loss formula. The free-space path loss, PL

free

, that is referred to in the table

can be written as

)5.0(20log+46.4)(20log

1010free c

fdPL +=

(4.24)

The distribution of the shadow fading is log-normal, and the standard deviation for each scenario is given

in the table.

Frequency dependencies of WINNER path-loss models

The path loss models shown in Table 4-4 are based on measured data obtained mainly at 2 and 5 GHz.

These models have been extended to arbitrary frequencies in the range from 2 – 6 GHz with the aid of the

path loss frequency dependencies defined below. Following various results from the open literature, as

WINNER II D1.1.2 V1.2

Page 44 (82)

[RMB+06, CG+99, JHH+05, Rudd03, SMI+02, KI04, YIT06], the following frequency extensions are

employed for the frequency coefficient C shown in (4.23)

(1)

For all LOS deployment scenarios, and for all distances smaller than or equal to the breakpoint

distance, d'

: C = 20. Beyond the breakpoint distance, the frequency dependence is defined by

the formulas in Table 4-4.

(2)

For rural NLOS environments: C =20;

(3)

For urban and suburban NLOS macrocells: C = 23;

(4)

For urban and suburban NLOS microcells: C = 23;

(5)

For indoor environments: C =20;

(6)

For indoor-to-outdoor and outdoor-to-indoor environments: C is the same as in the

corresponding outdoor scenario;

(7)

For fixed NLOS feeder scenarios: in urban and suburban scenarios C =23, otherwise C =20.

Table 4-4 Summary table of the path-loss models

Scenario Path loss [dB] Shadow

fading

std [dB]

Applicability range,

antenna height default

values

LOS A = 18.7, B = 46.8, C = 20 σ

= 3 3m < d < 100m,

= h

= 1... 2.5m

NLOS

A = 36.8, B = 43.8, C = 20 and

X = 5(n

- 1) (light walls)

X = 12(n

- 1) (heavy walls)

= 4 same as A1 LOS,

is the number of walls

between the BS and the

MS (n

> 0 for NLOS)

A = 20, B = 46.4, C = 20, X = 5n

NLOS

light walls:

heavy walls:

A = 20, B = 46.4, C = 20, X = 12n

= 6

= 8

same as A1 LOS,

is the number of walls

between BS and MS

FL For any of the cases above, add the floor loss

(FL), if the BS and MS are in different floors:

FL = 17+4(n

- 1), n

> 0

is the number of floors

between the BS and the

MS (n

> 0)

A2 NLOS

intwb

PLPLPLPL











=−+= +=

inin

inoutBb

dPL

ddPLPL

5.0

))cos(1(1514

)(

= 7

3m<d

out

< 1000m,

= 3(n

-1) + 2m

= 1.5,

See

for explanation of

parameters

LOS

A = 22.7, B = 41.0 , C = 20

( )

0.5log7.2)'(log3.17

)'(log3.1745.9)(log0.40

1010

10110

cMS

hdPL +−

= 3

10m < d

< d'

BP 4)

< d

< 5km

= 10m, h

= 1.5m

NLOS

),(),,(min

1221

ddPLddPLPL =

where

)0.5/(log3

)(log105.1220)(

ljjkLOS

dnndPL

ddPL ++−+

and

84.1,0024.08.2max

, PL

LOS

is the

path loss of B1 LOS scenario and k,l ∈ {1,2},.

= 4 10m < d

< 5km,

w/2 < d

< 2km

w =20m

(street width)

=10m, h

=1.5m

When

0<d

< w/2 , the LOS PL

is applied.

B2 NLOS Same as B1. σ

= 4

WINNER II D1.1.2 V1.2

Page 45 (82)

LOS A = 13.9, B = 64.4, C =20 σ

= 3 5m < d < 100 m,

= 6 m, h

= 1.5 m

NLOS A = 37.8, B = 36.5, C =23 σ

= 4 Same as B3 LOS

B4 NLOS Same as A2, except antenna heights. 3m<d

out

< 1000m,

=10m, h

=3(n

-1)+1.5m

B5a LOS A = 23.5, B = 42.5, C = 20

= 4 30m < d < 8km

= 25m, h

= 25m

B5c LOS Same as B1 LOS, except antenna heights (h

the relay antenna height). σ

= 3 10m < d < 2000m

=10m, h

(=h

)=5m

B5f NLOS A = 23.5, B = 57.5, C =23 σ

= 8 30m < d < 1.5km

= 25m, h

= 15m

LOS

A = 23.8, B = 41.2, C =20

( )

0.5log8.3)(log2.16

)(log2.1665.11)(log0.40

1010

cMS

hdPL +−

= 4

= 6

30m < d < d

< d < 5km,

= 25m, h

= 1.5m

NLOS

( )

0.5log23)(log83.5

46.31)(log)(log55.69.44

1010

cBS

dhPL ++

= 8 50m < d < 5km,

= 25m, h

= 1.5m

LOS

A = 26, B = 39, C =20

( )

0.5log0.6)(log0.14

)(log0.1447.13)(log0.40

1010

cMS

hdPL

+− −+=

= 4

= 6

10m < d < d'

BP 4)

< d < 5km

= 25m, h

= 1.5m

NLOS

( )

0.5log23)(log83.5

46.34)(log)(log55.69.44

1010

cBS

dhPL ++

= 8 Same as C1 NLOS

C3 NLOS Same as C2 NLOS Same as C2 NLOS

C4 NLOS

MSininoutC

hddd PLPL 8.05.04.17)(

where PL

is the path-loss function of C2

LOS/NLOS scenario. (Use LOS, if BS to wall

connection is LOS, otherwise use NLOS)

= 10 Same as C2 NLOS

See

for explanation of

parameters.

=25

, h

=3n

1.5m

LOS

A =21.5, B = 44.2, C =20

( )

0.5log5.1)(log5.18

)(log5.185.10)(log0.40

1010

cMS

hdPL +−

= 4

= 6

10m < d < d

< d < 10km,

= 32m, h

= 1.5m

NLOS PL=25.1log

(d )+55.4–0.13(h

–25)log

(d /100)

–0.9(h

–1.5)+ 21.3log

/5.0)

= 8 50m < d < 5km,

= 32m, h

= 1.5m

D2a LOS Same as D1 LOS

Actual A1 NLOS scenario (Corridor-to-Room)

Optional A1 NLOS scenario (Room-to-Room through wall)

is the B1 path loss, PL

is the C2 path loss, d

out

is the distance between the outdoor

terminal and the point on the wall that is nearest to the indoor terminal, d

is the distance from

the wall to the indoor terminal,

is the angle between the outdoor path and the normal of the

wall. n

is the floor index (the ground floor has index 1) .

= 4 h'

/c , where f

is the centre frequency in Hz, c = 3.0× 10

m/s is the

propagation velocity in free space, and h'

and h'

are the effective antenna heights at the BS

and the MS, respectively. The effective antenna heights h'

and h'

are computed as follows:

= h

– 1.0 m, h'

= h

– 1.0 m, where h

and h

are the actual antenna heights, and

the effective environment height in urban environments is assumed to be equal to 1.0 m.

The distances d

and d

will be defined below in Figure 4-3.

The breakpoint distance, d

, is computed as follows: d

= 4 h

/c , where h

, h

, f

and c have the same definition as under item 4).

WINNER II D1.1.2 V1.2

Page 46 (82)

The NLOS path loss model for scenario B1 is dependent on two distances, d

and d

. These distances are

defined with respect to a rectangular street grid, as illustrated in Figure 4-3, where the MS is shown

moving along a street perpendicular to the street on which the BS is located (the LOS street). d

is the

distance from the BS to the centre of the perpendicular street, and d

is the distance of the MS along the

perpendicular street, measured from the centre of the LOS street.

Figure 4-3 Geometry for d

and d

path-loss model

4.3.1 Transitions between LOS/NLOS

The WINNER channel model allows transitions between different propagation conditions, the most

important of which are transitions between LOS and NLOS within the same WINNER scenario. In the A1

(indoor) and B1 (urban microcell) scenarios, transitions from LOS to NLOS can occur as a result of the

MS turning from the corridor or street in which the BS is located (the LOS corridor/street) into a

perpendicular corridor or street. An analysis of this specific case has indicated that such transitions can be

adequately modelled by using the A1 or B1 LOS and NLOS path loss models defined in Table 4-4. Let d

and d

denote the distances along the LOS corridor/street and the perpendicular corridor/street,

respectively, as illustrated in Figure 4-3. The A1 LOS path loss model is then considered to be applicable

for values of d

smaller than 3F

, where F

represents the radius of the first Fresnel zone (for definition of

Fresnel zones see [Sau99, sec 3.3.1] ). For values of d

greater than 3F

, the A1 NLOS path loss model

can be used. For the B1 scenario, a better fit to measured data was obtained by choosing the NLOS/LOS

transition distance equal to 10F

. It is noted that, in most cases, reasonably good results can also be

obtained by setting the transition distance equal to half the width of the LOS corridor or street, as

reflected by the path loss model for B1 NLOS in Table 4-4.

4.4 Parameter tables for generic models

Table 4-5 provides parameter values corresponding to the WINNER generic channel models. Parameter

values related to elevation angles are provided in

Table 4-6

WINNER II D1.1.2 V1.2

Page 47 (82)

Table 4-5 Table of parameters for generic models.

NOTE! With arrival and departure directions we consider downlink case, i.e. departure refers to BS and arrival refers to MS.

The path loss models for the C1 LOS and D1 LOS scenarios contain separate shadowing standard deviations for distances smaller and greater than the breakpoint distance, respectively.

* The sign of the shadow fading term is defined so that increasing values of SF correspond to increasing received power at the MS.

AoD and AoA refer to azimuth angles at the indoor and outdoor terminals, respectively. Parameter values for the B4 and C4 scenarios are identical.

∇

In case column A2/B4/C4 contains two parameter values, the left value corresponds to A2/B4 microcell and the right value to C4 macrocell.

∆

In case column A2/B4/C4 contains two parameter values, the left value corresponds to A2 Indoor-to-Outdoor and the right value to B4/C4 Outdoor-to-Indoor.

A1 A2/B4/C4

B1 B3 C1 C2 D1 D2a

Scenarios LOS NLOS NLOS LOS NLOS LOS NLOS LOS NLOS LOS NLOS LOS NLOS LOS

-7.42 -7.60 -7.39/

-6.62

∇

-7.44

-7.12 -7.53 -7.41 -7.23 -7.12 -7.39 -6.63 -7.80 -7.60 -7.4

Delay spread (DS)

log

([s])

0.27 0.19 0.36/

0.32

∇

0.25 0.12 0.12 0.13 0.49 0.33 0.63 0.32 0.57 0.48 0.2

1.64 1.73 1.76 0.40

1.19 1.22 1.05 0.78 0.90 1 0.93 0.78 0.96 0.7 AoD spread (ASD)

log

([°])

0.31 0.23 0.16 0.37 0.21 0.18 0.22 0.12 0.36 0.25 0.22 0.21 0.45 0.31

1.65 1.69 1.25 1.40 1.55 1.58 1.7 1.48 1.65 1.7 1.72 1.20 1.52 1.5 AoA spread (ASA)

log

([° ])

0.26 0.14 0.42 0.20 0.20 0.23 0.1 0.20 0.30 0.19 0.14 0.18 0.27 0.2

Shadow fading (SF)

[dB]

3 4 7 3 4 3 4 4/6

8 4/6

8 4

7 N/A N/A 9 N/A 2 N/A 9 N/A 7 N/A 7 N/A 7

K-factor (K) [dB]

6 N/A N/A 6 N/A 3 N/A 7 N/A 3 N/A 6 N/A 6

ASD vs DS 0.7 -0.1 0.4 0.5 0.2 -0.3 -0.1 0.2 0.3 0.4

0.4

-0.1 -0.4 -0.1

ASA vs DS 0.8 0.3 0.4 0.8 0.4 -0.4 0 0.8 0.7 0.8

0.6

0.2 0.1 0.2

ASA vs SF -0.5 -0.4 0.2 -0.5 -0.4 -0.2 0.2 -0.5 -0.3 -0.5

-0.3

-0.2 0.1 -0.2

ASD vs SF -0.5 0 0 -0.5 0 0.3 -0.3 -0.5 -0.4 -0.5

-0.6

0.2 0.6 0.2

DS vs SF -0.6 -0.5 -0.5 -0.4 -0.7 -0.1 -0.2 -0.6 -0.4 -0.4 -0.4 -0.5 -0.5 -0.5

ASD vs ASA 0.6 -0.3 0 0.4 0.1 0.3 -0.3 0.1 0.3 0.3 0.4 -0.3 -0.2 -0.3

ASD vs

-0.6 N/A N/A -0.3 N/A 0.2 N/A 0.2 N/A 0.1 N/A 0 N/A 0

ASA vs

-0.6 N/A N/A -0.3 N/A -0.1 N/A -0.2 N/A -0.2 N/A 0.1 N/A 0.1

DS vs

-0.6 N/A N/A -0.7 N/A -0.3 N/A -0.2 N/A -0.4 N/A 0 N/A 0

Cross-Correlations *

SF vs

0.4 N/A N/A 0.5 N/A 0.6 N/A 0 N/A 0.3 N/A 0 N/A 0

Delay distribution Exp Exp Exp Exp Uniform

≤800ns Exp Exp Exp Exp Exp Exp Exp Exp Exp

AoD and AoA distribution Wrapped Gaussian

Delay scaling parameter r

3 2.4 2.2 3.2  1.9 1.6 2.4 1.5 2.5

2.3 3.8 1.7 3.8

11 10 9 9 8 9 6 8 4 8 7 12 7 12

XPR

[dB]

4 4 11 3 3 4 3 4 3 4 3 8 4 8

Number of clusters

12 16 12 8 16 10 15 15 14 8 20 11 10 8

Number of rays per cluster 20 20 20 20 20 20 20 20 20 20 20 20 20 20

Cluster ASD 5 5 8 3 10 5 6 5 2 6 2 2 2 2

Cluster ASA 5 5 5 18 22 5 13 5 10 12 15 3 3 3

Per cluster shadowing std ζ [dB] 6 3 4 3 3 3 3 3 3 3 3 3 3 3

Correlation

distance [m] DS 7 4 21/10

∆

9 8 3 1 6 40 40 40 64 36 64

ASD 6 5 15/11

∆

13 10 1 0.5 15 30 15 50 25 30 25

ASA 2 3 35/17

∆

12 9 2 0.5 20 30 15 50 40 40 40

SF 6 4 14/7

∆

14 12 3 3 40 50 45 50 40 120 40

6 N/A N/A 10 N/A 1 N/A 10 N/A 12 N/A 40 N/A 40

WINNER II D1.1.2 V1.2

Page 48 (82)

Table 4-6 Table of elevation-related parameters for generic models.

