An Application of Virtual MIMO Technology in WSN … · An Application of Virtual MIMO Technology...

International Journal ofIntelligent Engineering & Systems

http://www.inass.org/

An Application of Virtual MIMO Technology in WSN Routing ProtocolAlgorithms

Ying Liang∗, Yongxin Feng

School of Information Science and Engineering, Shenyang Ligong University, Shenyang, China∗ Corresponding author’s Email: [email protected]

Abstract: MIMO (Multiple-Input Multiple-Out-put) system is a core technology used in 802.11n which used mul-tiple antennas to suppress fading. In the field of WSN network, routing protocols must be designed to minimizeenergy consumption for the nodes’ energy supply is constrained. Using MIMO technique in WSN could make use ofthe energy accumulation in the receiving nodes to increase the receiver signal intensity, thus reducing bit error rateand improve the reliability of data communication. As the MIMO-based WSN could supply higher data transfer ratefor the same transmitting power and bit error rate, it also efficiently reduces the time of the node sending data andimproves communication performance of the whole network. In this paper, based on dynamic collaborative virtualantenna MIMO technology, a new energy efficient routing protocol is proposed. The simulation results indicate thatour protocol can obviously reduce the energy consumption in the specific application environment.

Keywords: MIMO; WSN; routing protocol; virtual antenna

1. Introduction

1.1 Overview of wireless sensor networksWireless Sensor Networks (WSNs) are composed of

a large number of battery-powered sensor nodes thathave the ability to sense the physical environment,compute the obtained information and communicateusing the radio interfaces. Because sensor nodes aregenerally deployed on a large and wild area, they arepowered by embedded battery. As it is difficult tochange or recharge the battery, energy-efficient mech-anism for wireless communication on each sensor nodeis so crucial for wireless sensor networks. Generally,WSNs have the following characteristics:

(1) Limited hardware resourcesDue to the price, volume and power restriction of

node, in addition to limited energy capacity, sensornodes have also limited processing and storage capac-ities.

(2) Huge number of nodes

Sensor node deployment in WSNs is application de-pendent and can be scattered randomly in an intendedarea or dropped massively over an inaccessible or hos-tile region. This request the node must have very strongfault tolerance and survivable.

(3) Self-organizationThe network organization does not depend on any

fixed network infrastructure. The nodes, after havingbeen deployed, coordinate their respective actions bya set of rules, which can quickly and automatically toform an independent network.

(4) Dynamic topologyIn WSN, some nodes could run out of the network

for failing or being blocked due to lack of power, orhave physical damage or environmental interference.New nodes may be added for the growing networksize. Moreover, the mobility of sensor nodes is some-times necessary in many applications. Therefore net-work topology changes frequently.

The purpose of the wireless sensor network is to

International Journal of Intelligent Engineering and Systems, Vol.4, No.1, 2011 34

collect information timely and reliably. In order toachieve this goal, protocols about the network trans-mission and topology control must be designed basedon the characteristics of the network itself. Since thepremise of the WSNs working well is that the sensornodes have sufficient energy, the design of networkprotocols must take into consideration the energy con-sumption factor and need to trade-off between improv-ing network performance and reducing energy con-sumption. Energy efficiency is not only the key to ex-tend the network lifetime, but also an important foun-dation to ensure the network is available and reliable,which meeting a certain quality of service.

1.2 MIMO technology overviewThe origin of MIMO technology has been a long

time, back in 1908, Marconi has proposed it to over-come channel fading. In the 70’s, some people havesuggested to using MIMO technology for wireless mo-bile communications systems, which is completed bythe AT & T Bell Laboratories scholars in the 90’s. In1995, Teladar theoretically derived from the MIMOcapacity under the fading condition. In 1996, Foshinihave proposed a new MIMO processing algorithm D-BLAST (Diagonal Bell Laboratories Layered SpaceTime). In 1998, a space-time code for MIMO was dis-cussed by Tarokh. Using V-BLAST (Vertical-BLAST)algorithm, Wolinansky et al have established a MIMOexperimental system in 1998 and achieved over 20bps/-HZ frequency spectrum utilization. The above worksattract much attention of scholars from various coun-tries and make the MIMO Technology has been devel-oping rapidly [1].

