KR102165862B1 - Methods and apparatuses for aggregating node data in wireless sensor network - Google Patents

Abstract

The present invention relates to a method and apparatus for aggregating node data in a wireless sensor network. According to an embodiment of the present invention, a method for aggregating data performed by a cluster head in a wireless sensor network comprises the steps of: clustering at least one sensor node among a plurality of sensor nodes based on data similarity; clustering node data by deleting redundant node data from node data of the at least one clustered sensor node; and transmitting the integrated data from which the redundant data has been deleted to a sink node. Embodiments of the present invention may reduce communication costs through efficient aggregation of data and transmission of the aggregated data to a base station (BS).

Description

Node data collection method and apparatus in wireless sensor network {METHODS AND APPARATUSES FOR AGGREGATING NODE DATA IN WIRELESS SENSOR NETWORK}

The present invention relates to a method and apparatus for gathering node data in a wireless sensor network.

A wireless sensor network (WSN) is composed of a large number of sensor nodes that sense surrounding information and transmit the sensed information. In a wireless sensor network, sensor nodes are connected to each other to process information distributedly like grid computing, and unlike personal computers with a limited interface, they have excellent situational awareness with various types of sensors and a large number of sensor nodes. For this reason, wireless sensor networks are evaluated as the most suitable technology for ubiquitous environments. Therefore, wireless sensor networks are used in various fields such as environmental control in intelligent buildings or factories, automatic control of production processes, logistics management, management of goods and information in hospitals, remote sensing of patient conditions, and military control.

The wireless sensor network expands radially around the sink node and forms a network by itself. After that, it depends on the purpose of the application, but the main focus is to collect the information of a specific area into a sink node. In order to transmit data to the sink node, each sensor node receives a data packet from a plurality of sensor nodes located nearby, or transmits the received data packet back to the sink node through the neighboring sensor nodes. In addition, the sensing data detected by the sensor node is transmitted to the sink node through the surrounding sensor node.

As such, the wireless sensor network enables efficient data collection and transmission in an Internet of Things (IoT) environment. However, in general, a wireless sensor network is composed of a plurality of sensor nodes. Therefore, a significant amount of redundant data and anomalies are generated, and the generated redundant data and anomaly values are transmitted to other nodes located nearby. These redundant data and outliers cause wireless sensor network performance degradation.

Embodiments of the present invention aim to provide a method and apparatus for aggregating node data in a wireless sensor network, which can reduce communication costs through efficient aggregation of data and transmission of the aggregated data to a base station (BS).

According to an embodiment of the present invention, there is provided a data aggregation method performed by a cluster head in a wireless sensor network, comprising: clustering at least one sensor node from among a plurality of sensor nodes based on data similarity; Clustering node data by deleting redundant node data from node data of the at least one clustered sensor node; And transmitting the integrated data from which the redundant node data has been deleted to a sink node, a method of setting node data in a wireless sensor network may be provided.

In the clustering of the sensor nodes, the sensor nodes may be clustered based on a maximum cosine similarity based on a cosine similarity function.

The clustering of the node data may further include removing outliers from the node data of the clustered sensor node through an inter-quartile range (IRQ) analysis.

In the step of removing the outlier, the upper and lower limit ranges for identifying the outliers are calculated using a difference value between the first and third quartiles of the node data, and the calculated upper and lower limit ranges are Outliers can be detected as criteria.

In the clustering of the node data, redundant data may be deleted from the node data of the clustered sensor node based on a self-organized map (SOM) neural network.

The node data of the clustered sensor nodes may be arranged in a multidimensional grid in the self-organizing map neural network.

In the clustering of the node data, a Euclidean distance between all weights and input vectors of the self-organizing map neural network is calculated and designated as a winner node, and redundant data overlapping with the node data of the designated winner node may be deleted.

In the clustering of the node data, the winner node may be designated by adjusting a matching weight for an input vector closest to the designated winner node.

In the clustering of the node data, weights of neighboring nodes around the designated winner node may be adjusted.

