Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more ➡
Download
Standard view
Full view
of .
Add note
Save to My Library
Sync to mobile
Look up keyword
Like this
2Activity
×
0 of .
Results for:
No results containing your search query
P. 1
A Game Theoretic Formulation for Strategic Sensing in Cognitive Radio Networks: Equilibrium Analysis and Fully Distributed Learning

A Game Theoretic Formulation for Strategic Sensing in Cognitive Radio Networks: Equilibrium Analysis and Fully Distributed Learning

Ratings: (0)|Views: 115|Likes:
Published by ijcsis
In cognitive radio (CR) systems, the licenced bands are opened to secondary users (SU), when they are not used. To support this reuse the CR users must sense the environment, and use the spectrum when its not utilized by primary users (PU). For these reasons, the spectrum sensing is an important topic in cognitive radio networks (CRN). In this paper, we consider a distributed opportunistic access (D-OSA), in which CR users attempt to access a channel licensed to a primary network. In this context, we formulate the problem of designing the equilibrium sensing time in a distributed manner, in order to maximize the throughput of CR users while guarantying a good protection to the PU. Next, we study the Nash equilibrium of the system, we also propose a combined learning algorithm for continuous actions that is fully distributed, and allows to the CR users to learn their equilibrium payoffs and their equilibrium sensing time. Finaly we give a comparison between the proposed solution and a centralized one. The simulation results show that the system can learn the sensing time and converge to a unique Nash equilibrium, which come to prove the theoretical study. The comparison between the proposed solution and the centralized one shows an expected result, which is the higher performances of the centralized method, in terms of throughput and sensing time, but as we can see in the simulation results, the difference is slight.
In cognitive radio (CR) systems, the licenced bands are opened to secondary users (SU), when they are not used. To support this reuse the CR users must sense the environment, and use the spectrum when its not utilized by primary users (PU). For these reasons, the spectrum sensing is an important topic in cognitive radio networks (CRN). In this paper, we consider a distributed opportunistic access (D-OSA), in which CR users attempt to access a channel licensed to a primary network. In this context, we formulate the problem of designing the equilibrium sensing time in a distributed manner, in order to maximize the throughput of CR users while guarantying a good protection to the PU. Next, we study the Nash equilibrium of the system, we also propose a combined learning algorithm for continuous actions that is fully distributed, and allows to the CR users to learn their equilibrium payoffs and their equilibrium sensing time. Finaly we give a comparison between the proposed solution and a centralized one. The simulation results show that the system can learn the sensing time and converge to a unique Nash equilibrium, which come to prove the theoretical study. The comparison between the proposed solution and the centralized one shows an expected result, which is the higher performances of the centralized method, in terms of throughput and sensing time, but as we can see in the simulation results, the difference is slight.

More info:

categoriesTypes, Research
Published by: ijcsis on Jul 30, 2013
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See More
See less

