JP6008572B2 - Exercise support system and exercise support method - Google Patents

Description

  The present invention relates to a system that supports a user's exercise, and more particularly, to a technique for presenting an appropriate exercise based on a predicted daily exercise amount.

  In order to prevent lifestyle-related diseases such as diabetes and hypertension, it is necessary to maintain exercise habits in daily life.

  Patent Document 1 describes physical activity amount recording means for recording the physical activity amount performed for physical activity exceeding the reference intensity so that the breakdown of the exercise amount and the daily activity amount can be understood, and the physical activity to be performed in a predetermined unit period. Based on the target storage means for storing the target of each of the amount and the amount of exercise, the amount of physical activity and the amount of exercise recorded by the physical activity amount recording means, and the target stored in the target storage means, and Disclosed is an exercise support device comprising display means for displaying a goal achievement level of each exercise amount.

  Patent Document 2 includes a sensor that measures momentum, a storage unit that accumulates the measured momentum and measurement date, and a display unit that displays information about the momentum, and achieves a probability of achieving a predetermined target momentum As a probability, an evaluation value storage unit that is stored in advance at each time as an evaluation value, which is an exercise amount integrated value corresponding to the achievement probability set in advance, and the exercise amount accumulated in the storage unit and the measurement date and time at predetermined time intervals A momentum integrating unit that calculates the momentum integrated value, and a comparison unit that compares the momentum integrated value and the evaluation value obtains an evaluation value corresponding to the current time at a predetermined timing, and A sensor device that is determined to be capable of achieving the target momentum is disclosed.

  Patent Document 3 discloses an action suggesting device that estimates human behavior based on biological information such as sleep time, number of steps, and amount of exercise, and presents improvement means for improving the health index and the like.

JP 2008-104758 A JP 2011-197890 A International Publication No. 2010/146811 Pamphlet

  In the pedometer disclosed in Patent Document 1, since only a pattern of a single average number of steps in the past is used, the wearer takes a pattern of behavior that varies from day to day, and the pattern of increase in the number of steps varies from day to day. In some cases, there is a problem that an error in the achievement level to be calculated becomes large.

  In the sensor device disclosed in Patent Document 2, when the behavior pattern is determined in advance by day of the week or the like, the accuracy of prediction is improved. However, when a non-periodic behavior pattern is shown, there is a problem that a prediction error becomes large.

  In the sensor device disclosed in Patent Document 3, the type of action correlated with the target can be known, but the amount of action (exercise) necessary to actually achieve the target and the content of the action are presented. There was a problem not to be.

  Therefore, the present invention provides an exercise support system that accurately predicts the amount of exercise.

A typical example of the invention disclosed in the present application is as follows. In other words, exercise support system includes a sensor terminal that the user wears, a movement support system Ru and a communicatively coupled computer with the sensor terminal, said sensor terminal, having at least an acceleration sensor, wherein The computer calculates time series data of momentum of a user wearing the sensor terminal based on time series data of acceleration detected by the acceleration sensor, and based on the time series data of the calculated momentum, the sensor The total daily momentum of the user wearing the sensor terminal is estimated by time-series estimation of the behavior of the user wearing the terminal and the multivariate analysis using the ratio of the estimated time-series behavior within a predetermined time as a parameter. Predict.

  According to the representative embodiment of the present invention, the momentum can be accurately predicted.

It is a block diagram which shows the structure of the exercise assistance system of embodiment of this invention. It is a block diagram which shows the structure of the sensor device of embodiment of this invention. It is a block diagram which shows the structure of the portable information terminal of embodiment of this invention. It is a block diagram which shows the structure of the server of embodiment of this invention. It is a figure explaining the structure of the sensor data of embodiment of this invention. It is a figure explaining the structure of the momentum data of embodiment of this invention. It is a figure explaining the structure of the action content data of embodiment of this invention. It is a figure explaining the structure of the prediction exercise amount data of embodiment of this invention. It is a figure explaining the structure of the user information of embodiment of this invention. It is a figure explaining the example of the exercise assistance monitor screen of embodiment of this invention. It is a figure explaining the relationship between the walking speed and momentum used by embodiment of this invention. It is a figure explaining the example of the living activity monitor screen of embodiment of this invention. It is a figure explaining the example of the weight monitor screen of embodiment of this invention. It is a figure explaining the example of the exercise assistance monitor screen of embodiment of this invention. It is a sequence diagram of a process in the exercise support system of the embodiment of the present invention. It is a flowchart of the process of the measurement program of embodiment of this invention. It is a flowchart of the process of the action analysis program of embodiment of this invention. It is a flowchart of a process of the all-day exercise amount prediction program of embodiment of this invention. It is a figure explaining the result of having predicted the exercise | movement amount of the day by embodiment of this invention. It is a figure explaining the comparison result of the accuracy of prediction shown in FIG. It is a flowchart of a process of the exercise content recommendation program of embodiment of this invention.

  FIG. 1 is a block diagram showing a configuration of an exercise support system according to an embodiment of the present invention.

  The exercise support system according to the present embodiment provides information via a sensor device 1 worn by a user 4, a portable information terminal 2 communicating with the sensor device 1, and a wireless base station 5 connected to a network (for example, the Internet) 6. A server 3 that communicates with the terminal 2; In addition, the exercise support system according to the present embodiment transmits sensor data from the ultra-compact and low-power sensor device 1 that is slow to process but always worn by a person to the large-sized server 3 that has high processing capacity and high storage capacity. By collecting, large-scale analysis can be processed at high speed.

