KR100959808B1 - Method of constituting cluster by sensor network - Google Patents

Abstract

The present invention relates to a method for forming a cluster of sensor networks using fuzzy logic, and more particularly, to a method for forming a cluster of sensor networks using fuzzy logic that selects a cluster head based on a distance from a sensor node that is close to the remaining energy of the sensor node. It is about. The present invention can provide a method for forming a cluster of sensor networks using fuzzy logic that can efficiently extend the life of a sensor network by selecting a sensor node having a large amount of energy and having a close distance to an adjacent sensor node as a cluster head.

Wireless, sensor network, cluster head, energy, distance

Description

Cluster formation method of sensor network {METHOD OF CONSTITUTING CLUSTER BY SENSOR NETWORK}

The present invention relates to a cluster formation method of a sensor network using fuzzy logic, and more particularly, to a cluster formation method of a sensor network using fuzzy logic that selects a cluster head based on energy and distance of a sensor node.

Sensor networks have computational and storage capabilities that can be used in areas such as remote monitoring of agriculture, transportation, inventory tracking, seismic monitoring, and more. Such a sensor network basically consists of a plurality of sensor nodes with limited energy and a sink that collects information from the plurality of sensor nodes. In particular, the energy problem is one of the biggest considerations in designing the sensor network because the sensor node is almost impossible to replace the battery once it is placed in the target area. Therefore, it is essential to design energy-efficient sensor networks in hardware and operating systems, as well as communication and routing protocols.

In order to solve this problem, the cluster routing scheme has been widely used as one of the methods for achieving energy efficiency in the routing protocol. Cluster routing is a mechanism that divides the sensor network into sub-networks called clusters, where each cluster elects a cluster head so that the cluster head collects, merges, and delivers the information within the cluster to the sink. The cluster routing technique greatly reduces the overhead incurred by all nodes communicating with the sink because only the cluster head merges and forwards the information to the sink. However, the cluster routing technique has a disadvantage in that energy consumption is concentrated at the cluster head.

For this reason, researches on efficient cluster head selection techniques are being actively conducted. A representative technique is LEACH (Low Energy Adaptive Cluster Hierarchy), which uses a local clustering method based on a probability model. In the case of LEACH, the cluster head is evenly performed by applying a probabilistic model to distribute the energy consumption concentrated in the cluster head.

However, LEACH relies only on probabilities for election of cluster heads, so that cluster heads may be adjacent to each other, resulting in inefficient clusters. In addition, LEACH does not consider the energy remaining situation of the sensor node, so that the sensor node with relatively low energy is elected as the cluster head, which may shorten the lifespan of the sensor network. In addition, LEACH has a problem in that when the distance between the elected cluster head and the neighboring sensor nodes is far, the neighboring sensor nodes consume a lot of energy to transmit information to the cluster head, thereby shortening the lifespan of the sensor network.

An object of the present invention is to provide a method for forming a cluster of sensor networks using fuzzy logic that can increase the lifetime.

In order to achieve the above object, the present invention provides a sensor network in which a plurality of sensor nodes are configured in at least one cluster in a sensor field, the area distance being the sum of energy remaining of the plurality of sensor nodes and the distance of the plurality of sensor nodes. A method of forming a cluster of sensor networks using fuzzy logic, the method comprising: selecting a cluster head based on the step, and forming a cluster with the cluster head and sensor nodes adjacent to the cluster head.

The area distance is the sum of the distances between the sensor node satisfying the cluster head support condition and the sensor node existing within a predetermined distance,

이고,The predetermined distance is

ego,

는

이며,remind

Is

Is,

는

이고,remind

Is

ego,

는

이며,remind

Is

Is,

는 상기 센서 필드의 면적이고,remind

Is the area of the sensor field,

은 상기 센서 필드 내의 센서 노드 개수이다.remind

Is the number of sensor nodes in the sensor field.