In the clustering of the node data, whether or not the node data of the designated winner node overlaps with the node data of the designated winner node may be determined using a cosine similarity function with the node data of the designated winner node.

Meanwhile, according to another embodiment of the present invention, a communication module for communicating with a sensor node in a wireless sensor network; A memory for storing at least one program; And a processor connected to the communication module and the memory, wherein the processor executes the at least one program to cluster at least one sensor node from among a plurality of sensor nodes based on data similarity; Clustering node data by deleting redundant node data from node data of the at least one clustered sensor node; And transmitting the integrated data from which the redundant node data has been deleted to a sink node, a node data aggregation apparatus may be provided in a wireless sensor network.

The processor may cluster the sensor nodes based on a maximum cosine similarity based on a cosine similarity function.

The processor may remove outliers from node data of the clustered sensor node through inter-quartile (IRQ) analysis.

The processor calculates upper and lower limit ranges for identifying the outliers by using a difference value between the first and third quartiles of the node data, and calculates the outliers based on the calculated upper and lower limit ranges. Can be detected.

The processor may delete redundant data from the node data of the clustered sensor node based on a self-organized map (SOM) neural network.

The node data of the clustered sensor nodes may be arranged in a multidimensional grid in the self-organizing map neural network.

The processor may calculate a Euclidean distance between all weights of the self-organized map neural network and an input vector to designate a winner node, and delete redundant data overlapping with node data of the designated winner node.

The processor may designate the winner node by adjusting a matching weight for an input vector closest to the designated winner node.

The processor may adjust the weights of neighboring nodes around the designated winner node.

The processor may determine whether or not the node data of the designated winner node overlaps with the node data of the designated winner node by using a cosine similarity function with the node data of the designated winner node.

Meanwhile, according to another embodiment of the present invention, as a non-transitory computer-readable storage medium including at least one program executable by a processor, when the at least one program is executed by the processor, the processor causes: Clustering at least one sensor node among a plurality of sensor nodes based on data similarity, clustering node data by deleting redundant node data from node data of the at least one clustered sensor node, and clustering the redundant node data A non-transitory computer-readable storage medium may be provided containing instructions for causing the deleted aggregated data to be transmitted to the sink node.

Embodiments of the present invention can reduce communication costs through efficient aggregation of data and transmission of the aggregated data to a base station (BS).

1 is a diagram showing a sensor node clustering configuration in a wireless sensor network environment according to an embodiment of the present invention.
2 is a flowchart illustrating a method of setting node data in a wireless sensor network according to an embodiment of the present invention.
3 is a diagram showing a self-organizing map neural network according to an embodiment of the present invention.
4 is a diagram showing an IRQ applied to an embodiment of the present invention.
5 is a diagram showing a neuron grid display in a self-organizing map environment according to an embodiment of the present invention.
6 shows a data set used in a simulation according to an embodiment of the present invention.
7 shows a simulation environment according to an embodiment of the present invention.
8 and 9 are U matrices of a data set showing weight distances between neighboring neurons according to an embodiment of the present invention.
10 and 11 are diagrams showing the number of data points associated with each neuron and cluster according to an embodiment of the present invention.
12 and 13 are diagrams showing a weight plan for each input vector of a data set according to an embodiment of the present invention.
14 is a diagram showing a comparison of the number of data removed in an experiment according to an embodiment of the present invention and a prior art.
15 and 16 are diagrams illustrating a comparison of energy consumption for each data set in an experiment according to an embodiment of the present invention and a prior art.
17 and 18 are diagrams showing a comparison of network lifetime times for each data set in an experiment according to an embodiment of the present invention and a prior art.
19 and 20 are diagrams showing a comparison between a real-time node and a data set in each round in an experiment according to an embodiment of the present invention and a prior art.
FIG. 21 is a diagram showing comparison of accuracy of data similarity measurement in an experiment according to an embodiment of the present invention and a prior art.
22 is a diagram showing a comparison of clustering accuracy between an embodiment of the present invention and an experiment according to the prior art.

Since the present invention can make various changes and have various embodiments, specific embodiments will be illustrated in the drawings and described in detail.

