KR20170126237A - System and method for recommending user's exercise intensity by analyzing and learning physiologic signal - Google Patents

Abstract

The present invention relates to a system and method for recommending exercise intensity for each user through physiologic signal analysis and learning. The system according to the present invention includes a user terminal which collects and analyzes the physiologic information of the user while the user take exercise, determines a physical anomaly state of the physiologic information collected from the user based on the boundary value of each step of the user stored in advance, and provides exercise intensity recommendation information to the user, and a server which configures a learning model for the corresponding user by analyzing the physiological information of the user, and provides the boundary value of each step calculated by applying the physiological information received from the user terminal to the learning model in a predetermined time unit, to the user terminal. Accordingly, the present invention can provide a convenient and stable exercise intensity recommendation service to the user.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system and method for recommending a user's exercise intensity by analyzing a biological signal,

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system and method for recommending exercise intensity for each user through bio-signal analysis and learning.

Social members are increasingly interested in health to improve their quality of life, and the demand for fitness clubs is increasing accordingly.

However, it often happens that you are in danger because of excessive exercise without considering your own physical condition. In this case, you may not be able to cope with the danger yourself unless you have someone to help you around.

Recently, smart devices that manage the health status of users are also emerging. Such a smart device stores the user's biometric information, analyzes the stored data, measures the exercise amount, and manages the body change or the target exercise amount based on the exercise.

However, most of the smart devices do not induce the exercise intensity to be adjusted by reflecting the user's physical condition in real time, but merely measure and manage the exercise amount completed by the user and present the target value.

Korean Patent Publication No. 10-2015-0043531

SUMMARY OF THE INVENTION It is an object of the present invention to provide an apparatus and method for setting a user's body abnormality state by variably setting a reference value for judging a user's body anomaly state by reflecting feedback data measured in real- To provide exercise intensity recommendation information corresponding to the exercise intensity recommendation information and the exercise intensity recommendation information.

The technical problems of the present invention are not limited to the above-mentioned technical problems, and other technical problems which are not mentioned can be understood by those skilled in the art from the following description.

According to another aspect of the present invention, there is provided an exercise intensity recommendation system for a user through bio-signal analysis and learning, the method comprising: collecting and analyzing biometric information of a user during a user's exercise; A user terminal for judging a physical anomaly state of biometric information collected from a user based on the user and providing exercise intensity recommendation information to a user, and a learning model for the user by analyzing the biometric information of the user, And a server for applying the calculated boundary value to the user terminal by applying the biometric information received from the user terminal to the learning model.

The threshold value of each step includes heart rate information corresponding to a plurality of steps classified according to a user's physical anomaly state.

Wherein the user terminal comprises: a sensor unit for measuring a user's biometric information when the user requests the exercise intensity recommendation service; an information analyzing unit for analyzing user's heart rate data from sensor data of the sensor unit; An abnormality determination unit for comparing a boundary value of each step of the user registered in the information DB to determine a user's body anomaly state, and a controller for providing step-by-step exercise intensity recommendation information corresponding to the user's body anomaly state .

The user terminal may further include an information management unit for transmitting the sensor data of the sensor unit to the server in a predetermined time unit, receiving the updated threshold values for each step based on the sensor data, and updating the health information DB .

And the information management unit adjusts the threshold value for each step registered in the health information DB based on the updated threshold value for each step.

The controller may output a strong alarm sound or a warning sound for a predetermined period of time if it is determined that the user is in a dangerous state as a result of the user's physical anomaly state determination. If the alarm sound or the alarm sound is not released, And an emergency communication is attempted by the emergency.

The learning model is characterized in that a reference value of decision trees constituting a random forest classification model based on a random forest classification (RFC) model and a feedback model is an RFC-feedback mixed model in which user's sensor data is feedbacked and changed variably do.

The server receives the sensor data of the user's heart rate from the user terminal and labels it according to a predefined criterion to form a learning data set of the feedback model.

Wherein the server arbitrarily extracts a predetermined number of sample data from the learning data set of the feedback model and arranges the sampled data in a tree, and distributes the distribution data satisfying a predetermined condition through a Gini impurity of the sample data arranged in the tree, And determining a boundary value for each step for classifying a heart rate according to a user's physical anomaly state by learning a decision tree of the random forest classification model.

