Journal of Chemical and Pharmaceutical Research, 2016, 8(4): Research Article

Transcription

1 Avalable onlne Journal of Chemcal and Pharmaceutcal Research, 2016, 8(4): Research Artcle ISSN : CODEN(USA) : JCPRC5 Vrtual Force Coverage Enhancement Optmzaton Algorthm Based on Node Energy Balance n Wreless Sensor Networks Guozeng Zhao Department of Computer and nformaton Engneerng, Luoyang Insttute of Scence and Technology, Luoyang, Chna ABSTRACT In ths paper, a vrtual force coverage enhancement algorthm based on node energy balance s proposed. Through the vrtual force between node and node and the target regon between moble node postons, nodes n the network fnally a ratonal dstrbuton and unform, and mprove the network coverage performance, wth a mnmum of sensor nodes to maxmze the coverage area. Reducng the montorng blnd area had n wreless sensor networks, reducng the cost of the network. Fnally, the feasblty of the algorthm s verfed by the expermental data. Key words: WSN; vrtual force; coverage; energy INTRODUCTION Wreless sensor networks (WSN) [1-2] s an open network formed by a large number of sensor nodes, whch are characterzed by ther ablty to perceve, compute, communcate and store. Wdely used n natonal defense mltary, ntellgent transportaton, medcal and health, and envronmental montorng and ntellgent home and other engneerng felds [3-5]. Coverage problem s a basc problem n wreless sensor network research, and t s also the core problem [6-8]. In WSN, the target area s covered s drectly related to the overall performance of sensor networks, the network s able to accurately and effcently complete the task has a crtcal nfluence. In order to mprove the network's overall coverage, reduce the number of sensor nodes n the network, usually need to target the coverage area of n-depth study, the node layout more reasonable [9-11]. In recent years, wth the progress of scence and technology, the development of moble robots technology promotes the emergence of moble sensor nodes, makes the target coverage area n randomly deployed sensor nodes accordng to the actual stuaton of heavy deployment was made possble, ths new technology wll also wth the emergence of moble sensor nodes and become a hot ssue. Therefore, how to use the least sensor nodes to complete the coverage and connectvty of the desgnated area and to restran the excessve energy consumpton of nodes s a challengng problem. Ths paper draws on the research dea of vrtual force. Accordng to the dfferent drecton of energy consumpton of nodes n the applcaton, the wreless sensor coverage problem s studed by usng the node's energy consumpton as the target nstead of the movng dstance. PROBLEM DEFINITION AND NETWORK MODEL Defnton 1: the dstance between any two nodes called (, ) d ( s, s ) < 2R, R s the radus of node sensng, called S and d s s node S and S node to the neghbor node. S Eucldean dstance, when 788

