KR102076419B1 - APPARATUS AND METHOD FOR ESTIMATING OCCUPANCY USING IoT INFORMATION - Google Patents

Abstract

Disclosed are an apparatus and a method for estimating occupation by using IoT information. The apparatus for estimating occupation by using IoT information can comprise: an interface part collecting a plurality of IoT information sensed in a target space during a learning period; a model generation part that gives priority in the target space to each piece of the plurality of IoT information, and generates a plurality of classification models by using first IoT information to which priority within a predetermined ranking is given; and a processor that selects an optimal classification model based on the results of verifying the plurality of classification models, and during an estimation period, applies second IoT information sensed and collected in the target space to the optimal classification model to estimate occupation information in the target space.

Description

Apparatus and method for occupancy estimation using IoT information {APPARATUS AND METHOD FOR ESTIMATING OCCUPANCY USING IoT INFORMATION}

The present invention relates to a technique for estimating whether to be occupied in an arbitrary space by using various Internet of Things (IoT) information.

The manager of the building (management system) can take advantage of the presence in the building, more effectively manage the energy efficiency of the building.

Existing methods (eg, carbon dioxide sensor, infrared motion sensor, image image analysis, etc.) are used to determine whether they are occupied due to inaccuracy, installation and maintenance costs, limitations of sensing range, and privacy infringement. The use is limited.

Accordingly, there is a need for a technology capable of confirming occupancy information without high cost while maintaining accuracy.

The present invention collects a plurality of IoT information from a plurality of sensors (for example, sensors that periodically sense the temperature, humidity, carbon dioxide concentration, power consumption, etc.) installed in the target space, and uses the collected plurality of IoT information By estimating the occupancy information in the target space, an object of the present invention can be easily obtained without installing an additional occupancy sensor.

In addition, the present invention considers the influence of the IoT information on the occupancy estimation of the target space, each of the plurality of IoT information sensed in the target space, give priority according to the correlation degree of the occupancy estimation in the target space By estimating occupancy information in the target space using IoT information, which is given priority within a predetermined rank, an object of the present invention is to increase the speed and accuracy of occupancy information estimation.

In order to achieve the above object, the occupancy estimation apparatus using IoT information includes an interface unit for collecting a plurality of IoT information sensed in a target space during a learning period, and each of the plurality of IoT information, priorities in the target space. A model generation unit for generating a plurality of classification models using the first IoT information to which the priority is given within a predetermined rank and a result of verifying the plurality of classification models, and selecting an optimal classification model. The processor may include a processor configured to select second IoT information sensed and collected in the target space during the estimation period and to apply the optimal classification model to the optimal classification model.

In order to achieve the above object, the method for estimating occupancy using IoT information includes collecting a plurality of IoT information sensed in a target space during a learning period, and prioritizing the priority in the target space with each of the plurality of IoT information. Assigning, generating a plurality of classification models using first IoT information to which priority is given within a predetermined rank, and selecting an optimal classification model based on a result of verifying the plurality of classification models. And estimating occupancy information in the target space by applying the second IoT information sensed and collected in the target space during the estimation period to the optimal classification model.

According to the present invention, a plurality of IoT information is collected from a plurality of sensors (for example, sensors that periodically sense temperature, humidity, carbon dioxide concentration, power consumption, etc.) pre-installed in a target space, and the collected plurality of IoT information By estimating the occupancy information in the target space, it is possible to easily obtain the occupancy information without installing an additional occupancy sensor.

Further, according to the present invention, in consideration of the influence of the IoT information on the occupancy estimation of the target space, each of the plurality of IoT information sensed in the target space, give priority in the target space, within a predetermined rank By estimating the occupancy information in the target space by using the IoT information to which priority is given, the accuracy of estimating the occupancy information can be increased.

In addition, the present invention, when combined with MPC (Model Predictive Control) technology in the building, based on the simulation results reflecting the real-time rehabilitation situation, it is possible to control the buildings and facilities more accurately and efficiently than before.

In addition, the present invention can be easily utilized in a security system, elderly care system, and the like by estimating the presence of occupants in real time without a separate occupancy sensor.

In addition, the present invention is embedded in the IoT sensor in the form of a unit module for measuring the indoor environment and energy use, it is possible to easily provide additional room information.

