A NOVEL ENERGY EFFICIENCY PROTOCOL FOR WSN BASED ON OPTIMAL CHAIN ROUTING
- 1. Proceedings of the 10th
INDIACom; INDIACom-2016; IEEE Conference ID: 37465
2016 3rd
International Conference on “Computing for Sustainable Global Development”, 16th
- 18th
March, 2016
Bharati Vidyapeeth's Institute of Computer Applications and Management (BVICAM), New Delhi (INDIA)
Copy Right © INDIACom-2016; ISSN 0973-7529; ISBN 978-93-80544-20-5 488
A NOVEL ENERGY EFFICIENCY PROTOCOL FOR
WSN BASED ON OPTIMAL CHAIN ROUTING
Arun Agarwal
PhD Research Scholar, GGSIPU
New Delhi, India
Email: arun.261986@gmail.com
Khushboo Gupta
PhD Research Scholar, UPTU
Lucknow, India
Email:khushboogupta2806@gmail.com
K P Yadav
Director, KCCITM,
Greater Noida|, India
Email:drkpyadav732@gmail.com
Abstract – Energy Efficiency in Wireless Sensor Network is
one of the most significant aspects of routing in these
networks. WSN consist of thousands of sensory nodes densely
distributed over wide geographical network. As these nodes
are deployed in remote areas where recharging is not
possible, even if it is possible it will incur high cost. So there
is a need of a protocol which facilitates less energy dissipation
and thereby enhances the overall performance of the
network. We surveyed several protocols such as LEACH,
PEGASIS, ACT etc. and concluded that important
performance measures are First Node Die (FND), Half Node
Alive (HNA) and Last Node Alive (LNA). Values for above
mentioned parameters vary for different protocols. In this
paper we present a new protocol Energy Efficient Optimal
Chain Protocol (EEOC) which outperforms all above
mentioned protocols. We compared the results of all these
protocols with EEOC and found that with respect to FND,
HNA and LNA EEOC performs way better than others.
Keywords – ACT, EEOC, Energy Efficiency, LEACH,
PEGASIS, Routing Algorithm, Simulation, WSN
Abbreviation:
1. ACT Arranging Cluster Range and Transmission
Ranges
2. EEOC Energy Efficient Optimal Chain
3.LEACH Low energy adaptive clustering hierarchy
4.PEGASIS Power-Efficient GAthering in Sensor
Information Systems
5. FND First Node Die
6. LNA Last Node Alive
7. HND Half Node Die
8. LNA Last Node Alive
9. SN Sensor Node
10. BS Base Station
11. MTE Multipath transmission energy efficient
protocol
12. DCP Direct communication protocol
13. WSN Wireless Sensor Networks
14. CH Cluster Head
I. INTRODUCTION
Wireless Sensor Networks [1] consists of thousands of
energy constrained Sensor Nodes (SNs) [2] which are spatially
distributed over various geographical locations for performing
their task that focus on simple data gathering applications.
These SNs gathers data by sensing the environment and then
send the same using wireless channel to remote Base Station
(BS). SNs are self-configured but have constrained resources
and therefore have limitations in computation and
communication abilities. Moreover, these nodes are battery
operated often limiting energy availability as battery may get
depleted in the network. Although WSN are generally
stationary after deployment but replacement of the battery
may not be feasible due different geographical regions. Hence
there is a requirement of effective and efficient solutions
during various operations of data collection, processing,
communication and storage. It has been studied in literature
that in comparison to processing, communication from SNs to
BS dissipates more energy as data transmission depends upon
the area between source and destination.
In recent years, various routing protocols [4] based upon
clustering have been developed and deployed to reduce the
network traffic from SNs to BS. Cluster-based routing
protocols [3] are based upon hierarchical routing [17] in which
nodes are arranged in different hierarchical levels sustained by
parameters such as energy levels, processing capability etc.
The low energy nodes can be used to observe and perform the
sensing in the proximity of the target while the higher energy
nodes can be used to process the data and send the
information. Thus by creating cluster and assigning special
tasks to CHs, the overall scalability, energy efficiency and
network lifetime can be increased. These CHs can be chosen
from the deployed SNs or can be more powerful than other
SNs in terms of energy, processing power, bandwidth, and
memory. In this paper, we have considered the node
heterogeneity for hierarchical cluster based routing. The
concept of advanced node has been introduced which are the
nodes having large initial energy relative to other nodes in the
network.
