Upload
9841731748
View
221
Download
0
Embed Size (px)
Citation preview
8/7/2019 Mobility of Wsn
http://slidepdf.com/reader/full/mobility-of-wsn 1/6
A Data Delivery Mechanism to Support Mobile
Users in Wireless Sensor Networks
Euisin Lee, Younghwan Choi, Soochang Park, Donghun Lee, and Sang-Ha KimDepartment of Computer Engineering
Chungnam National University220 Gung-dong, Yuseong-gu, Daejeon, 305 764, Republic of Korea
Emails: {eslee, yhchoi, winter, dhlee}@cclab.cnu.ac.kr and [email protected]
Abstract — Wireless sensor networks traditionally consist of
sensors perceiving data and sinks gathering the data. In addition,
users receive required information from the sinks via
infrastructure networks. The users, however, should receive the
information from the sinks through multi hop communications
of disseminated sensor nodes if such users move into the sensor
networks without infrastructure networks. Unlikely, the
previous works only considered mobility of sinks, which function
users. Nevertheless, it is difficult for such approaches about
mobility to exploit the existing data-centric routing algorithms
and also for the mobile sinks to function as gateways to connect
with infrastructure. To improve the shortcomings, we suggest a
novel viewpoint of mobility for wireless sensor networks and
propose a novel architecture and mechanism to support the
mobility with multiple static sinks in this paper. The multiple
static sinks, which are connected with each other via
infrastructure, provide high throughput and low latency.
Furthermore, they improve hotspot problems and prolong
network lifetime. The proposed mechanism finally is evaluated
by simulation results about throughput, latency, and network
lifetime.
Index Terms — wireless sensor networks, mobility, users, and
multiple static sinks
1. Introduction
Wireless sensor networks traditionally consist of sensors
perceiving data and sinks gathering the data. In addition,users receive required information from the sinks viainfrastructure networks. A mobility model in wireless sensor
networks, mentioned above, can classify according to eachobjects. Namely, it is a mobility of sensor node, a mobility of a sink, and a mobility of a user. Hence, deciding a mobility of
what kind of object for sensor networks that suits sensor networks according to various applications is important.
Recently, applications transmitting data to moving users
inside sensor fields such as rescue in disaster area or maneuver in the war zone are on the rise in large-scale sensor
networks [5]. But, until now, such researches support amobility of user by only two forms for these applications.First form is same with figure 2. It supports a mobility of a
user on the assumption that the user communicates directlysinks through infrastructure networks, namely, internet likecommunication systems in traditional sensor networks [1].
But, in applications such as rescues in disaster area or maneuvers in the war zone, circumstances without
infrastructure networks except sensor networks are moreactually. Because, infrastructure networks in these
applications cannot be used because they are damaged as aresult of the war or the disaster. Hence an assumption that auser and a sink can directly communicate through internet has
a problem that is not actually. Therefore communication between the user and the sink inside sensor fields is supported
by only sensor nodes.Second form is same with figure 3. It identifies a user with a
sink. So it supports a mobility of the user by reflectingmovement of the user with the direct movement of the sink
[5~9]. But, researches for this form have also various problems. First of all, they cannot use existing effective datacollection algorithms [2][3][4] between a sink and sensor
nodes based on data in static sink sensor networks. Because,such algorithms can hardly be exploited due to location changeof sink which collects data if sinks in sensor networks have
Figure 1. Typical wireless sensor networks model
Figure 2. A model to support
mobility of user through infra
structure networks, namely,
internet & satellite.
Figure 3. Mobile sink model
ISBN 978-89-5519-131-8 93560 - 245 Feb. 12-14, 2007 ICACT2007
8/7/2019 Mobility of Wsn
http://slidepdf.com/reader/full/mobility-of-wsn 3/6
in the outskirts of the war zone deploy sensor nodes in the war zone and collect locations and movements of enemies fromsensor nodes. Through collected data, the headquarters
elaborate a plan of operations and delivery the operations tosoldiers in the war zone. Then, soldiers carry out the
operations and by extension, will collect directly data fromsensor nodes to obtain the latest information.
3. Description of Mechanism
A. Overview of Mechanism
In our mechanism, if a user intends to obtain information onmoving inside sensor networks, the user disseminates interestto the nearest sink via sensor nodes, and then the user receivesresults of interest from the sink. Also, if the nearest sink of theuser changes the user requests the results to new the nearest
sink and receives the results from the new sink.
This paper makes the following assumptions:• A user can communicate multiple static sinks through
only sensor nodes, because networks within sensor fieldsare without infrastructure networks.