The selecting of the cluster head based on the remaining energy of the plurality of sensor nodes and the regional distance that is the sum of the distances of the plurality of sensor nodes may include supporting the cluster head by at least one sensor node of the plurality of sensor nodes. And selecting a cluster head among sensor nodes supporting the cluster head.

The at least one sensor node of the plurality of sensor nodes supporting the cluster head may include calculating a cluster head support condition of the plurality of sensor nodes, selecting a cluster head of a sensor node that satisfies the cluster head support condition. Calculating an opportunity value and comparing the cluster head election opportunity value of one sensor node that satisfies the cluster head support condition with a cluster head election opportunity value of another sensor node proximate to the one sensor node.

을 생성하는 단계와, 상기 랜덤 값

와 클러스터 헤드를 지원한 센서 노드의 비율과 센서 네트워크에서 요구되는 최적의 클러스터 헤드의 비율을 곱한

를 비교하는 단계와, 상기 랜덤 값

가 상기

보다 작을 때 클러스터 헤드를 지원하는 단계를 포함할 수 있다. 하지만 상기 랜덤 값

가 상기

보다 클 때 클러스터 헤드 메시지를 수신 대기하는 상태로 전환하는 단계를 포함할 수 있다. 이때, 상기

는

이다.Computing the cluster head support condition of the plurality of sensor nodes, the random value of each of the plurality of sensor nodes

Generating a random value;

Multiplied by the ratio of sensor nodes supporting the cluster head to the optimal cluster head ratio required by the sensor network.

Comparing and random values

Remind

Supporting the cluster head when smaller. But the random value above

Remind

Transitioning to a state where the cluster head message is received when larger. At this time, the

Is

to be.

는

이며,In addition,

Is

Is,

는

이고,remind

Is

ego,

는

이며,remind

Is

Is,

는

이고,remind

Is

ego,

는 상기 센서 필드의 면적이며,remind

Is the area of the sensor field,

은 상기 센서 필드 내의 센서 노드 개수이고,remind

Is the number of sensor nodes in the sensor field,

는 1이상이고

이하이다. 이때, 상기

는 2.5인 것이 가장 효과적이다.remind

Is 1 or more

It is as follows. At this time, the

2.5 is most effective.

The calculating of the cluster head election opportunity value of the sensor node satisfying the cluster head support condition may include calculating the cluster head election opportunity value of the sensor node satisfying the cluster head support condition using a non-fuzzy technique. Include.

The cluster head election opportunity value is calculated by a fuzzy if-then rule, and the fuzzy if-then rule divides the energy remaining of the sensor node into low, medium, and high, and stops the local distance of the sensor node. And close to each other, and the cluster head election opportunities are classified into very small, small, somewhat small, moderate small, normal, moderate large, somewhat large, large, and very large based on the energy remaining amount and the local distance.

The cluster head election opportunity value is the area of the graph according to the fuzzy if-then rule,

이고,The area of the graph according to the fuzzy if-then rule is

ego,

는 퍼지 if-then 규칙에 따른 최저값이 0이며 최대값이 100인 클러스터 헤드 선출 기회값이고, 상기

는 퍼지 if-then 규칙 결과의 멤버쉽 함수이다.remind

Is a cluster head election opportunity value having a minimum value of 0 and a maximum value of 100 according to the fuzzy if-then rule.

Is the membership function of the fuzzy if-then rule result.

Comparing the cluster head election opportunity value of one sensor node satisfying the cluster head support condition with the cluster head election opportunity value of another sensor node proximate to the one sensor node, the cluster head support message is transmitted to the other sensor node in proximity. And storing the ID of the one sensor node in cluster head information of one sensor node satisfying the cluster head support condition and storing the cluster head election opportunity value of the one sensor node in cluster head election opportunity value information. Receiving a cluster head support message of another sensor node adjacent to the one sensor node; and when the cluster head election opportunity value of the one sensor node is greater than the cluster head election opportunity value of the other sensor node. The cluster head information of the node and the head election Maintaining the opportunity value information.

