KR20220122367A - Method of providing olfactory environment information service using olfactory information classification process based on neural network learning - Google Patents

Abstract

In a service providing method according to an embodiment of the present invention, the reactive pattern information is collected from the olfactory environment sensor device having a single sensor module for measuring the reactive pattern information of the volatile organic compound from the olfactory environment sensor device, and the reactive pattern information is collected. The information is applied to one or more pre-machine-learned olfactory environment sensor information neural network models to obtain olfactory environment information, and olfactory environment service information is configured based on the olfactory environment information and transmitted to a user terminal corresponding to the olfactory environment sensor device. do.

Description

A method for providing olfactory environment information service using a neural network learning-based olfactory information classification process

The present invention relates to a service providing apparatus and an operating method thereof. More specifically, the present invention relates to an olfactory environment sensor device using a neural network learning-based olfactory information classification process, an olfactory environment information service providing device, and an operating method thereof.

The environment can be viewed as natural conditions that can directly or indirectly affect living things. Such an environment may contain gas particles distributed in the air, and humans inhale the gaseous particles present in the air through the nose, and the olfactory receptors in the nose are chemically stimulated to recognize and identify them as smells through the olfactory nerve. .

To such olfactory nerve stimulation, a person can respond emotionally and feel mental or emotional stability. It can also help you get a good night's sleep and reduce body fatigue.

On the other hand, in the case of organic compounds constituting a harmful environment, such as cigarette smell and sewer smell, an olfactory nerve stimulus may cause discomfort. In the case of such odor, hydrogen sulfide, mercaptans, amines, and other stimulatory gaseous substances stimulate the olfactory nerve, giving discomfort and disgust. may cause

In order to recognize and cope with such odor generation, various systems using sensors for each gaseous compound have emerged. For example, in Korean Patent No. 10-2056354, a plurality of sensor modules arranged at a set interval on a substrate to measure the concentration of odor generation contained in the self-introduced gas and to discharge the gas to the outside thereof; Disclosed is a gas sensor array for measuring odor, which includes a gas inlet pipe that transmits gas flowing into a sensor module and a gas outlet pipe that transmits gas flowing out from the plurality of sensor modules.

However, such a current odor measurement system has to separately install different sensor modules for sensing each gas characteristic, which not only increases the manufacturing cost, but also requires an integrated understanding of the operation method or condition of each module, and the environment If this is changed, the usability deteriorates, and the lifespan of each sensor module is short, so there is a problem that it needs to be replaced frequently.

In particular, since it is possible to measure only the gas for each installed sensor module, there is a problem that it is very difficult to distinguish between various and complex odors. ham exists.

Furthermore, even if such a system is installed in a space such as a hospital room that requires different immediate countermeasures for each odor when a specific odor occurs, real-time information about various odor sensing is provided to users who have difficulty understanding the meaning of gas sensing information. It is too difficult to expect a response.

In particular, although proper comprehensive management and response to various odors generated in the facility management area such as each ward, classroom, and office are required, it is appropriate at the present time to grasp only limited sensor information unless the manager personally visits and takes charge of it. And the situation is not being made in response.

The present invention was devised to solve the above problems, and as it provides an olfactory environment information service using a neural network learning-based olfactory information classification process, a single sensor module-based olfactory environment sensor with a simple structure is the only one based on machine learning. An olfactory environment sensor device that can sense various substances to provide a simple olfactory environment situation and long lifespan while lowering manufacturing costs, and a service providing device that can provide various environmental information services and countermeasure notification services based on this And an object of the present invention is to provide a method for operating the same.

In a method of operating a service providing apparatus according to an embodiment of the present invention for solving the above problems, the reactive pattern information from the olfactory environment sensor device having a single sensor module for measuring the reactive pattern information of volatile organic compounds receiving; obtaining olfactory environment information by applying the reactive pattern information to one or more pre-machine-learned olfactory information neural network models; and constructing olfactory environment service information based on the olfactory environment information and transmitting the olfactory environment service information to a user terminal corresponding to the olfactory environment sensor device.

A service providing apparatus according to an embodiment of the present invention for solving the above-described problems, receiving the reactive pattern information from an olfactory environment sensor device having a single sensor module for measuring reactive pattern information of volatile organic compounds communication department; an olfactory environment information processing unit configured to obtain olfactory environment information by applying the reactive pattern information to one or more pre-machine-learned olfactory information neural network models; and a service providing unit that configures olfactory environment service information based on the olfactory environment information and transmits it to a user terminal corresponding to the olfactory environment sensor device.

