A Cluster-Based Data Routing for Wireless Sensor ... - Semantic Scholar
A Cluster-Based Data Routing for Wireless Sensor Networks Hao-Li Wang and Yu-Yang Chao Department of Computer Science and Information Engineering, National Chiayi University, Syuefu Rd., Chiayi City 60004, Taiwan {haoliwang,yychao94}@gmail.com
Abstract. To promise low energy consumption for wireless sensor network routing, a number of routing protocols have been published. A novel routing protocol called BeamStar was proposed in [6]. In BeamStar, each base station uses a directional antenna with power control to assist locating sensors. It shifts the burden of network control and management from resource-limited sensors to resource-abundant and more sophisticate base stations, but still has three serious problems. The different size between the nearer and farther region, base station's scanning time, and a great quantity of inter-node communication all waste precious energy. In the cause of solving these problems, we present a routing protocol, namely Cluster-based BeamStar (CBS), for wireless sensor networks. CBS follows cluster-based structure to decrease the volume of internode communication and we have reformed the scanning manner in BeamStar. Our simulations show that CBS is scalable, low-cost, and energy efficient. Keywords: cluster, wireless sensor network, routing, directional antenna, power control.
1 Introduction The recent advances in wireless communication, MEMS technology, and light-weight Operation System have greatly reduced the size and cost of wireless sensor devices. Because having the features of low-cost, non-infrastructure and fault-tolerance, wireless sensor networks can operate in kinds of environments and be used in military, disaster, and surveillance usage [1]. As shown in Fig. 1, a wireless sensor network consists of a large number of sensors, randomly deployed in an interested region for detecting and monitoring tasking [2]. These tiny sensing devices have processing and wireless communication capabilities, which enable them to gather information from the environment and to generate and transmit report messages to base station. According to [3], the key networking challenges in wireless sensor network can divide into four broad categories: limiting radio operation, data management, geographic routing, and system monitoring maintenance. In terms of routing, because of severe energy constraints and unreliable network nodes, most routing protocols proposed for Mobile Ad Hoc networks are not suitable for wireless sensor networks, or A. Hua and S.-L. Chang (Eds.): ICA3PP 2009, LNCS 5574, pp. 129–136, 2009. © Springer-Verlag Berlin Heidelberg 2009
130
H.-L. Wang and Y.-Y. Chao
Fig. 1. Wireless sensor network architecture
cannot be used in wireless sensor networks without any modification [4]. In general, routing in wireless sensor network can be classified into three categories depending on the network structure: (1) (2) (3)
Flat-based: Direct diffusion [5] and BeamStar [6]. Hierarchical-based: LEACH [7] and Fault-tolerant clustering of wireless sensor networks [8]. Location-based: GAF [9] and SPAN [10].
We presented here is a revised BeamStar routing protocol called Cluster-based BeamStar (CBS). The cluster architecture such as LEACH protocol is used to reduce the volume of inter-node communication and the energy load will be distributed evenly. The rest of the paper is organized as follows. Section 2 reviews the related work. Section 3 presents our proposed method. Section 4 describes the performance evaluation. Section 5 concludes the paper.