However, this is not intended to limit the present invention to a specific embodiment, it is to be understood to include all changes, equivalents, or substitutes included in the spirit and scope of the present invention.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. These terms are used only for the purpose of distinguishing one component from another component. For example, without departing from the scope of the present invention, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element. The term and/or includes a combination of a plurality of related listed items or any of a plurality of related listed items.

When a component is referred to as being "connected" or "connected" to another component, it is understood that it may be directly connected or connected to the other component, but other components may exist in the middle. Should be. On the other hand, when a component is referred to as being "directly connected" or "directly connected" to another component, it should be understood that there is no other component in the middle.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In the present application, terms such as "comprise" or "have" are intended to designate the presence of features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, but one or more other features. It is to be understood that the presence or addition of elements or numbers, steps, actions, components, parts, or combinations thereof, does not preclude in advance.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs. Terms such as those defined in a commonly used dictionary should be interpreted as having a meaning consistent with the meaning in the context of the related technology, and should not be interpreted as an ideal or excessively formal meaning unless explicitly defined in this application. Does not.

Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the accompanying drawings. In describing the present invention, in order to facilitate an overall understanding, the same reference numerals are used for the same elements in the drawings, and duplicate descriptions for the same elements are omitted.

1 is a diagram showing a sensor node clustering configuration in a wireless sensor network environment according to an embodiment of the present invention.

As shown in FIG. 1, the wireless sensor network environment 100 according to an embodiment of the present invention includes a plurality of nodes 110, a plurality of cluster heads 120, and a sink node 130. However, not all of the illustrated components are essential components. It may be implemented by more components than the illustrated components, and may be implemented by fewer components.

Hereinafter, a detailed configuration and operation of each component of the wireless sensor network environment of FIG. 1 will be described.

An embodiment of the present invention is to extend the life of a wireless sensor network environment by reducing communication cost through efficient data transmission to aggregation and BS (Base Station).

Sensor node clustering in a wireless sensor network environment is configured as shown in FIG. 1.

In clustering of wireless sensor networks, sensor nodes 110 are generally grouped according to a set of rules. Here, the data similarity is expanded based on the cosine similarity function to form a cluster of similar sensor nodes as data. The topology of the wireless sensor network using an embodiment of the present invention is modeled by an undirected graph G = (S, E). Here, S is a set of sensor nodes, and E is a set of energies. When using a cluster-based data set, each cluster head (CH, Cluster Head, 120) integrates the data transmitted from the member node 110 and synchronizes the base station (BS) with one hop or multiple hops as shown in FIG. It is transmitted to the Sink node 130.

Assuming that there are n neighboring nodes (S1, S2, ..., Sn) within the wireless coverage of sensor node S0, the data object of S0 is D0 and the data object of the neighboring node is D1, D2, ..., Dn. to be. If there are N (≤n) data objects whose distance to D is less than ε among n data objects, the data density correlation of s_0 to CH, which is the ε-approximation of D, is as shown in [Equation 1] below. Is defined as

Where minPts is the threshold of the number of neighboring nodes, ε is the data threshold, and dΔ is the distance between D and the center of the data object in ε-neighborhood. d is N data object and s0. a1, a2, and a3 are weights and a1 + a2 + a3 = 1.

Meanwhile, the node may have the same data similarity as the two clusters as shown in q1 of FIG. 1. This makes it difficult to properly determine the cluster. Therefore, the cosine similarity function is used to estimate the similarity between the data of the cluster head CH and the target node, and each node is associated with the cluster head CH having the highest cosine similarity.

1, the data set related to CH1, CH3 is A = [a1, a2, ... , an], B = [b1, b2,… , bn] and C = [c1, c2,… , cn] and q1, respectively. To this end, the cosine-like function cs(a, b) is used to extend the data similarity measurement technique to accurately measure the similarity between two objects as shown in [Equation 2] below.

2 is a flowchart illustrating a method of setting node data in a wireless sensor network according to an embodiment of the present invention.

In step S101, the nodes and the data set are initialized.