2 Defnton 2: coverage 0, Rs d( s, s ) p( s, s ) = 1, Rs Re d( s, s ) (1) In R representaton for a node's sensng radus, s where d = d ( s, s ) ( Rs R e ), (, ) d ( s, s ) ( Rs Re ), R represents the dynamc parameters montorng sensor node, e d s s s the Eucldean dstance between sensor nodes; when s node s detected, otherwse, s not detected [5-8]. Defnton 3: The movng dstance of the sensor node energy consumpton of the moble and sensor nodes are closely related, we assume that the energy consumpton of the sensor nodes to moble unt dstance, sensor nodes n coverage n the process of energy consumpton can use formula (2) to say: E( s ) = d E (2) p Among them, d s the movng dstance of the node s n the process of coverng. Obectves of the algorthm s to make wreless sensor network to meet the heavy condtons of connectvty, energy consumpton mnmzaton, equalzaton of nodes n a sensor network, maxmze the wreless sensor network's overall coverage. COVERING ALGORITHM BASED ON VIRTUAL FORCE Stress analyss of node In the experment, usng mathematcal methods to the vrtual force nto movng dstance and add energy factors. Wreless sensor networks n the ntal random deployment, some nodes may are arranged n the far dstance montorng pont poston, even s arranged on the outsde of the whole montorng area, n order to make the random deployment of sensor nodes can be deployed n the target coverage area of nternal and need to n the target area arranged to attract the source, provde attractve to the surroundng nodes [12-14]. Assumng that the center of the target coverage area s the target coverage, we regard t as an attractve source of the target area, and t plays a role of attracton to the randomly deployed wreless sensor network nodes. In addton, sensor node and the target area center dstance farther, by the center of attracton s greater, ths can remote sensor nodes faster nto the target wthn the coverage area and mprove the algorthm's effcency. Ths paper uses Hooke's law to smulate the target regon of attracton. Hooke's law stpulates: the sprng and the sprng s proportonal to the change of shape. Smlar goals coverng the regon of attracton and node and a regonal center dstance proportonal to, accordng to Hooke's law formula F = k x, we defne target coverage regonal center of any sensor node attracton [9-10]: F ( s ) = k d( o, s ) (3) a a Among them, k s the target area center attracton coeffcent, Eucldean dstance s d( o, s ). Obvously, the a regonal center of attracton Fa ( s ) drecton s determned by the sensor node s ponts to the target area center. Ths paper studes the target coverage area grd dvson; each grd center provdes attractve target area to ts perpheral nodes. But f only the target area center attracton, then all the sensor nodes wll converge to the target area center, the entre network covers only to a central pont, no practcal sgnfcance. Therefore, n order to balance the regonal center of attracton cover so that nodes can be n force balance of the statonary state, we ntroduce the repulson between neghbor nodes, n order to enable the network to unform coverage. The ntroducton of the neghbor node repulson between the role s to ensure that sensor nodes can force balance, the coverage area of adacent nodes not excessve overlap. Ensure that the adacent sensor nodes to acheve the optmal dstance, mprove the network coverage. Assume that when the dstance between adacent sensor nodes s less than the best dstance, t wll produce the correspondng repulsve force greater than the optmum dstance, there s no such repulson. For the deployment of wreless sensor networks, sensor nodes to maxmze the network coverage at least are the ultmate purpose. In order to meet the above condtons, when the sensor node dstance between the hours and by the repulson should be more, such ablty and the smaller the dstance and the value center of the target area more attractve to mantan the state of equlbrum, the repulson between the repulson and 789

3 same-sex charged partcles are very smlar. Accordng to the formula of Coulomb's law, the repulsve force between the sngle hop neghbors can be obtaned as follows: k d d < d < d Fr( s, s) = 0, d > dopt 2 2 b (1/ 1/ opt ),0 opt (4) When the dstance between nodes s less than the best dstance d opt, t wll produce a correspondng repulsve force. Ths repulson wth the dstance between nodes ncreases exponentally, untl the dstance between the nodes s larger than the optmal dstance d opt, the repulsve force dsappears. In lne wth the actual stuaton, we defne the optmal dstance between adacent nodes s: d opt mn( R,2 r ), Clm ( A) < 1 = mn( R, 3 r ), Clm ( A) = 1 (5) Cover the area, become useless nodes, resultng n waste of resources, affect the performance of the network. Therefore, we must take some measures to ensure that the sensor nodes n the target coverage area wll not move the target coverage area. The concept of boundary repulson s ntroduced. Through the comprehensve effect of the three forces, sensor nodes are randomly deployed n the regon to the best poston to the target moble coverage. Energy s a basc problem n wreless sensor networks, sensor networks, should try to reduce and balance the network energy consumpton and maxmze the lfetme of the sensor network. The resdual energy of nodes and frcton combnaton put forward the concept of moble node frcton. Frcton formula for moble sensor nodes: F = 1/ ( µ E ), F 0 (6) c s Among them, µ s the coeffcent of frcton, E s the resdual energy of the sensor nodes, F c s the sensor nodes are subect to the ont force n addton to frcton. Seen from the above formula, when the frcton of moble sensor nodes and node's resdual energy s nversely proportonal, so that when the resdual energy of the nodes s small, receved from the sensor node frcton s large, t s dffcult to move, reduce the moble sensor nodes, energy consumpton wll be reduced. In ths way, the balance of the resdual energy of the entre sensor network s guaranteed, and the network lfetme s prolonged. It force sensor nodes, but also do not know how the role of sensor nodes n these moble forces. After the formaton of the network, the wreless sensor network nodes are randomly deployed, not all nodes are not n the target coverage area. When the node coverage area wthn the target, the use of four knds of vrtual force can ensure that the sensor nodes as much as possble to acheve the best poston. And when the nodes n the target coverage area outsde, n order to make the external node can faster nto the target coverage area, mprove the network performance; we assume that at ths tme all external nodes only by the target coverage of regonal center of attracton. The sensor node s expected to: Fa ( s ) + Fr ( s ) + Fb ( s ) + Fc ( s ), s S F( s) = Fa ( s ), s S (7) S represents a collecton of sensor nodes n the target coverage area. The stress analyss and the nature of the tangent functon of the sensor nodes are ntegrated, and the formula for the sngle step dstance of the sensor nodes s as follows: 2 d( s ) = arctan( F( s ) ) dmax (8) π The drecton of movement of the node s the drecton of the ont force. Each sensor node accordng to the vrtual force to move to a new poston, and update the coordnates (x, y) for the new coordnate ( x, y ), the followng formula: 790

