A Secant Location Estimation Algorithm for Wireless Sensor Networks

Transcription

1 A Secant Location Etimation Algorithm for Wirele Senor Network Tung-Han Lee, Yu-Jhong Fu, Lin-huang Chang Department of Computer an Information Science National Taichung Univerity Taichung, Taiwan, R.O.C. Abtract. The Location-bae ervice on Wirele Senor Network, mainly epen on the Receiver Signal Strength (RSS to meaure the itance between the ener an receiver. The location unaware enor noe en beacon perioically which capture by anchor enor noe. However, the etimate itance may incur large error, an will rik the localization preciion. A ecant location etimation algorithm in wirele enor network i propoe in thi paper. The mobile anchor noe can calculate the poition for poition unaware enor noe through the propoe location etimation cheme in the three-imenional coorinate. The imulation reult how the propoe algorithm i vali an effective. Keywor: mobile anchor, three-imenional localization, wirele enor network Introuction Wirele enor network have been ientifie a one of the mot prominent technologie for the t century. The embee wirele ytem are wiely ue in variou fiel, uch a inutrial application, life an builing automation, multimeia application. Therefore, there have been much reearch interet in WSN (Wirele Senor Network a they poe many intereting challenge in eigning optimal WSN framework for many ifferent application in recent year. Senor network are alo being ue for conitional bae maintenance, urveillance, computer augmentation an inventory tracking. Thu, the nee for location aware ervice ha increae manifol. There are everal eential iue in wirele enor network reearch area, uch a localization, eployment, energy conumption, coverage, etc. Localization i one of the mot important ubject becaue the location information i typically ueful for coverage, eployment, routing, location ervice, target tracking, an recue. Therefore, how to accurately know the location of each enor noe i one of important iue. In [], it i an example for maintenance of large mechanical 76

2 equipment an high-level equipment acce control an other application. It i not helpful without recoring the location of enor noe. In many outoor location application, configure a GPS receiver to ait poitioning of enor noe nee. However, the cot i too high an each enor noe ha a conierable egree of GPS poitioning error that leaing to a large gap of location information of enor noe. The current localization algorithm can be broaly ivie into two type: Range Bae, the main ue of itance or angle information to calculate the poitioning information. Range Free [7] [8] [9], without the nee to obtain precie information on the itance an angle to etimate the noe location. Range Bae location metho have Time of Arrival (TOA, Time Difference of Arrival (TDOA, Angle of Arrival (AOA, Receive Signal Strength Inicator (RSSI, [] [] [] [5] [6]. Range Bae technique in location ha more accurate reult than the Range Free, but the harware evice nee more accurate. Therefore, the harware evelopment cot i relatively high. The comparion between Range-free an Range-bae localization cheme, Range-free o not nee expenive equipment in the harware evice, which followe by receiving the beacon information to calculate the itance between other enor noe. Thu, in general, in the cae of the phyical layer allow, the more beacon noe will make it more precie location. In []-[9], the majority of the localization algorithm are bae on two-imenional pace for icuion. But the real-life application are bae on three-imenional pace, uch a natural iater monitoring, earch for peronnel location ytem. The two-imenional location algorithm ue for the three-imenional pace coul have large ifferent. The Range-free three-imenional location algorithm ha been propoe in [0] []. They alo ue flying anchor an GPS receiver, an thee flying anchor broacat their location meage continuouly. Accoring to the location meage, the urrouning noe are able to calculate their own poition. The multiple flying anchor broacat their GPS poition meage parallel. Other enor noe calculate their own poition accoring to the poition meage. However, the reource of wirele enor noe (i.e., the performance of microproceor, memory ize, etc. are very limite. The complicate computation for localization may excee the computational buren of enor noe. Furthermore, the energy conumption of WSN noe coul increae, becaue of each location unaware noe receive poition location meage continually. In thi paper, the ingle mobile anchor with a GPS receiver to coorinate it own poition in unknown territory. The mobile anchor can fly through the ening pace to collect beacon from location unaware enor noe, an calculate the accuracy poition through the Secant Location Etimation Algorithm (SLEA in the threeimenional coorinate ytem. There i one of example on mountain recue application, each hiker ha one WSN tranmitter which broacating beacon meage perioically. The ingle mobile anchor i uitable for track, earch an location for the lot contant hiker by uing the propoe localization algorithm in thi paper. The remainer of the paper i organize a follow; in ection, we ecribe our aumption an the etail of the propoe technique. Simulation reult are preente in ection. Section conclue the paper. 77

3 Secant Location Etimation Algorithm In wirele enor network, enor noe are uually eploye uniformly acro the eployment area. The anchor noe are more powerful noe with more avance capabilitie, uch a GPS to receive it own location. Figure how the ecant location algorithm for ingle mobile anchor noe. Fig.. Secant location etection Algorithm In Fig., the point S i the location unaware enor noe, however, enor noe en the ignal area viible a a phere. The ingle mobile anchor noe flie through S noe ignal tranmiion range. The ecant of the phere will certainly cro with the urface of phere by two virtual reference point, uch a p an p. Four virtual reference point from P to P are electe from ecant. Four electe virual reference point are locate on the phere urface of S noe ignal tranmiion range. Thu, the Receiver Signal Strength from four virtual reference point are equivalent. In the other wor, four virtual reference point have ame itance to the location unaware enor noe. We aume that the coorinate of noe S i (x, y, z an four virtual reference point coorinate i (x, y, z, (x, y, z, (x, y, z an (x, y, z repectively. Equation (-( are ue to calculate the three-imenional itance between noe S an viual reference point. The itance from to i can be obtaine from following expreion. ( x x ( ( x x ( + + ( ( x x ( y y ( z ( x x ( 78

