KR20180123254A - Non-Intrusive Appliance Load Monitoring Method and System based on IoT Service Interworking - Google Patents

Abstract

The present invention provides a method and a system for identifying a non-contact home appliance based on internet of things (IoT) service interworking. According to an embodiment of the present invention, the method for identifying a device predicts a probability for each device status and power usage state information by using measured values detected by IoT service interworking sensors installed in a space in which at least one device is included. Therefore, the present invention can improve an identification rate of a home appliance without an additional cost such as separate development of the sensor by using influence of the sensors and the home appliance for an IoT service.

Description

TECHNICAL FIELD [0001] The present invention relates to a non-contact appliance identification method and system based on IoT service interworking,

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to energy management technology, and more particularly, to a method and system for identifying a power usage state of a home appliance.

Energy management systems such as BEMS (Building Energy Management System), FEMS (Facility Energy Management System) and HEMS (Home Energy Management System) are being implemented.

Such an energy management system is intended to save energy by controlling devices that unnecessarily use electrical energy. In order to maximize its effect, identification of the power-consuming devices should take precedence.

To this end, there is a contactless method which uses a smart plug and a non-contact method which identifies a household electric appliance based on a total electric power amount. However, the former has an inefficiency due to a cost problem, have.

In the latter case, it is possible to assume the complement of the sensor to improve the device recognition rate, but there is an additional cost problem such as the separate development of the sensor.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a method and system for identifying use of home appliances using a sensor that is linked to and connected to an IoT service.

According to an aspect of the present invention, there is provided a method for identifying a device, the method comprising the steps of: using sensor data, which are measured values sensed by a plurality of sensors installed in a space including at least one device, A first calculation step of calculating sensor data based state information which is information predicted by a state probability; And estimating power usage status information of at least one device using the calculated sensor data based status information, wherein the plurality of sensors include IoT service interworking sensors.

And, IoT service interworking sensors can provide sensor value through IoT service platform and open API.

In addition, the IoT service may include at least one of a commercial IoT service and a standard IoT service.

The IoT service may include a service for providing sensor values of a plurality of sensors through a smart device.

In addition, the plurality of sensors may further include sensors that are individually installed without interlocking the IoT service.

The prediction step may predict the power usage state information of at least one device by using the calculated sensor data based state information and the previously calculated device state information together.

The first calculating step may calculate sensor data based status information using the influence degree matrix showing the degree of the measurement value of each sensor in the state prediction of each device for each device, have.

The first calculating step uses only the measured values that exceed the specific threshold value among the measured values sensed by the plurality of sensors installed in the space including at least one device as valid sensor data, It is possible to calculate the sensor data based status information which is information predicting the star probability.

According to another aspect of the present invention, there is provided an energy management system comprising: a receiver for receiving sensor data, which are measured values sensed by a plurality of sensors installed in a space including at least one device; And estimating power usage status information of at least one device by using the sensor data based status information calculated by the sensor data based on the sensor data, ; And the plurality of sensors include IoT service interlock sensors.

According to various embodiments of the present invention, it is possible to improve the identification ratio of home appliances without additional cost such as the development of sensors separately using the influence of sensors and appliances for commercial IoT service.

In addition, according to various embodiments of the present invention, the home appliance is identified based on the sensor data collected from the IoT service such as a mobile communication company, thereby ensuring independence from the sensor and expanding the supply in the future Since it is possible to link with the expected IoT sensor, the interworking cost is reduced and more sensors can be expected to be used.

Brief Description of the Drawings Fig. 1 is a view provided in a conceptual description of a non-contact home appliance status identification method based on an IoT service interworking sensor,
FIG. 2 is a diagram illustrating a structure of a device state identification module implemented in a HEMS;
FIG. 3 is a conceptual diagram of a method of identifying a home electric appliance according to an embodiment of the present invention,
4 is a block diagram of a HEMS in accordance with an embodiment of the present invention.
FIG. 5 is a flowchart illustrating a method of identifying a home appliance according to an exemplary embodiment of the present invention;
FIG. 6 is a graph showing influences of a home appliance and a sensor according to an embodiment of the present invention,
FIG. 7 is a schematic view illustrating a process of obtaining an influence matrix between a home appliance and a sensor according to an embodiment of the present invention; FIG.
FIG. 8 is a diagram illustrating a process of applying a threshold value to a sensor measurement value according to an embodiment of the present invention, and FIG.
9 is a diagram illustrating a concept of state retention probability of a home appliance according to an embodiment of the present invention.

