KR20160128759A - Wireless sensor network device and method for controlling the same - Google Patents

Abstract

The present invention relates to a wireless sensor network device and a method for controlling the same, and more specifically, to a wireless sensor network device which uses a plurality of sensor nodes and estimates kinds and locations of objects, and a method for controlling the same. The wireless sensor network device, according to one embodiment of the present invention, comprises: the sensor node which is placed in a region and senses at least one object which exists in the region; a communication unit which receives information sensed by the sensor node from the sensor node; and a control unit which estimates a location of the at least one object which exists in the region, on the basis of the sensed information. The control unit extracts labeled data and unlabeled data on the basis of the sensed information, and estimates a kind and a location of an object in the region on the basis of a semi-supervised learning method using the labeled data and the unlabeled data.

Description

TECHNICAL FIELD [0001] The present invention relates to a wireless sensor network device and a control method thereof,

The present invention relates to a wireless sensor network device and a control method thereof, and more particularly, to a wireless sensor network device and a control method thereof capable of estimating the type and position of an object using a plurality of sensor nodes.

Conventionally, techniques such as Time Of Arrival (TOA), Time Difference Of Arrival (TDOA), and Angle OF Arrival (AOA) have been used as techniques for estimating the position of an object. However, these techniques require expensive sensor nodes such as radar. In such an expensive sensor node, the power consumption is inevitably increased and the volume is increased.

Conventionally, Fast Furrier Transfer (FFT) techniques have been applied to acquired data to identify objects. However, applying this technique requires a high-capacity data storage space and a high-performance processor.

As described above, in order to estimate the position of an object or identify an object by a conventional method, high-performance equipment is required, so it is difficult to estimate the position of the object or identify the object using the low-cost sensor node.

Accordingly, in recent years, there is a need to develop a method for estimating the position of an object or identifying an object using a low-cost sensor node.

It is an object of the present invention to provide a wireless sensor network apparatus and a control method thereof capable of estimating the position of an object using a low-cost sensor node.

It is another object of the present invention to provide a wireless sensor network device and a control method thereof, which can estimate the position of an object using both identified data and unidentified data sensed from a plurality of sensor nodes.

A wireless sensor network apparatus according to an embodiment of the present invention includes a sensor node disposed in a certain area and sensing at least one object existing in a certain area, information received by the sensor node from the sensor node And a control unit for estimating a position of at least one object existing in the area based on the sensed information, wherein the control unit controls the display unit to display labeled data based on the sensed information, Extracting the unlabeled data and the semi-supervised learning method using the labeled data and the unlabeled data to determine the type of object in the area and the position Is estimated.

In an exemplary embodiment, the sensor node may include a PIR sensor for sensing infrared rays reflected or emitted from an object, a seismic sensor for sensing vibration generated by an object, And a received signal strength indicator (RSSI) measuring unit for measuring a strength of a signal, wherein the sensed information includes data sensed by at least one of the PIR sensor, the vibration sensing sensor, and the RSSI measuring unit .

In an embodiment, the labeled data is data sensed by at least one of the PIR sensor and the vibration detection sensor, and the control unit is configured to determine, based on the data corresponding to the strength of the signal received through the RSSI measurement unit, The probability distribution diagram is calculated, and the unlabeled data is extracted using the calculated object probability distribution diagram.

In an embodiment, the control unit calculates the object probability distribution diagram by applying filtering to data corresponding to the strength of the received signal.

A method of controlling a wireless sensor network device according to an embodiment of the present invention includes the steps of sensing at least one object located in a certain area in a sensor node disposed in an area, Extracting labeled data and unlabeled data based on the sensed information, and extracting unlabeled data based on the sensed information and semi-map learning using the labeled data and unlabeled data semi-supervised learning method for estimating a type of an object in the area and a position of the object in the area.

