An Improved Weighted Centroid Localization Algorithm

Transcription

1 Internatonal Journal of Future Generaton Communcaton an Networng Vol.6, No.5 (203), pp An Improve Weghte Centro Localzaton Algorthm L Bn, Dou Zheng*, Nng Yu an Ln Yun College of Informaton an Communcaton Engneerng Harbn Engneerng Unversty Harbn, Chna ouzheng@hrbeu.eu.cn Abstract As one of the most mportant applcatons of wreless sensor networ, postonng technology has become a extremely hot research fle. Tang nto account of the poston of beacon noess, they mae an effect on the postonng accuracy of unnown noes. Therefore, n ths paper, t proposes a new weghte centro localzaton algorthm base on the tratonal algorthm. For the selecton of weght, the stance between beacon noes an unnown noes, an the sle length of the trangle are use to forme as the weghte factor. Experment Smulaton results show that ths algorthm ncrease localston accuracy than that of tratonal algorthm. Keywors: Wreless Sensor Networs;Postonng Technology;Weghte Centro Localzaton;Noes Poston. Introucton Wreless sensor networs (WSN) s consere to be one of the most mportant technologes n the 2st century that have a huge nfluence on the proucton an lfe. It has broa prospects n the mltary, envronment, health, famly, an other commercal fels. Noe localzaton has been a topc of actve research n recent years. The postonng algorthm can be ve nto range-base an range-free. Range-base localzaton algorthm nee to measure the relatve stance or orentaton between aacent noes at frst, then use the trangulaton metho, trlateraton or maxmum lelhoo estmaton metho to calculate the unnown noe's locaton. Rangng technques use commonly nclue the receve sgnal strength ncator (RSSI) metho, the tme of arrval (TOA) [] metho, the tme fference of arrval (TDOA) an the angle of arrval(aoa) [2] metho. Range-free localzaton algorthm usng the atonal nformaton between noes to estmate the stance to calculaton noe locatons. It nclues centro algorthm, DV-HOP algorthm, Amorphous locaton algorthm [3] an APIT algorthm [4]. 2. Tratonal Centro Localzaton Algorthm Moel Tratonal centro localzaton algorthm s one of the Range-free localzaton algorthms. The beacon noes wthn the wreless communcaton of unnown noe form a polygon. We regar the centro of the polygon as the unnown noe s locatons. It s shown n Fgure. ISSN: IJFGCN Copyrght c 203 SERSC

2 Internatonal Journal of Future Generaton Communcaton an Networng Vol.6, No.5 (203) D E C F B A Fgure. Schematc of Centro Localzaton Algorthm For sx gven anchors: A( x a,ya), B( x b,yb), C( x c,yc), D( x,y), E( x e,ye), F( x,y ), the coornates of the unnown noe can be calculate as follows : f f x a+x b+x c+x +x e+xf x= () 6 y a+y b+y c+y +y e+yf y= (2) 6 Centro algorthm can be acheve only wth the noe connectvty, calculaton metho s smple, fast, ths metho s smple. However, we nee more beacon noe to mprove the accuracy. 3. Weghte Centro Localzaton Algorthm Theoretcal proof, the larger the RSSI value that the unnown noe receve from the beacon noe, the smaller the stance between two noes, then the beacon noes have greater nfluence on unnown noe, an vce versa. In aton, when the beacon noes form trangles or polygons, ts postonng accuracy s hgher than that any eployment n the area to be postone. In orer to reflect the effect that the beacon noes poston has on the unnown noes, we use the weghte centro localzaton algorthm. There are a total of n beacon noes n the wreless range of the unnown noe N.Any three of these beacon noes can form a trangle. The moel shown n Fgure 2 llustrates the algorthm. 46 Copyrght c 203 SERSC