However, the present invention is not limited thereto, and when the cluster head election opportunity value of the one sensor node is smaller than the cluster head election opportunity value of the other sensor node, the cluster head information and the head election opportunity value information of the one sensor node are read. Changing to another sensor node.

The selecting of the cluster head among the sensor nodes supporting the cluster head may include identifying a sensor node having the largest cluster head election opportunity value among the sensor nodes satisfying the cluster head support condition, and the cluster head election opportunity value. Sending a cluster head message for this largest sensor node to an adjacent sensor node.

The present invention can provide a method for forming a cluster of sensor networks using fuzzy logic that can efficiently extend the life of a sensor network by selecting a sensor node having a large amount of energy and having a close distance to an adjacent sensor node as a cluster head.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

However, the present invention is not limited to the embodiments disclosed below, but will be implemented in various forms, and only the embodiments are intended to complete the disclosure of the present invention, and to those skilled in the art to fully understand the scope of the invention. It is provided to inform you. Like reference numerals in the drawings refer to like elements.

1 is a configuration diagram of a sensor network according to the present invention, Figure 2 is a conceptual diagram showing the operation of the sensor network according to the time according to the present invention, Figure 3 shows an example of calculating the local distance of the sensor network according to the present invention 4 is a conceptual diagram illustrating a algorithm of a cluster configuration step of a sensor network according to the present invention.

As shown in FIG. 1, the sensor network according to the present invention includes a plurality of sensor nodes N provided in the sensor field SF, and a cluster head CH communicating with the plurality of sensor nodes N. . It also includes a sink SK in communication with the cluster head CH. In this case, a plurality of sensor nodes N are distributed in an area where data collection is required, that is, a sensor field SF, and a sensor node satisfying a specific condition among the plurality of sensor nodes N is selected as the cluster head CH. do. In addition, the selected cluster head CH collects information from neighboring sensor nodes N and transmits the information to the sink SK. The sink SK stores the information collected from the sensor field SF and transmits the information to the user. to provide. In this case, the sink SK may transmit the collected information to the user through the Internet.

The sensor node N is a basic element constituting a sensor network for observing a physical phenomenon in the sensor field SF and transmitting information about it. The sensor node N may include smart dust or the like.

The cluster head CH is a sensor node that satisfies a specific condition among the sensor nodes N, and transmits information received from the sensor node to the sink SK. In this case, according to the present invention, the remaining sensor energy among the sensor nodes N is higher than that of other sensor nodes, and the sensor node having the closest distance between the sensor nodes N is selected as the cluster head CH.

The sink SK receives the information collected by the sensor node N through the cluster head CH and delivers it to the user. In this case, the sink SK may transmit the information of the sink SK to the user by means such as, for example, the Internet.

In the sensor network according to the present invention having such a configuration, as shown in FIG. 2, a round consisting of a cluster configuration step and a data transmission step is repeated at a predetermined period. At this time, in the cluster forming step, a new cluster head CH is selected to form a cluster, thereby dispersing energy consumption concentrated on the cluster head CH. In addition, the data transmission step collects the data based on the formed cluster and transmits the data to the sink (SK). At this time, as shown in FIG. 4, the cluster configuration step includes a cluster head support step, a cluster head selection step, and a cluster formation step.

The cluster head support step includes calculating cluster head support conditions of a plurality of sensor nodes (S1), calculating cluster head election opportunity values of sensor nodes satisfying the cluster head support conditions (S2), and cluster head support. Comparing the cluster head election opportunity value of one sensor node that satisfies the condition with the cluster head election opportunity value of another sensor node in close proximity to the one sensor node (S4, S5, S6, S7), the cluster head support condition Comparing the cluster head election opportunity value of one sensor node satisfying with the cluster head election opportunity value of another sensor node close to the one sensor node (S4, S5, S6, S7) supports the cluster head to another sensor node in proximity. Transmitting the message (S4) and the cluster head information of one sensor node that satisfies the cluster head support condition; Storing an ID of the sensor node and storing the cluster head election opportunity value of the one sensor node in cluster head election opportunity value information (S5), and receiving a cluster head support message of another sensor node in proximity to one sensor node (S6); And maintaining the cluster head information and the head election opportunity value information of the one sensor node when the cluster head election opportunity value of one sensor node is greater than the cluster head election opportunity value of the other sensor node (S7). Include.