1.3 The advantages of MIMO technologyEnergy consumption module of sensor node include

the sensor module, processor module and wireless com-munication module. With advances in integrated cir-cuit technology, the processor and sensor module ismore and more low-power, and thus, wireless commu-nication module becomes the largest energy-consumingpart in a sensor node.

In wireless communications, the relationship betweenenergy consumption and communication distances canbe expressed as:

E = kDn (1)

where n is the path loss factor, and the range is 2 <n < 4. The value of n depends on many factors such asthe deployment environment of sensor nodes, quality

of the antenna and so on. The parameter k is a con-stant, D is the communication distance. According tothe formula (1), with the communication distance in-creases, the energy consumption of wireless commu-nication will increase dramatically. Therefore, in thecurrent wireless sensor networks based on SISO (sin-gle input and single output), many protocols use re-ducing the single-hop communication distance to savetransmission energy of nodes. When we use MIMOtechnique in WSN, it could make use of the energyaccumulation in the receiving nodes to increase thereceiver signal intensity, thus reduce bit error rate andimprove the reliability of data communication. Thatis, for the same transmitting power and bit error rate,the communication distance in MIMO-based WSN ismuch larger than that in SISO-based WSN, which effi-ciently reduces the time of the node sending data andimproves communication performance of the wholenetwork.

2. Related Research and Question Statement

The research of using MIMO technology in wire-less sensor networks first began in 2004 and currentlythere are some scholars have started working at someof the challenging issues. Literature [2] analyzed theextra energy cost of the communication among coop-eration nodes and achieved the total energy consump-tion in cooperative MIMO-based wireless sensor net-work. The experimental results show that, even inconsidering the additional traffic load, collaborativeMIMO communications mode still save more energythan SISO communications mode. Literature [3] ana-lyzed the performance of MIMO-based wireless sen-sor network in the case of the same channel interfer-ence. Literature [4] analyzed the diversity gain andmultiplexing gain provided by MIMO and proposedthe method of using different gains methods in MIMOaccording to different application requirements.

Some researchers dedicated to study the challeng-ing issues for WSNs using collaborative MIMO, suchas how to select collaboration nodes, synchronization,routing and so on. Literature [5] proposed a coop-erative node selection algorithm. Each node selectsthe node from its neighbors as its cooperation nodewhich has the smaller ratio of its transmission energyconsumption to its residual energy. Taking into ac-count the similarity of characteristics of collaborativeMIMO and clustering WSN, literature [6] proposedusing cooperative MIMO techniques in cluster-basedwireless sensor networks, which select collaborationnodes from cluster-heads to form collaborative MIMO.


In order to reduce the total energy consumption in thenetwork, literature [7] proposed a method of optimiz-ing the transmission rate and the size of cluster at thesame time. However, the algorithms proposed basedon the same assumption, that is, nodes are denselydistributed in the network to ensure cluster-heads findenough cooperative nodes. In practice, with the run-ning of network, some nodes could run out of the net-work for failing or being blocked due to lack of power,or have physical damage or environmental interfer-ence. When the network deployment becomes sparserbut the amount of information to be transferred is huge,the collaboration nodes could consume more energythan the normal nodes, which aggravates the energyimbalance in the network. Though some protocolsuse the method of shortening the period of clusteringto solve the problem, frequent clustering could wastemore energy. In view of the above insufficiency, thepaper proposes a dynamic collaborative virtual MIMO-based routing protocol (DCVM).

3. DCVM Routing Protocol

The basic idea is that it includes the node’s residualenergy additional information when sending a pack-age, then according to energy left information, theMIMO systems is formed by selecting a certain per-centage of nodes as virtual antennas dynamically, na-mely, DCVM routing protocol.

