WO2010146811A1 - Behavior suggestion device and method - Google Patents

Abstract

Provided is a behavior suggestion device that infers a person's behavior on the basis of biological information such as amount of sleep, number of steps walked, and amount of exercise, and presents improvement means that improve health indicators and the like for that person. From the person's biological information and behavior history (information) obtained from sensor nodes over a long period of time, pieces of biological information (sensor data) happening at the same times of day and having the same characteristics are classified as habitual behavior, and a behavior guide generation program on a personal computer computes correlations between the times of, lengths of, and transitions between habitual behaviors and target indicator data that the user wants to improve. The main behaviors of the user are displayed chronologically; simultaneously, a display device presents optimal behavior times and lengths for improving the target indicators, and more concrete behavior examples and comments for implementing said optimal behavior times and lengths. Said optimal behavior times and lengths, behavior examples, and comments are generated using a behavior history inputted and recorded by the user in the past, and behavior history data inputted by other people.

Description

Action suggestion apparatus and method

The present invention relates to a technology that can be worn on a person's body and that presents characteristics of life by presenting human behavior from information on a sensor terminal that measures biological information and behavioral states, and presents improvement means.

It has been pointed out that disruption of life rhythms such as modern night life and lack of sleep can cause lifestyle-related diseases. On the other hand, pedometers and activity meters that measure people's movements and steps are generally recognized. As introduced in Non-Patent Document 1, a small sensor node equipped with a sensor, wireless, microcomputer, etc., called Mote (registered trademark), is also in a practical stage. This sensor node is activated only when sensing or wireless communication is necessary, and otherwise it is operated for a long time with a small internal battery, with a small diameter of 3cm, by turning off the power and reducing power consumption. Yes, it is easy for people to wear. Therefore, in daily life, it has become possible to easily collect information on activities, sleep cycles, etc. in daily life, that is, life rhythm.

In Patent Document 1, simultaneously reporting (displaying) a life rhythm obtained by inputting an individual's calorie consumption, heart rate, body temperature, sleep depth, fitness club, hospital record, and other biological information relating to weight and other health Discloses means for recognizing the cause of the quality of biological information.

In Patent Document 2, when a person travels at a high speed for a long distance to overseas or the like, a deviation between a person's life rhythm (biological rhythm) and day / night rhythm is detected from an acceleration sensor or an ambient light sensor to correct jet lag. Thus, a means for emitting stimulation light to a living body is disclosed.

Japanese Patent Laid-Open No. 9-103413 Japanese Patent Laid-Open No. 10-68787

Even if movements and steps are measured with a sensor as shown in Non-Patent Document 1, it is difficult to wear multiple sensors at all times for 24 hours, such as a waist like a pedometer or an arm like a wristwatch. What you will find in practical use. However, only information obtained from a single sensor quantitatively evaluates values such as sleep time, number of steps, and amount of exercise, and it has been difficult to effectively promote life improvement. For example, in a diet, it is not always possible to effectively reduce the amount of weight by simply increasing the number of daily steps.

In Patent Document 1, a plurality of life rhythm indicators and biological information such as body weight are simultaneously displayed to present information that leads to improvement of life to the user, but a specific improvement method suitable for each individual is presented. Means to do is not disclosed.

Also, Patent Document 2 discloses a means for correcting sleep rhythm related to jet lag, but does not disclose means for improving health indicators obtained from other biological information.

An object of the present invention is to provide an action suggesting apparatus and a method for presenting means for estimating a person's action based on biological information such as sleep time, number of steps, and amount of exercise, and improving the health index and the like. . *

In order to achieve the above object, in the present invention, a processing unit is provided with a processing unit and a storage unit, and the processing unit collects personal ecological information as an action proposing device that makes an action proposal to an individual based on personal biometric information, Based on the collected biological information, individual behavior is discriminated and behavior element data is extracted, and individual behavior data is generated from this behavior element data for a predetermined period. These behavior element data, habit behavior data, and individual Provide an action proposal device and method for generating an individual action guideline based on the target index data set in advance and making an individual action proposal based on the generated action guideline.
Moreover, the action proposal apparatus and method further provided with the display part which displays an action guideline and action proposal as an action proposal apparatus of this invention are provided.
Further, in the behavior proposing device of the present invention, the processing unit performs correlation analysis between habitual behavior data and target index data for a plurality of predetermined periods, and habit behavior data having high correlation in the correlation analysis result is used as an action guideline. An action suggesting apparatus and method for adding and displaying on a display unit are provided.

