46 51

ISSN: 2277 – 9043
International Journal of Advanced Research in Computer Science and Electronics Engineering
Volume 1, Issue 6, August 2012

A Cooperative Multi-Hop Clustering Protocol
for Energy Saving of Mobile Devices
P. Narasimhaiah M.Sc (M. Tech), Mrs. S. Vasundra M. Tech. (Ph. D)

Abstract— Wireless local area network (WLAN), or IEEE Because of the limited power Bluetooth supports low
802.11, has become a most popular because of its sufficient bandwidth of 2mbps for range of 10 meters. This work
bandwidth and well-constructed infrastructures. But, a serious presents the idea of using the co-existence of WLAN and
problem of WLAN is its considerable energy consumption. Bluetooth in a single mobile platform to solve the energy
Mobile devices are driven by limited battery power, so it is consumption problem of WLAN interface.
essential to reduce the power consumption due to WLAN
interface without degrading its performance. This can be Previous several works used Bluetooth as
achieved by using the co-existence of WLAN and Bluetooth in a secondary radio to reduce the energy consumption [3], [4],
single mobile platform to solve the power consumption [5]. Mainly Bluetooth is used as a connected channel
problem in WLAN based communication systems. This work is between mobile devices and the WLAN access points (AP).
based on Multi-Hop Clustering proposed to increase the life In On-Demand paging [3], and wake on wireless [7],
time of network. This protocol selects the cluster head
Bluetooth is used as a wake up channel. Bluetooth is also
considering three factors, maximum energy, the number of
used as a data communication channel when application
neighbouring nodes, and distance to the access point. In this
paper, a cluster is a Bluetooth Personal Area Network (PAN), need low data rates to prolong the power off time of the
which consists of one cluster head and several regular nodes WLAN. This requires the hardware and software
.The cluster head acts as gateway between the PAN and the modification to Wireless infrastructures. Clustering make
WLAN and enabling the regular nodes to access the WLAN. the nodes share their WLAN interfaces with each other, and
This paper presents a Cooperative Multi-Hop Clustering Clustering performed periodically in a distributed manner
Protocol, which dynamically reforms clusters according to the based on the energy usage and bandwidth need of nodes.
each nodes bandwidth requirement, energy use, and Clustering is performed independent of WLAN APs,
application type. This protocol does not require the
therefore there is no need to made changes to the existing
modifications of existing wireless infrastructures. This protocol
infrastructures.
simulated for large network more than 200 nodes. But results
demonstrate that, this approach is effective in reducing the
Compare to the previous works, this approach is
power consumption of the WLAN.
based on Multi-Hop Clustering. Clustering is commonly
used in sensor networks for load balancing [10],[12], or
Keywords—Bluetooth, bandwidth, communication system, energy efficiency [8], [9], [12]. In this work clustering make
energy consumption, Multi-Hop. nodes that share their WLAN with each other. A cluster is
Bluetooth Personal Area Network (PAN) [13] as shown in
I. INTRODUCTON Fig. 1(a), that can consist of one cluster head (CH) and
several regular nodes (RNs). CH coordinates the nodes
Wireless local area network (WLAN) [11], or IEEE 802.11 within the cluster, and forward the packets from PAN to the
is well-known because of its well developed infrastructures WLAN APs, and vice-versa. CH keep their WLAN interface
and sufficient bandwidth. But a serious problem with the on to provide links to the WLAN AP, and allow RNs to use
WLAN interface of mobile devices is, it consume more only Bluetooth and turn WLAN interface off in order to
power relative to the other devices like processor, memory, save energy. Clustering is performed periodically in a
display and disk. The relative power consumption of the distributed manner based on the energy usage and
WLAN interface depends on mobile devices and it varies bandwidth need of the node.
from 5-10% in laptops and, more than 50% in PDAs [4], [6].
Mobile devices are driven by limited battery power. It is In this work clustering is performed independently
necessary to device a solution to reduce the energy of WLAN APs. Therefore, this approach does not require
consumption of WLAN interface without reducing the any changes to the existing infrastructures. Because of the
performance and modifying hardware and software large difference between the communication range WLAN
infrastructures. and Bluetooth, only devices close to the CH can use the
low-power radio to communicate CH and RN, and vice-
The Bluetooth is mainly designed for low energy versa. On the other hand, in this case most devices are not
consumption, requires only 10% of the WLAN energy [6]. needed to connect CH directly and each device in the cluster
46
All Rights Reserved © 2012 IJARCSEE