3.1 System modelLEACH algorithm is a typical topology adaptive clus-

tering algorithm. Its implementation is periodic, eachround divides into the clustering establishment stageand the stable data communication stage. During theclustering stage, the neighbor node dynamic forms thecluster, the cluster head selected according to certainprinciples. During the data communication stage, nodessend date to the cluster head, then, cluster head fusethese data and send it to the sink node, LEACH clus-tering algorithm achieve a good energy saving effec-tive, obtained the widespread attention. This articledraws on the idea of LEACH algorithm.

This network model shown in Figure 1, the networkis a sparse deploying; Nodes are randomly deployedin a square monitor area for collect data, in above net-work model, according to the role it plays, nodes canbe divided into different three kinds: head node, co-ordination antenna node and normal node. In this net-work model, the gathered data is fused by head node,after fusing, the date will be sent to base station by an-tenna nodes with MIMO-based technology. The vir-

Figure 1 Network Model

tual antenna node is dynamic selects from its neigh-bor node by cluster head node under certain limitingconditions. When head node communicate with basestation, the antenna node and head node dynamic col-laborative send package to base station with MIMOpatters, by reducing the send error rate and convert toidle quickly, this can save sending energy of networks.In addition to the head node and the antenna node, theremaining nodes shall be normal nodes. Each normalnodes belonging to a cluster, and normal node directcommunication with the head node.

Similar to LEACH, DCVM also based on the pe-riodic operation, in each round, first select the clusterhead, in the data collection phase, according to the ad-ditional residual energy information to form antennanode dynamically, then processing data transmission.When data transmission, firstly, normal nodes senddata to head node with the way of SISO. When thecluster head receives and fusion the data send fromnormal node, then send the data to base station withthe way of MIMO which is coded with STBC space-time code algorithm.

In this network model, assumption that Energy andcapacity of the base station is not limits and the basestation is of multiple antennas, base station receivespace-time coding data by multi-antenna way, and getthe original data by space-time decoding. In order tomore clearly describe and analysis the energy con-sumption of network protocols, make the followingassumptions:

(1) In the cluster, the distances of normal nodes com-munication between clusters head is shorter, commu-nication channel is additive white Gaussian noise chan-nel (AWGN) with the path loss factor of 2. The com-munication distance between cluster head node and


base station is longer. Communication channel is fre-quency non-selective slow fading channel with the pathloss factor of 4.

(2) In order to avoid the system is too complicated,ignore the baseband signal processing modules, in-cluding source coding, pulse shaping and so on.

(3) The STBC pace-time coding method is used whi-ch the encoding rate is 1.

(4) Each node can work in four different states, na-mely, sending, receiving, idle and sleep states.

During node transmission data phase, the node con-sumes energy is the biggest, in the accepting state, thenode only receives the data, consumes energy is small,In the idle phase, nodes monitor the use of wirelesschannel and check whether there is data sending to it,the energy consumption is smaller. In the sleep state,nodes turn off communication module, and then theenergy consumption is smallest.

3.2 Clustering algorithm descriptionIn the wireless sensor network, each node has a glob-

ally unique ID number; the ID number is MAC ad-dress of the hardware node. DCVM protocol is di-vided into four stages, namely, clustering, selection ofthe cluster head node, dynamic selection of the vir-tual antennas and data transmission. They are listedas follows:

3.2.1 Clustering stageThe selection of Cluster head is similar to the LEACH

protocol, in each round, according to its own gener-ated random number and some certain rules to run forcluster head. Constraint rules as follows:

P(i) =

{n

N−n∗(c mod Nn )

i f f lag(i) = 1

0 i f f lag(i) = 0

}(2)

where N is the total number of nodes in the network,n is the number of total clusters. If the node has beencluster head in the last round, then f lag(i) = 0, oth-erwise the f lag(i) = 1, only the residual energy isgreater than the threshold of the residual energy, thenode has a chance to become a cluster head. Thisvalue decided by energy consumption in one round,this stage is the same with the LEACH protocol.

