Multiple Service providers sharing Spectrum using Cognitive Radio in Wireless Communication Networks

Size: px

Start display at page:

Download "Multiple Service providers sharing Spectrum using Cognitive Radio in Wireless Communication Networks"

Lynette Jefferson
5 years ago
Views:

1 1 Multiple Service providers sharig Spectrum usig Cogitive Radio i Wireless Commuicatio Networks R. Kaiezhil ad Dr. C. Chadrasekar Abstract The curret utilizatio of the spectrum is quite iefficiet; cosequetly, if properly used, there is o shortage of the spectrum that is at preset available. Therefore, it is aticipated that more flexible use of spectrum ad spectrum sharig betwee radio systems will be key eablers to facilitate the successful implemetatio of future systems. Cogitive radio, however, is kow as the most itelliget ad promisig techique i solvig the problem of spectrum sharig. I this paper, we cosider the techique of spectrum sharig amog users of service providers to share the licesed spectrum of licesed service providers. It is show that the proposed techique reduces the call blockig rate ad improves the spectrum utilizatio. Idex Terms, call blockig, CR, service provider,spectrum sharig, spectrum utilizatio,system efficiecy. 1 INTRODUCTION N owadays, there are a lot of wireless applicatios sharig the same medium. This overload leads to a lack of spectrum i give frequecy bads. At this time, the allocatio of the spectrum is mostly static ad based o licesig, as i the case of the FM-bad. The purpose of this work is to preset aother approach of sharig the same medium by meas of a dyamic allocatio of the spectrum istead of the static allocatio which is believed to cause spectrum iefficiecy ad scarcity, sice the wireless users demads chage both temporally ad spatially. Dyamic spectrum sharig is a promisig approach for reusig the uderutilized spectrum, i which the spectrum is shared amog primary ad secodary (ulicesed) users (SUs) to improve spectrum flexibility ad, therefore, efficiecy. The cogitive radio (CR) provides a solutio to this dyamic spectrum access (DSA). CR techology is a key eabler for both real time spectrum markets ad dyamic sharig of licesed spectrum with ulicesed devices. The CR device is able to perform spectrum acquisitio, either through purchasig (i cleared spectrum) or sesig (i vacat chaels of geographical iterleaved spectrum), over a rage of frequecy bads ad operate i this spectrum at times ad locatios where it is able to trasmit i a o-iterferig basis. The overall spectral efficiecy of a system ca be improved with good coexistece properties, good spectrum sharig capabilities, as well as with flexibility i the spectrum use. Capabilities to share spectrum with other systems will sigificatly icrease the efficiecy as well as acceptability of the system. The overall spectral efficiecy of the work ca be also icreased with a flexible use of spectrum that adapts to the spatial ad temporal variatios i the traffic ad eviromet characteristics. The flexibility ad scalability of the system is importat also i order to simplify the etwork deploymet uder spectrum arragemets that may vary from regio to regio. Built-i capabilities for flexibility ad sharig may siificatly ease the task of spectrum idetificatio for the system. R. Kaiezhil is curretly pursuig Ph.D i Computer Sciece, Periyar Uiversity, Salem, Idia, PH kaiezhil@yahoo.co.i. I this paper, we propose a scheme about the efficiet spectrum utilizatio (ie spectrum sharig) amog service providers via static CR odes where the uder utilized spectrum of a particular service provider ca be shared by the overloaded service providers with a coordiatio amog them. This also proposes deployig a etwork of fixed cogitive radio (CR) odes to maitai spectrum sharig across multiple service providers operatig i the same geographical area. These CR odes estimate the spectrum utilisatio i a give area, ad cooperate to provide the spectrum usage iformatio for the overloaded ifrastructures of service providers. The algorithm SBAC is set to the beyod 3rd geeratio (B3G) wireless commuicatio systems. I have already proposed the Miimizig the Iterferece ad Performace of QoS based o spectrum sharig usig CR odes[1] ad the overall spectral efficiecy is carried out i the preset work. The paper is orgaized as follows; Sectio 2 ad 3 defies cogitive radio ad proposes a system model approach for its implemetatio. I sectio 4, Performace aalysis has bee ivestigated to improve the system efficiecy. Sectio 5 presets the simulatio results ad implemetatio issues. Fially, coclusios are preseted i Sectio 6. 2 COGNITIVE RADIO The key eablig techology of dyamic spectrum access techiques is cogitive radio (CR) techology, which provides the capability to share the wireless chael with licesed users i a opportuistic maer. The term, cogitive radio, ca formally be defied as follows : A Cogitive Radio is a radio that ca chage its trasmitter parameters based o iteractio with the eviromet i which it operates. From this defiitio, two mai characteristics of the cogitive radio ca be defied as follows: Cogitive capability: It refers to the ability of the radio techology to capture or sese the iformatio from its radio eviromet. Through this capability, the portios of the spectrum that are uused at a specific time or loca-