1 is a diagram illustrating an example of a network including an apparatus for estimating occupancy using IoT information according to an embodiment of the present invention.
2 is a diagram showing the configuration of an occupancy estimation apparatus using IoT information according to an embodiment of the present invention.
3 is a view for explaining a process of estimating the occupancy in the occupancy estimation apparatus using the IoT information according to an embodiment of the present invention.
4 is a diagram illustrating an example of selecting first IoT information of a part of a plurality of IoT information collected from a plurality of sensors in an apparatus for estimating occupancy using IoT information according to an embodiment of the present invention.
5 is a flowchart illustrating a method of estimating occupancy using IoT information according to an embodiment of the present invention.

Hereinafter, various embodiments of the present invention will be described in detail with reference to the accompanying drawings and the contents described in the accompanying drawings, but the present invention is not limited or limited to the embodiments.

1 is a diagram illustrating an example of a network including an apparatus for estimating occupancy using IoT information according to an embodiment of the present invention.

Referring to FIG. 1, the network 100 may include a plurality of sensors 110 and an occupancy estimating apparatus 120 (hereinafter, abbreviated as an occupancy estimating apparatus) using IoT information.

The plurality of sensors 110 may be installed in a target space (for example, an office, a laboratory, or a lecture room) to check whether the occupants are occupied. It may include.

The plurality of sensors 110 may transmit the Internet of Things (IoT) information generated as the sensing result to the occupancy estimation apparatus 120.

The occupancy estimation apparatus 120 collects a plurality of IoT information sensed in the target space during the learning period from the plurality of sensors 110, and uses the collected plurality of IoT information to collect the occupancy information in the target space. It can be estimated. In this case, the occupancy estimation apparatus 120 selects IoT information satisfying the set criteria from a plurality of IoT information, generates a classification model using the selected IoT information, and then uses the classification model in the target space. Information on occupancy can be estimated.

In detail, the occupancy estimation apparatus 120 assigns a priority in the target space to each of the plurality of IoT information, selects first IoT information to which priority is given within a predetermined rank, and provides a plurality of classification models. Can be generated. Thereafter, the occupancy estimation apparatus 120 selects an optimal classification model based on the results of verifying the plurality of classification models, and during the estimation period, the optimal IoT IoT information sensed and collected in the target space for the optimal classification model. By applying to the classification model of can be estimated the information on the occupancy in the target space.

Meanwhile, the occupancy estimation apparatus 120 may periodically receive a plurality of IoT information from the plurality of sensors 110, and may make different IoT information satisfying the set criteria for each period. The occupancy estimator 120 updates IoT information satisfying the set criterion at each cycle and uses the data to estimate the occupancy information in the target space, thereby accurately estimating the occupancy information in the target space. Create an environment that is

2 is a view showing the configuration of the physical room estimation apparatus according to an embodiment of the present invention.

Referring to FIG. 2, the occupancy estimation apparatus 200 according to an embodiment of the present invention may include an interface unit 210, a model generator 220, and a processor 230.

The interface unit 210 may periodically collect (eg, every minute) a plurality of IoT information sensed in the target space during the learning period (eg, 20 minutes). Here, the plurality of IoT information is sensed through a sensor installed in the target space, and may include, for example, temperature, humidity, carbon dioxide concentration, power consumption, and the like.

According to an embodiment, if the plurality of IoT information collected during the learning period does not satisfy a predetermined criterion, the interface unit 210 invalidates the plurality of collected IoT information and sets a new learning period. By re-collecting a plurality of new IoT information for the target space.

For example, the interface unit 210 collects a plurality of IoT information about temperature for one minute period during the first learning period '20 minutes', and the difference between the highest temperature and the lowest temperature among the plurality of IoT information for the collected temperature is a reference value. Within 1 degree, the IoT information collected during the first learning period is invalidated, the second learning period '20 minutes' is newly set, and a plurality of pieces of IoT information about temperature may be newly collected in a 1 minute period.

In other words, the interface unit 210 generates an classification model that can be adapted to various field situations by excluding IoT information with too little variation and inducing learning to use IoT information with appropriate variation. You can make it.

The model generator 220 may assign a priority in the target space to each of the plurality of IoT information and generate a plurality of classification models by using first IoT information to which priority is given within a predetermined rank. have. In this case, the model generator 220 assigns a score by converting the influence of the IoT information on the occupancy estimate of the target space into a numerical value, including the influence of the season, and assigns the scores in the order of the high score. The priority may be given for each.

When the plurality of IoT information is n (n is a natural number), the model generator 220 may generate at least 2 ⁿ -1 classification models in consideration of all IoT information combinations.