3 Internatonal Journal of Future Generaton Communcaton an Networng Vol.6, No.5 (203) Fgure 2. Schematcs of Weghte Centro Localzaton Algorthm In lterature [6], the algorthm taes the recprocal of the sum of stance as the weght. As shown n the followng formulas: x = xa ( ) + xb ( ) + xc ( ) a+b b+c a+c a b b c a c (3) y ya ( ) + yb ( ) + yc ( ) a+b b+c a+c = a b b c a c Where ( x,y) s the coornates of the noe N that estmate by the trangle, ABC. a, b, c are the stance of the noe N to noes A, B, C. ( xa, ya), ( xb, yb) an ( x, y ) are the ntersecton pont of the three crcles by usng trlateraton. c c (4) x = y = [ x ( )] = a() + b() +c() [ ] + + = a() b() c() [ y ( )] = a() + b() +c() [ ] + + = a() b() c() (5) (6) Where, ( x, y ) s the approxmate coornates of the unnown noe, 3 =C n, x,, c() y s are the the coornate of unnown noe calculate by the trangle, a(), b() stance of the unnown noe to the three vertces of the trangle. The weght factor reflect the smaller the stance between unnown noe an beacon noes, the greater the nfluence that beacon noes have on the unnown noe. Copyrght c 203 SERSC 47

4 Internatonal Journal of Future Generaton Communcaton an Networng Vol.6, No.5 (203) In lterature [7], the algorthm taes the sum of the recprocal of the stance as the weght. The weght factor n formulas (3), (4), (5), (6), become + a, + b b, c +, + +. a c a() b() c() In lterature [6, 7],the propose algorthm taes avantage of the stance nformaton between the noes, but the two weghts not reflect the nfluence that beacon noes eployment have on the unnown noe. Lterature [8] presents a new metho. It regars the angles as the weghts. It s shown max - mn as w= - α α π trangle. Then the coornates of the unnown noe can be calculate by the followng formulas., where αmax an α mn are the maxmum an mnmum angle of a x= y= = = = wx = w wy w (7) (8) Where x an y are the coornate of unnown noe calculate by the trangle. In lterature [9], the algorthm taes the angles an the ege length of a trangle as the weghtng factor. It s shown as follows: β = A B C (9) AB+ AC L= BC W= β +L (0) () Tae W nstea of w nto the formulas (7) an (8), we can get the approxmate coornates of the noe. The postonng algorthms propose n lterature [8, 9] o not conser the stance mpact on the postonng. 48 Copyrght c 203 SERSC

5 Internatonal Journal of Future Generaton Communcaton an Networng Vol.6, No.5 (203) 4. An mprove Weghte Centro Localzaton Algorthm Ths paper presents a new weght whch compensates for the lac of the weght presente n the above algorthms. Calculate as follows: xa ( ) + xb ( ) + xc ( ) a+b b+c a+c x = a b b c a c ya ( ) + yb ( ) + yc ( ) a+b b+c a+c y= a b b c a c (2) (3) y= x= [ x ( )] = a() b() c() L [ ] L = a() b() c() [ y ( )] = a() b() c() L [ ] L = a() b() c() (4) (5) Where L s the rato of the largest se length an the smallest se length of the trangle, Other parameters have the same efnton wth (3), (4), (5), (6). 5. Implementaton Process of Algorthm The mprove weghte centro localzaton algorthm can be calculate as follows: Frst, choose three ponts arbtrarly from all of the beacon noes n the wreless communcaton range of the unnown noe. Suppose the number of beacon noes s m, the 3 trangle s, then we can get =C m. Secon, calculate the estmate coornates of the unnown noe etermne by -trangle, =, 2, 3,,. (See Equaton (9), (0)). Thr, calculate the average coornates(x0, y0) an stanar evaton σ of the set of coornates. Flter the ata that has a large error whch excee 3σ. Fourth, seconary weghtng to the processe coornates to obtan the approxmate coornates of the unnown noe. (See Equaton (), (2)). Copyrght c 203 SERSC 49

