An Energy Efficient Routing Scheme for Wireless

0 downloads 0 Views 499KB Size Report
Energy Adaptive Clustering Hierarchy (LEACH) [5] and Threshold-Sensitive Energy Efficient Sensor. Network Protocol (TEEN) [6], which employ different.
Fifth International Conference on Computational Science and Applications

An Energy Efficient Routing Scheme for Wireless Sensor Networks Ki Young Jang, Kyung Tae Kim, Hee Yong Youn School of Information and Communication Engineering Sungkyunkwan University, 440-746, Suwon, Korea [email protected] ,[email protected], [email protected] Abstract

classified into flat routing protocol, hierarchical routing protocol, and location-base routing protocol [2]. The representative flat routing protocol is Sensor Protocols for Information via Negotiation (SPIN) [3] and Direct diffusion [4]. With flat routing protocol, all nodes use a same routing technique with equal opportunity and they have same role and function. The representative hierarchical routing protocols are LowEnergy Adaptive Clustering Hierarchy (LEACH) [5] and Threshold-Sensitive Energy Efficient Sensor Network Protocol (TEEN) [6], which employ different hierarchical structure based on the clustering approach. The Geographic Adaptive Fidelity (GAF) [7] and Geographic and Energy Aware Routing (GEAR) [8] are the representative location-based routing protocol. They use position information of sensor nodes in establishing the routing path to the base station. With LEACH, the representative hierarchical routing protocol, the aggregated data in the cluster head are transmitted to the base station. The lifetime of the network is increased by selecting the head node randomly which takes the responsibility for the transmission of data to the base station. Because of the rotational selection of head node, the energy load of data transmission among the sensors is evenly distributed throughout the network. However, the cluster head selection algorithm of LEACH does not guarantee maximum lifetime of each node since the cluster head nodes are chosen randomly without considering the status on the energy of the nodes. LEACH-C [9] thus employs a cluster head selection algorithm electing the head node according to the ratio of total energy remaining in the network to the energy remaining in each node. Even though it improves the network lifetime, the overhead for finding out the total energy in the network is still not negligible. In this paper we propose an energy efficient routing scheme which further improves the LEACH and LEACH-C scheme. The proposed scheme includes three mechanisms. First, the node having the largest remaining energy is selected as the cluster head when the energy level drops below 50% of the initial energy. Second, each node receiving the advertisement

Since sensor nodes have limited energy, energy efficient routing is very important for wireless sensor network (WSN). In this paper we propose a routing scheme which significantly enhances the performance of existing schemes such as LEACH and LEACH-C. The proposed scheme selects the cluster head not randomly but considering the remaining energy when the energy level drops below 50% of the initial energy. For each node, the cluster head to join is determined by not only the signal power but also the remaining energy of the cluster head. Also, data transmission occurs when the context satisfies the preset condition. NS-2 simulation shows that the proposed scheme outperforms LEACH and LEACH-C by 37% and 30%, respectively, in terms of the lifetime of the sensor network.1

1. Introduction Due to recent technological advances in the manufacturing of small and low-cost sensors, the wireless sensor network can be applied to numerous applications. It consists of a large number of sensor nodes which have the characteristics of low price, low power, and deployment without a predetermined structure. Here the adjoining sensor nodes sense similar information, and the whole activity of the sensor network is not affected seriously even if some sensor nodes fail or get partially damaged. Nowadays, the routing protocol reflecting the unique characteristic of wireless sensor network receives a great deal of attention [1]. The routing protocol for sensor network proposed so far can be 1

This research was supported by the Ubiquitous Autonomic Computing and Network Project, 21st Century Frontier R&D Program in Korea, R01-2006-000-10954-0, Basic Research Program of the Korea Science & Engineering Foundation. Corresponding author: Hee Yong Youn

0-7695-2945-3/07 $25.00 © 2007 IEEE DOI 10.1109/ICCSA.2007.10

399

Authorized licensed use limited to: Guangdong Univ of Tech. Downloaded on December 24, 2008 at 21:54 from IEEE Xplore. Restrictions apply.

random number. A node can become a cluster head again after 1/k round, and the elected cluster head node broadcasts a message containing its own information to the neighbor nodes.

message from the cluster heads select its cluster head based on the cost value which is determined by the signal power and distance to respective cluster head. Third, the data transmission at each node according to the TDMA schedule is allowed when the context satisfies the preset condition. Otherwise, they put themselves into the Sleep mode for saving the energy. The proposed mechanisms will increase the lifetime of each sensor node, which eventually increases the lifetime of the entire sensor network. NS-2 simulation shows that the proposed scheme outperforms LEACH and LEACH-C by 37% and 30%, respectively, in terms of the lifetime of the sensor network. The rest of the paper is organized as follows. Section 2 reviews the hierarchical routing protocol using LEACH, and Section 3 introduces the proposed scheme. Section 4 presents the result of performance evaluation of the proposed scheme along with LEACH and LEACH-C using NS-2. Finally, Section 5 concludes the paper.