The processor 230 selects an optimal classification model based on the result of verifying the plurality of classification models, and during the estimation period (for example, 10 minutes), the second IoT information sensed and collected in the target space, By applying the optimal classification model, it is possible to estimate the occupancy information in the target space. Here, the estimation period may be shorter than the learning period, for example, may be 1/2 of the learning period.

In the selection of the optimal classification model, the processor 230 may receive the priorities for each of the plurality of classification models when the actual occupancy information in the target space is received through the interface unit 210 during the learning period. Applying the first IoT information, and among the plurality of classification models, the classification model having the highest number of matches between the occupancy information and the actual occupancy information obtained as a result of applying the first IoT information as the optimal classification model. You can choose.

For example, the processor 230 applies the collected temperature values to the first to fifth classification models for each minute during the 20-minute learning period, and acquires 20 information and objects obtained as a result of applying the 20 temperature values. A classification model having the highest number of matches between actual occupancy information in the space may be selected as the optimal classification model. Here, the 'real room information' may be detected by the sensor 110 installed in the target space, and the interface unit 210 may collect 20 room information actually checked every minute from the sensor 110. . The result of applying 20 temperature values to the first classification model is 18, and the number of matches between the 20 occupant information and the actual occupancy information in the target space is 18, and the 20 temperature values are applied to the second to fifth classification models, respectively. When the number of matches between the 20 occupant information obtained as a result value and the actual occupant information in the target space is 15, 14, 15, and 16, respectively, the processor 230 may select the first classification model as the optimal classification model. .

In this case, the processor 230 increases and adjusts the learning period when the number of matches does not exceed a predetermined value in the classification model selected as the optimal classification model, and the plurality of IoT information sensed during the adjusted learning period. By using, it is possible to reselect the optimal classification model.

For example, in the above example, if the number of coincidences between the 20 occupant information and the actual occupant information for the first to fifth classification models does not exceed 15, which are all set values, the processor 230 may adjust the learning period for an additional 5 minutes. have. The interface unit 210 may further collect IoT information during the added learning period of 5 minutes, and the model generator 220 updates the existing first to fifth classification models based on the collected IoT information. The sixth classification model can be further generated. Thereafter, the processor 230 compares the number of matches between the 20 occupancy information and the actual occupancy information with respect to the updated first to fifth classification models and the newly generated sixth classification model, and the fifteen matching numbers are set values. It is also possible to reselect one classification model that exceeds 0 as the optimal classification model.

Meanwhile, when the second IoT information does not belong to the first IoT information related to generating the optimal classification model, the processor 230 converts an estimation period to a learning period, and then models the generation unit ( Through 220, the plurality of classification models may be regenerated using the second IoT information.

For example, the optimal classification model (attribute: temperature) is selected based on the IoT information about 'temperature' collected during the learning period, and is converted to the estimation period, and the IoT information collected in the converted estimation period is related to the temperature but is directly If the humidity does not belong to the processor 230, the processor 230 converts the estimated period into a learning period, and the model generator 220 may regenerate a plurality of classification models by further using IoT information about humidity. . Thereafter, the processor 230 may again select an optimal classification model (attributes: temperature & humidity) based on the result of verifying the regenerated plurality of classification models.

The learning period and the estimation period may be defined as a window span. The occupancy estimation apparatus 200 may move the window span as time passes, continuously repeating the learning and estimation of the occupancy using the plurality of IoT information, thereby providing reliable occupancy information in the target space in real time. have.

3 is a view for explaining the process of estimating the room in the room estimation apparatus according to an embodiment of the present invention.

Referring to FIG. 3, the occupancy estimation apparatus may collect a plurality of pieces of IoT information x _{1 to} x ₂₃ sensed by a plurality of sensors installed in a target space during the learning period 300. Here, the plurality of IoT information is sensed through a sensor installed in the target space, and may include, for example, temperature, humidity, carbon dioxide concentration, power consumption, and the like.

The occupancy estimation apparatus may assign a priority in the target space to each of the plurality of IoT information (310).

The occupancy estimation apparatus may generate a plurality of classification models using the first IoT information (x _{1 to} x4 ₄ ) to which priority is given within a predetermined rank (320).

The occupancy estimation apparatus may select an optimal classification model based on the results of verifying the plurality of classification models (330). In this case, when the actual occupancy information is received in the target space during the learning period 300, the occupancy estimation apparatus applies the first IoT information to which the priority is assigned to each of the plurality of classification models, and the plurality of classifications. Among the models, a classification model having the highest number of matches between the occupancy information obtained as a result of applying the first IoT information and the actual occupancy information may be selected as the optimal classification model.

