KR101024038B1 - Cluster configuration method of cluster sensor network and sensor network to which the method is applied - Google Patents

Abstract

The present invention relates to a configuration method for controlling the size of a cluster in a sensor network of a cluster structure, by maximizing the life of the sensor network by adjusting the size of the cluster to equalize the energy consumption of the cluster heads included in each cluster of the sensor network. It is about how to adjust the cluster.

In a method for adjusting a cluster size of a sensor network according to the present invention, the method for adjusting a cluster of a sensor network having a cluster structure includes determining a distance between each sensor node constituting the cluster and a sink node of the sensor network. And determining the size of each cluster in consideration of the energy consumption of the sensor node and the characteristics of the sensor network according to the determined distance.

According to the cluster sizing method as described above, by adjusting the size of the cluster so that the energy consumption of the cluster head constituting each cluster is the same, it is possible to maximize the life of the sensor network.

Cluster, Sensor Network, Zigbee, CSMA / CA

Description

A cluster configuration method of a cluster sensor network and a sensor network to which the method is applied {Method of configuring clustered wireless sensor networks and clustered wireless sensor networks thereby}

The present invention relates to a method for adjusting the size of a cluster in a sensor network of a cluster structure, and to a method for configuring a cluster to maximize the life of the sensor network by adjusting the size of the cluster according to the energy consumption of each node of the sensor network.

Recently, with the rapid development of wireless communication technology and communication equipment, the ubiquitous networking era has arrived. In the ubiquitous networking, a sensor network using sensors with low power consumption and low production cost has been in the spotlight.

Among the sensor networks, ZigBee has a short communication distance but has an advantage of easily extending communication coverage through a clustering structure, and thus has been widely applied to various fields such as home network, home appliance control, and industrial automation.

ZigBee's sensor networks do not require sophisticated placement algorithms due to the inexpensive sensor nodes, and they can change their topology on their own. Each sensor node does not have an energy supply line and operates only on batteries. Since it is not easy to exchange or recharge the battery of each sensor node, energy consumption problem is the biggest problem of sensor network.

See Seema Bandyopadhyay, Edward J. Coyle, "Minimizing communication costs in hierarchically-clustered networks of wireless sensors", Computer Networks, Vol. 44, pp. 1-16, 15 Jan. 2004 In ", we calculated the number of cluster heads that minimize the energy consumption, assuming that the energy consumption of the cluster network is proportional to the communication distance.

In "W. Heinzelman, A. Chandrakasan, H. Balakrishnan," Energy-Efficient Communication Protocol for Wireless Microsensor Networks, "Proc. Of the 33rd Annual Hawaii International Conference on System Sciences, pp. 3005-3014, Jan 2000. In order to reduce the load, the proposed method was proposed by repeatedly changing the role of the cluster head between cluster members.

In "W. Heinzelman, A. Chandrakasan, H. Balakrishnan," An application specific protocol architecture for Wireless Microsensor Networks ", IEEE Trans. On Wireless Communications, Vol. 1, no. 4, pp. 660-670, 2002. The cluster members exchanged their remaining energy information, and proposed a method of selecting the cluster member with the most energy as the cluster head.

However, the above methods have the disadvantage that additional energy is consumed by causing another computing load on low-complex sensor nodes, and a bottleneck caused by a larger energy consumption of the cluster head near the sink node than other cluster heads. ), The problem of shortening the lifespan of the entire sensor network still occurred.

The present invention is to solve the above problems by adjusting the size of the cluster according to the distance between the sync node and the sensor node constituting the sensor network so that the energy consumption of the cluster head included in each cluster to be the same, the life of the sensor network It is to provide a sensor network consisting of a cluster control method and a scaled cluster of sensor networks that can be maximized.

In order to achieve the above object, the method of controlling a cluster of a sensor network of the present invention is a method of adjusting a cluster of a sensor network of a cluster structure, the distance between each sensor node constituting the cluster and the sink node of the sensor network. And determining the size of each cluster in consideration of the energy consumption of the sensor node and the characteristics of the sensor network according to the determined distance.

In addition, the cluster control method of the sensor network of the present invention in order to achieve the above object in the sensor network of the cluster structure consisting of a sensor node including a cluster head (CH) and the cluster member (CM), comprising the cluster Determining a distance between each sensor node and a sink node of the sensor network, and determining the size of each cluster such that energy consumption of the cluster head included in each cluster is equal according to the determined distance.