According to another aspect of the present invention, there is provided a method for recommending exercise intensity for each user through bio-signal analysis and learning, comprising the steps of: The user terminal analyzes the biometric information measured by the user while the user is exercising and compares the biometric information of the user with the boundary value of each step of the user registered in advance, The method comprising the steps of: the user terminal providing step-by-step exercise intensity recommendation information corresponding to the user's physical anomaly state; receiving, when the server receives biometric information of the user collected by the user terminal, Which is calculated by applying the biometric information to the learning model, The step of providing to the user terminal, and the user terminal, characterized by including the step of updating each step, the boundary values of the user registered in advance.

The threshold value of each step includes heart rate information corresponding to a plurality of steps classified according to a user's physical anomaly state.

The method according to the present invention may further comprise the steps of outputting a strong alarm sound or a warning sound for a predetermined period of time if it is determined that the user terminal is in a dangerous state as a result of the user's physical anomaly state determination, Or if the warning sound is not released, an attempt is made to make emergency contact with the pre-registered emergency network.

The learning model is characterized in that a reference value of decision trees constituting a random forest classification model based on a random forest classification (RFC) model and a feedback model is an RFC-feedback mixed model in which user's sensor data is feedbacked and changed variably do.

The method according to the present invention further comprises the step of, when the server receives the sensor data of the user's heart rate from the user terminal, labeling the received sensor data according to a predefined criterion to construct a learning data set of the feedback model And further comprising:

The method may further include the step of the server extracting a predetermined number of sample data from the learning data set of the feedback model and arranging the extracted sample data in a tree, calculating a genetic impurity degree of the sample data arranged in the tree Determining a boundary value for each step for classifying a heart rate according to a user's body anomaly state by learning a decision tree of the random forest classification model according to the extracted variance value, And further comprising:

According to the present invention, it is possible to variably set a reference value for determining a user's body anomaly state by reflecting feedback data measured in real time while the user is exercising, and thereby, based on a reference value that varies according to a body anomaly state, It is possible to provide the exercise intensity recommendation information corresponding to the abnormal state, thereby providing a convenient and stable exercise intensity recommendation service to the user.

FIG. 1 is a diagram showing a configuration of a user-specific intensity recommendation system through bio-signal analysis and learning according to the present invention.
FIG. 2 is a detailed block diagram of the user terminal of FIG. 1. Referring to FIG.
3 is a view showing a detailed configuration of the server of FIG.
FIG. 4 and FIG. 5 are views illustrating an embodiment to be referred to in describing a modeling operation of a user-specific intensity recommendation system through bio-signal analysis and learning according to the present invention.
6 to 8 are flowcharts illustrating a method of recommending a user-specific exercise intensity by analyzing and learning bio-signals according to the present invention.
9 is a diagram illustrating a configuration of a computing system to which an apparatus according to the present invention is applied.

Hereinafter, some embodiments of the present invention will be described in detail with reference to exemplary drawings. It should be noted that, in adding reference numerals to the constituent elements of the drawings, the same constituent elements are denoted by the same reference numerals even though they are shown in different drawings. In the following description of the embodiments of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the difference that the embodiments of the present invention are not conclusive.

In describing the components of the embodiment of the present invention, terms such as first, second, A, B, (a), and (b) may be used. These terms are intended to distinguish the constituent elements from other constituent elements, and the terms do not limit the nature, order or order of the constituent elements. Also, unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the relevant art and are to be interpreted in an ideal or overly formal sense unless explicitly defined in the present application Do not.

FIG. 1 is a diagram showing a configuration of a user-specific intensity recommendation system through bio-signal analysis and learning according to the present invention.

As shown in FIG. 1, the exercise intensity recommendation system for each user through bio-signal analysis and learning according to the present invention may include a user terminal 100 and a server 200.

The user terminal 100 may be a portable terminal that is provided with a sensor for collecting biometric information of a user, for example, heart rate information, and may be any terminal that supports communication with the server 200 .

For example, the user terminal 100 may be a smartphone, a tablet personal computer, a personal digital assistant (PDA), a portable multimedia player (PMP), an MP3 player, a wearable device, smart watch).

The user terminal 100 may analyze the user's biometric information collected through the sensor and provide the user with health information, alarm, or alarm according to the analysis result through the output means of the user terminal 100.

For example, the user terminal 100 may provide the user with exercise intensity recommendation information according to the user's heart rate.

Hereinafter, the detailed configuration of the user terminal 100 will be described in detail with reference to FIG.