6 Internatonal Journal of Future Generaton Communcaton an Networng Vol.6, No.5 (203) 6. Algorthm Smulaton Smulatonng the weghte centro localzaton algorthm n the MATLAB platform. Where, the regon to be etecte s 64m 64m,the number of unnown noes s 30 an beacon noes s 7.The followng fgures showe are the results obtane by one tme of smulaton. Fgure 3. Smulaton Dagram of Improve Algorthm Fgure 4. Smulaton Dagram n Lterature [6] Fgure 5. Smulaton Dagram n Lterature 50 Copyrght c 203 SERSC

7 Internatonal Journal of Future Generaton Communcaton an Networng Vol.6, No.5 (203) Fgure 6. Smulaton Dagram n Lterature [8] Fgure 7. Smulaton Dagram n Lterature [9] After 00 tmes smulaton, we can obtane the average error of each algorthm, where the mprove algorthm s 2.33m,the algorthm propose n lterature [6] s 0.4m, the algorthm propose n lterature [7] s 2.54m,the algorthm propose n lterature [8] s 2.87m,the algorthm propose n lterature [9] s 2.57m. 7. Concluson In ths artcle, we put forwar a new weghte centro localzaton algorthm to mprove the accuracy. We tae the stance between noes an the se length of trangles forme by beacon noes as the weghtng factor. Smulaton results show that the new algorthm has mprove the postonng accuracy greatly. Acnowlegements Ths wor s supporte by the Naton Nature Scence Founaton of Chna No , Nature Scence Founaton of Helongang Provnce of Chna No. QC202C069 an the Copyrght c 203 SERSC 5

8 Internatonal Journal of Future Generaton Communcaton an Networng Vol.6, No.5 (203) Funamental Research Funs for the Central Unverstes No. HEUCFZ29, No. HEUCF3080 an No. HEUCF3087. References [] W. Meng, X. Wenong an X. Lhua, Dstrbute algorthm for hybr TOA/DOA-base source localzaton, Inustral electroncs an applcatons, IEEE, (20), pp [2] Y. Sun Lee, J. Woo Par an L. Baroll, A localzaton algorthm base on AOA for a-hoc sensor networs, Moble nformaton systems, (202), pp [3] L. Yng, An aaptve mult-hop stance localzaton algorthm n WSN, Manufacturng Automaton, vol., (20), pp [4] W. Jzeng an J. Hongxu, Improvement on APIT localzaton algorthms for wreless sensor networs, Networs securty, wreless communcatons an truste computng, IEEE, (2009), pp [5] H. Yan-l, Research on Centro Localzaton Algorthm for Wreless Sensor Networs base RSSI, Computer Smulaton, vol. 5, no. 5, (20), pp [6] C. We-e, L. Wen-feng, S. Heng an Y. Bng, Weghte Centro Localzaton Algorthm Base on RSSI for Wreless Sensor Networs, Journal of Wuhan Unversty of technology, vol. 2, no. 4, (2006), pp [7] Y. Hu-xa, Weghte centro localzaton algorthm wth weght correcte base on RSSI for wreless sensor networ, Electronc Test, vol., no., (202), pp [8] L. Jn, Research on Range-base Localzaton Algorthm of Wreless Sensor Networs, Aeronautcal Computng Technque, vol. 6, no., (2009), pp [9] W. Wen-an, L. Jn, L. He-ln an K. Bng, An mprove weghte trlateraton localzaton algorthm, Journal of Zhengzhou Unversty of Lght Inustry(Natural Scence), vol. 3, no. 6, (202), pp [0] X. Chen, J. He, B. Le an T. Jang, An Improve Localzaton Algorthm of Noes n Wreless Sensor Networ, Journal of Networs, vol. 7, no., (202), pp [] H. Lv, Y. Wu an B. Peng, A localzaton metho of wreless sensor networ base on two-hop focus, Procea Engneerng, vol. 5, (20), pp [2] M. V. Ramesh, P. L. Dvya, R. V. Kularn an R. Mano, A Swarm Intellgence base strbute localzaton technque for Wreless Sensor Networ, IEEE Avances n Computng, Communcatons an Informatcs, (202), pp Copyrght c 203 SERSC