‘The research leading to these results was derived from the

European Community’s Seventh Framework Programme (FP7) under

Grant Agreement number 248454 (QoSMOS)’.

Quality Of Service and MObility driven cognitive radio Systems

Overview of the EU-project QoSMOS PMSE Workshop, DLR, Berlin

Michael Fitch

7th December 2011

Two trends are occurring:

1. Cells are becoming smaller...

2

Higher data

rates and

more users

Large signal

strength

Large

spectrum use

Smaller cells

(including self

-install)

Approx linear relationship

Need to re-use

spectrum over

shorter distances

Self-organising

networks

Planning is infeasible

Non-linear relationship

Impossible transmit

powers over long distances

Licensed

Unlicensed EG WiFi

Bluetooth

EG GSM, 3G

Shared EG TV White

Space

Flexible radios

Intelligent use

of spectrum

resources

Low power

Unplanned

Interference

High Power

Radio Planning

Spectrum shortage

Medium Power

Dynamic planning

My 5 year vision is

very flexible and

reconfigurable

user terminals.

And managed use

of spectrum.

2. Regulation is changing to allow spectrum sharing

- to enable more efficient use of the spectrum

Response to these trends….

• Spectrum sharing using Cognitive Radio

as the enabling technology

– Collaborative projects, such as QoSMOS and

Cambridge TVWS group,

– BT In-house trials and use-case assessment

– QoSMOS to provide technology and business

models

QoSMOS at a glance

• Quality of Service and MObility driven cognitive radio

Systems

• To develop critical technologies, value chain and

regulatory environment for opportunistic use of spectrum

• Is an FP7 Integrating Project

– Call 4 objective ICT-2009.1.1; The Network of the Future, part

(b): Spectrum-efficient radio access to Future Networks

– Duration is 36 months from January 2010

• Budget

– 1198 PMs

– Total = 14.5M€, EC contribution = 9.4M€

Date, slide number

Partners

7 December 2011, 6

Participant no. Participant organisation name Part. short

name

Country

1 (Coordinator) British Telecommunications PLC BT United Kingdom

2 Telenor ASA TEL Norway

3 Commissariat à l’Energie Atomique CEA France

4 Oulun Yliopisto UOULU Finland

5 Technische Universität Dresden TUD Germany

6 Instituto de Telecomunicões IT Portugal

7 NEC Technologies (UK) Ltd NTUK United Kingdom

8 Agilent Technologies Belgium NV AGILENT Belgium

9 Thales Communications SA TCF France

10 University of Surrey UNIS United Kingdom

11 NEC Corporation NEC Japan

12 Fraunhofer-Gesellschaft zur Förderung

der angewandten Forschung e.V.

Fraunhofer Germany

13 TST Sistemas SA TST Spain

14 Alcatel-Lucent Deutschland AG ALD Germany

15 Budapesti Műszaki és

Gazdaságtudományi Egyetem

BME Hungary

Objectives

• The main objective is to provide a platform for efficient

radio access to future networks

• Under this are two S & T objectives

– Cognitive Wireless Access Provision [measurable criteria]

• Platform aspects

• Intelligence aspects

– Network Support Provision [measurable criteria]

• And two non-S & T objectives

– Use-case development [guidelines on marketing]

– Preparation of regulatory policies [response of regulators]

7 December 2011, 7

7 December 2011, 8

Concept

An upper cognitive

manager that

manages the spectrum

portfolio

A lower cognitive

manager that

allocates resource

A significant novelty is a two-step

spectrum management process

Wanted outcomes

• to develop the critical technologies to allow spectrum sharing

• to establish confidence of regulators, primary and other secondary users that spectrum sharing can be achieved without causing harmful interference

• to provide a forum that encourages framework alignment

across Europe so that the market is big enough for

equipment that give a high user satisfaction at the right

price

• to give terminal deployment guidelines – antenna

spacing etc

• to give network deployment guidelines – database

integration etc

Role of the QoSMOS advisory

board Advisory board:

ANFR

BNetzA

RA-NL

AT4wireless

WinnF

BBC

Microsoft

NXP

Ofcom UK

Ofcom Swiss

SWR

ETSI RRS

Bosch

Steering and

deliverable reviews QoSMOS

Five meetings

aligned with project

milestones

Dissemination route

The advisory group remains open to other organisations

QoSMOS rationalised scenarios Scenario Range LoS Datarate Mobile Suitable

Frequency Dynamic backhaul 10 km Maybe High

(10–

50Mbit/s)

No >2GHz if LoS,

<1GHz if non-

Los

Cellular extension in White

Space

0.1 – 10 km No Med (2 –

10Mbit/s)

Yes >1GHz if <1km

Rural Broadband 1 – 10 km Maybe Med No >2GHz if LoS,

<1GHz if non-

Los

Cognitive ad hoc Network 1 – 1000 m No Med Yes >2GHz if

<50m

Direct Terminal-to-Terminal in

Cellular

10 – 1000 m No Low

(<2Mbit/s)

No >2GHz if

<50m

Cognitive femtocell 1 – 100 m No Med Maybe >2GHz if

<50m

Rationalisation was carried out through questionnaires to Stakeholders

in the value chain and includes technical and Commercial feasibility.

