KR102476825B1 - Method and apparatus for providing IoT service based on data platform - Google Patents

Abstract

An IoT service providing method based on a data platform according to an embodiment of the present invention includes the step of receiving, by an IoT service providing device, a plurality of types of sensing data from a plurality of IoT devices, and the IoT service providing device using different types of sensing data. Analyzing the correlation/correlation between the sensed data and determining the weight of each type of the sensed data by the IoT service providing device using the result of analyzing the correlation/correlation, and providing the IoT service The method may include providing a first IoT service by using the sensing data to which the weight is reflected.

Description

Method and apparatus for providing IoT service based on data platform {Method and apparatus for providing IoT service based on data platform}

The present invention relates to a method and apparatus for providing an IoT service based on a data platform. More specifically, a method for maximizing the usability and availability of IoT services by using a data platform capable of receiving sensing data from one or more IoT devices and converging the received sensing data, and performing the method It's about the device.

As the spread and use of IoT devices have recently been activated, the types of IoT services that can be provided using IoT devices are also diversifying. In other words, as IoT technology spreads in various fields such as medical care, logistics, energy, social safety, and smart home, many IoT service providers develop IoT applications for their services and operate services.

However, most conventional IoT devices have been developed with a focus on one activity. For example, each manufacturer has developed IoT devices to provide these services by focusing on one service, such as sleeping, exercising, or watching TV. As a result, the IoT environment has been built around the IoT devices made by each manufacturer.

For example, even though IoT devices are intended to provide the same service, the built-in sensors and collected data may be different depending on the manufacturer. This may cause a problem of duplicate data collection or different recognition of the same situation. For example, suppose that one user uses a first IoT device and a second IoT device made by different manufacturers for sleep management. At this time, the analysis result of the first IoT device and the analysis result of the second IoT device may come out different. This may cause limitations in terms of usability for users.

In addition, since some IoT devices do not share data even though they have the same sensors and collect the same data, if a problem occurs in a specific IoT device, the IoT service provided by the IoT device may not be provided smoothly. have. This can create limitations in terms of usability for users.

For example, suppose that one user uses a first IoT device for sleep management and a second IoT device for exercise management. At this time, even if the sensors embedded in the first IoT device and the second IoT device are the same and the collected information is the same, if the first IoT device fails, the sleep management service cannot be received. If the data collected by the first IoT device and the second IoT device can be shared, even if the first IoT device fails, a sleep management service may be provided based on the data collected by the second IoT device.

Accordingly, there is a need for a method to share and converge the data collected by each IoT device centering on the IoT service, rather than the IoT environment centered on the IoT device created by each manufacturer.

A technical problem to be solved by the present invention is to provide a method and apparatus for providing an IoT service based on a data platform.

The technical problems of the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the description below.

An IoT service providing method based on a data platform according to an aspect of the present invention for solving the above technical problem includes receiving, by an IoT service providing device, a plurality of types of sensing data from a plurality of IoT devices, and providing the IoT service. Analyzing, by a device, a correlation/correlation relationship between the different types of the sensing data, and determining a weight of each type of the sensing data by using a result of the IoT service providing device analyzing the correlation/correlation relationship and providing, by the IoT service providing device, a first IoT service by using the sensing data to which the weight is reflected.

In an embodiment, analyzing the correlation/correlation relationship may include applying logistic regression analysis and principal component analysis to the sensing data of different types.

In another embodiment, determining the weight of the sensing data may include obtaining a proportional coefficient of each type of the sensing data and determining the weight using the proportional coefficient.

In another embodiment, determining the weight by using the proportional coefficient may include determining a reciprocal of the proportional coefficient as the weight.

In another embodiment, the providing of the first IoT service may include, when reception of the first type of sensing data among the plurality of types of sensing data is omitted, a correlation/correlation relationship with the first type of sensing data. It may include providing a first IoT service requiring reference to the first type of sensing data by using another type of sensing data having .

In another embodiment, the providing of the first IoT service requiring reference to the first type of sensing data may include restoring a value of the first type of sensing data using the other type of sensing data. It may include generating sensing data.

In another embodiment, the plurality of IoT devices include a first IoT device to an n-th IoT device providing the first IoT service, and providing the first IoT service includes the first IoT device. and providing the first IoT service by summing sensing data measured by nth IoT devices according to the weight.