The server 200 receives the sensor data of the user's biometric information from the user terminal 100 and analyzes the sensor data to provide the user terminal 100 with a stepwise boundary value corresponding to the user's biometric state.

Here, the server 200 receives the sensor data of the user's biometric information every predetermined period and learns the corresponding data to construct an RFC (Random Forest Classifier) -feedback mixed model, and transmits the RFC-feedback mixed model to the user terminal 100) may be applied to calculate the stepwise boundary value corresponding to the user's biological condition.

The server 200 provides the user terminal 100 with boundary values for each step of the calculated user. In this case, the user terminal 100 analyzes the biological condition of the user on the basis of the stepwise boundary values provided from the server 200. Hereinafter, the detailed configuration of the server 200 will be described in detail with reference to FIG.

FIG. 2 is a detailed block diagram of the user terminal of FIG. 1. Referring to FIG.

2, a user terminal 100 according to the present invention includes a control unit 110, a sensor unit 120, an interface unit 130, a communication unit 140, a storage unit 150, an information analysis unit 160, An abnormality determination unit 170, and an information management unit 180. Here, the control unit 110 may process signals transmitted between the respective units of the user terminal 100. [

The sensor unit 120 may include at least one sensor for collecting biometric information from a user's body in contact with the user's body. Here, the sensor may include a heart rate sensor that measures a user's heart rate.

For example, when the user performs the exercise using the treadmill, the sensor unit 120 may operate in the exercise intensity recommendation mode at the request of the user. At this time, the sensor unit 120 is started when the user terminal 100 is switched to the exercise intensity recommendation mode, and thus the user can measure the heart rate from the user's body during the exercise. Of course, the sensor unit 120 can be started when the predetermined button is operated by the user even though the sensor unit 120 is not in the exercise intensity recommendation mode.

The interface unit 130 may include input means for receiving a control command from a user and output means for outputting health state information of the user.

Here, the input means may correspond to a key button implemented outside the user terminal 100, and may correspond to a soft key implemented on the display. Further, the input means may be an input means such as a mouse, a joystick, a jog shuttle, or a stylus pen.

The output means may include a display on which biometric information and health status information measured by a user, an alarm or a longitude message, and the like are displayed, and may include a speaker for outputting an alarm sound or a warning sound or a vibration sensor for outputting an alarm through vibration .

Here, the display may be used as an input device in addition to an output device when a sensor for sensing a touch operation is provided. That is, when a touch sensor such as a touch film, a touch sheet, or a touch pad is provided on the display, the display operates as a touch screen. In this case, the input means and the output means can be integrated.

The display may be a liquid crystal display (LCD), a thin film transistor liquid crystal display (TFT LCD), an organic light-emitting diode (OLED), a flexible display, , A field emission display (FED), and a 3D display (3D display).

The communication unit 140 may include a communication module for wireless Internet access or short range communication. Here, the wireless Internet technology includes a wireless LAN (WLAN), a wireless broadband (Wibro), a Wi-Fi, a World Interoperability for Microwave Access (WIMAX), a High Speed Downlink Packet Access And may include Bluetooth, ZigBee, Ultra Wideband (UWB), Radio Frequency Identification (RFID), Infrared Data Association (IrDA), and the like.

The storage unit 150 stores data and programs necessary for the user terminal 100 to provide the user with the exercise intensity recommendation service.

In addition, the storage unit 150 may include a user information DB 151 storing unique information of a user who can identify the user, sensor data collected in real time by the sensor unit 120, and the like. The storage unit 150 may further include a health information DB 155 for receiving and storing the stepwise boundary values generated based on the sensor data corresponding to the user from the server 200. [

Here, the storage unit 150 may be a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (for example, an SD or XD memory A random access memory (SRAM), a read-only memory (ROM), a programmable read-only memory (PROM), an electrically erasable programmable read-only memory (EEPROM) Erasable Programmable Read-Only Memory).

The information analysis unit 160 analyzes sensor data collected by the sensor unit 120, for example, heart rate data, and provides the analysis result of the heart rate data to the abnormality determination unit 170. Here, the sensor data collected by the sensor unit 120 may be data collected while the user is exercising.

The abnormality determination unit 170 compares the heart rate data analysis result of the information analysis unit 160 with the threshold value of each step stored in the health information DB 155 to determine whether the user is abnormal.

