SPECTRUM MANAGEMENT IN COGNITIVE RADIO NETWORKS

I. F. AKYILDIZ
Broadband Wireless Networking Laboratory School of Electrical and Computer Engineering Georgia Institute of Technology Tel: 404404-894894-5141; Fax: 404404-894894-7883 Email: ian@ece.gatech.edu Web: http://www.ece.gatech.edu/research/labs/bwn

I.F. Akyildiz, W. Y. Lee, M.C. Vuran and S. Mohanty, NeXt Generation/Dynamic Spectrum Access/Cognitive Radio Wireless Networks: A Survey, Survey, Computer Networks (Elsevier) Journal, Sept. 2006. I. F. Akyildiz, W.Y. Lee, M. C. Vuran, and S. Mohanty,
"A Survey on Spectrum Management in Cognitive Radio Networks, Networks,

IEEE Communications Magazine, April 2008.

IFA2008

AVIGNON

FIXED SPECTRUM ASSIGNMENT

IFA2008

AVIGNON

Fixed Spectrum Utilization

Maximum Amplitudes
Heavy Use

Amplitude (dBm)

Heavy Use

Sparse Use

Medium Use

Frequency (MHz)
IFA2008

AVIGNON

Fixed Spectrum Utilization

PSD (Power Spectrum Density)

GHz
Freq (GHz) Utilization(%) 0~1 54.4 1~2 35.1 2~3 7.6 3~4 0.25 4~5 0.128 5~6 4.6

Measurements show that there is wide range of spectrum utilizations across 6 GHz of spectrum

IFA2008

AVIGNON

COGNITIVE RADIO NETWORKS; DYNAMIC SPECTRUM ALLOCATION NETWORKS (DSANs);

xG INITIATIVE

Dynamic Spectrum Allocation

IFA2008

AVIGNON

COGNITIVE RADIO
A Cognitive Radio is the key enabling technology for Dynamic Spectrum Access!! Capability to use or share the spectrum in an opportunistic manner -> SPECTRUM HARVESTING Dynamic spectrum access techniques allow the CR to operate in the best available channel
IFA2008

AVIGNON

Ultimate Objective of Cognitive Radio

CR enables the usage of temporally unused spectrum Spectrum Hole or White Space. If this band will be used by a licensed user, CR moves to another spectrum hole Alters its transmission power level or modulation scheme to avoid interference
IFA2008

AVIGNON

Spectrum Hole Concept

Power
Frequency Spectrum Holes

Spectrum Occupied by Licensed Users

Time

IFA2008

AVIGNON

FEATURES OF COGNITIVE RADIO

Senses RF Environment and modifies frequency, power or modulation Dynamic Spectrum Selection Adaptive Modulation Adaptive Power Control Real Time Spectrum Management Significantly Increases Spectrum Efficiency
IFA2008

AVIGNON

10

OVERALL VIEW
tive p a Ad ocol t Pro te Sui
WiMAX WiMAX
Applications TCP/UDP IP MAC Error Control Physical

Dynamic Spectrum Allocation

Cellular (2,3,4-G) Mesh Networks

Sensor Networks

Wireless LANs & MANs

IFA2008

AVIGNON

11

COGNITIVE RADIO NETWORK COMMUNICATION FUNCTIONALITIES

Application Control

Application
Handoff Delay, Loss

QoS Requirements

Transport

Reconfiguration Routing Information/ Reconfiguration Scheduling Information/ Reconfiguration

Spectrum Mobility Function Link Layer Spectrum Delay Sharing

Routing Information

Network Layer

Spectrum Decision Function

Link Layer Spectrum Sensing

Sensing Information

Physical Layer

Sensing Information / Reconfiguration

Handoff Decision, Current and Candidate Spectrum Information

IFA2008

AVIGNON

12

WHAT IS A COGNITIVE RADIO?