2 2 tio ca be idetified. Cosequetly, the best spectrum ad appropriate operatig parameters ca be selected. Recofigurability: The cogitive capability provides spectrum awareess whereas recofigurability eables the radio to be dyamically programmed accordig to the radio eviromet.. CR etworks are evisioed to provide high badwidth to mobile users via heterogeeous wireless architectures ad dyamic spectrum access techiques. This goal ca be realized oly through dyamic ad efficiet spectrum maagemet techiques. Fig.2 CR Nodes sesig the available chaels Fig. 1. Spectrum Hole Cocept The ultimate objective of the cogitive radio is to obtai the best available spectrum through cogitive capability ad recofigurability as described before. Sice most of the spectrum is already assiged, the most importat challege is to share the licesed spectrum without iterferig with the trasmissio of other licesed users as illustrated i Fig 1. The cogitive radio eables the usage of temporarily uused spectrum, which is referred to as spectrum hole or white space. If this bad is further utilized by a licesed user, the cogitive radio moves to aother spectrum hole or stays i the same bad, alterig its trasmissio power level or modulatio scheme to avoid iterferece as show i Fig Proposed CR Nodes Sesig With the capability of sesig the eviromet ad fidig the available spectrum dyamically, CR techique ca help to implemet spectrum sharig ad improve the spectrum utilizatio efficiecy. I the proposed work, CR egies such as lear ad decisio does the work of Spectrum Sesig. Here, CR odes helps to provide the list of available chaels ad it checks the availability of the chaels. Mobile odes get the iformatio of availability of chaels ad ss it to BS through the CR odes as show i Fig 2. Whe CR odes moitor chaels i a give area, ecoomically, the umber of deployed CR odes should be miimal, ad, meawhile, these CR odes ca fully cover the give area ad properly estimate the chael utilisatio. 3 PROPOSED SPECTRUM SHARING TECHNIQUES 3.1 Process of Specrtum Sharig Durig the peak hours, the commuicatio of the users will be blocked due the umber of active users is greater tha the maximal umber of users ie the ifrastructure is foud to be overloaded(because of the chael scarcity). At the same time, ifrastructure of the other service providers might be i the uder-loaded status. Hece, the available chaels of the uder-loaded service providers ca be utilized by the overloaded service providers. This may vary accordig to the services offered by the customers example Wireless iteret for laptops, mobile commuicatios ad so o. So this may lead to the differece i the traffic of the active users across service providers. Here, both of these two service providers operate cell-based wireless etworks. Therefore implemetig spectrum sharig amog service providers would highly improve the spectrum efficiecy ad it also reduces the call blockig rate ad Co- Chael Iterferece. I order to reduce the Co-Chael Iterferece ad to remove the eed of equippig CRs i each ifrastructure of service providers, we propose a method that implemets a spectrum sharig amog service providers via CR odes i a log term spectrum assigmet scheme. These CR odes are distributed regularly withi a area of iterest. Each CR ode seses the surroudig eviromet ad moitors the chael usage withi its sesig rage of differet service providers. To avoid co-chael Iterferece, CR odes provides the list of the chael availability of each cell for the overloaded ifrastructure of service provider. CR odes are coected to each other via wire or they commuicate wirelessly to form a etwork. This etwork is called as spectrum maagemet etwork ad it coexists with the wireless etworks operated by differet service providers. I this techique, oly a limited umber of CR odes are required, ad either users or service providers eed to sese the eviromet for available spectrum. I the proposed techique, each user subscribes to a specific service provider that is assiged fixed frequecy bads. Whe oe or more ifrastructures (e.g. base statios) of a service provider are overloaded, they use extra available chaels (for commuicatio) which are licesed to other service providers. The overloaded ifrastructures obtai the chael availability iformatio from surroudig CR odes. CR odes are deployed to estimate the chael utilisatio ad provide the chael usage iformatio for ifrastructures upo their requests. The ifra-