The objective of this paper is to design and develop a novel
energy efficient protocol for WSNs. The solution is
- 2. A NOVEL ENERGY EFFICIENCY PROTOCOL FOR WSN BASED ON OPTIMAL CHAIN ROUTING
Copy Right © INDIACom-2016; ISSN 0973-7529; ISBN 978-93-80544-20-5 489
implemented and the performance of the protocol is evaluated
by comparing it with the existing routing protocols like
LEACH, PEGASIS and ACT.
The organisation of the paper as follows: Section II addresses
the work Literature Survey carried out in this field. Section III
gives Problem Statement. Section IV gives a brief description
on system design and implementation followed by Algorithms
of routing protocols in Section V. Section V contains the
Simulation Results. Section VI gives the conclusions and future
works enhancement possible.
II. LITERATURE SURVEY
This section is the description of various routing algorithms
[13] as proposed for WSN. It also includes brief comparison of
each one of them with the advantages and disadvantage of the
underlying mechanism (Table 1).
Direct communication protocol (DCP) which involves direct
communication of each node to the base station and thereby
result in very large dissipation of energy in each round for
those nodes which are very situated far away from the BS [7].
Because it follows radio model for energy dissipation where
the dissipated energy is directly proportional to square of the
distance between source and the sink. Thus depleting the area
far away from the BS. Apart from these limitations DCP is still
very simple to implement and it does not impart any
complexities for calculation in each round and thereby
decreasing the overall delay.
Multipath transmission energy efficient protocol (MTE)
underlying architecture relies on the basis of multipath
propagation which involves relaying of data packets from
farthest node to its neighbour node and repeating the same
process until data arrives at the node very close to the BS
whose responsibility is to transfer the collected data to the BS.
This is a efficient protocol as it collects information from
farthest nodes too and imparting the load on each node in
between but imposing an additional work on nodes closer to BS
thus resulting into a situation where nodes closer to BS dies
very early thus the situation of first node die occurs very soon
and performance degrades.
Next protocol which removes all the disadvantages in DCP and
MTE by applying a phenomenon of clusters and assigning a
leader with each cluster designated as cluster head (CH).
LEACH (Low energy adaptive clustering hierarchy) [5, 9]
divides the entire area under consideration into number of
clusters with each cluster having a CH which communicate
directly to the BS. The remaining nodes will transmit data to
the CH whose responsibility is to aggregate it and then transmit
it directly to the BS. To avoid extra load on CH in every round
LEACH proposed a mechanism for load sharing among all the
nodes resides in the cluster itself by applying a probabilistic
strategy which states that if a given node is CH in current round
than it will not become CH in the coming rounds until and
unless all the nodes remaining in the cluster will not become
CH at least once. This approach will give fair chance to every
node to become a CH. It achieves a factor of 8 reduction in
energy dissipation as compared with conventional routing
protocols. But as it applies that CH have to directly
communicate with BS, clusters far away will die quickly as
compared to the one situated closer to BS.
PEGASIS (Power-Efficient GAthering in Sensor Information
Systems) is another protocol used for routing in WSN which is
based upon constructing a chain applying some greedy
strategy for formation of the chain and transmitting data
packets starting from the farthest node to the CH along the
path as directed by the chain constructed. The main idea of
this is to receive and transmit only to the close neighbors thus
saving the overall energy dissipation [11]. This approach will
distribute the energy node evenly among all the sensor nodes
in the network which are organize to form a greedy chain.
It performs the data fusion at every node except the farthest
node in the chain. It will strictly allow a node to become the
CH only when the entire remaining node will become the CH
at least once. It will attain an improvement of two times the
number of round compared to leach and will provide a near
optimal performance.