A Cognitive Radio is a radio that can change its transmitter parameters based on interaction with the environment in which it operates. (Federal Com Commission2005) FCC (Non-Federal Use of the Spectrum)
IFA2008

AVIGNON

13

Microsoft, Google, Dell, HP

KNOWS: Kognitiv Networking Over White Spaces
http://research.microsoft.com/netres/projects/KNOWS/

Prototypes designed to identify wireless microphone, NTSC and Digital TV signals Undergoing testing at FCC
IFA2008

AVIGNON

14

Adaptrums Testbed
http://www.adaptrum.com/home

Experiments undertaken in the region of 500 MHz 700 MHz Undergoing testing at FCC

IFA2008

AVIGNON

15

Georgia Tech: OCRA Testbed

http://www.ece.gatech.edu/research/labs/mag/cognitive_radio.htm

IFA2008

AVIGNON

16

Physical Architecture of the Cognitive Radio (Cognitive Radio Transceiver)

Receiver

Modem
Radio Frequency (RF) Analog-to-Digital Convertion (A/D)

Processor
Data Processing

Baseband Processing

Transmit

Control Parametrization (SDR) (Reconfiguration)

IFA2008

AVIGNON

Physical Architecture of the Cognitive Radio (Wideband RF/Analog FRONTFRONT-END Architecture)

Wideband Antenna
Low Noise Amplifier LNA
Channel Selection Filter

Analog-to-Digital Converter

Mixer

AGC
Automatic Gain Control

A/D

RF Filter

Voltage Controlled VCO Oscillator

PLL

IFA2008

AVIGNON

From user

To user

RF Front-End (Tx/Rx)

Cognitive Radio Network Architecture

Spectrum Band Unlicensed Band
Primary Network Access
CR User

Spectrum Broker

Licensed Band I
Primary User Primary Base-station CR Base-station Other Cognitive Radio Networks

Licensed Band II

CR Network Access

Primary Network

Cognitive Radio Network (With Infrastructure)

IFA2008

AVIGNON

CR Ad Hoc Networks Architecture

Spectrum Band
Unlicensed Band CR User Licensed Band I
Primary User Primary Base-station

Licensed Band II

CR Ad Hoc Access

Primary Network

IFA2008

Cognitive Radio Network (Without Infrastructure) AVIGNON

20

Cognitive Radio Network Architecture

Spectrum Band

Unlicensed Band
Primary Network Access

CR User

Spectrum Broker

Licensed Band I
Primary User Primary Base-station

Licensed Band II

CR Ad Hoc Access

CR Base-station

Other Cognitive Radio Networks

CR Network Access

Primary Network

Cognitive Radio Network (Without Infrastructure)

Cognitive Radio Network (With Infrastructure)

21

IFA2008

AVIGNON

Primary Network
An existing network infrastructure which has an access right to a certain spectrum band.
e.g., Common cellular systems and TV broadcast networks.

IFA2008

AVIGNON

22

Primary User
(or Licensed User)
Has a license to operate in a certain spectrum band. REMARK: Primary users do not need any modification or additional functions for co-existence with CR users

IFA2008

AVIGNON

23

Cognitive Radio Network

(or Dynamic Spectrum Access Network, or Secondary Network or Unlicensed Network)

* Does not have license to operate in a desired band.

* Hence, the spectrum access is allowed only in an opportunistic manner !

IFA2008

AVIGNON

24

Cognitive Radio User

(or Unlicensed User, Secondary User)

has no spectrum license

Hence, additional functionalities are required to share the licensed spectrum band !

IFA2008

AVIGNON

25

SPECTRUM MANAGEMENT FRAMEWORK

1) Determine which portions of the spectrum is available and detect the presence of licensed users when a user operates in a licensed band

(Spectrum Sensing)
2) Select the best available channel (Spectrum

Decision)

3) Coordinate access to this channel with other users (Spectrum Sharing) 4) Vacate the channel when a licensed user is detected (Spectrum Mobility Spectrum Handoff)
IFA2008

AVIGNON

26

COGNITIVE CYCLE
Transmitted Signal

Radio Environment
RF Stimuli

Spectrum Mobility

Primary User Detection Decision Request

Spectrum Sensing
Spectrum Hole

Spectrum (Channel) Characterization

Spectrum Sharing

Channel Capacity

Spectrum Decision

IFA2008

AVIGNON

27

What is Spectrum Sensing ?

How to detect spectrum holes by the COGNITIVE RADIO so that it can adapt itself to its environment !!

