14-06-2009, 12:54 AM
Mobility Metric based LEACH-Mobile Protocol
G. Santhosh Kumar
#1
, Vinu Paul M V
*2
, K. Poulose Jacob
#3
#
Department of Computer Science, Cochin University of Science and Technology
Cochin “ 682 022, Kerala, INDIA
1,3
{san,kpj}[at]cusat.ac.in
*
Centre for AI and Robotics
Bangalore, INDIA
2
vinumanayil[at]gmail.com
Abstractâ€Cluster based protocols like LEACH were found
best suited for routing in wireless sensor networks. In mobility
centric environments some improvements were suggested in the
basic scheme. LEACH-Mobile is one such protocol. The basic
LEACH protocol is improved in the mobile scenario by ensuring
whether a sensor node is able to communicate with its cluster
head. Since all the nodes, including cluster head is moving it will
be better to elect a node as cluster head which is having less
mobility related to its neighbours. In this paper, LEACH-Mobile
protocol has been enhanced based on a mobility metric
remoteness for cluster head election. This ensures high success
rate in data transfer between the cluster head and the collector
nodes even though nodes are moving. We have simulated and
compared our LEACH-Mobile-Enhanced protocol with LEACH-
Mobile. Results show that inclusion of neighbouring node
information improves the routing protocol.
I. I
NTRODUCTION
Wireless
sensor networks
[1][2] are promising
unprecedented levels of access to information about the
physical world, in real time. Many areas of human activity are
starting to see the benefits of utilizing sensor networks. Some
of the real deployments include UC Berkleyâ„¢s Smart Dust,
MITâ„¢s
µ
-Adaptive Multi-domain Power aware Sensors and
UCLAâ„¢s Wireless Integrated Sensor Networks. In almost all
such cases, sensor networks are statically deployed. In static
networks, the mobility of sensors, users and the monitored
phenomenon is totally ignored. The next evolutionary step for
sensor networks is to handle mobility in all its forms. One
motivating example could be a network of environmental
monitoring sensors, mounted on vehicles used to monitor
current pollution levels in a city. In this example, the sensors
are moving, the sensed phenomenon is moving and the users
of the network move as well.
The dynamic nature of mobile wireless sensor networks
introduces unique challenges in aspects like data management,
accuracy and precision, coverage, routing protocols, security
and software support.
Many of the above mentioned
challenges related to a static deployment of the sensors are
well addressed by the researchers. One of the most important
constrains on sensor nodes is the route enabling when the
nodes keep moving. It has been reported that the clustering
mechanisms and hierarchical routing make huge improvement
in sensor networks in terms of energy consumption and
efficient data gathering. Such improvement is due to the
structure of the network, assumed before the deployment of
the sensor nodes. Once the network becomes dynamic we do
not have the freedom to pre-assume such structures. The
conventional routing protocols for static sensor networks are
to be optimized once mobility is introduced. To study the
performance of routing protocols under such conditions, we
have to consider the mobility patterns and associated mobility
metrics.
In this paper we propose an improvement to the LEACH-M
protocol, which is suitable for mobile wireless sensor
networks. The basic idea of this LEACH-Mobile-Enhanced
(LEACH-ME) protocol is to make sure as much as possible
that the cluster heads are from the group of mobile nodes
having minimum node mobility or they are in a group motion
with the other cluster members (as in RPGM model [3]). By
doing the modified election process for cluster heads or
modified rotation of duty of cluster heads, the protocol makes
sure that the clusters are disturbed minimally in the event of
movement of cluster heads.
II. RELATED WORK
This section briefly outlines the related work in mobile
sensor network (MSN) and LEACH protocol improvements.
A. Mobile Sensor Network (MSN) Research
Researchers have only recently started to study the sensor
movement and unique attributes of mobile sensor networks
since the sensor networks were originally assumed to consist
of only static nodes. It has been suggested [4] that the mobility
of sensor nodes improves the sensing coverage. Robotic Fleas
project in Berkeley [5], Robomote [6] and Parasitic Mobility
[7] were attempts to enable mobility in sensor networks. It is
shown that Ëœdata muleâ„¢ [8] approach can be used to efficiently
collect the data by minimizing the data delivery latency with
minimum energy consumption in a controlled mobile sensor
network. A number of approaches exploiting mobility for data
collection can be found in [9]. Adaptive Sampling and
Prediction (ASAP) [10] is a real world application of MSN
where a fleet of undersea mobile sensor nodes coordinate and
collect measurements of ocean without human intervention. A
sensor network based adaptive navigation system is discussed
in [11]; where sensors equipped on vehicles collect real-time
traffic information and exchanges among the neighbour
vehicles.