ISSN: 2277 – 9043

connected to the CHs indirectly through the other nearby guarantee that node is more than d hops away from the
devices. This increases the range of PAN and most devices leader, where d is a value selected for heuristic. The main
can obtain the benefit of energy saving as shown in Fig. objectives of this protocol are:
1(b).
1. Elects cluster heads in such a manner as to
favour their re-election in future rounds..
2. Cluster heads are well distributed.
3. Energy efficiency of wireless network is
improved by exploiting Bluetooth.
4. Clusters are dynamically configured to meet
the bandwidth need of all nodes.
5. Provide the load balancing among cluster
heads to insure a fair distribution among
cluster heads.
6. Reduce the transition over heads.

A Cooperative Multi-Hop Clustering Protocol
simulated for large networks and evaluates the performance.
This result demonstrates that Cooperative Multi-Hop
clustering protocol is effective in reducing the power
Fig. 1(a). A Single-Hop Clustering. consumption of WLAN-based communication system.

II. SYSTEM MEODL

The mobile devices considered in this paper are popular user
terminals like the smart phones. Here mobile device simply
referred as a node. The following properties are assumed
about the nodes:

1. Each node should capable of measure its
energy.
2. Each node has one WLAN interface and one
Bluetooth interface.
3. There is at least one WLAN AP in the field.
4. The WLAN AP do not have Bluetooth
interface.
5. All Bluetooth interfaces have the same range
of communication.
6. Each node must know the total bandwidth
need, and free bandwidth of the Bluetooth link.
Fig. 1(b). A Multi-Hop Clustering.
The main goal of this protocol is to reduce the energy
In this approach mobile devices are arbitrarily consumption in wireless networking applications. To
controlled by their users, this needs consideration of node achieve this goal design a general clustering protocol. The
mobility as well as a large variance of bandwidth need for essential operation in clustering is to select a set of cluster
various applications. However, because all devices have heads from the set of nodes in the network. Cluster heads
equal importance, changing the role of CH among all are responsible for coordinating the nodes within the cluster,
devices is necessary to distribute the energy consumption. and aggregation of the data from cluster members, and send
Nodes also can be turned off at any time and powered again the data to the APs. Clusters are formed to meet the user
depending on the user‟s need, which require the requirements and the following requirements:
consideration of unexpected link failure.
1. To minimize the number of cluster heads.
This paper presents a distributed Cooperative 2. To minimize the number, and size of the data
Multi-Hop Clustering Protocol. This work started with the structures need.
aim of generalizing the clustering heuristics so that a node is 3. Clustering should be completely distributed.
either a CH or a regular node atmost d-hops away from the 4. For each cluster the sum of bandwidth
CH. A new distributed leader election heuristic proposed to
requirements of all regular nodes within the
47