IFA2008

AVIGNON

28

Classification of Spectrum Sensing Techniques

Spectrum Sensing
Transmitter Detection Receiver Detection Interference Temperature Management

Matched Filter Detection

Energy Detection

Cyclostationary Feature Detection

IFA2008

AVIGNON

29

Limitations of the Transmitter Detection

Receiver Uncertainty Problem
Interference
CR Transmitter Range

Shadowing Problem
Hidden Terminal Problem due to Shadowing

Primary Transmitter Range

Primary Base-station

CR User Interference CR User

Cannot detect the transmitter Primary Transmitter Range CR Transmitter Range

Primary User

Interference due to uncertainty of receiver location

Primary Base-station

Primary User

Cannot detect the transmitter

IFA2008

AVIGNON

30

Transmitter Detection
Non-Cooperative vs Cooperative Detection
Detection Method Transmitter Detection Detection Behavior Transmitter Detection

Matched Filter Detection

Energy Detection

Cyclostationary Feature Detection

Non-Cooperative Detection

Cooperative Detection

IFA2008

AVIGNON

31

Cooperative Detection
Detect the primary user correctly
Primary Base-station CR User 3 BUSY

Shadowing Cannot detect the primary user due to the obstacles

Multi-path fading
Weak signals are received due to the multi-path fading may not detect the primary user CR User 2 IDLE Primary User

CR User 1 IDLE

IFA2008

AVIGNON

By exchanging their sensing information, CR users can detect the primary user under fading and shadowing environments
32

Primary Receiver Detection

Spectrum Sensing
Transmitter Detection

Receiver Detection
Cyclostationary Feature Detection

Interference Temperature Management

Matched Filter Detection

Energy Detection

IFA2008

AVIGNON

33

Primary Receiver Detection

B. Wild and K. Ramchandran, Detecting Primary Receivers for Cognitive Radio Applications in Proc. IEEE DySPAN, pp. 124-130, Nov. 2005.

Primary Base-station

Local Oscillator (LO) Leakage Power

CR users detect the LO leakage power for the detection of primary users instead of the transmitted signals

When primary users receive the signals from the transmitter, they emit the LO leakage power.

CR User

Primary User

IFA2008

AVIGNON

34

Interference Temperature Management

Spectrum Sensing
Transmitter Detection Receiver Detection

Interference Temperature Management

Matched Filter Detection

Energy Detection

Cyclostationary Feature Detection

IFA2008

AVIGNON

35

Interference Temperature Model

o

Licensed Signal
New Opportunities for Spectrum Access

Minimum Service Range with Interference Cap

Service Range at Original Noise Floor

Power Interference at Temperature Limit Receiver

Original Noise Floor

Distance from Licensed Transmitting Antenna
IFA2008

AVIGNON

36

Open Research for Spectrum Sensing

Sensing Parameter Optimization (Based on Transmitter / Energy Detection )

W. Y. Lee and I. F. Akyildiz, Akyildiz, Optimal Spectrum Sensing Framework for Cognitive Radio Networks, Networks,

IEEE Trans. on Wireless Communications, Oct. 2008. 2008.

IFA2008

AVIGNON

Open Research for Spectrum Sensing

Optimization of Cooperative Sensing: Determine the # of cooperating users and spectrum bands Cooperative Feature (Cyclostationary) Detection Scheme: For faster detection time and less complexity Hybrid Transmitter & Receiver Detection: Based on the activities both in uplink and downlink channels RAT (Radio Access Technology)-Adaptive Spectrum Sensing: Consider radio access technologies (TDMA, CDMA, FDMA, etc ) of primary networks
IFA2008

AVIGNON

Spectrum Decision
Transmitted Signal

Radio Environment
RF Stimuli

Spectrum Mobility

Primary User Detection Decision Request

Spectrum Sensing
Spectrum Hole

Spectrum (Channel) Characterization

Spectrum Sharing

Channel Capacity

Spectrum Decision

IFA2008

AVIGNON

39

Spectrum Decision
Stage 1 Spectrum Characterization
RF information
Interference Path