  The sensor device 1 measures the movement of the user 4 and biological information by being worn by the user 4 at all times or during the main activity time of the daytime. The sensor device 1 transmits the measured information to the portable information terminal 2 by short-range wireless communication such as Bluetooth (Bluetooth is a registered trademark, the same applies hereinafter) or wired communication such as USB (Universal Serial Bus). Send. The mobile information terminal 2 is connected to the base station 5 at home or on the go by using public wireless communication (for example, WiMAX, 3G, LTE, etc.) or wireless LAN (WiFi) used in a mobile phone or the like, so that the sensor device 1 The information measured in (1) is transmitted to the server 3 connected to the network 6, and the information analyzed by the server 3 and the accumulated information are received.

  FIG. 2 is a block diagram showing the configuration of the sensor device 1.

  The sensor device 1 includes a microcomputer 101, an acceleration sensor 105, a display unit 106, an RTC (real time clock) 107, a short-range wireless communication unit 108, a flash memory 109, a power circuit 113, a battery 114, a charging circuit 115, a charging terminal 117, a communication Line 118 and power line 119 are included.

  The acceleration sensor 105 is a sensor that measures the movement and direction of a person. The display unit 106 is composed of a liquid crystal display panel, for example, and displays various types of information. An RTC (real time clock) 107 provides current time information and calendar information. The short-range wireless communication unit 108 communicates with the portable information terminal 2 by Bluetooth or the like. The flash memory 109 is a large-capacity nonvolatile memory element. The battery 114 is a power source for driving the sensor device 1. The power supply circuit 113 generates power necessary for each element of the sensor device 1 from the output of the battery 114. The charging circuit 115 is a circuit that charges the battery 114 with power supplied from the outside. The charging terminal 117 connects an external power source and the charging circuit 115. The communication line 118 is a communication path (for example, a bus) between elements. The power line 119 supplies power to each element.

  The microcomputer 101 executes a measurement program 103 and a sensor data transmission program 104 recorded in advance in a ROM 102 which is a nonvolatile storage element, and operates according to these programs.

  The measurement program 103 stores sensor data 112, which is information measured by the acceleration sensor 105, and a time stamp 111, which is time information at the time of measurement acquired from the RTC 107, in the time stamp storage of the sector 110, which is a rewrite unit of the flash memory 109. It stores in the area and sensor data storage area. In addition, the measurement program 103 controls and cuts off the power supplied from the power supply circuit 113 for each element except during necessary operations such as measurement, recording, and communication, thereby reducing power consumption. As a result, even if the battery 114 is small enough to be worn by a person, the sensor device 1 can continuously operate for a long time without being frequently charged.

  The sensor data transmission program 104 automatically determines the timing designated by the user 4 or the timing at which communication with the portable information terminal 2 is possible, starts communication, and uses the sensor data recorded in the flash memory 109 as the portable information terminal 2. Send to. Since a time stamp is added to the sensor data, even if it is transmitted after a lapse of time after measurement, the portable information terminal 2 and the server 3 that have received this information rearrange the sensor data in time series and record them. be able to.

  FIG. 3 is a block diagram showing the configuration of the portable information terminal 2.

  The portable information terminal 2 includes a processor (CPU) 201, a display unit 202, a sensor 203, a RAM 204, a power line 205, a short-range wireless communication unit 206, a wide-area wireless communication unit 207, a communication line 208, a battery 209, a power supply circuit 210, and a charging circuit. 211, a charging terminal 212, an input unit 213, an AUDIO circuit 217, and a sound output terminal 218.

  The processor 201 executes a program recorded in the flash memory 208, which is a large-capacity nonvolatile storage element, and controls other elements. The display unit 202 is configured by a liquid crystal display panel, for example, and displays information instructed by the program. The input unit 213 is an input element such as a keyboard or a touch panel that is operated by a user to instruct execution of a program or the like.

  The sensor 203 includes a magnetic sensor that detects the tilt of the portable information terminal 2, an acceleration sensor that detects movement, a temperature sensor that detects temperature, a pressure sensor that detects atmospheric pressure, a sensor that detects a position (for example, a GPS module), and the like. Including. The RAM 204 is a volatile storage element that temporarily records information necessary for processing by the processor 201. The short-range wireless communication unit 206 communicates with the sensor device 1 or the like by Bluetooth or the like. The wide area wireless communication unit 207 communicates with the server 3 by public wireless communication (for example, WiMAX, 3G, LTE, etc.) or wireless LAN (WiFi).

  The battery 209 is a power source for driving the portable information terminal 2. The power supply circuit 210 generates power necessary for each element of the portable information terminal 2 from the output of the battery 209. The charging circuit 211 is a circuit that charges the battery 209 with power supplied from the outside. The charging terminal 212 connects an external power source and the charging circuit 211. The communication line 208 is a communication path (for example, a bus) between elements. The power line 205 supplies power to each element.

  The AUDIO circuit 217 generates a sound to be output (for example, a walking pitch sound) according to an instruction from the processor 201. The sound output terminal 218 can be connected to a speaker (earphone), and outputs the sound generated by the AUDIO circuit 217 as an acoustic signal.

  The portable information terminal 2 relays the sensor data output from the sensor device 1 and transmits it to the server 3, and displays content such as data received from the server 3 via the network 6 on the display unit 202.

  The flash memory 208 stores a plurality of programs for controlling the operation of the portable information terminal 2. The sensor data transmission / reception program 213 controls the short-range wireless communication unit 206 and the wide-area wireless communication unit 207 to temporarily record the sensor data 220 received from the sensor device 1 in the sensor data storage area and then transmit it to the server 3. To do. As a result, even if a state in which the server 3 cannot be temporarily connected continues, sensor data is not lost.

