Hindawi Publishing CorporationInternational Journal of Distributed Sensor NetworksVolume 2013, Article ID 269215, 6 pageshttp://dx.doi.org/10.1155/2013/269215
Research ArticleThe Cluster-Heads Selection Method considering EnergyBalancing for Wireless Sensor Networks
Choon-Sung Nam,1 Young-Shin Han,2 and Dong-Ryeol Shin3
1 Interaction Science Institute, Sungkyunkwan University, Seoul 110-745, Republic of Korea2Division of Information Technology, Sungkyul University, Anyang 440-746, Republic of Korea3 College of Information and Communication Engineering, Sungkyunkwan University, Suwon 430-742, Republic of Korea
Correspondence should be addressed to Dong-Ryeol Shin; [email protected]
Received 21 March 2013; Revised 2 October 2013; Accepted 2 October 2013
Academic Editor: Tai-hoon Kim
Copyright © 2013 Choon-Sung Nam et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.
Wireless sensor networks (WSNs) are self-organizing networks in which sensor nodes with limited resource are scattered in anarea of interest to gather information. WSNs need to have effective node’s energy management methods for stable and seamlesscommunication. As one of a number of good technical solutions, a clustering technique has been issued and proposed amongresearchers for reducing energy consumption in WSNs. Also, it can prevent the problem of data duplication by the sensor nodes.Generally, to reduce WSNs’ energy consumption as much, cluster heads (CHs) are selected dynamically based on cluster rotationmechanism. However, the CH that is already previously selected could not be selected again unless the round process is over eventhough the node has more energy than others. Following this fact, in WSNs, there is a kind of irregular energy consumption statusamong sensor nodes because of CHs’ overhead energy usages. To solve this problem, in WSN networks, any sensor node shouldbe a candidate to be CH without any exception even if the node is already chosen before. Therefore, in this paper, we will establishand propose an energy balanced CH selection mechanism and the distribution of sensor node’s energy consumption in WSNs forequal and stable energy management.
1. Introduction
In wireless sensor networks (WSNs), the large number ofsensors is deployed over a wide range to inspect or mea-sure their environmental performance. Generally, the majorresponsibility of these sensor nodes in WSN is to detect andcollect WSN’s environmental data and to send its data intoWSN network’s external end users. Because the sensor nodeinWSNs should be operated stably in the irregular deployingfieldswhose environment is so difficult to be approached or tobemuch dangerous, mostly, the unattended operation systemwhich is automatically operated by itself without additionaloperator’s action is needed on WSNs communication. Tomake our goal closer, trusted and stable WSNs operationworks in irregular deploying environment, WSNs need tohave the self-organizing mechanism which allows the nodeto implement its network topology by itself and also need theoptimizing battery management mechanism. In particular,
because the sensor nodes in WSNs have so limited powerresources during the operation, the sensor nodes have to dolow-powered communication as much [1]. So, following thisfact, the issue, which is mostly concerned about in WSNsresearch area, is that how tomanage sensor’s energy resourcesefficiently. The maximum or largest energy consumption ofnodes mostly occurred on WSN’s sensed data transmissionphrase. The sensed data which is detected by a certain nodeis equally the same thing as its neighbor nodes’ one. Thus,WSNs need to have a data aggregation mechanism to preventduplicated data [2–4]. Normally, a clustering technique ismentioned as one of the most typical data aggregation mech-anisms in WSNs because it is possible to avoid duplicatingdata transmission amongnodes by grouping the nodes, whichdetected similar events during transmission into local cluster[5–9]. To set local cluster as well, the cluster-head (CH)selection process should be preceded mostly. CH, typically,can be selected among deployed or separated nodes in
2 International Journal of Distributed Sensor Networks
WSNs. Unlike other local cluster’s member nodes, CH hasusually lots of energy consumption because of its additionalenergy usages, data aggregation collection. Thus, to migratewith this problem, WSNs need to have new CH rotationalselection algorithm which concerns CH’s irregular energyconsumption problem.