Selection cluster head algorithm as follows:(1) When a node receives the trigger information

from the base station, each node generates a randomtime rand(t), which has a ceiling limit value, in net-work start or running period, the limit value is sendfrom base station to each nodes by broadcasts ways.Through the limit value, each cluster node knows the

length of cluster head selection time, according to itsrandomly generated time rand(t) to run for cluster hea-der selection.

(2) After the random time is ready, then turns onown timer, when the timer reaches the random timerand(t), within this time, if the node not listen to elec-tion information from other nodes, indicates that nonode becomes cluster head. Then, in the rand(t) time,the node broadcast cluster head election informationby using CSMA/CA channel access method. Symbol-izes that this node success becomes a cluster header.The transmit power value of cluster header selectioncome from base station by broadcast ways. The valueof all nodes are the same, the value determines thecommunication range of the cluster header.

(3) If a node’s timer has not reached the time whicha random time rand(t) generates by this node, if thenode has receive the message of cluster head nodeelection, it shows that there has a node successfullybecomes cluster header within the communication ran-ge. Then the node to give up running, so ensure thatcluster in the range of the communication has only onecluster header, meanwhile, according to the receivedheader election signal strength, the node estimates thedistance from cluster header. And the distance andthe cluster header ID are stored in the set of clusterheader; therefore, the nodes need a collection to stor-age cluster header, the table includes the cluster headID and the distance to the cluster header. Until theend of selection of the cluster header, for any clus-ter header election message, all the similar methodadopted to estimate the distance to cluster header. Andstore the cluster header ID to the collection of clusterheader.

(4) All node’s timer will be turn off, when the timerreaches the max limit value which generated in clusterheader selection stage. Then, each node who failedto run for cluster header, Select their nearest clusterheader as its cluster header in which storage in thecollection of cluster header. For join the cluster, itsends a request message to cluster header by usingCSMA/CA channel. The information include clusterheader ID, the node ID and the node’s residual energyinformation, this could as basis of selection collabo-ration node. Each cluster header receives a requestmessage from normal nodes, according to the signalstrength of the message to estimate the distance to itsmember, then stores these distance information in themember table. This table has recorded each membernode’s ID and its residual energy information. Afterthe cluster head is selected and the cluster is formed,


then the network enter working stages.

3.2.2 Working stagesIn this study, the network environment is Data per-

ceived frequently. Therefore, in working stages, alarge number of data package could received. Thedata packet contains the node’s residual energy infor-mation Ei−le f t . Cluster header could modify mem-ber node’s residual energy dynamically according toEi−le f t . Meanwhile, a number of nodes will be se-lected as virtual antennas nodes, these nodes Formsa virtual antenna matrix. The amount of virtual an-tennas related to the number of antennas of the basestation. The cluster header send command packet toits members, this command packets including: nodeID, Wake-up command, the antenna assignment in-formation such as commands. The node who receivesthe antenna command act as a antenna node. In or-der to decrease the energy consumption for send com-mand package which caused by the cluster header up-date the list of virtual antenna frequently with chang-ing of the network information, a lock is added to thecollection of antenna node, if within the lock period,the collection doesn’t update although some nodes hasmore residual energy. Before data transmission, clus-ter header using STBC encode the data package ac-cording to the antenna nodes in collection. A gen-eral round options is 1/nT , and T is each round timeslot. When working, in pre-allocate time slot, thenodes in each cluster send data package to its clusterheader according to the send power which calculatorsby cluster header, then the node enter the sleep statefor saving energy. When the cluster header receivedata which sent from its member complete, it couldfusion the data according to a certain rules for reduc-ing data redundancy and saving the sending energy.Then, based on the virtual antenna matrix informa-tion encode sending data with STBC algorithm anddetermine which the antenna node is selected, clus-ter header will broadcast the compressed data to thevirtual antenna node. Then the antenna node sendsthe encoding data to the base station with the way ofmulti-antenna collaborative.