  The exercise history / recommended content display program 214 receives from the server 3 the exercise amount and the action content history of the user 4 calculated by the server 3, and receives the received data as the action history 240, the exercise amount 230, the predicted exercise amount 250, and the recommended exercise. The content 260 is stored in the flash memory 208, and the data stored in the flash memory 208 is displayed on the display unit 202 in response to the operation of the user 4 or when the sensor 203 detects a predetermined situation. Since prediction of exercise amount and recommendation of exercise content require large-scale calculations such as regression analysis using a large amount of past data, processing should be performed by the server 3 having better processing capacity and storage capacity than the portable information terminal 2 Is desirable. However, since the portable information terminal 2 has the all-day exercise amount prediction program 215 and the exercise content recommendation program 216, the estimated exercise amount and the recommended exercise content can be calculated only by the portable information terminal 2.

  FIG. 4 is a block diagram showing the configuration of the server 3.

  The server 3 is a computer having a processor (CPU) 301, a RAM 302, a LAN communication unit 303, a power supply circuit 304, a storage 305, a communication line 311, and a power supply line 312. The server 3 can perform large-scale analysis processing using a large amount of data by the processor 301 that is faster than the sensor device 1 and the portable information terminal 2.

  The processor 301 executes a program stored in the RAM 302. The RAM 302 is a volatile storage element that temporarily records information necessary for processing by the processor 301. The LAN communication unit 306 connects the server 3 to the network 6, controls communication with other devices, and is connected to the network 6 through a communication cable 313.

  The power supply circuit 304 generates power to be supplied to each element. The power cable 314 supplies power supplied from the outside to the power circuit 304. The power line 312 supplies power to each element of the server 3. The communication line 311 is a communication path (for example, a bus) between elements. The storage 305 is a large-capacity nonvolatile storage device such as a magnetic disk device, and analyzes a program that executes arithmetic processing necessary to control the server 3, sensor data transmitted from the sensor device 1, and sensor data. Record the results.

  The sensor data reception program 306 receives the sensor data transmitted from the sensor device 1 via the portable information terminal 2 and the network 6 and stores the received sensor data 320 in the storage 305. The behavior analysis program 307 calculates the behavior type for each time of the user from the received sensor data, and stores the behavior content 340 and the amount of exercise 330 in the storage 305.

  The all-day exercise amount prediction program 308 uses the ratio of the action content 340 up to a certain point in time, the combination of the action content 340, the change pattern of the exercise amount 330, the change pattern of the number of steps, etc. Learn the relationship with the amount of momentum achieved in advance by regression analysis. Through this regression analysis, the amount of exercise achieved for the day is predicted from the momentum and behavior history data up to a certain point in time, the predicted amount of exercise 350 is stored in the storage 305, and each time new sensor data 320 is added, the predicted amount of exercise is calculated. Can be updated.

  By storing the parameter (prediction variable 370) necessary for the regression analysis learned by the all-day exercise amount prediction program 308 in the prediction variable storage area, it is possible to omit learning time before the subsequent prediction. By repeatedly updating the prediction variable 370 (for example, with a daily cycle), the prediction accuracy can be maintained even if the user's behavior pattern changes.

  The exercise content recommendation program 309 compensates for the exercise amount predicted to be insufficient from the difference between the achieved exercise amount of the day predicted at a certain point in the user's day and the user's target exercise amount preset in the user information 380. The recommended exercise content is generated and the recommended exercise content 360 is stored in the storage 305.

  The recommended exercise content is, for example, by showing that the amount of exercise is increased by increasing the walking speed according to the distance and the number of steps expected to walk on that day, without greatly changing the daily behavior, Necessary exercises can be supplemented on commuting routes. In addition, if the shortage of momentum cannot be compensated for by increasing the walking speed within a feasible range, it may be suggested that the shortage of momentum is compensated, for example, by adding distance and / or number of steps. it can.

  The WEB service program 310 transmits data stored in the storage 305 in response to a request from the portable information terminal 2 via the network 6. However, the information that can be transmitted is limited to the information analyzed from the sensor data measured by the sensor device 1 of the user holding the portable information terminal 2, and access to other information is blocked by a known security technique. In addition, user information can be input through an input interface provided by the WEB service program 310, and the user can input height, age, sex, weight, and the like and store the user information 380 in the storage 305. The target exercise amount and the target weight can be calculated from the user information.

  FIG. 5 is a diagram for explaining the configuration of the sensor data 320.

  The sensor data 320 is a collection of accelerations measured by the acceleration sensor 105 at a predetermined timing (for example, at intervals of 50 milliseconds) for a predetermined period (for example, 1 second) and stored in the storage 305 of the server 3. Is done. One record of the sensor data 320 includes a device ID 321, a measurement date 322, and sensor measurement values (temperature 323 and acceleration 324). The device ID 321 is identification information for uniquely identifying the sensor device 1 that has measured the sensor data. The measurement date and time 322 is the date and time when the sensor data is measured. The sensor measurement values 323 and 324 are physical quantities measured by the sensor device 1. It can be identified that each record of the sensor data 320 is data measured at the measurement date 322 by the device with the device ID 321.

  The sensor data 220 stored in the portable information terminal 2 is data having the same configuration as the sensor data 320 stored in the server 3 or data obtained by encrypting data having the same configuration.

  FIG. 6 is a diagram for explaining the configuration of the momentum data 330.

  The exercise amount data 330 is a summary of the exercise amount, the number of steps, and the step speed calculated from the sensor data for a predetermined period (for example, 1 minute), and is stored in the storage 305 of the server 3. One record of the exercise amount data 330 includes a user ID 331, a device ID 332, a date and time 333, an exercise amount 334, a step count 335, and a step speed 336.

  The user ID 331 is identification information for uniquely identifying the user who owns the sensor device. The device ID 332 is identification information for uniquely identifying the sensor device 1, and uses the same identification information as the device ID 321 of the sensor data 320.