Conventionally, if IoT services are managed on a 1:N basis centering on IoT devices, according to some embodiments of the present invention, IoT devices are managed on a 1:N basis on IoT services. That is, since sensing data collected from different IoT devices can be shared and converged centering on IoT service, the usability of data collected by IoT devices can be increased.

In addition, since IoT devices are managed 1:N centering on IoT services, even if a failure occurs in a specific IoT device, IoT services can be provided to users using other IoT devices. Since IoT services can be provided to users reliably, availability can be secured.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

1 is a diagram for explaining the number of wearable devices that can be worn for each body part.
2 is a diagram for explaining data collected by a wearable device that can be worn for each body part.
3 to 5 are exemplary diagrams for explaining data collected by IoT devices and IoT services provided by IoT devices.
6 is an exemplary diagram for explaining a method of providing an IoT service based on a data platform according to an embodiment of the present invention.
7A to 7B are diagrams for explaining a method for securing availability in an IoT service providing method based on a data platform according to an embodiment of the present invention.
8A to 8C are diagrams for explaining regression analysis that can be used in one embodiment of the present invention.
9 is a diagram for explaining weights between sensing data that may be used in an embodiment of the present invention.
10A to 11B are diagrams for explaining a logistic function of logistic regression analysis that can be used in an embodiment of the present invention.
12 is a block diagram of an IoT service providing device based on a data platform according to an embodiment of the present invention.
13 is a flowchart of an IoT service providing method based on a data platform according to an embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Advantages and features of the present invention, and methods for achieving them, will become clear with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and only the present embodiments make the disclosure of the present invention complete, and common knowledge in the art to which the present invention belongs. It is provided to fully inform the holder of the scope of the invention, and the present invention is only defined by the scope of the claims. Like reference numbers designate like elements throughout the specification.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used in a meaning commonly understood by those of ordinary skill in the art to which the present invention belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined. Terminology used herein is for describing the embodiments and is not intended to limit the present invention. In this specification, singular forms also include plural forms unless specifically stated otherwise in a phrase.

As used herein, "comprises" and/or "comprising" means that a stated component, step, operation, and/or element is the presence of one or more other components, steps, operations, and/or elements. or do not rule out additions.

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

1 is a diagram for explaining the number of wearable devices that can be worn for each body part.

Referring to FIG. 1 , it can be seen that the number of wearable devices that can be mounted from head to toe in a human shape is displayed for each part. The figures in Figure 1 are extracted from statistical data as of the end of 2014. The drawing of FIG. 1 is briefly summarized in a table as follows. As can be seen from the table, IoT devices that can be worn on the head and wrist are most active.

body part number of wearable devices Head 42 Neck 11 Chest 7 pieces Torso 12 Arms 8 Wrist 85 Hands 5 pieces Waist 9 Legs 7 pieces Feet 8

As illustrated in FIG. 1 , one of the most commercialized fields among IoT devices is a field related to wearable devices, which measures a person's activity and provides customized services. For example, Fitbit's product line, Samsung's gear series and smart band, and Sony's smart watch and smart band products.

There are many IoT devices on the market, but each manufacturer and product are different, so there is no service or product that combines them. Accordingly, if IoT services can be provided by building a data platform that collects sensing data collected by IoT devices and acting as a kind of ecosystem, the usability and availability of IoT services can be strengthened.

In FIG. 1, wearable devices are mainly displayed, but IoT devices targeted by the data platform of the present invention are not limited to wearable devices. In addition, IoT devices installed for smart homes at home or IoT devices installed for automation of production processes in factories may also be targets. Hereinafter, for convenience of understanding, the data platform will be described with a focus on wearable devices that are currently commercialized.

2 is a diagram for explaining data collected by a wearable device that can be worn for each body part.

Referring to FIG. 2 , there are various wearable devices that can be attached to each body part, and information collected by each can be seen here. For example, a device that can be worn on the head can collect data on sleep time by detecting brain waves, and a device that can be worn on the neck can detect when a user suddenly falls down (fall detection). have.

Bands that can be worn on the arm can also measure blood pressure or calories. In addition, bands that can be worn on the wrist may measure stress or total activity. In addition, a patch-type device attached to the chest area can measure electrocardiography (ECG).