Step 2: Cluster set-up When a node receives an advertisement messages from the cluster heads, it sends the join-request message (Join-REQ) to the cluster head which it chooses as its cluster head based on the received signal power. This message includes the node’s ID and cluster head’s ID. Figure 1 illustrates the clustering operation of LEACH.

2. The Related Work In this section we review the operation of hierarchical routing protocol using LEACH and discuss the factors related to energy consumption of wireless sensor network.

Figure 1. Clustering of LEACH

2.1 LEACH

Step 3: Schedule creation

In the LEACH scheme, the cluster heads are chosen from the nodes with uniform distribution. The TDMA time schedule is adopted between the cluster head and member nodes to avoid collision, and the base station classifies the cluster heads by CDMA schedule. Also, each cluster head assumes direct communication to the base station. The operation of LEACH consists of four steps which are cluster head election step, cluster set-up step, schedule creation step, and data transmission step.

Each cluster head transmits a schedule to its own member nodes in the cluster after it creates a TDMA schedule. After this step, the clustering formation process is over. Step 4: Data transmission The member nodes in each cluster transmit data to their head node within the time slot allocated by the TDMA schedule. Each cluster head sends the data to the base station, which are collected from the member nodes and then aggregated by data fusion. Figure 2 shows the structure of one round of operation of LEACH protocol.

Step 1: Cluster head election

k P (n) =

1 1 − k * ( r mod ) k 0

if

n ∈ C (n )

(1)

otherwise

Here k and r is the probability to become a cluster head and round number, respectively. Round is the unit of time in LEACH. C is the set of nodes that have been cluster heads in the last 1/k rounds. Each node compares a random number between 0 and 1 with the calculated threshold value, and elects itself to be a cluster head in case of the threshold is bigger than the

Figure 2. One round of LEACH operation LEACH operates according to the time unit called round. Each round consists of set-up phase in which clusters are formed and steady-state phase consisting of TDMA frames. In the steady-state phase, each member

400

Authorized licensed use limited to: Guangdong Univ of Tech. Downloaded on December 24, 2008 at 21:54 from IEEE Xplore. Restrictions apply.

consumption between the sensor nodes. LEACH-C considers the remaining energy of all nodes to choose the cluster heads but does not consider the remaining energy of respective node. As a result, they may select the nodes of insufficient energy as cluster heads, which limits the lifetime of sensor nodes and eventually entire network. Hence, we propose a new cluster head selection scheme minimizing the chance of selecting the nodes of low energy as cluster heads. For this, we employ both the LEACH approach and a new approach, and apply one approach according to the condition of the remaining energy in each node. When the remaining energy is larger than 50% of the initial energy in a node, the LEACH algorithm is applied as represented in Equation (1). Otherwise, the proposed approach which considers the remaining energy in each node is applied as shown in Equation (4).

node transmits data during its own timeslot and reduces the energy consumption by entering the Sleep mode during the remaining timeslots.

2.2 Energy Consumption in Sensor Network Figure 3 shows the propagation model of wireless sensor network.

Figure 3. The propagation model of wireless sensor network Energy consumption for data increases in proportion to the square of size of packet transmitted, while energy for reception rises as the packet size shown in Equation (2) and (3).

transmission distance and consumption increases as

ETx (k , d ) = k ⋅ Eelec + k ⋅ ε amp ⋅ d 2 E Rx (k ) = k ⋅ E elec

Pi (n ) = 2k *

Eresidual Einit

(4)

Here k denotes the probability to become a cluster head, Eresidual does the remaining energy in the node, and Einit does the initial energy of the node, respectively. In case of the Pi(n) value is larger than a random number between 0 and 1, it becomes a cluster head. If there is relatively large remaining energy in a node (say more that 50% of Einit), the probability that the node becomes the cluster head will be high since LEACH algorithm is applied. On the contrary, if the remaining energy is small, the probability that the node becomes a cluster head will be decreased by using the proposed scheme. Consequently, the lifetime of entire sensor network can be extended.

(2) (3)

Generally, a lot of radio communications using a radio frequency module result in large energy consumption in the sensor nodes. Therefore, radio communication needs to be reduced as much as possible.