TABLE I. COMPARISION OF VARIOUS ROUTING PROTOCOL IN WSN
Routing
Protocol
Advantage Disadvantage
DCP [6]
Simple & Easy to
implement
Overall Delay is reduced
Large Energy
dissipation
Reduced Network
Lifetime
MTE[8] Multipath propagation
Load Balancing
Overload on nodes
near to BS
Delay Increases
BCDCP[12]
Centralized Control
Reduced Energy
Dissipation
Overload on BS
Bottleneck
Congestion
LEACH[6]
Cluster Based Approach
Load Sharing using
probabilistic CH
determination
CH are heavily loaded
Reduction in overall
network lifetime
PEGASIS[14]
Uses greedy chain
strategy for
communication
Near optimal
performance
Delay in
determination of CH
Heavy Load on CH
ACT[15]
Prolonged Network
Lifetime
Cross Level Transmission
Complex to determine
no. of clusters
Overload on nodes
near to BS
At last we are describing the ACT (Arranging Cluster Range
and Transmission Ranges) which will become a basis for our
proposed work also which is the adaptive clustering protocol
that reduces the load on the cluster near the base station by
reducing the cluster size near to base station as they are the
one which exhaust their power quickly because they are
heavily loaded. ACT determines the cluster size and cluster
radios and thus calculating total number of clusters in the area
defined. Another approach that distinguishes it with LEACH
is that there is a direct transmission from CH to BS in LEACH
whereas ACT proposed that CH will transmit their available
data up into the level where each cluster belongs to one or
another level such as first level is the one nearest to the BS
- 3. Proceedings of the 10th
INDIACom; INDIACom-2016; IEEE Conference ID: 37465
2016 3rd
International Conference on “Computing for Sustainable Global Development”, 16th
- 18th
March, 2016
Copy Right © INDIACom-2016; ISSN 0973-7529; ISBN 978-93-80544-20-5 490
[16].ACT also suggested cross level transmission to improve
the overall quality of the network.
ACT enhances the performances of entire network by
increasing the overall network life time which may be stated as
number of round before the first sensor node uses up its energy
in the network.
III. PROPOSED PROBLEM
ACT assumes the following to build up its own system model:
The positions of SNs and BS are fixed.
The energy of all SNs kept initially the same.
SNs are uniformly dispersed in the sensor field with
density dn.
In each round a SN transmits one unit of data to the CH
A CH aggregates data and then forwards to BS.
SNs can recognize their geographical position and the BS
position via exchanging information
SNs can adjust their energy levels. For transmitting data
to BS directly, the maximum energy level can also be
used.
The covered area is a W X L rectangle, where W is the
width and L is the length.
It will assume that in each round every sensor node will
transmit its data to CH, thus imposing an overhead on CH with
each received data thus reducing the energy level of CH very
soon and thereby reducing the overall network lifetime.
Whereas if we use another data aggregation scheme like
PEGASIS, it will create a dynamic chain in each round
collecting information from each sensory node and forwarding
the data up only to its neighbor up in the chain to the last node
which directly transmit this aggregated data to the CH.
This reduces the overall load on CH as it receives from only
two nodes that are designated as last node in their own chain.
But still there exists a problem with PEGASIS that CH has to
transmit their collected data directly to the BS thus reducing the
network lifetime.
ACT deploys multi hop communications in which less energy
is consumed from each sensor node. The load on each CH is
balanced by considering cluster sizes and CH locations. Energy
dissipation in ACT is very much lower than other available
protocols but still does not enhances the overall network
lifetime. Our research measures the network lifetime which
may be defined as the number of rounds before the first sensor
node avail up its energy in the network. We suggest that even
though the CHs are selected and multi hop communication is
there but the CH will be less utilized as it will dissipate less
energy in receiving data from all nodes in its cluster. Rather it
will be very much beneficial that we will create a chain as used
in PEGASIS so that energy dissipation in CH is only due to
two nodes not by all the nodes in its cluster. Thereby reducing
the overall energy dissipation and enhances the overall network
lifetime.
IV. SYSTEM MODEL AND IMPLEMENTATION
A. NETWORK MODEL
Figure 1 shows the organization of SNs, CHs and BS at
various levels in the network. CHs are chosen based upon
some criteria for monitoring the SNs. Each CH sends the data
to the BS.
Fig 1. Network Model
Our network model for simulation will assume the following:
WSN include several static sensory nodes and a
single base station.
Sensor nodes are located uniformly in network field.
Position of base station is fixed and it is centrally
located.
Sensory nodes are having limited battery power.