• Multiple static sinks are deployed in an arbitrary
position in the outskirts of sensor fields connected with
infrastructure networks as internet.
• Multiple static sinks can directly communicate other
sinks via infrastructure networks.
• The data which one sink collects is aggregated by the
sink. All sinks share the aggregated data via infrastructurenetworks.
To implement the proposed mechanism, we need to addressthe following phases: dissemination of sink announcementmessage, interest dissemination of user, data collection of sink,
information sharing of multiple static sinks, mobility supportof user, and information propagation of sink. We detail each phase to next section.
B. Dissemination of sink announcement messageTo the initial stage of sensor network, if a sink is located in
an arbitrary position in the outskirts of sensor fields which isconnected with infrastructure networks as internet it has
flooded a sink announcement message to announce itself inside the whole sensor fields like figure 5. As a result of
flooding a sink announcement message, every sensor nodeshave known hop counts and next hop neighbor sensor node to
each sink. And every sensor nodes have known the nearestsink from location of themselves through hop counts to eachsink.
C. Interest dissemination of the user While moving inside the sensor fields, if a user wants to
collect a data from sensor fields, the user selects the nearestsensor node from location of itself as first agent. And the user delivers an interest to the first agent. The first agent which is
delivered the interest from user forwards the interest to a nexthop neighbor node toward the nearest sink. The next hopsensor node which is delivered the interest also forwards a
next hop neighbor node toward the nearest sink. This processis continued until the sink. So, the sink receives the interest of the user. Also, a back route from the sink to the user for the
interest has established through this process.
D. Data collection of the sink Sensor networks with a static sink are a network that
sensing data from sensor nodes should be transmitted to the
static sink through multi-hop communication. Routingalgorithms to collect data in sensor networks with a static sink are used in scenario of various types, for example, a scenario
generating data by periods, a scenario generating a minorityevent, and a scenario detecting a moving object, etc.
Hence, a user will can select and use the most appropriaterouting algorithm with static sink according to a scenario.Such research was advanced already plentifully [2][3][4]. Sowe will not mention anymore in this paper. Therefore, we use
one in the existing routing algorithm as the routing algorithmto collect data in this paper.
E. Information sharing of multiple static sinksAs shown in Fig. 1, a sink in typical sensor networks takes
charge of the function as gateways for connection withinfrastructure networks [1]. And various papers in relation tomultiple static sinks also indicate the connection between a
sink and an infrastructure network and the connection betweenall sinks as an assumption [10 - 13]. Therefore, in this paper, itis a sufficient propriety that all sinks placed in the edge of a
sensor field can communicate with the other sinks via
Figure 5. Dissemination of sink announcement
message
Figure 6. Mobility support of the user Figure 7. Information propagation of the sink
ISBN 978-89-5519-131-8 93560 - 247 Feb. 12-14, 2007 ICACT2007
8/7/2019 Mobility of Wsn
http://slidepdf.com/reader/full/mobility-of-wsn 4/6
infrastructure networks. Hence, in proposed mechanism, asink which is delivered an interest from a user collects datafrom sensor fields and aggregates the collected data. Next, the
sink will share aggregated information with the other sinksthrough infrastructure networks.
F. Mobility support of the user
The user may move to other place after sending interest tosink by agent. In this case, the user selects another agent andcreates a new connection from the original agent to newselected agent. The user can receive the aggregatedinformation from sink through this connection. Thus, mobility
of the user is guaranteed.The user can obtain the information of neighbor nodes in
radio range of the first agent by reply message received from
the first agent. If the user moves out of RF range of the firstagent, then the user will retransmit the agent selection messagewith same RF range of the agent. All sensor nodes which have
received this message send a reply message which contains theinformation of neighbor node in their radio range to the user.
As shown in Figure 6, the user checks whether there is sensor nodes which are a neighbor node both in RF range of firstagent and in RF range of the sensor node sent the replymessages. Among the sensor nodes which have the connection
node in their radio range, the user selects the nearest sensor node from user as second agent. The user settles this selected
sensor node as second agent and informs the connection nodes between the first agent and the second agent to the secondagent. Second agent creates the connection to the first agent
through the connection node informed by user. By this way,the path from the first agent to the user can be createdaccording to movement of user.
G. Information propagation of the sink A sink delivers the aggregated information of the collected
data to first agent through the connected path. The first agentdelivers the aggregated information to last agent through theconnection of agents. Last agent deliveries the aggregated
information to the user.As shown in figure 7, if the nearest sink from the user
changes, the user requests the information to the new nearestsink and receives the information from the new nearest sink.