Hereinafter, the present invention will be described in detail with reference to the drawings.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram illustrating a concept of a data-based non-contact home appliance status identification method collected from IoT service interworking sensors. As shown in FIG. 1, the Home Energy Management System (HEMS) collects the values of the sensors connected to the commercial IoT service platform through the open API.

Thus, the ON and OFF states of the home appliances are detected based on the probabilities, and unnecessary home appliances are turned off through the presence or absence of current home users or behavior patterns.

On the other hand, the user can receive such information through the HEMS using a smart phone or monitor the sensor values by connecting to the commercial IoT service homepage. When IoT sensors connected to mobile communication companies are installed in the home, sensors, HEMS and users are interconnected through a service platform and it is very easy to utilize sensor data.

2 is a diagram illustrating a structure of an Appliance Status Identification Module implemented in the HEMS. As shown in FIG. 2, the device identification module is a module for managing a commercial and standard IoT service platform (IoT service platform) through a matrix (Appliance Consumption Power Matrix) and IoT_IF_Agents using total home power consumption information obtained through a power meter, (Sensor-Appliance Influence Matrix) using the sensor value received from the sensor-appliance matrix, thereby deriving the final instrument state identification result.

Here, the IoT_IF_Agents are connected to commercial IoT services such as KT, SKT, and LGU +, and international and industry standard IoT services such as oneM2M, OCF, and AllJoyn, and these agents collect sensor values through services through the normal operation.

The concept of the home appliance identification method according to the above process is shown in more detail in FIG. According to the method of identifying a home appliance according to the embodiment of the present invention, the HEMS 100 can use the total power based state information, the sensor data based state information, and the previously calculated home appliances The device state information is used together to predict the power usage state information of at least one household electric appliance.

In this case, the total power-based state information indicates a method of predicting the current household appliances from the total power consumption amount. For example, the non-intrusive appliance load monitoring (NIALM) based on the total power consumption have. The HEMS 100 calculates the total power based state information by applying the device use identification algorithm through the power total amount based device identification measured through the power meter of the distribution board shown in FIG.

In addition, the sensor data based status information is information that predicts the probability of each of the home appliances using the sensor data obtained by using the measured values of various sensors of the hybrid sensor. The HEMS 100 detects whether or not the device is used through the hybrid sensor-based device usage situation measurement (acoustic appliance, electrothermal appliance, lighting appliance, etc.) measured through the composite sensor including the acoustic sensor, the illuminance sensor and the temperature sensor shown in FIG. And the sensor data base state information is calculated by applying the identification algorithm.

Here, the composite sensor further includes individual sensors separately installed for purposes other than the IoT service interworking in addition to the IoT service interworking based sensors.

The HEMS 100 can collect the status of each of the household appliances (audio, set-top, electric heater, lighting, etc.) by using the total power based status information, the sensor data based status information and the previously calculated home appliance status information together On, off, river, middle, and medicine).

Prediction of the power consumption per household appliance through the household appliance identification method can be expressed as the total power consumption [P (t)] of the household appliances collected from the power meter of the distribution board during a specific time period t, To the consumed electric power (p _i ) per household appliance to predict the household appliances currently in use.

_{P (t) = p 1 (} t) + p 2 (t) + p 3 (t) ... + p m (t)

The above equation represents the total power consumption [P (t)] divided by the individual power consumption of m household appliances and expressed as a sum value.

The HEMS 100 uses the learning results of the power consumption patterns of the home appliances to calculate the total power based state information.

The HEMS 100 calculates sensor data based state information, which is information that predicts the probability of each home appliance by using sensor data, which are measurement values sensed by a plurality of sensors installed in a space including home appliances . The sensor data consists of a combination of acoustic information, illuminance information and temperature information, and sensors (IoT service interlock sensors + individual sensors) for collecting the information are required as shown in Fig.

However, the above listing of the information constituting the sensor data is merely illustrative. It is, of course, possible to implement sensor data by excluding at least one of them, replacing at least one with other information, or including at least one other information.

Hereinafter, a method for identifying the household appliance of the HEMS 100 will be described in detail. 4 is a block diagram of a HEMS 100 in accordance with an embodiment of the present invention.

The HEMS 100 according to an embodiment of the present invention includes a receiving unit 110, a processor 120, a storage unit 130, and an output unit 140, as shown in FIG.

The power meter 10 measures the total power consumption of a space including at least one household appliance. The composite sensor 20 includes at least two of an acoustic sensor, an illuminance sensor, and a temperature sensor, which are implemented by an IoT service interlocking sensor and / or an individual sensor, and is installed in a space including at least one household appliance, And generates sensor data serving as a basis of the state information.