Generally, CH selection algorithm leads WSNs to givetheir nodes almost equal opportunities to become CHby using mathematical probability process. Applying thisconcept, WSNs possibly do stable energy balancing duringcommunication as well. But there is an operational limitationthat WSNs must concern two measuring criteria during CHselection process, the energy consumption on data collectionphrase and local cluster’s size. Furthermore, since the processof nodes’ battery residual value collection requires a lot ofadditional energy consumption, it is impossible to calculatean entireWSNnetwork energy usage state to choose a suitableCH node. So, in WSNs, nodes must check their currentenergy states by themselves and choose a suitable CH basedon their calculation analysis results asmuch.Therefore, in thispaper, we propose the CH self-selection mechanism basedon nodes’ energy residual value comparison algorithm tomigrate these problems.
2. Backgrounds
Many clustering algorithms have been proposed to select asuitable CH by many researchers. Among these algorithms,LEACH [10] is one of the most typical proposals that leadWSNs to disperse sensor nodes’ energy problems as formedlocal cluster. To be detailed, this algorithm is able to solvenodes’ energy traffic problem by pointing a certain CHout, with a certain numeric probability, per a certain roundtime, one time selection per one round. After forming localcluster, LEACH-C [10] puts CH on the center of local clusterto reduce the cost of energy transmission among membernodes. But this proposal requires additional energy cost toselect a new CH and to track a new trajectory for findingout the geographical local information of local cluster. Themajor characteristic of deterministic cluster-head selection[11] is to calculate the level of nodes’ remaining energy asthe acceptable probability rate of CH selection. However, thismechanism requires additional energy consumption becauseall nodes should share their current energy states when CHselection process. Other CH selection proposals [12, 13] alsohave the same additional energy costs to select CH becauseof sharing nodes’ current energy information. To overcomethis limitation, all nodes have to apply probability to the CHselection-making process by estimating the probability as CHcandidates like LEACH algorithm.
LEACH circulates a CH to distribute all sensor nodes’energy consumption and manages a local cluster by the CH.
LEACH circulates a CH to distribute all sensor nodes’energy consumption and manages a local cluster by the CH.It is composed of a “set-up” phase for clustering and “steady-state” for the time divisionmultiple access (TDMA) frame. Inaccordance with (1), all sensor nodes can be elected as a CHusing a threshold (𝑇(𝑛)) in the “set-up” phase. A node with a
lesser value than the threshold can be elected as a CH after itchooses a random value between “0” and “1.” The “G” meansthe set of nodes that can be selected as a CH.The “otherwise”means that node “i” has been a cluster head. Consider
𝑇 (𝑖) ={
{
{
𝑝
1 − 𝑝 ∗ (𝑟 mod (1/𝑝)), 𝑖 ∈ 𝐺,
0, otherwise.(1)
In (2), “i” is the number of nodes, “p” is the ratio of CHs,“r”is the current round, and “G” is the set of nodes that are notselected as a CH. Each node can be selected as a CH onceper 1/p rounds. A higher number of rounds mean that thereis a higher probability that a node will be selected as a CH.After 1/p rounds, all nodes would be selected as a CH, bysetting the threshold as “0.” But LEACH cannot guaranteeequal clustering.This can cause an imbalance of a local cluster[11]. 𝐸
𝑖(𝑡) is the current energy of node “i” with time “t,”
𝐸total(t) is the total energy of all nodes with time “t,” and 𝑃𝑖(𝑡)
is the CH probability of node “i” with time “t”. Cosider
𝑃𝑖 (𝑡) = min{
𝐸𝑖 (𝑡)
𝐸total (𝑡)𝑝, 1} . (2)
This method is for energy distribution, as all sensor nodeswould be selected as a CH after 1/p rounds. This helps nodesachieve efficient energy conservation, since nodes that havehigh residual energy are elected as CHs. However, it does notconsider unequal energy consumption of nodes by unequalclusters. The elected CH is not reelected as the CH during 1/prounds, although the node hasmore energy than the others inthe cluster. For example, if there are N nodes, the number ofCHs isN∗p, k.The average number of nodes in a local clusteris N/k. We assume that the probability that a local cluster hasN/k nodes is P. The probability should be P, 1/p to manageP from 1 round to 1/p rounds. But the probability is remote.Thus, a method to overcome unequal clustering is needed.
The soft-threshold method [14] is used to select adaptivecluster heads among normal sensor nodes by using a constant𝜀 value. However, this technique does not have guaranteedequal energy consumption among nodes, because it does notconsider the energy status of nodes, especially in CHs.