을 발생하게 되며,

을 P_opt와 비교하여 특정 센서 노드만이 클러스터 헤드(CH)를 지원하도록 한다. 이때, P_opt는 다음과 같은 수식(수식 1)으로 정의된다.The sink SK is started when it sends a cluster configuration beacon message to all sensor nodes N. At this time, if all sensor nodes support the cluster head CH, the sensor network speed decreases and the energy of the sensor node N is reduced, so that only the sensor nodes satisfying a specific condition support the cluster head CH. That is, each sensor node (N) is a random number between 0 and 1 when the cluster configuration beacon message is received.

Will occur,

Is compared with P _{opt so that} only specific sensor nodes support the cluster head (CH). At this time, P _opt is defined by the following equation (Equation 1).

[Equation 1]

는 클러스터 헤드(CH)를 지원한 센서 노드의 비율이며, P는 센서 네트워크에 요구되는 최적의 클러스터 헤드(CH)의 비율이다. 만약, 센서 노드(N)에서 발생된 랜덤한 수인

이 P_opt 값보다 크다면 해당 센서 노드는 클러스터 헤드(CH) 지원을 포기하고 클러스터 형성 단계에서 이후 선출될 클러스터 헤드(CH)로부터의 클러스터 헤드 메시지를 수신 대기한다. 물론, 센서 노드에서 발생된 랜덤한 수인

이 P_opt 값보다 작다면 해당 센서 노드의 클러스터 헤드 선출 기회값을 계산한다.At this time,

Is the ratio of the sensor nodes supporting the cluster head CH, and P is the ratio of the optimal cluster head CH required for the sensor network. If, the random number generated at the sensor node (N)

If it is larger than this P _opt value, the sensor node abandons the cluster head CH support and listens for a cluster head message from the cluster head CH to be elected later in the cluster formation step. Of course, the random number generated at the sensor node

If this value is smaller than the P _opt value, the cluster head election opportunity value of the sensor node is calculated.

의 범위는 1≤

≤

로 정의된다.In addition, the minimum value of P _opt for determining the ratio of the cluster head support sensor node must be at least greater than P indicating the optimal cluster head ratio, and the maximum value of P _opt exceeds 1. therefore,

The range of 1≤

≤

Is defined as

가 2.5일 때 평균적으로 선출된 클러스터 헤드의 수가 최적의 클러스터 수와 동일해져 네트워크의 수명을 극대화 할 수 있었다.In addition, the optimal P was calculated as follows (Equation 2), the simulation results

Is 2.5, the average number of elected cluster heads is equal to the optimal number of clusters, which maximizes the lifetime of the network.

[Equation 2]

At this time, the first order radio model was used to calculate the energy consumed for communication in the sensor network, and it was consumed for transmission and reception through the following equation (Equation 3, Equation 4, Equation 5, Equation 6). To calculate the energy.

[Equation 3]

[Equation 4]

[Equation 5]

[Equation 6]

은 전송되는 정보의 길이(bit)를 의미하며,

는 두 센서 노드사이의 거리를 의미한다.

와 는

를 기준으로 라디오 신호 증폭기(Radio frequency Amplifier)에서 비트 당 소모되는 에너지를 의미하며,

는 송수신 회로에서 비트 당 소비되는 에너지 소비량을 의미한다. 이때,

는 자유공간(Free Space) 환경에서 전송 거리 d에 따른 전달 손실(Propagation Loss)을 고려하여 라디오 신호 증폭시에서 소모되는 에너지를 의미한다. 또한,

는 다중경로간섭(Multi Path Pading) 환경에서 전송 거리 d에 따른 전달 손실(Propagation Loss)을 고려하여 라디오 신호 증폭시에서 소모되는 에너지를 의미한다. 일반적으로

이면 자유공간환경이 적용되며,

이면 다중경로간섭환경이 적용된다.At this time,

Means the length (bit) of the transmitted information,

Is the distance between two sensor nodes.