A1 A2/B4

/C4

Scenarios LOS NLOS NLOS

0.88 1.06 0.88

Elevation AoD

spread (ESD)

0.31 0.21 0.34

0.94 1.10 1.01

Elevation AoA

spread (ESA)

0.26 0.17 0.43

ESD vs DS 0.5 -0.6 N/A

ESA vs DS 0.7 -0.1 0.2

ESA vs SF -0.1 0.3 0.2

ESD vs SF -0.4 0.1 N/A

Cross-

Correlations

ESD

vs ESA 0.4 0.5 N/A

Elevation AoD and AoA

distribution Gaussian

Cluster ESD 3 3 3

Cluster ESA 3 3 3

ESD and ESA refer to elevation angle spreads at the indoor and outdoor terminals, respectively.

System level simulations require estimates of the probability of line-of-sight. For scenarios A2, B2,

B4, C2 and C3, the LOS probability is approximated as being zero. For the remaining scenarios, LOS

probability models are provided in Table 4-7. These models are based on relatively limited data sets

and/or specific assumptions and approximations regarding the location of obstacles in the direct path,

and should therefore not be considered exact.

If the terminal locations are known with respect to a street grid or floor plan, which can be the case in

grid-based scenarios such as A1 (indoor) and B1 (urban microcell), the WINNER channel model

provides the option to determine the existence of NLOS/LOS propagation conditions

deterministically.

Table 4-7 Line of sight probabilities

Scenario

LOS probability as a function of distance d [m] Note

( )











>−−−

≤

=5.2,)(log61.024.119.01

5.2,1

LOS

)36/exp()36/exp(1)1,/18min( dddP LOS −+−−⋅=





















−

−

≤

=10,

exp

10,1

LOS

For big factory halls,

airport and train stations.











−= 200

exp d

LOS

)63/exp()63/exp(1)1,/18min( dddP

LOS











−= 1000

exp d

LOS

4.4.1 Reference output values

Table 4-8 and Table 4-9 provide median values of the large-scale parameters produced by the

WINNER channel model for various scenarios. The values in Table 4-9 were computed under the

WINNER II D1.1.2 V1.2

Page 49 (82)

assumption that the maximum cell radii for microcells and macrocells are 200 and 500 m,

respectively, and that the distribution of user terminals over the cell area is uniform. The median

values are dependent on cell radii, thus the tabulated values are not universal in bad urban scenarios.

Table 4-8: Median output values of large-scale parameters.

Scenario DS (ns) AS at BS (º)

AS at MS (º)

ES at BS (º)

ES at MS (º)

LOS

40 44 45 8 9

A1 NLOS

25 53 49 11 13

A2/B4

/C4 NLOS

49/240

∇

58 18 10 10

LOS

36 3 25

B1 NLOS

76 15 35

LOS

27 17 38 21.2

B3 NLOS

39 12 50 22.3

LOS

59 6 30

C1 NLOS

75 8 45

LOS

41 10 50

C2 NLOS

234 8 53

LOS

16 6 16

D1 NLOS

37 9 33

D2 LOS

39 5 32

AS at BS denotes indoor azimuth spread and As at MS denotes outdoor azimuth spread

∇

In case column A2/B4/C4 contains two parameter values, the left value corresponds to A2/B4

microcell and the right value to C4 macrocell.

Table 4-9: Median output values of large scale parameters for bad urban scenarios.

Scenario DS (µs) AS at BS

(º) AS at MS

(º) Power of

the 1

cluster

(dB)

Power of

the 2

cluster

(dB)

Delay of

the 1

cluster

(µ s)

Delay of

the 2

cluster

(µ s)

B2 0.48 33 51 -5.7 -7.7 1.1 1.6

C3 0.63 17 55 -9.7 -13.0 3.1 4.8

4.5 CDL Models

Although the clustered delay line (CDL) model is based on similar principles as the conventional

tapped delay line model, it is different in the sense that the fading process for each tap is modelled in

terms of a sum of sinusoids rather than by a single tap coefficient. The CDL model describes the

propagation channel as being composed of a number of separate clusters with different delays. Each

cluster, in turn, is composed ofa number of multipath components (rays) that have the same delay

values but differ in angle-of-departure and angle-of-arrival. The angular spread within each cluster can

be different at the BS and the MS. The offset angles represent the Laplacian PAS of each cluster. The

average power, mean AoA, mean AoD of clusters, angle-spread at BS and angle-spread at MS of each

cluster in the CDL represent expected output of the stochastic model with parameters listed in Table

4-8. Exceptions are the fixed feeder link models in scenario B5, for which no stochastic models have

been defined.

Parameter tables for the CDL models are given in Section 6 of this document.

WINNER II D1.1.2 V1.2

Page 50 (82)

5. Channel Model Usage

The purpose of this chapter is to discuss issues concerning usage of the WINNER channel model for

simulations.

5.1 System level description

5.1.1 Coordinate system

System layout in the Cartesian coordinates is for example the following:

Figure 5-1: System layout of multiple base stations and mobile stations.

All the BS and MS have (x,y) coordinates. MS and cells (sectors) have also array broad side

orientation, where north (up) is the zero angle. Positive direction of the angles is the clockwise

direction.

Table 5-1: Transceiver coordinates and orientations.

Tranceiver Co-ordinates Orientation [°]

BS1 cell1 (x

bs1

) Ω

cell2 (x

bs1

) Ω

cell3 (x

bs1

) Ω

BS2 cell4 (x

bs2

) Ω

cell5 (x

bs2

) Ω

cell6 (x

bs2

) Ω

MS1 (x

ms1

) Ω

ms1

MS2 (x

ms2

) Ω

ms2

MS3 (x

ms3

) Ω

ms3

Both the distance and line of sight (LOS) direction information of the radio links are calculated for the

input of the model. Distance between the BS

and MS

WINNER II D1.1.2 V1.2

Page 51 (82)

)()(

kikiki

MSBSMSBSMSBS

yyxxd −+−=

. (5.1)

The LOS direction from BS

to MS

with respect to BS antenna array broad side is (see Figure 5-2)











<Ω−°−













−

≥Ω−°+













−

iki

BSMSBS

BSMS

BSMSBS

BSMS

MSBS

when ,90arctan

(5.2)

The angles and orientations are depicted in the figure below.

MSBS ,

BSMS ,

Figure 5-2: BS and MS antenna array orientations.

Pairing matrix

is in the example case of Figure 5-2 a 6

3 matrix with values

n,m

∈ {0,1}. Value 0

stands for link celln to MSm is not modelled and value 1 for link is modelled.













3,62,61,6

3,22,21,2

3,12,11,1

mscmscmsc

χχχ

χχχ χχχ

MMM

(5.3)

The pairing matrix can be applied to select which radio links will be generated and which will not.

5.1.2 Multi-cell simulations

5.1.2.1 Single user (Handover)

A handover situation is characterized by a MS moving from the coverage are of one BS to the

coverage area of another BS. Figure 5-3 illustrates this setup.

WINNER II D1.1.2 V1.2

Page 52 (82)

Figure 5-3: Handover scenario.

There are two base-stations or cells denoted c1 and c2, and one mobile station. Thus, while there is

only one mobile station in the scenario, each location of the mobile on its path is assigned a unique

label ms1 to msM. This is equivalent to a scenario with multiple mobile stations at different positions

ms1 to msM. Path-loss will be determined according to the geometry and large-scale parameters

correlate properly. The resulting procedure is as follows:

Set base station c1 and c2 locations and array orientations according to geometry.

Set MS locations ms1 to msM and array orientations along the route. Choose the distance

between adjacent locations according to desired accuracy.

Set all the entries of the pairing matrix to 1.

Generate all the radio links at once to obtain correct correlation properties. It is possible to

generate more channel realizations, i.e. time samples, for each channel segment afterwards.

This can be done by applying the same values of small scale parameters and restoring final

phases of the rays.

Simulate channel segments consecutively to emulate motion along the route.

It is also possible to model even more accurate time evolution between locations as described in

section 3.4. The clusters of current channel segment (location) are replaced by clusters of the next

channel segment one by one.

5.1.2.2 Multi-user

The handover situation from the previous section was an example of single-user multi-cell setup.

Other cases of such a setup are for example found in the context of multi-BS protocols, where a MS

receives data from multiple BS simultaneously.

The extension to multiple users (and one or more base stations) is straightforward. Because location

and mobile station index are treated equivalently, it follows that all locations of all mobiles have to be

defined. Consider the drive-by situation in Figure 5-4.

WINNER II D1.1.2 V1.2

Page 53 (82)

Figure 5-4: Drive-by scenario (with multiple mobile stations).

Here, M locations of mobile station 1, and N locations of mobile station 2 are defined yielding a total

of M+N points or labels. The resulting procedure is as follows.

Set BS c1 and c2 locations and array orientations according to layout.

Set MS locations ms11 to ms2N and array orientations according to layout.

Set the links to be modelled to 1 in the pairing matrix.

Generate all the radio links at once to obtain correct correlation properties. It is possible to

generate more channel realizations, i.e. time samples, for each channel segment afterwards.

This can be done by applying the same values of small scale parameters and restoring final

phases of the rays.

Simulate channel segments in parallel or consecutively according to the desired motion of the

mobiles.

5.1.3 Multihop and relaying

Typically, the links between the MSs and the links between the BSs are not of interest. Cellular

systems are traditionally networks where all traffic goes through one or more BS. The BS themselves

again only talk to a BS hub and not between them.

Multihop and relaying networks break with this limitation. In multihop networks, the data can take a

route over one or more successive MS. Relaying networks, on the other hand, employ another level of

network stations, the relays, which depending on the specific network, might offer more or less

functionality to distribute traffic intelligently. The WINNER channel model can be used to obtain the

channels for multihop or relaying scenarios, as described below.

Figure 5-5: Multihop and relaying scenarios.

WINNER II D1.1.2 V1.2

Page 54 (82)

In the example figure above the signal from MS1 to BS3 is transmitted via MS3 and BS2 act as a

repeater for BS1. These scenarios can be generated by introducing a BS-MS pair into position of a

single BS serving as a relay or into position of a single MS serving as a multihop repeater. In these

cases one can apply path-loss models of feeder scenarios described in section 3.2.4. The resulting

procedure is as follows.

Set base station BS1 to BS3 locations and array orientations according to layout.

Set mobile locations MS1 to MS3 and array orientations according to layout.

Add extra base station BS4 to position of MS3 and extra mobile MS4 to position of BS2 with

same array orientations and array characteristics as MS3 and BS2 respectively.

Set the BS

MS pairing matrix to













0010

0100

0001

1000

Generate all the radio links at once.

Simulate the channel segments in parallel.

5.1.4 Interference

Interference modelling is an application subset of channel models that deserves additional

consideration. Basically, communication links that contain interfering signals are to be treated just as

any other link. However, in many communication systems these interfering signals are not treated and

processed in the same way as the desired signals and thus modelling the interfering links with full

accuracy is inefficient.

A simplification of the channel modelling for the interference link is often possible but closely linked

with the communication architecture. This makes it difficult for a generalized treatment in the context

of channel modelling. In the following we will thus constrain ourselves to giving some possible ideas

of how this can be realised. Note that these are all combined signal and channel models. The actual

implementation will have to be based on the computational gain from computational simplification

versus the additional programming overhead.

AWGN interference

The simplest form of interference is modelled by additive white Gaussian noise. This is sufficient for

basic C/I (carrier to interference ratio) evaluations when coupled with a path loss and shadowing

model. It might be extended with e.g. on-off keying (to simulate the non-stationary behaviour of

actual transmit signals) or other techniques that are simple to implement.

Filtered noise

The possible wideband behaviour of an interfering signal is not reflected in the AWGN model above.

An implementation using a complex SCM or WIM channel, however, might be unnecessarily

complex as well because the high number of degrees of freedom does not become visible in the noise-

like signal anyway. Thus we propose something along the lines of a simple, sample-spaced FIR filter

with Rayleigh-fading coefficients.

Pre-recorded interference

A large part of the time-consuming process of generating the interfering signal is the modulation and

filtering of the signal, which has to be done at chip frequency. Even if the interfering signal is detected

and removed in the communication receiver (e.g., multi-user detection techniques) and thus rendering

a PN generator too simple, a method of pre-computing and replaying the signal might be viable. The

repeating content of the signal using this technique is typically not an issue as the content of the

interferer is discarded anyway.

Exact interference by multi-cell modelling

Interference situations are quite similar to multi-cell or multi-BS situations, except that in this case the

other BSs transmit a non-desired signal which creates interference.

WINNER II D1.1.2 V1.2

Page 55 (82)

5.2 Space-time concept in simulations

5.2.1 Time sampling and interpolation

Channel sampling frequency has to be finally equal to the simulation system sampling frequency. To

have feasible computational complexity it is not possible to generate channel realisations on the

sampling frequency of the system to be simulated. The channel realisations have to be generated on

some lower sampling frequency and then interpolated to the desired frequency. A practical solution is

e.g. to generate channel samples with sample density (over-sampling factor) two, interpolate them

accurately to sample density 64 and to apply zero order hold interpolation to the system sampling

frequency. Channel impulse responses can be generated during the simulation or stored on a file

before the simulation on low sample density. Interpolation can be done during the system simulation.

To be able to obtain the deep fades in the NLOS scenarios, we suggest using 128 samples per

wavelength (parameter '

SampleDensity

' = 64). When obtaining channel parameters quasi-stationarity

has been assumed within intervals of 10-50 wavelengths. Therefore we propose to set the drop

duration corresponding to the movement of up to 50 wavelengths.

5.3 Radio-environment settings

5.3.1 Scenario transitions

In the channel model implementation it is not possible to simulate links from different scenarios

within one drop. This assumes that all propagation scenarios are the same for all simulated links. The

change of the scenario in time can be simulated by changing the scenario in the consecutive drop.

Similarly, to obtain different scenarios within radio-network in the same drop, multiple drops could be

simulated – one for each scenario. Afterwards, merging should be performed.

5.3.2 LOS\NLOS transitions

Mix of LOS and NLOS channel realizations can be obtained by first calculating a set of LOS drops

and after it a set of NLOS drops. This can be done by setting the parameter '

PropagCondition

' to 'LOS'

and later to 'NLOS'.