Here, the boundary value for each step is calculated by the server 200 and can be classified into five levels according to the abnormal state corresponding to the user's heart rate, for example. The threshold value for each step may vary depending on the age and physical condition of the user.

For example, the threshold value for each step for a male in his thirties is less than 1 - 40 bpm, 2 - 100 bpm, 3 - 150 bpm, 4 - 190 bpm and 5 - 190 bpm .

As described above, the abnormality determination unit 170 classifies the abnormal state of the user into five levels as shown in Table 1 below based on the boundary values of the respective steps stored in the health information DB 155, To determine the current state of the user.

In Table 1, steps 1 and 5 are life-threatening stages because heart rate is too low or too high. Accordingly, the abnormality determination unit 170 transmits the result to the control unit 110 immediately after the data is generated, and the control unit 110 transmits a message containing the exercise intensity recommendation information such as "Stop exercise" And if you do not turn off the alarm or beep within a certain period of time, you can try to make an emergency call to the designated emergency.

The second stage is the stage where the user's heart rate is low and the exercise intensity is too weak. Accordingly, the abnormality determination unit 170 transmits the result to the control unit 110 when the data is generated, and the control unit 110 displays a message including the exercise intensity recommendation information such as " Vibration and the like.

Step 3 is a step where the user's heart rate is within the reference range and the strength of the exercise is appropriate. Accordingly, the abnormality determination unit 170 transmits the result to the control unit 110 during the data generation, and the control unit 110 outputs a message including the exercise intensity recommendation information "Maintain the current exercise intensity" can do.

Step 4 is the stage where the user's heart rate is higher than the reference range and the intensity of the exercise is high. Accordingly, the abnormality determination unit 170 transmits the result to the control unit 110 when the data is generated, and the control unit 110 outputs a message including the exercise intensity recommendation information such as "lower exercise intensity" Vibration and the like.

For example, if the average heart rate of the male user in the thirties and the average heart rate in the male user in the thirties are less than 40 bpm or 190 bpm or more as a result of analyzing the heart rate from the sensor data collected by the sensor unit 120, Or step 5, so it can be judged to be in a dangerous state.

On the other hand, when the heart rate is 100 bpm or more and less than 150 bpm, the abnormality determiner 170 can determine that the exercise intensity is appropriate because the user's heart rate corresponds to three levels. If the heart rate is 40 bpm or more and less than 100 bpm or 150 bpm or more and less than 190 bpm, the abnormality determiner 170 determines that the exercise intensity is somewhat weak or strong because the user's heart rate is 2 or 4 .

Accordingly, the abnormality determiner 170 can determine the abnormal state of the user and transmit the result to the controller 110. The control unit 110 may generate a message including the exercise intensity recommendation information corresponding to the body anomaly state of the corresponding step on the display of the interface unit 130 based on the result delivered from the abnormality determination unit 170 have.

If the abnormal state of the body is in a dangerous state in the first or fifth step as a result of the determination by the abnormality determination unit 170, the control unit 110 may output an alarm sound or a warning sound through the speaker. If the state in which the alarm sound or the alarm sound is output is not released until a predetermined time elapses, the control unit 110 attempts emergency contact with a registered contact, for example, a family member, a designated hospital,

For example, if a user in their thirties does not release the warning sound for one minute in a state where the heart rate is 190 bpm or more during the running exercise using the treadmill and a warning message is output together with the warning message, the control unit 110 It can be determined that the user is in a dangerous state and that there is no person to help the person around. Accordingly, the control unit 110 may attempt emergency contact through the pre-registered emergency network so that the user can escape from the dangerous state.

According to the embodiment, the controller 110 may cause the alarm sound to be output even when the body abnormality state is the second or fourth step.

The information management unit 180 manages data transmission / reception with the server 200.

The information management unit 180 calls the sensor data including the sensor data collected by the sensor unit 120 and the user information from the user information DB 151 and transmits the called user data to the communication unit 140 through the communication unit 140 To the server (200). At this time, the information management unit 180 may transmit the user data in units of a predetermined time, for example, 12 hours or 24 hours.

In addition, the information management unit 180 may receive the updated user health information based on the user data transmitted to the server 200. At this time, the information management unit 180 updates the updated user health information to the health information DB 155. Here, the user health information received by the server 200 analyzes the user's physical condition from the server 200 from the information included in the user data, and includes the boundary values (heart rate) of each step calculated by the analysis result do.

Accordingly, the abnormality determination unit 170 can determine the abnormal state of the user by applying the updated boundary value for each step.