In one embodiment of the present invention, the step of extracting includes extracting at least one of data related to infrared rays and data related to vibrations included in the sensed information with the labeled data, Calculating an object probability distribution diagram using data corresponding to the strength of a signal, and extracting the unlabeled data using the calculated object probability distribution diagram.

According to the present invention, it is possible to estimate the position of an unidentified object using a machine learning technique (semi-teaching learning method) and filtering. Therefore, the present invention can propose a new object position estimation method applying both machine learning technique and filtering.

In addition, the present invention makes it possible to estimate the position of an unidentified object using both labeled data and unlabeled data. Accordingly, the present invention solves the problem that there is a limitation in the real time property by using only a small number of labeled data in the related art.

In addition, the present invention can simultaneously perform object identification and position estimation by configuring a wireless sensor network using a low-cost sensor node. Accordingly, the present invention can solve the problem that conventionally had to use an expensive sensor node to identify an object or estimate a position.

In addition, the present invention configures a wireless sensor network using a low-cost sensor node and performs various military applications by simultaneously performing object identification and position estimation. That is, the present invention makes it easy to construct a sensor network over a wide area such as a border line by using a low-cost sensor node, and it is possible to identify an object (target) Can be strengthened.

1 is a conceptual diagram illustrating a wireless sensor network device according to an embodiment of the present invention.
2 is a block diagram illustrating a configuration of a wireless sensor network device according to an embodiment of the present invention.
3 is a flowchart illustrating a method of controlling a wireless sensor network device according to an exemplary embodiment of the present invention.
Fig. 4 is a flowchart for explaining the control method described in Fig. 3 in more detail.
5, 6, and 7 are conceptual diagrams illustrating simulation results of estimating the position of an object according to a wireless sensor network device and a control method thereof according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, wherein like reference numerals are used to designate identical or similar elements, and redundant description thereof will be omitted. The suffix "module" and " part "for the components used in the following description are given or mixed in consideration of ease of specification, and do not have their own meaning or role. In the following description of the embodiments of the present invention, a detailed description of related arts will be omitted when it is determined that the gist of the embodiments disclosed herein may be blurred. It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are intended to provide further explanation of the invention as claimed. , ≪ / RTI > equivalents, and alternatives.

Terms including ordinals, such as first, second, etc., may be used to describe various elements, but the elements are not limited to these terms. The terms are used only for the purpose of distinguishing one component from another.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between.

The singular expressions include plural expressions unless the context clearly dictates otherwise.

In the present application, the terms "comprises", "having", and the like are used to specify that a feature, a number, a step, an operation, an element, a component, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

Hereinafter, a wireless sensor network apparatus and a control method thereof according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

1 is a conceptual diagram illustrating a wireless sensor network device according to an embodiment of the present invention.

First, a wireless sensor network (WSN) of the present invention refers to a sensor node configured as a network. The ubiquitous paradigm that enables human-centered and anywhere access to computing environments is one of the technologies that have been actively researched worldwide.

The wireless sensor network device (100) of the present invention means a device constituting such a wireless sensor network.

Referring to FIG. 1, a wireless sensor network device 100 of the present invention includes at least one sensor node 110 and a control device 120.

The sensor node 110 may refer to a small device having a sensing function and a communication function for observing a physical phenomenon. The sensor node 110 is a basic element that constitutes a wireless sensor network, and a smart dust may be an example.

Referring to FIG. 1, the sensor node 110 may be located in any area 1. Specifically, at least one sensor node 110 may be disposed in any region 1.

The sensor node 110 may sense at least one object 2 existing in the region 1. Here, sensing the object 2 may mean sensing whether the object enters or departs from the region 1, a change in the external environment generated from the object, information related to the object, and the like have.

The sensor node 110 can transmit the sensed information to the control device 120 when sensing the object 2 existing in the region 1. At this time, the sensor node 110 can transmit the sensed information to the control device 120 through the wireless communication.