5.4 Bandwidth/Frequency dependence

5.4.1 Frequency sampling

The WINNER system is based on the OFDM access scheme. For simulations of the system, channel

realizations in time-frequency domain are needed. The output of WIM is the channel in time-delay

domain. The time-frequency channel at any frequency can be obtained by applying next two steps:

•

define a vector of frequencies where the channel should be calculated

•

by use of the Fourier transform calculate the channel at defined frequencies

5.4.2 Bandwidth down scaling

The channel models are delivered for 100 MHz RF band-width. Some simulations may need smaller

bandwidths. Therefore we describe below shortly, how the down-scaling should be performed. In

doing so we assume that the channel parameters remain constant in down-scaling as indicated in our

analyses.

5.4.2.1 Down-scaling in delay domain

There is a need for down-scaling, if the minimum delay sample spacing in the Channel Impulse

Response (CIR) is longer than 5 ns in the simulation. Five nanoseconds is the default minimum

spacing for the channel model samples (taps) and defines thus the delay grid for the CIR taps. For all

smaller spacings the model shall be down-scaled. The most precise way would be filtering by e.g. a

FIR filter. This would, however, create new taps in the CIR and this is not desirable. The preferred

method in the delay domain is the following:

Move the original samples to the nearest location in the down-sampled delay grid.

In some cases there are two such locations. Then the tap should be placed in the one that has

the smaller delay.

Sometimes two taps will be located in the same delay position. Then they should be summed

as complex numbers.

WINNER II D1.1.2 V1.2

Page 56 (82)

Above it has been assumed that the CIR samples are taken for each MIMO channel separately and that

the angle information has been vanished in this process. This is the case, when using the model e.g.

with the WIM implementation [WIN2WIM].

5.4.2.2 Down-scaling in frequency domain

If desired, the down-scaling can also be performed in the frequency domain. Then the starting point

will be the original CIR specified in the delay domain. This CIR is transformed in the frequency

domain for each simulation block. Then the transformed CIR can be filtered as desired, e.g. by

removing the extra frequency samples, and used in the simulation as normally.

The maximum frequency sampling interval is determined by the coherence bandwidth

, (5.4)

where

is the rms delay spread and C is a scaling constant related to fading distribution.

5.4.3 FDD modeling

In next steps we explain how to obtain both uplink and downlink channel of an FDD system with

bandwidths of 100 MHz. The center carrier frequencies are f

and f

+ ∆f

•

Define BS and MS positions, calculate the channel for one link, e.g. BS to MS at certain

carrier frequency

•

Save the small scale parameters

•

Exchange the positions of the BS and MS

•

Calculate the other link, in this example the MS to BS by:

Using saved small scale parameters

Randomizing the and initial phases of rays

Changing the carrier frequency to

∆+

5.5 Comparison tables of WINNER channel model versions

This section shows the main differences between the different versions of WINNER channel models

(Phase I (D5.4), Phase II Interim (D1.1.1), and Phase II Final (D1.1.2) models). Note! This section is

aimed as comparison of the different versions, not as the primary source of channel model parameters.

Table below shows which scenarios are available in the different versions. Note that all the scenarios

of Phase I have been updated in Phase II models.

Table 5-2:Availability of Generic and CDL models

Phase I Phase II

Scenario D5.4 D1.1.1 D1.1.2

Code Definition Generic

model CDL Generic

model CDL

A1 indoor office yes yes yes yes yes yes

A2 indoor-to-outdoor yes yes yes yes

B1 urban micro-cell yes yes yes yes yes yes

B2 bad urban micro-cell yes yes yes yes

B3 large indoor hall yes yes yes yes yes yes

WINNER II D1.1.2 V1.2

Page 57 (82)

B4 outdoor-to-indoor yes yes yes yes

B5a stationary feeder yes yes yes

B5b stationary feeder yes yes yes

B5c stationary feeder yes yes yes

B5d stationary feeder yes

B5f stationary feeder yes yes

C1 suburban macro-cell yes yes yes yes

C2 urban macro-cell yes yes yes yes

C3 bad urban macro-cell yes yes

C4 urban macro outdoor-to-indoor yes

C5 LOS feeder yes yes

D1 rural macro-cell yes yes yes yes yes yes

D2a moving networks yes yes yes yes

D2b moving networks yes yes

The features of Phase I model and Phase II model are compared in table below.

Table 5-3: Comparison of Features.

Phase I Phase II

D5.4 D1.1.1 D1.1.2

Feature generic

model CDL generic

model CDL

Number of main scenarios (see table

above) 7 7 13 13 14 14

Number of scenarios including sub-

scenarios (a,b,c,…) 10 10 16 16 18 18

Number of scenarios including sub-

scenarios and LOS/NLOS versions 15 15 21 21 24 24

Indoor-to-outdoor models yes yes yes yes

Outdoor-to-indoor models yes yes yes yes

Bad urban models yes yes yes yes

Moving networks models yes yes yes yes

Support of coordinate system yes yes yes

Support of multi-cell and multi-user

simulations yes yes yes

Support of multihop and relaying

simulations yes yes* yes yes* yes yes*

Correlation of large-scale parameters yes yes yes

Support of interference simulations yes yes yes

Time evolution yes yes

Reduced variability clustered delay

line (CDL) model for calibration, yes yes yes

WINNER II D1.1.2 V1.2

Page 58 (82)

comparisons, and fast simulations

CDL analyzed from measured PDP yes

CDL based on expectation values of

generic model yes yes

Intra-cluster delay spread yes yes yes yes

Far cluster option yes yes yes yes

Modelling of elevation yes yes

LOS as random variable yes yes

Moving scatterers yes yes

* With slight modification: AoD and AoA should be adjusted according to the network layout.

Table below shows the difference in parameter values.

Table 5-4: Comparison of parameters of Phase I and Phase II models

Phase I Phase II

D5.4 D1.1.1 D1.1.2

Parameter Unit Generic

model CDL Generic and

CDL model Generic and

CDL model

Frequency range GHz 5 5 2 – 6 2 – 6

Bandwidth MHz 100 100 100 100

Number of sub-paths per

cluster 10 10 20 20

A1 LOS delay spread ns 39.8 12.9 38.0 40

A1 NLOS delay spread ns 25.1 24.5 25.1 25

B1 LOS delay spread ns 36 19.5 41.7 36

B1 NLOS delay spread ns 76 94.7 81.3 76

B3 LOS delay spread ns 26.0 18.6 28.2

B3 NLOS delay spread ns 45.0 30.0 39.8

C1 LOS delay spread ns 1.6 29.6 58.9 59

C1 NLOS delay spread ns 55.0 61.5 75.9 75

C2 LOS delay spread ns 41

C2 NLOS delay spread ns 234.4 313.0 182.0 234

D1 LOS delay spread ns 15.8 20.4 15.8 16

D1 NLOS delay spread ns 25.1 27.8 25.1 37

D2 LOS delay spread ns 39

A1 LOS AoD spread º 5.5 5.1 43.7 44

A1 NLOS AoD spread º 20.0 23.2 53.7 53

B1 LOS AoD spread º 3 5.6 2.5 3

B1 NLOS AoD spread º 15 12.4 17.4 15

WINNER II D1.1.2 V1.2

Page 59 (82)

B3 LOS AoD spread º 26.4 3.7 30.2

B3 NLOS AoD spread º 38.0 3.0 39.8

C1 LOS AoD spread º 13.8 14.2 13.8 6

C1 NLOS AoD spread º 3.4 5.0 3.4 8

C2 LOS AoD spread º 10

C2 NLOS AoD spread º 8.5 8.0 8.5 8

D1 LOS AoD spread º 16.6 21.5 16.6 6

D1 NLOS AoD spread º 9.1 22.4 9.1 9

D2 LOS AoD spread º 5

A1 LOS AoA spread º 33.1 32.5 44.7 45

A1 NLOS AoA spread º 37.2 39.1 46.8 49

B1 LOS AoA spread º 25 37.1 25.1 25

B1 NLOS AoA spread º 35 36.4 39.8 35

B3 LOS AoA spread º 13.1 18.1 14.1

B3 NLOS AoA spread º 9.5 18.7 11.7

C1 LOS AoA spread º 40.7 45.8 40.7 30

C1 NLOS AoA spread º 46.8 53.0 46.8 45

C2 LOS AoA spread º 50

C2 NLOS AoA spread º 52.5 53.0 52.5 53

D1 LOS AoA spread º 33.1 24.0 33.1 16

D1 NLOS AoA spread º 33.1 17.9 33.1 33

D2 LOS AoA spread º 30

A1 LOS Shadow fading dB 3.1 3 3

A1 NLOS Shadow fading dB 3.5 6 6

B1 LOS Shadow fading dB 2.3 3 3

B1 NLOS Shadow fading dB 3.1 4 4

B3 LOS Shadow fading dB 1.4 2

B3 NLOS Shadow fading dB 2.1 2

C1 LOS Shadow fading dB 4.0 … 6.0 4 … 6 4 … 6

C1 NLOS Shadow fading dB 8.0 8 8

C2 LOS Shadow fading dB 8.0 8 4

C2 NLOS Shadow fading dB 8

D1 LOS Shadow fading dB 3.5 … 6.0 4 … 6 4 … 6

D1 NLOS Shadow fading dB 8.0 8 8

D2 LOS Shadow fading dB 2.5

WINNER II D1.1.2 V1.2

Page 60 (82)

5.6 Approximation of Channel Models

WINNER Generic model is aimed to be applicable for many different simulations and to cover high

number of scenarios with several combinations of large-scale and small-scale parameters. Generic

model is the most accurate model and is recommended to be used whenever possible. However, in

some simulations, channel model can be simplified (approximated) to reduce the simulation

complexity. It has to be done very carefully. When approximating the model, reality is reduced, and

the impact of the approximation has to be understood. The impact of the approximation depends on,

e.g., the transceiver system, algorithms, modulation, coding, multi-antenna technology, and required

accuracy of the simulation results. If someone is uncertain whether approximation affects on the

simulation results or not, it is better not to approximate. Therefore, the following approximation steps

can only be done by the simulation experts.

Firstly, we can approximate the model by assuming all the large scale parameters fixed to median

values. Furthermore, we can reduce the model by fixing the delays, but keep angles as random. The

third approximation can be done by freezing all propagation parameters to obtain so called Clustered

Delay Line (CDL) model. If, from a good reason, correlation model is desired, we can calculate

correlation matrices from the CDL model by fixing the antenna structure. Kronecker approach can

simplify the model even further, and finally, independent channels make the model very simple, but at

the same time very inaccurate. The approximation steps are shown below.

WINNER II Generic Model (D1.1.2)

Fixed large scale parameters

Constant delays, random angles ("CDL with random angles")

WINNER II CDL Model (D1.1.2)

Tapped Delay Line model (delays are taken from CDL) with MIMO Correlation Matrix

Tapped Delay Line model with TX and RX Correlation Matrix, MIMO correlation is

obtained via Kronecker product.

Tapped Delay Line model, zero correlation between MIMO channels.

WINNER II D1.1.2 V1.1

Page 61 (82)

6. Parameter Tables for CDL Models

In the CDL model each cluster is composed of 20 rays with fixed offset angles and identical power. In the

case of cluster where a ray of dominant power exists, the cluster has 20+1 rays. This dominant ray has a

zero angle offset. The departure and arrival rays are coupled randomly. The CDL table of all scenarios of

interest are give below, where the cluster power and the power of each ray are tabulated. The CDL

models offer well-defined radio channels with fixed parameters to obtain comparable simulation results

with relatively non-complicated channel models.

Delay spread and azimuth spreads medians of the CDL models are equal to median values given in Table

4-8. Intra cluster delay spread is defined in Table 4-2.

6.1 A1 – Indoor small office

The CDL parameters of LOS and NLOS condition are given below. In the LOS model Ricean K-factor is

4.7 dB.

Table 6-1 Scenario A1: LOS Clustered delay line model, indoor environment.

Cluster # Delay [ns] Power [dB] AoD [º] AoA [º] Ray power [dB]

1 0 5 10 0 -15.1

-16.9

0 0 -0.23*

-22.9**

2 10 -15.8 -107 -110 -28.8

3 25 -13.5 -100 102 -26.5

4 50 55 60 -15.1

-17.3

-19.1

131 -134 -25.1

5 65 -19.2 118 121 -32.2

6 75 -23.5 131 -134 -36.5

7 75 -18.3 116 -118 -31.3

8 115 -23.3 131 -134 -36.4

9 115 -29.1 146 149 -42.2

10 145 -14.2 102 105 -27.2

11 195 -21.6 -126 129 -34.6

12 350 -23.4 131 -134 -36.4

Cluster ASD = 5º

Cluster ASA = 5º

XPR = 11 dB

Power of dominant ray,

Power of each other ray

Figure 6-1: PDP and frequency correlation (FCF) of CDL model.

Table 6-2 Scenario A1: NLOS Clustered delay line model, indoor environment.

Cluster #

Delay [ns] Power [dB] AoD [º]

AoA [º] Ray power [dB]

1 0 -2.2 45 41 -15.2

2 5 -6.6 77 -70 -19.7

3 5 -2.1 43 39 -15.1

4 5 -5.8 72 66 -18.8

5 15 -3.3 54 -49 -16.3

6 15 -4.7 -65 59 -17.7

Cluster ASD = 5º

Cluster ASA = 5º

XPR = 10 dB

WINNER II D1.1.2 V1.1

Page 62 (82)

7 15 -4.1 -60 -55 -17.1

8 20 -8.2 85 -78 -21.2

9 20 25 30 -3.0 -5.2 -7.0 0 0 -13.0

10 35 40 45 -4.6 -6.8 -8.6 -104 95 -14.6

11 80 -10.0 95 86 -23.0

12 85 -12.1 -104 95 -25.1

13 110 -12.4 -105 -96 -25.4

14 115 -11.8 103 -94 -24.8

15 150 -20.4 -135 123 -33.4

16 175 -16.6 -122 -111 -29.6

Figure 6-2: PDP and frequency correlation (FCF) of CDL model.

6.2 A2/B4 – Indoor to outdoor / outdoor to indoor

Table 6-3 Scenario A2/B4: NLOS Clustered delay line model, indoor to outdoor environment.

Cluster #

Delay [ns] Power [dB]

*AoD

[º]

*AoA [º]

Ray power [dB]

1 0 5 10 -3.0 -5.2 -7.0 0 0 -13.0

2 0 -8.7 102 32 -21.7

3 5 -3.7 -66 -21 -16.7

4 10 -11.9 -119 37 -24.9

5 35 -16.2 139 -43 -29.2

6 35 -6.9 91 28 -19.9

7 65 70 75 -3.4 -5.6 -7.3 157 -49 -13.4

8 120 -10.3 -111 -34 -23.3

9 125 -20.7 157 -49 -33.7

10 195 -16.0 138 43 -29.1

11 250 -21.0 158 49 -34.0

12 305 -22.9 165 51 -35.9

** Cluster ASD = 8º

** Cluster ASA = 5º

XPR = 9 dB

* AoD refer to angles of the indoor terminal and AoA refer to outdoor terminal

** Cluster ASD refer to indoor terminal and Cluster ASA refer to outdoor terminal

Figure 6-3: PDP and frequency correlation (FCF) of CDL model.