B. ENERGY MODEL
The energy model for our proposed protocol will consist of
following energy components:
Initial energy Ei
Energy for data aggregation/sensing Ea
Energy for data receiving Er
Energy for data transmission Et
Residual Energy Eres
where Eres = Ei − (Ea+Er+Et)
C. OUR APPROACH
We have used this radio model for energy dissipation with
respect to each round and iteration there is a reduction of
energy that is being dissipated and for each node ordered from
zero to n-1, total energy dissipation may be given by the
following energy equations
E0 = initial – ETx to E1
E1 = initial – ERx from E0 – aggregation – ETx to E2
E2 = initial – ERx from E1 – aggregation – ETx to E3
o
o
o
En = initial – ERx from En-1 – aggregation – ETx to BS
As it is being shown in the figure 2 that it comprises of a
series of nodes connected with each other with a temporary
chain. In our work we have divided our whole area under
consideration into two equal halves. As node density is
uniform we have just partitioned the given area into almost
equal number of nodes where each node can have the location
- 4. A NOVEL ENERGY EFFICIENCY PROTOCOL FOR WSN BASED ON OPTIMAL CHAIN ROUTING
Copy Right © INDIACom-2016; ISSN 0973-7529; ISBN 978-93-80544-20-5 491
of every other node in its area. Thus we have suggested to form
an optimal chain based on inter distance between nodes where
it is being advised that nodes communicate to only that node
which is nearest from the given node constructing an optimal
chain. This optimality brings a huge improvement in overall
performance of the network as energy dissipated is directly
proportional to square of the distance and we are using a chain
where each node communicate to the one located at smallest
distance. At each step the given node receives data from its
predecessor node, aggregate it with its own data and then
transmit it along the chain to the next node coming in the chain.
This process starts from the two nodes situated at very far end
in both the regions and progress along the connections given by
that temporary chain up to the node nearest to the BS which is
being designated at CH shown in figure 2.
Fig 2. Temporary Chains in EEOC
In every iteration the residual energy of each node is calculated
and the determination of dynamic chain, two farthest nodes and
the node nearest to BS i.e. CH is done. In each iteration the
above said process is repeated calculating total number of dead
nodes in each round. If the number of dead nodes is one i.e. if
there exists a situation of first node die (FND) then the current
round is stored as it indicates FND used for analyzing network
performance. Similarly calculations is done for half node alive
(HNA) and last node alive (LNA).
The entire algorithm is executed multiple times so as to obtain
average values to compare the network lifetime and efficiency
of our proposed algorithm.
V. EVALUTION PARAMETER PSEUDO CODE FOR
EACH PROTOCOL
A. LEACH PROTOCOL
1. Initialize system model parameters
Base Station coordinates (50,300)
Area under consideration (100,100)
Number of nodes (n=100)
Initial energy values (Eo=0.5J)
Packet length (k=2000)
2. Create a random WSN consisting of n nodes randomly
spaced in the area defined above.
3. Set probability value p=0.05 for determination of CH
4. For each round where maximum numbers of rounds are 9999
repeat steps 5 to 9
5. Determine the CH for each cluster using the given
probability condition
'
)
1
mod(1
)(
p
rp
p
nT
If n € G
6. Determine number of dead nodes
If (dead==1)
Output first_node_die
If (dead==n)
Output last_node_alive
Exit from for loop step 4
7. Calculate the remaining energy values
8. Reassign probability values for determination of CH.
9. Plot x and y values to show no. of rounds and no. of nodes
alive.
10. EXIT
B. PEGASIS PROTOCOL
1. Initialize system model parameters
Base Station coordinates (50,300)
Area under consideration (100,100)
Number of nodes (n=100)
Initial energy values (Eo=0.5J)
Packet length (k=2000)
2. Create a random WSN consisting of n nodes randomly
spaced in the area defined above.
3. Repeat steps 4 to 11 till there exists a single node alive,
4. Determination of Greedy Chain and CH
5. Determine number of die nodes
If (dead==1)
Output first_node_die
If (dead==n)
Output last_node_alive
Exit from for loop step 3
6. Calculation of energy dissipation using aggregation and
data forwarding in each step.
7. Calculation of energy dissipation of CH.
8. Plot x and y values to show no. of rounds and no. of nodes
alive.
9. EXIT
C. ACT PROTOCOL
1. Initialize system model parameters
Base Station coordinates (50,300)
Area under consideration (100,100)
Number of nodes (n=100)
Initial energy values (Eo=0.5J)
Packet length (k=2000)
- 5. Proceedings of the 10th
INDIACom; INDIACom-2016; IEEE Conference ID: 37465
2016 3rd
International Conference on “Computing for Sustainable Global Development”, 16th
- 18th
March, 2016
Copy Right © INDIACom-2016; ISSN 0973-7529; ISBN 978-93-80544-20-5 492
2. Create a random WSN consisting of n nodes randomly
spaced in the area defined above.