02/11/2014

pdf

text

original

 
A Game Theoretic Formulation for StrategicSensing in Cognitive Radio Networks : EquilibriumAnalysis and Fully Distributed Learning
Sofia BouferdaRITM-ESTC-GREENTICENSEM-Hassan II UniversityCasablanca, Moroccosofiabouferda@gmail.comEssaid SabirENSEM-GREENTICHassan II University,Casablanca, Moroccoe.sabir@ensem.ac.maAawatif HayarENSEM-GREENTICHassan II University,Casablanca, Moroccoa.hayar@greentic.uh2c.maMounir RIFIRITM-ESTC-GREENTICHassan II University,Casablanca, Moroccorifi@email.com
 Abstract
—In cognitive radio (CR) systems, the licenced bandsare opened to secondary users (SU), when they are not used. Tosupport this reuse the CR users must sense the environment, anduse the spectrum when its not utilized by primary users (PU).For these reasons, the spectrum sensing is an important topicin cognitive radio networks (CRN). In this paper, we considera distributed opportunistic access (D-OSA), in which CR usersattempt to access a channel licensed to a primary network. In thiscontext, we formulate the problem of designing the equilibriumsensing time in a distributed manner, in order to maximize thethroughput of CR users while guarantying a good protectionto the PU. Next, we study the Nash equilibrium of the system,we also propose a combined learning algorithm for continuousactions that is fully distributed, and allows to the CR users tolearn their equilibrium payoffs and their equilibrium sensingtime. Finaly we give a comparison between the proposed solutionand a centralized one. The simulation results show that thesystem can learn the sensing time and converge to a uniqueNash equilibrium, which come to prove the theoretical study. Thecomparison between the proposed solution and the centralizedone shows an expected result, which is the higher performancesof the centralized method, in terms of throughput and sensingtime, but as we can see in the simulation results, the differenceis slight.Keywords— Distributed Spectrum sensing, Nash equilibrium,combined learning
I. I
NTRODUCTION
The great development of communication has createdhuge demand of radio spectrum, which has become a scarceresource. Nevertheless, it has been found that the majorlicensed bands are underutilized and some of the remainingbands are heavily used [1]. This is in a part due to theactual resource allocation strategies, which allocate a fixedfrequency band to a single licensed system.Cognitive radio is a promising solution to this radio spectrumwastage, it allows the simultaneous access to the samespectrum band by PU and secondary users (SU), by usingspatial, temporal and frequency spectrum holes left by idlesPUs [2]Spectrum sensing has been identified as a key enablingcognitive radio to not interfere with PUs, by reliabilitydetecting PUs signals. It motivates the concept of opportunisticspectrum access (OSA), which allows secondary cognitiveradio (CR) systems to opportunistically exploit the under-utilized spectrum. A basic component of OSA is a sensingstrategy at the MAC layer for spectrum opportunity tracking.Due to hardware limitations, energy constraints, a secondaryuser may not be able to sense all channels in the spectrumsimultaneously, and due to throughput constraints Secondaryusers can not sense channels for a long time, in general thelonger the sensing time, the less time available for the CRuser to transmit data during the rest of the time frame. In thiscase a sensing strategy for intelligent channel selection andsensing time optimisation is necessary.The presence of high priority primary users and therequirement that the cognitive users should not interferewith them define a medium access paradigm called cognitivemedium access. Cognitive medium access control is relevantto both the cognitive radio and the coexistence community.Cognitive medium access control protocols play an importantrole to exploit the spectrum opportunities, manage theinterference to primary users, and coordinate the spectrumaccess amongst secondary users [3], [4], [5]
 A. Literature review
Cognitive medium access control game formulations havebeen studied in [8], [9], [10] using different approaches.Cognitive medium access control coupled with state estimationand detection for identifying and exploiting instantaneousspectrum opportunities is one of the central challenges. Inaddition, the opportunistic spectrum access needs to the takeinto consideration energy constraints of sensing, exploration,exploitation and transmission. Then the question is how todevelop cognitive sensing and access strategies that learn fromobservations and offer improved fairness and performanceover time [9], [10] without central controllers or dedicatedcommunication/control channels. In [11] the authors exam-ined evolutionary game theoretic perspective of the classi-cal medium access control. Cognitive considerations are notdiscussed. [5] proposed memory management for distributedspectrum sharing under limited spectrum sensing.In [8] the authors examine cognitive medium access and the(IJCSIS) International Journal of Computer Science and Information Security,Vol. 11, No. 6, June 2013
25http://sites.google.com/site/ijcsis/ISSN 1947-5500
 