ISSN: 2277 – 9043

cluster must not exceed maximum data rate of The power consumption of anode using either
Bluetooth. WLAN or Bluetooth interface with data transfer rate „r‟ can
5. Communication between CHs and APs is be defined as fallows
single-hop.
𝑟 𝑟
6. At the end of the clustering process, each node PW ( r ) = PWA +
PWi (1— ) (1)
𝑅𝑊 𝑅𝑊
should be either cluster head or atmost d hops
𝑟 𝑟
away a cluster head. PB ( r ) = PBA + PBi ( 1— ) (2)
𝑅𝐵 𝑅𝐵
7. Clustering should be efficient in terms of
processing complexity and message exchange. Where RW is rate of WLAN and RB is rate of Bluetooth.
8. Re-elects CHs when possible.
9. Provides load balancing among cluster heads. By monitoring the idle time of WLAN and
Bluetooth channels during the period of time T, each node
III. THE MULTI-HOP COOPERATIVE CLUSTERING can estimate its FreeBWW and FreeBWB using a moving
PROTOCOL OPERATION average with weight α ∈ [0, 1] as follows:
𝑇𝑊
Several Multi-Hop Clustering algorithms are there. Among FreeBWW = α FreeBWW + (1—α) RW—R mar (3)
𝑇
them, the authors of [2] propose a heuristic to form a d-
dominating set. Within the cluster messages are exchanged 𝑇𝐵
FreeBWB = α FreeBWB + (1—α) R B — R mar (4)
𝑇
among the nodes via Bluetooth. When a node is newly
booted up, it becomes a CH and uses the Bluetooth for Where RW and, RB are the transmission rates of WLAN,
advertising its resource information periodically (say, every Bluetooth respectively. Rmar is a predefined constant used to
200 ms). The advertisement message contains the clustering maintain the free bandwidth to be slightly lower than the
cost, amount of bandwidth available for packet forwarding, actual available bandwidth.
ID, and the network address. Each node manages a set SCH,
which stores the information advertised by neighbouring Each CH waits for join request from other nodes
CHs. The nodes wait for a specific amount of time for CH for short time (say, 1second). The join message of a node
announcement (cluster joining offer) during the cluster includes the amount of required bandwidth, up on receiving
formation phase shown in Fig. 2. Hearing upon these offers, a join message, the CH compares its free bandwidth with the
each node selects the best-offer using the information required band width. If the CH has a sufficient amount of
received from the CH. The nodes that lie outside 1-hop free bandwidth for the sender, it will accept the request.
range of the CHs cannot hear any announcement directly.
They, however , receive offers from existing CMs to join a When there is no request, the CH counts the
cluster. number of RNs within the cluster. If there is at least one RN,
then CH keep continuous its current role. This allows the
RNs to select their next CH by themselves, which is
necessary for network stability. If the CH has no RNS, then
for saving energy it select the next CH by itself, The CH
selection process of Multi-hop Clustering cooperative
protocol executes the following Max-Min D-Cluster
Formation procedure.

Each node initiates 2d rounds of flooding. Each
node maintains a logged entry of the results of each flooding
round. The rounds are segmented in to the 1st d rounds and
the 2nd d rounds. The 1st d rounds are a flood max to
propagate the largest node Ids. After completion of the 1st d
rounds of flooding the 2nd d rounds of flooding begin, using
the values that exist at each node after the 1st d rounds. The
2nd d rounds of flooding are a flooding to allow the smaller
node Ids to reclaim some of their territory. After completion
of 2nd d rounds each node looks at its logged entries for the
2d rounds of flooding. The following rules explain the
logical steps that each node runs on the logged entries.
Fig. 2. A Multi-Hop Cluster Formation.
Rule 1: Each node check to see if it has received its
own original node Id in the 2nd d rounds of
flooding. If it has then it can declare itself as a CH
48

ISSN: 2277 – 9043

and skip the rest of the phase, otherwise proceed to uniformly dispersed into a 70×70 meters field. Because it is
rule 2. unrealistic to assume that all 100 nodes have the same
purpose of prolonging the group life time and consider
Rule 2: Each node looks for node pair, once a node individual networking in this solution. An ordinary WLAN
has identified all node pairs; it selects the minimum AP is located at (0, 0). Here assume that the WLAN and
node pair to be the CH. If node pair does not exist Bluetooth communication ranges are 100 and 10 meters
for a node then proceed to rule 3. respectively. Therefore, all nodes can communicate with the
access point anywhere in the field using their WLAN
Rule 3: Elect the maximum node Id in 1st d rounds
interface.
f flooding as the CH for the node.
A. EFFECT OF NODE DENSITY ON ENERGY
Floodmax- Each node locally broadcast its winner
value to all of its 1-hop neighbours. After all the Vary the number of nodes in the field from 10 to 100 to
neighboring nodes have been heard from for a study how Cooperative Multi-Hop Clustering protocol
single round, the node chooses the largest value works. Fig. 3. Show the effect of node density on
among its own winner value and the value communication energy consumption. In Cooperative Multi-
received in the round as its new winner. This Hop Clustering protocol energy consumption decreases as
process continues for d rounds. the node density increases, because more nodes can be
grouped into a cluster and share WLAN interface. A node
Floodmin-This follows Floodmax and also last d
belongs to a larger cluster can spend more time as RN.
rounds. It is the same as floodmax except a node
chooses smallest rather than the largest value as its
new winner.

Node Pairs: A node pair is any node Id that occurs
at least once a winner in both 1st and 2nd d rounds
of flooding for an individual node.

The heuristics has three logical stages, first the propagation
of larger node Ids via floodmax, second propagation of
smaller node Ids via floodmin, and third the determination
of cluster heads. The proposed heuristic provides an optimal
solution, when largest nodes Id are spaced d distance apart.

