Node Deployment Coverage in Large Wireless Sensor Networks

Transcription

1 Node Deploymet Coverage i Large Wireless Sesor Networks Gayarti Devi Professor, Dept. of CSE, ABIT, Cuttack , Odisha, Idia. Rajeeb Sakar Bal Seior Lecturer, Dept. of CSE, ABIT, Cuttack , Odisha, Idia. Abstract Curretly, wireless sesor etwork (WSN) is beig applied i a lot of differet applicatios. To a large extet the effectiveess of the WSNs depeds o the exposure provided by the sesor deploymet proposal. There are differet deploymet demads i differet eviromet. Firstly, we cover the existig deploymet method of sesor odes (SNs) i WSN. We ivestigate radom ad determiistic sesor ode (SN) deploymets for large-scale WSNs i the performace metric coverage. We cosider three regular fields circle, rectagle ad square for a uiform radom, a square grid, a patter-based Tri-Hexago Tilig (THT) ad Hexago Tilig (HT) ode deploymet. Fially, we aalyze tradeoffs betwee these performaces metric for each deploymet pla to show which approach is preferable uder what factors, e.g., the umber of SNs. Idex Terms Wireless Sesor Network (WSN), Sesor Node(SN), Tri-Hexago Tilig (THT), Hexago Tillig(TH). 1. INTRODUCTION A wireless sesor etwork ca be defied as a wireless commuicatio cosistig of coutless distributed devices usig sesors to moitor physical or evirometal coditios ad also trackig objects show i fig.1. of the WSN is performed ad defied as the WSN. Wireless etwork of sesors spread over a area. Such kid of etworks is usually used for moitorig applicatios iclude idoor or outdoor ad the trackig applicatios iclude trackig objects. The WSN is built of odes from a little to several hudreds or eve thousads where each ode is coected to oe (or sometimes several) sesors. The data is forwarded probably via multiple hops to a sik (occasioally deoted as cotroller or moitor) that ca use it locally or it is coected to ew etworks (e.g., the Iteret) through a gateway. The odes ca be statioary or movig. They ca be resposive of their positio or ot. They ca be homogeeous or ot [1] show i fig.. Fig. 1. The overall view of WSN. It is capable to provide the iformatio logged by the sesors to remote locatios ad it ca do so without the use of a big ad composite wired etworks. The whole setup of data beig set from the sesors distributed across a wide area by meas Fig.. The Overall view of WSN ad its Architecture. This is a traditioal sigle-sik WSN show i fig..the sigle-sik system suffers from the lack of scalability by icreasig the umber of odes, the quatity of data gathered by the sik icreases ad oce its ability is reached, the etwork size caot be augmeted. I additio, for reasos related to MAC ad routig aspects, etwork performace caot be measured idepedet from the etwork size show i fig. 3. ISSN: EverSciece Publicatios 19

2 1.1 Hardware of Sesor Node Fig. 3. The Sigle sik sceario i WSN. A more geeral sceario icludes multiple siks i the etwork show i fig.4. Give a level of ode solidity, a larger umber of siks will decrease the possibility of isolated groups of odes that caot sed their data due to ufortuate sigal propagatio coditios. I stadard, a multiple-sik WSN ca be scalable while this is evidetly ot true for a sigle sik etwork. Still, a multi-sik WSN does ot represet a isigificat extesio of a sigle sik cotaier for the etwork trick. I umerous cases odes sed the data composed to oe of the siks chose amog may which forward the data to the gateway headed for the fial user. From the protocol poit of view meas that a choice ca be doe based o a suitable criterio that could be least delay, greatest throughput, least umber of hops, etc. Thus, the existece of multiple siks esures better etwork performace with respect to the sigle sik case, but the cotact protocols must be more complex ad should be desiged accordig to appropriate criteria []. Fig. 4. The Multi sik sceario i WSN. Lastly, the costructio of WSNs may be categorized ito three classes are as hardware, wireless etworkig, ad applicatios. Fig. 5. The compoets of sesor ode i wireless sesor etwork. The geeral architecture ad the major compoets of a wireless sesor device (ode) are show i fig. 5. Geerally, a ode is composed of four basic compoets: a sesig uit, a processig uit, a trasceiver uit ad a power uit. Moreover, additioal compoets ca also be itegrated ito the sesor ode depedig o the applicatio. These compoets are i show i fig. 5 iclude: a locatio fidig system, a power geerator, ad a mobilizer. Sesig uit: The sesig uit is the mai compoet of a sesor ode that distiguishes it from ay other embedded system with commuicatio capabilities. Processig uit: The processig uit is the mai cotroller of the wireless sesor ode through which every other compoet is maaged. The processig uit may cosist of a o-board memory or may be associated with a small storage uit itegrated ito the embedded board. Trasceiver uit: The commuicatio betwee ay two wireless sesor odes is performed by the trasceiver uits. Power uit: Oe of the most importat compoets of a wireless sesor ode is the power uit. Usually, battery power is used, but other eergy sources are also possible. Each compoet i the wireless sesor ode is powered through the power uit ad the limited capacity of this uit requires eergy-efficiet operatio for the tasks performed by each compoet. Locatio fidig system: Most of the WSN applicatios, sesig tasks, ad routig techiques eed kowledge of the physical locatio of a ode. Thus, it is commo for a sesor ode to be equipped with a locatio fidig system. ISSN: EverSciece Publicatios 0