For example, while a male user in his thirties is running at a speed of 10 km / h on a running machine, a message is output indicating that the heart rate corresponding to the fourth stage is 160 bpm and the intensity of the exercise is lowered by the control unit 110 The information management unit 180 transmits the sensor data of the day collected by the sensor unit 120 to the server 200, assuming that the user has continued to stay for about 10 minutes.

In this case, the server 200 confirms that the user continues the exercise for 10 minutes in the fourth step while analyzing the user data, and sets the boundary value of the user in the range of 150 bpm to 190 bpm and 155 bpm to 195 bpm And updates the health information DB 155 of the user by adjusting the threshold value of the 5th step to 195 bpm or more. The health information of the updated user is transmitted to the user terminal 100. Accordingly, the abnormality determiner 170 determines the abnormal state of the user by applying the updated boundary values of the fourth and fifth steps.

3 is a view showing a detailed configuration of the server of FIG.

Referring to FIG. 3, the server 200 may include a signal processing unit 210, a communication unit 220, a storage unit 230, and a data analysis unit 240. The signal processing unit 210 may process the signals transmitted between the respective units of the server 200. [

The communication unit 220 may include a communication module that supports wireless Internet communication or short-range communication with one or more user terminals 100. [ The communication module can support wireless LAN (WLAN), Wibro, Wi-Fi, WiMAX, HSDPA, etc. and supports Bluetooth, ZigBee, UWB, RFID, (IrDA).

Here, the communication unit 220 can receive the sensor data collected in the user terminal 100, that is, the heart rate data, on a predetermined time basis.

The storage unit 230 includes a user information DB 231 for storing user data received from one or more user terminals 100 and a health information DB 233 for analyzing user data and storing model information applied to calculate boundary values for each step, (235).

The user information DB 231 classifies and stores user information and sensor data included in user data received from each user terminal 100 for each user and stores boundary value information Can be stored.

The health information DB 235 may store an RFC-feedback mixed model, set values and algorithms for analyzing the user's sensor data and calculating boundary values for each step as an analysis result.

Here, the storage unit 230 may be a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (for example, SD or XD memory), magnetic memory, magnetic disk, optical disk, RAM, SRAM, a ROM, a PROM, and an EEPROM.

The data analyzer 240 learns the RFC-feedback mixed model using the feedback data and the sample data from the user terminal 100.

Here, the RFC-feedback mixed model refers to a model in which reference values of decision trees constituting an RFC based on a Random Forest Classifier (RFC) model and a feedback model are variably changed by feedback of user's sensor data.

The feedback model is feedback data on the sensor data collected by the sensor unit 120 of the user terminal 100, and an embodiment of the feedback data can be represented as shown in FIG. The feedback data for the user may be applied to a learning data set for learning the decision tree constituting the RFC.

The RFC model refers to a classification model that makes decisions by combining decisions of a plurality of decision trees. A decision tree refers to a data structure capable of classifying a given data through a plurality of hierarchical branches as shown in FIG. In the embodiment of the present invention, a decision tree that can be classified into five levels in total based on the heart rate is applied, but the present invention is not limited thereto.

The RFC extracts a subset of the given data to generate multiple decision trees. We classify all the decision trees that are generated when we classify the data, and adopt the method that adopts the largest number of results.

The data analysis unit 240 can learn the decision tree of the RFC model using the feedback data. At this time, the data analysis unit 240 receives the sensor data on the heart rate from the user terminal 100, and labels the received sensor data according to a predefined criterion to make a learning data set .

Thereafter, the data analyzer 240 randomly extracts a predetermined number of sample data from the learning data set, generates a plurality of decision trees, and repeats this process to learn the decision tree. At this time, the data analyzer 240 may calculate a Gini impurity of the sample data arranged in the decision tree through a learning process of a plurality of stages to find a variance value satisfying a predetermined condition.

The data analyzer 240 calculates the boundary value for each step for classifying the heart rate into five levels according to the abnormal state of the user through the learning process.

If it is determined that adjustment of the boundary value of four levels among the boundary values of each stage of the male user in their thirties is necessary, the data analysis unit 240 may determine that the four levels of sensor data on the heart rate Is determined as a reference value (T _L4 ), and the boundary values of the fourth step and the preceding and following steps are adjusted upward by the resultant value obtained by applying the predetermined reference value to the following equation (1).