An olfactory environment sensor device according to an embodiment of the present invention for solving the above problems, a single sensor module for measuring the reactive pattern information of the volatile organic compound; and a communication module for transmitting the reactive pattern information to a server, wherein the server obtains olfactory environment information by applying the reactive pattern information to one or more pre-machine-learned olfactory information neural network models, and the olfactory environment information It is characterized in that the server configures olfactory environment service information based on the olfactory environment and transmits the information to a user terminal corresponding to the olfactory environment sensor device.

On the other hand, the method according to an embodiment of the present invention for solving the above problems may be implemented as a program for executing the method in a computer and a recording medium in which the program is recorded.

According to an embodiment of the present invention, the reactive pattern information is collected from the olfactory environment sensor device having a single sensor module for measuring the reactive pattern information of the volatile organic compound, and the reactive pattern information is used in one or more pre-machine-learned A service providing apparatus that obtains olfactory environment information by applying it to an olfactory information neural network model, configures olfactory environment service information based on the olfactory environment information and transmits it to a user terminal corresponding to the olfactory environment sensor device, and an operating method thereof; An olfactory environment sensor device used for this may be provided.

Accordingly, according to an embodiment of the present invention, it is possible to provide an olfactory environment information service using a neural network learning-based olfactory information classification process, and a single sensor module-based olfactory environment sensor with a simple structure senses various substances based on machine learning. A service providing device capable of providing an olfactory environment sensor device that can provide a simple olfactory environment situation identification and long lifespan while lowering manufacturing costs, and providing various environmental information services and countermeasure guide notification services based on this A method of its operation may be provided.

In particular, the service according to an embodiment of the present invention can be usefully used in indoor service systems sensitive to odors or specific odors such as hospital rooms, hotels, and service workplaces, and furthermore, certain harmful substances in external environments such as cleaning services and camping sites There is an advantage that can be applied to a management service system such as a safety inspection service for the environment.

1 is a conceptual diagram schematically illustrating an entire system according to an embodiment of the present invention.
2 is a block diagram illustrating an environment sensor device according to an embodiment of the present invention in more detail.
3 is a block diagram illustrating a service providing apparatus according to an embodiment of the present invention in more detail.
4 is a flowchart illustrating an operation of a service providing apparatus according to an embodiment of the present invention.
5 is a flowchart illustrating in more detail an operation of learning a neural network and providing a service of a service providing apparatus according to an embodiment of the present invention.
6 to 7 are diagrams illustrating experimental analysis of reactive pattern data collected according to an embodiment of the present invention.
8 is an exemplary diagram of a service screen interface provided to a user terminal according to an embodiment of the present invention.

The following is merely illustrative of the principles of the invention. Therefore, those skilled in the art will be able to devise various devices that, although not explicitly described or shown herein, embody the principles of the present invention and are included within the spirit and scope of the present invention. Further, it is to be understood that all conditional terms and examples listed herein are, in principle, expressly intended solely for the purpose of enabling the concept of the present invention to be understood, and not limited to such specifically enumerated embodiments and states. should be

Moreover, it is to be understood that all detailed description reciting the principles, aspects, and embodiments of the invention, as well as specific embodiments, are intended to cover structural and functional equivalents of such matters. It is also to be understood that such equivalents include not only currently known equivalents, but also equivalents developed in the future, i.e., all devices invented to perform the same function, regardless of structure.

Thus, for example, the block diagrams herein are to be understood as representing conceptual views of illustrative circuitry embodying the principles of the present invention. Similarly, all flowcharts, state transition diagrams, pseudo code, etc. may be tangibly embodied on a computer-readable medium and be understood to represent various processes performed by a computer or processor, whether or not a computer or processor is explicitly shown. should be

In addition, the clear use of terms presented as processor, control or similar concepts should not be construed as exclusively referring to hardware having the ability to execute software, and without limitation, digital signal processor (DSP) hardware, ROM for storing software. It should be understood to implicitly include (ROM), RAM (RAM) and non-volatile memory. Other common hardware may also be included.