4. Energy Analysis Model

In order to evaluate energy efficiency of DCVM pro-tocol, the network energy consumption model can bedescripted as follows:

(1) It is well known that sending and receiving com-munication energy consumption accounted for the ma-jority of the total energy consumption. Therefore, the

analysis of energy consumption, only consider the en-ergy consumption of sending and receiving, ignoringthe energy consumption which in idle and sleep state.

(2) As DCVM protocol is periodic working, afterthe cluster is established, it can repeat data transmis-sion within the round. So most of the energy con-sumed in data communication, but the energy con-sumed in cluster establishment is less, it can be ig-nored.

(3) Due to the range of cluster is not big and thesensing information has a certain correlation, it shouldbe fused, before sending, for reducing the transmis-sion energy consumption.

According to the description about data transmis-sion stage, the network’s energy consumption can bedivided into three parts: the communicate energy con-sumption between normal node and cluster header, thecommunicate energy consumption that clusters headersend data to base station with virtual antenna ways.

And each part is made up of the electric circuits con-sume energy and transmission consumes energy, De-tailed calculations as follows:

4.1 Cluster communication energy consumptionIn sensor networks, N nodes randomly deploying in

a square with length L, and the entire network is di-vided into n clusters, each cluster average N/n nodes.as the distance between node and cluster header iscloser, assume that the transmission fading factor is2, within the cluster, when a node send k bits data toits cluster header, the average energy consumption canbe expressed as:

ETotal = kEr + kEs +αKE[d2] (3)

where Es and Er respectively, per bit send/receive con-sume energy between node and cluster header,α is theparameter related to bit error rate. E[d2] is the math-ematical expectation which the square of the distancebetween node and cluster header, can be expressed as:

E[d2ch ch] =

∫ ∫(x2 + y2)ρ(x,y)dxdy

=∫ ∫

r2ρ(r,θ)drdθ(4)

where ρ(x,y) is distribution probability that any po-sition of a node in the cluster, in the entire cluster,assumption that node’s distributed probability is thesame. Then ρ(r,θ) independence in r and θ . ρ(r,θ)=

1(A2/n) . Assuming that each cluster is a circular area

with a radius equal to R,it has R = A√πn , when each


node send k bit data to cluster header, each communi-cation consumption energy can be expressed as:

Ecluster cn ch = (Nn−1)Ech ch

= (Nn−1)(KEct +Klε f s

A2

2πn+ kEcr)

(5)

4.2 Virtual MIMO energy consumptionAfter completion of space-time coding, the collaboration

nodes, will send a message to base station in MIMO way.If has the Nt node forms collaboration type transmissionantenna, Base station equipped with Nr receive antennas. itforms the MIMO communication mode. Where, the pathloss factor of the communication between cluster headerand base station is 4. And because the base station withoutenergy restriction, the base station’s power consumption isnot calculated. According to the above, the energy con-sumption which virtual node sends data to base station byMIMO way is:

Ech bs =Nn

krNc(Ect + lεmpE[d4ch bs]) (6)

which lεmp is the antenna parameters, this parameter has re-lated with error rate, antenna number Nt and base station’santenna number Nr. E[d4

ch bs] is the 4 powers mathematicalexpectation which the distance between cluster header andbase station. As the base station is located in the center ofthe network, assuming the base station height is H, thenE[d4

ch bs] can be expressed as:

E[d4ch bs] =

∫ ∫(x2 + y2 +H2)

2ρ(x,y)dxdy (7)

ρ(x,y) is the distribution probability of cluster header.