The energy adaptive cluster-head selection (EACHS)algorithm [15] compares the average energy of every nodewith its remaining energy and the energy consumed for datatransfer from the former round. The selection probability ofthe CH increases if it has more energy and decreases if it hasless energy; that is, the selection of the CH is based on theenergy level. 𝐸rest is the rest energy of a node in this round,𝐸diss is dissipation of a node in this round, and 𝐸ave is theaverage energy consumption of a node. Cosider
CHprob =𝑝
1 − ((𝑟 + 1) mod (1/𝑝))× [𝐸resi − 𝐸diss𝐸ave − 𝐸diss
] . (3)
However, as (1) shows, a node can be not selected if it was aCH in the same round. In addition, EACHS may consumeadditional energy to compare the threshold method, sinceit requires the exchange of information between nodes tocalculate the average energy of nodes. This mechanism also
International Journal of Distributed Sensor Networks 3
need to calculation of the probability that a node couldbe selected as a CH. So, we present the new cluster headselection algorithm to solve three problems. First, it haveto require noadditional energy consumption by exchangingenergy condition information. Second, it can re-select a nodeas a CH though it was elected in current round. Finally, it cancalculate the CH qualification by each node using remainingnode energy.
3. (Energy Balanced Cluster-Heads Selection)EBCHS
The previous section showed that the energy gap between aCH and a member node is large due to the additional cost ofthe CHs. Generally, a member node simply detects changesin its surrounding environment and transmits the senseddata to a CH. The amount of aggregated data produced bya CH depends on the number of its member nodes. Thus,the energy consumption of a CH is higher than that of theother member nodes. Tomanage the energy balance betweennodes, the WSNs require a novel CH selection algorithmthat considers the energy status of the CHs. Our proposal iscalled the energy balanced cluster-heads selection (EBCHS)algorithm and it uses a threshold, 𝑇(𝑖). As shown in (1), if ris 0, 𝑟 = 0, the probability of all sensor nodes, 𝑇(𝑖)
𝑟= 0, is
p because not all sensor nodes have been selected as a CH.The scattered nodes of WSNs are “N.” Thus, we revised thisequation. Equations (4) and (5) are based on LEACH andEACHS. The purpose of proposed equation is guaranteed toprevent additional communication cost and reelect CH rolebefore 1/p rounds. Energy consumption model is also basedon LEACH algorithm and is referenced by [10].
Like (1), if 𝑟 > 0, the threshold value of a nodeselected as a CH can be reduced by the amount of energyconsumed. This means that the consumption energy ratio,𝑇(𝑖)𝑟>1
, can calculate the previous threshold value, 𝑇(𝑖)𝑟−1
,and the current threshold values, T(i). T(i) consists of twothreshold values: 𝑇(𝑖)
𝑟 ch,𝑇(𝑖)𝑟 mem. 𝑇(𝑖)𝑟 ch is the amount ofenergy drain ratio of remaining energy in a CH, and𝑇(𝑖)
𝑟 memis the amount of energy drain of remaining energy in amember node. 𝐸mem diss is energy consumption of membernodes, 𝐸ch diss is energy consumption of cluster heads, and𝐸remaning is remaining energy of nodes. Cosider
𝑇(𝑖)𝑟=0 = {
𝑝
1 − 𝑝 ∗ (𝑟 mod (1/𝑝))= 𝑝, 𝑖 ∈ 𝑁 , (4)
𝑇(𝑖)𝑟>1 = {𝑇(𝑖)𝑟−1 − 𝑇(𝑖)𝑟 mem, 𝑇 (𝑖) > 0,
𝑇(𝑖)𝑟−1 + 𝑇(𝑖)𝑟 ch, 𝑇 (𝑖) ≤ 1,(5)
𝑇(𝑖)𝑟 mem =𝐸mem diss𝐸remaining
, (6)
𝑇(𝑖)𝑟 ch =𝐸ch diss𝐸remaining
. (7)
When 𝑇(𝑖)𝑟=0
, all nodes are selected as a CH at least onceduring 1/p rounds. In the next round of CH selection, thenodes’ threshold value that is used with CH selection is
different, as it is the CH energy consumed in its local cluster.This difference stems from the fact that the number ofmember nodes in local clusters varies. If a CH has fewermember nodes than the average number of member nodes,the threshold value is also lower. This means that the CHcan be reselected during 1/p rounds. This will result in abetter energy distribution for sensor networks and increasednetwork lifetime.