2 Related Works We will simply introduce one protocol in each category mentioned above. A. BeamStar BeamStar is proposed in [6] which is a brand-new design concept. It is motivated by two observations: (1) Since management and control are indispensable, especially in large scale networks, BeamStar shifted computational intensive and energy consuming routing control overhead from sensor nodes to base stations. (2) Sensor nodes spend most of the energy on sensing and data forwarding. In BeamStar, the base station scans the network using a power controlled directional antenna. The relative angle of the sensor to the base station is determined by the direction of the base station transmission, while the distance of a sensor to the base station can be uniquely determined by the power level used for a received control message. As shown in Fig. 2, a sensor node’s ID is defined as the combination of its RN and SN. The node’s directionality of the last base station transmission is called Sector Number (SN), and the lowest power level information that it can receive from the base station is called Ring Number (RN). When the event occurs, each node simply compares their ID and the sequence number of the packet, then either drops or rebroadcasts the packet. The forwarding strategy is depending on the ID so there is no need for maintaining a routing table. With BeamStar, each data packet is relayed in an interleaved, loop-free
A Cluster-Based Data Routing for Wireless Sensor Networks
131
Fig. 2. BeamStar location discovery process [6]
mesh constrained in a sector towards the BS, making data delivery robust to sensor failures and transmission error. B. LEACH Heinzelman, et al. [7] introduced a hierarchical algorithm for wireless sensor networks, called Low Energy Adaptive Clustering Hierarchy (LEACH). LEACH is a clustering-based routing protocol that utilizes alternately select cluster-heads to distribute the energy load among the sensor nodes in the network evenly. The operation of LEACH is separated into two phases: the setup phase and the steady state phase. In the setup phase, each node has elected itself a cluster-head by a certainly formula (more interpretations will see in [7]) for the current round. All elected cluster-heads broadcast an advertisement message to rest of the nodes that they are the new clusterheads in current round. According to the signal strength of the advertisement message, the non- cluster-head nodes decide on the cluster to which they want to belong. After receiving all the messages from the nodes that would like to be included in the cluster, the cluster-head node creates a TDMA schedule for each node and broadcasts to all the nodes in the cluster. Once the clusters are created and the TDMA schedule is fixed, the stead state phase starts. In stead state phase, sensor nodes start sensing and forwarding packets to the cluster-heads. When the event occurs, the node who have sensed will send the sensing data to its cluster-head. After all the data has been received, the cluster-head performs data aggregation to reduce the redundant data and sends the report packets to the base station. After a certain time, the next round begins and each node returns to the setup phase. LEACH randomly selects a few sensor nodes as cluster-heads and rotates this role to evenly distribute the energy load among the sensors in the network. C. GAF Xu et al. [9] propose a location-based ad hoc routing protocol, which is called geographical adaptive fidelity (GAF). GAF conserves energy by identifying nodes that are equivalent from a routing perspective and then turning off unnecessary nodes, keeping a constant level of routing fidelity. The network area here is divided into fixed zones and forms a virtual grid. Each node uses GPS-indicated location information to association itself with a virtual grid, where all nodes in the same grid square are equivalent with respect to forwarding packets. For each grid, nodes coordinate with each other to elect a node awake for a certain period of time, and then the rest go
132
H.-L. Wang and Y.-Y. Chao
to sleep. This node is responsible for monitoring and forwarding data to the BS on behalf of the nodes in the same zone. Thus, by turning off the unnecessary nodes in the network, GAF economizes energy and keeps the level of routing fidelity simultaneously. In GAF, nodes are in one of three states: discovery, active, and sleep. The initial state of each node is discovery. When in state discovery, a node turns on its radio and exchanges discovery messages to find other nodes in the same grid. After discovery state, the node enters active state and takes charge of forwarding data. While an awake node can determine some other equivalent node will handle routing, it powers down its radio and changes state to sleep. The highest remaining energy node will always be selected stay in the active state by node ranking. In order to handle mobility, each node estimates the time it expects to leave its grid and includes this information in the discovery message. The sleeping nodes in the same grid adjust their sleeping time accordingly. Before the active nodes leaves, sleeping nodes wake up and one of them becomes active. This change does not change the node rank, but nodes will sleep for a shorter time. GAF is implemented both for non-mobility and mobility of nodes. It performs at least as well as a normal ad hoc routing protocol for packet loss and route latency, and yet it can substantially conserve energy, allowing network lifetime to increase in proportion to node density.