3. Determination of number of clusters and respective radius.
4. Determination of number of nodes in each cluster.
5. Assigning a random node as CH satisfying the probability
condition as given in the proposed model.
6. Repeat steps 7 and 8 till total no. of died nodes is equals to n.
7. Calculate the following parameters for the proposed work in
each iteration
FND, HND & LNA
C1, C2, C3 representing no. of nodes in b/w FND to HNA
B1, B2, B3 representing no. of nodes in b/w HNA to LNA
8. Calculate the remaining energy values for each node in every
round.
9. Form a table which shows the values as calculated in step 7.
10. EXIT.
D. PROPOSED PROTOCOL (EEOC)
1. Initialize system model parameters
Base Station coordinates (50,300)
Area under consideration (100,100)
Number of nodes (n=100)
Initial energy values (Eo=0.5J)
Packet length (k=2000)
2. Create a random WSN consisting of n nodes randomly
spaced in the area defined above.
3. Repeat steps 4 to 11 till there exists a single node alive,
4. Determination of nearest node and two farthest nodes.
5. Determination of OPTIMAL Chain 1 in left of the area under
consideration.
6. Determination of OPTIMAL Chain 2 in right of the area
under consideration.
7. Determine number of die nodes
If (dead==1)
Output first_node_die
If (dead==n)
Output last_node_alive
Exit from for loop step 3
8. Calculation of energy dissipation consumption for left area
9. Calculation of energy dissipation consumption for right area
10. Calculation of energy dissipation consumption for node
nearest to Base Station i.e. CH
11. Plot x and y values to show no. of rounds and no. of nodes
alive.
12. EXIT
VI. SIMULATION AND RESULTS
We have done simulations to evaluate the performance of the
proposed EEOC protocol and the other three protocols,
LEACH, PEGASIS and ACT. First of all, we have list the
simulation parameters. Secondly, we present the simulation
results, which show the performance results under different
performance metrics. Finally, we discuss and analyses the
simulation results.
A. Simulation Parameters
TABLE 2. COMPARISION OF VARIOUS ROUTING PROTOCOL IN WSN
Parameter Value
Network Size 100 X 100
Base Station Location 50 X 300
Number of Sensor Nodes 100
Density 1 node/ 100 m2
Initial Energy of each sensor
nodes
0.5 J
Energy Transmission 50 nJ
Energy Reception 50 nJ
Energy Data Aggreegation 5 nJ
Packet Data Size 2000 bits
Eelec
50 nJ/bit
Єamp
100 pJ/bit/m2
B. Analysis Graph for FND
Fig 3. Analysis Graph of FND
2. Analysis Graph for HNA
Fig 4. Analysis Graph of HN
- 6. A NOVEL ENERGY EFFICIENCY PROTOCOL FOR WSN BASED ON OPTIMAL CHAIN ROUTING
Copy Right © INDIACom-2016; ISSN 0973-7529; ISBN 978-93-80544-20-5 493
3. Analysis Graph for LND
Fig 5. Analysis Graph of LND
VII. CONCLUSION AND FUTURE SCOPE
From the above simulation results it will be concluded that the
said protocol i.e. EEOC routing protocol for WSN provides
better results as compared from all other protocols such as
LEACH, PEGASIS and ACT. The performance comparison of
all these protocols are being shown in terms of simulation
results as depicted in figure 3, 4 and 5. All these will reflect
that in case of first node die, ACT will perform slightly better
than our proposed approach EEOC but in all other performance
criteria that is used to measure the performance and network
lifetime our proposed approach will outperform than rest of the
protocols analysed. EEOC will have drastic improvement over
all other protocols in case of half node alive and last node alive.
It will attain approximately 150 percent improvement as it is
clear from the simulation results. Also it may be concluded
from the study that these results can easily be used to decide for
which application the given protocols may be used. It will be
benefitted for application specific purposes where one have to
choose from available options which protocol is very effective
in the given scenario. Our study results in optimal coverage and
prolonged network lifetime.