4. Simulation Setup and Performance Analysis
4.1. Network Setup for Simulation. An ns2 [16, 17] simulatorwas used to evaluate the proposed algorithm. Ns2-allinone-2.27 was installed in cygwin of Windows XP SP3. The sensornetwork environment is as follows: a wireless channel, tworay ground radio-propagation model, physical wireless andMAC 802.15.4 protocol, Queue/DropTail/PriQueue, Omni-Antenna, and LEACH routing protocol. The data referencedby the sensor network references was from a real sensor node,MiCaz, based on the 802.15.4 data packet [18]. This is a 29-byte packet. We configured the network size to be 100m ×100m. The number of sensor nodes was N. The number ofCHs per round was 5% of the total number of nodes, as inLEACH. Thus, p is 5%, and the number of CHs per roundis 15. We set the data packet size as 34 bytes, the header as 5bytes, and the payload as 29 bytes. The sink node location is(0, 0). We assume that a node can transmit k (bits) data to asink or neighbor node 10m from a node.
4.2. Performance Analysis. We have to know the number ofsensor nodes in the range of a 100m × 100m network, sothat sensor nodes can communicate without generating anisolated node in the network; that is, we have to calculatethe number of nodes needed to avoid generating an isolationnode, when any node is put into the sensor network. Anisolation node is not generated if we have about 100 nodeswhen the nodes are put into the network in a grid pattern.However, we have to determine an isolation node by thefollowing experiment, when nodes are put into the networkrandomly. N, the number of sensor nodes in the network,is 300, 250, 200, 150, 100, and 50. Since each sensor nodeis placed in the network randomly, we can determine thenumber of isolation nodes, as shown in Figure 1, if the isola-tion node is defined as a node that is unable to communicatewith any other node.The higher the number of sensor nodes,the fewer the number of isolation nodes. When N is 50, thenetwork cannot function, because about 21 nodes, or 42% ofall nodes, are isolated.WhenN is 150, about 3 nodes, or 2% ofall nodes, are isolated. In that case, the network is not affected.However, over 300 sensor nodes are required for a normalnetwork environment, since isolation nodes cannot connectto any CHs or neighbor nodes.
The proposed method, energy balanced cluster-headsselection (EBCHS), is nearly equivalent to the previousmethod, in the first clustering round. Therefore, we willcompare the average energy consumption of nodes when𝑟 > 1. We assume that “1” round time is the time to selecta CH 20 times. Figure 2 compares the existing methods,
4 International Journal of Distributed Sensor Networks
50 100 150 200 250 3000
5
10
15
20
25
The number of sensor nodes in networks
The n
umbe
r of i
sola
tion
node
s
Figure 1: Average number of isolation nodes.
0 50 100 150 200 250 3001
2
3
4
5
6
7
8
9
10
ID of sensor nodes
The n
umbe
r of e
lect
ed cl
uste
r hea
ds
LEACHEACHSEBCHS
Figure 2: The number of cluster-heads counts per a node after 100rounds.
LEACH and EACHS, which determine the regular size ofa CH, as a regular probability p, and the proposed methodEBCHS. In Figure 2, while a node selected as a CH cannotbe reselected until the end of the round in the LEACH andEACHS methods, the node can be reselected in the EBCHSmethod, depending on its remaining energy. Therefore, thenumber of times that each node can be selected as a CH isdifferent; that is, each node can only be selected as a CH fivetimes during the 5∗1/p rounds of the LEACH and EACHSmethods. However, any node can be selected as a CH, even ifa nodewas selected as aCH in a previous round of the EBCHSmethod.Thus, there is awide variation in the number of timeseach node can be selected as a CH.
0 50 100 150 200 250 300ID of sensor nodes
Nod
e ene
rgy
cons
umpt
ion
×10−3
0.51
1.52
2.53
3.5
LEACH 50EACHS 50EBCHS 50
(a) After 50 rounds
0 50 100 150 200 250 300ID of sensor nodes
LEACH 100EACHS 100EBCHS 100
Nod
e ene
rgy
cons
umpt
ion
×10−3
2.53
3.54
4.55
5.5
(b) After 100 rounds
Figure 3: Energy consumption per a node after 50 and 100 rounds.
