networking cognitive radioscrtwireless.com/files/cognitiveradionetworks.pdfadapted from mitola,...

1

Cognitive Radio Technologies, 2007

1Cognitive RadioTechnologiesCCognitiveognitive RRadioadioTTechnologiesechnologies

CRTCRTCRTCRTCRTCRTCRTCRTCRTCRTCRTCRT

James Neel

[email protected]

(540) 230-6012www.crtwireless.com

MPRG Symposium

Session D-2

June 8, 2007

Networking Cognitive Radios

These Slides Available Online:http://www.crtwireless.com/Publications.html


2

James Neel• President, Cognitive Radio Technologies,

LLC

• PhD, Virginia Tech 2006– http://scholar.lib.vt.edu/theses/available/etd

-12082006-141855/

• Textbook chapters on:– Cognitive Network Analysis in

– Data Converters in Software Radio: A Modern Approach to Radio Engineering

– SDR Case Studies in Software Radio: A Modern Approach to Radio Engineering

– UWB Simulation Methodologies in An Introduction to Ultra Wideband Communication Systems

• SDR Forum Paper Awards for 2002, 2004 papers on analyzing/designing cognitive radio networks

• Email: [email protected] Cognitive RadioTechnologiesCCognitiveognitive RRadioadioTTechnologiesechnologies


2


3

Presentation Overview

• (32) Overview of Cognitive Radio

• (20) Interactive Decision Problem

• (44) A “Quick” Review of Game Theory

• (62) Designing Cognitive Radio Networks

• (26) Examples of Networked Cognitive

Radios

• (20) Future Directions in Cognitive Radio


4


• Overview of Cognitive Radio

• Interactive Decision Problem

• A “Quick” Review of Game Theory

• Designing Cognitive Radio Networks

• Examples of Networked Cognitive Radios

• Future Directions in Cognitive Radio

3


5

Overview of Cognitive Radio

Concepts, Definitions,

Implementations


6

Cognitive Radio: Basic Idea• Software radios permit network or

user to control the operation of a software radio

• Cognitive radios enhance the control process by adding– Intelligent, autonomous control of the radio

– An ability to sense the environment

– Goal driven operation

– Processes for learning about environmental parameters

– Awareness of its environment• Signals

• Channels

– Awareness of capabilities of the radio

– An ability to negotiate waveforms with other radios

Board package

(RF, processors)

Board APIs

OS

Software Arch

Services

Waveform Software

Co

ntr

ol

Pla

ne

4


7

•

No

inte

rfere

nce

•

Ne

go

tiate

Wave

form

s

•

•

•

“Aw

are

”

Ca

pa

bilitie

s

•

•

•

•

Le

arn

the

En

viro

nm

en

t

•

•

•

•

•

•G

oa

l Driv

en

•

•

•

•

•

•

•

“Aw

are

”

En

viro

nm

en

t

•

•

•

•

•

•

•

•

•

Re

ce

ive

r

•

•

•

•

•

•

•

•

•

•

Tra

nsm

itter

•

•

•

•

•

•

•

•

•

•

Ca

n s

en

se

En

viro

nm

en

t

••IEEE 1900.1

••IEEE USA

••NTIA

••Haykin

••ITU-R

••SDRF SIG

••VT CRWG

••SDRF CRWG

••Mitola

••FCC

Au

ton

om

ou

s

Ad

ap

ts

(Inte

llige

ntly

)Definer

Cognitive Radio Capability Matrix


8

Cognitive Radio Applications

5


9

Why So Many Definitions?• People want cognitive radio to be something

completely different– Wary of setting the hype bar too low

– Cognitive radio evolves existing capabilities

– Like software radio, benefit comes from the paradigm shift in designing radios

• Focus lost on implementation– Wary of setting the hype bar too high

– Cognitive is a very value-laden term in the AI community

– Will the radio be conscious?

• Too much focus on applications– Core capability: radio adapts in response changing operating

conditions based on observations and/or experience

– Conceptually, cognitive radio is a magic box


10

Used cognitive radio definition

• A cognitive radio is a radio whose control processes

permit the radio to leverage situational knowledge and intelligent processing to autonomously adapt

towards some goal.

• Intelligence as defined by [American Heritage_00] as

“The capacity to acquire and apply knowledge,

especially toward a purposeful goal.”

– To eliminate some of the mess, I would love to just call

cognitive radio, “intelligent” radio, i.e.,

– a radio with the capacity to acquire and apply knowledge

especially toward a purposeful goal

6


11

Overview of Implementation Approaches

How does the

radio become

cognitive?


12

Proposes and Negotiates New ProtocolsAdapts Protocols8

Generates New GoalsAdapts Plans7

Autonomously Determines Structure of

EnvironmentLearns Environment6

Settle on a Plan with Another RadioConducts Negotiations5

Analyze Situation (Level 2& 3) to Determine Goals (QoS, power), Follows Prescribed Plans

Capable of Planning4

Knowledge of Radio and Network Components, Environment Models

Radio Aware3

Knowledge of What the User is Trying to DoContext Awareness2

Chooses Waveform According to Goal. Requires Environment Awareness.

Goal Driven1

A software radioPre-programmed0

CommentsCapabilityLevel

Adapted From Table 4-1Mitola, “Cognitive Radio: An Integrated Agent Architecture for Software Defined Radio,” PhD Dissertation

Royal Institute of Technology, Sweden, May 2000.

Levels of Cognitive Radio Functionality

7


13

NormalUrgent

Level0 SDR1 Goal Driven2 Context Aware3 Radio Aware4 Planning5 Negotiating6 Learns Environment7 Adapts Plans8 Adapts Protocols

Allocate Resources

Initiate Processes

Orient

Infer from Context

Parse Stimuli

Pre-processSelect Alternate

Goals

Establish Priority

Plan

Normal

Negotiate

Immediate

LearnNew

States

Negotiate Protocols

Generate Alternate

Goals

Adapted From Mitola, “Cognitive Radio for Flexible Mobile Multimedia Communications ”, IEEE Mobile Multimedia Conference, 1999, pp 3-10.

Observe

Outside

World

Decide

Act

User Driven

(Buttons)Autonomous Determine “Best”

Plan

Infer from Radio Model

States

Determine “Best”

Known WaveformGenerate “Best”

Waveform

Cognition Cycle


14

OODA Loop: (continuously)

• Observe outside world• Orient to infer meaning of

observations

• Adjust waveform as needed to achieve goal

• Implement processes needed to change waveform

Other processes: (as needed)

• Adjust goals (Plan)• Learn about the outside

world, needs of user,…

Urgent

Allocate Resources

Initiate Processes

Negotiate Protocols

OrientInfer from Context

Select Alternate

Goals

Plan

Normal

Immediate

LearnNew

States

Observe

Outside

World

Decide

Act

User Driven

(Buttons)Autonomous


StatesGenerate “Best”

Waveform

Establish Priority

Parse Stimuli

Pre-process

Cognition cycle

Conceptual Operation[Mitola_99]

8


15

Implementation Classes

• Weak cognitive radio

– Radio’s adaptations

determined by hard coded

algorithms and informed by

observations

– Many may not consider this

to be cognitive (see

discussion related to Fig 6

in 1900.1 draft)

• Strong cognitive radio

– Radio’s adaptations

determined by conscious

reasoning

– Closest approximation is

the ontology reasoning

cognitive radios

In general, strong cognitive radios have potential to achieve both much better and much worse behavior in a network, but may not be realizable.


16

Brilliant Algorithms and Cognitive Engines

• Most research focuses on development of algorithms for:– Observation

– Decision processes

– Learning

– Policy

– Context Awareness

• Some complete OODA loop algorithms

• In general different algorithms will perform better in different situations

• Cognitive engine can be viewed as a software architecture

• Provides structure for incorporating and interfacing different algorithms

• Mechanism for sharing information across algorithms

• No current implementation standard

9


17

Security

User Model

Policy Model

WSGA

Evolver

|(Simulated Meters) – (Actual Meters)| Simulated

Meters

Actual Meters

Cognitive System Module

Cognitive System Controller

Chob

Uob

Reg

Knowledge BaseShort Term Memory

Long Term Memory

WSGA Parameter Set

Regulatory Information

Initial Chromosomes

WSGA Parameters

Objectives and weights

System Chromosome

maxmax

UUU

CHCHCH

USD

USD

•=

•=

Decision Maker

CE

-use

r inte

rface

User Domain

User preference

Local service facility

Policy Domain

User preference

Local service facility

Security

User data security

System/Network security

X86/Unix

Terminal

Radio-domain cognition

Radio Resource

Monitor

Performance API Hardware/platform API

Radio

Radio Performance

Monitor

WMS

CE-Radio Interface

Search SpaceConfig

Channel

Identifier

Waveform

Recognizer

Observation

Orientation

Action

Example Architecture from CWT

Decision

Learning

Models


18

DFS in 802.16h• Drafts of 802.16h defined

a generic DFS algorithm

which implements

observation, decision,

action, and learning

processes

• Very simple

implementation

Modified from Figure h1 IEEE 802.16h-06/010 Draft IEEE Standard for Local and metropolitan area networks Part 16: Air Interface for Fixed Broadband Wireless Access Systems Amendment for Improved Coexistence Mechanisms for License-Exempt Operation, 2006-03-29

Channel Availability

Check on next channel

Available?

Choose

Different Channel

Log of Channel

Availability

Stop Transmission

Detection?

Select and change to

new available channel

in a defined time with a

max. transmission time

In service monitoring

of operating channel

Channel unavailable for

Channel Exclusion time

Available?

Background In service

monitoring (on non-

operational channels)

Service in function

No

No

No

Yes

Yes

Start Channel Exclusion timer

Yes

Learning

Observation

Decision,

Action

Decision,

Action

Observation

10


19

Observation Sources

• RF Chain

– Signal

detection/classification

– Active ranging

• GPS

– Location, time

• Network

– Others’ observations

• Device sensors

– Biometrics, temperature

• User interfaces

BPSK


20

Orientation Processes

• Gives radio significance

of observations

– Does multipath profile

correspond to a known

location?