3 Proposed Routing Protocol 3.1 The Problems of BeamStar Though BeamStar protocol has reduced the burden of Sensor Node, three problems of this protocol was emerged. (1)Area of the farther region (i.e. the region with a larger RN) will larger, which means sensor nodes will not evenly distribute in each region. This drawback will lead to three differences. First, in accuracy, event locating becomes inaccurate with the increase of the area. Second, more sensor nodes have to broadcast for an event in a larger region. It will deplete much rare energy. Third, for the frequency of event occur, the larger area certainly more than the smaller one. That is to say, the farther regions have a higher probability to exhaust. (2)The number of base station’s scanning is enormous. Locating will consume a long period of time. For example, in fig. 2, base station needs to scan 15 times for a 5×3 area. It is inefficient in a dense events wireless sensor network. (3)The times of inter-node communication is huge. When the event occurs, source node generates a data packet and broadcasts it to its neighbors. The packet gradually broadcasts to base station and wastes much precious energy. 3.2 Cluster-Based BeamStar In order to improve the drawbacks mentioned above, a revised BeamStar routing protocol called Cluster-based BeamStar (CBS) is proposed. The protocol can divide into three phases:
A Cluster-Based Data Routing for Wireless Sensor Networks
133
：
Phase 1 Locating phase Similar to base station in BeamStar, base station in CBS performs powered controlled scan of the sensor network by using a directional antenna. Different from BeamStar, CBS makes change in the radius of region and the manner of scanning whole network. In order to keep the area of each region the same, CBS adjusts the radius R’ of region such as:
R′ =
iR
, ∀i = 1, 2,..., n
(1)
For example, the radius of region with ID{1, 1} is R, and the radius of region with ID{1, 2} will become 2 R, etc. We have also reformed the manner of scanning. As shown in Fig. 3(a) ~ (c). First, according to (1), the base station broadcasts the control message Ring Number to the whole network in order. Then, in Fig. 3(d) ~ (f), the base station will adjust its phase angle and broadcast Sector Number to the farthest region. All regions at the same phase angle can receive this control message and get their finally ID pair.
Fig. 3. An illustration of locating phase
：
Phase 2 Cluster building up phase After the locating phase, each region is regarded as a cluster such that the sensor nodes with the same ID are considered in the same cluster. The cluster members begin to exchange their remaining energy. Node in cluster with the largest remaining energy will be elected as the cluster-head. After a certain time, cluster-head broadcasts an advertisement message to its subordinates to declare it is the cluster-head in this round.
：
Phase3 Data transferring phase When the event occurs, the nodes which have sensed will broadcast the data around their own cluster. After all the data has been received, the cluster-head performs data aggregation to compress the data into a single signal. Only inter-cluster transmission
134
H.-L. Wang and Y.-Y. Chao
is needed in the following communication. The cluster-node will broadcast the control message to all the member nodes when its energy drops below a certain threshold. The nodes who have received this message will return their remaining energy to cluster-head. Then the most robust node was noticed to act as a cluster-head from the original cluster-head. An advertisement message was announced by the new clusterhead to its cluster member.
4 Comparison We have compared the performance of CBS to BeamStar in three aspects. 4.1 Difference of Sensor Nodes in Regions In our simulation, 100 nodes are randomly deployed in an area of radius 100. The length of RN is 20, 40, 60, 80, and 100 in BeamStar; r, 2 r, 3 r, 2r, and 5 r in ours which r 100/ . Base station is located in the center of the area. Fig. 4 shows the results of sensor nodes scatter in the deployed area. BeamStar has an obvious increase with the larger RN, and CBS keep a steady quantity in each region. It means nodes distributing are evenly. Furthermore, the regions which were distant from the base station will not increase their area can still keep the accuracy.
＝
4.2 Base Station Scanning Time As shown in Fig. 5, we set the maximum SN equals 20 and maximum RN is 10, i.e. 200 regions will be divided. Assume each time base station to scan a region is 1 second. BeamStar spends 200 seconds to scan the whole network, but CBS only needs 30 seconds. CBS shortens the scanning time substantially. 4.3 Data Transferring The Fig. 6 shows the different data transferring manners of BeamStar and CBS. Suppose that an event is detected by a sensor node in a region with ID{2, 3}. In BeamStar, this sensor node will generate a data packet and broadcast it to its neighbors. At 40
Number of sensor nodes
35
CBS
30
BeamStar
25 20 15 10 5 0 1
2
RN 3
4
5
Fig. 4. Node distributing in the region
A Cluster-Based Data Routing for Wireless Sensor Networks
135
200
Scanning time (sec.)