The receiving unit 110 receives the total power consumption value from the power meter 10 and receives sensor data from the composite sensor 20. [ The receiving unit 110 can receive the total power consumption value and the sensor data by wire or wirelessly.

The processor 230 executes the household appliance identification method shown in Fig. The processor 230 may output the execution result through the output unit 250. [

The storage unit 240 stores a power consumption pattern learning result database, past sensor data, total power consumption, and an influence matrix to be described later, and previously calculated home appliance status information.

The HEMS 100 described in FIG. 4 may be implemented as a device that is physically independent of itself, as a part of any device or system, or may be implemented as a program or framework installed in a computer or a server, And may be implemented in the form of software such as an application. In addition, each component of the HEMS 100 may be implemented as a physical component or as a functional form of software.

Hereinafter, with reference to FIG. 5, a detailed description will be given of a method of identifying a home electric appliance by the HEMS 100 described above. 5 is a flowchart illustrating a method of identifying a home appliance according to an exemplary embodiment of the present invention.

The HEMS 100 uses sensor data, which are measured values detected by a plurality of sensors installed in a space including at least one home appliance, to calculate sensor data based status information (S210).

At this time, the HEMS 100 calculates the sensor data based state information using the influence degree matrix that shows the degree of the measurement value of each sensor in the state prediction of each home appliance, .

The influence degree matrix will be described with reference to FIG. 6 and FIG.

6 is a graph showing influences of a home appliance and a sensor according to an embodiment of the present invention.

The HEMS 100 first generates the influence graph of the home appliance and the hybrid sensor. A change in the state of the appliance affects the environment of the surrounding space. For example, a lighting device may be illuminated when turned on and dimmed when turned off, so that the condition can be detected by the illuminance sensor. In addition, since a sound is emitted when the TV or sound device is turned on and no sound is emitted when the sound is turned off, it is possible to detect the state by the acoustic sensor. Thus, the composite sensor 20 collects sensor data that measures such environmental change. The HEMS 100 expresses the learned result based on the collected sensor data in the form of an influence graph shown in FIG. 6 and stores the result in the storage unit 130. Here, the influence graph is information representing the degree of change of the device state corresponding to the change of the surrounding environment by defining the state of the device corresponding to the user's behavior.

The influence graph expresses the relationship between the state of the appliance and the sensor data which detects the change of the surrounding environment. One independent influence graph is generated for each appliance. If each sensor value does not affect the instrument state change (subgraph with e = 0), do not create it.

The influence graph represents a sensor (circle in Fig. 6) that affects the state change according to the state of the home appliance (bold square in Fig. 6) (the general square in Fig. 6). If the value measured through the sensor affects a particular instrument condition, define the degree as e. The HEMS 100 generates an independent influence graph for all home appliances and stores the graph in the storage unit 130. [

Then, the HEMS 100 generates an affinity matrix using the affinity graph, which will be described with reference to FIG. FIG. 7 is a diagram schematically illustrating a process of obtaining an influence matrix between a home appliance and a sensor according to an embodiment of the present invention. Referring to FIG.

As shown in FIG. 7, the HEMS 100 separates a given influence graph into subgraphs of degree = 1. The separated subgraph consists of one sensor corresponding to the appliance state. Then, the HEMS 100 inputs the influence values corresponding to the K states in the i-th row and the j-th column corresponding to the indices of the household appliances and sensors corresponding to the subgraph in a vector form. Here, i denotes the i-th home appliance, j denotes the j-th sensor, and K denotes the total number of types of household appliances. Enter 0 for the element of the index where the subgraph does not exist. The HEMS 100 generates the affinity matrix E upon completion of generation of the influence graph generated from all the appliances. Then, the HEMS 100 measures the degree of the measurement value of each sensor in the state prediction of each home appliance by using the influence degree matrix E, t). This process can be expressed in concrete terms as follows.

The sensor data base state information A (t) can be expressed by Equation (1) and Equation (2) using the influence degree matrix E generated by the influence degree graph and the sensor data x of the measured equation (4) . Equation (3) represents the value of each element of the influence matrix E. Here, the total number of state types of household appliances is K, m is the number of household appliances, and n is the number of sensors. And, the influence degree matrix E is a matrix of m 占 n, and each element is a vector including K influence degree values.

가 된다. Specifically, the sensor data values measured at each composite sensor 20 at time t are stored in the x (t) vector of Equation (4). Therefore, when the set of all the sensor modules included in each composite sensor 20 is S _i , the vector of the measured sensor data is

.