In step S102, a node having the maximum energy from all sensor nodes is selected as the cluster head.

In step S103, the cluster head performs node clustering based on data similarity.

In step S104, the cluster head checks each node data and cluster head, and checks whether a join is established.

In step S105, as a result of the confirmation, a node having similar data to the cluster head 1 (CH1) joins the cluster head 1.

In step S106, 50 nodes join the group of cluster head 1 (CH1).

In step S107, cluster head 1 (CH1) aggregates data by applying IRQ and self-organizing map neural network to 50 nodes, and removes outliers and redundant data.

In step S108, cluster head 1 (CH1) performs data cleansing and transmits the integrated data 1 to a sink node.

In step S109, as a result of the confirmation, a node having similar data to the cluster head 2 (CH2) joins the cluster head 2.

In step S110, the 20 nodes join the group of cluster head 2 (CH2).

In step S111, cluster head 2 (CH2) aggregates data by applying IRQ and self-organizing map neural network to 20 nodes, and removes outliers and redundant data.

In step S112, the cluster head 2 (CH2) performs data cleansing and transmits the integrated data 2 to the sink node.

Thereafter, cluster head 3 (CH3) to cluster head n-1 (CHn-1) are repeatedly performed.

In step S113, the cluster head joins a node having similar data to the cluster head n (CHn) as a result of the confirmation to the cluster head n.

In step S114, n nodes join the group of cluster head n (CHn).

In step S115, the cluster head n (CHn) aggregates data by applying the IRQ and the self-organizing map neural network to n nodes, and removes outliers and redundant data.

In step S116, the cluster head n (CHn) performs data cleansing and transmits the integrated data n to a sink node.

And the sink node receives the transmitted aggregated data 1 to the aggregated data n.

3 is a diagram showing a self-organizing map neural network according to an embodiment of the present invention.

A data set method based on a self-organizing map neural network according to an embodiment of the present invention re-clusters data collected in each cluster. The self-organizing map neural network consists of a two-layer neural network, inputs and outputs. As shown in Figure 3, the input nerve layer X = ( x ₁ , x ₂ , ... x _d ) All received data of dimension_d indicated by ^T is the output nerve layer Y = ( y ₁ , y ₂ , … , y _m ) Is fully connected to The order of neural networks in the output layer, and T represents the number of iterations of the learning process.

The synaptic weight vector W _k = ( w _k , ₁ , w _k , ₂ , ..., w _k , _m ) is a directional link between the input layer X and the output layer Y. Here, k ∈{ 1, 2, ..., m ² } represents the index of the k-th node of the output layer. The training process of the self-organizing supervised neural network-based data set method repeatedly updates the synaptic weights of the neurons of the winner and neighbor. At each training step, a sample vector x _i,d is randomly selected from the input data set. As training progresses, the algorithm calculates the Euclidean distance between all weights and the input vector x _d , and the node with the weight vector closest to the input vector is BMU (best-matching unit), j ^* tagged as It is designated as Equation 3].

4 is a diagram showing an IRQ applied to an embodiment of the present invention.

4 shows a diagram of the interquartile (IRQ). In order to detect and remove outliers from the data, quartile (IRQ) analysis is used as shown in [Equation 4] below.

Here, Q ₁ and Q ₃ represent the first and third quartiles of the sample data, respectively, as shown in FIG. 4, and before determining whether the data point is an outlier, the potential of the data point is [ It is identified as in Equation 5]. Here, U _r and L _r represent the upper and lower limit ranges of data points for identifying outliers, respectively.

가 받아들여지면, x _i 는 이상 값이며, 입력 데이터 포인트 x _i 의 중복성을 측정하기 위해 코사인 유사성 함수가 하기의 [수학식 6]과 같이 사용된다.If hypothesis

When is accepted, x _i is an ideal value, and a cosine similarity function is used as shown in [Equation 6] below to measure the redundancy of the input data point x _i .

If cs(x _i , y _m ) = 1, the hypothesis ∂ is allowed and the data point x _i is redundant. As a result, outliers and redundant data are deleted by the following [Equation 7].