In addition, in order to achieve the above object, the sensor network of the present invention consists of a cluster whose size is adjusted by the cluster adjusting method of the sensor network according to the present invention.

The present invention having the above configuration has the effect of maximizing the life of the sensor network by adjusting the size of each cluster head according to the distance from the sink node.

In addition, by adjusting the size of the cluster so that the energy consumption of the cluster head included in each cluster is the same, it is possible to prevent the life of the specific cluster head is shortened quickly.

In addition, the cluster heads have the effect of maximizing the life of the sensor network by equalizing the power consumption per node at any position.

Hereinafter, an example according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a sensor network having a cluster structure. The sensor network of the cluster structure is a proposed network for efficiently managing a plurality of sensor nodes, and has an advantage of minimizing energy consumption between sensor nodes. In a clustered network, a cluster consists of a cluster head and cluster members, and data of cluster members are collected and transmitted to the sink node, thereby reducing energy consumption.

The sensor network of the cluster structure shown in FIG. 1 includes one Personal Area Network (PAN) coordinator, which is responsible for formation and maintenance of a PAN. In addition to the PAN coordinator, the sensor network includes a number of clusters, each consisting of a cluster head and cluster members.

Each cluster is provided with one cluster head CLH to relay communication between the sensor node and the PAN. The cluster member that needs to transmit the data transmits the data to the cluster head, and the cluster head collects the data of the cluster members and transmits the data to the PAN coordinator in a multi-hop manner.

In this process, the cluster head near the data acquisition node (assuming a PAN coordinator in the figure) is a bottleneck with a shorter lifespan than the cluster head farther from the destination data collection node because the battery consumes more energy due to high energy consumption due to frequent data transmission. (bottleneck) phenomenon occurs.

Sensor nodes included in a cluster communicate using multiple channels, and each cluster uses multiple channels to communicate within a cluster and with nodes of another cluster. For example, a Zigbee terminal node is a communications device built on the IEEE 802.15.4 standard, operating in 868-868.6 MHz in Europe and 902-928 MHz in North America and operating in the Industrial, Scientific and Medical (ISM) band globally. Do it. There are 16 channels in the 2.4 GHz band.

Each cluster uses any one of channels 11 to 24, and the cluster head communicates through the multi-hop between cluster heads using channels 25 or 26 for the existing channel. The reason for determining the communication channel between the cluster heads as 25 or 26 is to avoid interference with the WLAN operation channel.

Sensor nodes constituting the sensor network are usually distributed randomly in the field because the cost of the terminals constituting the sensor node is low, so a sophisticated placement algorithm is not required.

The present invention is a cluster control method for properly adjusting the size of the cluster according to the distance between the sink node and the sensor node in order to maximize the life of the sensor network with reference to Figure 2 ZigBee (node) node as an example hop number Accordingly, a cluster adjustment method for differently determining a probability of becoming a cluster head and adaptively adjusting the size of the cluster through the cluster head will be described below.

After the Zigbee terminal nodes are randomly distributed, the sink node broadcasts its hop value 0 in order from channels 11 to 24. ZigBee terminal nodes within the sync node's transmission distance r, i.e., 1-hop away, receive the message. The terminal nodes receiving the message set the hop value plus 1 in the message as their hop value and broadcast it again. This method is called a flooding technique, and Zigbee terminal nodes have their respective hop values by the flooding technique for a predetermined time (step S200).

The hop value setting method by flooding is used as a method for estimating the distance between the sink node and each Zigbee terminal node, that is, the sensor node, but is not limited thereto. In addition, if the location of each node is known, distance information can be found out immediately.

The zigbee terminal nodes that make up the cluster by the flooding technique have their respective hop values, and when the hop value of each zigbee terminal node is determined, the cluster head is set according to the probability value p to be the cluster head differentially set by the cluster adjustment method of the present invention. Or become a cluster member.

In the cluster adjustment method of the present invention, the cluster size is determined using the probability value p to be the cluster head, and p is differentially determined according to a distance from the sink node, for example, a hop value. As a result, the cluster size varies according to the distance, and the energy consumption of the cluster head constituting the network is the same, thereby maximizing the life of the sensor network. That is, the bottleneck phenomenon is effectively prevented, so that the life of the cluster head near the sink node is the same as the life of the cluster head located at the outer edge of the sensor network.