Wow Is

Mean energy consumed per bit in the radio frequency amplifier,

Denotes energy consumption per bit in the transceiver circuit. At this time,

Denotes energy consumed in amplifying a radio signal in consideration of propagation loss according to a transmission distance d in a free space environment. Also,

Denotes energy consumed when amplifying a radio signal in consideration of propagation loss according to a transmission distance d in a multipath padding environment. Generally

, The free space environment is applied,

The multipath interference environment is then applied.

In order to calculate the cluster head election opportunity value, the present invention defines a fuzzy if-then rule using three fuzzy sets of the energy remaining of the central sensor node Nu, the regional distance, and the cluster head election opportunity value. Referring to FIG. 3, the energy remaining amount refers to the energy remaining amount of the central sensor node Nu, and the local distance in the central sensor node Nu refers to the sum of the distances with the peripheral sensor nodes Nv existing within the distance r. . In this case, the center sensor node Nu refers to a sensor node supporting the cluster head CH, and the peripheral sensor node Nv refers to a sensor node located around the center sensor node Nu.

For example, in the case of FIG. 3, the local distance is between the first to sixth sensor nodes N1 to N6 and the central sensor node Nu, which are the peripheral sensor nodes Nv existing within the distance r at the central sensor node Nu. It can be found as the sum of the distances of.

[Formula 7]

In addition, the distance r means the average radius of the cluster, it is obtained as shown in Equation 8.

[Equation 8]

In this case, area means the area of the sensor field SF, n means the total number of sensor nodes, and P means the ratio of the required cluster head (CH).

5 to 7 are fuzzy set graphs defined in the sensor network according to the present invention. Fuzzy logic is intended to quantitatively express ambiguous expressions such as language. At this time, the X-axis of Figure 5 is the energy remaining, the X-axis of Figure 6 is the local distance, the X-axis of Figure 7 represents the cluster head election opportunity, the Y-axis of Figures 5 to 7 is about the size of the X axis membership It is shown.

Referring to FIG. 5, when the energy of the central sensor node applying the cluster head in the fuzzy set for the remaining energy is about 20, the absolute value is about 0.6 and the value is about 0.4. Medium). In addition, referring to FIG. 6, when the local distance is about 80, the absolute value corresponds to about 1 (Far) and about 0.4, which corresponds to medium. In this way, the present invention defines the remaining energy of the central sensor node as Low, Medium, and High, and the local distance is defined as Close, Medium, and Far. do. In addition, a cluster head election opportunity is defined as shown in FIG. 7.

Table 1 is a fuzzy if-then rule that expresses the relationship between energy levels, regional distances, and cluster head election opportunities.

TABLE 1

number Energy balance Area street Cluster Head Election Opportunities One Low Far Very Low 2 Low Medium Low 3 Low Close Rather Low 4 Medium Far Med Low 5 Medium Medium Med 6 Medium Close Medium High 7 High Far Rather High 8 High Medium High 9 High Close Very High

Referring to Table 1, for example, when the remaining energy of the central sensor node is Low and the local distance is Close, the chance of selecting a cluster head is somewhat low. In addition, when the remaining energy of the central sensor node is high and the local distance is close, the chance of selecting a cluster head is very high.

8 is an exemplary graph of obtaining cluster head election opportunities using fuzzy set and fuzzy if-then rules. At this time, the energy remaining amount of the central sensor node Nu is 20, and the local distance is 40.