  The amount of exercise 334, the number of steps 335, and the step speed 336 were obtained by analyzing sensor data 320 indicating the movement of the user (the user identified by the user ID 331) who has the sensor device 1 identified by the device ID 332, This is a value obtained by calculating the amount of exercise, the number of steps, and the walking speed for a predetermined time (for example, 1 minute) represented by the date and time 333. The amount of exercise 334 may be a value corresponding to METs that can be converted into calories consumed. When the walking speed 336 is calculated only from the acceleration sensor data, since the stride is not known, the walking speed 336 is a walking pitch (steps / second). Note that the walking speed (m / sec or m / min) can be calculated by multiplying the stride estimated from the height by the walking pitch.

  The exercise amount data 230 stored in the portable information terminal 2 is obtained by acquiring the exercise amount of a specific user from among the exercise amounts 330 stored in the server 3 and has the same configuration as the exercise amount 330 stored in the server 3. .

  FIG. 7 is a diagram illustrating the configuration of the action content data 340.

  The action content data 340 is stored in the storage 305 of the server 3, and one record of the action content data 340 includes a user ID 341, a device ID 342, a start date and time 343, an end date and time 344, and an action content 345. It may also include conditions (such as day of the week and weather) that affect the behavior of.

  The user ID 341 is identification information for uniquely identifying the user who owns the sensor device, and uses the same identification information as the user ID 331 of the exercise amount data 330. The device ID 342 is identification information for uniquely identifying the sensor device 1, and uses the same identification information as the device ID 321 of the sensor data 320. The start date / time 343 and the end date / time 344 are delimiters of the action content 345. The action content 345 is an estimated type of user action, and analyzes sensor data 320 indicating the movement of the user (the user identified by the user ID 342) who owns the sensor device 1 identified by the device ID 341. Can be obtained.

  FIG. 8 is a diagram for explaining the configuration of the predicted exercise amount data 350.

  The predicted exercise amount data 350 is stored in the storage 305 of the server 3, and one record of the predicted exercise amount data 350 includes a user ID 351, a date 352, an update date 353, a predicted exercise amount 354, and a predicted number of steps 355.

  The user ID 351 is identification information for uniquely identifying the user who owns the sensor device, and uses the same identification information as the user ID 331 of the exercise amount data 330. The predicted exercise amount 354 and the predicted number of steps 355 are the exercise amount and the number of steps predicted to be achieved on the date of the user's date 352 identified by the user ID 351. Date and time.

  FIG. 9 is a diagram illustrating the configuration of the user information 380.

  The user information 380 is stored in the storage 305 of the server 3, and one record of the user information 380 includes a user ID 381, a target exercise amount 382, a stride 383, a weight 384, a home position 385, and a nearest station position 386, and other Personal information (name, age, gender, etc.).

  The user ID 351 is identification information for uniquely identifying the user who owns the sensor device, and uses the same identification information as the user ID 331 of the exercise amount data 330. The target exercise amount 382 is a daily target exercise amount set for each user. The stride 383 is a user stride, and is used when the distance and the number of steps are converted. Note that the user information 380 may include height instead of the stride, and the stride may be estimated from the height. The body weight 384 is the weight of the user, and is used when the amount of exercise and calorie consumption represented by METs are converted. The home position 385 and the nearest station position 386 are the positions of the home and the nearest station, respectively, and are represented by, for example, longitude and latitude.

  FIG. 10 is a diagram for explaining an example of the exercise support monitor screen 401. The exercise support monitor screen 401 is obtained by displaying the past exercise history and recommended exercise content calculated based on the sensor data on the display unit 202 by the exercise history / recommended content display program 214 of the portable information terminal 2. is there.

  The exercise support monitor screen 401 notifies the user of the achievement status of his / her own exercise amount, and when the exercise amount of the day is predicted to be insufficient, the exercise content that can be specifically executed is notified. Help change and achieve goals. The exercise history display 405 indicates the transition of the past exercise amount. By converting the amount of exercise into calorie consumption, the user can compare the balance with the calorie intake, recognize the lack of exercise, and recognize that he is eating too much. In addition, by displaying the basal metabolism, which is the calorie consumed even if it does not move at all, and the calorie consumed by exercise separately, and by displaying it separately in the time zone (for example, morning and afternoon), its own behavior and consumption The user can easily recognize the relationship with calories.

  The current exercise amount display 406 indicates a result calculated based on the latest sensor data collected in the server 3 from the sensor device 1 worn by the user. By showing the predicted achievement momentum calculated from the action content of the day, the user can know the goal achievement status in the middle of the day, and can change the action to achieve the goal. When the predicted achievement exercise amount does not reach the target, the deficient exercise amount display 402 is notified of the deficient value, and the recommended walking speed display 403 and the recommended distance display 404 for compensating the deficiency are displayed.

  The recommended walking speed display 403 recommends an increase in the walking speed necessary to make up for the insufficient amount of exercise with the number of steps predicted to walk by the end of the day after the current time. In other words, it is a recommendation for achieving the goal without changing the daily moving means such as commuting and the moving route. On the other hand, if the shortage cannot be compensated for by simply changing the walking speed, the walking distance to be added is displayed on the recommended distance display 404. Note that at least one of walking speed (recommended walking speed display 403) or walking distance to be added (recommended distance display 404) may be displayed. That is, you may recommend an exercise that increases the walking distance without changing the walking speed.

  That is, the target exercise amount can be achieved by slightly changing the behavior according to the recommended walking speed display 403 and the recommended distance display 404. In addition, by converting the walking speed with the value displayed on the recommended walking speed display 403 to the sound of the recommended walking pitch 408 and outputting it from the speaker 407, the action content is notified more intuitively and the goal is achieved. Can help.

  The user can start outputting sound by operating the portable information terminal 2. Further, the portable information terminal 2 may start outputting sound when the sensor 203 detects that the user has started walking.