The probability value p to be the cluster head in each cluster is determined in consideration of the characteristics of the sink node and the sensor network. Hereinafter, an example of determining p of each cluster will be described in detail.

First, as described above, the present invention adaptively determines the size of the cluster according to the distance to the sink node. In other words, as the distance from the sink node gets closer, the data from the outside is crowded, so the cluster is set smaller to balance energy consumption.

Also, if p _N is defined as the probability of being the cluster head in the cluster at the Nth hop from the sink node, the following conditions must be met:

η : 지그비 단말노드의 밀도

η: Density of Zigbee terminal node

개가 된다) 중에 적어도 하나의 클러스터 헤드가 있어야 하는 조건이 표현된 것이고(

≥(1개 클러스터 헤드/

개 노드들)) 그 앞의 부등호 들은, 싱크노드로부터 가까울수록 클러스터 헤드가 될 확률이 높아야 한다는 조건을 나타낸 것이다. (N이 작을수록 싱크노드에 가깝다.)Here, the last inequality sign is the transmission distance r of the Zigbee terminal node (the number of nodes within the transmission distance is η since the terminal node density is η).

The condition that there must be at least one cluster head)

≥ (1 cluster head /

The inequalities in front of them represent a condition that the closer to the sink node, the higher the probability of becoming a cluster head. (The smaller N, the closer to the sink node.)

However, even if the above Equation 1 is satisfied, there may be a case where there are two Zigbee terminal nodes in one cluster and none in the other cluster. Therefore, the cluster size itself is used to be smaller than the transmission distance r of the Zigbee terminal node.

Since the Zigbee terminal node is randomly distributed, it satisfies the property of Voronoi tessellation, and the cluster size can be defined as the radius ρ of the circle centering on the cluster head as shown in FIG. 3, where the cluster size ρ is the Zigbee terminal node transmission distance r. The greater probability must be less than the very small value σ. In summary, the following equation holds.

는 클러스터 헤드의 밀도를 나타낸다. 이 관계는 클러스터 사이즈가 전송거리보다 클 확률을 (Pr{ρ>r})이 노드의 밀도(η), 클러스터 헤드가 될 확률(p) 그리고 전송거리의 제곱(r²)에 반비례하는 것을 나타내는 Voronoi의 성질로서 잘 알려져 있다. Where the right side protest of the equality relationship

Represents the density of the cluster head. This relationship indicates that (Pr {ρ> r}) is inversely proportional to the node's density (η), cluster head's probability (p), and the square of the transmission distance (r ² ). It is well known as the nature of Voronoi.

Equation 1 and Equation 2 The size of the cluster may be adjusted by determining p _N as the larger value of the probabilities calculated through the two equations. However, the number of cluster members included in the cluster and the throughput of the entire system are deeply related. As the number of cluster members increases, the probability of collision between each other increases and energy consumption increases due to retransmission when collision occurs. Therefore, the following equation can be further considered.

k is the number of cluster members maximizing normalized throughput. If there are cluster members smaller than k or larger than k in a cluster, the channel is inefficiently used or collisions increase, resulting in lowered normalized throughput. Therefore, as shown in Equation 4 below, Equation 1, Equation 2, and Equation 3 can be selected.

Determining the probability of becoming a cluster head in this way has the effect of scaling the cluster more accurately.

The p value of each cluster according to the hop value can be determined using the p _N determined as described above. The pMA of the cluster having different hop values is modeled by modeling the data traffic in the superframe structure having the active and inactive intervals through the protocol of CSMA / CA. The value can be determined. Figure 4 shows the renewal period of the superframe structure through which the data transmission amount of each cluster head is determined. Of course, this is an example, and may use characteristics of sensor networks other than throughput or throughput of other protocols.

After determining the value of p _{N, the} probabilities to be the remaining cluster heads must be sequentially obtained. The data communication is modeled to determine the probabilities of becoming each cluster head so that the energy consumption of each cluster head is equal.

As an example, assuming that a certain amount of data traffic is generated in one second, the frame structure is set to be a renewal period in which idle and transmission are infinitely repeated. In this case, if the probability that the cluster members transmit is P _tr , P _tr can calculate the probability that the cluster member successfully transmits the frame by using the renewal period and the number of slots. By using this, the number of frames to be transmitted to the cluster head can be obtained, and the same can be transmitted to the sink node through multi-hop. Therefore, the amount of frames transmitted by each cluster head for one second can be calculated using the hop number. Using the condition that energy consumption is equal and substituting the previously obtained p _N value, p _N-1 value can be obtained. The value of p _{N-1 is} again determined by the value of p _N-2 , and thus the probability of becoming the cluster head of each cluster can be obtained in this way.