4. Performance Evaluation
In this section, we evaluate the performance of a proposedmechanism through simulations. We first describe our simulation model and simulation metrics. We next evaluate
how environment factors and control parameters affect the performance of a proposed mechanism.
A. Simulation Model and Metrics
We implement the proposed mechanism in the Qualnetver.3.8 [14]. A sensor node’s transmitting and receiving power consumption rate are 0.66W and 0.39W respectively. Thetransceiver in the simulation has a 200m radio range. Eachinterest packet is 36 bytes long and the data packet has 64
bytes. The sensor network consists of 50 sensor nodes, whichare uniformly deployed in a 1000m x 1000m field.
The multiple static sinks are located in the outskirts of
sensor fields. The number of user is one. And the default speedof user is set to 10 m/sec. the user disseminates an interest at aninterval of 10 second. Every sensor nodes receive the interest
and generate only one sensing data for the interest. This isdefined as one interest round. Namely, one interest round is 10
second. The simulation lasts for 500 seconds.We use for metrics to evaluate the performance of the
proposed mechanism. The network lifetime is defined as thenumber of the interest round first sensor node die. The
residual energy is defined as the residual energy of sensor nodes at time, namely, the interest round that first sensor node
die. The data delivery ratio is the ratio of the number of successfully received reports at a user to the total number of reports generated by every sensor node. The delay is defined
as the average time between the time a sensor node transmits areport and the time a user receives the report.
We compare a static sink model without a mobile user to a
multiple static sinks model with a mobile user in the
simulation. We express a direct communication model between the user and the sink as ‘sink 1 and not user’ and a
user movement model as ‘sink 1 and user 1’ according to thenumber of sink in figures of performance evaluation. Here ‘1’is case that sink is one. If the number of sink changes, this
numerical value reflects in figures of performance evaluation.
B. Impact of the number of multiple static sinksWe first study the impact of the number of sinks on the
proposed mechanism’s performance. The number of sinks
varies from 1, 2, 3, to 4. And the number of sensor nodesvaries from 50, 100, 150, to 200. Sinks have a maximum speed
Figure 8. Network lifetime for the number of sink and sensor node Figure 9. Residual energy for the number of sink
ISBN 978-89-5519-131-8 93560 - 248 Feb. 12-14, 2007 ICACT2007
8/7/2019 Mobility of Wsn
http://slidepdf.com/reader/full/mobility-of-wsn 5/6
of 10mm/s. In this part, we compare one static sink without the
user to multiple static sinks with the user.Figure 8 shows the number of interest round, namely,
network lifetime. One static sink model without a mobile user
is of small number due to hotspot problem of sensor nodes
near the sink. But, the number of interest round of multiplestatic sinks model with mobile user is higher than the number of interest round of one static sink sensor network even thoughthe sinks are more than 3. Network lifetime, namely, interest
round, prolonged because energy consumption of sensor nodes became evenly. Figure 9 shows the residual energy attime that first sensor node dies in simulation circumstances of
50 sensor nodes As shown in Figure 9, the energyconsumption of sensor nodes become more evenly because itsolves hotspot problem due to addition of sink. Figure 10
shows the data delivery ratio. A model with mobile user islower than a model without user because it must deliveryinformation from sink to user. But, the hop count between sink
and user decreases due to addition of sink because data failsreduce. Therefore the data delivery ratio of a model withmobile user approaches a model without user. Figure 11 shows
the delay. The delay of a model with mobile user is longer thanthe delay of a model without user because it must delivery
information from sink to user. Therefore the data delivery ratioof a model with mobile user also approaches a model withoutuser.
C. Impact of the number of sensor nodesWe next evaluate the impact of the number of sensor nodes
on the proposed mechanism. The number of sensor nodes
varies from 50, 100, 150, to 200. As shown in Figure 8, 10, and11, the proposed mechanism never falls a performancenevertheless the number of sensor node increase.
D. Impact of the user’ mobility
We last evaluate the impact of the user’s moving speed onthe proposed mechanism. In the default simulation setting, wevary maximum speed of a user from 6, 8, 10, 12, to 20m/s. In
this part, we compare a network model of one static sink without the user to a network model of four static sinks withthe user. Figure 12 shows data delivery ratio when the user’
moving speed changes. Because the static sink model withoutthe user is not user, it indicates the same result independent of the user’s moving speed. While the multiple static sinks modelwith the user decrease according to increment of the user’smoving speed. But the data delivery ratio remains around 0.9 – 1.0 nevertheless the user move faster. Figure 13 shows the
delay about data delivery, which increases only slightly as the
user moves faster, because it increases the number of agentfrom the sink to the user. Figure 13 shows that the network
lifetime decreases as the user’ moving speed increases. Thefaster a user moves, the more a user needs the number of agentfor connection between the user and the sink.