In addition, a device that can be worn on the fingertip can measure oxygen saturation (SpO2), a device that can be worn on the waist can measure skin temperature, and a device that can be worn in a pants pocket can measure the user's steps with a pedometer. You may. In this way, various information can be collected through various IoT devices.

In order to collect data from IoT devices, sensors must be built in. In some cases, a single IoT device may have multiple sensors embedded in it. Looking at the sensors of some commercialized products, Table 2 below shows them.

IoT devices built-in sensor samsung gear fit Accelerometer, Gyro Sensor, Heart Rate Sensor samsung gear s Accelerometer, Gyro Sensor, Altitude Sensor, GPS, Heart Rate Sensor, Ambient Light Sensor, UV Sensor, Fitbit's wristband (Flex) 3-axis accelerometer Fitbit's smart watch (Surge) GPS, 3-axis accelerometer, 3-axis gyroscope, digital compass, optical heart rate monitor, altimeter, ambient light sensor

As can be seen in Table 2, some IoT devices have some overlapping built-in sensors. That is, data collected by IoT device may partially overlap. However, when providing IoT services using collected data, there are few cases in which data is shared even though the data collected is overlapped. As a result, the utilization of data is often poor.

3 to 5 are exemplary diagrams for explaining data collected by IoT devices and IoT services provided by IoT devices.

Referring to FIG. 3 , it can be seen that three IoT devices are illustrated. One is a band 110 that can be worn on a wrist, the other is a clip 120 that can be worn in a pocket of clothes, and the other is a watch 130 that can be worn on a wrist. The IoT device illustrated in FIG. 3 is based on the premise that one user uses a plurality of IoT devices.

In the case of the band 110, the collected data are body temperature and heart rate, and exercise management 220 is provided as a service using these data. For example, there may be a service for managing exercise time or calories consumed during exercise. In the case of the clip 120, the collected data is sleep time and body temperature, sleep management 210 is provided as a service based on sleep time, and exercise management 220 is provided as a service based on body temperature. In the case of the watch 130, the collected data is pulse rate, blood flow, and decibels, and exercise management 220 is provided as a service based on pulse rate and blood flow, and music control 230 is provided as a service based on decibels.

In the example of FIG. 3 , collected data is different from each other, but services provided are similar in many parts. All of the IoT devices of the band 110, clip 120, and watch 130 provide exercise management 220 by default, and in the case of the clip 120, sleep management 210 is additionally added, and in the case of the watch 130 It can be seen to further provide additional music controls 230 .

However, as in the example of FIG. 3, there is no function of sharing collected data with each other by communicating with each IoT device. It only collects data for each device and provides services independently of other IoT devices. As a result, the exercise management 220 service provided by the band 110 and the exercise management 220 service provided by the clip 120 may show different results, and conversely, if the clip 120 is broken In some cases, the sleep management (210) service cannot be received.

Like the boundary line shown around each device in FIG. 3, IoT devices and IoT services have conventionally been provided to users with a 1:N relationship centered on IoT devices. If IoT services and IoT devices are provided to users in a 1:N relationship centering on IoT services, user convenience will be further enhanced.

FIG. 4 is a re-display of FIG. 3 by dividing it into sensor measurement and service provision parts. Referring to FIG. 4 , it can be seen which IoT device provides the corresponding service for each IoT service. That is, if FIG. 3 is a drawing centered on IoT devices, FIG. 4 is a diagram drawn centering on IoT services.

Referring to FIG. 4 , only the clip 120 provides the sleep management 210 service. Similarly, the music control 230 service is provided only by the watch 130 . The sleep management 210 service refers to, for example, a function of automatically analyzing sleep time to check sleepiness, tossing and waking up, and providing an alarm with vibration. For example, the music control 230 service refers to a function of measuring the level of ambient noise and providing selection, play, or volume control of music in the smartphone.

The exercise management 220 service is an IoT service provided by the band 110, the clip 120, and the watch 130. The exercise management 220 service refers to a function of determining a type of exercise, such as running or cycling, measuring calorie consumption, and a function of setting an exercise goal and providing whether or not it has been achieved.