The above-described objects, features and advantages will become more apparent through the following detailed description in relation to the accompanying drawings, and accordingly, those of ordinary skill in the art to which the present invention pertains can easily implement the technical idea of the present invention. There will be. In addition, in the description of the present invention, when it is determined that a detailed description of a known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted.

Hereinafter, a preferred embodiment according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a conceptual diagram schematically illustrating an entire system according to an embodiment of the present invention, and FIG. 2 is a block diagram illustrating an olfactory environment sensor device according to an embodiment of the present invention in more detail.

1 and 2 , a system according to an embodiment of the present invention includes an olfactory environment sensor device 100 , a service providing device 200 , and a user terminal 300 .

First, the user terminal 300 described in this specification includes a mobile phone, a smart phone, a laptop computer, a digital broadcasting terminal, a personal digital assistant (PDA), a portable multimedia player (PMP), a navigation system, and the like. may be included, but the present invention is not limited thereto, and may be various devices capable of user input and information display other than that.

In addition, the service providing device 200 collects the reactive pattern information from the olfactory environment sensor device having a single sensor module for measuring the reactive pattern information of the volatile organic compound from the olfactory environment sensor device 100 , and the reactivity A user terminal ( 300) is sent.

Here, the service providing device 200 may register and manage user information of the user terminal 300 , management workplace information, and mapping information of the olfactory environment sensor device 100 , and receive from the olfactory environment sensor device 100 . It is possible to monitor the data to be used, process the monitored information and the olfactory environment service information into a service interface and provide it to the mapping registered user terminal 300 .

For example, the olfactory environment sensor device 100 may be provided at the top of each bed in a hospital room, and the mapping information of the olfactory environment sensor device 100 includes device identification information, hospital room number information, bed identification information in a hospital room, and bed location. Since information may be included, it is possible to sense odor information for each specific user in a specific ward.

To this end, the olfactory environment sensor device 100, the service providing device 200, and the user terminal 300 may include one or more communication modules for mutual information registration and data transmission/reception, and the communication module is wired or wireless. A communication function for communicating over a network may be provided.

In addition, each communication module remotely transmits and receives data through an Internet network, LAN, WAN, PSTN (Public Switched Telephone Network), PSDN (Public Switched Data Network), cable TV network, WIFI, mobile communication network and other wireless communication networks, etc. As for short range communication, Bluetooth (Bluetooth), RFID (Radio Frequency Identification), Infrared Data Association (IrDA), UWB (Ultra Wideband), ZigBee, NFC (Near Field Communication) ), using at least one of Wireless-Fidelity (Wi-Fi), Wi-Fi Direct, and Wireless Universal Serial Bus (USB) technologies may also be exemplified.

The olfactory environment sensor device 100 may include a single olfactory environment sensor module 110 for measuring the concentration of volatile organic compounds (VOCs), and the single olfactory environment sensor module 110 is a TVOC A semiconductor sensor may be exemplified.

The TVOC sensor reacts with about 20 to 37 dormant compounds including at least one of acetaldehyde, acetylene, benzene, butane, chloroform, ethylene, formaldehyde, methanol, propylene, and toluene, to produce organic compounds in the air. As a sensor outputting a concentration, it is possible to measure and output a relative change in smell according to a change in TVOC concentration.

Since this TVOC sensor is a single semiconductor sensor, it does not require separate olfactory environment sensor modules, so the equipment and installation cost of the olfactory environment sensor device 100 can be greatly reduced compared to conventional methods, and can be easily installed in various locations. It has the advantage that it can be installed.

However, the TVOC single sensor itself has a feature that it cannot distinguish the type of smell or the gas other than the simple concentration. The olfactory environment sensor information neural network model according to an embodiment of the present invention can be built in advance by using the concentration response pattern information according to the sensor output, and real-time received through the communication module 120 of the olfactory environment sensor device 100 . By applying the reactive pattern information to the neural network model, it is possible to classify and identify odors by each organic compound, and provide an information service to the user terminal 300 based on the identified odor classification result information.

That is, based on the detection process based on a neural network model trained by deep learning on the reactivity pattern information for each time section of the TVOC recognized in response to each odor information, the TVOC single sensor alone detects a complex concentration change pattern and provides odor classification information. was confirmed according to the experiment to be recognizable, which will be described in more detail later.