– Really just hypotheses

testing

• Algorithms

– Data mining

– Hidden Markov Models

– Neural Nets

– Fuzzy Logic

– Ontological Reasoning

ΣΣΣΣa

Threshold Logic Unit

f (a)

w1

w2

wn

x1

x2

xn

11


21

Decision Processes

• Purpose: Map what radio

believes about network state to an adaptation

• Guided by radio goal and

constrained by policy

– May be supplemented with

model of real world

• Common algorithms (mostly heuristics)

– Genetic algorithms

– Simulated annealing

– Local search

– Case based reasoningFigure from Fig 2.6 in I. Akbar, “Statistical Analysis of Wireless Systems Using Markov Models,” PhD Dissertation, Virginia Tech, January 2007


22

Learning Processes• Informs radio when situation is not like

one its seen before or if situation does not correspond to any known situation

• Logically, just an extension to the orientation process with – a “none of the above” option– Increase number of hypotheses after

“none of the above”– Refine hypotheses and models

• Algorithms:– Data mining– Hidden Markov Models– Neural Nets– Fuzzy Logic

– Ontological Reasoning– Case based learning– Bayesian learning

• Other proposed learning tasks– New actions, new decision rules, new

channel models, new goals, new internal algorithms

12


23Modified from Table 13.1 in M. Kokar, The Role of Ontologies in Cognitive Radio in Cognitive Radio Technology, ed., B. Fette, 2006.

Knowledge Representation• Issue:

– How are radios “aware” of their environment and how do they learn from each other?

• Technical refinement:– “Thinking” implies some language for thought.

• Proposed languages:– Radio Knowledge Representation Language– XML– Web-based Ontology Language (OWL)


24

Points to Remember

• Used cognitive radio definition– a radio with the capacity to acquire and apply

knowledge especially toward a purposeful goal

• Key Implementation aspects– Techniques have been proposed and prototyped for

all of the core cognitive radio functionalities (observe, orient, decide, learn, act)

– Major research efforts will be driven by applications• Standardizing ontologies for common applications

• Refining classification methods for particular applications

• Standardizing software architectures/APIs

1


1





• Approaches to Designing Cognitive Radio

Networks

• Applications of Networked Cognitive Radios

• Research and Future Directions


2

The Problem with Networked Cognitive Radios

Concept, Examples,

and Modeling


2


3

OODA Loop: (continuously)

• Observe outside world• Orient to infer meaning of

observations

• Adjust waveform as needed to achieve goal

• Implement processes needed to change waveform

Other processes: (as needed)

• Adjust goals (Plan)• Learn about the outside

world, needs of user,…

Urgent

Allocate Resources

Initiate Processes

Negotiate Protocols


Select Alternate

Goals

Plan

Normal

Immediate

LearnNew

States

Observe

Outside

World

Decide

Act

User Driven

(Buttons)Autonomous



Waveform

Establish Priority

Parse Stimuli

Pre-process

Cognition cycle

Conceptual Operation[Mitola_99]


4

The Interaction Problem

• Outside world is determined by the interaction of numerous cognitive radios

• Adaptations spawn adaptations

Outside

World

3


5

Dynamic Spectrum Access Pitfall• Suppose

– g31>g21; g12>g32 ;g23>g13

• Without loss of generality– g31, g12, g23 = 1

– g21, g32, g13 = 0.5

• Infinite Loop!– 4,5,1,3,2,6,4,…

Interf.

Chan.

(1.5,1.5,1.5)(0.5,1,0)(1,0,0.5)(0,0.5,1)(0,0.5,1)(1,0,0.5)(0.5,1,0)(1.5,1.5,1.5)

(1,1,1)(1,1,0)(1,0,1)(1,0,0)(0,1,1)(0,1,0)(0,0,1)(0,0,0)

Interference Characterization

0 1 2 3 4 5 6 7

1

2

3


6

Implications

• In one out every four deployments, the

example system will enter into an infinite

loop

• As network scales, probability of entering

an infinite loop goes to 1:

– 2 channels

– k channels

• Even for apparently simple algorithms,

ensuring convergence and stability will

be nontrivial

( ) ( ) 31 3 / 4

n Cp loop ≥ −

( ) ( ) 111 1 2n kC

kp loop+− +≥ − −

4


7

Locally optimal decisions that lead to globally undesirable networks

• Scenario: Distributed SINR maximizing power control in a single cluster

• For each link, it is desirable to increase transmit power in response to increased interference

• Steady state of network is all nodes transmitting at maximum power

Power

SINR

Insufficient to consider only a

single link, must consider

interaction


8

Potential Problems with Networked Cognitive Radios

Distributed

• Infinite recursions

• Instability (chaos)

• Vicious cycles

• Adaptation collisions

• Equitable distribution of

resources

• Byzantine failure

• Information distribution

Centralized

• Signaling Overhead

• Complexity

• Responsiveness

• Single point of failure

5


9

1. Steady state characterization

2. Steady state optimality

3. Convergence

4. Stability/Noise

5. Scalabilitya1

a2

NE1

NE2

NE3

a1

a2

NE1

NE2

NE3

a1

a2

NE1

NE2

NE3

a1

a2

NE1

NE2

NE3

a3

Steady State CharacterizationIs it possible to predict behavior in the system?

How many different outcomes are possible?

OptimalityAre these outcomes desirable?Do these outcomes maximize the system target parameters?

ConvergenceHow do initial conditions impact the system steady state?What processes will lead to steady state conditions?

How long does it take to reach the steady state?

Stability/NoiseHow do system variations/noise impact the system?Do the steady states change with small variations/noise?

Is convergence affected by system variations/noise?

ScalabilityAs the number of devices increases,

How is the system impacted?

Do previously optimal steady states remain optimal?

Network Analysis Objectives

(Radio 1’s available actions)

(Ra

dio

2’s

ava

ilab

le a

ctio

ns)

focu

s


10

Why focus on OODA loop, i.e., why exclude other levels?• OODA loop is implemented

now (possibly just ODA loop as little work on context awareness)

• Changing plans

– Over short intervals plans don’t change

– Messy in the general case (work could easily accommodate better response equivalent goals)

• Negotiating

– Could be analyzed, but protocols fuzzy

– General case left for future work

• Learning environment

– Implies improving observations/orientation. Over short intervals can be assumed away

– Left for future work

• Creation of new actions, new goals, new decision rules makes analysis impossible

– Akin to solving a system of unknown functions of unknown variables

– Most of this learning is supposed to occur during “sleep” modes

• Won’t be observed during operation

6


11

General Model (Focus on OODA Loop Interactions)

• Cognitive Radios • Set N

• Particular radios, i, j

Outside

World


12


Actions• Different radios may

have different capabilities

• May be constrained by policy

• Should specify each radio’s available actions to account for variations

• Actions for radio i – Ai

Act

7


13


Decision Rules

• Maps observations to actions– di:O→Ai

• Intelligence implies that these actions further the radio’s goal– ui:O→ℜ

• The many different ways of doing this merit further discussion

Decide

Implies very simple,

deterministic function,

e.g., standard

interference function


14

Modeling Interactions (1/3)

Radio 2

Actions

Radio 1

Actions

Action Space

u2u1

Decision

Rules

Decision

Rules

Outcome Space

:f A O→Informed by

Communications

Theory

( )1 2ˆ ˆ,γ γ

( )1 1u γ ( )2 2ˆu γ

8


15

Modeling Interactions (2/3)

• Radios implement actions, but observe outcomes.

• Sometimes the mapping between outcomes and actions is one-to-one implying f is invertible.

• In this case, we can express goals and decision rules as functions of action space.– Simplifies analysis

• One-to-one assumption invalid in presence of noise.


16

Modeling Interactions (3/3)• When decisions are made

also matters and different radios will likely make decisions at different time

• Tj – when radio j makes its adaptations– Generally assumed to be an

infinite set

– Assumed to occur at discrete time

• Consistent with DSP implementation

• T=T1∪T2∪⋅⋅⋅∪Tn

• t ∈ T

Decision timing classes

• Synchronous

– All at once

• Round-robin

– One at a time in order

– Used in a lot of analysis

• Random

– One at a time in no order

• Asynchronous

– Random subset at a time

– Least overhead for a

network

9


17

Cognitive Radio Network Modeling Summary

• Radios

• Actions for each radio

• Observed Outcome

Space

• Goals

• Decision Rules

• Timing

• i,j ∈N, |N| = n

• A=A1×A2×⋅⋅⋅×An

• O

• uj:O→ℜ (uj:A→ℜ)

• dj:O→Ai (dj:A→ Ai)

• T=T1∪T2∪⋅⋅⋅∪Tn


18

DFS Example

• Two radios

• Two common channels

– Implies 4 element action space

• Both try to maximize Signal-to-

Interference Ratio

• Alternate adaptations

10


19

Items to Remember

• Cognitive radios introduce interactive decision problems

• When studying a cognitive radio network should identify– Who are the decision makers

– Available adaptations of the decision makers

– Goals guiding the decision makers

– Rules being used to formulate decisions

– Any timing information

1


1









2

A Whirlwind Review of Game Theory

Normal form games, Nash equilibria, Pareto efficiency, Improvement Paths, Noise


2


3

Game

1. A (well-defined) set of 2 or more players2. A set of actions for each player.3. A set of preference relationships for each

player for each possible action tuple.

• More elaborate games exist with more components but these

three must always be there.

• Some also introduce an outcome function which maps action

tuples to outcomes which are then valued by the preference

relations.

• Games with just these three components (or a variation on

the preference relationships) are said to be in Normal form

or Strategic Form


4

Set of Players (decision

makers)

• N – set of n players consisting of players “named” 1, 2, 3,…,i, j,…,n

• Note the n does not mean that there are 14 players in every game.

• Other components of the game that “belong”to a particular player are normally indicated by a subscript.

• Generic players are most commonly written as i or j.

• Usage: N is the SET of players, n is the number of players.

