Int. J. Communications, Network and System Sciences, 2012, 5, 708-713
http://dx.doi.org/10.4236/ijcns.2012.510073 Published Online October 2012 (http://www.SciRP.org/journal/ijcns)
A Cellular-Assisted Efficient Handover Algorithm for
Wireless Sensor Networks
Jun Zhu1, Hui Gao1,2, Yuling Ouyang2
1Key Laboratory of Intelligent Computing & Singnal Processing, Ministry of Education Anhui University, Hefei, China
2Shanghai Research Center for Wireless Communications, Shanghai, China
Received August 10, 2012; revised September 15, 2012; accepted September 25, 2012
ABSTRACT
Heterogeneou s network convergence and handov er have become very hot in recent years. This paper proposed an eff i-
cient handover scheme in Multi-PAN Wireless Sensor Networks (WSNs). A number of edge nodes are set at the edge
of each Personal Area Networks (PANs). A user equipment (UE), which has WSN and cellular network interface, acts
as sensor node or mobile cluster head in WSN area. Thus, edge early warning can be acquired from edge nodes and
neighbor channel information can be acquired with BS-assistance. Simulation results show that low transmission inter-
rupted delay and low energy consumption can be achieved compared with conventional scheme in WSN.
Keywords: Efficient Handover; Edge Early Warning; Special Beacon; Wireless Sensor Networks
1. Introduction
With the development of WSNs [1-3] and cellular net-
works, modern technologies have been prepared for cre-
ating the convergent communication infrastructures,
through which WSNs and cellular networks convergence
have become inevitable. These heterogeneous networks
appear in many application areas [4], especially in the
health care. In these applications it is usually needed for
mobile sensor nodes to frequent handover and to transmit
data between Multi-PANs.
If sensor nodes move between Multi-PANs, the nodes
are required to scan all channels to find the channel of
nearby PAN which sensors move to, and then hand over
to the target transmission channel. Because of scanning
all channels, it will not only bring energy cost, but also
will interrupt the data transmission for a long time and
influence the transmission delay and QoS. So how to use
the UE or cellular network assisted to realize and opti-
mize efficient handover between Multi-PANs in Zig-Bee
wireless sensor network is a key problem to be solved.
And because sensor nodes have limited energy and are
not convenient to replace battery, how to reduce energy
consumption is also an important problem for WSN (e.g.,
SMAC [5], TMAC [6]).
There are some handover protocols in WSN. Standard
ZigBee handover is in specification IEEE 802.15.4 [7]
and ZigBee specification [8,9]. If an associated node
wants to leave the PAN, it shall send a disassociation
notification command to its coordinator. Sensor node
scans all channels from lowest channel number to the
highest in order with peak energy. For the duration of the
scan, the node shall suspend beacon transmission and
data transmission. Until selecting a channel of nearby
PAN which it wants to move, the mobile sensor sends an
association request command to the coordinator of the
PAN. Richard et al. in [10] proposed a ha ndover mecha-
nism for WSNs. Sensor nodes can hand over in certain
areas to achieve connecting from a mobile cluster head to
a fixed and main-powered cluster head which connected
to a permanent power supply for saving energy. The mo-
bile sensor, mobile cluster head and fixed cluster head
are on the same channel. It however doesn’t consider
sensor nodes moving in different PANs and channels
scenario.
This paper proposed an efficient handover scheme in
Multi-PAN WSN. In the scheme, each PAN has one
channel for communication in WSN area. UE and mobile
nodes form a mobile cluster, they move in one PAN and
attach to the fixed node to communicate. The mobile
nodes have to transmit data continuously wherever they
move. When the mobile nodes move from one PAN to
another, they will trigger handover. Through coordina-
tion of the BS, edge nodes can be set at the edge of PAN
to achieve edge early warning and UE can acquire the
neighbor PAN channel information with cellular link to
achieve efficient handover and reduce transmission in-
terrupted delay.
The remainder of this paper is organized as follows.
Section 2 introduces the scenario and the efficient hand-
C
opyright © 2012 SciRes. IJCNS