The olfactory environment sensor information neural network model according to an embodiment of the present invention may be configured by a deep learning-based machine learning model construction method, and is mapped to the reactivity pattern information of the TVOC sensor and the reactivity pattern information of the TVOC sensor. A first neural network learning model obtained by performing LSTM (Long Short-Term Memory) learning of a Recurrent Neural Network (RNN) method based on the olfactory environment classification information, the reactivity pattern information of the TVOC sensor, and the reactivity of the TVOC sensor A second neural network learning model obtained by performing learning of a convolutional neural network (CNN) method based on olfactory environment classification information mapped to pattern information, reactivity pattern information of the TVOC sensor, and reactivity pattern information of the TVOC sensor. It may be selectively configured by including a third neural network learning model obtained by performing FNN (Fully Connected Network) learning based on the olfactory environment classification information.

In addition, the service providing apparatus 200 may combine and calculate each olfactory environment information derived by applying real-time responsive pattern information to each constructed neural network learning model to finally determine olfactory classification result information. The olfactory classification result information may be configured differently according to the classification criteria of odors such as cigarette smell, ammonia smell, and the like.

Also, the service providing apparatus 200 may configure a service interface including the olfactory classification result information, and may provide the service interface to the user terminal 300 . For example, the service interface may include notification service information corresponding to the olfactory classification result information, and the notification service information may include guide action information preset in response to the olfactory classification result information.

For example, the user of the user terminal 300 may be a business manager of a hotel, a hospital, a restaurant, a cafe, a PC room, a karaoke room, a gym, etc., and the smell generated in each hospital room, guest room, business place, etc., which is difficult for the manager to directly grasp Information can be easily grasped through the service interface at a low installation cost, and a related response action can be immediately performed in real time.

In addition, it is sometimes difficult for the manager of each business site to know what kind of response is desirable even if a specific smell is detected. For example, even if a bad odor occurs, various countermeasures such as ventilation, cleaning, direction, deodorization, disinfection, closing the door, and laundry may vary according to the priority according to the classification. In other words, washing is an act of sucking the smell of clothes, and closing the door is only to buy time to perform temporary isolation from external odors, so it may not be a priority action.

To solve this problem, the service providing apparatus 200 according to an embodiment of the present invention may provide preset guide behavior information in response to olfactory classification result information through a service interface, and the provided guide behavior information is The guide behavior information indexed by mapping the olfactory classification result information to the constructed guide behavior information mapping table may be included.

Accordingly, the business site manager can easily recognize and manage what kind of odor or scent is emitted at the current location, and can immediately identify and respond to necessary measures for bad odors and the like.

On the other hand, although the olfactory environment sensor device 100 described above is described as a single device, it is included as one component of the fragrance generator 400 that discharges the fragrance generated from one or more fragrance modules or fragrance capsules mounted on the body. You may. For example, the fragrance module may be mounted in a plurality, according to a user setting set corresponding to the user terminal 300 or schedule information input from the user terminal 300, provided by the service providing apparatus 200 According to the fragrance scheduling information, the combination of each individual fragrance module and the intensity of scent diffusion may be adjusted.

Accordingly, the service providing apparatus 200 determines a combination of a fragrance module for improving the olfactory environment information and a fragrance diffusion intensity based on the olfactory environment information, and the fragrance generator 400 in which the olfactory environment sensor device 100 is mounted. ) can also be sent.

3 is a block diagram illustrating a service providing apparatus according to an embodiment of the present invention in more detail.

Referring to FIG. 3 , the service providing apparatus 200 according to an embodiment of the present invention includes a communication unit 210 , a control unit 220 , a database 230 , a device manager 240 , an olfactory environment information processing unit 250 and and a service providing unit 260 .

The communication unit 210 may include one or more communication modules for wirelessly or wired communication with at least one of the user terminal 300 and the olfactory environment sensor device 100 .

In addition, the communication unit 210 may receive registration information from the user terminal 300 and transmit it to the device manager 240 , and receive reactive pattern information from the olfactory environment sensor device 100 , and receive the olfactory environment information processing unit 250 . to the user interface, and the user interface information configured by the service providing unit 260 may be transmitted to the user terminal 300 .