As described above, the probability value of the cluster head according to the number of hops is determined, and the size of the cluster is adjusted through this (step S220).

As described above, the size of the cluster can be adjusted according to the distance from the sink node, thereby preventing bottlenecks and maximizing the lifetime of the entire network.

The cluster control method according to the present invention can adjust the size of the cluster so that the energy consumption of the cluster head is equal. First, the energy consumption of the cluster constituting the sensor network is calculated through modeling and the like. After establishing the relationship with the size in consideration of the characteristics of the sensor network, the size of the cluster can be determined so that the energy consumption of the cluster head of each cluster is equal using the established relationship.

Specific methods for this may be a method of adjusting the size of the cluster under the condition that energy consumption is equal by determining p according to the hop value described above, but is not limited thereto.

5A and 5B show a network of a conventional cluster structure in which the number of hops is displayed and a network of a cluster structure according to the present invention, in which one cluster head 32 exists for each cluster, and the size of the cluster 30 is generally central. It can be seen that the larger the distance away from the sink node. That is, it can be seen that the size of the cluster is inversely proportional to the distance from the sink node.

6 and 7 are graphs comparing the transmission or total power consumption according to the number of hops with the sink node, respectively.

6 and 7, the transmission and total power consumption of the conventional cluster head decreases as the number of hops increases, so that the energy consumption of the cluster head having a hop number of 1 differs by almost 750 μJ compared to the cluster head having a hop number of 10. Able to know. In contrast, the present invention can be seen that only the cluster head having a hop number of 1 is slightly higher in energy consumption and the energy consumption is almost the same.

Therefore, it can be seen that the bottleneck does not occur when the size of the cluster is adjusted by the cluster adjusting method according to the present invention. Of course, this has the effect of maximizing the lifetime of the entire network.

The size is determined first by considering the density and the size, and the sensors are arranged by using artillery or airplane.

The configuration of the present invention has been described above by specific embodiments such as specific components and limited embodiments and drawings, but it is provided to help general understanding of the present invention. The present invention is not limited to the above embodiments, and those skilled in the art may add various forms of modifications and variations from this description.

Accordingly, the spirit of the present invention should not be limited to the described embodiments, and all subjects having equivalent or equivalent modifications to the claims as well as the following claims are intended to be included in the scope of the present invention. will be.

1 is a diagram illustrating a sensor network having a cluster structure.

2 is a flowchart showing step by step an embodiment of a method for adjusting a cluster according to the present invention;

3 is a diagram showing the size of a cluster shown through a circle centered on the cluster head in a cluster formed by Voronoi tessellation,

4 is a diagram illustrating a renewal period of a single cluster operation for calculating energy consumption of a cluster head,

5A and 5B are diagrams illustrating a network of a conventional cluster structure in which hop numbers from a sink node are displayed, and a cluster structure according to the present invention, respectively.

6 is a graph comparing the transmission power consumption of the cluster head according to the number of hops with the sink node as compared with the conventional art.

7 is a graph comparing the total power consumption of the cluster head according to the number of hops with the sink node.

<Explanation of symbols for the main parts of the drawings>

Cluster: 30 Cluster Head: 32

Claims (21)

In the cluster control method of the sensor network for adjusting the cluster of the sensor network of the cluster structure, Each cluster is composed of a cluster head (CH) and a cluster member (CM), Determining, at the cluster head (CH), a distance between each sensor node constituting the cluster and the sink node of the sensor network; And determining the size of each cluster in consideration of the energy consumption of the sensor node and the characteristics of the sensor network according to the determined distance. The size of each cluster is determined using the probability of being a cluster head (CH) among sensor nodes constituting each cluster, Placement of the sensor network is made according to a Poisson process, The probability of becoming the cluster head (CH) is a cluster control method of the sensor network, characterized in that determined using the nature of the Voronoi Tessellation of the cluster among the characteristics of the sensor network. The method of claim 1, The sensor network is a multi-hop wireless network, And a distance between the sink node and each sensor node is determined using a hop value. 3. The method of claim 2, The hop value of each sensor node is determined by using a floating technique. delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete

Images