Muneeb Ali et al [12] discuss a mobility management
service layer in SensorNet Protocol which is a cross layer
978-1-4244-2963-9/08/$25.00 © 2008 IEEE
ADCOM 2008
248
Authorized licensed use limited to: COCHIN UNIVERSITY OF SCIENCE AND TECHNOLOGY. Downloaded on February 7, 2009 at 06:48 from IEEE Xplore. Restrictions apply.Page 2
approach where mobility information is stored in a database
so that it is visible across all layers. A new concept called
network dynamics is introduced in [13] to solve mobility
management issues. This work is an earlier attempt to
formulate laws that govern mobility motivated by classical
dynamics that study the movement of objects. In short,
mobility of sensor nodes is of great importance and there is an
uprising research trend towards leveraging node mobility to
enhance network performance in terms of energy efficiency,
coverage, lifetime, localization and fault tolerance.
B. LEACH Protocol enhancements
Low Energy Adaptive Clustering Hierarchy (LEACH) [14],
is one of the most popular hierarchical routing protocols for
wireless sensor networks. The idea is to form clusters of the
sensor nodes based on received signal strength indicator
(RSSI) and use local cluster heads as routers to the sink.
LEACH has motivated the design of several other protocols
which try to improve upon the cluster-head selection process
[15][16].
LEACH protocol does not consider the mobility of sensor
nodes. In mobility centric environments an agent based data
collection scheme is put forward in [17]; where a mobile agent
effectively process the data and saves the total energy spent by
the network. LIMOC [18] is a scheme to enhance the life time
of network in which energy rich moving cluster heads
collaborate intelligently each other to route the data to the
base station. The Enhancement to LEACH to support mobility
is introduced as LEACH-Mobile, in short LEACH-M [19].
The basic idea in LEACH-M is to confirm whether a mobile
sensor node is able to communicate with a specific cluster
head.
III. LEACH MOBILE ENHANCED PROTOCOL
We adopt the proposals in the LEACH-M protocol and
extend it by proposing remoteness concept for cluster head
election. This section explains the cluster head election and
maintenance of both LEACH-M and LEACH-ME protocols.
A. LEACH Routing Phases
The LEACH operations are mainly in two major phases -
Set-up phase and Steady-state phase. Set-up phase is the initial
one and this is the phase where all cluster formation takes
place. This phase is relatively short compared to the steady-
state phase. In this phase, one of the basic ideas in LEACH-
ME is to confirm the election of specific cluster heads which
either have no node movement or minimum relative node
movement.
In the steady-state phase, the cluster head and non-cluster
head nodes receive a particular message at a given time slot
according to TDMA time schedule of sensor cluster, and then
reorganize the cluster with minimum energy consumption.
The steady state phase does the actual data transfer between
the sensing node and the sink.
B. Cluster Head Election and Maintenance in LEACH-M
LEACH-M uses the same set-up procedure used in the
basic LEACH protocol. In LEACH, the nodes organize
themselves into local clusters, with one node acting as the
local base station or cluster-head. If the cluster heads are
chosen a priori and fixed throughout the system lifetime, as in
conventional clustering algorithms, it is easy to see that these
sensors chosen to be cluster-heads would die quickly due to
overloading, ending the useful lifetime of all nodes belonging
to those clusters. Thus LEACH includes randomized rotation
of the high-energy cluster-head position such that it rotates
among the various sensors in order not to drain the battery of a
single sensor. In addition, LEACH performs local data fusion
to compress the amount of data being sent from the clusters
to the base station, further reducing energy dissipation and
enhancing system lifetime.