• N \ i = 1,2,…,i-1, i+1 ,…, n All players in Nexcept for i

3


5

Actions

Ai – Set of available actions for player i

ai – A particular action chosen by i, ai ∈ Ai

A – Action Space, Cartesian product of all Ai

A=A1× A2×· · · × An

a – Action tuple – a point in the Action Space

A-i – Another action space A formed from

A-i =A1× A2×· · · ×Ai-1 × Ai+1 × · · · × An

a-i – A point from the space A-i

A = Ai × A-i

A1= A-2

A2 = A-1

a

a1 = a-2

a2 = a-1

Example Two Player

Action Space

A1 = A2 = [0 ∞)

A=A1× A2

b

b1 = b-2

b2 = b-1


6

Preference Relation expresses an individual player’s desirability of

one outcome over another (A binary relationship)

*

io o

o is preferred at least as much as o* by player i

i

Preference Relationship (prefers at least as much as)

i Strict Preference Relationship (prefers strictly more than)

~i “Indifference” Relationship (prefers equally)

*

io o *

io o

iff *

io o

but not

*~io o *

io o

iff *

io o

and

Preference Relations (1/2)

4


7

Preference Relationship (2/2)

• Games generally assume the relationship

between actions and outcomes is

invertible so preferences can be

expressed over action vectors.

• Preferences are really an ordinal

relationship

– Know that player prefers one outcome to

another, but quantifying by how much introduces difficulties


8

Preference Relation then defined as

*

ia a

Maps action space to set of real numbers.

iff ( ) ( )*

i iu a u a≥

*

ia a iff ( ) ( )*

i iu a u a>

*~i

a a iff ( ) ( )*

i iu a u a=

:iu A →R

A mathematical description of preference relationships.

Utility Functions (1/2)(Objective Fcns, Payoff Fcns)

5


9

Utility Functions (2/2)

By quantifying preference relationships all sorts of valuable

mathematical operations can be introduced.

Also note that the quantification operation is not unique as

long as relationships are preserved. Many map preference

relationships to [0,1].

Example

Jack prefers Apples to Oranges

JackApples Oranges ( ) ( )Jack Jacku Apples u Oranges>

a) uJack(Apples) = 1, uJack(Oranges) = 0

b) uJack(Apples) = -1, uJack(Oranges) = -7.5


10

Variety of game models• Normal Form Game <N,A,ui>

– Synchronous play

– T is a singleton

– Perfect knowledge of action space, other players’ goals (called utility functions)

• Repeated Game <N,A,ui,di>– Repeated synchronous play of a normal form game

– T may be finite or infinite

– Perfect knowledge of action space, other players’ goals (called utility functions)

– Players may consider actions in future stages and current stages• Strategies (modified di)

• Asynchronous myopic repeated game <N,A,ui,di,T>– Repeated play of a normal form game under various timings

– Radios react to most recent stage, decision rule is “intelligent”

• Many others in the literature and in the dissertation

6


11

NormalUrgent

Allocate Resources

Initiate Processes


Establish Priority

PlanNormal

Negotiate

Immediate

LearnNew

States

Goal

Adapted From Mitola, “Cognitive Radio for Flexible Mobile Multimedia Communications ”, IEEE Mobile Multimedia Conference, 1999, pp 3-10.

Observe

Outside

World

Decide

Act

Autonomous


States

\

Utility function

Arguments

Utility Function

Outcome Space

Action Sets

Decision

Rules

Cognitive radios are naturally modeled as players in a game


12

Radio 2

Actions

Radio 1

Actions

Action Space

u2u1

Decision Rules

Decision

Rules

Outcome Space

:f A O→Informed by

Communications

Theory

( )1 2ˆ ˆ,γ γ

( )1 1u γ ( )2 2ˆu γ

Interaction is naturally modeled as a game

7


13

Some differences between game models

and cognitive radio network model

Can learn O (may know or learn A)Knows AKnowledge

Not invertible (noise)

May change over time (though relatively

fixed for short periods)

Has to learn

Invertible

Constant

Knownf : A →O

Cardinal (goals)OrdinalPreferences

Cognitive RadioPlayer

• Assuming numerous iterations, normal form game only has a single stage.– Useful for compactly capturing modeling components

at a single stage

– Normal form game properties will be exploited in the analysis of other games


14

1. Steady state characterization

2. Steady state optimality

3. Convergence

4. Stability/Noise

5. Scalabilitya1

a2

NE1

NE2

NE3

a1

a2

NE1

NE2

NE3

a1

a2

NE1

NE2

NE3

a1

a2

NE1

NE2

NE3

a3

Steady State CharacterizationIs it possible to predict behavior in the system?

How many different outcomes are possible?

OptimalityAre these outcomes desirable?Do these outcomes maximize the system target parameters?

ConvergenceHow do initial conditions impact the system steady state?What processes will lead to steady state conditions?

How long does it take to reach the steady state?

Stability/NoiseHow do system variations/noise impact the system?Do the steady states change with small variations/noise?

Is convergence affected by system variations/noise?

ScalabilityAs the number of devices increases,

How is the system impacted?

Do previously optimal steady states remain optimal?

Network Analysis Objectives

(Radio 1’s available actions)

(Ra

dio

2’s

ava

ilab

le a

ctio

ns)

focu

s

8


15

Steady-states

• Recall model of <N,A,di,T> which we characterize with the evolution function d

• Steady-state is a point where a*= d(a*) for all t ≥t *

• Obvious solution: solve for fixed points of d.

• For non-cooperative radios, if a* is a fixed point under synchronous timing, then it is under the other three timings.

• Works well for convex action spaces– Not always guaranteed to exist

– Value of fixed point theorems

• Not so well for finite spaces– Generally requires exhaustive search


16

An action vector from which no player can profitably unilaterally deviate.

( ) ( ), ,i i i i i iu a a u b a− −≥An action tuple a is a NE if for every i ∈ N

for all bi ∈Ai.

Definition

Note showing that a point is a NE says nothing about the

process by which the steady state is reached. Nor anything about its uniqueness.

Also note that we are implicitly assuming that only pure

strategies are possible in this case.

“A steady-state where each player holds a correct expectation of the other players’ behavior and acts rationally.” - Osborne

Nash Equilibrium

9


17

Examples

• Cognitive Radios’

Dilemma

– Two radios have two

signals to choose

between n,w and N,W

– n and N do not overlap

– Higher throughput from

operating as a high

power wideband signal

when other is

narrowband

• Jamming Avoidance

– Two channels

– No NE

(-1,1)(1,-1)1

(1,-1)(-1,1)0

10

Jammer

Transmitter


18

Nash Equilibrium as a Fixed

Point

• Individual Best Response

• Synchronous Best Response

• Nash Equilibrium as a fixed point

• Fixed point theorems can be used to establish existence of NE (see dissertation)

• NE can be solved by implied system of equations

( ) ( ) ( ) ˆ : , ,i i i i i i i i i i iB a b A u b a u a a a A− −= ∈ ≥ ∀ ∈

( ) ( )ˆ ˆi

i NB a B a

∈= ×

( )* *ˆa B a=

10


19

Best Response Analysis

( ) ( )ˆ / 1ic B K N i N= − + ∀ ∈

( )i k i i

k N

u c B c c Kc∈

= − −

∑Goal

( )\

ˆ / 2i i k

k N i

c B c B K c∈

= = − −

∑Best Response

Simultaneous System of

Equations

( )ˆ / 6ic B K i N= − ∀ ∈Solution

Generalization


20

Significance of NE for CRNs

• Why not “if and only if”?– Consider a self-motivated game with a local maximum and a hill-climbing

algorithm.

– For many decision rules, NE do capture all fixed points (see dissertation)

• Identifies steady-states for all “intelligent” decision rules with the same goal.

• Implies a mechanism for policy design while accommodating differing implementations

– Verify goals result in desired performance

– Verify radios act intelligently

Autonomously Rational Decision Rule

11


21

Optimality

• In general we assume the existence of some design objective function J:A→R

• The desirableness of a network state, a, is the value of J(a).

• In general maximizers of J are unrelated to fixed points of d.

Figure from Fig 2.6 in I. Akbar, “Statistical Analysis of Wireless Systems Using Markov Models,” PhD Dissertation, Virginia Tech, January 2007


22

Example Functions

• Utilitarian– Sum of all players’ utilities

– Product of all players’utilities

• Practical– Total system throughput

– Average SINR

– Maximum End-to-End Latency

– Minimal sum system interference

• Objective can be unrelated to utilities

Utilitarian Maximizers

System Throughput Maximizers

Interference Minimization

12


23

Price of Anarchy (Factor)

• Centralized solution always at least as good as distributed solution– Like ASIC is always at least as good as

DSP

• Ignores costs of implementing algorithms– Sometimes centralized is infeasible (e.g.,

routing the Internet)

– Distributed can sometimes (but not generally) be more costly than centralized

Performance of Centralized Algorithm Solution

Performance of Distributed Algorithm Solution

≥≥≥≥ 1

9.6

7


24

Implications

• Best of All Possible Worlds– Low complexity distributed algorithms with low anarchy factors

• Reality implies mix of methods– Hodgepodge of mixed solutions

• Policy – bounds the price of anarchy

• Utility adjustments – align distributed solution with centralized solution

• Market methods – sometimes distributed, sometimes centralized

• Punishment – sometimes centralized, sometimes distributed, sometimes both

• Radio environment maps –”centralized” information for distributed decision processes

– Fully distributed• Potential game design – really, the panglossian solution, but only

applies to particular problems

13


25

Pareto efficiency (optimality)

• Formal definition: An action vector a* is Pareto efficient if there exists no other action vector a, such that every radio’s valuation of the network is at least as good and at least one radio assigns a higher valuation

• Informal definition: An action tuple is Pareto efficient if some radios must be hurt in order to improve the payoff of other radios.