180 160
CBS
140
BeamStar
120 100 80 60 40 20 0 20
40
60
80
100 120 140 160 180 200
Number of region
Fig. 5. Scanning frequency of two protocols
Fig. 6. Data transferring of BeamStar (a) and CBS (b)
the four boundaries of this region, only the region with ID{2, 2} will further relay this packet. By this method, the packet will be eventually delivered to the base station. CBS just need to broadcast in the region with ID{2, 3}. After cluster-head in this region has received the data packet. Only inter-cluster transmission is needed in the following communication. As seen in TABLE I, inheriting from BeamStar, CBS performs better than LEACH in scalability. By contrast, CBS reduces the volume of inter-node communication than BeamStar and the energy load among the all the sensor nodes has evenly distributed. When the event occurs frequently, the cluster-head selection will more satisfied because it was depending on the remaining energy instead of the time interval. The cluster-heads in CBS also perform data aggregation to reduce the redundant. Table 1. Comparison with three protocols
Scalability Power usage Inter-node communication Cluster-head selection Data aggregation
BeamStar Good Low More
LEACH Low Good Less
CBS Good Good Less
×
Time-based
Energy-based
No
Yes
Yes
136
H.-L. Wang and Y.-Y. Chao
5 Conclusion Although BeamStar protocol shifted computational intensive and energy consuming routing control overhead from sensor nodes to base stations. It still has three lethality shortcomings mentioned in Section 3. We propose a new routing protocol named CBS to overcome there drawbacks. It is unnecessary to store a routing table and equip with additional hardware (such as GPS). CBS selects cluster-head alternately which also evenly distribute the energy load among the sensor nodes in the network. We also improve the scanning manner to decrease the scanning frequency and shorten the locating time. It will efficiently prolong the lifetime of the wireless sensor network. The proposed routing protocol is scalable, low-cost and energy efficient for wireless sensor networks.
References 1. Xu, N.: A Survey of Sensor Network Applications, http://enl.usc.edu/~ningxu/ 2. Al-Karaki, J.N., Kamal, A.E.: Routing Techniques in Wireless Sensor Network: A Survey. Proceedings of IEEE Wireless Communications 11(6), 6–28 (2004) 3. Deepak, G., Alberto, C., Yan, Y., Wei, Y., Jerry, Z., Deborah, E.: Networking Issues in Sensor Networks. Journal of Parallel and Distributed Computing 64(7), 799–814 (2004) 4. Tian, D., Georganas, N.D.: Energy efficient routing with guaranteed delivery in wireless sensor networks. Proceedings of Wireless Communications and Networking 3, 1923–1929 (2003) 5. Intanagonwiwat, C., Govindan, R., Estrin, D.: Directed diffusion a scalable and robust communication paradigm for sensor networks. In: Proceedings of the 6th annual international conference on Mobile computing and networking (August 2000) 6. Shiwen, M., Hou, Y.T.: BeamStar: A New Low-cost Data Routing Protocol for Wireless Sensor Networks. In: Proceedings of Global Telecommunications Conference, 2004, vol. 5, pp. 2919–2924 (2004) 7. Wendi, R.H., Anantha, C., Hari, B.: Energy-efficient communication protocol for wireless microsensor networks. In: Proceedings of the 33rd Annual Hawaii International Conference, vol. 2, p. 10 (2000) 8. Gupta, G., Younis, M.: Fault-tolerant clustering of wireless sensor networks. In: Proceedings of Wireless Communications and Networking, 2003, vol. 3, pp. 1579–1584 (2003) 9. Xu, Y., Heidemann, J., Estrin, D.: Geography-informed Energy Conservation for Ad-hoc Routing. In: Proceedings of 7th Annual ACM/IEEE Int’l. Conf. Mobile Comp. and Net, pp. 70–84 (2001) 10. Chen, B., et al.: SPAN: an Energy-efficient Coordination Algorithm for Topology Maintenance in Ad Hoc Wireless Networks. Proceedings of Wireless Networks 8(5), 481–494 (2002) 11. Smith, T.F., Waterman, M.S.: Identification of Common Molecular Subsequences. J. Mol. Biol. 147, 195–197 (1981)
Abstract. To promise low energy consumption for wireless sensor network routing, a number of routing protocols have been published. A novel routing protocol called BeamStar was proposed in [6]. In BeamStar, each base station uses a directional antenna with power control to assist locating sensors. It shifts the burden of network control and management from resource-limited sensors to resource-abundant and more sophisticate base stations, but still has three serious problems. The different size between the nearer and farther region, base station's scanning time, and a great quantity of inter-node communication all waste precious energy. In the cause of solving these problems, we present a routing protocol, namely Cluster-based BeamStar (CBS), for wireless sensor networks. CBS follows cluster-based structure to decrease the volume of internode communication and we have reformed the scanning manner in BeamStar. Our simulations show that CBS is scalable, low-cost, and energy efficient. Keywords: cluster, wireless sensor network, routing, directional antenna, power control.