Sensors elect themselves to be local cluster-heads at any
given time with a certain probability. These cluster head nodes
broadcast their status to other sensors in the network. Each
sensor node determines to which cluster it wants to belong by
choosing the cluster-head that requires the minimum
communication energy
2
. Once all the nodes are organized into
clusters, each cluster-head creates a schedule for the nodes in
its cluster. This allows the radio components of each non-
cluster-head node to be turned off at all times except during its
transmit time, thus minimizing the energy dissipated in the
individual sensors.
C. Cluster Head Election and Maintenance in LEACH-ME
In LEACH the election and cluster head rotation makes
sure that the cluster heads do not die due to prolonged extra
work. This is done by the random rotation of the cluster head
duty across the nodes in the cluster by considering the energy
level of the nodes. In view of mobility centric environment,
the election of a cluster or the job rotation of the cluster head
on purely energy level, without considering the node mobility
can cause serious problem. A node with sufficiently rich
energy level, taking over the duty of cluster head possessing
high mobility, may move out of the cluster, causing the cluster
to become headless. The situation causes the cluster to go for
a new cluster head. But again the mobility of the nodes is not
considered causing the same process to repeat.
To cope with the situation of cluster head going out of
reach due to mobility, the head rotation process needs to
consider the nodeâ„¢s mobility. The nodes need to maintain
certain additional information to make room for handling
mobility. Following are some of the information the node
should maintain [20]:
¢
Role: to indicate if the sensor is acting as a Cluster head
CH (value=1) or as a participating node (value=0) in the
zone
¢
Mobility Factor: calculated based on the number of
times a node changes from one cluster to another or on
the basis of remoteness.
¢
Members List: if the node is a cluster head, a list which
contains references to the nodes associated with its
Cluster.
¢
TDMA Schedule: Time slot information, when data need
to be collected from the sensor nodes by the cluster head.
249
Authorized licensed use limited to: COCHIN UNIVERSITY OF SCIENCE AND TECHNOLOGY. Downloaded on February 7, 2009 at 06:48 from IEEE Xplore. Restrictions apply.Page 3
The node needs to maintain all these four information, in
which the mobility factor is the one with prime importance for
the election of cluster head. There are different approaches to
calculate mobility factor. One approach is to calculate the
transitions the node makes across the cluster and the other one
is through the concept of remoteness introduced in [21]. In our
proposed scheme we primarily focus on the second method
for the cluster head election.
1) Mobility factor based on transition count
The node associated to a cluster in motion may break its
association to the cluster head and create a new association
with a new cluster head in its new territory. The mobility
factor is calculated based on the number of times the node
moves from one cluster to another.
2) Mobility factor through the Concept of Remoteness
Mobility measure should have a linear relationship with
link change rate. If all the nodes in the cluster are in group
motion like in RPGM, even though the nodes are in motion,
the average page link change is minimal, maintaining high spatial
dependency. The node movement in such scenarios does not
make any breakage of association with the cluster head. So
remoteness can be treated as a measure of mobility factor.
Let
1
,...
3,
2,
1,
0
),
(
-
=
N
i
t
n
i
, where N is the number of nodes,
represents the location vector of node i at time t
and
|)
(
)(
|
)(
t
n
t
n
t
d
i
j
ij
-
=
, the distance from node i to j at time
t. Then the remoteness from node i to node j at time t
is
))
(
(
)(
t
d
F
t
R
ij
ji
=
, where F is the function of remoteness.
For a simple choice of F as identity function, the remoteness is
just the distance between the nodes.
As a node moves relative to the other nodes, remoteness
remains proportionate to its previous values. But as the node
moves in a manner, in which its speed and angular deviation
from the current state are not predictable, remoteness changes
in time. Thus the definition of relative mobility measure in
terms of remoteness of a node as a function of time with
respect to its immediate neighbors is
?
-
=
'
-
=
1
0
|)
(
|
1
1
)(
N
j
ij
i
t
d
N
t
M
(1)
In order to calculate d
ij
(t), from i
th
node to all its j
th
neighboring nodes, the broadcast medium may be used. In
LEACH protocol all nodes in a cluster are time synchronized
with the cluster head. The TDMA schedule issued by the
cluster head are complied by the nodes. Each node uses its
time slot given by the schedule to communicate to the cluster
head. To reduce energy consumption during the other time
slots not intended for a node, the node goes to sleep mode.