WINNER II D1.1.2 V1.1

Page 63 (82)

6.3 B1 – Urban micro-cell

The parameters of the CDL model have been extracted from measurements with chip frequency of 60

MHz at frequency range of 5.3 GHz. In the LOS model Ricean K-factor is 3.3 dB.

Table 6-4 Scenario B1: LOS Clustered delay line model.

Cluster # Delay [ns] Power [dB] AoD [º] AoA [º] Ray power [dB]

1 0 0.0 0 0 -0.31

-24.7

2 30 35 40 -10.5

-12.7

-14.5

5 45 -20.5

3 55 -14.8 8 63 -27.8

4 60 65 70 -13.6

-15.8

-17.6

8 -69 -23.6

5 105 -13.9 7 61 -26.9

6 115 -17.8 8 -69 -30.8

7 250 -19.6 -9 -73 -32.6

8 460 -31.4 11 92 -44.4

Cluster ASD = 3º

Cluster ASA = 18º

XPR = 9 dB

Power of dominant ray,

Power of each other ray

Figure 6-4: PDP and frequency correlation (FCF) of CDL model.

Table 6-5 Scenario B1: NLOS Clustered delay line model.

Cluster #

Delay [ns] Power [dB]

AoD

[º]

AoA [º] Ray power [dB]

1 0 -1.0 8 -20 -14.0

2 90 95 100 -3.0 -5.2 -7.0 0 0 -13.0

3 100 105 110 -3.9 -6.1 -7.9 -24 57 -13.9

4 115 -8.1 -24 -55 -21.1

5 230 -8.6 -24 57 -21.6

6 240 -11.7 29 67 -24.7

7 245 -12.0 29 -68 -25.0

8 285 -12.9 30 70 -25.9

9 390 -19.6 -37 -86 -32.6

10 430 -23.9 41 -95 -36.9

11 460 -22.1 -39 -92 -35.1

12 505 -25.6 -42 -99 -38.6

13 515 -23.3 -40 94 -36.4

14 595 -32.2 47 111 -45.2

15 600 -31.7 47 110 -44.7

16 615 -29.9 46 -107 -42.9

Cluster ASD =10º

Cluster ASA = 22º

XPR = 8 dB

WINNER II D1.1.2 V1.1

Page 64 (82)

Figure 6-5: PDP and frequency correlation (FCF) of CDL model.

6.4 B2 – Bad Urban micro-cell

Table 6-6 Scenario B2: NLOS Clustered delay line model, bad urban, microcell

Cluster

# Delay [ns] Power [dB] AoD

[º] AoA [º] Ray power [dB]

1 0 5 10 -3.0 -5.2 -7.0 0 0 -13.0

2 35 -5.4 20 -46 -18.4

3 135 140 145 -5.0 -7.2 -9.0 40 -92 -15.0

4 190 -8.2 25 57 -21.2

5 350 -21.8 40 -92 -34.8

6 425 -25.5 -44 -100 -38.5

7 430 -28.7 -46 -106 -41.7

8 450 -20.8 39 90 -33.8

9 470 -30.7 -48 -110 -43.7

10 570 -34.9 -51 -117 -47.9

11 605 -34.5 -51 -116 -47.5

12 625 -31.5 -48 -111 -44.5

13 625 -35.3 -51 -118 -48.3

14 630 -37.5 53 121 -50.5

Cluster ASD = 10º

Cluster ASA = 22º

15 1600 -5.7 -110 15 -18.7

16 2800 -7.7 75 -25 -20.7 3º

3º

XPR = 8 dB

Figure 6-6: PDP and frequency correlation (FCF) of CDL model.

6.5 B3 – Indoor hotspot

The CDL parameters of LOS and NLOS condition are given below. In the LOS model Ricean K-factor is

2 dB.

WINNER II D1.1.2 V1.1

Page 65 (82)

Table 6-7 Scenario B3: LOS Clustered delay line model.

Cluster # Delay [ns] Power [dB] AoD [º] AoA [º] Ray power [dB]

1 0 0.0 0 0 -0.32

-24.5

2 0 5 10 -9.6 -11.8

-13.6

-23 -53 -19.6

3 15 -14.5 -34 -79 -27.6

4 25 -12.8 -32 -74 -25.8

5 40 -13.7 33 76 -26.8

6 40 45 50 -14.1

-16.4

-18.1

-35 80 -24.1

7 90 -12.6 32 -73 -25.6

8 130 -15.2 -35 80 -28.2

9 185 -23.3 -43 -100 -36.4

10 280 -27.7 47 -108 -40.7

Cluster ASD = 5º

Cluster ASA = 5º

XPR = 9 dB

Power of dominant ray,

Power of each other ray

Figure 6-7: PDP and frequency correlation (FCF) of CDL model.

Table 6-8 Scenario B3: NLOS Clustered delay line model.

Cluster

Delay [ns] Power [dB] AoD

[º]

AoA [º] Ray power [dB]

1 0 -6.6 -16 -73 -19.6

2 5 10 15 -3.0 -5.2 -7.0 0 0 -13.0

3 5 -11.0 -21 -94 -24.0

4 10 15 20 -4.3 -6.5 -8.2 -10 -46 -14.3

5 20 -7.1 17 75 -20.1

6 20 -2.7 -10 -46 -15.7

7 30 -4.3 -13 -59 -17.3

8 60 -14.1 -24 107 -27.1

9 60 -6.2 -16 71 -19.2

10 65 -9.1 19 86 -22.1

11 75 -5.5 -15 67 -18.5

12 110 -11.1 -21 95 -24.1

13 190 -11.8 22 98 -24.8

14 290 -17.0 -26 117 -30.1

15 405 -24.9 -32 142 -37.9

Cluster ASD = 6º

Cluster ASD = 13º

XPR = 5 dB

WINNER II D1.1.2 V1.1

Page 66 (82)

Figure 6-8: PDP and frequency correlation (FCF) of CDL model.

6.6 C1 – Urban macro-cell

The CDL parameters of LOS and NLOS condition are given below. In the LOS model Ricean K-factor is

12.9 dB.

Table 6-9 Scenario C1: LOS Clustered delay line model, suburban environment.

Cluster # Delay [ns] Power [dB] AoD [º] AoA [º] Ray power [dB]

1 0 5 10 0.0 -25.3

-27.1

0 0 -0.02

-33.1

2 85 -21.6 -29 -144 -34.7

3 135 -26.3 -32 -159 -39.3

4 135 -25.1 -31 155 -38.1

5 170 -25.4 31 156 -38.4

6 190 -22.0 29 -146 -35.0

7 275 -29.2 -33 168 -42.2

8 290 295 300 -24.3

-26.5

-28.2

35 -176 -34.3

9 290 -23.2 -30 149 -36.2

10 410 -32.2 35 -176 -45.2

11 445 -26.5 -32 -159 -39.5

12 500 -32.1 35 -176 -45.1

13 620 -28.5 33 -165 -41.5

14 655 -30.5 34 -171 -43.5

15 960 -32.6 35 177 -45.6

Cluster ASD = 5º

Cluster ASA = 5º

XPR = 8 dB

Power of dominant ray,

Power of each other ray

Figure 6-9: PDP and frequency correlation (FCF) of CDL model.

WINNER II D1.1.2 V1.1

Page 67 (82)

Table 6-10 Clustered delay-line model for Scenario C1 NLOS

Cluster

Delay [ns] Power [dB] AoD

[º]

AoA [º] Ray power [dB]

1 0 5 10 -3.0

-5.2

-7.0 0 0 -13.0

2 25 -7.5 13 -71 -20.5

3 35 -10.5 -15 -84 -23.5

4 35 -3.2 -8 46 -16.2

5 45 50 55 -6.1

-8.3

-10.1

12 -66 -16.1

6 65 -14.0 -17 -97 -27.0

7 65 -6.4 12 -66 -19.4

8 75 -3.1 -8 -46 -16.1

9 145 -4.6 -10 -56 -17.6

10 160 -8.0 -13 73 -21.0

11 195 -7.2 12 70 -20.2

12 200 -3.1 8 -46 -16.1

13 205 -9.5 14 -80 -22.5

14 770 -22.4 22 123 -35.4

Cluster ASD = 2º

Cluster ASA = 10º

XPR = 4 dB

Figure 6-10: PDP and frequency correlation (FCF) of CDL model.

6.7 C2 – Urban macro-cell

The CDL parameters of LOS and NLOS condition are given below. In the LOS model Ricean K-factor is

7.0 dB.

Table 6-11 Scenario C2: LOS Clustered delay line model.

Cluster # Delay [ns] Power [dB] AoD [º] AoA [º] Ray power [dB]

1 0 0.0 0 0 -0.08

-30.6

2 0 5 10 -16.2

-18.4

-20.2

-24 -120 -26.2

3 30 -15.3 26 129 -28.3

4 85 -16.7 -27 -135 -29.7

5 145 150 155 -18.2

-20.4

-22.2

26 -129 -28.2

6 150 -18.2 28 141 -31.2

7 160 -15.3 26 -129 -28.3

8 220 -23.1 -32 -158 -36.1

Cluster ASD = 6º

Cluster ASA = 12º

XPR = 8 dB

Power of dominant ray,

Power of each other ray

WINNER II D1.1.2 V1.1

Page 68 (82)

Figure 6-11: PDP and frequency correlation (FCF) of CDL model.

Table 6-12 Scenario C2: NLOS Clustered delay line model.

Cluster

Delay [ns] Power [dB] AoD

[º]

AoA [º] Ray power [dB]

1 0 -6.4 11 61 -19.5

2 60 -3.4 -8 44 -16.4

3 75 -2.0 -6 -34 -15.0

4 145 150 155 -3.0 -5.2 -7.0 0 0 -13.0

5 150 -1.9 6 33 -14.9

6 190 -3.4 8 -44 -16.4

7 220 225 230 -3.4 -5.6 -7.4 -12 -67 -13.4

8 335 -4.6 -9 52 -17.7

9 370 -7.8 -12 -67 -20.8

10 430 -7.8 -12 -67 -20.8

11 510 -9.3 13 -73 -22.3

12 685 -12.0 15 -83 -25.0

13 725 -8.5 -12 -70 -21.5

14 735 -13.2 -15 87 -26.2

15 800 -11.2 -14 80 -24.2

16 960 -20.8 19 109 -33.8

17 1020 -14.5 -16 91 -27.5

18 1100 -11.7 15 -82 -24.7

19 1210 -17.2 18 99 -30.2

20 1845 -16.7 17 98 -29.7

Cluster ASD = 2º

Cluster ASA = 15º

XPR = 7 dB

Figure 6-12: PDP and frequency correlation (FCF) of CDL model.

6.8 C3 – Bad urban macro-cell

The CDL parameters of NLOS condition are given below.

Table 6-13 Scenario C3: NLOS Clustered delay line model, bad urban, macrocell

WINNER II D1.1.2 V1.1

Page 69 (82)

Cluster

Delay [ns] Power [dB] AoD

[º]

AoA [º] Ray power [dB]

1 0 -3.5 -9 -52 -16.5

2 5 -8.9 14 -83 -22.0

3 35 -4.6 -10 -60 -17.6

4 60 -9.2 -14 -85 -22.2

5 160 165 170 -3 -5.2 -7 0 0 -13.0

6 180 -1.7 -6 -36 -14.7

7 240 -2.7 7 46 -15.7

8 275 -7 -12 74 -20.0

9 330 -5.9 11 68 -18.9

10 335 -6.7 -12 -72 -19.7

11 350 355 360 -4.3 -6.5 -8.3 -10 -62 -14.3

12 520 -5.3 -10 -64 -18.3

13 555 -4.9 -10 -62 -17.9

14 555 -9.4 14 85 -22.4

15 990 -12.3 16 -98 -25.3

16 1160 -12.2 16 -97 -25.2

17 1390 -20.8 21 127 -33.8

18 1825 -25.4 -23 140 -38.4

Cluster ASD = 2º

Cluster ASA = 15º

19 4800 -9.7 -135 25 -22.7

20 7100 -13 80 40 -26.0 2º 2º

XPR = 7 dB

Figure 6-13: PDP and frequency correlation (FCF) of CDL model.

6.9 C4 – Outdoor to indoor (urban) macro-cell

The CDL parameters of NLOS condition are given below.

Table 6-14 Scenario C4: NLOS Clustered delay line model, outdoor to indoor (urban) macro-cell

Cluster #

Delay [ns] Power [dB] AoD [º]

AoA [º] Ray power [dB]

1 0 5 10 -3.0 -5.2 -7.0 0 0 -13.0

2 15 -6.9 28 -91 -19.9

3 95 -3.6 -20 65 -16.6

4 145 -16.2 43 -139 -29.3

5 195 -8.5 -31 101 -21.5

6 215 -15.9 43 -138 -28.9

7 250 -6.9 28 -91 -19.9

8 445 -14.1 -40 130 -27.1

9 525 530 535 -3.8 -6.0 -7.8 45 -146 -13.8

10 815 -13.6 -39 128 -26.6

11 1055 -17.8 45 -146 -30.8

12 2310 -32.2 -61 196 -45.2

Cluster ASD = 5º

Cluster ASA = 8º

XPR = 9 dB

WINNER II D1.1.2 V1.1

Page 70 (82)

Figure 6-14: PDP and frequency correlation (FCF) of CDL model.

6.10 D1 – Rural macro-cell

The CDL parameters of LOS and NLOS condition are given below. In the LOS model Ricean K-factor is

5.7 dB.

Table 6-15 Scenario D1: LOS Clustered delay line model, rural environment.

Cluster # Delay [ns] Power [dB] AoD [º] AoA [º] Ray power [dB]

1 0 5 10 0.0 -15.0

-16.8

0 0 -0.23*

-22.8**

2 20 -15.5 17 44 -28.5

3 20 -16.2 17 -45 -29.2

4 25 30 35 -15.3

-17.5

-19.2

18 -48 -25.3

5 45 -20.5 -19 50 -33.5

6 65 -18.9 18 -48 -31.9

7 65 -21.1 -19 51 -34.2

8 90 -23.6 -20 -54 -36.6

9 125 -26.1 -22 57 -39.1

10 180 -29.4 23 -60 -42.4

11 190 -28.3 -22 59 -41.3

Cluster ASD = 2º

Cluster ASA = 3º

XPR = 12 dB

Power of dominant ray,

Power of each other ray

Figure 6-15: PDP and frequency correlation (FCF) of CDL model.