1 Introduction The recent advances in wireless communication, MEMS technology, and light-weight Operation System have greatly reduced the size and cost of wireless sensor devices. Because having the features of low-cost, non-infrastructure and fault-tolerance, wireless sensor networks can operate in kinds of environments and be used in military, disaster, and surveillance usage [1]. As shown in Fig. 1, a wireless sensor network consists of a large number of sensors, randomly deployed in an interested region for detecting and monitoring tasking [2]. These tiny sensing devices have processing and wireless communication capabilities, which enable them to gather information from the environment and to generate and transmit report messages to base station. According to [3], the key networking challenges in wireless sensor network can divide into four broad categories: limiting radio operation, data management, geographic routing, and system monitoring maintenance. In terms of routing, because of severe energy constraints and unreliable network nodes, most routing protocols proposed for Mobile Ad Hoc networks are not suitable for wireless sensor networks, or A. Hua and S.-L. Chang (Eds.): ICA3PP 2009, LNCS 5574, pp. 129–136, 2009. © Springer-Verlag Berlin Heidelberg 2009
130
H.-L. Wang and Y.-Y. Chao
Fig. 1. Wireless sensor network architecture
cannot be used in wireless sensor networks without any modification [4]. In general, routing in wireless sensor network can be classified into three categories depending on the network structure: (1) (2) (3)
Flat-based: Direct diffusion [5] and BeamStar [6]. Hierarchical-based: LEACH [7] and Fault-tolerant clustering of wireless sensor networks [8]. Location-based: GAF [9] and SPAN [10].
We presented here is a revised BeamStar routing protocol called Cluster-based BeamStar (CBS). The cluster architecture such as LEACH protocol is used to reduce the volume of inter-node communication and the energy load will be distributed evenly. The rest of the paper is organized as follows. Section 2 reviews the related work. Section 3 presents our proposed method. Section 4 describes the performance evaluation. Section 5 concludes the paper.