4 Fx 2 x + arctan( F( s ) ) dmax, Fc Fth I F( s ) 0 x = Fxy π x, Fc < Fth U F( s ) = 0 Fy 2 y + arctan( F( s ) ) dmax, Fc Fth I F( s ) 0 y = Fxy π y, Fc < Fth U F( s ) = 0 (9) Vrtual force node coverage algorthm When we need to montor the target coverage area n a random deployment of wreless sensor network nodes, the algorthm begns to run. Frst, each sensor node to broadcast the stuaton to transmt ther poston nformaton and resdual energy nformaton, and through the snk node wll send the nformaton to the control termnal, the control termnal after gettng the nformaton accordng to the formula of the moble node dstance of the sensor, smulate the movement of sensor nodes, adustng the sensor node poston, repeat the process, untl the sensor nodes are not moble, the termnaton of the algorthm. The control termnal of each sensor node had locaton nformaton s sent to the sensor network, each sensor node accordng to the nformaton receved by the moble to the correspondng poston. In order to make the algorthm converge faster and avod the state of nfnte loop, we need to set a threshold value of Cth. Sensor nodes on a smulated moble, namely algorthm for cycle tme, when the algorthm cycles to reach the threshold CTH, regardless of the results, to the end of the algorthm, sensor nodes are not moble, mprove the algorthm's effcency. In addton, consderng the resdual energy of sensor nodes not much, f to contnue the nodes of moble, so that the node energy faster depleton, the regon of the montorng node wll become vulnerablty montorng, affect the performance of the whole network. Therefore, the threshold value of a frcton force s provded, and the sensor node does not move when the frcton force of the sensor node s greater than the threshold value. In ths way, the servce tme of the sensor nodes can be extended as far as possble, whch mproves the performance and the lfetme of the whole sensor network, and also balances the energy consumpton of the whole sensor network. Algorthm process and ratonalty analyss The precedng dscusson and research are n two-dmensonal space, and n the practcal applcaton, the nodes of wreless sensor network s deployed n 3D space. Therefore, n the research of three dmensonal sensor nodes deployment algorthm has more practcal sgnfcance. In three dmensonal spaces, t s assumed that the target coverage area s a three-dmensonal cube space. Sensor nodes need to deployed n the regon of the cube and at the same tme, based on the assumpton that the front, sensor nodes n whch by the combned acton of the varous forces, under the acton of these forces n the nodes correspondng moble, the sensor network to acheve the best state. In the three-dmensonal space, the force of the node s smlar to that n the two-dmensonal space, the movng dstance of the node s the same, and the dfference s the expresson of the movng coordnate, because of the need to consder the movement of the Z axs n the three-dmensonal space. Solutons for the perceptual model of wreless sensor network after the ntal deployment, sensor node poston dstrbuton unreasonable, resultng n vulnerablty montorng, node resource waste and coverage area was smaller s proposed n ths paper. On the bass of the tradtonal vrtual force model s ntroduced based on node resdual energy of sensor nodes move frcton, thus to reduce waste of network resources, mprove network performance and ncrease coverage of sensor nodes at the same tme, reduce and balance the node energy consumpton and prolong the network lfetme. In addton, the now boomng wreless multmeda sensor networks, ths paper presents the dea of combned sensor nodes, the deployment of multmeda sensor nodes transformaton for sensor node deployment problem, reduce the dffculty of algorthm, the algorthm s used more and more wdely. SIMULATION ANALYSIS AND PERFORMANCE TEST In order to further verfy the algorthm's effectveness, the expermental platform s MATLAB7. In length of 1000 meters square area n randomly deployed 40 wreless sensor network node, target montorng area edge s 800 meters square area. The sensng radus of all the sensor nodes s 90 meters, and the communcaton radus of the 10 9 sensor nodes R s also 90 meters. The basc parameters of the vrtual force algorthm: k a = 1, k b = 10, k c = 10, 791