3 Mobilizer: A mobilizer may sometimes be eeded to move sesor odes whe it is ecessary to carry out the assiged tasks. Mobility support requires extesive eergy resources ad should be provided efficietly. Power geerator: While battery power is mostly used i sesor odes, a additioal power geerator ca be used for applicatios where loger etwork lifetime is essetial. For outdoor applicatios, solar cells are used to geerate power. Similarly, eergy scavegig techiques for thermal, kietic, ad vibratio eergy ca also be used [3]. 1. Wireless etworkig I [4]. a etwork maagemet system desiged for WSNs must take ito accout the uique properties of WSNs. The followig criteria are used to estimate the WSN maagemet systems: Lightweight operatio. A system should be able to ru o sesor odes without cosumig too much eergy or iterferig with the operatio of the sesor odes. Lightweight operatio prologs etwork lifetime. Robustess ad fault tolerace. WSNs are proe to etwork dyamics such as dropped packets, odes dyig, becomig discoected, powerig o or off, ad ew odes joiig the etwork. A maagemet system should be resiliet to etwork dyamics by recofigurig the etwork as required. Adaptability ad resposiveess. A system should be able to retrieve ad adapt to the curret etwork states or chagig etwork coditios icludig chages i etwork topology, ode eergy level, ad the coverage ad exposure bouds of WSNs. Miimal data storage. A data model used to represet maagemet data must be extesible ad able to accommodate iformatio eeded to perform the maagemet fuctios, but must also respect the memory costraits of WSNs. Scalability. A system should operate efficietly i ay etwork size. 1.3 WSN Applicatios The recet advaces i WSN, the applicatios ca be classified ito two categories: moitorig ad trackig show i fig.6. The moitorig applicatios iclude idoor/outdoor evirometal moitorig, health ad welless moitorig, power moitorig, ivetory locatio moitorig, factory ad process automatio, ad seismic ad structural moitorig. The trackig applicatios iclude trackig objects, aimals, humas, ad vehicles [5]. Fig. 6. The Applicatios of WSN.. NODE DEPLOYMENT All the related works that have bee doe by other researchers that are related to the curret research problem should be summarized i this sectio. Times New Roma fot with size 10 must be used i this sectio. Sub topic should be writte as A Wireless Sesor Network (WSN) ca be self-possessed of homogeeous or heterogeeous odes, which possess the same or differet commuicatio ad computatio capabilities, respectively. Although some works cosider heterogeeous odes, a lot of existig works ivestigate ode placemet i the circumstace of homogeeous WSNs. Less complexity ad a better maageability are the most valuable effects of homogeeity. Hece, we cosider homogeeous odes i WSNs. These odes ca be deployed over a etwork i radom or determiistic maer. While the radom ode deploymet is preferable i may applicatios, if possible, other deploymets should be ivestigated sice ufortuate ode deploymet ca raise the complexity of other problems i WSNs..1 Classificatio of Deploymet I WSN, the deploymet of sesor odes maily cocetrated i the static ad the dyamic deploymet. I static deploymet, accordig to the optimizatio strategy, it chooses the best locatio ad the locatio of the sesor odes has o chage i the life spa of the WSN. I curret, this deploymet icludes the determiistic deploymet ad the radomly deploymet. The self-orgaized algorithms are proposed by wakig up some sesor odes or makig some sesor odes sleep. This algorithm used the essetial potetial field to make the sesor odes move positios ad chage ISSN: EverSciece Publicatios 1