When the energy remaining amount of the central sensor node is 20, the central sensor node has a low energy level of about 0.6 and a medium of about 0.4 according to (a) of FIG. 8. In addition, when the area distance is 40, the area distance is close to about 0.5 and close to about 0.8 according to FIG. 5 (b). In this case, the cluster head election opportunity is calculated by applying 2, 3, 5, and 6 of the fuzzy if-then rules of Table 1. In FIG. 8, (a) rule 3, (b) rule 6, (c) rule 2 and (d) rule 5, the cluster head election opportunity is represented in shades. (E) of FIG. 8 is the final result which summed the result of (a), (b), (c), (d) of FIG. These results are expressed by fuzzy logic and are difficult to use as general figures. Therefore, the present invention calculates the final cluster head election opportunity value through the following equation (Equation 9) using a center of area (COA), which is a defuzzification technique.

[Equation 9]

와

는 각각 도 8(e)에 도시된 if-then 규칙 결과의

축과 멤버쉽 함수를 의미한다.At this time,

Wow

Are each of the result of the if-then rule shown in Fig. 8 (e).

Refers to the axis and membership functions.

The center sensor node Nu calculates the cluster head election opportunity value as described above, and then transmits its cluster head support message to the surrounding sensor nodes Nv with the calculated cluster head election opportunity value. After sending the cluster head support message, the center sensor node Nu sets 'cluster head (CH) = your ID (Nu)' and 'cluster head election opportunity value = your cluster head election opportunity value' as an internal variable. do. When the cluster head support message is received from the peripheral sensor node Nv, the cluster head CH is selected when the cluster head election opportunity value of the peripheral sensor node Nv is large compared to its own 'cluster head election opportunity value'. ) Is set as the peripheral sensor node Nv, and the cluster head election opportunity value is also set as the cluster head election opportunity value of the peripheral sensor node Nv.

The cluster head election step includes identifying a sensor node having the largest cluster head election opportunity value among the sensor nodes satisfying the cluster head support condition (S8), and providing a cluster head message for the sensor node having the largest cluster head election opportunity value. And transmitting to an adjacent sensor node (S9).

In the cluster head election step, the sink SK transmits a cluster head election beacon message to all sensor nodes after a predetermined time after the cluster head support step is completed. The center sensor node Nu receiving the cluster head elected beacon message checks whether the 'cluster head CH' is its ID Nu (S8). At this time, if the 'cluster head CH' is its ID Nu, the central sensor node Nu transmits a cluster head message that it is a cluster head to the peripheral sensor nodes Nv (S9).

The cluster forming step may include receiving a cluster head message from the central sensor node Nu (S10), transmitting a cluster join message to the central sensor node Nu (S11), and sending a cluster support message to the peripheral sensor node (Nu). Receiving from Nv (S12), and adding a peripheral sensor node (Nv) to the cluster node list (S13).

The other peripheral sensor nodes Nv except the center sensor node Nu selected as the cluster head receive the cluster head message (S10) and transmit the cluster join message to the cluster head CH of the closest cluster (S11). The cluster head CH receives the cluster support message from the peripheral sensor node Nv and adds the peripheral sensor node Nv to the cluster node list each time.

In the cluster head forming step, when the sink SK transmits a data transmission beacon message indicating the start of the data transmission step to all sensor nodes, all sensor nodes in the network finish forming the cluster head and complete cluster formation.

Thereafter, the information collected by the sensor node performs a data transmission step of transmitting to the sink SK through the cluster head. Of course, the round is repeated from the cluster configuration step again after the aforementioned round step is completed.

FIG. 9 is a diagram illustrating a simulation result in which clusters are formed by randomly distributing 400 sensor nodes in a 200 × 200 sensor field. At this time, Figure 9 (a) is a simulation results of the cluster formation of the sensor network according to the present invention, Figure 9 (b) is a simulation results of the cluster formation of the sensor network applying LEACH (Low Energy Adaptive Cluster Hierarchy).

Referring to FIG. 9 (a), the present invention shows that the cluster head CH is evenly distributed throughout the network, so that clusters of similar sizes are efficiently formed. However, in the case of LEACH, as shown in FIG. 9 (b), a node is concentrated in one cluster head (CH) so that energy consumption may be more concentrated in the cluster head (CH) (①). In addition, LEACH may cause the cluster heads (CH) to be adjacent to each other, thereby creating the cluster asymmetrically (②), which in turn lowers energy efficiency.