Therefore even though a node is in the radio range of its
neighboring nodes, it can not hear the information sent by its
immediate neighbors. In order for nodes to hear
simultaneously, the cluster head gives an extra time slot as
shown in Figure 1.
During the period of extra time slot, called ACTIVE slot,
all nodes need to send their broadcast IDs. As all nodes are
time synchronized with cluster head and use radio propagation,
the node i can make use of the ID broadcast of all the nodes it
hears and calculate d
ij
(t).
Figure.1 TDMA time slots in LEACH-ME protocol
Let beacon sent by a neighboring node was at the start of
ACTIVE time slot t
1
and received at time t
2
. The distance d
ij
(t)
= Radio velocity * |t
2
-t
1
| .
Upon receiving the information from all the nodes, itâ„¢s
possible to calculate the mobility factor for N neighbors
through equation (1). The node with least mobility factor is
considered for the next cluster head, provided the energy level
of that node is not below the threshold. Also the transition
count for the node is checked to be minimal among all of its
neighbors.
The method is explained in steps as given below. We
denote {a} as the normal node, c as the cluster head. The
following steps illustrate cluster head election process.
1. Cluster head c sends ACTIVE message to all its cluster
members to wake up simultaneously.
ACTIVE: c ? {a}: wake up
2. Upon receiving the ACTIVE message, all cluster
members broadcast their IDs with time-stamp. All
cluster member nodes set time-out to receive broadcast
of their entire neighboring node IDs.
The
ID_broadcast helps individual node to know its
neighbors.
ID_broadcast: {a}? NEIGHBORS: know_neighbors
3. Once the broadcast ID timer expires, each node
calculates the remoteness based on the IDs received
and the time at which the IDs are received. The
calculated remoteness information is broadcast by each
node. The process helps to know the remoteness of
neighbors of each other.
remoteness:{a}?NEIGHBORS: know_remoteness
4. Once all the remoteness values of neighbors are
received nodes can go for cluster head election, where
the node with minimal mobility factor is elected as
cluster head, provided its energy level is not below the
threshold.
Initial creation of clusters is based on certain random
selection. The number of cluster heads is based on a suggested
percentage of cluster heads for the network. Normal figure is
5% of the total number of nodes. In view of mobility, the
figure can go high depending on the spatial dependency factor
and the speed with which the node moves. A probable figure
is of the order of 5 -15 % of the total number of nodes.
It should be noted that the cluster head election need not be
done at every TDMA time slot. ACTIVE time slot can be
introduced periodically after a certain number of regular
TDMA periods. The periodicity can be decided based on the
active mobility of the nodes.
Slot
0
Slot
1
Slot
2
¦¦
Slot
N-1
ACTIVE
250
Authorized licensed use limited to: COCHIN UNIVERSITY OF SCIENCE AND TECHNOLOGY. Downloaded on February 7, 2009 at 06:48 from IEEE Xplore. Restrictions apply.Page 4
D. ACTIVE slot deciding phase
Calling ACTIVE slot in regular basis without considering
the nature of the mobility of the nodes can cause extra loss of
energy to the nodes and hence cause threat to the life of nodes.
So the selection of periodicity of ACTIVE slot in TDMA
schedule should be flexible based on the mobility nature of the
nodes. It is desirable to have a measure to decide the
periodicity of the ACTIVE slot. The approach followed is the
transition count as measure to periodicity decision. For a
specific cluster the average transition count of members
decide the slot frequency.
The node which migrates from one cluster to the other
cluster during steady state phase need to have a count of
number of number of such transition it made. The concept is
stated earlier as mobility factor based on transition count.
In order to have the average transition count of the cluster
there should be certain information with the cluster head
regarding the individual transition count of the node members.
But there is no additional time slot available to communicate
the transition count of the nodes to the cluster head. To
resolve this, each node get a data request from the cluster head
need to sent back data along with transition count information
to the cluster head. Cluster head need to process the transition
count information separately.
The decision of including
ACTIVE slot in the next TDMA cycle is taken based on the
average transition count calculated for the last few cycles.