Here, int (x) is a maximum integer smaller than x.

As shown in Equation (1), if the duration T _L4 of the fourth stage is less than 10, the data analyzer 240 increases the threshold values of the fourth stage and the preceding and following stages by int (T _L4 / 2) If T _L4 is 10 or more, the boundary value of the 4 steps and the preceding and following steps can be adjusted upward by 10.

The data analyzer 240 transmits the calculated boundary value of each step of the user to the user terminal 100.

Here, the data analyzer 240 updates the information stored in the user information DB 231 when the boundary value of each step calculated based on the sensor data of the user is different from the previously calculated value, and transmits the updated information to the user terminal 100 ).

The operation flow of the exercise intensity recommendation system for each user through bio-signal analysis and learning according to the present invention will be described in more detail as follows.

6 is a flowchart illustrating an operation of a user terminal according to the present invention.

Referring to FIG. 6, the user may be provided with the exercise intensity recommendation service by logging into the user terminal 100 before performing the exercise such as the treadmill.

When the user information for log-in of the exercise intensity recommendation service is input (S110), the user terminal 100 authenticates the user and constructs a user information DB corresponding to the authenticated user (S120). If there is a history in which the user has previously received the exercise intensity recommendation service, the 'S120' procedure may be omitted.

Upon completion of the user login, the user terminal 100 may collect and store sensor data on the heart rate of the user through the heart rate sensor while the user is exercising (S130).

Then, the user terminal 100 analyzes the user's body information from the sensor data stored in the step 'S130' (S140), and determines whether the user's body is abnormal (S150). In step 'S150', the user terminal 100 compares the result of analyzing the heart rate data with the threshold value of each step stored in the health information DB, based on the threshold values classified into five levels according to the abnormal state corresponding to the user's heart rate It is possible to judge whether or not the user is abnormal in the body.

Of course, when there is no history in which the user is provided with the exercise intensity recommendation service, the boundary value at each step may be defined as a default default value, and then, based on the sensor data collected by the sensors of the user terminal 100, .

If the user's body abnormality is determined to be normal in step S150, the user terminal 100 outputs a message indicating that the exercise intensity is appropriate to provide exercise intensity recommendation information to the user (S181). At this time, the user terminal 100 updates the user data on the user's body anomaly determination result (S185), and subsequently collects the sensor data (S130).

If it is determined that the user's body abnormality is not in the first or fifth step S160, that is, if the user's attention abnormality is in the second or fourth step, A message indicating that the exercise intensity is somewhat weak or strong is output to provide the user with exercise intensity recommendation information (S181). In this case, the user terminal 100 may output a vibration or a weak alarm at the time of message output. The user terminal 100 updates the user data on the user's body anomaly determination result (S185), and subsequently collects the sensor data (S130).

If it is determined in step S150 that the user is in the warning state in step 1 or step 5 (S160), the user terminal 100 displays a message indicating that the exercise intensity is too weak or too strong and dangerous, Vibration, an alarm sound, or a warning sound to provide the user with exercise intensity recommendation information (S160).

In the case of step 1 or step 5, if the alarm sound or the alarm sound is not released until the predetermined time? T elapses (S170, S180), the user terminal 100 may determine that the user is in a dangerous state. Accordingly, the user terminal 100 makes an emergency contact with the pre-registered emergency network to inform the user of the dangerous state.

On the other hand, if the alarm sound or the alarm sound output in the case of the first or fifth step is released before the predetermined time? T elapses (S170, S180), the user terminal 100 judges that the user is not in a dangerous state . In this case, the user terminal 100 generates and outputs a message related to the related information (S181), and the user terminal 100 updates the user data on the user's abnormal physical condition (S185) (S130).

Here, the updated user data can be transmitted to the server 200. In this case, since it is confirmed that the user is not in a dangerous state even in the warning state, the boundary value for each step of the heart rate according to the user's exercise state is transmitted to the server 200 ). ≪ / RTI >

FIG. 7 is a flowchart illustrating a data update operation in a user-specific intensity recommendation system through bio-signal analysis and learning according to the present invention.

Referring to FIG. 7, the user terminal 100 transmits user data including sensor data collected by the sensors and abnormal state information of the user to the server 200 as shown in FIG. 6 (S210). Here, the user terminal 100 may transmit user data to the server 200 at a predetermined time period.