WINNER II D1.1.2 V1.1

Page 71 (82)

Table 6-16 Scenario D1: NLOS Clustered delay line model, rural environment.

Cluster #

Delay [ns] Power [dB]

AoD

[º]

AoA [º] Ray power [dB]

1 0 5 10 -3.0 -5.2 -7.0 0 0 -13.0

2 0 -1.8 -8 28 -14.8

3 5 -3.3 -10 38 -16.3

4 10 15 20 -4.8 -7.0 -8.8 15 -55 -14.8

5 20 -5.3 13 48 -18.3

6 25 -7.1 15 -55 -20.1

7 55 -9.0 -17 62 -22.0

8 100 -4.2 -12 42 -17.2

9 170 -12.4 20 -73 -25.4

10 420 -26.5 29 107 -39.5

Cluster ASD = 2º

Cluster ASA = 3º

XPR = 7 dB

Figure 6-16: PDP and frequency correlation (FCF) of CDL model.

6.11 D2a – Moving networks

The CDL parameters of LOS condition are given below. In the LOS model Ricean K-factor is 7 dB.

Table 6-17 Scenario D2: LOS Clustered delay line model, MRS-MS, rural

Cluster # Delay [ns] Power [dB] AoD [º] AoA [º] Ray power [dB]

1 0 0.0 0.0 0.0 -0.12

-28.8

2 45 50 55 -17.8

-20.1

-21.8

12.7 -80.0 -27.8

3 60 -17.2 -13.6 86.0 -30.2

4 85 -16.5 13.4 84.4 -29.5

5 100 105 110 -18.1

-20.4

-22.1

-13.9 87.5 -28.1

6 115 -15.7 -13.0 -82.2 -28.7

7 130 -17.7 -13.9 87.5 -30.8

8 210 -17.3 13.7 86.2 -30.3

Cluster ASD = 2º

Cluster ASA = 3º

XPR = 12 dB

Power of dominant ray,

Power of each other ray

Figure 6-17: PDP and frequency correlation (FCF) of CDL model.

WINNER II D1.1.2 V1.1

Page 72 (82)

6.12 Fixed feeder links - Scenario B5

For the stationary feeder scenarios only CDL models have been created. The CDL models are based on

the parameters in the tables below which are derived mostly from literature. Note that the CDL models

only approximate the selected parameters. Basically any antenna pattern can be used with the models

However, for the B5 scenario at distances larger than 300 meters the 3 dB beamwidth

3dB

of one of the

link ends should be smaller than 10 degrees while the other is smaller than 53 degrees.

6.12.1 Scenario B5a

The clustered delay-line model for the rooftop to rooftop case is given in table below. In stationary

scenarios, i.e. B5, the Doppler shifts of the rays are not a function of the AoAs. Instead, they are obtained

from the movement of the scatterers. In B5 we let one scatterer per cluster be moving while the others are

stationary. The Doppler frequency of the moving scatterers is also included in tables below.

Table 6-18 Parameters selected for scenario B5a LOS stationary feeder: rooftop to rooftop.

Parameter Value

Power-delay profile Exponential (non-direct paths).

Delay-spread 40ns

K-factor 10dB

XPR 30dB

Doppler A peak centreed around zero Hz with most energy within

0.1 Hz.

Angle-spread of non-direct components. Gaussian distributed clusters with 0.5 degrees intra angle-

spread. Composite angle-spread 2 degrees. Same in both

ends.

Table 6-19 LOS Clustered Delay-Line model. Rooftop-to-rooftop.

cluster

delay

[ns] Power

[dB] AoD [º]

AoA [º]

Freq. of

one

scatterer

mHz

K-

[dB] XPR = 30dB, MS speed N/A

1 0 -0.39 0.0 0.0 41.6 21.8 -0.42

-35.2

2 10 -20.6 0.9 0.2 -21.5 -33.61

3 20 -26.8 0.3 1.5 -65.2 -39.81

4 50 -24.2 -0.3 2.0 76.2 -37.21

5 90 -15.3 3.9 0.0 10.5 -28.31

6 95 -20.5 -0.8 3.6 -20.2 -33.51

7 100 -28.0 4.2 -0.7 1.3 -41.01

8 180 -18.8 -1.0 4.0 2.2 -31.81

9 205 -21.6 5.5 -2.0 -15.4 -34.61

10 260 -19.9 7.6 -4.1 48.9

-∞

Number of rays /cluster = 20

Ray Power [dB]

-32.91

cluster AS at MS [º] = 0.5

cluster AS at BS [º] = 0.5

Composite AS at MS [º] = 0.76

Composite AS at BS [º] =1.13

Power of dominant ray,

Power of each other ray

Clusters with high K-factor will have 21 rays.

WINNER II D1.1.2 V1.1

Page 73 (82)

6.12.2 Scenario B5b

The clustered delay-line model for range1, range2 and range3 (i.e. path loss < 85 dB, 85 dB < path loss <

110 dB, path loss > 110dB), is given in tables below.

Table 6-20 Parameters selected for scenario B5b LOS stationary feeder: street-level to street-level.

Parameter Value

Shadow-fading

free

3dB,

rr ≤

beyond

=7dB,

rr >

Range definition Range 1: Loss <85, Range 2: 85<Loss<110, Range 3: Loss>110.

Power-delay profile Exponential (of non-direct paths).

Delay-spread Range 1: 30ns. Range 2: 110ns. Range 3: 380ns.

K-factor Range 1: 10. Rang2: 2. Range 3: 1.

XPR 9dB.

Doppler The spectrum has a peak at 0Hz and most of it's power within an

few Hz.

Angle-spread of non-direct

components. Clusters are uniform distributed [0,360]. Intra-cluster spread is

2degrees.

Table 6-21 Clustered delay-line model street-level to street-level range 1.

cluster #

delay

[ns] Power

[dB] AoD [º]

AoA [º] Freq.

of one

scatterer

mHz

K-factor

[dB]

XPR = 9dB, MS speed N/A

1 0 -0.37 0.0 0.0 744 20.0 -0.41

-33.4

2 5 -15.9 -71.7 70.0 -5 -28.91

3 15 -22.2 167.4 -27.5 -2872 -35.21

4 20 -24.9 -143.2 106.4 434 -37.91

5 40 -26.6 34.6 94.8 295 -39.61

6 45 -26.2 -11.2 -94.0 118 -39.21

7 50 -22.3 78.2 48.6 2576 -35.31

8 70 -22.3 129.2 -96.6 400 -35.31

9 105 -29.5 -113.2 41.7 71 -42.51

10 115 -17.7 -13.5 -83.3 3069 -30.71

11 125 -29.6 145.2 176.8 1153 -42.61

12 135 -26.6 -172.0 93.7 -772 -39.61

13 140 -23.4 93.7 -6.4 1298 -36.41

14 240 -30.3 106.5 160.3 -343 -43.31

15 300 -27.7 -67.0 -50.1 -7 -40.71

16 345 -34.8 -95.1 -149.6 -186 -47.81

17 430 -38.5 -2.0 161.5 -2287 -51.51

18 440 -38.6 66.7 68.7 26 -51.61

19 465 -33.7 160.1 41.6 -1342 -46.71

20 625 -35.2 -21.8 142.2 -61

-∞

Number of rays/cluster = 20

Ray Power [dB]

-48.21

cluster AS at MS [º] = 2

cluster AS at BS [º] = 2

Composite AS at MS [º] =22.4

Composite AS at BS [º] = 26.2

Power of dominant ray,

Power of each other ray

Clusters with high K-factor will have 21 rays.

WINNER II D1.1.2 V1.1

Page 74 (82)

Table 6-22 Clustered delay-line model street-level to street-level range 2.

cluster #

delay

[ns] Power

[dB] AoD [º]

AoA [º] Freq. of

one

scatterer

mHz

K-factor

[dB]

XPR = 9dB, MS speed N/A

1 0 -1.5 0.0 0.0 744 13.0 -1.7

-27.7

2 5 -10.2 -71.7 70.0 -5 -23.21

3 30 -16.6 167.4 -27.5 -2872 -29.61

4 45 -19.2 -143.2 106.4 434 -32.21

5 75 -20.9 34.6 94.8 294 -33.91

6 90 -20.6 -11.2 -94.0 118 -33.61

7 105 -16.6 78.2 48.6 2576 -29.61

8 140 -16.6 129.2 -96.6 400 -29.61

9 210 -23.9 -113.2 41.7 71 -36.91

10 230 -12.0 -13.5 -83.3 3069 -25.01

11 250 -23.9 145.2 176.8 1153 -36.91

12 270 -21.0 -172.0 93.7 -772 -34.01

13 275 -17.7 93.7 -6.4 1298 -30.71

14 475 -24.6 106.5 160.3 -343 -37.61

15 595 -22.0 -67.0 -50.1 -7 -35.01

16 690 -29.2 -95.1 -149.6 -186 -42.21

17 855 -32.9 -2.0 161.5 -2288 -45.91

18 880 -32.9 66.7 68.7 26 -45.91

19 935 -28.0 160.1 41.6 -1342 -41.01

20 1245 -29.6 -21.8 142.2 -61

-∞

Number of rays/cluster = 20

Ray Power [dB]

-42.61

cluster AS at MS [º] = 2

cluster AS at BS [º] = 2

Composite AS at MS [º] =42.8

Composite AS at BS [º] = 50.2

Power of dominant ray,

Power of each other ray

Clusters with high K-factor will have 21 rays.

Table 6-23 Clustered delay-line model street-level to street-level range 3.

cluster #

delay

[ns] Power

[dB] AoD [º]

AoA [º] Freq. of

one

scatterer

mHz

K-factor

[dB]

XPR = 9dB, MS speed N/A

1 0 -2.6 0.0 0.0 744 10.0 -3.0

-26.0

2 10 -8.5 -71.7 70.0 -5 -21.51

3 90 -14.8 167.4 -27.5 -2872 -27.81

4 135 -17.5 -143.2 106.4 434 -30.51

5 230 -19.2 34.6 94.8 295 -32.21

6 275 -18.8 -11.2 -94.0 118 -31.81

7 310 -14.9 78.2 48.6 2576 -27.91

8 420 -14.9 129.2 -96.6 400 -27.91

9 630 -22.1 -113.2 41.7 71 -35.11

10 635 -10.3 -13.5 -83.3 3069 -23.31

11 745 -22.2 145.2 176.8 1153 -35.21

12 815 -19.2 -172.0 93.7 -772 -32.21

13 830 -16.0 93.7 -6.4 1298 -29.01

14 1430 -22.9 106.5 160.3 -343

-∞

Number of rays/cluster = 20

Ray Power [dB]

-35.91

cluster AS at MS [º] = 2

cluster AS at BS [º] = 2

Composite AS at MS [º] =52.3

Composite AS at BS [º] = 61.42

WINNER II D1.1.2 V1.1

Page 75 (82)

15 1790 -20.3 -67.0 -50.1 -7 -33.31

16 2075 -27.4 -95.1 -149.6 -186 -40.41

17 2570 -31.1 -2.0 161.5 -2287 -44.11

18 2635 -31.2 66.7 68.7 26 -44.21

19 2800 -26.3 160.1 41.6 -1342 -39.31

20 3740 -27.8 -21.8 142.2 -61 -40.81

Power of dominant ray,

Power of each other ray

Clusters with high K-factor will have 21 rays.

6.12.3 Scenario B5c

Model for B5c scenario is same with B1 LOS. Difference is that in B5c both the environment and both

link ends are stationary except two clusters, which represent moving vehicles. In these two clusters all the

rays have different non-zero Doppler frequency.

Table 6-24 B5c Clustered Delay-Line model.

cluster #

delay

[ns] Power

[dB] AoD [º]

AoA [º] Freq. of

one

scatterer

[mHz]

K-factor

[dB]

XPR = 9dB, MS speed N/A

1 0 0 0 0 -127 3.3 -1.67

-18.0

2 30 -11.7 5 45 385 -24.71

3 55 -14.8 8 63 -879 -27.81

4 60 -14.8 8 -69 ++ -27.81

5 105 -13.9 7 61 +++ -26.91

6 115 -17.8 8 -69 -735 -30.81

7 250 -19.6 -9 -73 -274 -32.61

8 460 -31.4 11 92 691

-∞

Number of rays/cluster =

Ray Power [dB]

-44.41

cluster AS at MS [º] = 18

cluster AS at BS [º] = 3

Composite AS at MS [º]

=45.0

Composite AS at BS [º] = 4.5

Power of dominant ray,

Power of each other ray

Clusters with high K-factor will have 21 rays.

Frequency for 20 scatterers in Hz is {45.0, 45.5, 46.0, 46.5, … , 54.5}

+++

Frequency for 20 scatterers in Hz is {-55.0, -55.5, -56.0, -56.5, … , -64.5}

6.12.4 Scenario B5f

Model for B5f scenario is NLOS version of B5a model.

Table 6-25 Parameters selected for scenario B5f NLOS stationary feeder: rooftop to rooftop.

Parameter Value

Power-delay profile Exponential (non-direct paths).

Delay-spread 85ns

K-factor -∞ dB

XPR 10dB

Doppler A peak centreed around zero Hz with most energy

within 0.1 Hz.

WINNER II D1.1.2 V1.1

Page 76 (82)

Table 6-26 B5f Clustered Delay-Line model. Rooftop-to-rooftop NLOS.

cluster

delay

[ns] Power

[dB] AoD [º]

AoA [º]

Freq. of

one

scatterer

[mHz]

K-

[dB] XPR = 10dB, MS speed N/A

1 0 -0.1 0.0 0.0 41.6 -13.11

2 10 -5.3 0.9 0.2 -21.5 -18.31

3 20 -11.5 0.3 1.5 -65.2 -24.51

4 50 -8.9 -0.3 2.0 76.2 -21.91

5 90 0.0 3.9 0.0 10.5 -13.01

6 95 -5.2 -0.8 3.6 -20.2 -18.21

7 100 -12.7 4.2 -0.7 1.3 -25.71

8 180 -3.5 -1.0 4.0 2.2 -16.51

9 205 -6.3 5.5 -2.0 -15.4 -19.31

10 260 -4.6 7.6 -4.1 48.9

-∞

Number of rays /cluster = 20

Ray Power [dB]

-17.61

cluster AS at MS [º] = 0.5

cluster AS at BS [º] = 0.5

Composite AS at MS [º] = 2.33

Composite AS at BS [º] =2.87

WINNER II D1.1.2 V1.1

Page 77 (82)

7. References

[3GPPSCM] 3GPP TR 25.996, "3

Generation Partnership Project; technical specification group radio

access networks; spatial channel model for MIMO simulations (release 6)", V6.1.0.