2 Related Works We will simply introduce one protocol in each category mentioned above. A. BeamStar BeamStar is proposed in [6] which is a brand-new design concept. It is motivated by two observations: (1) Since management and control are indispensable, especially in large scale networks, BeamStar shifted computational intensive and energy consuming routing control overhead from sensor nodes to base stations. (2) Sensor nodes spend most of the energy on sensing and data forwarding. In BeamStar, the base station scans the network using a power controlled directional antenna. The relative angle of the sensor to the base station is determined by the direction of the base station transmission, while the distance of a sensor to the base station can be uniquely determined by the power level used for a received control message. As shown in Fig. 2, a sensor node’s ID is defined as the combination of its RN and SN. The node’s directionality of the last base station transmission is called Sector Number (SN), and the lowest power level information that it can receive from the base station is called Ring Number (RN). When the event occurs, each node simply compares their ID and the sequence number of the packet, then either drops or rebroadcasts the packet. The forwarding strategy is depending on the ID so there is no need for maintaining a routing table. With BeamStar, each data packet is relayed in an interleaved, loop-free
A Cluster-Based Data Routing for Wireless Sensor Networks
131
Fig. 2. BeamStar location discovery process [6]
mesh constrained in a sector towards the BS, making data delivery robust to sensor failures and transmission error. B. LEACH Heinzelman, et al. [7] introduced a hierarchical algorithm for wireless sensor networks, called Low Energy Adaptive Clustering Hierarchy (LEACH). LEACH is a clustering-based routing protocol that utilizes alternately select cluster-heads to distribute the energy load among the sensor nodes in the network evenly. The operation of LEACH is separated into two phases: the setup phase and the steady state phase. In the setup phase, each node has elected itself a cluster-head by a certainly formula (more interpretations will see in [7]) for the current round. All elected cluster-heads broadcast an advertisement message to rest of the nodes that they are the new clusterheads in current round. According to the signal strength of the advertisement message, the non- cluster-head nodes decide on the cluster to which they want to belong. After receiving all the messages from the nodes that would like to be included in the cluster, the cluster-head node creates a TDMA schedule for each node and broadcasts to all the nodes in the cluster. Once the clusters are created and the TDMA schedule is fixed, the stead state phase starts. In stead state phase, sensor nodes start sensing and forwarding packets to the cluster-heads. When the event occurs, the node who have sensed will send the sensing data to its cluster-head. After all the data has been received, the cluster-head performs data aggregation to reduce the redundant data and sends the report packets to the base station. After a certain time, the next round begins and each node returns to the setup phase. LEACH randomly selects a few sensor nodes as cluster-heads and rotates this role to evenly distribute the energy load among the sensors in the network. C. GAF Xu et al. [9] propose a location-based ad hoc routing protocol, which is called geographical adaptive fidelity (GAF). GAF conserves energy by identifying nodes that are equivalent from a routing perspective and then turning off unnecessary nodes, keeping a constant level of routing fidelity. The network area here is divided into fixed zones and forms a virtual grid. Each node uses GPS-indicated location information to association itself with a virtual grid, where all nodes in the same grid square are equivalent with respect to forwarding packets. For each grid, nodes coordinate with each other to elect a node awake for a certain period of time, and then the rest go
132
H.-L. Wang and Y.-Y. Chao
to sleep. This node is responsible for monitoring and forwarding data to the BS on behalf of the nodes in the same zone. Thus, by turning off the unnecessary nodes in the network, GAF economizes energy and keeps the level of routing fidelity simultaneously. In GAF, nodes are in one of three states: discovery, active, and sleep. The initial state of each node is discovery. When in state discovery, a node turns on its radio and exchanges discovery messages to find other nodes in the same grid. After discovery state, the node enters active state and takes charge of forwarding data. While an awake node can determine some other equivalent node will handle routing, it powers down its radio and changes state to sleep. The highest remaining energy node will always be selected stay in the active state by node ranking. In order to handle mobility, each node estimates the time it expects to leave its grid and includes this information in the discovery message. The sleeping nodes in the same grid adjust their sleeping time accordingly. Before the active nodes leaves, sleeping nodes wake up and one of them becomes active. This change does not change the node rank, but nodes will sleep for a shorter time. GAF is implemented both for non-mobility and mobility of nodes. It performs at least as well as a normal ad hoc routing protocol for packet loss and route latency, and yet it can substantially conserve energy, allowing network lifetime to increase in proportion to node density.