Stage 2 Decision

Primary Network Information

Loss Wireless Link Error Link layer delay

Primary User Activity Holding Time

Single Spectrum Decision

MultiSpectrum Decision

IFA2008

AVIGNON

40

SINGLE SPECTRUM DECISION

Each CR user selects only one spectrum band according to the application requirements
Occupied by primary users Idle spectrum band

Frequency(Hz)

CR user A
CR user A CR user B

CR user B Spectrum Handoff

IFA2008

AVIGNON

41

Multi-Spectrum Decision
CR users select multiple non-contiguous spectrum bands and use them simultaneously for the transmission.
Occupied by primary users Idle spectrum band Frequency(Hz)

Sub-channels for CR user A CR user A CR user B

Sub-channels for CR user B

Spectrum Handoff

IFA2008

AVIGNON

42

Open Research for Spectrum Decision

Spectrum Decision and Admission Control for Real-time / Best Effort Applications
W. Y. Lee and I. F. Akyildiz, Akyildiz,
QoS Aware Spectrum Decision for Cognitive Radio Networks, Networks,

Submitted for publication, publication, July 008.

IFA2008

AVIGNON

Open Research for Spectrum Decision

Primary Network Modeling and Analysis Primary User Behavior Predictivity Models CR Channel Characterization / Estimation: In terms of QoS parameters (throughput, PER, delay, jitter, etc)
IFA2008

AVIGNON

Open Research for Spectrum Decision

Spectrum Decision with Reconfiguration: Determine not only spectrum bands (operating frequency, bandwidth) but also transmission parameters (transmission power, modulation, channel coding, upper layer protocols, etc)

IFA2008

AVIGNON

Spectrum Sharing
Transmitted Signal

Radio Environment
RF Stimuli

Spectrum Mobility

Primary User Detection Decision Request

Spectrum Sensing
Spectrum Hole

Spectrum (Channel) Characterization

Spectrum Sharing

Channel Capacity

Spectrum Decision

IFA2008

AVIGNON

46

Spectrum Sharing
Spectrum Sharing similar to MAC Problems Multiple CR users try to access the spectrum Access must be coordinated (to prevent collisions in overlapping portions of the spectrum) Uniqueness Coexistence with licensed (primary) users Wide range of available spectrum
IFA2008

AVIGNON

47

SPECTRUM SHARING CLASSIFICATION

Intra-Network SS

Inter-Network SS

IFA2008

AVIGNON

48

Intra-Network Spectrum Sharing

Centralized Spectrum Sharing

IFA2008

AVIGNON

49

Intra-Network Spectrum Sharing

Spectrum sharing entity Sending local observations Sending spectrum allocations Spectrum sharing entity

Distributed Spectrum Sharing (Non-Cooperative)

Distributed Spectrum Sharing (Cooperative)

IFA2008

AVIGNON

50

Inter-Network Spectrum Sharing

Spectrum Broker (or Spectrum Server)
CR Network B CR Network B CR Network A Sending Local Observations Sending Spectrum Allocations Spectrum Sharing Entity

CR Network C

CR Network A CR Network C

Centralized Spectrum Sharing

Distributed Spectrum Sharing

IFA2008

AVIGNON

51

Why Game Theory?

[Spectrum Sharing in CR networks]
CR users have a common interest to have the spectrum resources as much as possible. However, CR users have competing interests to maximize their own share of the spectrum resources. i.e., the activity of one CR user can impact the activities of the others Also CR users rational decisions require anticipating rivals responses
IFA2008

AVIGNON

52

Why Game Theory?

Provides an efficient distributed spectrum sharing scheme. Provides the well-defined equilibrium criteria for the spectrum sharing problem to measure the optimality in various network scenarios.