Transition count beyond the threshold decides the ACTIVE
slot induction.
The method explained is put in steps as given below.
1. Cluster head sends data request to the respective nodes
in their TDMA time slot. If the TDMA cycle does not
contain ACTIVE slot, then the data request is sent with
the active flag as zero.
REQ_Data/active = 0: c ? {a}: get data
2. Upon receiving the data request from cluster head, the
cluster member sent its data along with transition count
for the last few cycles to the cluster head.
DATA: {a} ? c: sent data and transition count
3. Once all the cluster member data available, the cluster
head calculate the average transition count for the last
few cycles and decide whether it is above the threshold
decided earlier. If the value is above the threshold, then
all the cluster members are intimated about the
inclusion of ACTIVE slot in the next TDMA cycle by
setting active flag in the REQ_Data
REQ_Data/active=1: c ? {a}: get data and reschedule
4. Upon receiving the data request with active flag set,
the cluster members need to reschedule the TDMA
time slot accordingly to include the ACTIVE frame.
E. Steady State phase in LEACH-ME
In the set-up phase of LEACH, the clusters are organized
and cluster heads are elected. Configuration formed in the set-
up phase is used to transfer monitored data to the base station
during the steady state phase. Because of that, it can not
accommodate the alteration of cluster by mobile sensor nodes
during the steady-state phase. It is possible to resolve this
problem by a simple and traditional method that adds
membership declaration of mobile nodes to typical LEACH
protocol. In LEACH-M scheme, the non-cluster head nodes
instead of sending the data to the cluster head in their allotted
time slot in the TDMA schedule wait for a request (REQ_Data)
from the cluster head to send data.
In the vicinity of mobility it may happen that the
REQ_Data sent to a particular node by the cluster head is not
received by the node, since it is moved to a new location
which is not in the radio range of its current cluster head.
After sending the REQ_Data, if no response is obtained from
the node before the time slot allotted for that node, the node
will be marked as mobile-suspect. If the same thing repeats
for the next time slot allotted for the same node, then the
suspect node is declared as mobile and the time slot for that
node is deleted from the TDMA schedule.
On the other hand, if the node doesnâ„¢t receive any
REQ_Data from the cluster head when it is awake, it marks
itself as suspect of non-member of cluster. During the next
frame slot allotted to this node, if the same thing repeats, then
it takes the decision that it is not a member of the cluster.
Figure.2 Message sequences for cluster join of a mobile node
Once a node becomes a non-member in any of the cluster,
it looks for a cluster to join by sending a broadcast
JOIN_REQ. The cluster head hearing the JOIN_REQ allots a
time slot in its TDMA schedule and broadcasts it to all the
member nodes including the new member. Upon receiving the
new TDMA schedule the mobile node now becomes part of
the cluster and uses the new cluster schedule. The sequences
of messages are shown in Figure 2.
IV.R
ADIO
M
ODEL FOR
LEACH-ME
The first order radio model used in LEACH and LEACH-M
is used for LEACH-ME, where radio dissipates
bit
nanoJoule
E
elec
/
50
=
to drive the transmitter and the
transmit-amplifier dissipates
2
/
/
100
m
bit
picoJoule
elec
=
e
. It is
assumed that radio can be turned on or off as and when
required, to save energy. Also the radio spends the minimum
energy required to reach the destination. The transmission cost
251
Authorized licensed use limited to: COCHIN UNIVERSITY OF SCIENCE AND TECHNOLOGY. Downloaded on February 7, 2009 at 06:48 from IEEE Xplore. Restrictions apply.Page 5
of LEACH-M is different from LEACH-ME because of the
additional effort to calculate the remoteness at the ACTIVE
slot.
Assuming k-bit message is sent on normal and k
active
is sent
on ACTIVE slot, transmission and receiving cost for a
distance of d for k-bit can be calculated as follows
Transmitting cost for LEACH-M:
)
,
(
)
(
)
,
(
d
k
E
k
E
d
k
E
amp
Tx
elec
Tx
Tx
-
-
+
=
(2)
2
*
*
*
d
k
k
E
elec
elec
e
+
=
(3)
For N nodes in the cluster, the total transmission cost per
TDMA cycle is:
2
1
*
*
*
*)
1
(
)
,
(
i
N
i
elec
elec
cluster
Tx
d
k
k
E
N
d
k
E
?