• Important note– Like design objective function, unrelated to fixed

points (NE)

– Generally less specific than evaluating design objective function


26

Example Games

a1

b1

a2 b2

1,1 -5,5

-1,-15,-5

a1

b1

a2 b2

1,1 -5,5

3, 35,-5

Legend Pareto Efficient

NE NE + PE

14


27

The Notion of Time and Imperfections in Games and Networks

Extensive Form Games, Repeated Games, Convergence Concepts in Normal Form Games, Trembling Hand Games, Noisy Observations


28

Model Timing Review• When decisions are

made also matters and different radios will likely make decisions at different time

• Tj – when radio j makes its adaptations– Generally assumed to be

an infinite set

– Assumed to occur at discrete time

• Consistent with DSP implementation

• T=T1∪T2∪⋅⋅⋅∪Tn

• t ∈ T

Decision timing classes

• Synchronous– All at once

• Round-robin– One at a time in order

– Used in a lot of analysis

• Random– One at a time in no order

• Asynchronous– Random subset at a time

– Least overhead for a network

15


29

Repeated Games

Stage 1

Stage 2

Stage k

Stage 1

Stage 2

Stage k

• Same game is repeated

– Indefinitely

– Finitely

• Players consider discounted payoffs across multiple stages

– Stage k

– Expected value over all

future stages

( ) ( )k k k

i iu a u aδ=

( )( ) ( )0

k k k

i i

k

u a u aδ∞

=

=∑


30

Myopic Processes

• Players have no knowledge about utility functions, or expectations about future play, typically can observe or infer current actions

• Best response dynamic – maximize individual performance presuming other players’ actions are fixed

• Better response dynamic – improve individual performance presuming other players’ actions are fixed

• Interesting convergence results can be established

16


31

Paths and Convergence• Path [Monderer_96]

– A path in Γ is a sequence γ = (a0, a1,…) such that for every k ≥ 1 there exists a unique player such that the strategy combinations (ak-1, ak) differs in exactly one coordinate.

– Equivalently, a path is a sequence of unilateral deviations. When discussing paths, we make use of the following conventions.

– Each element of γ is called a step.

– a0 is referred to as the initial or starting point of γ.

– Assuming γ is finite with m steps, am is called the terminal point or ending point of γ and say that γ has length m.

• Cycle [Voorneveld_96]– A finite path γ = (a0, a1,…,ak) where ak = a0


32

Improvement Paths

• Improvement Path– A path γ = (a0, a1,…) where for all k≥1,

ui(ak)>ui(a

k-1) where i is the unique deviator at k

• Improvement Cycle– An improvement path that is also a cycle

– See the DFS example

γ2

γ1

γ3

γ4γ5

γ6γ2

γ1

γ3

γ4γ5

γ6

17


33

Convergence Properties

• Finite Improvement Property (FIP)

– All improvement paths in a game are finite

• Weak Finite Improvement Property (weak

FIP)

– From every action tuple, there exists an improvement path that terminates in an NE.

• FIP implies weak FIP

• FIP implies lack of improvement cycles

• Weak FIP implies existence of an NE


34

Examples

a

b

A B

1,-1

-1,1

0,2

2,2

Game with FIP

a

b

A B

1,-1 -1,1

1,-1-1,1

C

0,2

1,2

c 2,12,0 2,2

Weak FIP but not FIP

18


35

Implications of FIP and weak

FIP

• Assumes radios are incapable of reasoning ahead and must react to internal states and current observations

• Unless the game model of a CRN has weak FIP, then no autonomously rational decision rule can be guaranteed to converge from all initial states under random and round-robin timing (Theorem 4.10 in dissertation).

• If the game model of a CRN has FIP, then ALL autonomously rational decision rules are guaranteed to converge from all initial states under random and round-robin timing.– And asynchronous timings, but not immediate from definition

• More insights possible by considering more refined classes of decision rules and timings


36

Decision Rules

19


37

Absorbing Markov Chains and Improvement Paths

• Sources of randomness– Timing (Random, Asynchronous)

– Decision rule (random decision rule)

– Corrupted observations

• An NE is an absorbing state for autonomously rational decision rules.

• Weak FIP implies that the game is an absorbing Markov chain as long as the NE terminating improvement path always has a nonzero probability of being implemented.

• This then allows us to characterize – convergence rate,

– probability of ending up in a particular NE,

– expected number of times a particular transient state will be visited


38

Connecting Markov models,

improvement paths, and decision rules

• Suppose we need the path γ = (a0, a1,…am) for convergence by weak FIP.

• Must get right sequence of players and right sequence of adaptations.

• Friedman Random Better Response– Random or Asynchronous

• Every sequence of players have a chance to occur• Random decision rule means that all improvements have a chance to

be chosen

– Synchronous not guaranteed

• Alternate random better response (chance of choosing same action)– Because of chance to choose same action, every sequence of

players can result from every decision timing.

– Because of random choice, every improvement path has a chance of occurring

20


39

Convergence Results (Finite Games)

• If a decision rule converges under round-robin, random, or synchronous timing, then it also converges under asynchronous timing.

• Random better responses converge for the most decision timings and the most surveyed game conditions.– Implies that non-deterministic procedural cognitive radio

implementations are a good approach if you don’t know much about the network.


40

Trembling Hand (“Noise” in Games)

• Assumes players have a nonzero chance of making an error implementing their action.

– Who has not accidentally handed over the wrong

amount of cash at a restaurant?

– Who has not accidentally written a “tpyo”?

• Related to errors in observation as erroneous observations cause errors in implementation (from an outside observer’s perspective).

21


41

Noisy decision rules

• Noisy utility ( ) ( ) ( ), ,i i iu a t u a n a t= +

Trembling

Hand

Observation

Errors


42

Implications of noise

• For random timing, [Friedman] shows game with noisy random better response is an ergodic Markov chain.

• Likewise other observation based noisy decision rules are ergodic Markov chains– Unbounded noise implies chance of adapting (or not adapting) to

any action

– If coupled with random, synchronous, or asynchronous timings, then CRNs with corrupted observation can be modeled as ergodic Makov chains.

– Not so for round-robin (violates aperiodicity)

• Somewhat disappointing– No real steady-state (though unique limiting stationary

distribution)

22


43

DFS Example with three access points

• 3 access nodes, 3 channels, attempting to operate in band with least spectral energy.

• Constant power• Link gain matrix

• Noiseless observations

• Random timing

12

3


44

Trembling Hand

• Transition Matrix, p=0.1

• Limiting distribution

23


45

Noisy Best Response

• Transition Matrix, N(0,1) Gaussian Noise

• Limiting stationary distributions


46

Comment on Noise and Observations

• Cardinality of goals makes a difference for cognitive radios– Probability of making an error is a function of the difference

in utilities

– With ordinal preferences, utility functions are just useful fictions

• Might as well assume a trembling hand

• Unboundedness of noise implies that no state can be absorbing for most decision rules

• NE retains significant predictive power– While CRN is an ergodic Markov chain, NE (and the

adjacent states) remain most likely states to visit

– Stronger prediction with less noise

– Also stronger when network has a Lyapunov function

– Exception - elusive equilibria ([Hicks_04])

24


47

Items to Remember

• NE are always fixed points for self-interested adaptations– But may not be ALL fixed points

• Many ways to measure optimality

• Randomness helps convergence

• Unbounded noise implies network has a theoretically non-zero chance to visit every possible state

1


1









2

Designing Cognitive Radio Networks to Yield Desired Behavior

Policy, Cost Functions, Global Altruism, Potential Games


2


3

Potential Problems with Networked Cognitive Radios

Distributed

• Infinite recursions

• Instability (chaos)

• Vicious cycles

• Adaptation collisions

• Equitable distribution of

resources

• Byzantine failure

• Information distributionDistribution of Trusted Accurate

Information

Decision Interaction

Timing


4

Working with Interactive Decisions

• Design network to be a potential game

– Any self interested decision process will converge

• Limit decisions to processes known to converge

– Best responses in a supermodular game

• Limit effects of interactions

– Policy

• Eliminate interaction

– Centralize decision making

– Collaboration

– Repeated game with punishment

3


5

Policy

• Concept: Constrain the available actions so the worst cases of distributed decision making can be avoided

• Not a new concept –– Policy has been used since

there’s been an FCC

• What’s new is assuming decision makers are the radios instead of the people controlling the radios


6

Policy applied to radios instead of humans

• Need a language to convey policy– Learn what it is

– Expand upon policy later

• How do radios interpret policy– Policy engine?

• Need an enforcement mechanism– Might need to tie in to humans

• Need a source for policy– Who sets it?

– Who resolves disputes?

• Logical extreme can be quite complex, but logical extreme may not be necessary.

Policies

frequency

mask

4


7

• Detection

– Digital TV: -116 dBm over a 6 MHz channel

– Analog TV: -94 dBm at the peak of the NTSC

(National Television System Committee) picture carrier

– Wireless microphone: -107 dBm in a 200 kHz

bandwidth.

• Transmitted Signal

– 4 W Effective Isotropic Radiated Power (EIRP)

– Specific spectral masks

– Channel vacation timesC. Cordeiro, L. Challapali, D. Birru, S. Shankar, “IEEE 802.22: The First Worldwide Wireless Standard based on Cognitive

Radios,” IEEE DySPAN2005, Nov 8-11, 2005 Baltimore, MD.

802.22 Example Policies


8

Repeated Games

Stage 1

Stage 2

Stage k

Stage 1

Stage 2

Stage k

• Same game is repeated

– Indefinitely

– Finitely

• Players consider discounted payoffs across multiple stages

– Stage k

– Expected value over all

future stages

( ) ( )k k k

i iu a u aδ=

( )( ) ( )0

k k k

i i

k

u a u aδ∞

=

=∑

5


9

Impact of Strategies

• Rather than merely reacting to the state of the network, radios can choose their actions to influence the actions of other radios

• Threaten to act in a way that minimizes another radio’s performance unless it implements the desired actions

• Common strategies– Tit-for-tat

– Grim trigger

– Generous tit-for-tat

• Play can be forced to any “feasible” payoff vector with proper selection of punishment strategy.


10

Impact of Communication on Strategies

nada

c

Nada C

0,0 -5,5

-1,15,-5

N

-100,0

-100,-1

n -1,-1000,-100 -100,-100

• Players agree to play in a certain manner

• Threats can force play to almost any state

– Breaks down for finite number of stages

6


11

Improvement from Punishment

A. MacKenzie and S. Wicker, “Game Theory in Communications:

Motivation, Explanation, and Application to Power Control,” Globecom2001,

pp. 821-825.