In addition, the HEMS 100 uses only measurement values exceeding a specific threshold value, which are detected by a plurality of sensor modules installed in a space including at least one home appliance, as valid sensor data. Specifically, the HEMS 100 defines an activation function to effectively activate the sensor value only when the measured value in the sensor module exceeds a threshold for influencing the state of the corresponding appliance . Applying the activation function to the sensor data corresponds to Equation (5), and y in Equation (5) represents the result of applying the activation function f to the sensor data.

In addition, since the sensor values measured in each sensor module include characteristics (temperature: ° C, sound: dB) of different sensor modules, the HEMS 100 also performs a process of normalizing the values, (5).

)으로 정규화하여 최종 결과 벡터 y를 산출하게 된다. That is, as represented in equation (5), HEMS 100 applies an activation function according to a given threshold value for each measured sensor value (x _i (t)). The result is stored in the sensor measurement maximum value (

) So as to calculate the final result vector y.

FIG. 8 is a diagram illustrating a process of applying a threshold to a sensor measurement value according to an exemplary embodiment of the present invention. Referring to FIG. As shown in FIG. 8, the HEMS 100 measures a value exceeding a specific threshold value among the measured values sensed by a plurality of sensor modules installed in a space including at least one home appliance using the activation function f Only the values are used as valid sensor data. Specifically, FIG. 8 shows a process of extracting a final result vector y using the measured sensor data x (t). In Fig. 8, the activation function is defined as a max function. The activation function can be defined differently depending on the characteristics of the sensor and the home appliance. A (t) can be calculated as Equation (1) by applying the degree e that affects the state of the appliance to the y vector thus extracted.

However, the use of the max function as an activation function is merely an example, and the HEMS 100 can apply a threshold value to the sensor data by using various functions (e.g., a step function, etc.) .

Through this process, the HEMS 100 uses the sensor data, which are the measurement values sensed by the plurality of sensors installed in the space including at least one home appliance, The sensor data base state information is calculated.

Referring back to FIG. 5, the HEMS 100 calculates the total power based state information, which is information for predicting the probability of each at least one home appliance using the change pattern of the total power consumption (S220).

As described above, the total power-based state information represents state information calculated by a method of predicting what kind of household appliances currently in use from the total power consumption amount. The non-intrusive method (NIALM: Appliance Load Monitoring). The HEMS 100 can calculate the total power based state information by applying the non-contact identification method based on the total power consumption based on the total power amount measured through the power meter 10 of the distribution board. At this time, the calculated total power based state information is represented by D (t).

Then, the HEMS 100 calculates a value obtained by multiplying the calculated total power based state information D (t) by the first weight? ₁ and a value obtained by multiplying the calculated sensor data base state information A (t) by the second weight? ₂ (T) representing the probability value for each state of the household appliance by adding all the values of the previously calculated home appliance state information S (t-1) multiplied by the third weight? ₃ and the state maintaining probability? (S230). Here, the home appliance state information S (t) is a value including all the probabilities of all the states to which the home appliance can correspond, and corresponds to a matrix expressed by the following equation (6).

(6)

(6)

(7)

(7)

Here, all the elements of S (t), D (t), A (t) and S (t-1) are m × K columns, Represents a probability value. The total number of state types of the household appliances is K, m is the number of household appliances, and n is the number of sensors.

The first weight ω ₁ , the second weight ω _2, and the third weight ω ₃ may vary depending on the type of the space including at least one home appliance, the number of sensors, and the distribution of the sensors. For example, the HEMS 100 may set the first weight ω ₁ , the second weight ω _2, and the third weight ω ₃ to be different depending on whether the space including the appliances is a company space, an apartment space, or a housing space . Also, the HEMS 100 can be set such that the second weight? ₂ becomes larger as the number of sensors is increased or the distribution density of the sensor is higher. The number of the sensor more or the distribution density of the sensor is high the information by the sensor because the accuracy is relatively high, HEMS (100) includes a first weight ω ₁ and a second weight ω ₂ and the third weighting ω ₃ of the associated with the sensor The second weight < RTI ID = 0.0 > ₂ < / RTI >

The state retention probability pi indicates a probability that the state of the previous home appliance state information S (t-1) is maintained as it is. That is, since the probability that the previous home appliance state information S (t-1) will remain unchanged as the state maintaining probability? Increases, the previous home appliance state information S (t-1) is multiplied by the state maintaining probability? If the appliance is a type of household appliance that is kept in one state, it is possible to increase the weight of the previous appliance state information by multiplying the state maintaining probability? By the old appliance state information S (t-1). Accordingly, the state retention probability? Varies depending on the state of each appliance, and the HEMS 100 obtains the state retention probability pi by state of each appliance.