J. ZHU ET AL. 709
over scheme we proposed in detail which includes edge
node decision algorithm and procedure of the proposed
scheme. Section 3 is the results and analysis of the simu-
lation. Section 4 concludes this paper.
2. Efficient Handover Algorithm
2.1. System Structure
The efficient handover scheme is based on the scenario
which is shown in Figure 1. We consider the convergent
scenario for wireless sensor network (WSN) and cellular
network. In the cellular system, the user equipments
(UEs) ar e und er th e co ntr o l of a cellular base station (BS).
In the coverage area of a cellular network, there also ex-
ists a group of wireless sensor nodes composing the
WSN networks. In the convergent scenario, the cellular
UEs are dual-mode and have both WSN and cellular in-
terfaces, and they can act as mobile cluster heads for the
WSN. The UEs can provide backhaul access for the
WSN nodes under the help of 3GPP-LTE BS. The WSN
is based on Multi-PANs topology for beacon-enabled
mode. Each PAN has one channel for communication.
Some sensor nodes and UEs are moving in the WSN area
but connect with the WSN all the time. WSN control unit
(e.g., BS) keeps the topology of WSN and knows cover-
age area of all PANs and overlapping area of PANs.
2.2. Edge Node Decision Algorithm
Edge node is located at the edge of its PAN or in the
overlapping area of two PANs, and it is used to the edge
early warning for mobile sensors. The edge node is se-
lected by BS after organized the WSN or after topology
and usually has strong commu nication ability.
Edge node sets two beacon formats. If there is no mo-
bile node, edge node will send the normal beacon. If
there are some mobile nodes, edge node will be triggered
and send the special beacon containing edge flag. In bea-
con payload, the bit 16 and 17 are reserved. So we add
edge flag in beacon payload which occupied 1 bit, which
is shown in Table 1 [7] and Table 2.
ED ED
ED
Service Center
BS/eNodeB
ED WSN area
ED
ED ED
ED
ED
PAN 1
PAN 2
moving
Figure 1. System application scenario.
Table 1. Beacon frame format of edge node.
MAC Payload
Octets: 2 Variable Variable Variable
MHR
Superframe Specification GTS fields Pending Address fields Beacon Payload
MFR
Table 2. Special beacon payload format of edge node.
Bits: 0 - 7 8 - 11 12 - 15 16 17 18
Protocol ID Stack prof ile Nwk Protocol version Edge flag Reserved Router capacity
Bits: 19 - 22 23 24 - 87 88 - 111 112 - 119
Device depth End device capacity Nwk extended
PANID Tx offset (optional) nwkUpdateID

J. ZHU ET AL. 710
During the network initial phase, each Gateway (GW)
obtains approximately location of each sensor node,
forms the topology of PAN and reports the information
to BS. BS saves the approximately location, range and
neighbor PAN information of all PANs and uses the re-
lated information of PANs to determine the edge node.
The st eps are as follows.
BS searches the nodes of neighbor PAN at the double
transmission range of node in one PAN (e.g. PAN 1) and
saves the nodes of neighbor PAN in edge neighbor node
list. Then BS determines node in edge neighbor node list
as candidate edge node of PAN 1 and selects the edge
node from candidate edge node in PAN 1. If the distance
between two candidate edge nodes in the same PAN is
below the threshold, BS will select one node as edge
node; If BS will select both nodes as edge nodes. After
determining all edge nodes in Multi-PAN area, BS will
send edge-set signaling to inform them.
2.3. Edge Node Information Table
When all edge nodes are selected, BS saves a PAN chan-
nel information table including PAN ID, edge node ID,
location, channel and neighbor channel. According to the
information in the PAN channel table, BS knows which
neighbor channel the edge node has. The edge node in-
formation table is shown in Table 3. The detailed pa-
rameters in this table include but not limited to below
elements:
 PAN ID is the 16-bit address of a PAN in WSN net-
work.
 Edge node ID is the 16-bit network address of a BS-
selected edge node in WSN.
 Edge node location is the approximate location in-
formation of the edge node in a PAN.
 Edge node neighbor channel is the neighbor PAN
communication channel of the edge node.
2.4. Procedure of Efficient Handover Algorithm
As mentioned above, we used an efficient handover
method in Multi-PAN WSN. The complete procedure is
as follows:
Step 1: When organizing the WSN, BS determines
which sensor to be edge node and informs them with
edge-set signal i ng vi a WS N li nks.
Step 2: When UE and its mobile cluster move to the
edge of PAN, its topology will be changed and its pa rent
node will change into edge node. At the same time, edge
node discovers that there is mob ile node connecting to it.
So the edge node is triggered and sends the special bea-
con periodically.
Step 3: UE receives the special beacon containing
edge flag, which is edge early warning for mobile cluster.
When the special beacon is received by mobile node of
mobile cluster, it will report the information to UE.
Step 4: UE sends request containing PAN ID and edge
node ID about channel information of neighbor PAN to
BS with cellular interface. BS queries the neighbor chan-
nel table and replies neighbor PAN channel information
and to UE with cellular interface.
Step 5: UE informs each node of mobile cluster about
neighbor PAN neighbor chann el information. The mobile
cluster scans the obtained neighbor channel and hand-
overs.
All the procedure of efficient handover scheme is de-
scribed in Figure 2.
Table 3. Edge node information table.
PAN ID Edge node ID Edge node
location Edge node neighbor channel
(1)
Mobile Node NMobile UEEdge no deGateway BS
Send s peci a l
bea con
Determine edge node
Neighbor channel information request
Neighbor channel informati on reply
Overhear beacon
fr om e dge node
(2)
(5)
(4)
Discover m o bile
UE connected
(3 )
Inform m obile ndoes
Figure 2. Signaling for procedure of efficient handover scheme.