  FIG. 11 is a diagram for explaining the relationship between the walking speed and the amount of exercise.

  The relationship between the walking speed and the amount of exercise shown in FIG. 11 is data necessary for the server 3 to generate the recommended walking speed display 403. As shown in the figure, for example, when the normal walking speed is 100 [m / min], the amount of exercise can be increased by increasing or decreasing the walking speed. For example, in order to calculate the required walking speed for the insufficient exercise amount, the required exercise amount per movement distance can be calculated by dividing [the insufficient exercise amount] by the [predicted movement distance of the day]. Then, the recommended walking speed can be determined by selecting the corresponding walking speed from the graph of the figure.

  FIG. 12 is a diagram for explaining an example of the living behavior monitor screen 601. The living behavior monitor screen 601 shows details of the daily behavior of the user.

  On the living behavior monitor screen, the details of the user's behavior are shown by a step count transition display 607, an exercise amount transition display 608, and a behavior content display 609. Further, the predicted achievement momentum (604, 605) calculated based on these data is displayed. The action content display 609 indicates a breakdown or transition of the action contents by dividing the section divided by the change of the action contents by a color corresponding to specific action contents (for example, sleep, work, walking, desk work, etc.). .

  The values displayed in the current step count display 602 and the current exercise amount display 603 are calculated based on the data of the step number transition display 607 and the exercise amount transition display 608. Furthermore, the server 3 performs regression analysis on the past data of the step transition display 607, the exercise amount transition display 608, and the action content display 609, thereby calculating values of the predicted achievement step display 604 and the predicted achievement exercise amount display 605.

  FIG. 13 is a diagram for explaining an example of the weight monitor screen 501. The weight monitor screen 501 shows the transition of the weight value input by the user.

  The user can know the effect of weight loss by achieving the exercise goal on the weight monitor screen 501. The body weight transition display 505 shows the transition of the body weight in the past week or day, and the user recognizes the relation between the increase and decrease of the body weight and the amount of exercise by classifying the point color by achievement or non-achievement of the target exercise amount. It has become easier.

  The weight change display 502 shows the recent weight change, and the target weight display 503 shows the difference between the target weight and the current weight. The target weight is recorded in the user information 380 of the server 3, and an appropriate weight is set by a known algorithm based on the user's age and height. The goal achievement display 504 indicates the goal achievement status of the latest week or day (for example, five of the seven weeks achieved the goal).

  FIG. 14 is a diagram for explaining an example of the exercise support monitor screen 701. The exercise support monitor screen 701 shows an example in which the recommended exercise content shown in FIG. 10 is specifically displayed on a map.

  The portable information terminal 2 acquires the current position from the sensor 203, maps the shortest route 706 to a preset target point (for example, home), and the recommended route 707 for achieving the recommended distance 404 into a map. This is specifically shown at 705. Thereby, the user can easily execute the content of the recommended distance display 404.

  FIG. 15 is a sequence diagram of processing in the exercise support system according to the embodiment of the present invention. In the processing in which the server 3 analyzes the data measured by the sensor device 1 and displays the data on the portable information terminal 2, FIG. 1, the order of communication between the portable information terminal 2 and the server 3 and the execution of the program is shown.

  When the sensor device 1 executes the measurement program 103, the sensor 203 constantly measures the physical quantity, and the measurement value is recorded in the flash memory 109. The sensor device 1 executes the sensor data transmission program 104 at the timing of the data transmission start event 804 (detection of a user operation or connection with the portable information terminal 2), and carries the sensor data recorded in the flash memory 109 as a portable device. Transmit to the information terminal 2. After transmitting the sensor data, the sensor device 1 starts the measurement program 103 again (or executes the measurement program 103 in parallel with the sensor data transmission program 104).

  The portable information terminal 2 executes the sensor data transmission / reception program 213 to receive the sensor data transmitted by the sensor device 1 and transmits the received data to the server 3.

  The server 3 executes the sensor data reception program 306, receives sensor data from the portable information terminal 2, executes the behavior analysis program 307, and records the calculated behavior content in the behavior content 340.

  After the execution of the behavior analysis program 307 is completed, the all-day momentum prediction program 308 is executed, and a regression analysis is performed by using a combination of the momentum / step count pattern and behavior content calculated by the behavior analysis program 307, a breakdown, etc. as a feature amount. The achieved exercise amount (for example, the number of steps) is recorded in the predicted exercise amount 350.

  Next, the exercise content recommendation program 309 is executed, and when the value calculated by the all-day exercise amount prediction program 308 is smaller than the target value, the recommended exercise content (for example, the walking speed and distance) necessary to make up for the shortage is obtained. The calculated recommended exercise content is calculated, and the recommended exercise content 360 is stored in the storage 305.

  The WEB service program 310 can always accept access from the network, and transmits screen data for displaying values calculated by the server 3 in response to the request. However, by authenticating the user who owns the sensor device 1, only the user is permitted to access his / her sensor data and data calculated based on the sensor data. The exercise history / recommended content display program 214 of the portable information terminal 2 accesses the server 3 to acquire data on the amount of exercise, the predicted achievement amount of exercise, and the recommended exercise content to be notified to the user, and displays them on the display unit 202.

  FIG. 16 is a flowchart of the process of the measurement program 103. The measurement program 103 is executed by the microcomputer 101 of the sensor device 1, acquires the acceleration measured by the sensor 105 at a predetermined time interval (for example, a constant cycle), and stores the measured sensor data 112 in the flash memory 109.

  First, when the power of the sensor device 1 is turned on or a reset operation is performed (for example, a reset button is pressed) and the program is restarted, the microcomputer 101 and other elements are initialized (S101).