5. Conclusion
In this paper, we propose a novel sensor network model and
a novel mechanism to support mobility of users in wirelesssensor networks based on multiple static sinks In proposednetwork model, because multiple static sinks can
Figure 11. Delay for the number of sink and sensor node Figure 12. Data delivery ratio for user speed
Figure 12. Data delivery ratio for user speed Figure 13. Delay for user speed
ISBN 978-89-5519-131-8 93560 - 249 Feb. 12-14, 2007 ICACT2007
8/7/2019 Mobility of Wsn
http://slidepdf.com/reader/full/mobility-of-wsn 6/6
communicate with the other sinks as short hop via
infrastructure networks, the user receives the information withhigher data delivery ratio and faster time. And the lifetime of the sensor networks increase because the balance energy
consumption of sensor nodes is possible.We verified that the lifetime of sensor networks is
prolonged because a use of multiple static sinks decreases a
consumption of sensor nodes. Also, we verified that a
performance about the data delivery ratio and the delay never falls nevertheless a communication between the user and the
sink for guaranteeing movement of the user is supported byonly sensor nodes without infrastructure networks, namely,internet.
REFERENCES[1] I.F. Akyildiz, S. Weilian, et al., "A survey on sensor networks,"
Communications Magazine, IEEE Journal Vol. 40, pp. 102-114, Aug.
2002.
[2] C. Intanagonwiwat, R. Govindan, and D. Estrin, "Directed diffusion: A
scalable and robust communication paradigm for sensor networks,"
ACM/IEEE Mobicom Conference, 2000.
[3] W.R. Heinzelman, J. Kulik, and H. Balakrishnan, "Adaptive Protocols for
Information Dissemination in Wireless Sensor Networks," ACM/IEEEMobicom Conference 99, Aug. 1999.
[4] W. Heinzelman, A. Chandrakasan and H. Balakrishnan,
“Energy-Efficient Communication Protocol for Wireless Microsensor
Networks,” Proc. 33rd Hawaii Int’l. Conf. Sys. Sci., Jan. 2000.
[5] F. Ye, Haiyun Luo, et al., “A Two-Tier Data Dissemination Model for
Large-scale Wireless Sensor Networks,” ACM/IEEE MobiCOM 2002,
Sept. 2002.
[6] K. Hwang, J. In, et al., "Dynamic sink oriented tree algorithm for efficient
target tracking of multiple mobile sink users in wide sensor field," IEEE
VTC2004-Fall 2004, Sep. 2004.
[7] S. Kim, S. Son, et al., “SAFE: A Data Dissemination Protocol for
Periodic Updates in Sensor Networks,” Distributed Computing Systems
Workshops 2003, 23rd International Conference.
[8] H. L. Xuan and S. Lee, “A Coordination-based Data Dissemination
Protocol for Wireless Sensor Networks,” IEEE ISSNIP 2004, Dec. 2004.
[9] S. R. Gandham, M. Dawande, et al., "Energy Efficient Schemes for Wireless Sensor Networks with Multiple Mobile Base Stations," IEEE
GLOBECOM 2003, Dec. 2003.
[10] Henri Dubois-Ferriere, Deborah Estrin, and Thanos Stathopoulos,
“Efficient and Practical Query Scoping in Sensor Networks,” IEEE
International Conference on Mobile Ad-hoc and Sensor Systems 2004,
Oct. 2004.
[11] Abhimanyu Das and Debojyoti Dutta, “Data Acquisition in Multiple-sink
Sensor Networks,” ACM SIGMOBILE Mobile Computing and
Communications Review 2005
[12] E. Ilker Oyman and Cem Erso, “Multiple Sink Network Design Problem
in Large Scale Wireless Sensor Networks,” Communications, 2004 IEEE
International Conference on, Jun. 2004.
[13] Seung Jun Baek, Gustavo de veciana, and Xun su, “Minimizing Energy
Consumption in Large-Scale Sensor Networks Through Distributed Data
Compression and Hierarchical Agrregation,” Selected Areas in
Communications, IEEE Journal 2004, Aug. 2004.[14] Scalable Network Technologies, Qualnet, [online] available:
http://www.scalable-networks.com.
[15] Hui Dai and Rechard Han, “A node-centric load balancing algorithm for
wireless sensor networks,” Global Telecommunications Conference,
2003. GLOBECOM '03. IEEE, Dec. 2003.
ISBN 978-89-5519-131-8 93560 - 250 Feb. 12-14, 2007 ICACT2007