However, even though a node selected as a CH executesadditional work several times over in EBCHS, it is possi-ble to distribute the energy of the CH, because a CH isselected depending on the residual energy. Figure 3 showsthe maximum (0.00295 joule) and minimum (0.00143 joule)energy consumed in LEACH (50) and themaximum (0.00318joule) and minimum (0.00068 joule) energy consumed inEACHS (50) after each algorithm was executed around 50times. Both methods incur an energy difference betweennodes, so it is hard to balance energy consumption betweennodes. Conversely, in the proposed method, EBCHS (50), amaximum energy of 0.00270 joule and a minimum energyof 0.00172 joule are consumed; namely, EBCHS is better fordispersion of energy consumption. In addition, the maxi-mum and minimum energies are less in EBCHS than in anyothermethodon a round-by-roundbasis, so energy efficiencyis managed equally. After 100 rounds, energy gap betweennodes of EBCHS (100) is just 0.00085 joule. But energy gapof LEACH (100) is 0.00201 joule and EACHS (100) is 0.00272joule. LEACH (50), EACHS (50), and EBCHS (50) are theenergy status of nodes after round 50. In the same sense,LEACH (100), EACHS (100), and EBCHS (100) are energystatus of nodes after round 100.
If the energy deviation between nodes is large, a nodemay be dead due to energy depletion. This can be checkedby comparing the number of dead nodes, depending on theenergy consumption at each round. In Figure 4, the first nodedies at 357 in the case of LEACH. The round of (first nodedead) FND is 379 in the case of EACHS. Conversely, theround of FND is 478 in the case of EBCHS. As Figure 4 shows,
International Journal of Distributed Sensor Networks 5
350 400 450 500 550 6000
50
100
150
200
250
300
The number of rounds
The n
umbe
r of d
ead
node
s
LEACHEACHSEBCHS
Figure 4: Increasing the number of dead nodes per a round.
the round in which the number of dead nodes becomes10% of all nodes is 432 in the LEACH, 432 in EACHS,and 507 in EBCHS. EBCHS has more rounds, until thenumber of dead nodes becomes 70% of all nodes. However,nodes are dead consistently after 490 rounds due to unequalenergy consumption. Nevertheless, EBCHS consumes lessCH energy and distributes less energy between nodes than doothers, because at least 50% of nodes should bemaintained tooperate a network.
EBCHS can have equal node energy consumption,because of the energy distribution of nodes through duplicateselection as a CH.Thus, the energy of each node is consumedat constant rate. On the other hand, in EACHS and LEACH,there is a high energy difference between the nodes andCHs, because these methods do not consider the energyconsumption of the CHs. the first node dead (FND) roundof EACHS and LEACH is shorter than that of EBCHS dueto the huge energy consumption of the CHs. After FND, theenergy consumption of the CHs is reduced, because they havefewer nodes than the number of nodes in EBCHS. Thus, asimplied by Figure 4, the number of dead nodes in EACHS andLEACH increases more gradually than in EBCHS. However,based on the previous isolation node experiment, we knowthat the network status is normal when the number ofnodes exceeds 150.Therefore, EBCHS can maintain a normalnetwork better than other mechanisms by using a minimumof 530 rounds.
5. Conclusion
Clustering methods for wireless sensor networks have beenproposed by many researchers. The energy consumption ofCHs is higher thanmember nodes, asCHs incur an additionalcost to manage, collect, and aggregate data from membernodes. It is necessary to distribute CH energy consumptionusing a CH election method. We achieved this by electingCHs according to the residual energy of sensor nodes. Thiswas compared via simulations. The proposed method hasa better distribution of energy consumption between nodes
than the othermethods, and the round inwhich the first nodedies is later than that of LEACH and EACHS.More nodes arealive in stable networks.
Acknowledgment
This research was supported by Basic Science ResearchProgram through the National Research Foundationof Korea (NRF) funded by the Ministry of Education(2013R1A1A2063180).
References
[1] M. Tubaishat and S. Madria, “Sensor networks: an overview,”IEEE Potentials, vol. 22, no. 2, pp. 20–23, 2003.