[AGV98] Pauli Aikio, Ralf Gruber and Pertti Vainikainen, Wideband radio channel measurements

for train tunnels. VTC 1998.

[AHY06] M. Alatossava, V-M. Holappa, J. Ylitalo, "Outdoor to indoor MIMO radio channel

measurements at 5.25 GHz – characterization of propagation parameters", EUCAP,

November 2006, Nice, France.

[AI00] Farrokh Abrishamkar , James Irvine, Comparison of Current Solutions for the Provision

of Voice Services to Passengers on High Speed Trains. IEEE VTC-Fall, VTC 2000. 24 -

28 Sept. 2000.

[APM02] A. Algans, K. I. Pedersen, P.E. Mogensen,"Experimental analysis of the joint statistical

properties of azimuth spread, delay spread and shadow fading", IEEE J. Selected Areas in

Comm., Vol. 20, pp.523-531, 2002.

[AP02] J. B. Andersen and K. I. Pedersen, "Angle-of-arrival statistics for low resolution

antennas," IEEE Trans. Antennas Propagat., vol. 50, pp. 391–395, Mar. 2002.

[B+03] R. J. C. Bultitude et al., "A propagation-measurement-based evaluation of channel

characteristics and models pertinent to the expansion of mobile radio systems to

frequencies beyond 2 GHz, " IEEE Trans. VT, vol. 56, no. 2, pp. 382-388, Mar. 2007

[Bul02] Bultitude, R.J.C., "A Comparison of Multipath-Dispersion-Related Micro-Cellular Mobile

Radio Channel Characteristics at 1.9 GHz and 5.8 GHz", in Proc. ANTEM'02, Montreal,

Jul. 31 – Aug. 2, 2002, pp. 623-626.

[BB89] R.J.C. Bultitude, and G.K.Bedal, "Propagation characteristics on microcellular urban

mobile radio channels at 910 MHz," IEEE J. Select. Areas Commun, vol.7, no. 1, 1989,

pp. 31-39.

[Bul02] R.J.C. Bultitude., "Estimating frequency correlation functions from propagation

measurements on fading radio channels: A critical review," IEEE J. Select. Areas

Commun. Vol. 20, no. 6, August, 2002, pp. 1133-1143.

[BBK04] M.D. Batariere, T.K. Blankenship, J.F. Kepler, T.P. Krauss,"Seasonal variations in path

loss in the 3.7 GHz band", IEEE RAWCON, pp. 399-402, 2004.

[BBK+02] M.D. Batariere, T.K. Blankenship, J.F. Kepler, T.P. Krauss, I. Lisica, S. Mukthvaram,

J.W. Porter, T.A. Thomas, F.W. Vook: "Wideband MIMO mobile impulse response

measurements at 3.7 GHz", IEEE 55

VTC, pp. 26-30, 2002.

[BHS05] D. Baum, J. Hansen, J. Salo, G. Del Galdo, M. Milojvic, P. Kyösti: "An Interim Channel

Model for Beyond-3G Systems", IEEE VTC'05, April 2005.

[Cal+07] G. Calcev, D. Chizhik, B. Goeransson, S. Howard, H. Huang, A. Kogiantis, A. F.

Molisch, A. L. Moustakas, D. Reed and H. Xu, "A Wideband Spatial Channel Model for

System-Wide Simulations", IEEE Trans. Vehicular Techn., March 2007.

[CBW95]

D. J. Cichon, T. C. Beckcr, W. Wiesbeck, "Determination of Time-Variant Radio Links in

High-Speed Train Tunnels by Ray Optical Modeling". AP-S 1995.

[Cha03] Ashok Chandra, Propagation of 2000 MHz Radio Signal into A Multi-Storeyed building

Through Outdoor-Indoor Interface. The 14" IEEE 2003 International Symposium on

Personal,lndoor and Mobile Radio Communicatlon Proceedings. 2003.

[C02] M. Celidonio, "Outdoor-indoor propagation measurements in the 3.6-4.2 GHz band",

IEEE 13th PIMRC, Vol. 2, pp. 644-648, 2002.

[CG99] T.-S. Chu, and L. J. Greenstein, "A quantification of link budget differences between the

cellular and PCS bands," IEEE Trans VT, vol. 48, no. 1, pp. 60-65, Jan. 1999.

[CKC03] A. Chandra, A. Kumar, P. Chandra, "Propagation of 2000 MHz radio signal into a multi-

storeyed building through outdoor-indoor interface", Proceedings on 14th PIMRC, Vol. 3,

pp. 2983-2987, 2003.

WINNER II D1.1.2 V1.1

Page 78 (82)

[COST231] European Comission: European cooperation on the field of scientific and technical

research (EURO-COST 231): "Digital mobile radio towards future generation systems",

Final report, http://www.lx.it.pt/cost231/, Bruxelles, 1999.

[DDA00] Michael Doehler, Monica Dell' Anna, A. H. Aghvami, Pdf - Transformation for the

Outdoor-Indoor Propagation Model. VTC 2000-Spring. 2000, 15 – 18 May 2000 Pages

1646-1650 Vol. 3.

[deJKH02] Y.L.C.deJong. M.H.J.L Koelen, and M.H.J.Herben, ", "A building transmission model for

improved propagation prediction in urban microcells," IEEE Trans., Veh. Technol., vol.

53, no. 2, 2002.

[DRX98] G. Durkin, T.S Rappaport, H". Xu:"Measurements and models for radio path loss and

penetration in and around homes and trees at 5.85 GHz, IEEE Trans. Comm., Vol. 46,

pp.1484-1496, 1998.

[EGT+99] V. Erceg , L.J. Greenstein, S. Tjandra, S. Parkoff, A. Gupta, B. kulic, A. Julius, R.

Bianci.,"An empirically based path loss model for wireless channels in suburban

environments", IEEE J. Sel. Areas Comm., Vol. 17, pp. 1205-1211, 1999.

[Erc+01] V. Erceg, et al., "Channel Models for Fixed Wireless Applications" (IEEE802.16.3c-

01/29r4), IEEE P802.16, Broadband Wireless Working Group, 2001.

[ESB+04] V. Erceg, P. Soma, D.S. Baum, S. Catreux, "Multiple-input multiple-output fixed wireless

radio channel measurements and modeling using dual-polarized antennas at 2.5 GHz" in

IEEE Trans. on Wireless Communications, Vol 3, Nov 2006.

[FDS+94] Foster, H.M.; Dehghan, S.F.; Steele, R.; Stefanov, J.J.; Strelouhov, H.K., "Microcellular

measurements and their prediction", IEE Colloquium on "Role of Site Shielding in

Prediction Models for Urban Radiowave Propagation" (Digest No. 1994/231), Nov. 1994,

pp. 2/1 - 2/6.

[Fuj03] T. Fujii, "Path loss prediction formula in mobile communication -an expansion of

"SAKAGAMI" path loss prediction formula-," Trans. IEICE, Japan, J86-B, 10, pp. 2264-

2267, 2003.

[GEA03] S. Geng et al., " Measurements and Analysis of Wideband Indoor Radio Channels at 60

GHz," Proc. of 3rd ESA Workshop on Millimeter Wave Technology and Applications,

Espoo, Finland, May 21-23, 2003.

[GEY97] L.J. Greenstein, V. Erceg, Y.S. Yeh, M.V. Clark,"A new path-gain/delay spread model for

digital cellular channels", IEEE Trans. Veh. Tech., Vol. 46, pp. 477-485, 1997.

[GRZ07] Guo Rui Zhang, "Measurement and Characterisation of Delay, Angular and Frequency

Dispersion and their Evolution at Mobile Receivers in 2.25 GHz Microcells," M.Sc.

Thesis. Report, Dept. of Systems and Computer Engineering, Carleton University,

Ottawa, to be submitted, spring 2007.

[GTD+04] M. Ghaddar, L. Talbi, T.A. Denidni, A. Charbonneau, "Modeling human body effects for

indoor radio channel using UTD", in Canadian Conference on Electrical and Computer

Engineering, Vol. 3, Pages: 1357- 1360, 2-5 May 2004

[Gud91] M. Gudmundson, "Correlation model for shadow fading in mobile radio systems",

Electron. letter, vol. 27, pp. 2145-2146, Nov. 1991.

[ITU] ITU Rec. ITU-R P.1238-4.

[ITU-R] Rec. ITU-R P.1546-2, "Method for point-to-area predictions for terrestrial services in the

frequency range 30 MHz to 3000 MHz"

[JHH+05] T. Jämsä, V. Hovinen, L. Hentilä, and J. Iinatti, "Comparisons of Wideband and Ultra-

wideband channel measurements " IEEE IWS2005/WPMC2005, Aalborg, Denmark.

[KeM90] J.M Keenan, A.J. Motley, "Radio coverage in buildings", British Telecom Tech. Journal,

vol.8, no.1, Jan. 1990, pp.19-24.

[KI04] K. Kitao, S. Ichitsubo,"Path loss prediction formula for microcell in 400 MHz to 8 GHz

band", Electronics Letters, Vol. 40, No. 11, 2004.

[KJ07] Pekka Kyösti and Tommi Jämsä, "Complexity Comparison of MIMO Channel Modelling

Methods", ISWCS'07, Trondheim, Norway, October 2007.

WINNER II D1.1.2 V1.1

Page 79 (82)

[KKM02] J. F. Kepler, T.P. Krauss, S. Mukthvaram:"Delay spread measurements on a wideband

MIMO channel at 3.7 GHz", IEEE 56th VTC, pp. 2498-2502, 2002.

[KBM+06] Sandra Knoerzer, Michael A. Baldauf, Juergen Maurer, Werner Wiesbeck, "OFDM for

Multimedia Applications in High-Speed Trains: Channel Model Including Different

Antenna Types.", 2006.

[KMV+05] Sandra Knoerzer, Juergen Maurer, Sven Vogeler, Karl-Dirk Kammeyer and Werner

Wiesbeck, "Channel Modeling for a High-Speed Train OFDM Communication Link

Supporting High Data Rates". ITST 2005.

[KP02] M. B. Knudsen, Member, IEEE, and G. F. Pedersen, "Spherical Outdoor to Indoor Power

Spectrum Model at the Mobile Terminal", IEEE Journal on Selected Areas in

Communications, vol. 20, no. 6, August 2002.

[KRB00] A. Kuchar, J-P. Rossi, E. Bonek:"Directional macro-cell channel characterization from

urban measurements", IEEE Trans. Antennas and Propagation, Vol. 48, pp.137-145, 2000.

[KSL+02] K. Kalliola, K. Sulonen, H. Laitinen, O. Kivekäs, J. Krogerus, P. Vainikainen:"Angular

power distribution and mean effective gain of mobile antenna in different propagation

environments", IEEE Trans. Veh. Techn., Vol. 51, pp.823-838, 2002.

[KZV99] J. Kivinen, X. Zhao and P. Vainikainen, "Wideband Indoor Radio Channel Measurements

with Direction of Arrival Estimations in the 5 GHz Band," Proc. of IEEE Vehicular and

Technology Conference (VTC'99), pp. 2308-2312, Netherlands, 1999.

[Lan02] J. Nicholas Laneman "Cooperative Diversity in Wireless Networks: Algorithms and

Architectures," PhD thesis, sep. 2002

[Medav] http://www.channelsounder.de

[MEJ91] P.E. Mogensen, P. Eggers, C. Jensen:"Urban area radio propagation measurements for

GSM/DCS 1800 macro and micro cells", ICAP 91, pp. 500-503, 1991.

[MHA+04] J. Medbo, F. Harryson, H. Asplund, J-E. Berg, "Measurements and analysis of a MIMO

macrocell outdoor-indoor scenario at 1947 MHz", IEEE 59th VTC, Vol. 1, pp. 261-265,

2004.

[MKA02] Masui, H.; Kobayashi, T.; Akaike, M., "Microwave path-loss modeling in urban line-of-

sight environments," IEEE Journal on Selected Areas in Communications, Vol. 20, Iss. 6,

Aug. 2002, pp. 1151-1155.

[MOT02] Y. Miura, Y. Oda, T. Taka, "Outdoor-to indoor propagation modelling with the

identification of path passing through wall openings", IEEE 13th PIMRC, Vol. 1, pp. 130-

134, 2002

[MRA93] L. Melin, M. Rönnlund, R. Angbratt: "Radio wave propagation – a comparison between

900 and 1800 MHz", IEEE 43

VTC conference, pp. 250-252, 1993.

[NAP05] H. T. Nguyen, J. B. Andersen, G. F. Pedersen, "Characterization of the indoor/outdoor to

indoor MIMO radio channel at 2.140 GHz", Department of Communication Technology,

Aalborg, Denmark, 2005.

[OP04] C. Oestges, and A.J. Paulraj, "Propagation into buildings for Broadband Wireless

Access," IEEE Trans. Veh. Techn., vol. 53, No. 2, pp. 521-526, March 2004.

[OBL+02] J. Ojala, R. Böhme, A. Lappeteläinen and M. Uno, "On the propagation characteristics of

the 5 GHz rooftop-to-rooftop meshed network," IST Mobile & Wireless

Telecommunications Summit 2002, Jun. 2002, Thessaloniki, Greece.

[OC07] C. Oestges, and B. Clerckx, "Modeling outdoor macrocellular clusters based on 1.9-GHz

experimental data," IEEE Trans. Veh. Tech., vol. 56, No. 6, November 2007.

[OVC06] C. Oestges, D. Vanhoenacker-Janvier, B. Clerckx , "Channel Characterization of Indoor

Wireless Personal Area Networks", in IEEE Transactions on Antennas and Propagation,

Vol. 54, Issue 11, Pages:3143 – 3150, Nov. 2006.

[OKT+04] Ohta, G.I.; Kamada, F.; Teramura, N.; Hojo, H., "5 GHz W-LAN verification for public

mobile applications - Internet newspaper on train and advanced ambulance car

Consumer Communications and Networking Conference, 2004. CCNC 2004. First IEEE.

Volume , Issue , 5-8 Jan. 2004 Page(s): 569 – 574.

WINNER II D1.1.2 V1.1

Page 80 (82)

[OOK+68] Y. Okumura, E. Ohmori, T. Kawano, K. Fukuda:"Field strength and its variability in VHF

and UHF land-mobile radio services", Review of the Electrical Comm. Lab., Vol. 16, No

9. 1968.