  Next, the previous time synchronization timing is acquired, and it is determined whether the correct time is set in the RTC 107. As a result, if it is estimated that the correct time is set, the process proceeds to step S103. On the other hand, if the time synchronization timing is past the predetermined time and there is a possibility that the correct time has not been set, the process proceeds to step S121, the time information is received through communication with the portable information terminal 2, and a new time is set. Wait until Thus, since the sensor device 1 can attach a correct time stamp to the sensor data, the measured sensor data can be arranged in time series without depending on the timing at which the portable information terminal 2 and the server 3 receive the sensor data. it can.

  Next, the output voltage of the battery 114 is determined. As a result, if the voltage is lower than the predetermined threshold, the process proceeds to step S104. On the other hand, if the voltage is higher than the predetermined threshold, it is determined whether the sensor data 112 to be transmitted is stored in the flash memory 109 (S122). Specifically, since the sensor data 112 is transmitted in units of sectors, it is determined whether there is a sector for which an untransmitted flag is set. If the sensor data 112 to be transmitted is stored, the sensor data 112 is transmitted to the portable information terminal 2 (S123).

  Next, it waits for an interrupt from the RTC 107. This interruption is issued at a predetermined cycle (for example, 50 milliseconds), and the sensor 105 measures acceleration according to this interruption. Note that while waiting for an interrupt, in order to reduce power consumption, the microcomputer 101 and other elements are powered off or set to a power-saving sleep mode. After receiving the interrupt, the process proceeds to step S105.

  Next, data measured from the acceleration sensor 105 is acquired and stored in the buffer of the microcomputer 101 (processing S105).

  Next, it is determined whether the data measured by the sensor has accumulated for a predetermined amount (for example, 20) or for a predetermined time (for example, 1 second) (S106). As a result, if the accumulated data is smaller than the predetermined amount, the process returns to step S103. On the other hand, if the accumulated data is greater than or equal to the predetermined amount, the process proceeds to step S107. By collectively processing the sensor data 112 (time-series acceleration data, etc.), it is possible to improve the efficiency of data compression and encryption processing and reduce power consumption.

  Next, the data stored in the buffer is compressed by a known algorithm (S107), and the compressed sensor data is encrypted by a known algorithm (S108). By encrypting the sensor data, reading of the sensor data can be prevented even if the sensor device 1 is lost.

  Next, it is determined whether or not the encrypted data stored in the buffer exceeds the size of the sector 110 that is a rewrite unit of the flash memory 109 (S109). As a result, if the encrypted data exceeds the size of the sector 110, the data accumulated in the buffer is stored in the flash memory 109, and the data stored in the flash memory 109 is erased from the buffer (S110). Return.

  FIG. 17 is a flowchart of the process of the behavior analysis program 307. The behavior analysis program 307 is executed by the CPU 301 of the server 3 and analyzes the amount of exercise and the behavior content from the sensor data measured by the sensor device 1.

  When the behavior analysis program 307 starts (S201), the sensor data 320 newly received from the sensor device 1 is read from the storage 305 (S202).

  Thereafter, the amount of exercise is calculated from the read sensor data 305 for a certain period of time (for example, the amplitude of acceleration for 1 minute) and recorded in the amount of exercise data 330 (S203). The amount of exercise can be calculated as a METs value that can be converted into calories burned from the integrated value of the amplitudes of the values measured by the acceleration sensor.

  Next, the number of steps in the same period as the momentum calculation period is calculated from the read sensor data 305 (for example, acceleration cycle) and recorded in the momentum data 330 (S204). The number of steps can be calculated by a known algorithm that detects and counts periodic vibrations in acceleration sensor data.

  Next, the action content is identified from the read sensor data and exercise amount data 330 and recorded in the action content data 340 (S205). For example, since the action content of a person is mainly divided by walking, the action content can be divided using the walking data recorded in the exercise amount data 330. Further, in the divided section, the user's behavior (walking, sleeping, deskwork, etc.) can be identified mainly by the inclination and periodicity that can be understood from the momentum and acceleration data. Then, the analysis result is recorded in the action content data 340.

  Thereafter, the behavior analysis program is terminated (S206).

  FIG. 18 is a flowchart of the process of the all-day exercise amount prediction program 308. The all-day exercise amount prediction program 308 is executed by the CPU 301 of the server 3, and features such as combinations and breakdowns of the pattern of the exercise amount (steps) data 330 read from the exercise amount data 330 until the middle of the day and the action content data 340, Learn the relationship with the amount of exercise achieved in a day by regression analysis. Also, the amount of exercise achieved per day is predicted from the sensor data until the middle of the day.

  When the all-day exercise amount prediction program 308 starts (S301), it is checked that the learned regression analysis parameters are recorded in the prediction variable 340 (S302), and the learned regression analysis parameters are stored in the prediction variable 340. If recorded, the process proceeds to step S303. On the other hand, if the learned regression analysis parameter is not recorded in the prediction variable 340, the process proceeds to step S310, and the parameter learning process is executed based on the past data.

  Thereafter, it is determined whether the learned regression analysis parameter recorded in the prediction variable 340 is within a predetermined effective period (S303). As a result, if the regression analysis parameters are calculated before a predetermined period (for example, one month) or more, changes in behavior patterns or the like during that period are not reflected, which may reduce the prediction accuracy. is there. For this reason, it progresses to process S310 and performs a learning process again. On the other hand, if a parameter within the expiration date is recorded, the process proceeds to step S304, and a prediction process is executed.

  In the process S310, the action content data 340 for a past fixed period is read, and in the process S311, the exercise amount data 230 for a past fixed period is read. As the period of the data to be read out, a period (for example, 1 month to 3 months) in which a sufficient amount for the regression analysis is obtained may be selected.