3 3 structures process the iformatio received from CR odes to select the optimum chaels based o the chael associated metrics such as iterferece level, cost ad the probability of chael beig available for a certai time duratio. 3.2 Operatios of CR Nodes Suppose, if the ifrastructure of oe service provider is foud to be overloaded, it ss the request to the adjacet CR odes regardig the chael usage availability i its coverage area. It gives the list of available chael from the relevat adjacet CR odes ad it also provides the relevat iformatio associated with each chael such as the average sigal to oise plus iterferece. The overloaded ifrastructure ow selects the proper chael to use. Thus, users ca commuicate with the overloaded ifrastructure over the ew chaels after they are iformed with the chael availability. This requires both ifrastructure ad users to be equipped with radios capable of operatig over differet frequecy bads. If the ifrastructure is ot foud to be overloaded or the traffic is free, the the chaels should be released ie users would stop usig these chaels. The mai operatio of CR odes is to periodically sese the eviromet ad estimate the chael utilizatio withi their sesig rage. Oce it receives the chael availability iformatio from ifrastructures, CR odes would respose to these ifrastructures with a set of available chaels ad relevat iformatio. The mai challeges that we facig durig the spectrum sharig are: How to decide the availability of a chael for a CR ode? (ii) How to select the optimum chaels for usage based o the iformatio provided by CR odes for the ifrastructure? The solutios are give by the determiatio of chael availability i CR odes ad i a ifrastructure ie CR odes seses the uused chaels ad a overloaded ifrastructure would receive the chael usage iformatio from its adjacet CR ode after request. This ca be proceeded by the SBAC Algorithm(Selectio of Best Available Chael). Thus, implemetig those CR odes ad deployig them as a etwork should be doe with the egotiatio betwee service providers based o the cost ad the etwork maagemet policy. This successful deploymet provides the risk ad the cost of operatig the CR etwork. It fouds to be a challegig task of deployig the CR etwork with the coexistece with the curret wireless etworks i a geographical situatio. It also icreases the cost ad complexity of wireless etwork maagemet. If CR odes commuicate with each other wirelessly, the it requires extra wireless resources ad it also icreases the overhead of wireless etworks. Algorithm : SBAC BS m (mobile ode s request to Base Statio) CR BS CR CR curret _ chael _ available _ list CR prob curret _ chael _ available / total _ chael ch_frq if frmax< ch_frq frmax frq if frmi> ch_frq frmi ch_frq iter frmax frmi cost t *60* c ch _ u (10* * prob) *log * iter 3 cost cu _ list[ cu _ cout ] ch _ u if ( cu _ cout! 0) cu _ list if maxi cu _ list maxi ch _ list chael _ maxi maxi 3.3 Simulatio Results of Spectrum Sharig If the ifrastructure of oe service provider is foud to be overloaded, it ss the requests to the adjacet CR Nodes regardig the availability of Chael. Chael availability ca be determied by sig service request to the BS. BS receives service request from the mobile odes ad it will s the chael request to the CR ode.fig.3 shows the BS before sig request to the CR Nodes.