The occupancy estimation apparatus may estimate occupancy information in the target space by applying the second IoT information sensed and collected in the target space to the optimal classification model during the estimation period 301.

4 is a diagram illustrating an example of selecting first IoT information of a part of a plurality of IoT information collected from a plurality of sensors in an occupancy estimation apparatus according to an embodiment of the present invention.

Referring to FIG. 4, the occupancy estimation apparatus collects a plurality of IoT information (eg, Light_Power, TEMP, CO _2, etc.) from a plurality of sensors installed in a target space, and collects some first IoT information among the plurality of IoT information. By selecting, it can be used as data for estimating the occupancy information in the target space.

When selecting the first IoT information, the occupancy estimation apparatus may assign a priority in the target space to each of the plurality of IoT information and select first IoT information to which priority is given within a predetermined rank. . In this case, the occupancy estimating apparatus assigns a score by converting the influence of the IoT information on the occupancy estimate of the target space into a numerical value, including the influence according to the season, and the priority for each of the plurality of IoT information in order of the highest score. You can assign a rank.

Here, the occupancy estimation apparatus utilizes the information gain used in information theory based on entropy, which is a measure for evaluating the impurity of the output values in the dataset, for example. Can be assigned.

Specifically, the occupancy estimation apparatus calculates an information gain through Equation 1, and then applies the calculated information gain to Equation 2 to calculate a gain ratio. Can be. The occupancy estimation apparatus utilizes the calculated gain ratio as the score, assigns a priority in the target space to each of a plurality of IoT information, and selects first IoT information to which priority is given within a predetermined rank. Can be.

,

는 출력변수(재실여부, 1 또는 2) 값의 종류를,

는 입력변수(예컨대, IoT 정보가 21일 경우, 1과 21 사이의 정수)의 종류를 의미할 수 있다.here,

,

Is the type of output variable (whether it is lost, 1 or 2),

May mean a type of an input variable (eg, an integer between 1 and 21 when IoT information is 21).

는 데이터 세트D(복수의 IoT 정보) 중 임의로 선택한 데이터의 출력변수 값이

번째 출력변수의 값일 확률이고

는

번째 입력변수를 기준으로 분류된 데이터 세트를 의미한다.Also,

Is the output variable value of randomly selected data among data set D (multiple IoT information)

Is the value of the second output variable

Is

The data set classified based on the first input variable.

Therefore, the larger the gain ratio, the higher the influence on the classification of the output variable.

Meanwhile, the gain ratio for the input variable (IoT information) may be different for each season (for example, winter, transition period, spring, and summer). Accordingly, the occupancy estimation apparatus may calculate a seasonal gain ratio for an input variable (for example, 21 IoT information). For example, the top four input variables having a high impact among the plurality of input variables may be seasonally determined. Can be selected. For example, the occupancy estimation apparatus may select input variables of Light_Power, PC_Usage, ILLUM, and TEMP from winter among 21 input variables, and may select input variables of Light_Power, EHP_Usage, ILLUM, and CO2 in summer.

Looking at the four input variables selected by season, lighting power was the most influential input variable in winter and summer, but the influence was low in season and spring. This can be understood as a result of changing behavior patterns of occupants in relation to lighting. For example, during the winter and summer vacations, most of the non-rooms are turned off, but during the semester or the spring, during temporary semesters such as lectures and meetings, the lights are often turned off and the objects are often empty. Can be judged. In addition, it can be seen that the influence of the energy consumption of the electric heat pump (EHP) fell only in winter, which can be understood as the central space radiator and auxiliary heating equipment are used for heating in addition to the EHP during the winter.

5 is a flowchart illustrating a method of estimating occupancy using IoT information according to an embodiment of the present invention.

Referring to FIG. 5, in operation 510, the occupancy estimation apparatus may collect a plurality of IoT information sensed in a target space during a learning period. In this case, if the plurality of IoT information collected during the learning period does not satisfy a predetermined criterion, the occupancy estimation apparatus invalidates the plurality of collected IoT information and sets a new learning period to the target space. Can re-collect a plurality of new IoT information.