However, the exercise management 220 service is an IoT service provided by the band 110, the clip 120, and the watch 130, but the sensing data used by each IoT device is different. In the case of the band 110, the exercise management 220 service is provided using body temperature and heart rate, and in the case of the clip 120, only body temperature is used, and in the case of the watch 130, exercise management 220 using pulse and blood flow. ) to provide services.

Although there are several IoT devices that provide the exercise management 220 service, accuracy is inevitably different because each service is provided. In order to provide the exercise management 220 service, usability will increase if the IoT service can be provided by considering all of the body temperature, heart rate, pulse rate, and blood flow collected by each IoT device.

Assume that the clip 120 is out of order among the three IoT devices as illustrated in FIG. 5 . In this case, among the data collected by the clip 120, data related to sleep time cannot be used. Of course, since the clip 120 collects the body temperature, if the clip 120 breaks down, the body temperature collected by the clip 120 cannot be used, but there is no problem because the band 110 collects the body temperature.

If the clip 120 is broken in this way, there is no problem with other IoT services, but the sleep management 210 service cannot be used. At this time, if another device can collect the sleep time or collect information that can replace the sleep time, the sleep management 210 service can be provided to the user even if the clip 120 is broken. . In other words, the availability of IoT services will increase.

Examples of conventional IoT devices and IoT services provided by the IoT devices were reviewed through FIGS. 3 to 5 . The present invention proposes an IoT service providing method using a data platform in order to increase the usability and availability of the IoT service. A data platform used in an embodiment of the present invention will be reviewed through FIG. 6 .

6 is an exemplary diagram for explaining a method of providing an IoT service based on a data platform according to an embodiment of the present invention.

Conventionally, as shown in the example of FIG. 3, the IoT device and the IoT service have a 1:N relationship, that is, the IoT service is provided centered on the IoT device. In contrast, the present invention further includes a component having a mediation function called a data platform 300 between an IoT device and an IoT service in order to build an IoT environment centered on the IoT service.

The data platform 300 serves to collect and process the information collected by the IoT device 100 and provide it to the IoT service 200. If the conventional FIG. 3 is a pipeline structure, FIG. 6 may be regarded as a structure like an hourglass. Information collected by each IoT device 100 is converged in the data platform 300 and provided to the IoT service 200, thereby ensuring usability and availability.

7A to 7B are diagrams for explaining a method for securing availability in an IoT service providing method based on a data platform according to an embodiment of the present invention.

Referring to FIG. 7A , it can be seen that correlation/correlation between data collected and transmitted by each IoT device within the data platform 300 is analyzed. That is, it can be known from the data platform 300 that sleep time can be estimated using body temperature and decibels, and heart rate can be estimated using pulse and blood flow rate.

For example, if the body temperature is constant and the decibel is low, it is considered that a person is in a sleep state, and the sleep time can be calculated by analyzing the time when the body temperature is kept constant and the decibel is low. Through this, sleep time cannot be directly measured because there is no sensor for measuring brain waves, but sleep time can be measured indirectly using a sensor for measuring body temperature and decibels.

Of course, if there is a clip 120 device capable of directly measuring sleep time through brain wave analysis, etc., it is not necessary to indirectly measure sleep time using body temperature and decibels. However, as described in FIG. 5 above, when the user wants to prevent a situation in which the user cannot receive the sleep management 210 service when the clip 120 is broken, such a correlation/correlation relationship can be used.

Referring to FIG. 7B , as in the case of FIG. 5 , it can be seen that data directly measured for sleep time cannot be used because the clip 120 device is out of order. Nevertheless, in the case of FIG. 7B , the sleep time can be virtually calculated and utilized using the body temperature and decibels. That is, the sleep management service 210 may be provided to the user using the indirectly estimated sleep time.

As seen in Table 2, considering that there are many overlapping sensors embedded in IoT devices, IoT services can be equally applied to users by utilizing sensor information from other devices despite the failure of one device. . To this end, the data platform 300 used in an embodiment of the present invention receives sensing data from each IoT device and analyzes correlation/relationship.

In other words, by analyzing the dependencies between data, even if a situation in which one data is unavailable occurs, the unavailable data can be restored from other data with dependencies. That is, in order to provide the IoT service, as illustrated in FIG. 7B , a virtual sleep time node may be created and information of the virtual sleep time node may be utilized.