J. ZHU ET AL. 711
3. Analysis and Simulation
3.1. Analysis of Efficient Handover Algorithm
WSNs always have the question of energy consumption.
Devices will be battery powered, and battery replacement
or recharging in relatively short intervals is impractical.
Therefore, the power consumption is of significant con-
cern. Energy of cellular networks has large advantage
comparing with WSN. So it’s a great improvement for
WSN to transfer energy consumption from the sensor
networks to the cellular networks.
The energy consumption of conventional method in
WSN is included in Equation (1):
_handover wsnrequestscan
PMPN request PMP

 
_quest scan
PIP
_handover wsnrequestscanrequest
TMTNTMT

(1)
 Prequest means the energy consumption that sensor
node which has disassociation with its coordinator,
would send request signaling to its original channel
for association again. M is the hop number of request
signaling which is sent to original coordinator.
 Pscan means the energy consumption that sensor node
which has disassociation with its coordinator, scans
each channel from lowest number to highest in order
to find a appropriate one to association. The device is
expected to scan in 16 channels. N is the number of
scanning channels.
 request
P means the energy consumption that sensor
node which has found a appropriate channel, would
send a association request to the coordinator. M’ is
the hop number of request signaling which is sent to
new coordinator.
Thus the energy consumption of proposed method is
included in Equation (2):
__handovernewedge informUEre
PP (2)
 Pedge_inform means the energy consumption that eNB
sends edge-set signaling to inform the edge node.
 PUE_request means the energy consumption that UE
sends neighbor channel request with edge node ID to
eNB for itself or mobile cluster.
 Pscan means the energy consumption that sensor node
scans the target channel for only one superframe pe-
riod. I is the number of mobile nodes in mobile clus-
ter.
According to the calculation, N*Pwsn > Pnew (N > 4).
The proposed method is suitable for frequent handover or
mobile cluster handover. The proposed method would be
used and guarantee the power consuming in the cellular
link should be smaller than normal sensor network rou-
tine.
When sensor node moving in the WSN area, transmis-
sion with less interrupted delay is also very important.
The time spent on handover in conventional method is
shown in Equation (3):

 (3)
 Trequest means the time spent on sending a association
request to the original coordinator. M is th e hop num-
ber which signaling sen t to the original coordinator.
 Tscan means the time spent on scanning each channel,
which is [a Base Super frame Duration

(2n + 1)]
symbols, where the n is the value of the Scan Dura-
tion parameter, which at the ranges from 0 to 14. It
means scanning each channel will spent 8 ms to 200
ms, which is a long time for transmission interruption.
And the number of scan ni n g chan nel N is random.
 request
T

means the time spent on sending a association
request to the new coordinator. M’ is the hop number
which signaling sent to the new coordinator.
The time spent on handover in proposed method is
shown in Equation (4):
__ _handover newUE requesteNB replyscanrequest
TTTTT


 (4)
 TUE_request means the time UE sends a neighbor chan-
nel request with edge node ID to eNB for itself or
mobile cluster.
 TeNB_reply means the time eNB sends neighbor channel
response signaling back to UE.
 Tscan means the time spent on scanning one c han n el .
 request T

means the time spent on sending a associa-
tion request to the coordinator.
A signaling interactive time is about 3 ms - 5 ms [11]
on the physical layer of air interface in TD-LTE. Ac-
cording to the above process, connection time is about 9
ms - 15 ms which calculating the number of signaling
interaction. The transmission date time is 4 ms - 6 ms
(two child frame length), so the consuming time is 13 ms
- 21 ms in cellular, which 18 ms being typical.
According to the calculation, Thandover_wsn > Thandover_new,
the proposed method has reduced a large of interrupted
transmission delay. UE would be used and guarantee the
time consuming in the cellular link should be smaller
than normal sensor network routine.
3.2. Simulation Scenario Description
To evaluate the effectiveness of efficient handover sche-
me, we measure the power consumption and average
scanning delay when mobile nodes hand over. We com-
pare the results of our proposed scheme and traditional
scheme. Here we take standard ZigBee handover as tra-
ditional scheme. For the convenience of analysis, some
assumptions are as follow:
 There are 2 numbers of PANs. In each PAN, there are
5 mobile nodes and a UE forming a mobile cluster
and N numbers of static sensor nodes.
 The number of channel is 16 (ZigBee).
 The scanning delay of each PAN channel is 30 ms.