In future we may also improve the first node die component of
the EEOC protocol, so that it will be used in any specific
application as the one which is optimal among all the available
protocols designed. Another task which we will focus is of
changing the cluster architecture hence to provide some better
results and changing the entire procedure so as to suit
heterogeneous environment also.
REFERENCES
[1] F. L. LEWIS, D.J. Cook and S.K. Das: Wireless Sensor
Networks, in Wiley, New York, 2004 John Wiley & Sons, Inc.,
Publication.
[2] Waltenegus Dargie and Christian Poellabauer: Fundamentals
of Wireless Sensor Networks - Theory and Practices, Wiley
series on wireless and mobile computing, 2010 John Wiley &
Sons, Inc., Publication.
[3] J. Hong, J. Kook, S. Lee, D. Kwon, S. Yi, T-LEACH: the
method of threshold-based cluster head replacement for
wireless sensor networks, Information Systems Frontiers
(2009) 513–521.
[4] J.N. Al-Karaki, A.E. Kamal, Routing techniques in in wireless
sensor networks, IEEE Wireless Communication 11 (6)
(2004) 6- 28.
[5] Y.C. Chang, Z.S. Lin,J.L. Chen Cluster based self-
organization management protocols for wireless sensor
networks, IEEE Transaction on Consumer Electronics
52(1)(2006) 75-80.
[6] W.R. Heinzelman, A. Chandrakasan, H. Balakrishnan,
Energy-efficient communication protocol for wireless
microsensor networks, in: Proceedings of Hawaii
International Conference on System Sciences (HICSS),
January 2000, pp. 1–10.
[7] G. Xing, T. Wang, Z. Xie, and W. Jia, “Rendezvous Planning
inWireless Sensor Networks with Mobile Elements,” IEEE
Trans.Mobile Computing, vol. 7, no. 12, pp. 1430-1443, Dec.
2008.
[8] Long Cheng, Sajal K. Das, Jiannong Cao, Canfeng Chen and
Jian Ma, Minimum Transmission Multicast Routing Protocol
for WSN : 2010 39th International Conference on Parallel
Processing.
[9] H.M. Abdulsalam, L.K. Kamel, W-LEACH: Weighted Low
Energy Adaptive Clustering Hierarchy aggregation algorithm
for data streams in wireless sensor networks, in: Proceedings
of IEEE International Conference on Data Mining Workshops
(ICDMW), 2010, pp. 1–8.
[10] M.O. Farooq, A.B. Dogar, G.A. Shah, MR-LEACH: Multi-
hop Routing with Low Energy Adaptive Clustering
Hierarchy, in: Proceedings of IEEE International Conference
on Sensor Technologies and Applications (SENSORCOMM),
July 2010, pp. 262–268.
[11] F. Bounbdallah, N. Bouabdallah, R. Boutaba, On Balancing
energy consumption in wireless wireless sensor networks,
IEEE Transaction on Vehicular Technology 58(6) (2009)
2909-2924.
[12] Y. Liu, J. Gao, Y. Jia, L. Zhu, A cluster maintenance
algorithm based on LEACH-DCHS protocol, in: Proceedings
of IEEE International Conference on Networking,
Architecture, and Storage, June 2008, pp. 165–166.
[13] S.D. Muruganathan, D.C.F. Ma, R.I. Bhasin, A.O. Fapojuwo,
A centralized energy-efficient routing protocol for wireless
sensor networks, IEEE Communications Magazine 43 (3)
(2005) 8–13.
[14] S. Lindsey, C. Raghavendra, PEGASIS: power-efficient
gathering in sensor information systems, in: Proceedings of
IEEE Aerospace Conference, vol. 3, 2002, pp. 1125–1130.
[15] HWei Kuang Lai , Chung Shuo Fan, Lin Yan Lin: Arranging
cluster sizes and transmission ranges for wireless sensor
networks, in: Elseviere Information Sciences 183 (2012) 117–
131.
[16] M.B. Yassein, A. Al-zou’bi, Y. Khamayseh, W. Mardini,
Improvement on leach protocol of wireless sensor network
(vleach), International Journal of Digital Content Technology
and its Applications 3 (2) (2009) 132–136.
[17] C. Alippi, R. Camplani, M. Roveri, An adaptive LLC-based
and hierarchical power –aware routing algorithm , IEEE
Transaction on Instrumentation and Measurement
58(9)(2009) 3347-3357.