는 이상치를 나타내고,

는 중복 데이터를 나타낸다.here,

Represents an outlier,

Represents redundant data.

5 is a diagram showing a neuron grid display in a self-organizing map environment according to an embodiment of the present invention.

As a result, the winner node y _j* is promoted by adjusting the matching weight w _j towards the nearest input vector x _i . In order to determine the neuronal closet for the input vector, the weights of all nodes in the neighborhood of y _j* as well as w _j* are adjusted. In addition, all neurons close to each other are arranged in a two-dimensional grid as shown in FIG. 5.

The synaptic weight of each heightened neuron is adjusted as shown in [Equation 8] below.

Here, α and t represent the learning rate factor and repetition of the learning process, respectively. The Gaussian neighbor function h _ci (t) represents how strongly neighboring neurons are connected as shown in FIG. 5 in the learning process. The Gaussian neighbor function h _ci (t) is represented by the following [Equation 9].

는 그 사이의 거리이다. 다음은 본 발명의 일 실시예에 따른 자체구성 지도 기반의 데이터 집합 방법에 대한 절차를 나타낸다.Where r _c and r _i represent the positions of winner neuron_c and neuron_i in the self-organizing map grid,

Is the distance between them. The following shows a procedure for a self-organized map-based data set method according to an embodiment of the present invention.

In terms of energy efficiency, the self-organizing map-based data set method according to an embodiment of the present invention achieves an energy-efficient data set using machine learning technology. This greatly reduces the amount of data, which has a significant impact on network traffic and energy efficiency. Here, the primary wireless model is adopted to model the power consumption of the sensor node for data reception and transmission. Based on this model, the node needs ε _elec = 50nJ to drive the circuit and ε _amp = 100 pJ/bit/m ² for the transmit amplifier. Therefore, the power consumption to receive 1 bit of data e _r is given by ε _elec . The power consumption e _t for transmitting 1 bit of data to a neighboring node is given by Equation 10 below.

Here, d is the distance between the sender and the receiver node.

It is assumed that data traffic from node i to j per unit time is f _i,j . The energy consumption of node i for data reception and transmission for node j can be expressed as e _r (i, j) and e _t (i, j) as shown in [Equation 11] below.

6 is a diagram illustrating a configuration of a node data aggregation device in a wireless sensor network according to an embodiment of the present invention.

As shown in FIG. 6, in the wireless sensor network according to an embodiment of the present invention, the node data aggregation device 200 includes a communication module 210, a memory 220, and a processor 230. However, not all of the illustrated components are essential components. The node data aggregation device 200 may be implemented by more components than the illustrated components, or the node data aggregation device 200 may be implemented by fewer components. The node data aggregation device 200 may be implemented by being included in the cluster head 120.

Hereinafter, a detailed configuration and operation of each component of the node data aggregation apparatus 200 of FIG. 6 will be described.

The communication module 210 communicates with a sensor node in a wireless sensor network.

The memory 220 stores at least one program.

The processor 230 is connected to the communication module and the memory 220. The processor 230 clusters at least one sensor node among a plurality of sensor nodes based on data similarity by executing at least one program, and deletes redundant node data from node data of at least one clustered sensor node. Thus, node data is clustered, and integrated data from which redundant data has been deleted is transmitted to the sink node.

The processor 230 may cluster the sensor nodes based on a maximum cosine similarity based on a cosine similarity function.

The processor 230 may remove outliers from node data of the clustered sensor node through an inter-quartile (IRQ) analysis.

The processor 230 calculates the upper and lower limit ranges for identifying the outliers by using the difference between the first and third quartiles of the node data, and detects the outliers based on the calculated upper and lower limit ranges. I can.

The processor 230 may delete redundant data based on a self-organized map (SOM) neural network from node data of a clustered sensor node.

The node data of the clustered sensor nodes may be arranged in a multidimensional grid in the self-organizing map neural network.

The processor 230 may calculate the Euclidean distance between all the weights of the self-organized map neural network and the input vector, designate a winner node, and delete redundant data overlapping with the node data of the designated winner node.