The control unit 220 may include a control circuit that generates a control signal for performing the overall operation of the service providing apparatus 200 and transmits it to each unit, wherein the control circuit is a method according to an embodiment of the present invention. It may include one or more microprocessors that drive the

The database 230 may be composed of one or more storage media for information storage, and store the user terminal 300 information managed by the service providing device 200, the olfactory environment sensor device 100 information, and the artificial intelligence neural network model database. may include

In addition, the device manager 240 may receive and register authentication information of the olfactory environment sensor device 100 from the user terminal 300 , and store the registered user terminal 300 and the olfactory environment sensor device 100 information. mapping, and storing and managing the mapping information in the database 230 .

Here, the user terminal 300 and the olfactory environment sensor apparatus 100 may each exist in plurality, and one or more olfactory environment sensor apparatuses 100 mapped to a plurality of user terminal 300 groups may also exist. A plurality of olfactory environment sensor devices 100 mapped to one user terminal 300 may also exist. The device manager 240 may include the mapping information in a service interface provided from the communication unit 210 to the service provider 260 .

For example, the mapping information may include device identification information, location information, and installation information of the olfactory environment sensor device 100 , and the device manager 240 performs the user terminal 300 through the service provider 260 . Through the service interface provided as , an accurate installation position of the olfactory environment sensor device 100 currently being monitored can be identified. For example, when the olfactory environment sensor device 100 is installed in a hospital, the mapping information may include room information in the hospital, installed bed identification information, bed location map information, and the like.

Meanwhile, the olfactory environment information processing unit 250 collects reactive pattern information received from the device-registered olfactory environment sensor device 100 and applies the collected reactive pattern information to a pre-built neural network model, thereby olfactory classification result information can be configured.

To this end, the olfactory environment information processing unit 250 may include a neural network model construction unit 251 and a classification result information construction unit 253, and the neural network model construction unit 251 includes a first neural network model construction unit ( 2511 ), a second neural network model construction unit 2513 , and a third neural network model construction unit 2515 .

Here, the neural network model building unit 251 divides the reactive pattern learning data corresponding to the olfactory environment sensor device 100 by time section, and learns and processes the olfactory environment classification information mapped to the reactive pattern learning data, so that one or more Deep learning-based neural network models can be built.

In particular, as various models are emerging in the current neural network model technology, the neural network model building unit 251 according to an embodiment of the present invention combines or sums the output signals of a plurality of neural network models by combining or summing the registered olfactory environment sensor. Real-time classification result information corresponding to the reactivity pattern information of the device 100 may be configured.

More specifically, the neural network model building unit 251 is configured to perform recurrent (RNN) based on olfactory environment classification information mapped to reactive pattern learning data of the TVOC sensor of the olfactory environment sensor device 100 and reactive pattern learning data of the TVOC sensor. It may include a first neural network learning model building unit 2511 for performing LSTM (Long Short-Term Memory) learning of a Neural Network) method.

In addition, the neural network model building unit 251 is a Convolutional Neural Network (CNN) based on the olfactory environment classification information mapped to the reactive pattern learning data of the TVOC sensor of the olfactory environment sensor device 100 and the reactive pattern learning data of the TVOC sensor. ) may include a second neural network model building unit 2513 that performs learning.

In addition, the neural network model building unit 251 is a fully connected network (FNN) based on the olfactory environment classification information mapped to the reactive pattern learning data of the TVOC sensor of the olfactory environment sensor device 100 and the reactive pattern learning data of the TVOC sensor. ) may include a third neural network learning model obtained by performing the learning method.

Accordingly, the classification result information configuration unit 253 combines the output of the first neural network learning model according to the application of the reactive pattern information, the output of the second neural network learning model, and the output of the third neural network learning model , the olfactory classification result information of the olfactory environment information may be finally determined. Here, as the combinational operation method, an average operation, a multiplication operation, a sum operation, or the like may be exemplified.

More specifically, the neural network model building unit 251 configures a window for each time section corresponding to the graph of the reactive pattern learning data, and configures the olfactory environment classification information mapped to the window-based data array as deep learning learning data. The first neural network model building unit 2511 , the second neural network model building unit 2513 , and the third neural network model building unit 2515 may be inputted to each, and a learning training process for improving the accuracy of odor determination may be commanded.

Here, in the case of the LSTM model built by the first neural network model building unit 2511, it may be a method of constructing a recursive neural network (RNN) that is optimized for time series data and configured to have a different processing structure for each layer.