[OTT+01] Y. Oda, R. Tsuchihashi, K. Tsunekawa, M. Hata,"Measured path loss and multipath

propagation characteristics in UHF and microwave frequency band for urban mobile

communications", VTC 2001 Spring, Vol.1, pp. 337-341, 2001.

[Pab04] Ralf Pabst et al. ,"Relay based deployment concepts for wireless and mobile broadband

radio," IEEE communication magazine, Sep. 2004, pp 80-87.

[Paj03] P. Pajusco:"Double characterisations of power angule spectrum in macrocell

environment", Electronics Letters, Vol. 39, pp. 1565-1567, 2003.

[Pap05] P. Papazian, "Basic transmission loss and delay spread measurements for frequencies

between 430 and 5750 MHz", IEEE Trans. Ant. Propagation, Vol. 53, pp. 694-701, 2005.

[PCH01] E. Perahia, D. Cox, S. Ho, "Shadow fading cross-correlation between base stations", IEEE

VTC, pp. 313-317, May 2001.

[PLB04] J.W. Porter, I. Lisica, G. Buchwald,"Wideband mobile propagation measurements at 3.7

GHz in an urban environment", IEEE Ant. Propagat. Intern. Symposium, Vol. 4, pp.

3645-3846, 2004.

[PLN+99] M. Pettersen, P. H. Lehne, J. Noll, O. Rostbakken, E. Antonsen, R. Eckhoff,

"Characterization of the directional wideband radio channel in urban and suburban areas",

IEEE 50

VTC, pp. 1454-1459,1999.

[PMF00] K. I. Pedersen, P.E. Mogensen, B.H. Fleury, "A stochastic model of the temporal and

azimuthal dispersion seen at the base statin in outdoor propagation environments", IEEE

Trans. Veh. Technol., Vol. 49, pp. 437-447, 2000.

[Psound] http://www.propsim.com

[PT00] J. W Porter and J. A Thweatt, "Microwave Propagation Characteristics in the MMDS

Frequency Band," in Proc. IEEE ICC'00, Jun. 2000, Vol. 3, pp. 1578-1582.

[RMB+06] M. Riback, J. Medbo, J.-E. Berg, F. Harrysson, and H. Asplund, "Carrier Frequency

Effects on Path Loss," IEEE VTC 2006-Spring, vol. 6, pp. 2717-2721, 2006.

[RSS90] T. Rappaport, S. Seidel, R. Singh.,"900-MHz multipath propagation measurements fot

U.S. digital cellular radiotelephone", IEEE Trans. Veh. Technol., Vol. 39, pp.132-139,

1990.

[RKJ05] T. Rautiainen, K. Kalliola, J. Juntunen,"Wideband radio propagation characteristics at 5.3

GHz in suburban environments", PIMRC Berlin, Vol. 2, pp. 868-872, 2005.

[RJK07] T. Rautiainen, J. Juntunen, K. Kalliola, "Propagation analysis at 5.3 GHz in typical and

bad urban macrocellular environments", VTC Dublin, April 2007.

[RWH02] T. Rautiainen, G. Wölfle, R. Hoppe, "Verifying path loss and delay spread predictions of

a 3D ray tracing model in urban environment", IEEE 56

Veh. Technol. Conf., Vol. 4, pp.

2470-2474, Sept. 2002.

[Rudd03] R.F. Rudd, "Building penetration loss for slant-paths at L-, S- and C-band." ICAP 2003,

31.3.-1.4, 2003.

[Sau99] S. Saunders "Antenna and propagation for communication systems concept and design",

Wiley, 1999.

[SG02] S. Salous, H. Gokalp,"Dual-frequency sounder for UMTS frequency-division duplex

channels", IEE Proc. Comm., Vol. 149, pp. 117-122, 2002.

[SRJ+91] S. Seidel, T. Rappaport, S. Jain, K. Lord, R. Singh,"Path loss, scattering, and multipath

delay statistics in four European cities for digital cellular and microcellular

radiotelephone", IEEE Trans. Veh. Technol., Vol. 40, pp. 721-730, 1991.

[SBA+02] Schenk, T.C.W., Bultitude, R.J.C., Augustin, L.M., van Poppel, R.H., and Brussaard, G.,

"Analysis of Propagation loss in Urban Microcells at 1.9 GHz and 5.8 GHz," in Proc.

URSI Commision F Open Symposium on Radiowave Propagation and Remote Sensing,

Garmisch-Patenkirchen, Germany, Feb. 2002.

WINNER II D1.1.2 V1.1

Page 81 (82)

[SCK05] N. Skentos, Constantinou and A. G Kanatas, "Results from Rooftop to Rooftop MIMO

Channel Measurements at 5.2 GHz," COST273 TD(05)59, Bologna, Jan. 19-21.

[SCT03] A. Seville, S. Cirstea and J.F. Taylor. Effects of propagation between the indoor and

outdoor environment. ICAP 2003. 31.3.-1.4. 2003.

[SG00] T. Schwengler, M. Gilbert:"Propagation models at 5.8 GHz - path loss and building

penetration", IEEE Radio and Wireless Conference 10-13 Sep. 2000, pp. 119-124.

[SJD94] E. Sousa, V. Jovanovic, C. Dainegault, "Delay spread measurements for the digital

cellular channel in Toronto", IEEE Trans. Veh. Technol., Vol. 43, pp. 837-846, 1994.

[SMI+00] H. Shimizu, H. Masui, M. Ishi, K. Sakawa, and T. Kobayashi, "LOS and NLOS Path-Loss

and Delay Characteristics at 3.35 GHz in a Residential Environment," IEEE Antennas and

Propagation Society International Symposium 2000, Vol. 2, Jul. 2000, pp. 1142 - 1145.

[SMI+02] K. Sakawa, H. Masui, M. Ishii, H. Shimizu, T. Kobayashi, "Microwave path-loss

characteristics in an urban area with base station antenna on top of a tall building", Int.

Zurich Seminar on Broadband communications, pp. 31-1 -31-4, 2002.

[SMJ+99] A. M. Street, J. G. O. Moss, A. P. Jenkins, D. J. Edwards, "Outdoor-indoor wideband

study for third generation communication systems", IEE National Conference on

Antennas and Propagation, pp. 128-131, 1999.

[SS01] S. Stavrou, S. R. Saunders, "A deterministic outdoor to indoor propagation modeling

approach", IEEE 54th VTC, Vol. 2, pp. 1097-1100, 2001.

[SMB01] M. Steinbauer, A. F. Molisch, and E. Bonek, "The double-directional radio channel,"

IEEE Antennas and Propagation Mag., pp. 51–63, August 2001.

[SV87] A. Saleh, and R. A. Valenzuela, A statistical model for indoor multipathpropagation,

IEEE J. Select. Areas Commun., vol. SAC-5, no. 2, Feb. 1987, pp. 128–137.

[Sva02] Svantesson, T., "A double-bounce channel model for multi-polarized MIMO systems," in

Proc. IEEE VTC'02-Fall, Vol. 2, Sep. 2002, pp. 691 – 695.

[TPE02] S. Thoen, L. Van der Perre, and M. Engels, "Modeling the Channel Time-Variance for

Fixed Wireless Communication", IEEE Communication Letters, Vol. 6, No. 8, Aug. 2002.

[VES00] F. Villanese, N.E. Evans, W.G. Scanlon, "Pedestrian-induced fading for indoor channels

at 2.45, 5.7 and 62GHz", in IEEE VTS-Fall VTC 2000, Vol. 1, Pages: 43-48 vol.1, 2000.

[VKV04] L. Vuokko, J. Kivinen, P. Vainikainen., "Results from 5.3 GHz MIMO measurement

campaign", COST273, TD(04)193, Duisburg, Germany, 20.-22.9.2004.

[WAE+04] S. Wyne, P. Almers, G. Eriksson, J. Karedal, F. Tufvesson, A. F. Molisch, "Outdoor to

indoor office MIMO measurements at 5.2 GHz", IEEE 60th VTC, pp. 101-105, 2004.

[WHL94] J.A. Wepman, J.R. Hoffman, L.H. Loew,"Characterization of macrocellular PCS

propagation channels in the 1850-1990 MHz band, 3rd Annual International Conference

on Universal Personal Communications, pp. 165-170, 1994.

[WHL+93] J.A. Wepman, J.R. Hoffman, L.H. Loew, W.J. Tanis, M.E. Hughes: "Impulse response

measurements in the 902.928 and 1850.1990 MHz bands in macrocellular environments",

2nd international conference on Universal Personal Communications, Vol. 2, pp. 590-594,

1993.

[WIN1D54] WINNER1 WP5: "Final Report on Link Level and System Level Channel Models"

Deliverable D5.4, 18.11.2005

[WIN1D72] WINNER WP7, System assessment criteria specification, v1.0.

[WIN2IR111] WINNER2 WP1: "Propagation Scenarios", Internal Report IR1.1.1, 22.5.2006.

[WIN2UCM] WINNER2 WP1: "Updated Channel Models", Internal, June 2006.

[WINNERII] WINNER II Contract, Annex I – "Description of Work", IST-4-027756, 25/10/2005.

[WAE+04] S. Wyne, P. Almers, G. Eriksson, J. Karedal, F. Tufvesson, and A. F. Molisch, Outdoor to

Indoor Office MIMO Measurements at 5.2 GHz. VTC 2004-Fall. 2004 IEEE 60

Volume

1, 26-29 Sept. 2004 Pages 101-105 Vol. 1.

WINNER II D1.1.2 V1.1

Page 82 (82)

[WH02] J. Weitzen, T. J. Lowe, "Measurement of angular and distance correlation properties of

log-normal shadowing at 1900 MHz and its application to design of PCS systems", IEEE

transactions on vehicular technology, vol. 51, No. 2, march 2002.

[WMA+05] S. Wyne, A. F. Molisch, P. Almers, G. Eriksson, J. Karedal, F. Tufvesson, "Statistical

evaluation of outdoor-to-indoor office MIMO measurements at 5.2 GHz", IEEE 61st

VTC, Vol. 1, pp. 146-150, 2005.

[WOT99] G. Woodward, I. Oppermann, J. Talvitie, "Outdoor-indoor temporal & spatial wideband

channel model for ISM bands", IEEE 50th VTC, Vol. 1, pp. 136-140, 1999.

[ZEA99] X. Zhao et al, "Diffraction over typical-shaped terrain obstacles," Journal of

Electromagnetic Waves and Applications, vol. 13, pp. 1691-1707, 1999.

[Zet05] P. Zetterberg, "Auto and Multi-Site Correlation of Large Scale Parameters: Model

Evolution", Internal WINNER document, Aug. 2005.

[Xia96] H.H. Xia:" An analytical model for predicting path loss in urban and suburban

environments", Seventh IEEE Int. Symposium PIMRC, Vol 1, pp. 19.23, 1996

[XBM+94] H. Xia, H. Bertoni, L. Maciel, A. Lindsey-Steward, R. Rowe, "Microcellular propagation

characteristics for personal communications in urban and suburban environments", IEEE

Trans. Veh. Technol., Vol. 43, pp- 743-752, 1994.

[YIT06] Yonezawa, Ishikawa, Takeuchi, "Frequency range extension of path loss prediction

formula for over-rooftops propagation in microwave band", IEEE International Symp.

Antennas Propagation, pp. 4747-4750, 2006.

[YMI+04] K. Yonezawa, T. Maeyama, H. Iwai, H. Harada:" Path loss measurement in 5 GHz macro

cellular systems and considerations of extending existing path loss prediction models",

IEEE WCNC, Vol. 1, pp. 279-283, 2004.

[ZKV+02] X. Zhao, J. Kivinen, P. Vainikainen, K. Skog:"Propagation characteristics for wideband

outdoor mobile communications at 5.3 GHz", IEEE Sel. Areas Comm., Vol. 20, pp. 507-

514, 2002.

[ZRK+06] X. Zhao, T. Rautiainen, K. Kalliola, P. Vainikainen, "Path-loss models for urban

microcells at 5.3 GHz, IEEE Antennas and Propag. Lett., Vol. 5, pp. 152-154, 2006.

... The channel propagation is modelled using the WINNER II B1 propagation model [20] for the low frequency (3.5 GHz) ultra-small cell access network as both the BSs and MSs are deployed outdoors. In the WINNER II models the propagation parameters may vary over time between the channel segments. ...

... where k, l ε{1, 2}. Other important parameters mentioned in Eq. 7 to Eq. 10 are further explained in [20]. During the uplink transmission, the effective signal strength at the receiver is obtained by accounting for the gains of MS and BS antennas, shadowing, path loss on the channel and interference from other users using the same resource blocks. ...

... Since, the system will always be in a particular state; therefore, the state probabilities should essentially satisfy the normalization equation [16], given by the following Eq. 20: n j1=0 n j2=0 P (j 1 , j 2 ) = 1 (20) The (n + 1) 2 equations along-with the normalization equation could be represented in a matrix format, given by: ...

Millimetre-wave ultra-dense high capacity networks by providing an important component of future 5G cellular systems, by providing extremely high capacity to end users. Disparate types of users coexist in such scenarios, which can make the heterogeneous network unfair in terms of allocation of resources. A mechanism is required for effective spectrum sharing and to achieve overall system fairness. In this paper, an analytical model is suggested, which is based on a two-dimensional Markov state-transition diagram, to help set the parameter values to control the issuance of resources in coexistence layouts. A restriction approach is further implemented to gain a fair balance of the Grade-of-Service for both user groups using the User Admission Control mechanism. The developed mechanism restricts access to various channel resources for users with complete choice to give a greater probability of access to different users with limited resource options. Various levels of restriction are investigated in order to offer a balanced low-blocking probability performance to both user groups in order to improve the overall network fairness. Also, the proposed approach could provide a precise level of Grade-of-Service guarantee for both the user groups if sufficient flexibility is available within the whole network. Our simulation results show that approximately 30% to 45% enhancement, in terms of grade of service (GoS), could be achieved in high to medium loads by restricting some users' flexibility.

... Indeed, some UEs have strong Line-of-Sight (LOS) components and can undergo spatially-correlated smallscale fading, with common propagation paths [18]. The geometric channel model is thus appropriate for simulating the sub-6 GHz radio environment, with recommended statistical parameters as in the COST 2100 [80] or the WINNER II frameworks [81]. ...

... The superposition of several paths results in correlation between antenna elements and temporal fading with corresponding Doppler spectrum. Further information about the Winner II channel model can be found in[81]. ...