IFA2008

AVIGNON

53

Intranetwork Spectrum Sharing

Local Bargaining
(Cao/Zheng, IEEE SECON 2005)

Belief Assisted Pricing

(J. Zhu and Ray Liu IEEE Com. Magazine, May 2007)

Repeated Spectrum Sharing Game Model

(R. Etkin, A. Parekh, and D. Tse, IEEE DySPAN 2005)

Auction Based Spectrum Sharing Game

(J. Huang, R. Berry, and M. L. Honig, ACM Monet Journal, 2006)

IFA2008

AVIGNON

Internetwork Spectrum Sharing

O, Ileri, D. Samardzija, and N. Mandayam IEEE DySPAN 2007

S. Gandhi, C. Buragohain L. Cao, H. Zheng, S. Suri

IEEE DySPAN 2007 M. Bennis and J. Lilleberg Proc. CISS 2007 X. Jing, D. Raychaudhuri IEEE DySPAN 2005

IFA2008

AVIGNON

Open Research for Spectrum Sharing Inter-Cell Spectrum Sharing based on Joint Spectrum and Power Allocation
W. Y. Lee and I.F. Akyildiz, Akyildiz, ``Joint Spectrum and Power Allocation for InterInter-Cell Spectrum Sharing in Cognitive Radio Networks,' Networks,' in Proc. IEEE DySPAN 2008, Chicago, IL, USA, October 2008.
IFA2008

AVIGNON

Open Research for Spectrum Sharing

Inter-User (Intra-Cell) Spectrum Sharing: RAT adaptive-CR MAC Protocol / User power control Inter-Network Spectrum Sharing: Spectrum Auctions among CR operators

IFA2008

AVIGNON

Spectrum Mobility
Transmitted Signal

Radio Environment
RF Stimuli

Spectrum Mobility

Primary User Detection Decision Request

Spectrum Sensing
Spectrum Hole

Spectrum (Channel) Characterization

Spectrum Sharing

Channel Capacity

Spectrum Decision

IFA2008

AVIGNON

58

Spectrum Mobility/Handoff
Get the Best Available Channel Channel concept
CR radio can capture the best available channel Spectrum Mobility/ obility/Handoff (a new type of handoff) handoff) is defined as the process when a CR user (SU) changes its frequency of operation When channel conditions become worse or a primary user (PU) appears, appears, spectrum handoff happens
IFA2008

AVIGNON

59

Spectrum Handoff Example

Power
Spectrum Handoff
Spectrum Hole

Frequency

Spectrum handoff is a multi-step process, it includes: Primary User Detection (Sensing issue) Channel Switching (Hardware issue) Resume Communication Time (Sharing issue)

Spectrum occupied Sensing issue Sharing issue by Licensed users

IFA2008

AVIGNON

60

Open Research for Spectrum Mobility

Unified User Handoff Scheme for Spectrum and User Mobilities (Layer-2 Mobility)

W. Y. Lee and I. F. Akyildiz, Akyildiz, "Spectrum"Spectrum-Aware Mobility Management in Cognitive Radio Cellular Networks Networks, submitted for publication, Oct. 2008.
IFA2008

AVIGNON

Open Research for Spectrum Mobility

Multi-Layer Mobility Management (Network / Transport / APP): Enable upper layer protocols to adapt to spectrum mobility. Base-Station / Cell Handoff: Provide Seamless Spectrum Transition of BaseStations, cooperating with inter-cell spectrum sharing
IFA2008

AVIGNON

MORE RESEARCH CHALLENGES

COMMON CONTROL CHANNEL PROBLEM
- Clustering

- Predefined in licensed band

- Reserved in licensed band as Coordinated Access Band (CAB)

- Located at the lower edge of unlicensed band with Common Spectrum Coordination Channel (CSCC) (CSCC) Protocol
IFA2008

AVIGNON

MORE RESEARCH CHALLENGES

Wideband RF front-end: Support Fast sensing/switching /reconfiguration Admission Control Framework: Cooperate with spectrum decision, spectrum mobility and inter-cell spectrum sharing

IFA2008

AVIGNON

MORE RESEARCH CHALLENGES

CR QoS Framework: Enable QoS guaranteeing communications for various service classes: conversational, streaming, interactive, and background. Primary User Activity Modeling / Estimation: Based on the RAT of primary networks CR Transmission Scheme with Error Control: Adaptive to primary user activity as well as channel condition
IFA2008

AVIGNON

MORE RESEARCH CHALLENGES

CR Security: Jamming, PU emulating attack, etc. PHY-Adaptive Spectrum Management Framework (Sensing / Decision / Sharing /Mobility): OFDM, MIMO, Directional Antenna, UWB, etc.
IFA2008