3. The Proposed Scheme In this section we propose a new cluster head selection algorithm considering the remaining energy of each node. The scheme for selecting the cluster head to join for the member nodes is also introduced, which is based on the cost value. We also propose an approach with which the decision on data transmission is made according to the sensed environment condition.

3.1 Cluster-head Election With hierarchical routing protocol, three steps are involved which are clustering, data transmission scheduling, and data transmission. LEACH chooses cluster heads probabilistically assuming equal energy

Figure 4. Proposed cluster election scheme flow chart

401

Authorized licensed use limited to: Guangdong Univ of Tech. Downloaded on December 24, 2008 at 21:54 from IEEE Xplore. Restrictions apply.

Figure 4 is showed that energy dissipation is very low when the percent of nodes that are cluster-heads maintained within 5% in case of LEACH protocol. Therefore, the number of cluster heads is above 4% scheme are very efficient in side of energy efficiency in the decision box of Figure 4.

the sensed data to the head node during its timeslot indicated by the TDMA time schedule. Note that the member nodes always transmit data during its allocated timeslot. Therefore, energy might be wasted unnecessarily when the sensed data are useless or unimportant. In order to solve this problem, we propose a scheme which lets the sensor nodes send the data to the cluster head only when a certain condition on the context is satisfied such as “Does the temperature exceed 30 degree?”. In its own timeslot, if the condition is not satisfied, the member node enters Sleep mode to reduce the energy consumption. Figure 6 shows the operational structure of the proposed scheme.

3.2 Forming Clusters In the case of LEACH, each node receiving the advertisement messages from the cluster heads decides the node to join as the cluster head depending on the received signal strength (RSS). We propose a cost function used for deciding the node to join, which includes the remaining energy and signal power of cluster head as shown in Equation (5).

Cost (i) =CH(i) remaining energy +CH(i) signal power

set-up steady-state phase phase chek transmission algorism energy_based chek transmission algorism at each node i, j, k ... at each node i, j, k ... cluster head ... ...... selection frame round

(5)

...

energy_based cluster head selection

...

Figure 6. The operational structure of the proposed scheme

Here CH(i) remaining energy and CH(i) signal power are the remaining energy and signal strength of Cluster Head(i), respectively. Each node receiving the advertisement messages calculates the cost value, and then joins the cluster head of the largest cost value by sending the join-request (Join-REQ) message to the corresponding cluster head.

4. Performance Evaluation We evaluate the lifetime of sensor network using NS-2. In the simulation the size of network is 1000m × 1000m and the size of a packet is 500byte. Also, the initial energy of each sensor node is 2J, the radio electronics energy is 50nJ/bit, and the radio amplifier energy is 100pJ/bit/m2. Table I shows the simulation environments. Table I. The simulation environments Parameter Value

Figure 5. Proposed join scheme to cluster head flow chart Figure 5 is showing the detail procedure both side of cluster head node and cluster member nodes to join to cluster head by sending Join-REQ message.

Size of network

1000m x 1000m

Size of packets

500 byte

Eelec(Radio electronics energy)

50 nJ/bit

Eamp(Radio amplifier energy) Einit(Initial energy of each node)

100 pJ/bit/m2 2J

Number of nodes

100

Figure 7 represents the simulation result n the network lifetime. Here, X-axis denotes the round which is the unit of time. Y-axis presents the number of active nodes. The simulation includes the cluster head selection phase, the data transmission phase from each node to the head nodes, and the data transmission phase from the head nodes to the base station.

3.3 Data Transmission Each cluster head decides a TDMA time schedule and informs the member nodes of the schedule after the cluster is formed in LEACH. The nodes then transmit

402

Authorized licensed use limited to: Guangdong Univ of Tech. Downloaded on December 24, 2008 at 21:54 from IEEE Xplore. Restrictions apply.

In this paper we have proposed a scheme that improves the LEACH and LEACH-C scheme in terms of energy efficiency for wireless sensor network. The proposed scheme consists of three algorithms. First, the new cluster head is elected considering the remainder of energy of each node in the sensor network when the energy level is below a certain value. Second, the cluster head to join is selected according to the cost value determined by the signal power of the cluster head and the distance to the cluster head. Third, data transmission from the sensor nodes to the cluster head is made when the preset condition on the context is satisfied. The results of simulation based on NS-2 show that the lifetime of sensor network is significantly improved compared to the LEACH and LEACH-C scheme. The effectiveness of the proposed scheme will be investigated for more complicated environment of multihop communication in the future.