The processor 230 may designate a winner node by adjusting a matching weight for an input vector closest to the designated winner node.

The processor 230 may adjust weights of neighboring nodes around the designated winner node.

The processor 230 may determine whether or not the node data of the designated winner node overlaps with the node data of the designated winner node using a cosine similarity function with the node data of the designated winner node.

7 shows a data set used in a simulation according to an embodiment of the present invention.

As shown in FIG. 7, a data set used in a simulation according to an embodiment of the present invention is divided into D ₁ and D ₁ .

8 shows a simulation environment according to an embodiment of the present invention.

As shown in FIG. 8, the simulation environment may include a neural network, algorithms, progresses, and plots.

9 and 10 are U matrices of a data set showing weight distances between neighboring neurons according to an embodiment of the present invention.

9 and 10, the U matrix of the data set shows that the weight distances between neighboring neurons are different.

11 and 12 are diagrams showing the number of data points associated with each neuron and cluster according to an embodiment of the present invention.

11 and 12, the number of data points associated with each neuron and cluster is different.

13 and 14 are diagrams showing a weight plan for each input vector of a data set according to an embodiment of the present invention.

As shown in FIGS. 13 and 14, weights from input 1 to input 14 are shown, respectively, and have different weights.

15 is a diagram showing a comparison of the number of data removed in an experiment according to an embodiment of the present invention and a prior art.

16 and 17 are diagrams showing a comparison of energy consumption for each data set in an experiment according to an embodiment of the present invention and a prior art.

18 and 19 are diagrams illustrating a comparison of network lifetime times for each data set in an experiment according to an embodiment of the present invention and a prior art.

20 and 21 are diagrams showing a comparison between a real-time node and a data set in each round in an experiment according to an embodiment of the present invention and the prior art.

22 is a view showing comparison of the accuracy of data similarity measurement in an experiment according to an embodiment of the present invention and a prior art.

23 is a diagram showing a comparison of clustering accuracy between an embodiment of the present invention and an experiment according to the prior art.

15 to 23, the results are evaluated by computer simulation to verify the results according to an embodiment of the present invention. This is performed by MATLAB Toolbox developed for self-organization map (SOM) and a simulator for clustering wireless sensor networks to evaluate the data reduction rate, energy efficiency, and network lifetime efficiency of the scheme according to an embodiment of the present invention.

Low energy adaptive clustering hierarchy (LEACH), which are three representative data set algorithms according to the prior art, perceptron neural network-based data aggregation (PNNDA), and security based on polynomial regression. A data aggregation technique (PRDA, Polynomial Regression based data aggregation) and a self-organized map-based data aggregation method (SOMDA, Self-Organized Map based Data Aggregation) according to an embodiment of the present invention were compared. The simulation consists of two steps: clustering of sensor nodes and data processing using SOM. In the cluster building step, the sensor node of the greatest energy is selected as the cluster head (CH), and the next node is associated with a CH having the greatest cosine similarity. In the second step, CH runs the SOM algorithm on the collected data to cluster and reduce multidimensional data. Seven metrics were conducted to evaluate the performance of various data set techniques, such as the removal rate of redundant and outlier data, data reduction rate, energy consumption of sensor nodes, real-time nodes of each round, network lifetime, clustering and data similarity accuracy.

As a result of real-time simulation based on seven metrics, redundant data removal using the self-organized map-based data set method according to an embodiment of the present invention consistently showed higher efficiency than the other three techniques.

Referring to FIGS. 14 to 22, it was found that the more outlier and redundant data, the higher the data reduction rate. As the size of the data set increases, energy consumption increases. According to the simulation results, according to the simulation results, the self-organizing map-based data set method (SOMDA) according to an embodiment of the present invention decreases energy consumption as the clustering and data set rounds continue. The performance was consistently superior to the three methods.

The self-organizing map-based data set method according to the embodiments of the present invention described above may be implemented as a computer-readable code on a computer-readable recording medium. The self-organizing map-based data set method according to embodiments of the present invention may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable recording medium.