• Throughput/unit power gains be enforcing a common received power level at a base station

• Punishment by jamming

• Without benefit to deviating, players can operate at lower power level and achieve same throughput


12

Instability in Punishment

• Issues arise when radios aren’t directly observing actions and are punishing with their actions without announcing punishment

• Eventually, a deviation will be falsely detected, punished and without signaling, this leads to a cascade of problems

V. Srivastava, L. DaSilva, “Equilibria for Node Participation in Ad Hoc Networks –An Imperfect Monitoring Approach,” ICC 06, June 2006, vol 8, pp. 3850-3855

7


13

Comments on Punishment

• Works best with a common controller to announce

• Problems in fully distributed system– Need to elect a controller

– Otherwise competing punishments, without knowing other players’ utilities can spiral out of control

• Problems when actions cannot be directly observed– Leads to Byzantine problem

• No single best strategy exists– Strategy flexibility is important

– Significant problems with jammers (they nominally receive higher utility when “punished”

• Generally better to implement centralized controller– Operating point has to be announced anyways


14

Cost Adjustments

• Concept: Centralized unit dynamically adjusts costs in radios’ objective functions to ensure radios operate on desired point

• Example: Add -12 to use of wideband waveform

( ) ( ) ( )i i iu a u a c a= +

8


15

Comments on Cost Adjustments

• Permits more flexibility than policy

– If a radio really needs to deviate, then it can

• Easy to turn off and on as a policy tool

– Example: protected user shows up in a channel, cost to use that channel goes up

– Example: prioritized user requests channel,

other users’ cost to use prioritized user’s channel goes up (down if when done)


16

Potential Games

9


17

Potential Games

time

Φ(ω

)

• Existence of a function (called the potential function, V), that reflects the change in utility seen by a unilaterally deviating player.

• Cognitive radio interpretation:– Every time a cognitive radio

unilaterally adapts in a way that furthers its own goal, some real-valued function increases.


18

Exact Potential Games

( ) ( ) ( ) ( ), , , , , ,i i i i i i i i i i i i i i i

V a a V b a u a a u b a a b A a A− − − − − −− = − ∀ ∈ ∀ ∈

Definition Exact Potential Game

A normal form game whose objective functions are

structured such that there exists some function P: A →ℜwhich satisfies the following property for all players:

In other words it must be possible to construct a single-

dimensional function whose change in value is exactly equal to

the change in value of the deviating player.

10


19

Example Potential Game (1/2)

a1

b1

a2 b2

1,1 0, 0

3, 30, 0

Coordination Game

( )

( )( )( )( )

1 2

1 2

1 2

1 2

1 ,

0 ,

0 ,

3 ,

a a a

a a bV a

a b a

a b b

=

==

= =

u1(a1,a2) - u1(b1,a2) = 1 = V(a1,a2) - V(b1,a2)

u2(a1,a2) – u2(a1,b2) = 1 = V(a1,a2) - V(a1,b2)

u1(b1,b2) - u1(a1,b2) = 3 = V(b1,b2) - V(a1,b2)

u2(b1,b2) – u2(b1,a2) = 3 = V(b1,b2) - V(b1,a2)

Note: V is not unique.

Consider V’ = V + c

where c is a constant.

Also note the relation

between CG Prop. 2

and V


20

Example Potential Game (2/2)

a1

b1

a2 b2

4,2 -1, 1

2, 13,-2

( )

( )( )( )( )

1 2

1 2

1 2

1 2

1 ,

0 ,

0 ,

3 ,

a a a

a a bV a

a b a

a b b

=

==

= =

u1(a1,a2) - u1(b1,a2) = 1 = V(a1,a2) - V(b1,a2)

u2(a1,a2) – u2(a1,b2) = 1 = V(a1,a2) - V(a1,b2)

u1(b1,b2) - u1(a1,b2) = 3 = V(b1,b2) - V(a1,b2)

u2(b1,b2) – u2(b1,a2) = 3 = V(b1,b2) - V(b1,a2)

The Same Potential!!

The Same NE!

Coordination Game

(In Equilibriums)

11


21

Comments on Second Example

a1

b1

a2 b2

1,1 0, 0

3, 30, 0

Coordination Game

a1

b1

a2 b2

4,2 -1, 1

2, 13,-2

Second Game

a1

b1

a2 b2

3,1 -1, 1

-1,-23,-2

Dummy Game

This is a property of all exact potential games.


22

Continuous Action Sets (1/2)

i

i i

V u

a a

∂ ∂=

∂ ∂

22 2ji

i j i j i j

uV u

a a a a a a

∂∂ ∂= =

∂ ∂ ∂ ∂ ∂ ∂

If objective functions are twice differentiable then a game is a

EPG iff

Let G be a game in which the strategy sets are closed intervals of

ℜ. Suppose the objective functions are continuously

differentiable. A function V is a potential iff V is continuously

differentiable and

for every i ∈ N

for every i, j ∈ N

EPG Property 5 (Shapley)

EPG Property 6 (Shapley)

12


23

Vector Operations

Consider the set of EPG EPG1, EPG2,…,EPGK, with player set

N and action space A, and objective functions , and

potential functions . Form a new game, G, with

player set N, action space A, and objective functions given by

1 2, , , K

i i iu u u…

1 2, , , KV V V…

1 1 2 2G K K

i i i iu u u u cα α α= + + + + . Then G is an EPG with an EPF

given by1 1 2 2 K KV V V Vα α α= + +


24

Example

13


25

Exact Potential Game Forms

• Many exact potential games can be recognized by the form of the utility function


26

Proving the BSI Relationship( ) ( ) ( )

\

,i ij i j i i

j N i

u a w a a h a∈

= −∑

( ) ( ), ,i i i i i iu a a u b a− −− =

( ) ( )i i i ih a h b+ −( )

( ) \ \

, ,ij i j ij i j

j N i j N i

w a a w b a∈ ∈

−∑ ∑

( ) ( ) ( )1

1

,i

ij i j i i

i N j i N

V a w a a h a−

∈ = ∈

= −∑∑ ∑

( ) ( ), ,i i i iV a a V b a− −− = ( ) ( )1 1

\ 1 \ 1

, ,k k

kj k j kj k j

k N i j k N i j

w a a w a a− −

∈ = ∈ =

−∑ ∑ ∑ ∑( ) ( )\ \

, ,ij i j ij i j

j N i j N i

w a a w b a∈ ∈

−∑ ∑

( ) ( )\ \

k k k k

k N i k N i

h a h a∈ ∈

+ −∑ ∑( ) ( )i i i ih a h b+ −

14


27

Steady-states• As noted previously, FIP implies existence of NE

• Existence in infinite games for continuous potential function on compact action space

• Generally a subset of NE (which is a subset of steady-states)

• Sometimes only steady-states are maximizers of V


28

Optimality

• If ui are designed so that maximizers of V are coincident

with your design objective function, then NE are also optimal.

• (*) Can also introduce cost function to utilities to move NE.

• In theory, can make any action tuple the NE– May introduce additional NE

– For complicated NC, might as well completely redesign ui

( ) ( )* *

0i i

V a NC a

a a

∂ ∂+ =

∂ ∂

( ) ( ) ( )*

i iu a u a NC a= +

V

a

15


29

FIP and Potential Games

• GOPG implies FIP ([Monderer_96])

• FIP implies GOPG for finite games

([Milchtaich_96])

• Thus we have a non-exhaustive search

method for identifying when a CRN game

model has FIP.

• Thus we can apply FIP convergence (and

noise) results to finite potential games.


30

Convergence in Infinite Potential Games

• ε-improvement path– Given ε >0, an ε-improvement path is a path such that for all

k≥1, ui(ak)>ui(a

k-1)+ ε where i is the unique deviator at step k.

• Approximate Finite Improvement Property (AFIP)– A normal form game, Γ, is said to have to have the

approximate finite improvement property if for every ε>0 there exists an such that the length of all ε-improvement paths in Γare less than or equal to L.

• [Monderer_96] shows that exact potential games have AFIP, we showed that AFIP implies a generalized ε-potential game.

16


31

Convergence Implications

• [Ermoliev] showed directional better response, averaged best response, best response in infinite games.

• Asynchronous convergence new, convergence for infinite games with FIP is new

• Associated convergence rates for bounded potential functions is new


32Theorem 3.4 in A. Medio, M. Lines, Nonlinear Dynamics: A Primer, Cambridge University Press, Cambridge, UK, 2001.

Direct Method for Discrete Time Systems

17


33

Stability of Potential games

• New result

– Applying the discrete-time version of Lyapunov function to isolated maximizersof V with Lyapunov function

– Indicates stability of any better response decision algorithm with round-robin or random timing

( ) ( ) ( )*VL a V a V a= − +


34

Potential Game Properties• Steady-states

– Finite game NE can be found from maximizers of V.

• Optimality– Can adjust exact potential games with additive cost function

(that is also an exact potential game)

– Sometimes little better than redesigning utility functions

• Game convergence – Potential game assures us of FIP (and weak FIP)

– DV satisfy Zangwill’s (if closed)

• Noise/Stability– Isolated maximizers of V have a Lyapunov function for

decision rules in DV

• Remaining issue:– Can we design a CRN such that it is a potential game for

the convergence, stability, and steady-state identification properties

– AND ensure steady-states are desirable?

18


35

What does game theory bring to the design of cognitive radio networks? (1/2)

• A natural “language” for modeling cognitive radio networks

• Permits analysis of ontological radios– Only know goals and that radios will adapt

towards its goal

• Simplifies analysis of random procedural radios

• Permits simultaneous analysis of multiple decision rules – only need goal

• Provides condition to be assured of possibility of convergence for all autonomously myopic cognitive radios (weak FIP)


36

Interference Reducing Networks

Designing desirable potential game cognitive radio networks

time

Φ(ω

)

19


37

Interference Reducing Networks

• Concept– Cognitive radio network is a potential game with a potential function

that is negation of observed network interference

• Definition– A network of cognitive radios where each adaptation decreases the

sum of each radio’s observed interference is an IRN

• Properties:– Optimal steady-state exists

– Network converges

– Every adaptation improves performance

– Stability for isolated minimizers of Φ• But are they isolated?