4 Due to four virtual reference point have ame itance to noe S. In Secant Location Etimation Algorithm, the coorinate of (x, y, z, (x, y, z, (x, y, z an (x, y, z are learne from GPS evice. Thu, from equation ( to ( are equal. We can obtaine x, y, z, from equation (5 which preent the coorinate of location unaware enor noe S (x, y, z. ( x x ( x ( x ( x x x x (5 Simulation Reult In thi ection, imulation reult are preente for SLEA cheme in the D coorinate ytem. A 500 * 500 * 500 m imulation environment wa evelope uing the Qualnet eveloping library []. 50 location unaware enor noe were ranomly eploye in the ening fiel with zero altitue. Each target noe ue the fixe tranmiion power (mw broacat beacon at regular interval (0.ec. The raio propagation moel wa bae on Free-pace propagation moel. The mobile anchor noe wa place in the center of the imulate fiel in the beginning. The moving pee of mobile anchor i aroun meter/ec. The other parameter etting for the imulation are ummarize in Table. Table. Secant location imulation parameter Tranmiion frequency.ghz Interference moel NO Pathlo moel Free pace RSSI (Bm -85 ~ -70 Beacon ( Fig. an illutrate the location error imulation reult. Figure repreent the relationhip between the receive ignal trength an the location error itance. The reult how the invere proportion between the RSS value an the location itance error in SLEA cheme. It i becaue that the lower ignal trength ha higher enitivity form backgroun noie. Figure how the relationhip of the receive ignal trength an the location error ratio. The location error ratio of SLEA i aroun % in o. econ beacon interval, which repreent a contant itribution inie the tranmiion range. The higher beacon interval reult higher itance error. In the other han, the horter beacon interval will caue higher energy ubume in beacon tranmiion. 79

5 The imulation reult emontrate that the ingle mobile anchor ha a horter movement itance approach to etimate the location unaware noe accuracy poition. The reult how that, the propoe SLEA approach accuracy of localization i a major avantage. Fig.. Location Etimation reult in itance error (meter Fig.. The proportion of Location Etimation Error Concluion an Future Work The motivation for thi paper ha been to how the accuracy with which wirele enor network can etimate the enor location by uing the propoe SLEA Secant Location Etimation Algorithm. SLEA cheme require at leat four viual reference point to receive noe beacon, but, the ingle mobile anchor noe of the cheme make it a much better eay implementation. The accuracy of the etimate location i proportional with the itance. The percentage error i aroun % ~ % at 0.S 80

6 beacon interval. The relocation of any noe can eaily be one if etecte within the noe beacon tranmiion range. The aaptability of the SLEA cheme for the variable communication range make it more uitable for eployment leaing to patchy region. The error-prone channel conition can be coniere a a future work. Acknowlegement Thi work wa upporte by National Taichung Univerity reearch grant NCTU990 an NSC reearch grant E Reference. Wei Xin, Wang Quani, Yang Chenghe, Yang Yongming, A Three-imenional Noe Localization Algorithm of High-preciion in Wirele Senor Network, ICIT '09 Proceeing of the 009 IEEE International Conference on Inutrial Technology, 009.J. Clerk Maxwell, A Treatie on Electricity an Magnetim, r e., vol.. Oxfor: Clarenon, 89, pp L. Cong an W. Zhuang, Hybri TDOA/AOA Mobile Uer Location for Wieban CDMA Cellular Sytem, IEEE Tran. Wirele Comm., vol., no., pp. 9-7, July 00.K. Elia, Title of paper if known, unpublihe.. M. McGuire, K.N. Platanioti, an A.N. Venetanopoulo, Location of Mobile Terminal Uing Time Meaurement an Survey Point, IEEE Tran. Vehicular Technology, vol. 5, no., pp , July 00Y. Yorozu, M. Hirano, K. Oka, an Y. Tagawa, Electron pectrocopy tuie on magneto-optical meia an platic ubtrate interface, IEEE Tranl. J. Magn. Japan, vol., pp. 70 7, Augut 987 [Diget 9th Annual Conf. Magnetic Japan, p. 0, 98]. N. Patwari, A.O. Hero, M. Perkin, N.S. Correal, an R.J. O Dea, Relative Location Etimation in Wirele Senor Network, IEEE Tran. Signal Proceing, vol. 5, no. 8, pp. 7-8, Aug Xiaobo Zhou, Liqiang Zhang an Qiang Cheng. (006 Lancape-: A robut localization cheme for enor network over complex terrain, In Proc. t IEEE International Conference Local Computer Network, page 9--6, November, Davie Merico an Roberto Biiani. (006 Poitioning, localization an tracking in wirele enor network, Technical report, DISCo, NOMADIS, March, Davie Merico an Roberto Biiani. (006 Poitioning, localization an tracking in wirele enor network, Technical report, DISCo, NOMADIS, March, Yu-Chee Teng, You-Chiun Wang. (008 Ditribute eployment cheme in mobile wirele enor network to enure multi level coverage, In IEEE Tranaction on Parallel an Ditribute Sytem, page 80--9, September, Jiang Jin, Yi Wang, Chen Tian, Wenyu Liu an Yijun Mo. (007 Localization an ynchronization for unerwater acoutic enor network, In Springer, Ubiquitou Intelligence an Computing 007, page 6--6, Hong Kong, China, July Chia-Ho Ou an Kuo-Feng Su. (008 Senor poition etermination with flying anchor in three imenional wirele enor network, In IEEE Tranaction on Mobile Computing, page , September QualNet imulator, 8