3 Proposed Routing Protocol 3.1 The Problems of BeamStar Though BeamStar protocol has reduced the burden of Sensor Node, three problems of this protocol was emerged. (1)Area of the farther region (i.e. the region with a larger RN) will larger, which means sensor nodes will not evenly distribute in each region. This drawback will lead to three differences. First, in accuracy, event locating becomes inaccurate with the increase of the area. Second, more sensor nodes have to broadcast for an event in a larger region. It will deplete much rare energy. Third, for the frequency of event occur, the larger area certainly more than the smaller one. That is to say, the farther regions have a higher probability to exhaust. (2)The number of base station’s scanning is enormous. Locating will consume a long period of time. For example, in fig. 2, base station needs to scan 15 times for a 5×3 area. It is inefficient in a dense events wireless sensor network. (3)The times of inter-node communication is huge. When the event occurs, source node generates a data packet and broadcasts it to its neighbors. The packet gradually broadcasts to base station and wastes much precious energy. 3.2 Cluster-Based BeamStar In order to improve the drawbacks mentioned above, a revised BeamStar routing protocol called Cluster-based BeamStar (CBS) is proposed. The protocol can divide into three phases:
A Cluster-Based Data Routing for Wireless Sensor Networks
133
：
Phase 1 Locating phase Similar to base station in BeamStar, base station in CBS performs powered controlled scan of the sensor network by using a directional antenna. Different from BeamStar, CBS makes change in the radius of region and the manner of scanning whole network. In order to keep the area of each region the same, CBS adjusts the radius R’ of region such as:
R′ =
iR
, ∀i = 1, 2,..., n
(1)
For example, the radius of region with ID{1, 1} is R, and the radius of region with ID{1, 2} will become 2 R, etc. We have also reformed the manner of scanning. As shown in Fig. 3(a) ~ (c). First, according to (1), the base station broadcasts the control message Ring Number to the whole network in order. Then, in Fig. 3(d) ~ (f), the base station will adjust its phase angle and broadcast Sector Number to the farthest region. All regions at the same phase angle can receive this control message and get their finally ID pair.
Fig. 3. An illustration of locating phase
：
Phase 2 Cluster building up phase After the locating phase, each region is regarded as a cluster such that the sensor nodes with the same ID are considered in the same cluster. The cluster members begin to exchange their remaining energy. Node in cluster with the largest remaining energy will be elected as the cluster-head. After a certain time, cluster-head broadcasts an advertisement message to its subordinates to declare it is the cluster-head in this round.
：
Phase3 Data transferring phase When the event occurs, the nodes which have sensed will broadcast the data around their own cluster. After all the data has been received, the cluster-head performs data aggregation to compress the data into a single signal. Only inter-cluster transmission
134
H.-L. Wang and Y.-Y. Chao
is needed in the following communication. The cluster-node will broadcast the control message to all the member nodes when its energy drops below a certain threshold. The nodes who have received this message will return their remaining energy to cluster-head. Then the most robust node was noticed to act as a cluster-head from the original cluster-head. An advertisement message was announced by the new clusterhead to its cluster member.
4 Comparison We have compared the performance of CBS to BeamStar in three aspects. 4.1 Difference of Sensor Nodes in Regions In our simulation, 100 nodes are randomly deployed in an area of radius 100. The length of RN is 20, 40, 60, 80, and 100 in BeamStar; r, 2 r, 3 r, 2r, and 5 r in ours which r 100/ . Base station is located in the center of the area. Fig. 4 shows the results of sensor nodes scatter in the deployed area. BeamStar has an obvious increase with the larger RN, and CBS keep a steady quantity in each region. It means nodes distributing are evenly. Furthermore, the regions which were distant from the base station will not increase their area can still keep the accuracy.
＝
4.2 Base Station Scanning Time As shown in Fig. 5, we set the maximum SN equals 20 and maximum RN is 10, i.e. 200 regions will be divided. Assume each time base station to scan a region is 1 second. BeamStar spends 200 seconds to scan the whole network, but CBS only needs 30 seconds. CBS shortens the scanning time substantially. 4.3 Data Transferring The Fig. 6 shows the different data transferring manners of BeamStar and CBS. Suppose that an event is detected by a sensor node in a region with ID{2, 3}. In BeamStar, this sensor node will generate a data packet and broadcast it to its neighbors. At 40
Number of sensor nodes
35
CBS
30
BeamStar
25 20 15 10 5 0 1
2
RN 3
4
5
Fig. 4. Node distributing in the region
A Cluster-Based Data Routing for Wireless Sensor Networks
135
200
Scanning time (sec.)