From the user's point of view, even if any one IoT device fails, the IoT service 200 can be provided without problems, so the availability of the IoT service 200 can be increased. In the prior art, if the IoT device was centered, the present invention is based on the IoT service, so the service can be stably provided by reducing the dependence of the IoT service on the IoT device.

To do this, it is necessary to know the correlation/correlation relationship between each sensed data. The data platform 300 of the present invention may utilize logistic regression analysis to analyze the correlation/correlation between each sensed data. In this regard, it will be examined in more detail in FIGS. 8A to 8C.

8A to 8C are diagrams for explaining logistic regression analysis that may be used in an embodiment of the present invention.

Logistic regression is a probability model proposed by DRCox in 1958 and is a statistical technique used to predict the probability of occurrence of an event using a linear combination of independent variables. For more information related to logistic regression, visit the wiki website https://en.wikipedia.org/wiki/Logistic_regression .

Referring to FIG. 8A , a process of analyzing correlation/correlation between sleep time, body temperature, and decibels can be seen. When the sleep time cannot be directly measured due to the failure of the clip 120, the sleep time can be estimated using body temperature and decibels. First, the body temperature and decibel coefficients, which form the main components, are calculated. That is, the correlation/relationship between the body temperature and the decibel is analyzed first.

Next, referring to FIG. 8B , it can be seen that the correlation/correlation between the sleeping time and the main component of the relationship between the body temperature and the decibel obtained in FIG. 8A is analyzed again. That is, inactive data may be inferred by first analyzing correlation/correlation for data other than the sleep time to be estimated, and projecting it onto an inactive data axis to be extracted.

Referring to FIG. 8C , it can be seen that the correlation/relationship between body temperature, decibels, and sleep time is expressed as one main component on a three-dimensional plane composed of body temperature, decibels, and sleep time. The accuracy of this correlation/correlation analysis increases as the amount of accumulated data increases. That is, as the data of body temperature, decibels, and sleep time accumulate, previous data can be analyzed to continuously update the correlation/relationship.

9 is a diagram for explaining weights between sensing data that may be used in an embodiment of the present invention.

Referring to FIG. 9 , it can be seen that the weights of each variable are displayed through correlation/relationship analysis among the four variables body temperature, heart rate, pulse rate, and blood flow to provide the exercise management 220 service. For convenience, the weight is displayed by scaling so that the sum becomes 1.

The weight between each variable can be obtained in the process of analyzing the correlation/correlation relationship of each variable. For example, when a correlation/correlation relationship between variable a and variable b is analyzed, the correlation/correlation relationship may have a ratio of 2:1. That is, the degree of influence on a specific indicator can be evaluated as the same as the degree to which a increases in the value of 2 and the degree to which b increases in the value of 1. In this case, since a smaller change in b has a larger effect, the reciprocal of the proportionality can be used as the weight. That is, the weight of a can be 1 and the weight of b can be 2.

By using this data platform, it is possible to provide an integrated IoT service by defining the degree of association and weight between sensing data. Since the IoT service is provided by reflecting all the sensing data related to a specific IoT service, the sense of difference between the data collected by the IoT device can be resolved.

Referring to FIG. 9 , in order to provide an exercise management service (220), the body temperature is weighted at 0.1, the heart rate is weighted at 0.5, the pulse rate is weighted at 0.3, and the blood flow is weighted at 0.2, and exercise management (220 ) service can be seen. Conventionally, as in the example of FIG. 3, the band 110 provides exercise management 220 service only with body temperature and heart rate, and the clip 120 provides exercise management 220 service only with body temperature. In the present invention, each IoT device IoT service is provided by integrating the data of Through this, the usability of IoT services can be improved.

10A to 11B are diagrams for explaining a logistic function of logistic regression analysis that can be used in an embodiment of the present invention.

The logistic function refers to the result of transforming the value of the dependent variable to be within the range [0, 1] regardless of the independent variable. In this case, the reliability of the dependent variable may vary according to the range of the independent variable. In the case of FIG. 10A , when estimating the sleep time with the main components of body temperature and decibels, it can be seen that the value of the sigmoid function changes based on a threshold of 0.5.