A non-transitory computer-readable storage medium including at least one program executable by a processor, wherein when the at least one program is executed by the processor, the processor causes: at least one of a plurality of sensor nodes based on data similarity. Commands for clustering one sensor node, clustering node data by deleting redundant node data from node data of the clustered at least one sensor node, and transmitting integrated data from which the redundant data has been deleted to a sink node. Including, a non-transitory computer-readable storage medium may be provided.

The above-described method according to the present invention may be implemented as a computer-readable code on a computer-readable recording medium. The computer-readable recording medium includes all kinds of recording media in which data that can be decoded by a computer system is stored. For example, there may be read only memory (ROM), random access memory (RAM), magnetic tape, magnetic disk, flash memory, optical data storage device, and the like. In addition, the computer-readable recording medium can be distributed to a computer system connected through a computer communication network, and stored and executed as code that can be read in a distributed manner.

Although described above with reference to the drawings and examples, it does not mean that the protection scope of the present invention is limited by the drawings or examples, and those skilled in the art It will be appreciated that various modifications and changes can be made to the present invention without departing from the spirit and scope.

Specifically, the described features may be implemented in digital electronic circuitry, or computer hardware, firmware, or combinations thereof. Features may be executed in a computer program product implemented in storage in a machine-readable storage device, for example, for execution by a programmable processor. And the features can be performed by a programmable processor executing a program of directives to perform the functions of the described embodiments by operating on input data and generating output. The described features include at least one programmable processor, at least one input device, and at least one output coupled to receive data and directives from the data storage system and to transmit data and directives to the data storage system. It can be executed within one or more computer programs that can be executed on a programmable system including the device. A computer program includes a set of directives that can be used directly or indirectly within a computer to perform a specific action on a given result. A computer program is written in any form of a programming language, including compiled or interpreted languages, and is included as a module, element, subroutine, or other unit suitable for use in another computer environment, or as a independently operable program It can be used in any form.

Suitable processors for execution of a program of directives include, for example, both general and special purpose microprocessors, and either a single processor or multiple processors of a different type of computer. Storage devices suitable for implementing computer program directives and data implementing the described features are, for example, semiconductor memory devices such as EPROM, EEPROM, and flash memory devices, magnetic devices such as internal hard disks and removable disks. Devices, magneto-optical disks, and all types of non-volatile memory including CD-ROM and DVD-ROM disks. The processor and memory may be integrated within application-specific integrated circuits (ASICs) or added by ASICs.

The present invention described above is described on the basis of a series of functional blocks, but is not limited by the above-described embodiments and the accompanying drawings, and various substitutions, modifications and changes within the scope not departing from the technical spirit of the present invention It will be apparent to those of ordinary skill in the art to which this invention pertains.

Combinations of the above-described embodiments are not limited to the above-described embodiments, and combinations of various types as well as the above-described embodiments may be provided according to implementation and/or need.

In the above-described embodiments, the methods are described on the basis of a flow chart as a series of steps or blocks, but the present invention is not limited to the order of steps, and certain steps may occur in a different order or concurrently with other steps as described above. I can. In addition, those of ordinary skill in the art understand that the steps shown in the flowchart are not exclusive, other steps are included, or one or more steps in the flowchart may be deleted without affecting the scope of the present invention. You can understand.

The above-described embodiments include examples of various aspects. Although not all possible combinations for representing the various aspects can be described, those of ordinary skill in the art will recognize that other combinations are possible. Accordingly, the present invention will be said to include all other replacements, modifications and changes falling within the scope of the following claims.