( ) ( )i

i N

Iω ω∈

Φ =∑

timeΦ

(ω)


38

Implementing IRNs

• Gather interference information from other devices in the network

• Conceptually obvious implementations

• Scales badly

• A “bureaucratic nightmare”

• Design network such that adaptations implement an IRN without gathering information on other devices’ interference

• Scales well – ideal solution

• Non-obvious how to implement

– Invisible hand of cognitive radio?

Implicit informationExplicit information

Image source: http://unbeknownst.net/images/bureaucracy.jpg

20


39

Globally Altruistic Networks (Explicit Information)

• Radio goal: minimize network interference

• Potential, Interference Function

• Unique benefit: Works for all waveform adaptations

• Unique drawback: Lots of overhead – may need functional radio environment map

• Proposed algorithms that satisfy GAN: [Sung], [Nie]

( ) ( )\

i i

k N j N k

u Iω ω∈ ∈

= −∑ ∑

( ) ( ) ( )\

i

k N j N k

V Iω ω ω∈ ∈

Φ = − =∑ ∑


40

Locally Altruistic Networks (Explicit Information)

• Let denote the set of radios which are close enough that i produces non-negligible interference.

• Goal: minimize interference of those within “range”

• Same interference and potential function as before (just eliminated terms for which Ii = 0)

• Benefit – Less overhead, just as generalizable

• Drawback – Need extra routine to identify

i N⊆I

( ) ( )\i i

i i

k j k

u Iω ω∈ ∈

= −∑ ∑I I

iI

21


41

Global Altruism: distributed, but more costly

• Concept: All radios distributed all relevant information to all other radios and then each independently computes jointly optimal solution– Proposed for spreading code allocation in Popescu04, Sung03

– Used in xG Program (Comments of G. Denker, SDR Forum Panel Session on “A Policy Engine Framework”) Overhead ranges from 5%-27%

• C = cost of computation

• I = cost of information transfer from node to node

• n = number of nodes

• Distributed– nC + n(n-1)I/2

• Centralized (election)– C + 2(n-1)I

• Price of anarchy = 1

• May differ if I is asymmetric


42

Improving Global Altruism

• Global altruism is clearly inferior to a centralized solution for a single problem.

• However, suppose radios reported information to and used information from a common database– n(n-1)I/2 => 2nI

• And suppose different radios are concerned with different problems with costs C1,…,Cn

• Centralized– Resources = 2(n-1)I + sum(C1,…,Cn)

– Time = 2(n-1)I + sum(C1,…,Cn)

• Distributed– Resources = 2nI + sum(C1,…,Cn)

– Time = 2I + max (C1,…,Cn)

22


43

Example Application: • Overlay network of secondary

users (SU) free to adapt power, transmit time, and channel

• Without REM:– Decisions solely based on link

SINR

• With REM– Radios effectively know everything

Upshot: A little gain for the secondary users; big gain for primary users

From: Y. Zhao, J. Gaeddert, K. Bae, J. Reed, “Radio Environment Map Enabled Situation-Aware Cognitive Radio Learning Algorithms,” SDR Forum Technical Conference 2006.


44

Comments on Radio Environment Map

• Local altruism also possible

– Less information transfer

• Like policy, effectively needs a common language

• Nominally could be centralized or distributed database

• Read more:

– Y. Zhao, B. Le, J. Reed, “Infrastructure Support – The Radio Environment MAP,” in Cognitive Radio

Technology, B. Fette, ed., Elsevier 2006.

23


45

Bilateral Symmetric Interference

• Two cognitive radios, j,k∈N, exhibit bilateral symmetric interference if

Source: http://radio.weblogs.com/0120124/Graphics/geese2.jpg

What’s good for the goose, is

good for the gander…

( ) ( ), ,jk j j k kj k k jg p g pρ ω ω ρ ω ω= ,j j k kω ω∀ ∈ Ω ∀ ∈ Ω• ωk – waveform of radio k

• pk - the transmission power of radio k’s waveform

• gkj - link gain from the transmission source of radio k’ssignal to the point where radio jmeasures its interference,

• - the fraction of radio k’s signal that radio j cannot exclude via processing (perhaps via filtering, despreading, or MUD techniques).

( ),k jρ ω ω


46

Proof:

( ) ( ) ( ), , ,ki k k i ik i i k ik i kg p g p bρ ω ω ρ ω ω ω ω= =

( ) ( )\

,i ik i k

k N i

u bω ω ω∈

= −∑

( ) ( )1

1

,i

ki k k i

i N k

V g pω ρ ω ω−

∈ =

= −∑∑

( ) ( )2Vω ωΦ = −

• By bilateral symmetric interference

• Rewrite goal

• Therefore a BSI game (Si =0) (an EPG)

• Interference Function

• Therefore unilateral deviations increase V and decrease Φ(ω) – an IRN

24


47

Situations where BSI occurs• Isolated Network Clusters

– All devices communicate with a common access node with identical received powers.

– Clusters are isolated in signal space

• Close Proximity Networks– All devices are sufficiently

close enough that waveform correlation effects dominate

• Controlled Observation Processes– Leverage knowledge of

waveform protocol to control observations to achieve BSI


48

Isolated Network Clusters

• In this operational scenario, the network consists of a set of clusters C for which the following operational assumptions hold:– Perhaps through judicious frequency or code reuse between

clusters, each radio i is operating in a cluster c∈C for which

is a subset of the cluster.

– The cluster head enforces a uniform receive power, rc, on all radios k for signals transmitted to the cluster head.

– Waveforms are restricted to those waveforms for which

– Cluster heads provide interference measurements to all client radios in the cluster.

• Therefore

( ) ( ), ,k i i kρ ω ω ρ ω ω=

( ) ( ), ,jk j j k kj k k jg p g pρ ω ω ρ ω ω=

25


49

Example Simulation

• Single cell

• 7 cognitive

radios

• 6 code dimensions

• Interference minimizing

• Round-robin


50

Close Proximity Networks

• In this operational scenario it is assumed that the radios are operating as an ad-hoc network in sufficiently close proximity and transmitting with sufficiently similar power levels that waveform correlation dominates the distance and transmitted power effects are negligible. Under these assumptions -Iiis equivalent to

• Further, assume

( ) ( )\

,i k i

k N i

u ω ρ ω ω∈

= −∑( ) ( ), ,k i i kρ ω ω ρ ω ω=

26


51

DFS Close Proximity Network Simulation

• Specific parameters

– Signal bandwidth = 1 MHz

– Channel bandwidth = 10

Mhz

– 10 (decision making) links

– Frequency discretized with

center frequencies every

0.1 MHz

– Random initial frequencies


52

Controlled Observation Process

• Concept:

– Control the radios’ observation processes so that they only observe signals where

• Is this possible to do with meaningful

results?

( ) ( ), ,jk j j k kj k k jg p g pρ ω ω ρ ω ω=

27


53

802.11 – A victim of its own success

• Extremely large number of 802.11 deployments

– Overlapping coverage

produces interference

and contention

– Reduces throughput

• Solution 1: Deploy David nationally

• Solution 2: Cognitive Radio and DFS


54

An IRN 802.11 DFS Algorithm• Suppose each access node

measures the received signal power and frequency of the RTS/CTS (or BSSID) messages sent by observable access nodes in the network.

• Assumed out-of-channel interference is negligible and RTS/CTS transmitted at same power

( ) ( )jkkkjkjjjk ffpgffpg ,, σσ =

( ) ( ) ( )\

,i i ki k i k

k N i

u f I f g p f fσ∈

= − = −∑

( )1

,0

i k

i k

i k

f ff f

f fσ

==

≠

Listen on

Channel LC

RTS/CTS

energy detected?Measure power of access node

in message, p

Note address

of access node, a

Update

interference table

Time for decision?Apply decision criteria for new

operating channel, OCUse 802.11h

to signal change

in OC to clients

yn

Pick channel tolisten on, LC

y

n

Start

28


55

A DFS simulation of the process

• 30 cognitive access nodes

• Upper 5 GHz 802.11 band

• Choose channel with

lowest interference

• One randomly selected

access node adapts at

each instance

• n=3 path loss exponent

• Random initial channels

• Randomly distributed

positions over 1 km2

• Random timing


56

Dynamic Frequency Selection

Final channels by access node

29


57

(lowest frequency that improves) (highest frequency that improves)

Noiseless suboptimal adaptations


58

Policy VariationsBy Channel By Radio (10 radios only low channels)

30


59

Local Frequency Preferences• Each radio has a

random real constant added to its observation of each channel

• Exact potential game– BSI + Self-motivated

• Equivalent to having legacy devices present– If legacy devices are

transmitting at the same power as cognitive radios, then sum of


60

Asynchronous Timing

• Best Response

– p = 0.02

• Not monotonic,

but still an

absorbing

Markov chain as FIP and

potential game

theory predicts

31


61

Noisy Observations

• Observations corrupted by clipped noise modeled as

Gaussian

– Mean = -90 dBm

– Var = - 90 dBm

• Not stable

• Why?

– Large number of

equilibria, not isolated

– Fails Lyapunov’s direct

method


62

Stabilized Process

• Threshold adaptation (ε-better response)– Only adapt if

adaptation expected to reduce interference by at least -85 dBm

• Stabilizes di:O→Anot di:A→A– Small variations

in observations

32


63

Statistics• 30 cognitive access nodes in European UNII

bands

• Choose channel with lowest interference

• Random timing

• n=3

• Random initial channels

• Randomly distributed positions over 1 km2

0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

70

Number of Access Nodes

Reduction in N

et

Inte

rfere

nce (

dB

)

Round-robin Asynchronous Legacy Devices

Reduction in Net Interference

Reduction in Net Interference


64

Items to Remember

• In addition to interactive decisions, timing and distribution of information are critical

• Policies are a good way to limit worst case scenarios

• Additive cost functions can shape behavior

• Collaboration and centralization can eliminate interactive decision problems

• Punishment can limit incentives to cheat on collaborative agreements– But is very sensitive to the design

• Under special conditions (bilateral symmetric interference), interactive decisions form a virtuous cycle

1


1











Flavors of 802.11,

802.16, 802.22, and radios that collaborate

Applications and Emerging Standards


2


3

• Explicitly opened up Japanese spectrum for 5

GHz operation

• Part of larger effort to force equipment to operate based on geographic region, i.e., the local policy

2.4822.448France

2.4732.447Spain

2.4952.473Japan

2.482.402Europe

2.482.402U.S.