E[d4ch bs] = 0.04A2 +0.33A2H2 +H4 (8)

The cluster energy average consumption is as follows:

Ecluster = Ecluster cn ch +Ech sch +Ech bs (9)

The total energy consumption in a round can be calcu-lated as follows:

Etotal = nEcluster

= k(2NEc +NA2lε f s

2πn−2nEc− A2lε f s

2π+NE f u +muM)

(10)

Optimal cluster number n can be calculated as follows:

nopt =A√

Nls√2π

√2Ec(bNc−1)+T

(11)

T = mlε f sd4max+

mlε f s(0.04A2 +0.33A2H2 +H4)Nc(12)

5. Convergence Rate of DVCMThe convergence speed is an important indicator of the

performance evaluation of algorithmsThe proposed virtualMIMO algorithms, the convergence rate with the nodes inthe cluster during the iterations. The convergence rate ofDVCM has relations with the iterations of clustering. If theclustering phase requires too much iteration, then the nodeswill consume more energy be because of more informationwant to exchange.

During the clustering phase of DVCM, the iteration ter-mination condition of nodes is CHprob(N) = 1. So as longas known the number of iterations of nodes N when theprobability value is 1, the convergence rate of DVCM canbe obtained. From the algorithm description, it’s knownthat After N iterations, CHprob(N) = 2N−1CHprob, CHprobis initial probability values which calculated according tothe formula. For Iteration termination conditions is CHprob(N) = 1, So N is iterations of the iteration. N is calculatedby the following formula

N = [log21

chprob]+1 (13)

From the above formula, draw a conclusion that the valueof N is closely related to the initial selection probabilityCHprob of cluster head. The smaller initial probability value,the greater the number of iterations. Therefore, furthersolving the maximum limit of N, the maximum number ofiterations can be obtained.

N = limCHinitial prob→Pmin

[log21

chprob]+1

= [log21

Pmin]+1

(14)

Pmin is a constant in the formula, Therefore, the conver-gence of DVCM is σ(1). When Pmin = 0.0005, the nodecan be terminated up to 12 times iteration.

6. Simulation AnalysisIn order to verify the energy efficiency of protocol, we

evaluate the performance of DVCM and compare it withthe LEACH via simulations. We consider 250 nodes thatare uniformly deployed within a square of 200 meter ×200 meter, and assume that each node has 100J of the ini-tial energy. Each node sends 2000 bits of information to itscluster head per round. Taking into account the overhead ofprocessing control information in MIMO-based network,we add 150 bits and 100 bits processing in DVCM andLEACH. The simulation results shown in Figure 2:

Figure 2 shows the total number of nodes that remainalive over the simulation time, where Nt = 2. If the net-work lifetime is defined as the running time until there areonly 180 nodes alive in the network, we can find the net-work lifetime of DVCM can outperform twice more thanLEACH. To have this result is due primarily to the man-ner of DVCM using MIMO communication in data trans-mission, which result in the lower BER and higher data


Figure 2 Network nodes running time and the relationshipbetween survival

Figure 3 The average residual energy of node and networkoperation time

throughput within the entire cluster, and thus can save moreenergy for node rapidly converting from the working stateto the sleeping state. In addition, you can also see from Fig-ure 2, the performance of DVCM further improved whenthe number of virtual antenna node is increased to 4. Thisis because more antenna nodes can lead to lower BER andhigher data throughput. Therefore, in sparse sensor net-works using MIMO technology can significantly improveenergy efficiency, thus extending the network lifetime.

From Figure 3, Simulation result can draw the conclu-sion, as the network operation time increased. DVCM pro-tocol is superior to LEACH protocol. Also show that num-ber of antennas is 4, DVCM protocol with 4 antennas is su-perior to DVCM protocol with 2 antennas. This is also be-cause of the lower error rate to decrease and data through-put to increase, leading to more energy efficient network.But the simulation also shows that the number of virtual an-tennas node over 10% of cluster nodes, The DVCM proto-col’s performance reduces deeply. Under data frequent sen-sation environment, this is because of that on the one hand,

Figure 4 The total number of data received at the BS pergiven amount of dead nodes

Figure 5 The total number of data received at the BS pergiven amount of energy consumed

this can increase the complexity of virtual antenna space-time coding, namely, the computational cost increase, onthe other hand, the packet is broken down into more packetdata transmission, And because the number of antenna in-crease, leading nodes send, the receive processing is redun-dant. So has counter-balanced the advantage which MIMObrings. Generally speaking, the best number of antenna isabout 10% of the nodes in the cluster.