=
-
+
-
=
e
(4)
Transmitting cost of LEACH-ME per TDMA cycle is
transmitting cost of LEACH-M per TDMA cycle added with
active slot cost.
?
=
-
+
-
=
N
i
i
elec
elec
cluster
Tx
d
k
k
E
N
d
k
E
1
2
*
*
*
*)
1
(
)
,
(
e
?
=
+
-
+
N
i
i
active
elec
active
elec
d
k
k
E
N
1
2
*
*
*
2
*
*)
1
(2
e
(5)
In active slots the k-active bits need to be sent twice, one
for ID transmission and other for remoteness transmission.
The extra energy dissipated is in the ACTIVE slots to achieve
awareness of the remoteness to elect cluster head. The number
of bits in active frame is assumed to be less than that of the
data frame bits.
Reception cost will be same in LEACH-M and LEACH-
ME
Reception cost:
)
(
)
(
k
E
k
E
elec
Rx
Rx
-
=
(6)
k
E
k
E
elec
Rx
*
)
( =
(7)
The radio channel is assumed to be symmetric for given
signal to noise ratio.
V. E
XPERIMENTAL
R
ESULTS
To evaluate the performance of LEACH-ME, we simulated
LEACH-M and LEACH-ME using 100 random nodes with
topology for a 100m x 100m network region. The base Station
is located at (50, 50) in the center of the 100m x 100m field.
We simulated the wireless sensor network to get the number
of data packets that are successful in reaching the base station.
The simulation is run by changing the mobility factor for
LEACH-M and LEACH-ME. We also simulated the amount
of energy dissipations for the data packets transmitted.
Figure 3 shows the average successful communication rate
for various mobility factors. At low mobility, the performance
of LEACH-M and LEACH-ME are comparable. But as the
mobility increases, there is a definite improvement in average
successful communication rate in LEACH-ME. As is obvious
from figure 4, at the mobility factor of 4.0 the successful
communication rate is 16%, which is better than the LEACH-
M. On the other hand the amount of energy dissipation as well
as computational overhead increases as mobility increases.
This is obvious from figure 5 and figure 6. At very high
mobility (mobility factor 4 and above), the overhead of
LEACH-ME is 22% more than that of the overhead of
LEACH-M
Figure 3 Average successful Communications of Leach-M and LEACH-ME
for various mobility factors
Figure 4 Performance of LEACH-ME over LEACH-M protocol.
252
Authorized licensed use limited to: COCHIN UNIVERSITY OF SCIENCE AND TECHNOLOGY. Downloaded on February 7, 2009 at 06:48 from IEEE Xplore. Restrictions apply.Page 6
Figure 5 Computational overhead by the LEACH-M and LEACH-ME
protocol against the mobility factor
Figure 6 Energy overhead of LEACH-M and LEACH-ME protocols against
the mobility factor.
VI.C
ONCLUSION
In this paper, we describe how the LEACH protocol can be
enhanced to handle mobility modulation. The paper makes use
of the proposals in LEACH-M protocol where nodes isolated
due to mobility from the cluster are reconnected to a new
cluster through appropriate mechanism. The proposed
LEACH-ME protocol follows the same reconnection
mechanism for the isolated node. It uses the concept of
remoteness for electing the cluster head.
The simulation experiment shows that the proposed
enhanced protocol outperforms LEACH-M in average
successful communication rate by a reasonable margin, at
very high mobility. It is also clear that to achieve the level of
extra performance, energy dissipation needs to be sacrificed at
a tolerable level.
R
EFERENCES
[1] F Akylidiz, W. Su, Y Sankarasubramanyam, and E. Cayirci, Wireless
Sensor Network: a survey, Computer Networks, vol. 38 no. 4, 2002,
pp. 393-422
[2] Carlos de Morrais Codeiro, Dhrarma Prakash Agarwal Ad-hoc and
Sensor Networks, Theory and Applications 1st ed, World Scientific
Publishing Co. Pvt. Ltd., Singapore 2006
[3] Xiaoyan Hong, Mario Gerla, Guangyu Pei and Ching- Chuan Chiang.