Then, the second neural network model building unit 2513 applies convolution-based Maxpooling, batch normalization, and dropout to prevent overfitting, and then, two-dimensional data After performing flatten processing to convert , into one-dimensional data, it is possible to perform fully connected dense layer processing for classification and to build a neural network model that performs classification processing. have.

On the other hand, the third neural network model building unit 2515 may be configured as a fully connected deep network (FCC), and performs neural network model construction based on a method of connecting and learning all neurons of each neural network. can do.

However, the above-described first neural network model building unit 2511, second neural network model building unit 2513, and third neural network model building unit 2515 are exemplary, and each neural network model building unit 251 has more performance. When good neural network model building processes are developed later, a specific neural network model building unit may be added or the existing model building unit may be replaced.

Meanwhile, the service providing unit 260 configures a service interface including the olfactory classification result information and transmits it to the user terminal 300 . The service interface may include notification service information corresponding to the olfactory classification result information, wherein the notification service information is a guide action preset in response to the olfactory classification result information and mapping information of the olfactory environment sensor device 100 . may contain information.

To this end, the service providing unit 260 may configure an olfactory environment service interface to be provided to the user terminal 300 in response to the olfactory classification result information, and the olfactory environment service interface is a graphic-based GUI (GRAPHIC USER INTERFACE). form may be visually or aurally output through the user terminal 300 .

In addition, the service providing unit 260 may configure various service interface dashboards in order to provide olfactory environment management and information services to the user terminal 300 , and the olfactory classification result information and the information through the service interface dashboard. Various additional service information corresponding to the olfactory classification result information may be provided.

In particular, the service providing unit 260 provides output information of the first neural network learning model used for the olfactory classification result information, output information of the second neural network learning model, and the third neural network learning to provide more accurate information. The output information of the model may be included in the olfactory environment service interface and transmitted to the user terminal 300 .

For example, the user terminal 300 checks the identification information in the management area of the olfactory environment sensor device 100, the location information and installation information in the management area through the service interface, and the smell corresponding to the current location in the management area. You can check the monitoring information, and you can check the guide for countermeasures such as ventilation, deodorization, and removal of pollutants corresponding to the specific odor shown in the classification result information, and take action immediately.

Meanwhile, the service providing unit 260 may provide a pre-set interworking service with the central management system of the institution, and may provide the service interface to the central management system of the institution. For example, the institution may be a hospital, and the central management system of the hospital may reflect the olfactory service interface information in the ward state information based on the olfactory service interface information, and the ward manager in which the olfactory service interface information is reflected in the ward state information. An interface may be provided as a separately set hospital manager terminal.

4 is a flowchart for explaining the operation of the service providing apparatus 200 according to an embodiment of the present invention.

Referring to FIG. 4 , the service providing apparatus 200 according to an embodiment of the present invention processes device registration of the user terminal 300 and the olfactory environment sensor apparatus 100 ( S101 ).

Then, the service providing apparatus 200 collects reactive pattern information of the olfactory environment sensor apparatus 100 ( S103 ).

Thereafter, the service providing apparatus 200 obtains olfactory environment information by applying the reactive pattern information to one or more olfactory environment sensor information neural network models (S105), and configures an olfactory environment service interface based on the obtained olfactory environment information, It is transmitted to the user terminal 300 corresponding to the olfactory environment sensor device (S107).

As described above, the olfactory environment sensor device 100 is described as a single device, but as one component of the fragrance generator 400 that discharges the fragrance generated from one or more fragrance modules or fragrance capsules mounted on the body. The bar may also include, the service providing apparatus 200, through the service providing unit 260, the fragrance scheduling information for improving the olfactory environment information, each olfactory environment sensor device 100 is mounted fragrance generator 400 It is possible to control the combination of each individual fragrance module and the intensity of scent diffusion by transmitting it to the communication module of Accordingly, the service providing apparatus 200 determines a combination of a fragrance module for improving the olfactory environment information and a fragrance diffusion intensity based on the olfactory environment information, and the fragrance generator 400 in which the olfactory environment sensor device 100 is mounted. ) can also be sent.

To this end, the service providing unit 260 may determine combination information and scent diffusing intensity information of one or more scent modules that compensate for the scent information classified in the classification result information, and scent scheduling information including the determined combination information and scent diffusing intensity information. By configuring, the olfactory environment sensor device 100 can be transmitted to the mounted fragrance generator 400 .