The controller 120 performs wireless communication with the sensor node 110 and can perform a series of control based on the information received from the sensor node 110. Specifically, the control device 120 may receive the sensed information transmitted from the at least one sensor node 110. Thereafter, the control device 120 can identify the object 2 existing in the region 1 or estimate the position with respect to the object 2 based on the sensed information.

The control device 120 may be any type of device capable of performing wireless communication with the sensor node 110. For example, the control device 120 may include a mobile phone, a smart phone A laptop computer, a desktop computer, an ultrabook, a personal digital assistant (PDA), a radar, a server, and the like.

In this specification, a separate control device 120 is separately described to distinguish the sensor node 110 from the sensor node 110. That is, the control device 120 may be a separate component included in the wireless sensor network device 100 of the present invention. However, the present invention is not limited to this, and the control device 120 may be the wireless sensor network device 100 of the present invention.

Hereinafter, the configuration of the sensor node 110 and the control device 120 included in the wireless sensor network device 100 of the present invention will be described more specifically with reference to FIG.

2 is a block diagram illustrating a configuration of a wireless sensor network device according to an embodiment of the present invention.

2, the sensor node 110 included in the wireless sensor network apparatus of the present invention includes a communication unit 112 for performing wireless communication, a passive infrared radiation sensor 114, a seismic sensor sensor 116, a Received Signal Strength Indicator (RSSI) measuring unit 118, a controller (not shown), and the like.

The communication unit 112 of the sensor node 110 may be configured to perform wireless communication with an external device. The external device may include devices capable of performing wireless communication with the sensor node 110 such as another sensor node 110, a control device 120, and another wireless sensor network device 130. The communication unit 112 may transmit information sensed by the sensors 114, 116, and 118 provided in the sensor node to the communication unit 130 of the control device 120 through wireless communication.

The PIR sensor 114 is a passive infrared sensor capable of detecting infrared rays emitted from an object or emitted from an object. The PIR sensor 114 outputs information related to infrared rays based on the detection of an infrared ray from an object (arbitrary object) existing around the sensor node 110, that is, within a predetermined range from the sensor node 110 (Generated).

For example, when an object is present within a predetermined range from the sensor node 110, if the infrared ray is detected from the object, the PIR sensor 114 can output (generate) '1' information. On the other hand, when the object is out of the predetermined range from the sensor node 110, it can output (generate) '0' information.

The vibration detection sensor 116 is a sensor for detecting vibration. When a vibration occurs by an object (for example, when a vibration occurs in a crust where a sensor node is installed due to movement of an object) Output (generate).

For example, the vibration detection sensor 116 outputs (generates) '1' information when vibration is detected, and outputs (generates) '0' information when vibration is not detected.

The PIR sensor 114 and the vibration sensor 116 output (generate) certain information based on the detection based on the detection, and after generating the information based on the detection, Which means to output (transmit) the above information to the device 120. [

The RSSI measuring unit 118 can measure the intensity of a signal received from the outside. The RSSI measuring unit 118 may be referred to as an RSSI sensor in view of sensing the intensity of a signal received from the outside.

Specifically, the RSSI measuring unit 118 may measure the intensity (RSSI or RSSI value) of a signal periodically received from the external sensor node 110 or the control device 120. At this time, the RSSI measuring unit 118 may measure the RSSI value when the RSSI value changes as an object enters the region where the sensor node 110 is installed.

Further, when an object is formed to emit an arbitrary signal, the RSSI measuring unit 118 may receive a signal emitted from the object and measure an RSSI value of the received signal.

A control unit (not shown) of the sensor node can control components constituting the sensor node. Specifically, when an object is sensed by at least one of the PIR sensor 114, the vibration sensor 116, and the RSSI measurement unit 118, the control unit (not shown) transmits the sensed information to the communication unit 112, To the control device 120 via the network.

The sensed information may include data sensed by at least one of the PIR sensor 114, the vibration sensing sensor 116, and the RSSI measurement unit 118.