4 4 Fig. 3 Before sig request to the CR Nodes CR ode receives the chael request ad ss the broadcast message to the adjacet CR ode. A eighbor CR ode receives the broadcast message ad also ss the available chael list to the BS. CR ode ad its eighbors, updates the chael availability list ad ss respose to the BS. Fig.4 shows the sig ad receivig the request ad respose from CR odes. Fig.5 Respose from CR odes Our scheme, idetifies 'available' chael list for each CR ode. Such a list shows which chael is available to use depig o the distace amog the CR ode ad High frequecy bad. Withi a give eighborhood, multiple CR odes may cot for access to oe or more of the available chaels. 4 PERFORMANCE METRICS I this sectio, we discuss about the performace metrics study the impact of spectrum sharig o the service providers iclude call blockig rate, system efficiecy, Reveue efficiecy, etc. Fig.4 Sig request to the CR odes ad respose from the CR odes. If BS receives the respose from the CR ode, it selects the available chael ad ss service reply with the allocated chael to the mobile odes. This shows the maximize utilizatio of a chael ad it offers several services such as iteret service, call service, multimedia service ad so o to the mobile odes. Fig.5 shows the maximum utilizatio of chael by sig the respose to the BS regardig the chael availability. 4.1 Call Blockig Rate The call blockig rate R is defied as the ratio of total blocked calls over total calls processed by all service providers ad correspods to: (t) R =lim processed (t) where total blocked calls at time t by all service providers is give by sp (t)= (t) ad the total calls processed is: sp processed i=1 i=1 (t)= (t) processed where sp is the umber of service providers. Here, the call would be blocked, if all the service providers are over-loaded. 4.2 System Efficiecy The spectrum efficiecy is defied as the umber of chaels used at time t for service provider, ad the umber of total chaels owed by the service provider. Higher spectrum efficiecy is aticipated compared to service provider, because the call blockig rate of a user-cetral system is lower; thus, more calls ca cotribute to the spectrum utilizatio.

5 4.3 System Utilizatio Efficiecy The Spectrum Efficiecy η (sp) s is defied as the ratio of average busy chaels over total chaels owed by service providers.

umber of total chaels owed by service provider. Higher Spectrum efficiecy is estimated because the call blockig rate is (sp) lower; thus more calls ca cotribute to the spectrum utilizatio. 4.

5 5 4.3 System Utilizatio Efficiecy The Spectrum Efficiecy η (sp) s is defied as the ratio of average busy chaels over total chaels owed by service providers. It correspods to ( sp) t t sp 1 busy ( ) ( ) s lim dt t 0 ( sp) N ch total ( t) (sp) where busy(t) is the umber of chaels used at time t for service provider ad N (sp) (sp) ch-total (t) is the total umber of total chaels owed by service provider. Higher Spectrum efficiecy is estimated because the call blockig rate is (sp) lower; thus more calls ca cotribute to the spectrum utilizatio. 4.4 Reveue(Cost) Efficiecy Withi the observatio time, cost is determied by the umber of processed calls ad the legth of call holdig time. We defie the metric c e to reflect the cost efficiecy. c e is the ratio of the cost eared withi the observatio time t over total iput traffic itesity for service provider sp, is defied as i c =c /E = α.e.t /E e () ( ) P α.t.η s ( i) (sp) = where α i is the uit price ($/secod/chael) for service provider sp ad c is the average icome withi the observatio time. 5 SIMULATION RESULTS I this sectio, we preset simulatio results o the performace of our proposed sesig framework. Chael assigmet mechaisms i the traditioal multi-chael wireless etworks typically select the 'Best' chael for a give trasmissio. I the proposed work, we are choosig the available chael with the high probability ad high-frequecy bad. To geerate utility performace measures, we assume: 1) Maximal five service providers share their spectrum, ad 100 odes are chose. Maximum limit of user per chael is 10. 2) Call arrival of each service provider is the heterogeeous process. 3) Traffic rates are correlated joitly-gaussia radom variables. 4) The ifrastructures f differet service providers are located at the same positio ad the cell radii is also the same. 5) The CR odes are preset at the vertices of the cells of the service providers. 6) Each CR ode has the ability of sesig its rage withi the coverage limits. 7) CR odes have the capability of detectig all the available chaels that are licesed to the other service providers. 8) Chael parameters such as iterferece level, the probability of beig available for a give time period ad cost are the same for all available chaels. We coduct simulatios to verify the potetial of the call arrival rate for differet service providers i terms of utility performace measures. Fig.6 Mea arrival vs Active users Fig.6 shows that, the active users are reduced whe there is a higher traffic i call blockig of differet service providers are highly correlated. If the call blockig is reduced amog the correlated service providers, the there would be icrease i the active users. Fig.7 Mea arrival vs Call blockig Fig.7 shows that at higher traffic rates, the call blockig rate is higher whe the traffic rates of differet service providers are highly correlated. Fig.8 shows that as the call blockig is reduced the the traffic load for five service providers is miimum.