In operation 520, the occupancy estimation apparatus may select first IoT information satisfying a set criterion among the plurality of IoT information, and generate a plurality of classification models using the selected first IoT information. In this case, the occupancy estimation apparatus may give a priority in the target space to each of the plurality of IoT information and generate a plurality of classification models using first IoT information to which priority is given within a predetermined rank. . Here, the occupancy estimation apparatus assigns a score by converting the influence of the IoT information on the occupancy estimate of the target space into a numerical value, including the seasonal influence, and assigns a score to the plurality of IoT information in order of high score. You can assign a rank.

In operation 530, the occupancy estimation apparatus may select an optimal classification model based on a result of verifying the plurality of classification models. In selecting the optimal classification model, the occupancy estimation apparatus receives actual occupancy information in the target space during the learning period, and applies first IoT information to which the priority is given to each of the plurality of classification models, Among the plurality of classification models, a classification model having the highest number of matches between the occupancy information obtained as a result of applying the first IoT information and the actual occupancy information may be selected as the optimal classification model.

In this case, the occupancy estimation apparatus increases and adjusts the learning period in the classification model selected as the optimal classification model, and adjusts the plurality of IoT information sensed during the adjusted learning period. Can be used to reselect the optimal classification model.

For example, if the number of coincidences between the 20 occupant information and the actual occupant information for the first to fifth classification models does not exceed 15, which are all set values, the occupancy estimation apparatus may further adjust the learning period by 5 minutes. The occupancy estimation apparatus may collect more IoT information during the additional learning period of 5 minutes, and further generate a sixth classification model while updating existing first to fifth classification models based on the collected IoT information. can do. Thereafter, the occupancy estimating apparatus compares the number of matches between the 20 occupant information and the actual occupancy information with respect to the updated first to fifth classification models and the newly generated sixth classification model, and selects 15 items having the set number of matches. More than one classification model may be reselected as an optimal classification model.

In operation 540, the occupancy estimation apparatus may estimate occupancy information in the target space by applying the second IoT information sensed and collected in the target space to the optimal classification model during the estimation period.

If the second IoT information does not belong to the first IoT information related to generating the optimal classification model, the occupancy estimating apparatus converts the estimation period into a learning period, and then the second IoT information. By using more, the plurality of classification models can be regenerated.

For example, the optimal classification model (attribute: temperature) is selected based on the IoT information about 'temperature' collected during the learning period, and is converted to the estimation period, and the IoT information collected in the converted estimation period is related to the temperature but is directly In the case of 'humidity' which does not belong to, the occupancy estimation apparatus may convert the estimation period into a learning period and regenerate a plurality of classification models by further using IoT information about humidity. Thereafter, the occupancy estimation apparatus may reselect an optimal classification model (attributes: temperature & humidity) based on the result of verifying the regenerated plurality of classification models.

The apparatus described above may be implemented as a hardware component, a software component, and / or a combination of hardware components and software components. For example, the devices and components described in the embodiments may include, for example, processors, controllers, arithmetic logic units (ALUs), digital signal processors, microcomputers, field programmable arrays (FPAs), It may be implemented using one or more general purpose or special purpose computers, such as a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to the execution of the software. For convenience of explanation, one processing device may be described as being used, but one of ordinary skill in the art will appreciate that the processing device includes a plurality of processing elements and / or a plurality of types of processing elements. It can be seen that it may include. For example, the processing device may include a plurality of processors or one processor and one controller. In addition, other processing configurations are possible, such as parallel processors.

The software may include a computer program, code, instructions, or a combination of one or more of the above, and may configure the processing device to operate as desired, or process independently or collectively. You can command the device. Software and / or data may be any type of machine, component, physical device, virtual equipment, computer storage medium or device in order to be interpreted by or to provide instructions or data to the processing device. Or may be permanently or temporarily embodied in a signal wave to be transmitted. The software may be distributed over networked computer systems so that they may be stored or executed in a distributed manner. Software and data may be stored in one or more computer readable storage media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed by various computer means and stored in a computer readable medium. The computer readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions stored in the media may be those specially designed and constructed for the purposes of the embodiments, or they may be of the kind well-known and available to those having skill in the computer software arts. Examples of computer readable storage media include magnetic media such as hard disks, floppy disks and magnetic tape, optical media such as CD-ROMs, DVDs, and magnetic disks such as floppy disks. Hardware devices specially configured to store and execute program instructions such as magneto-optical media and ROM, RAM, flash memory and the like. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware device described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

Although the embodiments have been described by the limited embodiments and the drawings as described above, various modifications and variations are possible to those skilled in the art from the above description. For example, the described techniques may be performed in a different order than the described method, and / or components of the described systems, structures, devices, circuits, etc. may be combined or combined in a different form than the described method, or other components. Or, even if replaced or substituted by equivalents, an appropriate result can be achieved.