It is necessary to rotate the CH role regularly to
balance the energy consumption among all nodes. For this,
each RN has a timer, RN-timer, which expires in every TRN
Fig. 3. The effect of node density on the communication
seconds. When timer expires, the RN calls the CH selection
energy consumption.
process to select its next CH.
Fig3. Illustrate the relation between the node density energy
Due to the mobility of nodes, the distance between
consumption. It shows that as the number of nodes increases
the Ch and RN and its neighbour becomes longer than the
in the cluster, energy consumption decreases compare to the
Bluetooth range, and if the inbetween nodes turned off the
single Hop. As in Cooperative Multi-Hop Clustering as the
Bluetooth. In these cases, the RN immediately changes its
number of nodes increases from 10 to 100 the energy
CH.
consumption decreases to 57%.
The nodes considered here have a large variety of
B. EFFECT OF NODE DENSITY ON THROGHPUT
applications resulting in time-varying bandwidth
requirements with huge variations. Therefore, Nodes should As shown in Fig. 4. the number of nodes in the cluster are
selectively use either Bluetooth or WLAN depending on the increases, the through put decreases. But here in comparison
requirement. If the RN bandwidth requirement is not to the previous Single-Hop system in our Multi-Hop system
satisfied by the CH and it finds new CH. throughput increases slightly.
IV. SIMULATION AND PERFORMANCE EVALIATION

To evaluate the performance of Cooperative Multi-Hop
Clustering Protocol for large scale mobile networks via
extensive simulation. Here assume that 100 nodes are

49

ISSN: 2277 – 9043

Fig. 4. Effect of Number of nodes on throughput.

C. EFFECT OF PACKET DELAY Fig. 6. Effect of number of packets on energy
consumption.
As shown in the Fig. 5. as the number of nodes in the
cluster increases, the number of packets to be send also V. CONCLUSION AND FEATUREWORK
increases. As the number of packets increases the packet
In this paper presented Cooperative Multi-Hop Clustering, a
delay also increases. But in comparison to Single-Hop in
energy efficient multi radio mobile network. Cooperative
Multi-Hop system the packet delay is less.
Multi-Hop Clustering uses Bluetooth to reduce the power
consumption of WLAN in mobile devices. It dynamically
reconfigures the cluster based on the bandwidth requirement
of application without degrading performance. The
Cooperative Multi-Hop Clustering uses the Max-Min D-
Cluster formation for cluster head election. The Cooperative
Multi-Hop clustering also provides load balancing among
cluster heads and efficient in terms of processing complexity
and message exchanges. We can easily extend to other
interfaces such as WiMAX/Bluetooth.

VI. REFERENCES

[1] Jong-Woon Yoo, Kyu HoPark, “Cooperative
Clustering Protocol for Energy Saving of
Mobile Devices with WLAN and Bluetooth
Interfaces”, IEEE Transactions on Mobile
Fig. 5. Effect of number of packets on packet Computing, VOL. 10, NO.5, APRIL 2011.
delay. [2] A. D. Amis. R. Prakash, T.P. H. Vuong and D. T.
Hyunh, “Max-Min D-Cluster Formation in
D. EFFECT OF PACKET DELAY Wireless Ad hoc Networks”, Proc. IEEE Infocom,
PPP .32-41, 2000.
As the number of nodes in cluster increases the number of [3] Y. Agarwal, C. Schurgers, and R. Gupta,
packets to be sends also increases. As the number of packets “Dynamic Power Management Using on Demand
Paging for Networked EmbeddedSystems,” Proc.
increases the total energy consumption increases. As shown
Asia South Pacific Design Automation Conf., vol.
in the Fig. 6. in comparison to the Single system in our 2, pp. 755-759, 2005.
Multi-Hop system energy consumption is reduced by 35%. [4] T. Pering, Y. Agarwal, R. Gupta, and R. Want,
“CoolSpots: Reducing the Power Consumption of
Wireless Mobile Deviceswith Multiple Radio
Interfaces,” Proc. ACM MobiSys, pp. 220-
232,2006.
[5] Y. Agrawal, T. Pering, R. Want, and R. Gupta,
“SwitchR: Reducing System Power Consumption
in a Multi-Client, Multi-Radio Environment,” Proc.
IEEE Int‟l Symp. Wearable Computers, pp. 99-
102, 2008.
50