Figure 7. The number of live nodes as time passes Notice from the figure that the proposed scheme allows a significantly larger number of live nodes than LEACH and LEACH-C. The LEACH and LEACH-C scheme take 60 rounds and 70 rounds for having the first dead node, respectively. However, the proposed scheme takes 420 round of operation for the first dead node to appear. In addition, The LEACH and LEACHC scheme take 430 rounds of operation for the last node to die. It is 640 rounds for the proposed scheme. The simulation results clearly demonstrate that the proposed scheme consumes the energy much more effectively than the LEACH and LEACH-C scheme.

6. References [1] I.F. Akyildiz, Su Weilian, Y. Sankarasubramaniam, and E. Cayirci, “A survey on sensor networks,” IEEE Communications Magazine, Vol. 40, No. 8, pp. 102-114, 2002. [2] J.N. Al-karaki and A.E. Kamal, “Routing techniques in wireless sensor networks: a survey,” IEEE Communication Magazine, Vol. 11, No. 6, pp.6-28, 2004. [3] W. Heinzelman, J. Kulik, and H. Balakrishnan, “Adaptive protocols for information dissemination in wireless sensor networks,” Proc. 5th ACM/IEEE Mobicom, pp. 174–85, 1999. [4] C. Intanagonwiwat, R. Govindan, and D. Estrin, “Directed diffusion: a scalable and robust communication paradigm for sensor networks,” Proc. ACM MobiCom 2000, pp. 56–67, 2000. [5] C. Schurgers and M.B. Srivastava, “Energy efficient routing in wireless sensor networks,” MILCOM Proc.Commun. for Network-Centric Ops.: Creating the Info. Force, 2001. [6] A. Manjeshwar and D. P. Agarwal, “TEEN: a routing protocol for enhanced efficiency in wireless sensor networks,” 1st Int’l. Wksp. on Parallel and Distrib. Comp. Issues in Wireless Networks and Mobile Comp., 2001. [7] Y. Xu, J. Heidemann, and D. Estrin, “Geographyinformed energy conservation for ad-hoc routing,” Proc. 7th Annual ACM/IEEE Int’l. Conf. Mobile Comp. and Net., pp. 70– 84, 2001. [8] Y. Yu, D. Estrin, and R. Govindan, “Geographical and energy-aware routing: a recursive data dissemination protocol for wireless sensor networks,” UCLA Comp. Sci. Dept. tech. rep., UCLA-CSD TR-010023, 2001. [9] W.B. Heinzelman, A.P. Chandrakasan, "An applicationspecific protocol architecture for wireless microsensor networks," IEEE Transaction on Wireless Communication, Vol. 1, No. 4, pp. 660-670, 2002. [10] C.-Y. Wan, A.T. Campbell, and L. Krishnamurthy, “Pump-slowly, fetch-quickly (PSFQ): a reliable transport protocol for sensor networks,” IEEE Journal on Selected

Figure 8. The time when a node dies first with varying initial energy in each node Figure 8 shows the time when a node dies for the first time in the network as the initial energy of each node is varied from 10J to 25J. Observe that the proposed scheme always outperform the other two schemes, while the node life time almost linearly grows as the initial energy increases.

5. Conclusions and Future Work

403

Authorized licensed use limited to: Guangdong Univ of Tech. Downloaded on December 24, 2008 at 21:54 from IEEE Xplore. Restrictions apply.

Areas in Communications, Vol. 23, No. 4, pp. 862-872, 2005. [11] W. Chonggang, K. Sohraby, H. Yueming, L. Bo, and T. Weiwen, “Issues of transport control protocols for wireless sensor networks,” Proceedings International Conference on Communications, Circuits and Systems, Vol. 1, Vol. 1, pp. 27-30, 2005. [12] UCB/LBNL/VINT Network Simulator [Online]. Avilable: http://www.isi.edu/nsnam/ns [13] K.T. Kim and H.Y. Youn.: EDACH: Energy-Driven Adaptive Clustering Hierarchy for Wireless Sensor Networks: Proceeding of The EUC 2005 Workshops, 2005, pp. 1098-1107. [14] K.T. Kim and H.Y. Youn.: PEACH: Proxy-Enable Adaptive Clustering Hierarchy for Wireless Sensor network: Proceeding of The 2005 International Conference On Wireless Network, June 2005, pp. 52-57.

404

Authorized licensed use limited to: Guangdong Univ of Tech. Downloaded on December 24, 2008 at 21:54 from IEEE Xplore. Restrictions apply.