The control unit 120 may include a communication unit 130, an output unit (not shown), an input unit (not shown), a memory (not shown), and a controller 180.

The control device 120 (or the wireless sensor network device 100) can perform wireless communication with at least one sensor node installed in a certain area through the communication unit 130 of the control device 120. [ However, the present invention is not limited to this, and the communication unit 130 of the control device 120 may be formed to perform wire communication with an external device, and may be configured to perform communication with an external device capable of communicating outside the sensor node.

The communication unit 130 of the control device 120 may receive the information sensed by the sensor node 110 from the sensor node 110. The received sensed information may be stored in a memory (not shown).

The controller 180 may control the components and the sensor node 110 configured in the controller 120. The controller 180 of the controller 120 can control the communication unit 130, the output unit, the input unit, the memory, and the like provided in the controller 120, (E.g., communication unit 112, PIR sensor 114, vibration sensor 116, RSSI measurement unit 118, and control unit (not shown) of the sensor node).

The control unit 180 estimates the position of at least one object existing in an area where the sensor node 110 is installed based on the information received through the communication unit 130 and sensed by the sensor node 110 .

Hereinafter, with reference to FIG. 3, it is assumed that the wireless sensor network device 100 according to the present invention is configured to detect, based on information sensed through at least one sensor node, at least one A method for estimating a position of an object with respect to an object will be described in more detail.

First, in the present invention, a step of sensing at least one object existing in a certain region is performed by a sensor node disposed in an area (S100). Specifically, the sensor node 110 can sense an object using at least one of the PIR sensor 114, the vibration detection sensor 116, and the RSSI measurement unit 118 described with reference to FIG.

The sensor node 110 may be configured to sense an object existing within a predetermined range from the sensor node 110. As described above, when the sensor node 110 senses an object within the predetermined range from the sensor node 110 through the PIR sensor 114, information related to the infrared ray (or an infrared ray is detected (For example, " 1 ").

Also, the sensor node 110 can sense (detect) the vibration generated by the object through the vibration detection sensor 116. In this case, the vibration detection sensor 116 may generate information related to the vibration (or information that can identify that the vibration is detected) (for example, '1').

Also, the sensor node 110 can measure the strength (RSSI) of a signal received from the outside. The data corresponding to the strength of the received signal may be an intensity value (or RSSI value) of the received signal.

Thereafter, in the present invention, the control device receives the information sensed by the sensor node from the sensor node (S200).

As described above, the sensed information may include data sensed by at least one of the PIR sensor, the vibration sensing sensor, and the RSSI measurement unit. For example, the sensed information may include information related to infrared rays, information related to vibration, data corresponding to intensity of a received signal, and the like.

In the present invention, by applying the machine learning method and filtering using the sensed information, identification of an object and estimation of the position of the object can be simultaneously performed. Here, the machine learning method refers to a method in which an algorithm is applied so that a wireless sensor network device (or a control device) can learn.

The wireless sensor network apparatus of the present invention can apply the semi-supervised learning method in the machine learning method. The semi-guidance learning method may mean a combination of a supervised learning method and an unsupervised learning method.

The semi-guidance learning method uses labeled data and unlabeled data for machine learning. Here, the labeled data may be data associated with a label capable of identifying what kind of data the corresponding data is, i.e., identified data. In addition, the unlabeled data may be unidentified data.

When an object moves in real time, such as the identification and location estimation of an unidentified object, at any one time, the number of labeled data is less than the number of unlabeled data. That is, the number of unlabeled data to be sensed at any one time is larger than the number of labeled data to be sensed at any one of the time points.

The wireless sensor network apparatus of the present invention can be applied to a case of using a low-cost sensor node because the object is identified or the position of the object is estimated by using the semi-map learning method using the labeled data and unlabeled data.