UpperLower

2.4 GHz

US

UNII Low 5.15 – 5.25 (4) 50 mW

UNII Middle 5.25 – 5.35 (4) 250 mW

UNII Upper 5.725-5.825 (4) 1 W

5.47 – 5.725 GHz released in Nov 2003

Europe

5.15-5.35 200 mW

5.47-5.725 1 W

Japan

4.9-5.091

5.15-5.25 (10 mW/MHz) unlicensed

5 GHz

802.11j – Policy Based Radio


4

From "Vision RFC", http://www.darpa.mil/ato/programs/XG/

DARPA XG Program – Policy Agility

• Issue: – How are radios “aware” of

their environment and how do they learn from each other?

• Technical refinement:– “Thinking” implies some

language for thought.

• Approaches– Radio Knowledge

Representation Language

– xG Policy Language

– Web-based Ontology Language (OWL)

• Spectrum agility and policy agility are envisioned

• Load new policy constraints "on the fly" using flash cards or internet

3


5

• Enhances QoS for Voice over Wireless IP (aka Voice over WiFi ) and streaming multimedia

• Changes

– Enhanced Distributed Coordination Function (EDCF)

• Shorter random backoffs for higher priority traffic

– Hybrid coordination function (orientation)

• Defines traffic classes

• In contention free periods, access point controls medium access (observation)

• Stations report to access info on queue size. (Distributed sensing)

802.11e – Almost Cognitive


6

• Dynamic Frequency

Selection (DFS)

– Avoid radars

• Listens and discontinues

use of a channel if a radar is

present

– Uniform channel utilization

• Transmit Power Control (TPC)

– Interference reduction

– Range control

– Power consumption Savings

– Bounded by local regulatory

conditions

802.11h – Unintentionally Cognitive

4


7

802.11h: A simple cognitive radio

Observe

– Must estimate channel characteristics (TPC)

– Must measure spectrum (DFS)

Orientation

a) Radar present?

b) In band with satellite??

c) Bad channel?

d) Other WLANs?

Decision

– Change frequency

– Change power

– Nothing

Action

Implement decision

Learn

– Not in standard, but most implementations should learn the environment to address intermittent signals

Outside

World

Observe

OrientDecide

Act

Learn


8

• Ports 802.11a to 3.65 GHz – 3.7 GHz (US Only) – FCC opened up band in July 2005

– Ready 2008

• Intended to provide rural broadband access• Incumbents

– Band previously reserved for fixed satellite service (FSS) and radar installations –including offshore

– Must protect 3650 MHz (radar)– Not permitted within 80km of inband government radar– Specialized requirements near Mexico/Canada and other incumbent users

• Leverages other amendments– Adds 5,10 MHz channelization

(802.11j)– DFS for signaling for radar

avoidance (802.11h)

• Working to improve channel announcement signaling

• Database of existing devices– Access nodes register at

http://wireless.fcc.gov/uls– Must check for existing devices at

same siteSource: IEEE 802.11-06/0YYYr0

802.11y

5


9

• Wireless Regional Area Networks (WRAN)– Aimed at bringing broadband access in rural and

remote areas

– Takes advantage of better propagation characteristics at VHF and low-UHF

– Takes advantage of unused TV channels that exist in these sparsely populated areas

• 802.22 is to define:– Physical layer specifications

– Policies and procedures for operation in the VHF/UHF TV Bands between 54 MHz and 862 MHz

– Cognitive Wireless RAN Medium Access Control

IEEE 802.22


10

802.22 Deployment Scenario• Devices

– Base Station (BS)

– Customer Premise Equipment (CPE)

• Master/Slave relation– BS is master

– CPE slave

• Max Transmit CPE 4W

Figure from: IEEE 802.22-06/0005r1

6


11

• Data Rates 5 Mbps – 70 Mbps

• Point-to-multipoint TDD/FDD

• DFS, TPC

• Adaptive Modulation– QPSK, 16, 64-QAM, Spread QPSK

• OFDMA on uplink and downlink

• Use multiple contiguous TV channels when available

• Fractional channels (adapting around microphones)

• Space Time Block Codes

• Beam Forming– No feedback for TDD (assumes channel reciprocity)

• 802.16-like ranging

Proposed PHY Features of

802.22


12

Possible MAC Features of

802.22

• 802.16 MAC plus the following

– Multiple channel support

– Coexistence

• Incumbents

• BS synchronization

• Dynamic resource sharing

– Clustering support

– Signal detection/classification routines

• Security based on 802.16e security

7


13

• Observation– Signal strength and feature detection

– Aided by distributed sensing (CPEs return data to BS)

– Digital TV: -116 dBm over a 6 MHz channel

– Analog TV: -94 dBm at the peak of the NTSC (National Television System Committee) picture carrier

– Wireless microphone: -107 dBm in a 200 kHz bandwidth.

– Possibly aided by spectrum usage tables

• Orientation– Infer type of signals that are present

• Decision– Frequencies, modulations, power levels, antenna choice (omni and

directional)

• Policies– 4 W Effective Isotropic Radiated Power (EIRP)

– Spectral masks, channel vacation times

C. Cordeiro, L. Challapali, D. Birru, S. Shankar, “IEEE 802.22: The First Worldwide Wireless Standard based on Cognitive Radios,”

IEEE DySPAN2005, Nov 8-11, 2005 Baltimore, MD.

Cognitive Aspects of 802.22


14

Sensing Aspects of 802.22• Region based sensing

– Remote aided sensing

• Algorithm:– Partition cell into disjoint

regions– For each region assign a

remote (Customer Premise Equipment)

• Example considered squares with 500 m sides

– CPE feeds back what it finds

• Number of incumbents• Occupied bands

Source: IEEE 802.22-06/0048r0

0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

70

80

90

100

8


15

802.16h• Draft to ballot Oct 06,

67% approve, resolving comments)

• Improved Coexistence Mechanisms for License-Exempt Operation

• Basically, a cognitive radio standard

• Incorporates many of the hot topics in cognitive radio

– Token based negotiation

– Interference avoidance

– Network collaboration– RRM databases

• Coexistence with non 802.16h systems

– Regular quiet times for other systems to transmit

From: M. Goldhamer, “Main concepts of IEEE P802.16h / D1,” Document Number:

IEEE C802.16h-06/121r1, November 13-16, 2006.


16

General Cognitive Radio Policies in 802.16h

• Must detect and avoid radar and other higher priority systems

• All BS synchronized to a GPS clock

• All BS maintain a radio environment map (not their name)

• BS form an interference community to resolve interference differences

• All BS attempt to find unoccupied channels first before negotiating for free spectrum– Separation in frequency, then separation in time

9


17

DFS in 802.16h• Adds a generic

algorithm for performing Dynamic

Frequency Selection in license exempt bands

• Moves systems onto unoccupied channels based on observations

• Works when there is no interaction

Generic DFS Operation Figure h1(fuzziness in original)


18

Adaptive Channel Selection

• Used when BS turns on

• First – attempt to find a vacant channel– Passive scan

– Candidate Channel Determination

– Messaging with Neighbors

• Second – attempt to coordinate for an exclusive channel

• If unable to find an empty channel, then BS attempts to join the interference community on the channel it detected the least interference

Figure h37: IEEE 802.16h-06/010 Draft IEEE Standard for Local and metropolitan area networks Part 16: Air Interface for Fixed Broadband Wireless Access Systems Amendment for Improved Coexistence Mechanisms for License-Exempt Operation, 2006-03-29

10


19

Collaboration in 802.16h

• BS can request interfering systems to back off transmit power

• Master BS can assign transmit timings

– Intended to support up to 3 systems (Goldhammer)

• Slave BS in an interference community can “bid” for interference free times via tokens.

• Master BS can advertise spectrum for “rent” to other Master BS

– Bid by tokens

• Collaboration supported via Base Station Identification Servers, messages, and RRM databases

• Interferer identification by finding power, angle of arrival, and spectral density of OFDM/OFDMA preambles

• Every BS maintains a database or RRM information which can be queried by other BS

– This can also be hosted remotely


20

Cognitive Radio Applications Automated Interoperability

Spectrum Trading

Collaborative TechniquesIntelligent Beamforming

Cheaper Radios?

Improved Link ReliabilityOpportunistic Spectrum

Utilization

Advanced Networking

Source Cluster Relay cluster Destination Cluster

First Hop Second Hop





Destination Cluster

11


21

Collaborative Radio

• A radio that leverages the services of other radios to further its goals or the goals of the networks.