6 Fig.8 Mea arrival vs Traffic Load As the mea call arrival icreases the chael utilizatio also icreases as i the Fig.9.

10 shows that, at high-traffic rates, the system efficiecy is lower whe the traffic rates of differet service providers are highly correlated. Whe the correlatio is lower, based o Fig.

10 Mea arrival vs System Efficiecy 6 CONCLUSION The Spectrum assiged to differet service providers is ot properly utilized with the same frequecy.

6 6 Fig.8 Mea arrival vs Traffic Load As the mea call arrival icreases the chael utilizatio also icreases as i the Fig.9. Higher Spectrum efficiecy is estimated because the call blockig rate is lower; thus, more calls ca cotribute to the spectrum utilizatio. Fig.9 Mea arrival vs Spectrum Efficiecy Fig.10 shows that, at high-traffic rates, the system efficiecy is lower whe the traffic rates of differet service providers are highly correlated. Whe the correlatio is lower, based o Fig.10, as the dropped calls decrease, thus, the total processed calls icrease. The system efficiecy decreases whe the traffic rate is beyod the system capacity. Fig.10 Mea arrival vs System Efficiecy 6 CONCLUSION The Spectrum assiged to differet service providers is ot properly utilized with the same frequecy. So some service providers may try to use the allocated spectrum fully, ad eve they eed more spectrums ad, which may ot be used by service providers fully. This paper provides a offer (way) to utilize ad also to share the licesed spectrum amog the service providers if they are uder utilized. Here, we discussed about the operatios of CR odes ad ifrastructures of service providers for spectrum sharig. From the proposed algorithm, we ca sese the rage of CR ode ad we decide ad select the optimal chael for spectrum sharig. I additio to this work, we also defie some performace metrics for spectrum sharig amog service provides i order to efficiet spectrum utilizatio. This techique removes the eed of sesig spectrum i each user. Thus, it reduces the cost, complexity ad battery power cosumptio of user devices. REFERENCES [1] R. Kaiezhil, Dr.C. Chadrasekar, S. Nithya Rekha Chael Selectio for Spectrum Sharig usig CR Nodes Iteratioal Proceedigs of Computer Sciece ad Iformatio Techology, ISSN: X, vol.20, pp , [2] R. Kaiezhil, Dr. C. Chadrasekar, Performace Evaluatio of QoS Parameters i Dyamic Spectrum Sharig for Heterogeeous Wireless Commuicatio Networks, Iteratioal Joural of Computer Sciece Issues (IJCSI), Volume 9, Issue 1, pp , Ja [3] R. Kaiezhil, Dr.C. Chadrasekar, S. Nithya Rekha, Dyamic Spectrum Sharig for Heterogeeous Wireless Network via Cogitive Radio, Proceedigs of the IEEE Iteratioal Coferece o Patter Recogitio, Iformatics ad Medical Egieerig, March 21-23, 2012, pp [4] R. Kaiezhil, Dr. C. Chadrasekar, S. Nithya Rekha, Performace Evaluatio of QoS Parameters i Spectrum Sharig usig SBAC Algorithm, IEEE Iteratioal Coferece o Advaces i egieerig, Sciece ad Maagemet (IEEE-ICAESM 2012), Nagapattiam, March 30-31, 2012(accepted). [5] R. Kaiezhil, Dr.C Chadrasekar, S. Nithya Rekha, Chael selectio for Selective Sesig usig Cogitive Radio Nodes, Itera-