In order to estimate the position of an object existing in a certain area using a semi-guidance learning method, the present invention extracts labeled data and unlabeled data based on the sensed information (S300).

The labeled data may be information related to infrared rays and information related to vibration among the information sensed by the sensor node. Specifically, the labeled data may be data sensed by at least one of the PIR sensor 114 and the vibration sensor 116 of the sensor node. The control unit 180 may extract information related to infrared rays and information related to vibration among the sensed information at the sensor node as labeled data.

The unlabeled data may be an object probability distribution calculated based on data corresponding to the intensity of a signal received at the sensor node among information sensed by the sensor node. Specifically, the control unit 180 may calculate the object probability distribution diagram based on data (RSSI value) corresponding to the intensity of the signal received through the RSSI measuring unit 118 of the sensor node. Then, the controller 180 can extract the unlabeled data using the calculated object probability distribution diagram.

Thereafter, in the present invention, a step of estimating a position of an object in the region is performed based on a semi-supervised learning method using the labeled data and the unlabeled data (S400).

Hereinafter, referring to FIG. 4, a method of extracting labeled data and unlabeled data and a semi-map learning method using the extracted labeled data and unlabeled data to estimate the position of an object The method will be described in more detail with reference to the accompanying drawings.

First, a method of extracting unlabeled data by the wireless sensor network device 100 of the present invention will be described.

As described above with reference to FIG. 3, the sensor node 110 may measure the intensity (RSSI or RSSI value) of a signal received from the outside through the RSSI measuring unit 118 (S110).

 Specifically, the RSSI measuring unit 118 may measure the intensity (RSSI or RSSI value) of a signal periodically received from the external sensor node 110 or the control device 120. At this time, the RSSI measuring unit 118 may measure the RSSI value when the RSSI value changes as an object enters the region where the sensor node 110 is installed. Further, when an object is formed to emit an arbitrary signal, the RSSI measuring unit 118 may receive a signal emitted from the object and measure an RSSI value of the received signal.

The control device 120 (or the wireless sensor network device 100) may receive data (RSSI value) from the wireless sensor 110 on the strength of the received signal included in the sensed information.

The control unit 180 of the control device 120 may calculate the object probability distribution diagram using data corresponding to the intensity of the received signal (S310). Thereafter, the controller 180 may extract the unlabeled data using the object probability distribution diagram (S320).

More specifically, the controller 180 may calculate the object probability distribution by applying filtering to data corresponding to the strength of the received signal.

Here, the object probability distribution map means that the position of the object is represented by a stochastic distribution. The control unit 180 may apply filtering to calculate the object probability distribution diagram. As a filtering method for calculating the object probability distribution diagram, a Kalman filter approximation method based on Gaussian process regression and Radial Basis Function (RBF) can be applied. Here, in the present invention, a calculation of a formula determining the real-time property can be applied to the Kalman filter.

The formula for obtaining the target probability distribution using the RBF-based Kalman filter is as follows.

The control unit 180 can model the target probability distribution value y (X, t) at the position X and the time t as shown in Equation (1).

, 상기

는 radial basis function(RBF),

는 p개의 센서 노드로부터 수신된 시간 t에서의 RSSI값일 수 있다.The Q

, remind

Radial basis function (RBF),

May be the RSSI value at time t received from the p sensor nodes.

The control unit 180 may model y (X, t) of Equation 1 as shown in Equation (2) using a linear markov process.

는 가우시안 노이즈로, 평균값은 0, 분산은

이다. 또한, 상기 h_t(X, Xt)는 시간t-1에서의 물체 확률 분포도 값과 시간 t에서의 물체 확률 분포도 값 사이의 연관함수일 수 있다.At this time,

Is the Gaussian noise, the average value is 0, the variance is

to be. Also, h_t (X, Xt) may be an associative function between the object probability distribution value at time t-1 and the object probability distribution value at time t.