• Cognitive radio enables the collaboration process– Identify potential collaborators

– Implies observations processes

• Classes of collaboration– Distributed processing

– Distributed sensing


22

• Cooperation: relay nodes transmit its own information + source information

• Relay Operation:– Half duplex

• Time division, Frequency division, Code division

• Dedicated version in 802.16j, could be logically extended to non-dedicated relays

Cooperative Communication:Classical Relay Channel

source

relay

destination

Decode and Forward

source

relay

destination

Amplify and Forward

ij ij i jy x nα= +

12


23

Cooperative Antenna Arrays• Concept:

– Leverage other radios to effect an

antenna array

• Applications:

– Extended vehicular coverage

– Backbone comm. for mesh

networks

– Range extension with cheaper

devices

• Issues:

– Timing, mobility

– Coordination

– Overhead

• Dissertation of Ramesh Chembil

source

destination

Transmit Diversity

Cooperative MIMO

Source Cluster Relay cluster






Destination ClusterSource Cluster Relay cluster






Destination Cluster

http://scholar.lib.vt.edu/theses/available/etd-12132006-142934/


24

• Distributed processing

– Exploit different

capabilities on different

devices

• Maybe even for waveform

processing

– Bring extra

computational power to

bear on critical problems

• Useful for most

collaborative problems

• Collaborative sensing

– Extend detection range by

including observations of

other radios

• Hidden node mitigation

– Improve estimation statistics

by incorporating more

independent observations

– Immediate applicability in

802.22, likely useful in future

adaptive standards

Other Opportunities for Collaborative Radio (1/3)

13


25

• Improved localization

– Application of

collaborative sensing

– EMS location services for ad-hoc networks

– Friend finders (Samsung

USA working on this)

• Reduced contention MACs

– Collaborative

scheduling algorithms

to reduce collisions

– Perhaps of most value

to 802.11

• Some scheduling

included in 802.11e



26

• Distributed mapping

– Gather information relevant to

specific locations from mobiles

and arrange into useful maps

– Coverage maps

• Collect and integrate signal

strength information from mobiles

• If holes are identified and fixed,

should be a service differentiator

– Congestion maps

• Density of mobiles should

correlate with traffic (as in

automobile) congestion

• Customers may be willing to pay

for real time traffic information

• Theft detection

– Devices can learn which

other devices they tend to

operate in proximity of and

unexpected combinations

could serve as a security

flag (like flagging

unexpected uses of credit

cards)

– Examples:

• Car components that expect

to see certain mobiles in the

car

• Laptops that expect to

operate with specific

mobiles nearby


14


27

Items to Remember

• Initial standards ignored interaction, primary focus was on avoiding incumbents

• More recent standards act in a distributed fashion when possible to find non-interactive states, but collaborate to resolve interaction problem

• By collaborating, cognitive radios can provide performance beyond the capabilities of a single device– Collaborative MIMO, Collaborative Sensing

1


1











Technology trends

related to cognitive radio

Research and Future Directions


2


3

Urgent

Allocate Resources

Initiate Processes

Negotiate Protocols


Select Alternate

Goals

Plan

Normal

Immediate

LearnNew

States

Observe

Outside

World

Decide

Act

User Driven

(Buttons)Autonomous



Waveform

Establish Priority

Parse Stimuli

Pre-process

Cognition cycle

Increasing Availability of Network Information Databases

• Databases becoming available– 802.11k RRM

– 802.11v Network management

– 802.16f Network Management Information Base (MIB)

– 802.16g Network management plane

– 802.16i Mobile Management Information Base

• Cognitive radios need environmental information to make intelligent decisions

• Should simplify information gathering process


4

Increasing Flexibility in Network Topology• Numerous emerging topology

standards– 802.11s Mesh

– 802.16 Ad-hoc

– 802.16j MMR

– 802.15 WPAN ad-hoc

– 802.15.5 WPAN mesh networks

– 802.22 MMR?

• Greater flexibility in network choice– 802.21

– UMA

• Value in being able to recognize when ad-hoc modes or alternate networks are available and advisable => cognitive radio

IP or

Ethernet

3


5

Increasing Flexibility of PHY

• TDD OFDMA

• Beam Forming

• MIMO

• Adaptive coding/rate

• Adaptive modulation

• PHY and networking flexibility provides additional variables for cognitive radio to control

• Improved RRM, device performance

Burst 7

Ranging/BW Request

Burst 3

Burst 4

Burst 8

Frame n

Fre

qu

ency

Time

Bu

rst

1,2

Burst 5

Burst 6

IEEE 802.22-06/0105r0


6

Emergence of “Cute” CR-type Features at Application Layer

• TeleNav– Uses GPS and Maps to

give directions

– Similar to some features described by Mitola

• GPS (perhaps network assisted/provided) will be a critical source of information– Policy, remembering

coverage holes, spectrum mapping

4


7

Increasing Sophistication of SDR Hardware

Bin Le, Tom Rondeau, Jeff Reed, Charles Bostian, “Past, Present, and Future of ADCs,” IEEE Signal Processing Magazine, November 2005

800 1000 1200 1400 1600 1800 2000 2200 2400 2600(MHz)

AMPS

IS-95 GSM GPS DCS1800 802.11b/gUMTS

TDD

UMTS

FDDPCS1900

Bottom line: MEMS based RFICs dramatically expanding range

Bitwave4200 MHz

Data converters improving exponentially

PP

PP

PP

PP

++PP PP PP PP

PPPPPP PP

PPPP ++

PPPPPP PP

PPPP PP PP

PPPP PP PP

PPPP PP PP

++PP

PP

PP

P

I

PP P

II

II

II

II

Processor

Switch Matrix

Inter-picoArray Interface or Asynchronous Data Interface

Example signal flows

IIIPPP

picoArray Architecture

Improving processor capabilities


8

Dominance of Weak CR and

Continuous Aggregation of Algorithms

• Strong CR research unlikely to yield useful AI

– More likely to yield scenario classification and

negotiation routines useful for CBR systems

• Weak CR being implemented now

• Large number of conceivable and proposed applications suitable for Weak CR

• Simplest integration path is a good SDR with upgradeable control processes

– Add applications as developed

– Can become quite sophisticated

• Key will be scenario/opportunity recognition

5


9

Research Opportunities

What topics should we

be experimenting with

now?


10

Tools

• New technologies generally need new tools

• Examples:– Policy interpreters – intelligently and

automatically combine policy languages from different vendors

– CBR programming suite

– Multi-core compilers

– Development kits

6


11

Standards

• Portability and interoperability would greatly benefit from standardization of key aspects of cognitive radio

• Examples:– Standard cognitive language

– RRM databases

– Measurements (Interference temperature was an attempt)

– Minimal competencies

– Cognitive engine

– Interference avoidance techniques

– Software architectures• Some of these will be addressed by the 1900 group


12

• Develop Cognitive WiMAX– Intelligent Radio Resource Management

– Intelligent Physical/MAC layers

– Intelligent Fault Detection

• Why?– Enable less trained and experience personnel to

manage and install

– Improve performance and reliability

– Lower deployment and maintenance costs

– Ability to leverage both licensed, unlicensed and refarmed spectrum

Apply CR Principles to Existing Standards

7


13

Some Example WiMAXFunctions

• Radio Resource Management– Frequency management

• Avoid interference at cell boundaries

• Co-utilize license and unlicensed bands

– Smart antenna management

• Appropriate algorithm for the situation

• Efficiently use limited smart antenna capabilities

– Load balancing and coverage adjustments between cells

– Optimum handoff strategy: Intra and Inter system

• Mapping Applications to QoS management

• Implications of cognitive mobile multi-hop relay (MMR) network

Many parameters in the standard can be adjusted

using cognitive radio principles.


14

Traditional Issues in a New Domain

• Cognitive radio is still a radio so traditional

research areas still apply

• Examples:

– Security

– Code verification/validation

– Code portability

– Code/circuit optimization

– Front end flexibility and performance

8


15

Reducing Power Consumption

• Keep min. power by varying– Modulation

– Coding

– Carrier

– Filtering

– Sample rate

– Algorithms

– Bias points

– Application

• Radio knowledgeable of power consumed – Weight depends on hardware TX and RX

• Don’t specifically care about bandwidth!


16

Speculation on the Future

9


17

Commercial Market

• 1-3 years– Start of several initiatives and standards explicitly leveraging

cognitive radio

– Initial devices will be procedural and will not have cognitive engines

– Many market niche solutions

– First cognitive networks from upgraded standards

• 3-7 years– General purpose cognitive radio solutions become available

– Significant expansion of unlicensed bands

– Emergence of cognitive certification bodies

– Hacking your own cognitive radio becomes popular in certain circles

– Privacy issues will emerge


18

Technologies

• Radios will not become “conscious”, but will be a lot more intelligent

• Implementations will bifurcate into low complexity limited application cognitive radios and high complexity general purpose cognitive radios– Some low complexity radios will leverage REMs

– Some will gather observations, and implement network decisions

– Some will implement simple procedural rules

• REMs will grow increasingly important

• Lacking a standard, commercializing ontological reasoning will prove more difficult than it first appears because of issues with verification and portability

10


19

Summary of Points to Remember


20

Points to Remember

• Used cognitive radio definition– a radio with the capacity to acquire and apply

knowledge especially toward a purposeful goal

• Key Implementation aspects– Techniques have been proposed and prototyped for

all of the core cognitive radio functionalities (observe, orient, decide, learn, act)

– Major research efforts will be driven by applications• Standardizing ontologies for common applications

• Refining classification methods for particular applications

• Standardizing software architectures/APIs

11


21

Points to Remember

• Cognitive radios introduce interactive decision problems

• When studying a cognitive radio network should identify– Who are the decision makers

– Available adaptations of the decision makers

– Goals guiding the decision makers

– Rules being used to formulate decisions

– Any timing information


22

Points to Remember

• NE are always fixed points for self-interested adaptations– But may not be ALL fixed points

• Many ways to measure optimality

• Randomness helps convergence

• Unbounded noise implies network has a theoretically non-zero chance to visit every possible state

12


23

Points to Remember

• In addition to interactive decisions, timing and distribution of information are critical

• Policies are a good way to limit worst case scenarios

• Additive cost functions can shape behavior

• Collaboration and centralization can eliminate interactive decision problems

• Punishment can limit incentives to cheat on collaborative agreements– But is very sensitive to the design

• Under special conditions (bilateral symmetric interference), interactive decisions form a virtuous cycle


24

Points to Remember

• Initial standards ignored interaction, primary focus was on avoiding incumbents

• More recent standards act in a distributed fashion when possible to find non-interactive states, but collaborate to resolve interaction problem

• By collaborating, cognitive radios can provide performance beyond the capabilities of a single device– Collaborative MIMO, Collaborative Sensing

13


25

Points to Remember

• Supporting technologies for cognitive radio rapidly developing

• Potential value to adding cognitive radio on top of existing standards

• Dubious that cognitive radio will be “conscious” radio

• Cognitive radio will tend to either be very low complexity or very high complexity– Bimodal distribution

networking cognitive radioscrtwireless.com/files/cognitiveradionetworks.pdfadapted from mitola,...

Documents