AVIGNON

MORE RESEARCH CHALLENGES

CR Deployment Scenarios: New network architecture or applications based on CR technology, Cooperation model between PU and SUs CR Network Management: O&M (Operation and Maintenance), Billing
IFA2008

AVIGNON

Proposal for a CR Network Management Framework

Billing
Security Management

Decision/ Policy Management

SLA Management (Inter/Intra) Service Adaptation QoS Management

System Monitoring

Intra-operator Resource Manager Inter-operator Resource Manager

Intra-operator Mobility Manager Inter-operator Mobility Manager

Reconfiguration

IFA2008

AVIGNON

Spectrum Management Mobility Management

68

Cognitive Radio Network Standards

IEEE 802.22 (Wireless

Regional Area Networks)

(Cognitive Radio for TV Bands/Channels) (2004)

IEEE P1900 (2005)

IFA2008

AVIGNON

69

What is a Nano-Network?
I.F. Akyildiz, F. Brunetti, C. Blazquez, Nanonetworking: A New Communication Paradigm Paradigm Computer Networks, Elsevier, Sept. 2008.

Interconnection of Multiple Devices at NanoNano-scale

Two Research Branches
Architectural Point of View: Electronic integration at nanonano-scale Examples: Carbon nanotubes and networksnetworks-ononchip (NoC) Communication Point of View: NanoNano-scale communication including transmitter and receiver devices, information, carrier and medium. Examples: Molecular Communication

IFA2008

AVIGNON

70

Nano-Machine Development

IFA2008

AVIGNON

71

NanoNano-machine Development: TopTop-Down Approach

Development of nano-scale objects by downscaling current existing microscale device components. Fabrication and assembly processes are at an early stage Only simple mechanical structures can be created

IFA2008

AVIGNON

72

NanoNano-machine Development: BottomBottom-up Approach

Nanomachines are developed using individual molecules as building blocks. Molecular manufacturing technologies able to assemble nanomachines molecule by molecule do not exist.

IFA2008

AVIGNON

73

NanoNano-machine Development: BioBio-hybrid Approach

Biological nano-machines found in nature are used as models to develop new nano-machines. They can also be used as building blocks integrating them into more complex nano-machines. Several biological structures found in living organisms can be considered as nano-machines. They can be found in biological systems where they are interconnected to perform more complex tasks.

IFA2008

AVIGNON

74

Communication Mechanisms for NanoNetworks?

Electromagnetic Waves

Electromagnetic communication could be used to transmit information from a micromicro-device to a nanonano-machine, but not among nanonano-machines With the current technology, acoustic communication cannot be used in nanonetworks

Acoustic

Communication techniques

Communication

NanoNano-mechanical Communication

Nanomechanical communication is currently not a solution for nanonetworks

Molecular

Molecular communication can be used to interconnect

IFA2008

Communication

AVIGNON

nanonano-machines, resulting in nanonetworks! nanonetworks!

75

A Possible Solution: Molecular Communication

Molecular communication is defined as the transmission and reception of information encoded in molecules

It is a completely new and interdisciplinary field that spans nano, bio and information technologies
IFA2008

AVIGNON

76

NanoNetworks vs Traditional Communication Networks

Features
Carrier: Signal type: Propagation speed: Medium conditions:

Traditional
Electromagnetic waves Electronic, optical, mechanical Sound or light Wired: almost immune Wireless: affect communication Electromagnetic field and signals High energy consumption

Molecular
Molecules Chemical Extremely low Affect communication

Noise: Other features:

Particles and molecules in medium Low energy consumption

77

IFA2008

AVIGNON

Nanonetworks Components

Traditional communication

Molecular communication
IFA2008

AVIGNON

78

Short-Range Communication

Molecular Motors Two possible communication systems

Calcium Ions
IFA2008

AVIGNON

79

Long-Range Communication using Pheromones

Features:

Communication Range

mm - km
Wet and dry Pheromones

Medium Carrier
IFA2008

AVIGNON

80

Research Challenges in Nanonetworks

1. Development of nanonano-machines, testbeds and simulation tools IFA2008

2. Theoretical

3. Architectures and Protocols

AVIGNON

81