Therefore, other implementations, other embodiments, and equivalents to the claims are within the scope of the claims that follow.

200: device for estimating occupancy using IoT information
210: interface unit
220: model generation unit
230: processor

Claims (12)

During the learning period, the interface unit for collecting a plurality of IoT information sensed in the target space, and receives the actual room information in the target space;
Each of the plurality of IoT information gives priority to the target space, and generates n classification models (where n is a natural number of 5 or more) using first IoT information to which priority is given within a predetermined rank. A model generator; And
The first IoT information is applied to each of the n classification models, and among the n classification models, a classification having the highest number of matches between the occupancy information and the actual occupancy information obtained as a result of applying the first IoT information. A processor that selects a model as an optimal classification model and estimates the occupant information in the target space by applying the second IoT information sensed and collected in the target space to the optimal classification model during the estimation period.
Including,
In the classification model selected as the optimal classification model, when the number of matches does not exceed the set value,
The interface unit,
During the learning period adjusted to be increased by the processor, further collecting IoT information,
The model generator,
Based on the collected IoT information, the n + 1 classification model is newly generated while updating the n classification models,
The processor is
For the updated n classification models and the newly generated n + 1 classification model, the number of matches between the occupancy information obtained as a result of applying the first IoT information and the actual occupancy information is compared, and the number of matches is Reselecting one classification model that is the highest and exceeding the set value as the optimal classification model
Device for estimating occupancy using IoT information. The method of claim 1,
The model generator,
A score is assigned by converting the influence of the IoT information on the occupancy estimate of the target space into a numerical value, including the seasonal influence, and the priority is given to each of the plurality of IoT information in order of the highest score.
Device for estimating occupancy using IoT information. The method of claim 1,
When the second IoT information does not belong to the first IoT information related to generating the optimal classification model,
The processor,
After converting the estimation period to the learning period, through the model generation unit to regenerate a plurality of classification models by further using the second IoT information.
Device for estimating occupancy using IoT information. delete delete The method of claim 1,
When the plurality of IoT information collected during the learning period does not satisfy a predetermined criterion,
The interface unit,
Invalidating a plurality of previously collected IoT information, setting a new learning period, and re-collecting a plurality of new IoT information for the target space;
Device for estimating occupancy using IoT information. During the learning period, collecting a plurality of IoT information sensed in a target space and receiving actual room information in the target space;
Assigning priorities in the target space to each of the plurality of IoT information;
Generating n classification models (where n is a natural number of 5 or more) using first IoT information to which priority is given within a predetermined rank;
The first IoT information is applied to each of the n classification models, and among the n classification models, a classification having the highest number of matches between the occupancy information and the actual occupancy information obtained as a result of applying the first IoT information. Selecting a model, the optimal classification model;
In the classification model selected as the optimal classification model, when the number of matches does not exceed the set value,
During the learning period adjusted to increase, further collecting IoT information;
Generating a new n + 1 classification model while updating the n classification models based on the collected IoT information;
For the updated n classification models and the newly generated n + 1 classification model, the number of matches between the occupancy information obtained as a result of applying the first IoT information and the actual occupancy information is compared, and the number of matches is Reselecting one classification model that is the highest and above the set value as the optimal classification model; And
Estimating occupancy information in the target space by applying the second IoT information sensed and collected in the target space to the optimal classification model during the estimation period.
Room estimation method using the IoT information comprising a. The method of claim 7, wherein
The step of assigning the priority,
Assigning a score by converting the influence of the IoT information on the occupancy estimate of the target space into a numerical value, including an influence according to a season; And
Assigning the priority to each of the plurality of IoT information in order of increasing score
Room estimation method using the IoT information comprising a. The method of claim 7, wherein
When the second IoT information does not belong to the first IoT information related to generating the optimal classification model,
Converting the estimated period into a learning period; And
Regenerating a plurality of classification models by further using the second IoT information;
Room estimation method using the IoT information further comprising. delete delete The method of claim 7, wherein
When the plurality of IoT information collected during the learning period does not satisfy a predetermined criterion,
Invalidating a plurality of previously collected IoT information, setting a new learning period, and re-collecting a plurality of new IoT information for the target space;
Room estimation method using the IoT information further comprising.

Images