10 is a lifetime graph of a sensor network according to the present invention and a sensor network of LEACH. At this time, the X-axis of the graph is the number of simulations, and the Y-axis is the round in which the sensor node where all the energy is consumed is first found. The sensor network lifetime is generally defined as the time until the first discovery of a sensor node that has run out of energy, which is represented by the number of rounds in FIG. 10. In addition, FIG. 10 performed 100 simulations.

Referring to FIG. 10, it can be seen that the present invention extends network life by about 22.7% over LEACH.

FIG. 11 is a diagram illustrating cluster formation results of a 439th round of a sensor network and a sensor network of LEACH according to the present invention. At this time, Figure 11 (a) is a form of cluster formation of the 439th round of the present invention, Figure 11 (b) is a form of cluster formation of the 439th round in LEACH.

As shown in FIG. 11 (a), the number of sensor nodes capable of operating due to the remaining energy in the 439th round is 193, and the number of cluster heads CH is 21. FIG. However, in the case of LEACH, as shown in FIG. 11 (b), the number of operable sensor nodes is 164 and the number of cluster heads CH is 13. As such, it can be seen that the present invention has a large number of sensor nodes that can operate in the 439th round than LEACH, thereby enabling normal sensor network operation.

As described above, the present invention can efficiently extend the lifespan of the sensor network by selecting a sensor node having a large amount of energy and having a close distance to the adjacent sensor node as the cluster head.

Although described above with reference to the drawings and embodiments, those skilled in the art can be variously modified and changed within the scope of the invention without departing from the spirit of the invention described in the claims below. I can understand.

1 is a block diagram of a sensor network according to the present invention.

Figure 2 is a conceptual diagram showing the operation over time of the sensor network according to the present invention.

3 is a conceptual diagram illustrating an example of calculating a local distance of a sensor network according to the present invention.

Figure 4 is a flow chart showing the algorithm of the cluster configuration step of the sensor network according to the present invention.

5 to 7 is a fuzzy set graph defined in the sensor network according to the present invention.

8 is an exemplary graph for finding cluster head election opportunities using fuzzy set and fuzzy if-then rules.

FIG. 9 illustrates simulation results of forming clusters by randomly distributing 400 sensor nodes in a 200 × 200 sensor field. FIG.

10 is a lifetime graph of the sensor network of the present invention and the sensor network of LEACH.

FIG. 11 is a view illustrating a cluster formation result of a 439th round of a sensor network and a sensor network of LEACH according to the present invention. FIG.

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

N: sensor node Nu: center sensor node

Nv: Peripheral Sensor Node CH: Cluster Head

CF: cluster SF: sensor field

Claims (14)

delete In a sensor network in which a plurality of sensor nodes are composed of at least one cluster in a sensor field, Selecting a cluster head based on an energy remaining of the plurality of sensor nodes and an area distance that is a sum of distances of the plurality of sensor nodes; Forming a cluster with the cluster head and sensor nodes adjacent the cluster head, The area distance is the sum of the distances between the sensor node satisfying the cluster head support condition and the sensor node existing within a predetermined distance, The predetermined distance is

ego, remind

Is

Is, remind

Is

ego, remind

Is

Is, remind

Is the area of the sensor field, remind

Is the number of sensor nodes in the sensor field. The method according to claim 2, Electing the cluster head based on the remaining energy of the plurality of sensor nodes and the area distance that is the sum of the distances of the plurality of sensor nodes, At least one sensor node of the plurality of sensor nodes supporting a cluster head; And selecting a cluster head from among the sensor nodes supporting the cluster head. The method according to claim 3, At least one sensor node of the plurality of sensor nodes supports the cluster head, Calculating cluster head support conditions of the plurality of sensor nodes; Calculating a cluster head election opportunity value of a sensor node satisfying the cluster head support condition; And comparing a cluster head election opportunity value of one sensor node satisfying the cluster head support condition with a cluster head election opportunity value of another sensor node in proximity to the one sensor node. . The method according to claim 4, Computing the cluster head support condition of the plurality of sensor nodes, Random values of each of the plurality of sensor nodes

Generating a; The random value

Multiplied by the ratio of sensor nodes supporting the cluster head to the optimal cluster head ratio required by the sensor network.