That is, when the main component values of body temperature and decibel are in the range of 2 and 8, as shown in FIG. 10A, sleep is determined to be normal, and reliability at this time is high. Conversely, as in the case of FIG. 10B, when the main component values of body temperature and decibel are in the range of -2 and -8, sleep is judged to be abnormal, and reliability is high even at this time.

However, when the threshold value is between 0.5 and -0.5, the reliability becomes relatively low. In the case of FIG. 11A, the main component values of body temperature and decibel are -4 and +6, and at this time, it can be seen that both the high reliability interval and the low reliability interval are included, and the ratio has many intervals with low reliability. That is, in the case of FIG. 11A , sleep time can be estimated using body temperature and decibels, but an estimated value with low reliability is obtained.

Compared to the case of FIG. 11A, in the case of FIG. 11B, the main components of body temperature and decibels are between 1h and 8, and sleep is judged as a transition phase. In this case, both the high reliability interval and the low reliability interval are included, but unlike the case of FIG. 11A, since the ratio of the high reliability interval is higher, it can be seen that the reliability is relatively higher than in the case of FIG. 11A.

That is, it is possible to estimate the value of another IoT device according to the correlation/relationship using the measurement value of another IoT device, and also provide reliability. The data platform 300 according to an embodiment of the present invention does not simply converge sensing data collected from each IoT device, but also provides reliability to secure all of the usability, availability, and reliability of IoT services.

12 is a block diagram of an IoT service providing device based on a data platform according to an embodiment of the present invention.

Referring to FIG. 12 , the IoT service providing device 10 based on the data platform of the present invention may include an IoT data receiving module 510, an IoT data convergence module 550, and an IoT service providing module 530.

The IoT data receiving module 510 serves to receive data measured using a sensor in each IoT device from a plurality of IoT devices (S1110). Next, correlation/correlation between the received data is analyzed (S1130). The process of analyzing correlation/relationship proceeds by transferring the received data to the IoT data convergence module 550.

The IoT data convergence module 550 analyzes correlation/correlation using the data received from the IoT data receiving module 510, calculates weights, and performs processes of obtaining reliability. First, principal components are extracted from among data requiring analysis (S1510). That is, the correlation/correlation between independent variables is first analyzed to lower the dimension to one variable, and then the correlation/correlation between the corresponding principal component and the dependent variable to be analyzed is obtained (S1530). The process of obtaining the correlation/correlation relationship while reducing the dimension of the independent variable was examined in detail with reference to FIGS. 8A to 8C.

Principal component analysis (PCA) is a statistical analysis method that reduces high-dimensional data to low-dimensional data. Principal component analysis can transform a sample into a sample with the same or lower dimensions than the original sample. If principal component analysis is used to approximate a set of three-dimensional or more data as a single straight line, correlation/correlation can be analyzed easily and efficiently.

In order to estimate inactive sensor data that cannot be used due to device failure, etc., the main component of the sample consisting of the remaining data is approximated with a single straight line and then projected onto the axis of the inactive sensor data. Coefficients of each sensor data can be calculated in the process of approximating the principal components with a single straight line. Inactive data can be inferred through the principal component projected onto the axis of the inactive sensor data.

After the principal components are analyzed to obtain the correlation/correlation between the principal components and the dependent variables that need to be analyzed, residuals are calculated based on the used principal components, and reliability is extracted and provided together with the estimated dependent variable values (S1550). In addition, since the processes of analyzing the correlation/correlation relationship so far have higher accuracy as data accumulates, principal component coefficients may be continuously updated by machine learning (S1570).

The task of analyzing the correlation/correlation between each sensed data of the IoT device may be performed in advance. Then, when a failure occurs in a specific IoT device, the sensing data collected by the IoT device can be processed and provided based on other sensing data. This can increase the availability of IoT services. In this process, user convenience can be improved by providing the reliability of IoT data that is converged and restored.

In addition, through correlation/relationship analysis, rather than providing a specific IoT service based on only one sensing data, the usability of the IoT service can be increased by collecting and providing the service by collecting the sensing data of other related IoT devices. In this process, user convenience can be enhanced by providing IoT services by reflecting the weight between each IoT sensing data.