5 sensor nodes n a sngle step movng maxmum dstance d max =5 and cycle tmes of algorthm s 100, the unt dstance the energy consumpton of the moble Ep=1, sensor node threshold frcton for 5. Fgure 1 shows the value data of 40 sensor nodes n wreless sensor networks at dfferent teratons under coverage. From Fg.1 t can be found that the number of teratons and the coverage rate s proportonal to growth and coverage s ndeed ncrease wth the ncrease of the number of teratons when the number of teratons to reach about 50 tmes when, coverage rate remaned stable and reached the optmal coverage. In practcal applcatons, has a great nfluence on the performance and cost of many network sensors, we always hope to mnmze the cost and maxmze the network coverage rate. Therefore, ths paper also analyzes the nfluence of the number of sensor nodes on the coverage rate. Fg.2 shows the coverage data of dfferent nodes. Fg.1 t=50, An ontology concept herarchy tree Fg.2 t=100, An ontology concept herarchy tree The expermental results show that the algorthm can balance the energy consumpton of sensor network nodes, elmnate the montorng blnd area n the target coverage area and the overlappng coverage area, and ncrease the coverage rate of the network. The effect of the number of teratons and the number of sensor nodes on the performance of the algorthm s verfed by experments. CONCLUSION Based on the characterstcs of the prevous algorthms, ths paper proposes a vrtual force coverage enhancement optmzaton algorthm based on energy balance. In ths algorthm, the target montorng area s dvded nto grd, whch can accelerate the convergence rate, and provde the bass for the study of uneven coverage. On the bass of the exstng vrtual force coverage model, ths paper smulates the resdual energy of the node as the frcton force, reduces the energy shortage node movement, balances the overall energy consumpton of the network, and prolongs the lfetme of the network. The coverage enhancement algorthm proposed n ths paper s carred out n detal by usng the Matlab7 platform, and the effectveness of the algorthm s verfed. Acknowledgements Ths work s supported by the Natural Scence and Technology Research of Foundaton Maor Proect of Henan Provnce under Grant (No , ) 792

6 REFERENCES [1] Zeyu S, Weguo W, Huanzhao W, Heng C, Xngfe X. EURASIP Journal on Wreless Communcatons and Networkng, v.32,n.6,pp.1-17,2014. [2]Wen Shaoe, Zhang Zhenyu, Yang Wenzhong. Computer engneerng and applcaton, v.51,n.7,pp.80-93, [3] Sun Zeyu, Lawrence Ng, Wang Huanzhao et al. Journal of electroncs, v.3,n.3,pp ,2015. [4] Nshmura C.E. and D.M.Conlon. Marne Technology Socety Journal,v.51,n.6,pp.13-23,2004. [5] C. Herrng, S. Kaplan..IEEE Personal Communcatons, v.10,n.8,pp.60-72,2000. [6] G J.Pote, W.J. Communcaton of the ACM, v.16,n.5,pp ,2000. [7] J. Comes, S. Martnez, T. Karatas, F. Bulb. IEEE Trans. On Robotcs and Automaton, v.20,n.2,pp , [8] X. L, P. Wan,. IEEE Transactons on Computers, v.52,n.6,pp ,2003. [9] X. Xu, X. We, and Z. Ye. IEEE Sgnal Process. Lett., v.19, n.3, pp , [10] X. We, Y. Yuan, and Q. Lng, IEEE Trans. Sgnal Process., v. 60,n.12, pp , [11] B. Farhang-Boroueny, IEEE Sgnal Process. Mag., v. 28, n.3, pp , [12] L. Vangelsta and N. Laurent, IEEE Trans. Commun., v. 49, n.4, pp , [13] P. Sohan, C. Sclet, and N. Lacalle, IEEE Trans. Sgnal Process., v. 50, n.5, pp , [14] T. Fusco and M. Tanda, IEEE Trans. Sgnal Process., vol. 55, no.5, pp ,