Above simulation results show that under the same num-ber of death nodes, LEACH algorithm for the Amount ofreceives data of base station which the receives of LEACHalgorithm is less than DVCM algorithm. This is becausethe DVCM using virtual MIMO technology improves theefficiency of data transmission, when transmission the sameamount of data, the algorithm of DVCM consumes lessenergy than consumes of the LEACH algorithm. There-fore, the same numbers of death nodes, the actual networkrunning time of DVCM is longer than the running time ofLEACH algorithm.


7. ConclusionIn this paper, DVCM routing protocol is proposed, then

the protocol is described in detail, DVCM protocol usinghierarchical network topology structure, combine MIMOtechnology with clustering topology structure, and intro-ducing dynamic virtual antenna mechanism. The prob-lem that cannot find enough nodes as collaborative node,when using MIMO technology in sparse wireless sensornetworks, is Solved. Thus succeeds applies the MIMOtechnology in the sparse network. And, the energy con-sumption is analyzed; the total consuming energy formulais obtained. By the simulation analysis, DVCM is con-firmed which is of the higher performance in the energyconservation aspect.

8. AcknowledgmentsThis work is supported by Doctoral Foundation of Liaon-

ing Province under Grant No.20081027.

References[1] C. Shuguang, A. J. Goldsmith and A Bahai, “Energy-

Efficiency of MIMO and Cooperative MIMO Tech-niques in Sensor Network”, IEEE Journal on SelectedAreas in Communications, vol. 22, Aug. 2004, pp.1089-1098.

[2] S. K. Jayaweera, “Virtual MIMO-Based Coopera-tive Communication for Energy-Constrained Wire-less Sensor Networks”, IEEE Trans. Wireless Com-mun., Vol. 5, May. 2006, pp. 984-989.

[3] A. M. Abbosh, D. Thiel, “Performance of MIMO-based wireless sensor networks with cochannel inter-ference”, Proc. International Conference on Intelli-gent Sensors, Sensor Networks and Information Pro-cessing, IEEE Press, Dec., 2005, pp. 115-119.

[4] S. K. Jayaweera, “An Energy-Efficient Virtual MIMOArchitecture Based on V-BLAST Processing for Dis-tributed Wireless Sensor Networks”, Proc. IEEE In-ternational Conference on Sensor and Ad-hoc Com-munications and Networks(SECON 04), IEEE Press,Oct.,2004, pp. 299-308.

[5] G. N. Bravos, A. G. Kanatas, “Energy efficiencyof MIMO-based Sensor Networks with a Coopera-tive Node Selection Algorithm”, Proc. IEEE Inter-national Conference on Communications(ICC 07),IEEE Press, Jun. 2007, pp. 3218-3223.

[6] K. R. Xu, Z. Y. Yang, et al, “An End to end Op-timized Cooperative Transmission Scheme for Vir-tual MIMO Sensor Networks”, Proc. IEEE Singa-pore International Conference on CommunicationSystems(ICCS 2006).

[7] S. P. Laxminatayana, V. Krishnamurthy, “OptimalRate and Number of Clusters in Cluster based Co-operative MIMO Sensor Network”, Proc. Signalssystems and computers, IEEE Press, 2005, pp. 1556-1560.


An Application of Virtual MIMO Technology in WSN … · An Application of Virtual MIMO Technology...

Documents

Transcript of An Application of Virtual MIMO Technology in WSN … · An Application of Virtual MIMO Technology...