The IoT service providing module 530 provides the IoT service based on the information provided by the IoT data convergence module 550 through correlation/relationship analysis (S1310). In this process, the service is provided by reflecting the weight between the sensing data (S1330). Conventionally, IoT service was provided based only on sensing data measured by each IoT device, but if the data platform 300 of the present invention is used, IoT service can be provided by collecting data collected by multiple IoT devices according to respective weights. can

13 is a flowchart of an IoT service providing method based on a data platform according to an embodiment of the present invention.

12 is a diagram of a general case in which IoT service is provided by collecting sensing data of IoT devices. If a specific IoT device fails in the state of FIG. 12, the IoT data that was lost due to the failure is virtually restored using the correlation/correlation relationship analyzed in advance by the IoT data convergence module 550 to provide IoT service. can

Referring to FIG. 13 , sensing data is collected from each IoT device (S2100). If the data of a specific IoT device is missing, it is checked whether data convergence is required (S2200). If data convergence is required, a virtual node is created to restore data of the missing IoT device based on data measured by other IoT devices (S2300).

Along with restoration, the reliability according to the confidence interval is calculated and provided together with the restored data (S2400). The IoT service providing module 530 provides the IoT service to the user with reference to the received restoration data and reliability (S2500). Through this, the availability and reliability of IoT services can be secured together.

Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art to which the present invention pertains can be implemented in other specific forms without changing the technical spirit or essential features of the present invention. you will be able to understand Therefore, the embodiments described above should be understood as illustrative in all respects and not limiting.

Claims (9)

Receiving, by an IoT service providing device, a plurality of types of sensing data from a plurality of IoT devices;
Analyzing, by the IoT service providing device, a correlation/relationship between the sensing data of different types;
determining, by the IoT service providing device, a weight of each type of the sensing data by using a result of analyzing the correlation/correlation relationship; and
Providing, by the IoT service providing device, a first IoT service using the sensing data to which the weight is reflected;
The step of analyzing, by the IoT service providing device, the correlation/relationship between the different types of the sensing data,
When the first type of sensing data is missing among the plurality of types of sensing data, the first sensing data and the second sensing data having a correlation/correlation relationship with the missing first type of sensing data among the plurality of types of sensing data. determining sensing data;
analyzing a correlation/correlation relationship between the determined first sensing data and the second sensing data, and projecting the analyzed result onto an axis of the missing first type of sensing data; and
estimating the missing sensing data of the first kind in the projected axis;
Determining the weight of the sensing data,
Obtaining a proportional coefficient of each type of the sensing data and determining the weight using the proportional coefficient,
A method of providing IoT services based on a data platform. According to claim 1,
Analyzing the correlation/correlation relationship,
Applying logistic regression analysis and principal component analysis to the different types of sensing data,
A method of providing IoT services based on a data platform. delete According to claim 1,
The step of determining the weight using the proportional coefficient,
Including the step of determining the reciprocal of the proportional coefficient as the weight,
A method of providing IoT services based on a data platform. According to claim 1,
Providing the first IoT service,
When reception of the first type of sensing data is omitted among the plurality of types of sensing data, other types of sensing data having a correlation/correlation relationship with the first type of sensing data are used to detect the first type of sensing data. Including the step of providing a first IoT service that requires reference to sensing data,
A method of providing IoT services based on a data platform. According to claim 5,
The step of providing a first IoT service requiring reference to the first type of sensing data,
Generating virtual sensing data obtained by restoring a value of the first type of sensing data using the other type of sensing data.
A method of providing IoT services based on a data platform. According to claim 6,
Generating the virtual sensing data,
Computing the reliability of the restored virtual sensing data using a logistic regression function;
The step of providing a first IoT service requiring reference to the first type of sensing data,
Including providing the above reliability together,
A method of providing IoT services based on a data platform. According to claim 7,
The step of calculating the reliability is,
Comprising the step of calculating the reliability using a ratio including an interval within a threshold value in the logistic regression function,
A method of providing IoT services based on a data platform. According to claim 1,
The plurality of IoT devices,
Including a first IoT device to an n-th IoT device providing the first IoT service,
Providing the first IoT service,
Providing the first IoT service by summing the sensing data measured by the first IoT device to the nth IoT device according to the weight,
A method of providing IoT services based on a data platform.

Images