  Thereafter, regression analysis is executed using the read data (S312). In regression analysis, the amount of exercise achieved in a day is predicted based on multiple variables (for example, the pattern of momentum until the middle of the day, the pattern of the number of steps, the ratio of action contents in unit time, the combination of action contents, etc.) Learning data to be created is recorded, and the created parameters are recorded in the prediction variable 340. In the regression analysis, not only the exercise amount pattern but also data such as the day of the week and the weather may be used as parameters.

  On the other hand, in the process S304, all data on the amount of exercise, the number of steps, and the action content up to a certain time (for example, the current time) on the day on which the achieved exercise amount is to be predicted is read. If the regression equation calculated in step S312 includes parameters for the conditions of the day (day of the week, weather, etc.), data of conditions such as the day of the week, weather, etc. are read (S305). After that, the predicted achievement momentum is calculated by applying the read data to the regression equation (S306). The number of steps achieved by learning and prediction processing based on similar regression analysis can also be predicted.

  Thereafter, the calculated predicted exercise amount and the value of the predicted achievement step count are recorded in the predicted exercise amount data 350 (S307), and the achieved exercise amount prediction process is terminated (S308).

  FIG. 19 shows the result of predicting the amount of exercise achieved per day. The prediction results shown in FIG. 19 were acquired by the all-day exercise amount prediction program 308 by learning actual regression parameters using sensor data for two months, and then by 12:00 noon over one month. The result of predicting the amount of exercise achieved per day from sensor data is shown.

  FIG. 20 shows a comparison result of the accuracy of prediction shown in FIG. In FIG. 20, the prediction result by the method of this embodiment using the sensor data acquired from 8:00 to the designated time and the prediction by the single regression analysis using only the latest one momentum which is the conventional method. The results are shown in comparison. As can be seen from the figure, the prediction method of the present embodiment shows an accuracy superior to that of the conventional method at 8:00 to 10:00 with little accumulation of sensor data of the day.

  FIG. 21 is a flowchart of the process of the exercise content recommendation program 309. The exercise content recommendation program 309 is executed by the CPU 301 of the server 3, calculates the deficit amount for the target exercise amount for the day from the value calculated by the all-day exercise amount prediction program 308, and presents the walking speed and distance to compensate for the calculated deficiency amount. To do.

  When the exercise content recommendation program 309 is started (S301), the target exercise amount is read from the user information 370 (S302), and the predicted achievement exercise amount is read from the predicted exercise amount data 350.

  Thereafter, the read-out predicted achievement momentum and the target exercise amount are compared (S303). If the prediction achievement exercise amount exceeds the target exercise amount, the exercise content recommendation process is ended (S309).

  On the other hand, if the target exercise amount is greater than the predicted achievement momentum, the target exercise amount is subtracted from the prediction achievement momentum to calculate the insufficient exercise amount (S304). Thereafter, the walking speed necessary to compensate for the insufficient momentum is calculated (S305). Specifically, the amount of exercise deficient per distance is calculated by dividing the deficit exercise amount by the distance traveled by the predicted number of steps calculated by the all-day exercise amount prediction program 308 (insufficient exercise amount ÷ predicted movement distance). Specifically, by calculating the exercise intensity (METs) obtained by dividing the deficit exercise amount by 24 hours and the body weight and the predicted distance by multiplying the predicted achievement step number by the stride, and dividing the calculated exercise intensity by the prediction distance. Calculate the lack of momentum per distance.

  By applying the insufficient momentum per distance to the relationship between the walking speed and the momentum per distance shown in FIG. 11, the necessary walking speed can be calculated.

  However, the walking speed to be presented is generally limited to a walking speed that can be executed. Specifically, when the calculated walking speed exceeds a predetermined upper limit value, the upper limit value of the walking speed is presented. And the walking distance which supplements the momentum which is insufficient when walking at the walking speed of the upper limit value is calculated, and the calculated walking distance is presented.

  Next, the achieved exercise amount based on the recommended walking speed calculated in step S305 is calculated (S306). Then, the calculated achieved momentum and the target exercise amount are compared, and if the achieved exercise amount is smaller than the target exercise amount, an additional exercise amount (for example, a movement distance) for compensating for the shortage is calculated (S307). The additional momentum is, for example, an additional movement distance, and can be calculated by a calculation (insufficient momentum ÷ momentum per distance) that divides the insufficient momentum by the momentum per distance at the recommended walking speed. The calculated recommended walking speed and distance are recorded in the recommended exercise content data 360.

  Thereafter, the exercise content recommendation program is terminated (S309).

  According to the embodiment of the present invention, it is possible to accurately predict the exercise amount of the day from the behavior pattern up to a certain point in time. In particular, since the amount of exercise of the day is predicted by multivariate analysis using the number of steps and / or behavior patterns, for example, the behavior pattern of the wearer of the sensor device does not depend on the day of the week, etc. Even in some cases, the momentum can be accurately predicted.

  Furthermore, since the exercise content (the number of steps or the walking distance) that compensates for the insufficient amount of exercise is presented, the subsequent action can be corrected and the achievement of the goal can be assisted.

  Furthermore, since an upper limit is set for the number of steps to be presented, the user is not guided to an excessive exercise.

  Furthermore, since the recommended walking speed is calculated using the correspondence data between METs / distance and walking speed, an appropriate walking speed for achieving the momentum can be presented.

  Furthermore, since the walking speed is presented by sound or vibration, the user can walk according to the presented sound or vibration, which is convenient.

  Furthermore, since the route for achieving the calculated walking distance is displayed on the map, the user only has to walk the displayed route, which is convenient.

  Although the present invention has been described in detail with reference to the accompanying drawings, the present invention is not limited to such specific configurations, and various modifications and equivalents within the spirit of the appended claims Includes configuration.