The final column on the right is my addition

Terminal for sub-system1

RAN of sub-system1

Spectrum Management

(WP6)

Sensing

WP3 Transceiver

WP4

BTS#1

User Plane

Protocol

Stack

(L1-L5)

Cognitive manager

(WP5)

Spectrum

Mngt Sensing Transceiver

WP6 WP

3

WP4

BTS#2 BTS#2

User Plane

Protocol

Stack

Cognitive manager

(WP5)

Spectrum

Mngt Sensing Transceiver

WP6 WP

3

WP4

User Plane

Protocol Stack

L1-L2

Country-wide regulation constraints database

User Plane

Protocol Stack L3

Spectrum Management

(WP6)

To other

Spectrum

Manager(s)

Core Network

of QoSMOS system

To other

terminals To other

terminals

RAN of sub-system2

Terminal for sub-system2

Cognitive control flow

(logical I/F)

Enablers

Cog. Mng

WP5

Internal I/F

Spectrum Management

(WP6)

Sensing

WP3 Transceiver

WP4

User Plane

Protocol Stack

L1-L2

Cog. Mng

WP5

Proof of Concepts (demos)

Radio enviroment and Sensing engine

WP3

Cooperative Sensing(Data fusion & distributed

algorithms)WP3, 4

Flexible Transceiver Architecture

WP4

CM-SMWP6

CM-RMWP5

PoC #5

PoC #4

PoC #1 PoC #2

PoC #3

Adaptation Layer

HW Platforms

Three examples of QoSMOS

outcomes

• Wireless microphone detection

• Filter block modulation modelling

– Prototype being built for demo in December in

March 2012

• TVWS availability database

7 December 2011, 14

Wireless microphone detection Spectrum generated by Matlab model to EN 300 422

• Teager Kaiser detector – Use the knowledge and the characteristics of FM modulation

Teager-Kaiser energy detector [Crowncom2010]

Y[s(k)]=[s(k)]2-s(k+1)s(k-1)

• Frequency domain detection – Narrowband condition:

Detection in the frequency domain

Efficient implementation based on FFT for the energy

detection

Wireless microphone detection

Filter bank transceiver

Subchannel frequency responses

= Frequency-shifted versions of a prototype

FBMC transceiver

Due to the overlapping of neighbouring sub-channels, orthogonality is needed.

Use of Offset-QAM model:

- Each QAM symbol is mapped to two consecutive subcarrier samples.

- Subcarrier sample sequences are oversampled by a factor of 2.

Spectral Properties of LTE and FBMC

• Example 5 MHz Bandwidth

– Requires shaping filter to meet ACLR specifications of LTE

Spectral Properties of LTE • Example 5 MHz Bandwidth

– Requires shaping filter to meet ACLR specifications of LTE

FCC TVWS mask

Ofcom proposed database

structure

Prototype TVWS availability database

Rural Scotland

London

Ipswich

• TVWS database provides – channels

– power levels

– duration of access

– directivity for any location UK

An exploitation from QoSMOS

- BT’s trials of broadband to ‘not-spots’

If D + E are too long (>5km) then

BB is not possible. Fibre to the Cabinet

will help only if D is not too long.

In many rural locations, premises are

connected via long copper lines

directly to the exchange. If this is

>5km then BB is not possible

Backhaul

Exchange House

15% of premises in the UK (2.75 million) cannot

receive broadband (2Mbit/s downlink). These are

called ‘not-spots’.

DSL bit-rates against distance (downlink)

0

10

20

30

40

50

60

70

80

90

100

0 1000 2000 3000 4000 5000

Distance (m)

Lin

e b

it-r

ate

(M

bit

/s)

ADSL

ADSL2+

VDSL2+

VDSL2 max rate = 200Mbit/s

ADSL uplink approx 450kbit/s (capped)

VDSL is approximately symmetrical

Little difference between them for > 4km line length

Distribution of ‘notspots’ direct distance

to nearest exchange

The average distance is

3km, which means the fixed

lines to many hotspots do not

take the shortest route.

Using TVWS, 2Mbit/s can be

delivered at 5km Non-LoS

and 8km LoS

>2Mbit/s Non-LoS >2Mbit/s LoS

So TVWS can potentially

solve the access problem

for a large proportion of not-spots

…correlate well with the areas of

Great Britain where TV White Space is

most available.

Example TV

Whitespace availability

The more rural areas of Great Britain, where

there are higher levels of poor broadband

performance due to line length….

Percentage of bad lines due to length

0 to 1%

1% to 2%

2% to 3%

3% to 5%

5% to 9%

9% to 11.4%

White = maximum

Red = minimum

TV whitespace spectrum versus notspot locations

TVWS CPE transceiver

to Ethernet

TVWS BS

Router

DSLAM

Backhaul

Ethernet

Up to 5km non line-of-sight

8km line-of-sight

The basic idea

To share wireless spectrum with

Digital TV Transmitters. This is

UHF between 470 – 790MHz

- Low diffraction and building penetration loss

Trial on the Isle of Bute with base-station at Kilchattan Bay

‘Final point’ is the House of the Marquis

We are trying it on the Isle of Bute

Base-station antenna at Kilchatten Bay

exchange. Antenna up to 12m height

Backhaul antenna facing mainland

CPE antenna – same as TV antenna

Height 4m. At the House of the Marquis.

Test site on the Isle of Bute

Path height profile

Hill tops are smooth

4Mbit/s downlink and

2Mbit/s uplink was achieved

on this 5km non line-of-sight

path

Ongoing work to understand UHF propagation over hills

18GHz backhaul link, 12km

West Kilbride to Kilchatten Bay exchange

Improving the backhaul connection to the base-station

incurs additional costs which we are evaluating

Conclusions and further work

• Cognitive Radio is an enabling technology for spectrum sharing and

small cells

• QoSMOS is developing key technologies to support spectrum

sharing with managed QoS and mobility

– Two examples are FBMC and database

• PMSE sharing in UK is via database

– Sensing is not sufficiently reliable

• An early exploitation is rural broadband

– Being trialled by BT

• Further work

– Scope for improving spectrum packing of wireless microphones ?

– PMSE going digital ?

7 December 2011, 32