A Group Mobility Model for Ad Hoc Wireless Networks, ACM
International Workshop on Modeling and Simulation of Wireless and
Mobile Systems (MSWiM), August 1999.
[4] B. Liu, P. Bras, O. Dousse, P. Nain, and D.F. Towsley, Mobility
improves coverage of sensor networks, in Proc. MobiHoc, 2005,
pp.300-308.
[5] S. Bergbreiter and K. S. J. Pister, Design of an autonomous jumping
microrobot, in Proc. IEEE International Conference on Robotics and
Automation (ICRA™07), Rome, Italy, Apr. 10“14, 2007, pp. 447“453.
[6] K. Dantu, M.H. Rahimi, H. Shah, S. Babel, A. Dhariwal, and G.S.
Sukhatme, Robomote: enabling mobility in sensor networks, in Proc.
IPSN, 2005, pp.404-409.
[7] M. Laibowitz and J.A. Paradiso, Parasitic Mobility for Pervasive
Sensor Networks, in Proc. Pervasive, 2005, pp.255-278.
[8] G. Anastasi, M. Conti, E. Monaldi, and A. Passarella, An Adaptive
Data-transfer Protocol for Sensor Networks with Data Mules, in Proc.
WOWMOM, 2007, pp.1-8.
[9] Ekici E., Yaoyao Gu, D. Bozdag, Mobility-based communication in
wireless sensor networks, IEEE Communications Magazine, vol. 44,
no. 7, 2006, pp. 56-62.
[10] (2006) Adaptive Sampling and Prediction. [Online]. Available:
http://engineering.princeton. edu/gallery/asap/index.htm
[11] Huifang Chen, Hiroshi Mineno, Yoshitsugu Obashi, Tomohiro
Kokogawa, Tadanori Mizuno, Adaptive Data Aggregation for
Clustered Wireless Sensor Networks, Ubiquitous Intelligence and
Computing (UIC), 2007, pp.475-484.
[12] M. Ali, T. Voigt, and Z. A. Uzmi, Mobility Management in Sensor
Networks, MSWSN / 2nd DCOSS, 2006, pp. 131“140
[13] K. Ma, Y. Zhang, and W. Trappe, Managing the Mobility of a Mobile
Sensor Network Using Network Dynamics, IEEE Trans. Parallel
Distrib. Syst., 2008, pp.106-120.
[14]
W.
Heinzelman, A. Chandrakasan, and H. Balakrishnan, Energy-
efficient routing protocols for wireless microsensor networks, in Proc.
33rd Hawaii Int. Conf. System Sciences (HICSS), Maui, HI, 2000.
[15] H. Yang and B. Sikdar, Optimal Cluster Head Selection in the
LEACH Architecture, in Proc. IPCCC, 2007, pp.93-100.
[16] Xiangning, Fan; Yulin, Song, Improvement on LEACH Protocol of
Wireless Sensor Network, in Proc. of SensorComm, 2007, pp.260-264.
[17] Jeong Hee-Jin; Nam Choon-Sung; Jeong Yi-Seok; Shin Dong-Ryeol,
A Mobile Agent Based LEACH in Wireless Sensor Networks, in
Proc. ICACT 2008, pp.75-78.
[18] T. Banerjee, B. Xie, J.H. Jun, and D.P. Agrawal, LIMOC: Enhancing
the Lifetime of a Sensor Network with Mobile Clusterheads, in Proc.
VTC Fall, 2007, pp.133-137.
[19] Do-Seong Kim, Yeong-Jee Chung, Self-Organization Routing
Protocol Supporting Mobile Nodes for Wireless Sensor Network, in
Proc. Computational Sciences IMSCCS, 2006, pp.622-626.
[20] Liliana M. Arboleda C.and Nidal Nasser, Cluster-based Routing
Protocol for Mobile Sensor Networks,in Proc. QShine, 2006.
[21] B. J. Kwak, N. O. Song, and L. E. Miller, A Canonical Measure of
Mobility for Mobile Ad Hoc Networks, in Proc. IEEE MILCOM2003,