ISSN: 2277 – 9043

[6] B. Chen, K. Jamieson, H. Balakrishnan, and R. [14] J. Broch, D. Maltz, D. Johnson, Y. Hu, and J.
Morris, “Span: An Energy-Efficient Coordination Jetcheva, “A Performance Comparison of Multi-
Algorithm for Topology Maintenance in Ad Hoc Hop Wireless Ad Hoc Network Routing
Wireless Networks,” Proc. ACM MobiCom, pp. Protocols,” Proc. ACM MobiCom, pp. 85-97,
85-96, 2001. 1998.
[7] E. Shih, P. Bahl, and M.J. Sinclair, “Wake on [15] S.H. Shah, K. Chen, and K. Nahrstedt, “Available
Wireless: An Event Driven Energy Saving Strategy Bandwidth Estimation in IEEE 802.11 Based
for Battery Operated Devices,” Proc. ACM Wireless Networks,” Proc. Workshop Bandwidth
MobiCom, pp. 160-171, 2002. Estimation (BEst ‟03), 2003.
[8] A. Chaman and S. Pierre, “On the Planning of [16] N. Golmie, N. Chevrollier, and O. Rebala,
Wireless Sensor Networks: Energy-Efficient “Bluetooth and WLAN Coexistence: Challenges
Clustering under the Joint Routing and Coverage and Solutions,” IEEE Trans. Wireless Comm., vol.
Constraint,” IEEE Trans. Mobile Computing, vol. 10, no. 6, pp. 22-29, Dec. 2003.
8, no. 8, pp. 1077-1086, Aug. 2009. [17] Y. Wu, X.-Y. Li, Y. Liu, and W. Lou, “Energy-
[9] O. Younis and S. Fahmy, “HEED: A Hybrid, Efficient Wake-Up Scheduling for Data Collection
Energy-Efficient, Distributed Clustering Approach and Aggregation,” IEEE Trans. Parallel and
for Ad Hoc Sensor Networks,” IEEE Trans. Mobile Distributed Systems, vol. 21, no. 2, pp. 275-287,
Computing, vol. 3, no. 4, pp. 366-379, Oct.-Dec. Feb. 2010.
2004. [18] Saeed Ebadi, Arsalan Va hi, Nader Vahdani Manaf,
[10] W. Heinzelman, A. Chandrakasan, and H. Saeed Rasouli, “A NEW MULTI-HOP AND
Balakrishnan, “An Application-Specific Protocol HIERARCHICAL CLUSTERING ALGORITHM
Architecture for Wireless Microsensor Networks,” FOR ENERGY SAVING IN WIRELESS
IEEE Trans. Wireless Comm., vol. 1, no. 4, pp. SENSOR NETWORK”, IEEE 2010.
660- 670, Oct. 2002. [19] Zhan Huawei, Zhou Yun, ” A Power Management
[11] IEEE 802.11 (Legacy), Wireless LAN Medium Protocol for Multi-hop Ad Hoc networks”, IEEE
Access Control (MAC) and Physical Layer (PHY) 2009.
Specifications, IEEE, 1997. [20] Long Cheng∗†, Sajal K. Das†, Mario Di
[12] M. Cheng, X. Gong, and L. Cai, “Joint Routing and Francesco†, Canfeng Chen‡ and Jian
Link Rate Allocation under Bandwidth and Energy Ma∗§,Scalable and Energy-Efficient Broadcasting
Constraints in Sensor Networks,” IEEE Trans. in Multi-hop Cluster-Based Wireless Sensor
Wireless Comm., vol. 8, no. 7, pp. 3770- 3779, July Networks”, IEEE ICC 2011.
2009.
[13] Bluetooth Special Interest Group, “Bluetooth
Personal Area Networking Profile,” Specification
of the Bluetooth System, Ver. 1.1, 2001.

P. Narasimhaiah M.Sc (M. Tech)
Dept. of C S E, C B I T, Proddatur- 516360

Mrs. S. Vasundra M. Tech. (Ph. D)
Dept. of C S E, JNTU College of Engineering (A)
Anantapur-515002

51

46 51

More Related Content

What's hot (7)

Viewers also liked (8)

Similar to 46 51 (20)

More from Ijarcsee Journal (20)

Recently uploaded (20)

46 51