Flavio Maschietti

In the context of 5G and 5G+ mobile networks, massive multi-antenna transmission is an established technique to manage multi-user interference and improve the network performance through beamforming and multiplexing gain. In the massive antenna regime, the leading forms of distributed cooperation that can be envisioned are i) the beam selection and alignment across multiple mobile users – in particular, at mmWave frequencies – and ii) the cooperation among base stations for user scheduling, whose centralized solution requires significant coordination and resource overhead. In this thesis, we focus on decentralized cooperative methods for massive multi-antenna transmission optimization that are implemented at the cooperating devicesthemselves.

... Moreover, the effective density at a given snapshot in a cell using a certain repetition profile is denoted as λ (r). Average noise power −118 dBm †2 †1 WINNER II channel model measurements [210] †2 Calculated from noise figure of 3 dB and bandwidth 180 kHz ...

Bisma Manzoor

Advancement in radio communication has been a vital part of technological evolution, where the recent emergence of the Internet-of-Things (IoT), an ecosystem of remotely connected devices, has revolutionized the ICT paradigm and changed the way machines and humans interact with their surroundings. The accelerating growth of IoT applications is making the world a better-connected place, however, it concurrently challenges the researchers and network operators to devise solutions that meet the demands of the expanding IoT networks. Most of the challenges are related to sustaining a massive number of devices while simultaneously ensuring deep coverage and prolonged battery life of the end IoT devices. Moreover, due to the ubiquity of IoT applications, it has become crucial to provide global network coverage around the world. The low-complexity IoT devices which are supported by Low Power Wide Area Network (LPWAN), and require efficient energy consumption, low-throughput, and good coverage are classified as mMTC (massive Machine Type Communication), is a vital service category in 5G. While many of the popular IoT access technologies operate in the unlicensed frequency spectrum, 3GPP in 2015 standardized cellular IoT access technologies among which Narrow Band-IoT (NB-IoT) is popularly adopted by Telecom operators. However, the process of acquiring a new licensed spectrum poses financial and administrative challenges, especially for emerging small and medium-sized operators. This opens a door for the examination of deployment of NB-IoT in the unlicensed frequency bands, which will aid in the broader adoption of NB-IoT. Furthermore, one of the key characteristics of NB-IoT is to provide extended coverage. The coverage improvement is achieved by a repetition mechanism according to which the device repeats the same message multiple times. However, this technique comes at the cost of increased energy consumption of IoT devices. As a result, there is a trade-off between the coverage and energy expenditure of devices, thus calling for careful analysis and investigation. Another challenge in the domain of IoT is to provide adequate connectivity to remote areas where terrestrial telecommunication infrastructure is hard to deploy, also during times of natural disasters e.g., tsunami, earthquake when the terrestrial communication fails. One promising means in enabling remote and global network coverage is the use of non-terrestrial infrastructure that includes satellite and UAV networks. Although communication via satellites was dominated by applications such as navigation, military, broadcasting, the recent advancement in technologies has paved the way for IoT communication over non-terrestrial networks (such as UAV and Satellite). However, coexistence of terrestrial IoT access technologies over the non-terrestrial networks requires proper investigation due to a distinct satellite-to-ground propagation channel between the satellite and IoT devices. This thesis aims at modeling and optimizing the characteristics of cellular IoT networks. To address the challenges mentioned above, we first develop a geometric model to analyze the coverage of a dense urban IoT network in 3D spatial dimensions. The model is built utilizing mathematical tools from stochastic geometry and is implemented and tested using simulations and Ray-Tracing methodologies. The model establishes the ground for further investigation and analysis of high capacity mMTC networks. After that, we examine the performance of NB-IoT operating in the unlicensed ISM frequency band under realistic interference scenarios while employing the packet-repetition feature to examine the extended coverage. The investigation is carried out by embedding the actual interference into the link-level simulations for NB-IoT. The interference measurements are captured from the ISM band in a dense urban environment of Melbourne CBD, Australia, using a software-defined radio. The results pave the way for the possible deployment of NB-IoT in the unlicensed spectrum. Furthermore, the developed framework is used to generate a coverage map of IoT devices employing the repetition scheme, which aids in capturing the performance of IoT repetitions. To further understand the impact of repetitions on energy expenditure of devices and the resource occupation, this research presents a new mathematical model for frame-repetition in LPWAN IoT networks. The model is developed for the IoT uplink and aims at obtaining the optimal repetition rate across an IoT cell. The work first captures the imbalance between the success, in terms of coverage probability, and the elevated interference, in a cell implementing a repetition scheme. The model then provides the flexibility of tuning the repetition profile of a cell to acquire an optimal performance zone. In addition, the model is expanded to examine the energy cost of the devices employing repetitions. The analysis is carried out for two diversity combining techniques which are: (i) Selection Combining and (ii) Maximal Ratio Combining. Therefore, we shed light on the methodology that aids in improving service availability, radio resource efficiency, and energy consumption. The theoretical analysis and formulas are verified using Monte-Carlo simulations and prove the plausibility of the proposed optimization approach. Finally, this work extends the application of the developed repetition model into the non-terrestrial network. The probability of coverage is analyzed by employing a non-terrestrial propagation channel in the uplink between the satellite and IoT devices located in the satellite's service area, accordingly, an optimal repetition rate is formulated based on the satellite orbital and antenna parameters.

... In this study, we propose a modeling strategy for RISempowered communication systems by considering the currently used technical specifications on sub-6 GHz bands [10], [11]. According to a common assumption in the literature, the most efficient use of an RIS is possible when it is placed close to the terminals. ...

Reconfigurable intelligent surface (RIS)-assisted communications is one of the promising candidates for next generation wireless networks by controlling the propagation environment dynamically. In this study, a channel modeling strategy for RIS-assisted wireless networks is introduced in sub-6 GHz bands by considering both far-field and near-field behaviours in transmission. We also proposed an open-source physical channel simulator for sub-6 GHz bands where operating frequency, propagation environment, terminal locations, RIS location and size can be adjusted. It is demonstrated via extensive computer simulations that an improved achievable rate performance is obtained in the presence of RISs for both near-field and far-field conditions.

... 2) BWP of a building given its layout: The BWP evaluation of a building is demonstrated in a typical office building assuming A1 scenario of WINNER II channel model [14,Fig. 2.1]. ...

Over 80% of wireless traffic already takes place in buildings. Like water, gas, and electricity, wireless communication is becoming one of the most fundamental utilities of a building. It is well known that building structures have a significant impact on in-building wireless networks. If we seek to achieve the optimal network performance indoors, the buildings should be designed with the objective of maximizing wireless performance. So far, wireless performance has not yet been considered when designing a building. In this paper, we introduce a novel and interdisciplinary concept of building wireless performance (BWP) to a wide audience in both wireless communications and building design, emphasizing its broad impacts on wireless network development and deployment, and on building layout/material design. We first give an overview of the BWP evaluation framework proposed in our state-of-the-art works and explain their interconnections. Then, we outline the potential research directions in this exciting research area to encourage further interdisciplinary research.

... Only the LOS fading channel is considered in this case. • Rural case: the path loss and shadowing of the rural scenario in the WINNER II model is used [37]. The layout is similar to the urban case, but with wider street lanes (5 m) and a greater grid size (1000 m × 1000 m) as well as less occlusion from the buildings. ...

Yi Yuan
Gan Zheng
Kai-Kit Wong
Khaled B. Letaief

This paper studies the allocation of shared resources between vehicle-to-infrastructure (V2I) and vehicle-to-vehicle (V2V) links in vehicle-to-everything (V2X) communications. In existing algorithms, dynamic vehicular environments and quantization of continuous power become the bottlenecks for providing an effective and timely resource allocation policy. In this paper, we develop two algorithms to deal with these difficulties. First, we propose a deep reinforcement learning (DRL)-based resource allocation algorithm to improve the performance of both V2I and V2V links. Specifically, the algorithm uses deep Q-network (DQN) to solve the sub-band assignment and deep deterministic policy-gradient (DDPG) to solve the continuous power allocation problem. Second, we propose a meta-based DRL algorithm to enhance the fast adaptability of the resource allocation policy in the dynamic environment. Numerical results demonstrate that the proposed DRL-based algorithm can significantly improve the performance compared to the DQN-based algorithm that quantizes continuous power. In addition, the proposed meta-based DRL algorithm can achieve the required fast adaptation in the new environment with limited experiences.

... We employ the WINNER II line-of-sight pathloss model [33], which results in that the pathloss is uniformly distributed between −59.4 dB and −74.6 dB. All wireless channels are subject to Rayleigh fading. ...

In this paper the radio channel characteristics of the 8 × 4 MIMO system consisting of a base station and a small terminal equipped with multiple antennas for indoor-indoor and outdoor-indoor scenarios are presented. We study the large-scale variation and small-scale characteristics of the measured channel coefficients. Although the mean received power is very much dependent on the measured location, the channel capacity seems to be unchanged when the receiver's location is altered. The data collected from different scenarios (e.g. measurement locations, antenna setting) were used to investigate the advantage of having the knowledge of the channel at both ends of the transmission link. It is shown that using the water filling algorithm there is indeed an increase in the channel capacity. At low SNR, the benefit of knowing the channel at both link ends observed in the measurement data is much higher than which can be obtained in the channel matrix with usual assumption on identical independently distributed components. Using the small-scale and large-scale information in the formulation of the channel capacity we show that in our measurement, the variation of the mean received power has a greater influence on the change of the overall system performance than the change in the environmental multipath scattering property.

The attenuation by a curvilinear-topped obstacle and multiple flat-topped obstacles are solved in the present paper based on Fresnel-Kirchhoff theory. The results have clear physical meaning and are simple to use, it only needs to change the signs of j in Vogler's multiple knife-edge attenuation function to get each of the individual field, and add them up to get the total field. Meanwhile, the attenuation by a wedge, a flat-topped obstacle with bevel sides and double knife edges with ground reflection are also solved by the attenuation formula of multiple knife edges with ground reflection given in the paper.

Richard Rudd

ABSTRACT This report describes measurements,made to determine the statistics of building entry loss, for slant paths at frequencies between 1 and 6 GHz. The work described was funded under the S@TCOM programme,of the British National Space Centre (BNSC). Measurements of building penetration loss have been made at one office and three domestic sites in England. Tests were made at 1.3 GHz, 2.4 GHz and 5.7 GHz, and a tethered balloon was used to explore a range of elevation angles. A total of 12,450 spot measurements were made, at 11 receiver locations. The mean measured value of penetration loss, averaged over all frequencies and test locations, was 11.2 dB. The mean loss at the highest frequency was some 3.5 dB greater than that at the lowest frequency. Some dependence,was found on elevation angle at the two higher frequencies.

Jussi Ojala
Ralf Böhnke
Antti Lappeteläinen
Masahiro Uno

The IST project MIND [1] aims to ease the creation and provision of broadband services and applications that are fully supported and customised when accessed by users in the future from a wide range of wireless access technologies. As a part of that, techniques for the delivery of broadband services are evaluated. The paper presents pathloss and channel model for the Rooftop-to-Rooftop environment in the 5 GHz band. Based on the obtained models the paper addresses the performance and usability of the H/2 physical layer for providing fixed wireless broadband access in the Rooftop-to-Rooftop environment. ,,1752'8&7,21 The IST project MIND envisions interesting business scenarios based on rooftop wireless routers providing residential broadband access. The wireless routers would feature full IP stack; thus, they would create a mesh network topology similar of today's wired Internet. On the physical layer the routers utilise OFDM similar to H/2. This paper presents the Rooftop-to-Rooftop channel behaviour in the 5 GHz band, which is a crucial factor in the feasibility and performance analysis of the usage of H/2 PHYsical layer. The paper is organised as follows. In Chapter II the key business considerations are enlisted together with the respective impacts of the technical realisation of the rooftop wireless access. In Chapter III a comprehensive picture of the measurement set-up and, thus, the limits of the applicability of the pathloss and channel models are given. The main results, the Rooftop pathloss and channel models are derived and presented in Chapter IV. Link layer PER performance of the Rooftop channel model is analysed and compared to BRAN channel models in Chapter V.

Ashok Chandra
Ambuj Kumar
P. Chandra

Mobile communication systems are being developed operating both in outdoor and indoor environments. This calls for establishment of effective communications. IEEE-802.11b also provides wireless connectivity in both of these environments. We have made study in the scenario when a mobile user enters into an indoor environment from outside environment and establishes communication inside. This paper reports propagation measurements for two situations at 2000 MHz. In order to simulate, a source at 2000 MHz is used to illuminate the building from outside. Theoretical models have been used to calculate penetration loss, Rician factor 'K' and other channel parameters. It has been observed that when the transmitter illuminates the building from outside, the received signals inside the building attenuates severely. The received signal distribution follows d<sup>-m</sup> power law. Measurements reveal that the received signals are attenuated by approximately 40-45 dB in the corridors of different floors. The movements of people greatly vary amplitudes of the transmitted signal and make the indoor channel nonstationary even when either both the transmitter and receiver are stationed at a particular position or transmitter is fixed and receiver moves. This aspect has also been reported in this paper.

T. Rautiainen
J. Juntunen
Kimmo Kalliola

We present propagation analysis results for so called typical and bad urban macrocellular scenarios measured at 5.3 GHz carrier frequency and 100 MHz chip rate in Helsinki. Propagation characteristics between these scenarios have been compared, and small and large scale channel parameters have been extracted for stochastic geometry based channel models.

Kenya Yonezawa
Hiroyasu Ishikawa
Y. Takeuchi

This paper presents a frequency range extended path loss prediction formula based on the measurement results using multiple frequency bands from 0.8 to 8 GHz, where the BS antenna height is lower than the surrounding buildings. Although the lower and narrower frequency bands are assigned to the current cellular/mobile systems, the higher and wider frequency bands should be allocated to the future mobile systems for providing broadband multimedia wireless communications services. And the future mobile systems are assumed to be the multi-mode systems including the existing systems in the lower band. The proposed path loss prediction formula should be applicable to wider frequency range (from 0.8 to 8 GHz) and will be able to contribute to the cell design for the future mobile systems

To study the carrier frequency effects on path loss, measurements have been conducted at four discrete frequencies in the range 460-5100 MHz. The transmitter was placed on the roof of a 36 meters tall building and the receive antennas were placed on the roof of a van. Both urban and suburban areas were included in the measurement campaign. The results show that there is a frequency dependency, in addition to the well known free-space dependency 20 log<sub>10</sub>(f), in most of the areas included in the measurements. In non line of sight conditions, the excess path loss is clearly larger at the higher frequencies than at the lower. A model capturing these effects is presented

Source: https://www.researchgate.net/publication/234055761_WINNER_II_channel_models

Posted by: marlysmarlysantolinie0268191.blogspot.com

Mss Sp 72 Pdf Free Download

Post a Comment

Formulir Kontak