The control unit 180 may perform filtering by applying the Kalman filter after converting Equations (1) and (2) into a vector format. Through this process, the controller 180 can enhance the real-time property of the position estimation of the object when the object moves according to the time.

By performing such filtering, the controller 180 can calculate the object probability distribution diagram as shown in Equation (3), and calculate the object probability distribution diagram using the object probability distribution diagram.

The control unit 180 may extract the unlabeled data using the object probability distribution diagram calculated through the equation (3) (S320).

Next, a method of extracting labeled data from the wireless sensor network device 100 of the present invention will be described.

As described above with reference to FIG. 3, the sensor node 110 may sense infrared rays reflected from an object or emitted from an object through the PIR sensor 114, and generate data related to infrared rays based on the sensing result. In addition, the sensor node 110 senses the vibration generated by the object using the vibration sensor, and generates data related to the vibration based on the sensing result (S120).

If the sensed information is received from the sensor node, the controller 180 may extract the infrared-related data and the vibration-related data included in the sensed information as labeled data at step S330.

Then, the control unit 180 can perform the position estimation on the object type and the object by applying the semi-map learning method using the labeled data and the unlabeled data. Specifically, the control unit 180 can perform (or simultaneously) an operation (operation) for identifying a type of an object and estimating a position for the object by applying a semi-guidance learning method.

Equation (4) represents a formula of a semi-map learning method.

Where V is the loss function, || f || _A is the reproducing kernel Hilbert space regularization, and || f || _I is the manifold regularization. The V, || f || _A and || f || _I are each defined according to a specific algorithm, for example, a semi-supervised support vector machine, a Semi-Supervised colocalization, and a Laplacian gaussian process. Since specific algorithms as described above are generally known techniques, a detailed description thereof will be omitted.

In addition, the V, | f || _A and | f || _I may have different values depending on the labeled data and the unlabeled data, and the controller 180 may use the semi- Optimized data (f * function) can be derived based on the formula of the map learning method.

With this configuration, the wireless sensor network device of the present invention can estimate the type of an object by applying the labeled data and the unlabeled data to the semi-map learning method, and estimate the position of the object.

Figs. 5 to 7 show the results of verifying (simulating) a method of identifying an object or estimating an object through the above-described configuration (method).

Referring to FIG. 5, FIG. 5 shows a result of estimating the position of an object when three objects 2 exist in a state where thirteen sensor nodes 110 are disposed in a certain area.

The lower part of FIG. 5 shows the RSSI value, the RBF function, and the predictive target distribution calculated using the Kalman filter. As shown in the lower diagram of FIG. 5, it can be seen that the probability that an object exists in the coordinates (x, y) where the object exists exists, that is, the probability distribution of the object is high.

Fig. 6 shows an entry position, an estimated path, and an actual path of the ground object and the public object when the ground object and the public object enter together in any area. The wireless sensor network device of the present invention derives the object's type identification, object position estimation as well as an estimated path (i.e., estimated path) of the object based on the semi-map learning method using the labeled data and the unlabeled data It is possible.

As shown in FIG. 6, the wireless sensor network apparatus of the present invention can identify the type of an object (whether it is a ground object or a public object) in an area. Labeled data may be used for the type of the object.

For example, when the object is within a predetermined distance, the control unit 180 can estimate that the object is a terrestrial object when infrared is not detected and vibration is detected. In addition, when the object is within a predetermined distance, the control unit 180 can estimate that the object is a public object when infrared rays are detected and vibration is not detected. However, the present invention is not limited to this, and the estimation method may be changed according to user setting or control of the control unit.

Also, as shown in FIG. 6, the wireless sensor network apparatus of the present invention can estimate the position of an object existing in an area. Here, unlabeled data can be used to estimate the position of the object. Specifically, the controller 180 calculates an object probability distribution diagram based on the RSSI value measured at the sensor node, and estimates the position of the object through the unlabeled data extracted from the object probability distribution diagram.