5 is a flowchart for explaining in more detail the operation of neural network learning and service provision of a service providing apparatus according to an embodiment of the present invention, and FIGS. 6 to 7 are experimental examples of reactive pattern data collected according to an embodiment of the present invention. A diagram showing the analysis.

First, referring to FIG. 5 , the service providing apparatus 200 according to an embodiment of the present invention provides a Recurrent Neural Network (RNN) method Long Short Short Circuit (LSTM) based on olfactory environment classification information mapped to reactive pattern learning data of a TVOC sensor. -Term Memory) learning is performed to build a first neural network learning model (S201).

In addition, the service providing apparatus 200 according to an embodiment of the present invention performs learning of a Convolutional Neural Network (CNN) method based on olfactory environment classification information mapped to reactive pattern learning data of a TVOC sensor to learn a second neural network learning model. to build (S203).

Thereafter, the service providing apparatus 200 according to an embodiment of the present invention learns the third neural network by performing FNN (Fully Connected Network) learning based on the olfactory environment classification information mapped to the reactive pattern learning data of the TVOC sensor. Build a model (S205).

In addition, the service providing apparatus 200 according to an embodiment of the present invention provides an output of the first neural network learning model and an output of the second neural network learning model according to the application of real-time responsive pattern information of the olfactory environment sensor device 100 . and the output of the third neural network learning model are combined and calculated to determine the olfactory classification result information of the olfactory environment information (S207).

Thereafter, the service providing apparatus 200 according to an embodiment of the present invention configures notification service information including preset guide behavior information in response to the olfactory classification result information ( S209 ).

6 and 7, as a problematic odor, mainly human feces

The learning data on odor, the smell of sweat from patients, and the smell of cigarettes were constructed based on the TVOC sensor output. Since the radius of spread of the odor is usually about 1m to 3m, the experiment was conducted using tobacco and ammonia in a water tank with a volume of 1m×1m×05m. The experimental water tank was partially opened to force a certain degree of convection, and through this, data similar to the actual environment could be obtained. Cigarettes lit for about 2 minutes, tissue paper soaked in ammonia, etc. were placed in the water tank containing the sensor, and the sensor data values were analyzed.

As shown in FIGS. 6 and 7 , it is possible to check the data measured after adding tobacco and ammonia, and after the initial value soars, the ECO2 and TVOC values fluctuate with a certain pattern in a specific section. can be checked Here, eCO2 is an equivalent CO2 value, and may indicate a CO2-based air quality predicted from TVOC, and the predicted ECO2 value may be provided through the above-described service interface.

The ECO2 or TVOC data of this time section may be composed of the above-described learning data for each time section, the first neural network model building unit 2511 , the second neural network model building unit 2513 , and the third neural network model building unit (2515) can be used for pattern analysis and learning using deep learning.

According to this test, after neural network learning processing on the training data, classification corresponding to each cigarette odor and ammonia odor pattern was performed, and, in particular, the second neural network model construction unit 2513 It was confirmed that the accuracy of 72.9% was calculated. It can be confirmed that the accuracy can be increased according to the accumulation of training data, the suitability of the neural network model for each feature, and the passage of time.

8 is an exemplary diagram of a service screen interface provided to a user terminal according to an embodiment of the present invention.

As shown in FIG. 8 , in particular, the service providing apparatus 200 according to an embodiment of the present invention may be usefully used in a system in which the cleanliness of a ward must be managed, such as a nursing hospital. For example, the mapping information of the olfactory environment sensor device 100 may include device identification information, hospital room information, and installed hospital bed (bed) location information, and such detailed information is transmitted from the user terminal 300 through the service screen interface. can be output. The user terminal 300 user may be a hospital facility manager or a terminal of a guardian using a management service.

The olfactory service screen interface output through FIG. 8 may include the above-described TVOC sensor information itself, and may also include ECO2 information predicted based thereon. According to the graph detail view input, sensor monitoring information for each ward and bed may be provided through the graph interface output.