7 tioal Joural of Computer Applicatios(IJCA), , Volume 39 No.3, pp , February [6] Simo Hayki, Cogitive Radio: Brai-Empowered Wireless Commuicatios, Life Fellow, IEEE Joural o Selected Areas i Commuicatios, Vol. 23 No. 2, pp , Feb [7] Ia F. Akyildiz, Wo-Yeol Lee, Kaushik R. Chowdhury, CRAHNs: Cogitive radio ad hoc etworks, Elsevier Ad Hoc Networks, pp , [8] V. Brik, E. Rozer, S. Baarjee, P. Bahl, DSAP: a protocol for coordiated spectrum access i: Proc. IEEE DySPAN 2005, pp , November [9] Akyildiz, I. F., Lee, W.-Y., Vura, M. C., ad Mohaty, S., Next geeratio / dyamic spectrum access / ogitive radio wireless etworks: a survey, Computer Networks (Elsevier), vol. 50, pp , September [10] Buddhikot, M., Kolody, P., Miller, S., Rya, K., ad Evas, J., DIM- SUMNet: ew directios i wireless etworkig usig coordiated dyamic spectrum access, i Proc. IEEE WoWMoM 2005, p. 7885, Jue [11] Ileri, O., Samardzija, D., ad Madayam, N. B., Dyamic property rights spectrum access: Flexible owership based spectrum maagemet, i Proc. IEEE DySPAN 2007, April [12] Lee, W.-Y. ad Akyildiz, I. F., Joit spectrum ad power allocatio for iter-cell spectrum sharig i cogitive radio etworks, i Proc. IEEE DySPAN 2008, Chicago, IL, USA, October [13] Lee, W.-Y. ad Akyildiz, I. F., Spectrum sharig framework for itrastructure-based cr etworks, submitted for publicatio, [14] Etki, R., Parekh, A., ad Tse, D., Spectrum sharig for ulicesed bads, IEEE Joural o Selected Areas i Commuicatios, vol. 25, pp , April [15] Huag, J., Berry, R. A., ad Hoig, M. L., Spectrum sharig with distributed iterferece compesatio, i Proc. IEEE DySPAN 2005, pp , November [16] Zhag, L., Liag, Y., ad Xi, Y., Joit beamformig ad power allocatio for multiple access chaels i cogitive radio etworks, IEEE Joural o Selected Areas i Commuicatios, vol. 26, pp , Jauary [17] Jia, J., Zhag, Q., ad She, X., HC-MAC: a hardware costraied cogitive MAC for effiiciet spectrum maagemet," IEEE Joural o Selected Areas i Commuicatios, vol. 26, p , Jauary [18] Kevi Fall, Kaa Varadha., The s Maual, The VINT Project, May 9, [19] Teerawat Issariyakul, Ekram Hossai, Itroductio to Network Simulator NS2 ISBN: e-isbn: , Spriger

Performance Analysis of Channel Switching with Various Bandwidths in Cognitive Radio

Performance Analysis of Channel Switching with Various Bandwidths in Cognitive Radio Performace Aalysis of Chael Switchig with Various Badwidths i Cogitive Radio Po-Hao Chag, Keg-Fu Chag, Yu-Che Che, ad Li-Kai Ye Departmet of Electrical Egieerig, Natioal Dog Hwa Uiversity, 1,Sec.2, Da-Hsueh