Accordingly, the server 200 calls the learned RFC-feedback mixed model corresponding to the user (S220), and the user data received in the step 'S210' is transmitted to the RFC-feedback mixed model To calculate a boundary value (heart rate) for each step for determining a user's body anomaly state (S230).

In step 'S230', the server 200 may analyze the user data for any one step and calculate a boundary value for the step.

Accordingly, the server 200 can update the threshold value of each step registered in the health information DB corresponding to the user up or down based on the threshold value of each step calculated in the step 'S230' (S240).

If the boundary value of each step of the user is adjusted in step S240, the server 200 transmits the health information including the boundary value of each step of the user to the user terminal 100 in step S250.

The user terminal 100 updates the boundary value of each step registered in the health information DB of the user terminal 100 based on the health information received in the process of 'S250' (S260). Accordingly, the user terminal 100 can perform the processes of FIG. 6 based on the updated boundary values at step S260.

8 is a flowchart showing a model learning operation of the server according to the present invention.

Referring to FIG. 8, the server 200 previously determines the number N of trees and the number M of sample data to be used in each tree for constructing a decision tree of an RFC.

The server 200 configures a learning data set using feedback data collected from the user terminal 100, that is, sensor data (S305).

Then, the server 200 generates N empty trees (S310), extracts M sample data from the learning data set configured from the user's sensor data (S320), and extracts N empty trees generated in the process 'S310' (S330).

Thereafter, the server 200 selects a subset of the variables (S340), arranges the sample data distributed to the empty tree with respect to the variable selected in the process 'S340' (S350) (S360).

The server 200 may repeatedly perform the steps 'S340' to 'S360' for each variable, and may search for a variance value satisfying a predefined stop condition based on the genie impurity calculated at the step 'S360' do.

If the predefined suspension condition is satisfied (S370), the user terminal 100 terminates the learning of the corresponding tree, and then performs S320 to S370 for the other empty tree.

If there is no empty tree (S380), the RFC-feedback mixed model is constructed using the learned decision tree (S390).

The user terminal and / or the server according to the present embodiment operating as described above may be implemented as an independent hardware device. Meanwhile, the user terminal and / or the server according to the present exemplary embodiment may be driven by at least one processor included in another hardware device such as a microprocessor or a general-purpose computer system.

9 is a diagram illustrating a computing system to which an apparatus according to the present invention is applied.

9, a computing system 1000 includes at least one processor 1100, a memory 1300, a user interface input device 1400, a user interface output device 1500, (1600), and a network interface (1700).

The processor 1100 may be a central processing unit (CPU) or a memory device 1300 and / or a semiconductor device that performs processing for instructions stored in the storage 1600. Memory 1300 and storage 1600 may include various types of volatile or non-volatile storage media. For example, the memory 1300 may include a ROM (Read Only Memory) and a RAM (Random Access Memory).

Thus, the steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by processor 1100, or in a combination of the two. The software module may reside in a storage medium (i.e., memory 1300 and / or storage 1600) such as a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a removable disk, You may. An exemplary storage medium is coupled to the processor 1100, which can read information from, and write information to, the storage medium. Alternatively, the storage medium may be integral to the processor 1100. [ The processor and the storage medium may reside within an application specific integrated circuit (ASIC). The ASIC may reside within the user terminal. Alternatively, the processor and the storage medium may reside as discrete components in a user terminal.

The foregoing description is merely illustrative of the technical idea of the present invention, and various changes and modifications may be made by those skilled in the art without departing from the essential characteristics of the present invention.

Therefore, the embodiments disclosed in the present invention are intended to illustrate rather than limit the scope of the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents should be construed as falling within the scope of the present invention.

100: user terminal 110:
120: sensor unit 130: interface unit
140: communication unit 150:
151: user information DB 155: health information DB
160: information analysis unit 170: abnormal judgment unit
180: Information management unit 200: Server
210: signal processor 220:
230: storage unit 231: user information DB
235: health information DB 240: data analysis unit

Claims (1)

A user who collects and analyzes the user's biometric information while exercising the user and determines the physical anomaly state of the biometric information collected from the user based on the boundary value of each stage of the user stored in advance and provides the user with exercise intensity recommendation information Terminal; And
Analyzing the biometric information of the user to construct a learning model for the user, applying biometric information received from the user terminal to the learning model in a predetermined time unit, and providing the calculated boundary values to the user terminal server
And a motion intensity recommendation system for each user through bio-signal analysis and learning.

Images