Comparing the The random value

Remind

Supporting the cluster head when smaller, remind

Is

The cluster formation method of a sensor network characterized by the above-mentioned. The method according to claim 4, Computing the cluster head support condition of the plurality of sensor nodes, Random values of each of the plurality of sensor nodes

Generating a; The random value

Multiplied by the ratio of the center sensor nodes supporting the cluster heads to the ratio of the optimal cluster heads required by the sensor network.

Comparing the The random value

Remind

Transitioning to listening for a cluster head message when it is greater; remind

Is

The cluster formation method of a sensor network characterized by the above-mentioned. The method according to claim 5 or 6, remind

Is

Is, remind

Is

ego, remind

Is

Is, remind

Is

ego, remind

Is the area of the sensor field, remind

Is the number of sensor nodes in the sensor field, remind

Is 1 or more

The cluster formation method of a sensor network characterized by the following. The method of claim 7, remind

2.5 is the cluster forming method of the sensor network, characterized in that. The method according to claim 4, Computing the cluster head election opportunity value of the sensor node that satisfies the cluster head support condition, And calculating a cluster head election opportunity value of a sensor node that satisfies the cluster head support condition using a non-fuzzy technique. The method according to claim 9, The cluster head election opportunity value is calculated by a fuzzy if-then rule, The fuzzy if-then rule divides the remaining energy of the sensor node into low, normal and high, The local distance of the sensor node is divided into stop and normal and close, Clustering method of the sensor network, characterized in that the cluster head selection opportunity based on the energy remaining amount and the local distance is divided into very little, little, little or little, normal little, normal, normal much, more or more, much, very much . The method according to claim 10, The cluster head election opportunity value is the area of the graph according to the fuzzy if-then rule, The area of the graph according to the fuzzy if-then rule is

ego, remind

Is an opportunity for cluster head election with a minimum value of 0 and a maximum value of 100 according to the fuzzy if-then rule, remind

Is a membership function of the fuzzy if-then rule result. The method according to claim 4, Comparing the cluster head election opportunity value of one sensor node that satisfies the cluster head support condition with the cluster head election opportunity value of another sensor node close to the one sensor node, Transmitting a cluster head support message to another neighboring sensor node; Storing the ID of the one sensor node in cluster head information of one sensor node satisfying the cluster head support condition and storing the cluster head election opportunity value of the one sensor node in cluster head election opportunity value information; Receiving a cluster head support message of another sensor node adjacent to the one sensor node; Maintaining the cluster head information and the head election opportunity value information of the one sensor node when the cluster head election opportunity value of the one sensor node is greater than the cluster head election opportunity value of the other sensor node. Clustering method of the sensor network. The method according to claim 4, Comparing the cluster head election opportunity value of one sensor node that satisfies the cluster head support condition with the cluster head election opportunity value of another sensor node close to the one sensor node, Transmitting a cluster head support message to another neighboring sensor node; Storing the ID of the one sensor node in cluster head information of one sensor node satisfying the cluster head support condition and storing the cluster head election opportunity value of the one sensor node in cluster head election opportunity value information; Receiving a cluster head support message of another sensor node adjacent to the one sensor node; When the cluster head election opportunity value of the one sensor node is smaller than the cluster head election opportunity value of the other sensor node, changing the cluster head information and the head election opportunity value information of the one sensor node to the other sensor node. Cluster forming method of the sensor network comprising a. The method according to claim 3, Electing a cluster head of the sensor nodes supporting the cluster head, Identifying a sensor node having a largest cluster head election opportunity value among the sensor nodes satisfying the cluster head support condition; And transmitting a cluster head message for a sensor node having the largest value of the cluster head election opportunity value to an adjacent sensor node.

Images