4 directios automatically i detectio area based o directioal sesig model. With the coclusio of multiple prior coverage of hot targets which eeded higher quality ecessities, the algorithm could maximize the coverage rate throughout the detectio area. As per Poisso distributio, also it proposed a strategy of WSN odes radomly deploymet. I this strategy, first established the model of WSN ode distributio ad fid the relatioship betwee the profit of coverage area ad the odes desity of target area. Fially, fid the best rage of odes desity to get the optimal deploymet [6]. I dyamic deploymet, It may be backed to the deploymet of the robot i WSN. I order to make the sesor etworks get the most performace, sesor odes eed automatically shift to proper locatio, ad the start to work. I the radom deploymet, amely radomly throw odes firstly, ad the usig a selectio of optimizatio algorithm for deploymet optimizatio. Such as Virtual force algorithm, virtual force orieted particles algorithm, simulated aealig algorithm, particle swarm optimizatio algorithm ad simulated aealig geetic algorithm [6].. Classifyig Coverage Schemes I [6], the developmet of eergy efficiet schemes itegratig coverage ad coectivity for WSN which depeded o the coverage objectives ad applicatios show i fig. 7. I WSN, the coverage ca be classified ito three categories area coverage, poit coverage, ad path coverage. Area coverage: I this coverage, the mai objective of the sesor etwork is to cover (moitor) a regio (the group of all space poits withi the sesor field), ad each poit of the regio eed to be moitored. Poit coverage: I this coverage, the objective is to cover a set of poit (target) with kow positio that eed to be moitored. Agai i this coverage scheme, it focuses o determiig sesor odes exact positios where guaratee efficiet coverage applicatio for a limited umber of immobile poits (targets). Geerally, it ca be solved as a particular case of the area coverage problem whe sesor odes umber may leave out of accout. Path coverage: I this coverage, the goal is to miimize or maximize the probability of udetected peetratio through the regio. I brief, we itroduce how to evaluate the coverage performace of a regio covered by WSN. Give a set of sesors deployed i a moitored regio, coverage-evaluatig problem is to determie if all poits i the regio is suitably k- covered, i the sese that all poit i the target area is eclosed by at least k sesors, where k is a give parameter. Fig. 7. The sesor ode surroud by o-peetrate obstacle.. K-Coverage I [6], k-coverage refers to the miimum k-coverage. A etwork is said to have k-coverage if each poit i it is covered by at least k sesors. As per [6] formulate k-coverage of the regio for mostly sleepig large-scale WSNs. Also, it is claimed that the critical value of the expressio p / log( p) where is the umber of sesors ad p is the probability of active sesors, is 1 form a sufficiet coditio for k-coverage. Although miimum k-coverage is valuable for surveillace kid of applicatios ad other kids of coverage such as a average k-coverage or the most k- coverage, may be more meaigful for other WSN applicatios. Moreover, it seems usuitable to measure k- coverage for performace cotrast due to its idividual iterest i the miimum coverage area of the etwork. For this reaso we ivestigate the relative frequecy of the exactly k-covered poits i ode deploymet strategies. 3. DEPLOYMENT MODEL I this paper, we examie three competitors of ode deploymet for a WSN: a uiform radom, a square grid, ad a Tri-Hexago Tilig (THT) or Hexagoal Tilig (TH). Sice the precedece of performace metrics varies i applicatio specific WSNs, it is valuable to ivestigate a set of them. For the duratio of desig phase of WSNs, the desiger kows the umber of sesor odes which are deployed i a give field i either radom or determiistic maer. A circular field with radius R is measured i our experimets. We itroduce three ode deploymet strategies together with their characteristics. A Uiform Radom: I [7] the uiform radom deploymet, each of the sesors has equal possibility of beig placed at ay poit iside a give field, as show i fig, 8, 9 ad 10. Therefore, the odes are dotted o locatios which are ot kow with certaity. I geeral, a uiform radom ISSN: EverSciece Publicatios

5 deploymet is assumed to be easy as well as cost-effective. Also, it is claimed that a uiform radom deploymet outperforms both the grid ad the Poisso distributio deploymets for k-coverage. Square Grid: I WSN, we cosider a grid based deploymet is a good deploymet ad especially for the coverage performace show i fig. 8(b). A grid deploymet of sesors i a circular field, where each of the grid poits hosts a sesor. The approximate legth of a uit square, d', ca be calculated as follows: The approximate area of a uit square with legth d' ca be computed by dividig the etire area of a give field havig radius R with the umber of cells k. We do ot kow the value of k, but it is approximately equal to ( 1) for the square grid i equatio 1.From this relatio, we derive Equatio 1 for, the sesig radius. However, sice we cosider a iitial adjustmet for a startig poit, Equatio 1 caot be applied directly. Accordig to simulatio results, Equatio gives more precise values tha Equatio 1. Although we use these equatios to fid out the (i.e., the legth of a square, d') give ad R, this formula allows the estimated calculatio of ay oe parameter out of,, ad R give the other two parameters. Fig. 8(a). Radom Fig. 8(b). Grid Fig. 8(c). Tri-Hexagoal Fig.8. The Node deploymet for Circle field. Durig the desig phase of WSNs, we kow the umber of sesor odes which are deployed i a give field i either radom or determiistic fashio. We cosider aother two field oe is rectagle ad other oe is square field with sides is cosidered i our experimets show i figure ad formula give i table 1. R ( 1) (1) R () Fig. 9(a). Radom Fig. 9(b). Grid Fig. 9(c). Hexagoal Fig. 9. The Node deploymet for Rectagle field. Tri-Hexago Tilig (THT): The third approach is based o tilig. A tilig is the coverig of the etire plae with figures which do ot partly cover or disappear ay gaps. The tiligs are also sometimes called tessellatios. Amog differet tilig we use a semi-regular tilig (which has exactly eight differet tilig) where each vertex uses the same set of usual polygos. A usual polygo has the same side legths ad iterior agles. A semi-regular tilig that uses triagle ad hexago i the two dimesioal plae called Tri-Hexago Tilig. The ame comes from goig aroud a vertex ad listig the umber of sides each regular polygo has show i fig. 8(c). I a way related to the square grid, a approximate formulatio for ca be foud for THT. This approximate solutio ca be computed usig Equatio 3 [7]. Fig. 10(a). Radom Fig. 10(b). Grid Fig. 10(c). Hexagoal Fig. 10. The Node deploymet for Square field. 4R 3 3 (3) ISSN: EverSciece Publicatios 3