Also, as shown in FIG. 6, the wireless sensor network apparatus of the present invention can estimate an object's estimated path through position estimation of a moving object in real time based on semi-map learning method using labeled data and unlabeled data .

As shown in FIG. 6, it can be seen that the position of each identified object and the estimated path do not differ greatly from the actual object's path to the object's position.

Figure 7 shows the error values derived from the simulation of Figure 6. Referring to the simulation results shown in FIG. 7, when the semi-guidance learning method using the labeled data and the unlabeled data according to the present invention is used, the estimated path and the actual path of the ground object and the public object are estimated to be about 25 cm It can be seen that a consistent result is obtained while having a small error value.

As described above, according to the present invention, it is possible to provide a method of identifying the type of an object and estimating the position of the object using a low-cost sensor node equipped with a PIR sensor, a vibration sensor, and an RSSI measuring unit.

In addition, according to the present invention, labeling data and unlabeled data are extracted based on information sensed from a sensor node, and the extracted labeled data and unlabeled data are applied to a semi-map learning method, The function of identifying the type and the function of estimating the position of the object may be performed together and the path of the object may be estimated.

Further, according to the present invention, by using the semi-map learning method using the labeled data and the unlabeled data, it is possible to improve the accuracy of the object type identification and object position estimation.

The present invention described above can be embodied as computer-readable codes on a medium on which a program is recorded. The computer readable medium includes all kinds of recording devices in which data that can be read by a computer system is stored. Examples of the computer readable medium include a hard disk drive (HDD), a solid state disk (SSD), a silicon disk drive (SDD), a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, , And may also be implemented in the form of a carrier wave (e.g., transmission over the Internet). Also, the computer may include a control unit 180 of the terminal. Accordingly, the above description should not be construed in a limiting sense in all respects and should be considered illustrative. The scope of the present invention should be determined by rational interpretation of the appended claims, and all changes within the scope of equivalents of the present invention are included in the scope of the present invention.

Claims (6)

A sensor node disposed in an area and sensing at least one object existing in the area;
A communication unit for receiving information sensed by the sensor node from the sensor node; And
And a control unit for estimating a position of at least one object existing in the area based on the sensed information,
Wherein,
Extracts labeled data and unlabeled data based on the sensed information,
Wherein a type of an object in a certain area and a position of the object are estimated based on a semi-supervised learning method using the labeled data and the unlabeled data.
The method according to claim 1,
The sensor node comprises:
A PIR sensor (PIR sensor) for detecting infrared rays reflected or emitted from an object;
A seismic sensor for sensing vibration generated by an object; And
And an RSSI (Received Signal Strength Indicator) measuring unit for measuring an intensity of a signal received from the outside,
Wherein the sensed information includes data sensed by at least one of the PIR sensor, the vibration sensor, and the RSSI measurement unit.
3. The method of claim 2,
Wherein the labeled data is data sensed by at least one of the PIR sensor and the vibration detection sensor,
Wherein,
Calculates an object probability distribution diagram based on data corresponding to an intensity of a signal received through the RSSI measurement unit, and extracts the unlabeled data using the calculated object probability distribution diagram.
The method of claim 3,
Wherein,
Wherein the object probability distribution diagram is calculated by applying filtering to data corresponding to the strength of the received signal.
Sensing at least one object located in an area in which a sensor node is located;
Receiving information sensed by the sensor node from the sensor node;
Extracting labeled data and unlabeled data based on the sensed information; And
A step of estimating a type of an object within the area and a position of the object based on a semi-supervised learning method using the labeled data and the unlabeled data; Way.
6. The method of claim 5,
Wherein the extracting comprises:
Extracting at least one of data related to infrared rays and data related to vibrations included in the sensed information with the labeled data;
Calculating an object probability distribution diagram by using data corresponding to the strength of a received signal included in the sensed information; And
And extracting the unlabeled data using the calculated object probability distribution diagram.

Images