[2] I. F. Akylidiz, W. Su, Y. Sankarasubanmaniam, and E. Cayirci,“A survey on sensor networks,” Communication Magazine, vol.40, pp. 102–114, 2002.
[3] T. Wang, L. Wang, Z. Xie, and R. Yang, “Data compressionalgorithm based on hierarchical cluster model for sensor net-works,” in Proceedings of the 2nd International Conference onFuture Generation Communication andNetworking (FGCN ’08),pp. 319–323, December 2008.
[4] M. Meng, J. Yang, H. Xu et al., “Query aggregation in wirelesssensor networks,” International Journal of Multimedia andUbiquitous Engineering, vol. 3, no. 1, pp. 19–26, 2008.
[5] R. Haider, M. Y. Javed, and N. S. Khattak, “Design and imple-mentation of energy aware algorithm using greedy routingfor sensor networks,” International Journal of Security and ItsApplications, vol. 2, no. 2, pp. 71–86, 2008.
[6] A. Nikdel, M. S. Kohshoori, and S. M. Jamei, “An intelligentand energy efficient area coverage protocol for wireless sensornetworks,” International Journal of Grid and Distributed Com-puting, vol. 4, no. 4, pp. 41–56, 2011.
[7] R. Sheikhpour, S. Jabbehdari, andA.Khadem-Zadeh, “Compar-ison of energy efficient clustering protocols in heterogeneouswireless sensor networks,” International Journal of AdvancedScience and Technology, vol. 36, pp. 27–40, 2011.
[8] R. A. Akhtar, A. A. Minhas, and S. Jabbar, “Energy awareintra-cluster routing forwireless sensor networks,” InternationalJournal for Hybrid InFormation Technology, vol. 3, no. 1, 2010.
[9] W. R. Heizelman, A. C. Asan, and A. Balakrishnan, “Energy-efficient communication protocols for wireless sensor net-works,” in Proceedings of the Hawaii International Conference onSystems Sciences, January 2000.
[10] W. B. Heinzelman, A. P. Chandrakasan, and H. Balakrish-nan, “An application-specific protocol architecture for wirelessmicrosensor networks,” IEEE Transactions onWireless Commu-nications, vol. 1, no. 4, pp. 660–670, 2002.
[11] M. J. Handy, M. Hasse, and D. Timmermann, “Low energyadaptive clustering hierarchy with deterministic cluster-headselection,” in Proceedings of the 4th International Workshopon Mobile and Wireless Communications Network, pp. 9–11,September 2002.
[12] O. Younis and S. Fahmy, “HEED: a hybrid, energy-efficient,distributed clustering approach for ad hoc sensor networks,”IEEE Transactions on Mobile Computing, vol. 3, no. 4, pp. 366–379, 2004.
[13] J. Ibriq and I. Mahgoud, “Cluster-based routing in wirelesssensor networks: issues and challenges,” in Proceedings of the
6 International Journal of Distributed Sensor Networks
International Symposium on Performance Evaluation of Com-puter and Telecommunication Systems (SPECTS ’04), pp. 759–766, July 2004.
[14] D. Rong, Y. Bing, Y. Lei, and W. Jiawei, “Soft threshold basedcluster-head selection algorithm for wireless sensor networks,”in Proceedings of the 3rd International Conference on SensorTechnologies and Applications (SENSORCOMM ’09), pp. 526–530, June 2009.
[15] Y. Liang and H. Yu, “Energy adaptive cluster-head selection forwireless sensor networks,” in Proceedings of the 6th InternationalConference on Parallel and Distributed Computing, Applicationsand Technologies (PDCAT ’05), pp. 634–638, December 2005.
[16] http://www.internetworkflow.com/downloads/ns2leach/ns-llinone-2.27.tar.gz.
[17] “MIT 𝜇AMPS project ns2 code extensions,” http://www-mtl.mit.edu/research/icsystems/uamps/research/leach/leach code.shtml.
[18] S. Sanka and G. Konchady, “Communication between wirelesssensor devices and GNU radio,” May 2009.
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporation http://www.hindawi.com
Journal ofEngineeringVolume 2014
Submit your manuscripts athttp://www.hindawi.com
VLSI Design
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation http://www.hindawi.com
Volume 2014
The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Modelling & Simulation in EngineeringHindawi Publishing Corporation http://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
DistributedSensor Networks
International Journal of