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 11, No. 6, June 2013relation to competitive multi-armed bandit problem. The armbandit problem is well understood for single CR, which wishesto opportunistically exploit the availability of resources in thespectrum. For multiple users however the outcomes are notclear. This difficulty is in particular due to the interactivebehavior of the underlying processes of decision makingin dynamic environment. The authors proposed a Bayesianapproach to identify a trade-off metric between exploring theavailability of other free channels/time slots and exploiting theopportunities identified. In [10], a game theoretic learning andpricing have been proposed. In the above references, stochasticformulations of the medium access problem are examined.These formulations often lead to intractable configurations.Authors in [19], proposed a new two-step game where sens-ing and opportunistic access are jointly considered. A fullcharacterization of the Nash equilibria and analysis of theoptimal pricing policy, from the network owner view, for bothcentralized setting and decentralized setting, are also provided.Next, a combined learning algorithm that is fully distributedand allows the cognitive users to learn their optimal payoffsand their optimal strategies in both symmetric and asymmetriccases is proposed.
 B. contribution
In this paper, we propose to associate game theory tolearning strategies in cognitive medium access, to find theequilibrium sensing time. To the best of our knowledge,this is the first paper devoted to analyze distributed sensing.Meanwhile, the related literature usually considers the sensingtime from an optimization perspective [18] [13] [5]. Incontrast to the classical literature of medium access games,which does not focus on the random nature of cognitive radios,we propose a fully distributed strategic learning to learn theequilibrium payoff and the associated equilibrium strategies.Moreover, we provide many insightful results to understandthe possible relationship between sensing time and transmitprobability. Next, we analyze the impact of starting pointand speed of learning on convergence to Nash equilibrium.Finally we propose a comparison in terms of sensing time andthroughput between the proposed solution and a centralizedone.
C. Organization of the paper 
This paper will be organized as follows: In section II,the system model, the main notations and spectrum sensingpreliminaries are presented. In section III, we describe theutility function of the game and equilibrium analysis, insection IV, we propose a distributed learning algorithm, andin section V, a comparison between the proposed solution anda centralized one. Performance evaluation and results analysisare provided in section V.II. S
YSTEM
M
ODEL
A
ND
M
AIN
N
OTATIONS
 A. System model
We consider a secondary network that coexists with a pri-mary network where each PU is licensed to transmit wheneverhe/she wishes for most of time except for the case when thechannel is occupied by another PU. The duration of a primaryframe is denoted T. We consider that we have N SU tryingto access the spectrum of the PU. Throughout this work thefollowing consideration are taken into account:
Energy-based spectrum sensing: the primary networactivity is determined by measuring the signal strengthtraveling over the channel. If the received signal powerexceeds some given threshold, the channel is declaredbusy, it is declared idle in the other case.
Imperfect sensing in the sense that SUs may declare abusy channel while it is idle (false alarm).
Random access for data transmission of CRs. Here, weconsider that SUs follow a slotted aloha-like protocol totransmit data.During primary user’s activities, each SU i receives some givensignal. SU i samples the received signal at sampling frequency
s
without loss of generality, we assume that all SUs use thesame sampling frequency. The discrete received signal at theSU i can be represented as:
i
=
h
i
.S 
(
t
) +
n
(
t
)
: Hypothesis H
1
(
Busy
)
n
(
t
)
: Hypothesis H
0
(
Idle
)
(1)Where h
i
is the channel gain experienced by SU i and n(t)is an circular symmetric complex Gaussian noise with mean 0and variance
[
|
n
(
t
)
2
|
] =
σ
. The channel state is consideredas the binary hypothesis test H
0
and
1
.
 B. Energy based spectrum sensing
Spectrum sensing is often considered as a detection prob-lem. Many techniques were developed in order to detect theholes in the spectrum band. Focusing on each narrow band,existing spectrum sensing techniques are widely categorizedinto energy detection [6] and feature detection [7]. Althoughthis is not a restriction of our work, we will use energydetection throughout the paper.Let
τ 
i
be the sensing time and N the number of consideredsamples. Thus, we have
= [
τ
s
]
.it follows that the averageenergy detected by SU i is:T
i
(
) =
1
1
|
i
(
t
)
|
2
(2)
C. Imperfect sensing
We consider throughout this work a scenario where thespectrum sensor has imperfect detection performance. Inother terms, each SU i has a false alarm probability
fi
, i.e.,the probability that the channel is sensed to be busy while itis actually idle. Let
ǫ
denotes the threshold which specifiesthe collision tolerance bound of PUs. Then:
fi
(
ǫ,τ 
i
) =
Pr
(
i
(
)
> ǫ
\
0
)
(3)2
26http://sites.google.com/site/ijcsis/ISSN 1947-5500
 
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 11, No. 6, June 2013We assume that the primary signal is i.i.d complex Phaseshift keying modulated, with zero mean. We also assume thatthe primary signal and noise are independent. Using energydetection, the false alarm probability
fi
(
ǫ,τ 
i
)
can be estimated by the following:
fi
(
ǫ,τ 
i
) =12
erfc