6 Table 1. Fot sizes of headigs. Table captios should always be positioed above the tables. Field Sides Square Grid Hexagoal Circular R Rectagular a, b R 4R r se ( 1) 3 3 R ab ( ab) Fig. 1(a). Radom Deploymet i Circular Square A bb ( bb) 4. RESULTS I this sectio, we perform a performace evaluatio for our three ode deploymet strategies. The primary factors for all experimets are: the umber of siks, the umber of odes, ad the sesig regio. For each deploymet, the odes are distributed over a circular, rectagle, ad square field shapes ad siks are positioed at the Ceter of Gravity. Usig the Java codig to display three shapes with ode deploymet strategies show i fig. 11. Fig. 1(b). Radom Deploymet i Rectagle. Fig. 1(c). Radom Deploymet i Square Fig. 1. The Radom Deploymet i three differet fields (Circular, Rectagular ad Square). Fig. 11. The Meu desig for three fields (Circular, Rectagular ad Square). Further, we build assumptios to compare our three strategies. Sice we do ot cosider boudary coditio i radom deploymet, square grid ad Tri-hexagoal tilig. Therefore, i the case of the radom deploymet for circular, rectagle ad square fields, we do a systematic samplig over the areas which are differig show i fig. 1, 13 ad 14. Fig. 13(a). Square Grid Deploymet i Circular. ISSN: EverSciece Publicatios 4

7 Fig. 14(b). Tri-hexagoal deploymet i Rectagular. Fig. 13(b). Square Grid Deploymet i Rectagle. Fig. 14(c). Tri-hexagoal deploymet i Square. Fig. 14. The Square Grid Deploymet i three differet fields (Circular, Rectagular ad Square). 5. CONCLUSION We study the complete process of wireless sesor etwork about the deploymet are summarized ad aalyzed. I the result sectio, we proposed three fields with three strategies. It seems that a radom deploymet i three fields is more or less comparable to a square grid for a large-scale etwork. Fig. 13(c). Square Grid Deploymet i Square Fig. 13. The Square Grid Deploymet i three differet fields (Circular, Rectagular ad Square). Fig. 14(a). Tri-hexagoal deploymet i Circular. REFERENCES [1] I.F.Akyildiz, W.Su, Y. Sakarasubramaiam, E. Cayirci, Wireless Sesor Networks: A Survey, Computer Networks (Elsevier) Joural, March 00. [] Chiara Buratti, Adrea Coti, Davide Dardari, Roberto Verdoe,"A Overview o Wireless Sesor Networks Techology ad Evolutio", , Sesors, 31 August 009. [3] Ia F. Akyildiz, Mehmet Ca Vura, "Wireless Sesor Networks", 010 Joh Wiley & Sos Ltd. [4] Wiie Louis Lee, Amitava Datta, ad Rachel Cardell-Oliver, "Network Maagemet i Wireless Sesor Networks" CSSE Techical Report UWA-CSSE Jue 006 WiMS. [5] Jeifer Yick, Biswaath Mukherjee, Dipak Ghosal "Wireless sesor etwork survey, Computer Networks 5 (008) [6] Haitao Zhag ad Cuipig Liu "A Review o Node Deploymet of Wireless Sesor Network" IJCSI Iteratioal Joural of Computer Sciece Issues, Vol. 9, Issue 6, No 3, November 01. [7] Wit Yi Poe, Jes B. Schmitt, "Node Deploymet i Large Wireless Sesor Networks: Coverage, Eergy Cosumptio, ad Worst-Case Delay", AINTEC 09, November 18 0, 009, Bagkok, Thailad. ISSN: EverSciece Publicatios 5