180 160
CBS
140
BeamStar
120 100 80 60 40 20 0 20
40
60
80
100 120 140 160 180 200
Number of region
Fig. 5. Scanning frequency of two protocols
Fig. 6. Data transferring of BeamStar (a) and CBS (b)
the four boundaries of this region, only the region with ID{2, 2} will further relay this packet. By this method, the packet will be eventually delivered to the base station. CBS just need to broadcast in the region with ID{2, 3}. After cluster-head in this region has received the data packet. Only inter-cluster transmission is needed in the following communication. As seen in TABLE I, inheriting from BeamStar, CBS performs better than LEACH in scalability. By contrast, CBS reduces the volume of inter-node communication than BeamStar and the energy load among the all the sensor nodes has evenly distributed. When the event occurs frequently, the cluster-head selection will more satisfied because it was depending on the remaining energy instead of the time interval. The cluster-heads in CBS also perform data aggregation to reduce the redundant. Table 1. Comparison with three protocols
Scalability Power usage Inter-node communication Cluster-head selection Data aggregation
BeamStar Good Low More
LEACH Low Good Less
CBS Good Good Less
×
Time-based
Energy-based
No
Yes
Yes
136
H.-L. Wang and Y.-Y. Chao
5 Conclusion Although BeamStar protocol shifted computational intensive and energy consuming routing control overhead from sensor nodes to base stations. It still has three lethality shortcomings mentioned in Section 3. We propose a new routing protocol named CBS to overcome there drawbacks. It is unnecessary to store a routing table and equip with additional hardware (such as GPS). CBS selects cluster-head alternately which also evenly distribute the energy load among the sensor nodes in the network. We also improve the scanning manner to decrease the scanning frequency and shorten the locating time. It will efficiently prolong the lifetime of the wireless sensor network. The proposed routing protocol is scalable, low-cost and energy efficient for wireless sensor networks.
References 1. Xu, N.: A Survey of Sensor Network Applications, http://enl.usc.edu/~ningxu/ 2. Al-Karaki, J.N., Kamal, A.E.: Routing Techniques in Wireless Sensor Network: A Survey. Proceedings of IEEE Wireless Communications 11(6), 6–28 (2004) 3. Deepak, G., Alberto, C., Yan, Y., Wei, Y., Jerry, Z., Deborah, E.: Networking Issues in Sensor Networks. Journal of Parallel and Distributed Computing 64(7), 799–814 (2004) 4. Tian, D., Georganas, N.D.: Energy efficient routing with guaranteed delivery in wireless sensor networks. Proceedings of Wireless Communications and Networking 3, 1923–1929 (2003) 5. Intanagonwiwat, C., Govindan, R., Estrin, D.: Directed diffusion a scalable and robust communication paradigm for sensor networks. In: Proceedings of the 6th annual international conference on Mobile computing and networking (August 2000) 6. Shiwen, M., Hou, Y.T.: BeamStar: A New Low-cost Data Routing Protocol for Wireless Sensor Networks. In: Proceedings of Global Telecommunications Conference, 2004, vol. 5, pp. 2919–2924 (2004) 7. Wendi, R.H., Anantha, C., Hari, B.: Energy-efficient communication protocol for wireless microsensor networks. In: Proceedings of the 33rd Annual Hawaii International Conference, vol. 2, p. 10 (2000) 8. Gupta, G., Younis, M.: Fault-tolerant clustering of wireless sensor networks. In: Proceedings of Wireless Communications and Networking, 2003, vol. 3, pp. 1579–1584 (2003) 9. Xu, Y., Heidemann, J., Estrin, D.: Geography-informed Energy Conservation for Ad-hoc Routing. In: Proceedings of 7th Annual ACM/IEEE Int’l. Conf. Mobile Comp. and Net, pp. 70–84 (2001) 10. Chen, B., et al.: SPAN: an Energy-efficient Coordination Algorithm for Topology Maintenance in Ad Hoc Wireless Networks. Proceedings of Wireless Networks 8(5), 481–494 (2002) 11. Smith, T.F., Waterman, M.S.: Identification of Common Molecular Subsequences. J. Mol. Biol. 147, 195–197 (1981)