ǫσ
2
1
 
τ 
i
s
(4)Where the complementary error function is given by:
erfc
(
x
) =2
√ 
π
 
x
e
t
22
dt
(5)
Fig. 1. Frame structure of the cognitive radio network.
Each SU i senses the channel for a duration
τ 
i
Fig. 1. Thisduration affects directly the false alarm probability
fi
(
ǫ,τ 
i
)
that the tagged SU i will experience. This relationship could bestated as: long sensing time leads to a more accurate sensingoperation (equivalently low false probability) but generatesshort transmission opportunity and high competition. Thus,the experienced throughput would be affected drastically bythe outcome of the sensing operation. The issue is then how tofind a good trade-off between sensing and received throughput.In [13], it is shown that there exists an optimal sensing timethat maximizes the throughput of CR users. Similar result canbe found in [14]. In both papers, authors consider the case of centralized control, in this work unlike the papers mentionedabove we consider equilibrium sensing time in a distributedframework.III. E
QUILIBRIUM ANALYSIS
In absence of primary network activity (i.e., under hypoth-esis
0
), the average Throughput of SU i is given by:
i
(
τ 
i
) =
τ 
i
τ 
i
(1
fi
(
ǫ,τ 
i
))
i
 
=
i
(1
j
(1
fj
(
ǫ,τ 
j
)))
(6)In the proposed game G each SU chooses an appropriatesensing time to maximize its utility function, which is equiva-lent to its throughput. In this context, it is important to ensurethe stability of the system. A concept which relates to thisissue is the Nash equilibrium. As definition in [15], a purestrategy profile
{
τ 
j
}
j
=
i
,j
∈ {
1
,..M 
}
is a Nash equilibriumof the proposed game if, for every player i (i.e. SU i):
i
(
τ 
i
,τ 
i
)
i
(
τ 
i
,τ 
i
)
i
{
1
,..M 
}
(7)A Nash equilibrium can be regarded as a stable solution, atwhich none of the users has the incentive to change its sensingtime
τ 
i
.
 A. Existence of the Nash equilibriumProposition 1:
Game G admits at least one Nash equilib-rium.
Proof 
: The conditions for the existence of Nash equilibriumin a strategic game are given in [16]:1) The set
τ 
i
is a nonempty, convex, and compact subsetof some Euclidean space for all i.2) The utility function
i
(
τ 
i
;
τ 
i
)
is continuous on
τ 
andquasi-concave on
τ 
i
.On one hand and according to the above description of thestrategy space, it is straightforward to see that it is nonempty,convex and compact. On the other hand, the utility function
i
is a continuous and quasi-concave function. Hence, gameG admits at least one Nash equilibrium.
i
(
τ 
i
,τ 
j
) =
τ 
i
τ 
i
(1
fi
(
ǫ,τ 
i
))
i
 
=
i
(1
j
(1
fj
(
ǫ,τ 
j
)))=
τ 
i
1
12
erfc
(
ǫσ
2
1)
 
τ 
i
s
i
 
=
i
1
j
1
12
erfc
(
ǫσ
2
1)
 
τ 
i
s

(8)Let G be the primitive of 
e
t
22
,
erfc
(
x
) =2
√ 
π
(
G
(+
)
G
(
x
))
erfc
(
x
) =
2
√ 
πG
(
x
) =
2
√ 
πe
x
22
(9)For simplicity of the derivation let put:
A
=
ǫσ
2
1
.
(10)
erfc
(
ǫσ
2
1)
is then written as follows:
erfc
(
A
 
τ 
i
s
)
(11)From (9) we can deduce the fisrt derivative of (11) :
erfc
(
A
 
τ 
i
s
) =
Af 
s
2
√ 
τ 
i
 
s
2
√ 
πe
2
τ is
2
=
A
 
s
τ 
i
πe
A
2
τ is
2
(12)3
27http://sites.google.com/site/ijcsis/ISSN 1947-5500

You're Reading a Free Preview

Download
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->