9 is a diagram illustrating a concept of state retention probability of a home appliance according to an embodiment of the present invention.

이다. 이전 가전기기의 상태에 상태 변경 확률을 적용하여 현재 가전기기 상태 식별 예측값에 반영한다. The state maintaining probability? Of the home appliance can be defined in the manner shown in FIG. 9 to express the degree to which the state of the previous home appliance affects the state of the current household appliance. The state retention probability pi, which is the probability that the state of the home appliance will not change

to be. Apply the state change probability to the state of the previous home appliance and reflect it in the current household appliance state identification prediction value.

Referring back to FIG. 2, the HEMS 100 calculates household appliance status information including probability values to be applied to each state of each appliance in all states of the appliance, calculates the highest value among the probability values of each state of the calculated appliance- It is predicted that the state corresponding to the probability value is the current state of the home appliance (S240).

Specifically, when the HEMS 100 finally calculates the home appliance state information S (t) at time t, it is determined that the state of the highest probability value for each appliance is the current state of the home appliance as shown in the following equation (8) .

(8)

(8)

Then, the HEMS 100 can estimate the power usage state information of at least one home appliance by calculating the current consumption power of the home appliance by applying the power consumption value corresponding to the predicted current state. Here, the power use state information is information indicating the power consumption amount of each appliance, and is information that is calculated by extracting the amount of power consumption corresponding to the current state of the home appliance predicted through the appliance state information. The HEMS 100 calculates the power usage state information of each of the home appliances at every cycle. The user checks the power usage state information calculated by the HEMS 100 to determine which home appliances are consuming the power at a glance . Further, since the HEMS 100 predicts the power use state information by using all of the power based state information, the sensor data based state information, and the previously calculated home appliance state information, the power consumption amount of the highly accurate home appliances .

In the present embodiment, the HEMS 100 uses the calculated total power based state information, the calculated sensor data based state information, and the previously calculated home appliance state information together to calculate the power use state information of at least one household appliance It is also possible to predict the power use state information of at least one home appliance using only the calculated sensor data based state information and to calculate the calculated sensor use state information and the previously calculated power use state information The power usage state information of at least one household electric appliance may be predicted using the calculated total power based state information and the calculated sensor data based state information. Of course.

It is needless to say that the technical idea of the present invention can also be applied to a computer-readable recording medium having a computer program for performing the function of the HEMS 100 and the method of identifying the home appliance according to the present embodiment. In addition, the technical idea according to various embodiments of the present invention may be realized in the form of a computer-readable programming language code recorded on a computer-readable recording medium. The computer-readable recording medium is any data storage device that can be read by a computer and can store data. For example, the computer-readable recording medium may be a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical disk, a hard disk drive, a flash memory, a solid state disk (SSD), or the like. In addition, the computer readable code or program stored in the computer readable recording medium may be transmitted through a network connected between the computers.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, It will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the present invention.

10: Power Meter
20: Composite sensor (IoT service interlock sensor + individual sensor)
100: Home Energy Management System (HEMS)
110: Receiver 120: Processor
130: storage unit 140: output unit

Claims (8)

A first calculation step of calculating sensor data based state information, which is information predicting probability of each device by using sensor data which are measured values sensed by a plurality of sensors installed in a space including at least one device, ; And
And estimating power usage status information of at least one device using the calculated sensor data based status information,
The plurality of sensors,
RTI ID = 0.0 > IoT < / RTI > service interlock sensors.
The method according to claim 1,
IoT service interlock sensors,
Wherein the sensor value is provided through an IoT service platform and an open API.
The method of claim 2,
The IoT service,
A commercial IoT service, and a standard IoT service.
The method of claim 2,
The IoT service,
And providing a sensor value of a plurality of sensors via a smart device.
The method of claim 2,
The plurality of sensors,
Further comprising sensors individually installed without IoT service interworking.
The method according to claim 1,
In the prediction step,
And the power usage state information of at least one device is predicted using the calculated sensor data based state information and the previously calculated device state information together.
The method according to claim 1,
The first calculating step includes:
Wherein the state information based on the sensor data is calculated by using the influence degree matrix showing the degree of the measurement value of each sensor in the state prediction of each device for each device and for each state of each sensor.
A receiving unit for receiving sensor data, which are measured values sensed by a plurality of sensors installed in a space including at least one device; And
A processor for calculating sensor data based state information, which is information that predicts a probability of each device state by using sensor data, and predicting power use state information of at least one device by using the calculated sensor data based state information; Lt; / RTI >
The plurality of sensors,
≪ / RTI > IoT service interlock sensors.

Images