100: wireless sensor network environment
110: sensor node
120: cluster head
130: sink node

Claims (21)

In a data set method performed by a cluster head in a wireless sensor network,
Clustering at least one sensor node from among the plurality of sensor nodes based on data similarity;
Clustering node data by deleting redundant node data from node data of the at least one clustered sensor node; And
Transmitting the integrated data from which the redundant node data is deleted to a sink node,
The clustering of the node data further comprises removing outliers from the node data of the clustered sensor node through inter-quartile (IRQ) analysis. The method of claim 1,
Clustering the sensor nodes,
Node data set method in a wireless sensor network, clustering the sensor nodes based on a maximum cosine similarity based on a cosine similarity function. delete The method of claim 1,
The step of removing the outliers,
Radio that calculates the upper and lower limit ranges for identifying the outliers using the difference value between the first and third quartiles of the node data, and detects outliers based on the calculated upper and lower limit ranges How to set node data in sensor network. The method of claim 1,
Clustering the node data,
A node data set method in a wireless sensor network for deleting redundant data based on a self-organized map (SOM) neural network from the node data of the clustered sensor node. The method of claim 5,
The node data of the clustered sensor nodes are arranged in a multidimensional grid in the self-organizing map neural network. The method of claim 5,
Clustering the node data,
A node data set method in a wireless sensor network, in which a Euclidean distance between all weights and input vectors of a self-organizing map neural network is calculated and designated as a winner node, and redundant data overlapping with the node data of the designated winner node is deleted. The method of claim 7,
Clustering the node data,
The node data set method in a wireless sensor network, wherein the winner node is designated by adjusting a matching weight for an input vector closest to the designated winner node. The method of claim 7,
Clustering the node data,
A method of setting node data in a wireless sensor network for adjusting weights of neighboring nodes around the designated winner node. The method of claim 7,
Clustering the node data,
A method of setting node data in a wireless sensor network to determine whether or not to overlap with the node data of the designated winner node using a cosine similarity function with the node data of the designated winner node. A communication module for communicating with a sensor node in a wireless sensor network;
A memory for storing at least one program; And
And a processor connected to the communication module and the memory,
The processor, by executing the at least one program,
Clustering at least one sensor node among a plurality of sensor nodes based on data similarity, but removing outliers through inter-quartile (IRQ) analysis from node data of the clustered sensor node,
Clustering node data by deleting redundant node data from node data of the at least one clustered sensor node,
A node data aggregation device in a wireless sensor network for transmitting integrated data from which the redundant node data has been deleted to a sink node. The method of claim 11,
The processor,
A node data aggregation device in a wireless sensor network for clustering the sensor nodes based on a maximum cosine similarity based on a cosine similarity function. delete The method of claim 11,
The processor,
Radio that calculates the upper and lower limit ranges for identifying the outliers using the difference value between the first and third quartiles of the node data, and detects outliers based on the calculated upper and lower limit ranges Node data set device in sensor network. The method of claim 11,
The processor,
Node data aggregation device in a wireless sensor network for deleting redundant data based on a self-organized map (SOM) neural network from the node data of the clustered sensor node. The method of claim 15,
The node data aggregation device in a wireless sensor network, wherein the node data of the clustered sensor nodes are arranged in a multidimensional grid in the self-organizing map neural network. The method of claim 15,
The processor,
A node data aggregation device in a wireless sensor network for calculating a Euclidean distance between all weights and input vectors of a self-organizing map neural network, designating it as a winner node, and deleting redundant data overlapping with the node data of the designated winner node. The method of claim 17,
The processor,
A node data aggregation apparatus in a wireless sensor network for designating the winner node by adjusting a matching weight for an input vector closest to the designated winner node. The method of claim 17,
The processor,
A node data aggregation device in a wireless sensor network for adjusting weights of neighboring nodes around the designated winner node. The method of claim 17,
The processor,
A node data aggregation apparatus in a wireless sensor network to determine whether or not to overlap with the node data of the designated winner node using a cosine similarity function with the node data of the designated winner node. A non-transitory computer-readable storage medium comprising at least one program executable by a processor, wherein the at least one program, when executed by the processor, causes the processor to:
Clustering at least one sensor node among a plurality of sensor nodes based on data similarity, but removing outliers through inter-quartile (IRQ) analysis from node data of the clustered sensor node,
Clustering node data by deleting redundant node data from node data of the at least one clustered sensor node,
A non-transitory computer-readable storage medium comprising instructions for causing the redundant node data to transmit the deleted aggregate data to a sink node.

Images