And, as shown in FIG. 8 , the odor classification determination result for each neural network and the final classification result information may be intuitively output through the interface. For example, the first neural network model can discriminate ammonia of intensity 3 based on the odor intensity index classified as intensities 0 to 5 with 67% accuracy, and the second neural network model can determine ammonia of intensity 3 by 70 It can be determined with an accuracy of %, and the third neural network model can determine the cigarette smell of intensity 2 with an accuracy of 36%. The service providing apparatus 200 may configure the final result obtained by combining the calculations as text information and provide it through the olfactory service interface. For example, as illustrated in FIG. 8 , classification result information such as “the smell of ammonia is weak” may be provided to the user.

In addition, as shown in FIG. 8 , the behavior guide information corresponding to the classification result information may be provided through the service interface, which is to be configured based on the behavior guide information mapping table preset in the service providing device 200 . can For example, as shown in FIG. 8 , the action guide information can be output through the service interface to provide an action guide that a user or administrator can take immediately, such as "find and remove a pollutant, start a deodorizer or use a deodorant". .

The above-described method according to various embodiments of the present invention may be implemented as a program and provided to each server or device in a state stored in various non-transitory computer readable media. Accordingly, the user terminal 100 may access the server or device and download the program.

The non-transitory readable medium refers to a medium that stores data semi-permanently, rather than a medium that stores data for a short moment, such as a register, cache, memory, etc., and can be read by a device. Specifically, the above-described various applications or programs may be provided by being stored in a non-transitory readable medium such as a CD, DVD, hard disk, Blu-ray disk, USB, memory card, ROM, and the like.

In addition, although preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above, and the technical field to which the present invention belongs without departing from the gist of the present invention as claimed in the claims In addition, various modifications may be made by those of ordinary skill in the art, and these modifications should not be individually understood from the technical spirit or perspective of the present invention.

Claims (7)

In the service providing method,
registering mapping information of the user terminal and the olfactory environment sensor device;
collecting sensor information of the olfactory environment sensor device;
obtaining olfactory environment information using an olfactory information classification process by applying sensor information of the olfactory environment sensor device to a pre-machine-learned olfactory information neural network model;
determining the olfactory classification result information by combining the olfactory environment information according to a preset process, configuring an olfactory environment service interface based on the olfactory classification result information, and transmitting the olfactory environment service interface to a user terminal corresponding to the olfactory environment sensor device; containing
HOW TO PROVIDE SERVICES. The method of claim 1,
The method further comprising the step of determining a combination of a fragrance module for improving the olfactory environment information and a scent diffusing intensity based on the olfactory environment information, and transmitting the olfactory environment sensor device to a mounted fragrance generator
HOW TO PROVIDE SERVICES. The method of claim 1,
The mapping information of the olfactory environment sensor device is,
Including device identification information, location information, and installation information of the olfactory environment sensor device
HOW TO PROVIDE SERVICES. The method of claim 1,
The olfactory environment sensor device,
As a single sensor module that measures the reactive pattern information of volatile organic compounds, it is equipped with a TVOC (Total Volatile Organic Compunds) sensor.
HOW TO PROVIDE SERVICES. 5. The method of claim 4,
The step of obtaining the olfactory environment information includes:
Using the concentration responsive pattern information according to the TVOC sensor output, one or more olfactory information neural network models are pre-built, and olfactory environment information is obtained by applying the responsive pattern information collected in real time from the TVOC sensor to the one or more neural network models. doing
HOW TO PROVIDE SERVICES. 6. The method of claim 5,
The one or more pre-machine-learned olfactory information neural network models are
A first neural network obtained by performing Recurrent Neural Network (RNN) long short-term memory (LSTM) learning based on the reactivity pattern information of the TVOC sensor and olfactory environment classification information mapped to the reactivity pattern information of the TVOC sensor learning model,
A second neural network learning model obtained by performing convolutional neural network (CNN) learning based on the olfactory environment classification information mapped to the responsive pattern information of the TVOC sensor and the responsive pattern information of the TVOC sensor;
A third neural network learning model obtained by performing FNN (Fully Connected Network) learning based on the olfactory environment classification information mapped to the responsive pattern information of the TVOC sensor and the responsive pattern information of the TVOC sensor.
HOW TO PROVIDE SERVICES. 7. The method of claim 6,
The step of obtaining the olfactory environment information includes:
To determine the olfactory classification result information of the olfactory environment information by summing the output of the first neural network learning model, the output of the second neural network learning model, and the output of the third neural network learning model according to the application of the reactive pattern information including steps
How the service providing device works.

Images