J. ZHU ET AL. 712
The number of scanning channels in conventional
scheme is random.
3.3. Analysis of Simulation Results
We first measure the average scanning delay of our pro-
posed scheme and traditional scheme.
The simulation results of latency are shown in Figure
3, where the plus marker represents for latency using tra-
ditional scheme, while circle marker represents the la-
tency after using efficient handover scheme with BS as-
sisted. It can be seen from the simulation results that the
proposed scheme improves network performance for
each handover.
We also measure the power consumption of our pro-
posed scheme and traditional scheme.
The simulation results of power consumption are
shown in Figure 4, where the plus marker represents for
050 100150 200250 300 350400 450 500
0
50
100
150
200
250
300
350
400
450
500 Lat ency of handover
t
Lat ency
Convent i onal scheme
E ffic i e nt handover scheme
Figure 3. Performance comparisons with respect to average
delay time.
0100 200300 400 500600
0
10
20
30
40
50
60
70 P ower c onsumpti on of handover
t
P ower c onsumpti on
Convent i onal schem e
E fficient ha ndover scheme
Figure 4. Performance comparisons with respect to power
consumption.
power consumption using traditional scheme, while cir-
cle marker represents power consumption using efficient
handover scheme. It can be seen from the simulation
results, the circle marker has a slow grow after the in-
flection point t = 100, that because mobile nodes in the
mobile cluster has received neighbor PAN channel in-
formation from UE and need not to scan all the channels
for handover. So the addition of handover mechanism
with BS assisted can help to reduce the handover power
consumption of mobile cluster or frequent handover
power consumption of mobile nodes. Because mobile
nodes in mobile cluster will obtain the target PAN chan-
nel information instead of scanning channels to find an
appropriate PAN.
4. Conclusion
In this paper, we propose an efficient handover mecha-
nism in Multi-PAN area. We design the edge node used
for edge early warning; then we use BS to manage the
Multi-PAN channel information and use UE to acquire
neighbor PAN channel information to reduce the scan-
ning channel time and power consumption. The simula-
tion results indicate that the presented scheme can reduce
transmission interrupted delay and power consumption
when mobile sensor nodes hand over in Multi-PAN un-
der the help of cellular network.
5. Acknowledgements
This work was supported by the National Natural Sci-
ence Foundation of China (No. 61071168) and the 211
Project of Anhui University, the Anhui Provincial Natu-
ral Science Foundation (No. 11040606Q06).
REFERENCES
[1] D. Culler, D. Estrin and M. Srivastava, “Overview of
Sensor Networks,” IEEE Computer, Vol. 37, No. 8, 2004,
pp. 41-49. doi:10.1109/MC.2004.93
[2] J. Yick, B. Mukherjee and D. Ghosal, “Wireless Sensor
Network Survey,” Computer Networks, Vol. 52, No. 12,
2008, pp. 2292-2330. doi:10.1016/j.comnet.2008.04.002
[3] V. Rajaravivarma, Y. Yang and T. Yang, “An Overview
of Wireless Sensor Network and Applications,” Proceed-
ings of 35th Southeastern Symposium on System Theory,
Morgantown, 22 April 2003, pp. 432-436.
[4] A. Wheeler, “Commercial Applications of Wireless Sen-
sor Networks Using Zigbee,” IEEE Communications Ma-
gazine, Vol. 45, No. 4, 2007, pp. 70-77.
doi:10.1109/MCOM.2007.343615
[5] W. Ye, J. Heidemann and D. Estrin, “Medium Access
Control with Coordinated Adaptive Sleeping for Wireless
Sensor Networks,” IEEE/ACM Transactions on Networ-
king, Vol. 12, No. 3, 2004, pp. 493-506.
[6] T. Van Dam and K. Langendoen, “An Adaptive Energy-

J. ZHU ET AL. 713
Efficient MAC Protocol for Wireless Sensor Networks,”
Proceedings of the 1st International Conference on Em-
bedded Networked Sensor Systems, 2003, pp. 171-180.
doi:10.1145/958491.958512
[7] IEEE-TG15.4, “Part 15.4: Wireless Medium Access Con-
trol (MAC) and Physical Layer (PHY) Specifications for
Low-Rate Wireless Personal Area Networks (LR-WPANs),”
IEEE Standard for Information Technology, New York, 8
September 2006.
[8] ZigBee Alliance, “ZigBee Specification,” 2007 ZigBee
Standards Organization, San Ramon, 2008.
[9] E. Karapistoli, F.-N. Pavlidou, I. Gragopoulos and I.
Tsetsinas, “An Overview of the IEEE 802.15.4a Stan-
dard,” IEEE Communications Magazine, Vol. 48, No. 1,
2010, pp. 47-53. doi:10.1109/MCOM.2010.5394030
[10] R. M. Miller-Smith, “Handover Mechanism for Sensor
Networks,” US2010/0285807 A1, 2010.
[11] R1-060932. CATT, RITT, TD-Tech, “EUTRA TDD Ran-
dom Access Procedure,” Athens, 2006, pp. 27-31.