DESCRIPTION OF SYMBOLS 1 Sensor device 2 Portable information terminal 3 Server 4 User 5 Base station 6 Network 101 Microcomputer 102 Non-volatile memory | storage part (ROM)
107 Real-time clock (RTC)
201 processor (CPU)
204 Random access memory (RAM)
230 Exercise amount data 240 Action history data 250 Predicted exercise amount data 260 Recommended exercise content data 301 Processor (CPU)
303 Local Area Network (LAN) Communication Unit 304 Random Access Memory (RAM)
330 Exercise amount data 340 Action history data 350 Predicted exercise amount data 360 Recommended exercise content data 370 User information data 407 Speaker

Claims (14)

A sensor terminal worn by the user;
An exercise support system comprising a computer communicably connected to the sensor terminal,
The sensor terminal has at least an acceleration sensor,
The calculator is
Based on the time series data of the acceleration detected by the acceleration sensor, calculate the time series data of the momentum of the user wearing the sensor terminal,
Based on the time series data of the calculated momentum, the behavior of the user wearing the sensor terminal is estimated in time series,
An exercise support system for predicting a total daily exercise amount of a user wearing the sensor terminal by multivariate analysis using a ratio of the estimated time-series behavior within a predetermined time as a parameter. The exercise support system according to claim 1,
The calculator calculates time-series data of the number of steps of the user wearing the sensor terminal based on the time-series data of acceleration detected by the acceleration sensor,
An exercise support system that predicts the total daily exercise amount of a user wearing the sensor terminal by multivariate analysis using time series data of the number of steps as a parameter. A sensor terminal worn by the user;
A computer communicably connected to the sensor terminal;
The sensor terminal has at least an acceleration sensor,
The calculator is
Based on the time series data of the acceleration detected by the acceleration sensor, calculate the time series data of the momentum of the user wearing the sensor terminal,
Predicting the total daily exercise amount of the user wearing the sensor terminal by multivariate analysis using the time series data of the exercise amount as a parameter,
Calculating a deficit momentum by subtracting the predicted total momentum from a preset target momentum;
Calculate the deficit momentum per distance by dividing the calculated deficit momentum by the predicted movement distance calculated by regression analysis of past momentum and past actions;
Calculate the walking speed in the predicted travel distance from the calculated insufficient momentum per distance ,
An exercise support system characterized by presenting the calculated walking speed. The exercise support system according to claim 3,
When the calculated walking speed exceeds a predetermined upper limit value, the calculator presents the upper limit walking speed, and calculates a walking distance corresponding to an insufficient momentum when walking at the upper walking speed speed. An exercise support system characterized by doing. The exercise support system according to claim 3,
The computer calculates the walking speed by using correspondence data of a walking speed, a unit time, and an exercise amount per unit weight. The exercise support system according to claim 3,
A portable terminal connected to the computer;
The exercise support system, wherein the portable terminal presents the calculated walking speed by sound or vibration. The exercise support system according to claim 3,
A portable terminal connected to the computer;
An exercise support system, wherein the portable terminal generates data for displaying the calculated walking distance on a map. An exercise support method in an exercise support system that supports user exercise,
The exercise support system includes a sensor terminal worn by a user and a computer connected to the sensor terminal so as to communicate with each other.
The sensor terminal has at least an acceleration sensor,
The method
The computer calculates time series data of momentum of a user wearing the sensor terminal based on time series data of acceleration detected by the acceleration sensor, and based on the time series data of the calculated momentum, A first step of estimating the behavior of the user wearing the sensor terminal in time series;
A second step in which the computer predicts the total daily exercise amount of the user wearing the sensor terminal by multivariable analysis using the estimated time-series behavior within a predetermined time as a parameter; An exercise support method characterized by comprising. The exercise support method according to claim 8,
In the first step, the calculator calculates time-series data of the number of steps of the user wearing the sensor terminal based on the time-series data of acceleration detected by the acceleration sensor,
In the second step, the computer predicts the total daily exercise amount of the user wearing the sensor terminal by multivariable analysis using time series data of the number of steps as a parameter. An exercise support method in an exercise support system that supports user exercise,
The exercise support system includes a sensor terminal worn by a user and a computer connected to the sensor terminal so as to communicate with each other.
The sensor terminal has at least an acceleration sensor,
The method
The calculator calculates time series data of momentum of a user wearing the sensor terminal based on time series data of acceleration detected by the acceleration sensor;
The computer predicts the total daily exercise amount of the user wearing the sensor terminal by multivariate analysis using the time series data of the exercise amount as a parameter;
The calculator calculates a deficit momentum by subtracting the predicted total momentum from a preset target momentum;
Calculating the deficit momentum per distance by dividing the calculated deficit momentum by the predicted movement distance calculated by regression analysis of past momentum and past behavior;
Calculating a walking speed in the predicted travel distance from the calculated insufficient momentum per distance ;
A step of presenting the calculated walking speed. The exercise support method according to claim 10,
In the step of calculating at least one of the walking speed and walking distance, the calculator presents the walking speed of the upper limit value when the calculated walking speed exceeds a predetermined upper limit value, and the step of the upper limit value is calculated. An exercise support method characterized by calculating a walking distance corresponding to a momentum that is insufficient when walking at high speed. The exercise support method according to claim 10,
In the step of calculating at least one of the walking speed and the walking distance, the calculator calculates the walking speed using correspondence data between the walking speed and the amount of exercise per unit time and unit weight. Support method. The exercise support method according to claim 10,
The exercise support system has a mobile terminal connected to the computer,
The method includes the step of the mobile terminal presenting the calculated walking speed by sound or vibration. The exercise support method according to claim 10,
The exercise support system has a mobile terminal connected to the computer,
The method includes the step of generating data for the portable terminal to display the calculated walking distance on a map.