That is, in order to achieve the above object, in a preferred embodiment of the present invention, from the biological information and action history of a person over a long period of several weeks to several months obtained by information input means such as an acceleration sensor, the same time zone. Categorizing biometric information with the same characteristics as a kind of habitual behavior, habitual behavior data such as time, length of time and transition relationship of habitual behavior, weight, stress, life satisfaction that the user wants to improve Calculate the correlation with target index data such as business efficiency, display the main actions of the user in time series, and at the same time, propose the action time and length of time that are optimal for improving the target index Present on the display. In addition, we use action history that we entered and recorded more specific action examples and comments to realize the optimal action time and length of time, and action history entered by others. Generate and present.

According to the present invention, the user can easily recall his / her own life pattern and action content for a certain period from the display of the characteristics of the action, the time zone and the length of time, and the combination thereof, and at the same time, Know how to improve the best behavior to achieve better indicators. Furthermore, it is possible to correct the behavior by receiving feedback on the progress of whether or not the presented optimum behavior can be executed.

It is a block diagram of the system which produces | generates the action guideline and action proposal concerning a 1st Example. It is a block diagram which shows the structure of the program for updating the database of the system concerning a 1st Example. It is a figure which shows the structural example of the database which stores the sensor data of the system concerning a 1st Example. It is a figure explaining the method of calculating the momentum of the system concerning a 1st Example. It is a figure which shows the structural example of the database which stores the exercise | movement amount and step count of the system concerning a 1st Example. It is a figure which shows the structural example of the action definition database which stores the conditions for discriminating action from sensor data of the system concerning a 1st Example. It is a figure which shows the structural example of the database which stores the action element which is the action discriminate | determined from the sensor data of the system concerning a 1st Example. It is a figure which shows the process of producing | generating the habit action which classified the daily similar action of the system concerning a 1st Example, and a display example. It is a figure which shows the structural example of the database which stores the habit action which classified the daily similar action of the system concerning a 1st Example. It is a figure which shows the structural example of the database which stores the action log | history input by the user of the system concerning a 1st Example. It is a figure which shows an example of the habit action display screen of the system concerning a 1st Example. It is a plot figure which shows the correlation of a certain user's target parameter | index (work efficiency) and morning rest time analyzed with the system concerning a 1st Example. It is a figure which shows an example of the action guideline display screen of the system concerning a 1st Example. It is a figure which shows the example of the action proposal display screen of the system concerning a 1st Example. It is a figure which shows the example of the action log | history input screen of the system concerning a 1st Example. It is a figure which shows the flowchart which shows the detail of the action discrimination | determination process of the system concerning a 1st Example. It is a figure which shows the flowchart which shows the detail of the action guideline production | generation process of the system concerning a 1st Example. It is a block diagram of the sensor node of the system concerning a 1st Example. It is an external view of the sensor node of the system concerning a 1st Example. It is a figure which shows the example of a screen display of the sensor node concerning a 1st Example. It is a figure which shows the example of event input operation of the sensor node concerning a 1st Example. It is a block diagram of the system configuration which analyzes by a server concerning the 2nd example. FIG. 10 is a block diagram of a system configuration in which analysis is performed by a server and the result is browsed by a personal computer (PC) browser according to a third embodiment. It is a block diagram which shows an example of an internal structure of PC in each Example. It is a figure which shows an example of the flowchart of the process of the action proposal program in each Example.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. In this specification, a program executed by a processing unit of a computer such as a PC or a server may be expressed as a “unit” or “means”. For example, the “behavior guideline generating program” is referred to as “behavior guideline generating unit” or “behavior guideline generating means”.

FIG. 1 is a diagram showing a main configuration of a system for generating action guidelines and action proposals according to the first embodiment. In this system configuration, the sensor information which is the data measured by the sensor node 1 worn by the user 8 is received, and an index that the user 8 aims to improve day by day, such as work efficiency and results, satisfaction In order to propose a method for improving the index representing the degree, the past target index data 310 and the behavior element data 300 that records the past behavior determined from the sensor data are analyzed to calculate an improvement guideline, and the display device 3 and the like To the output device.

As will be described later, the sensor node 1 is a shape suitable for a person to wear, for example, a wristwatch-type sensor terminal, and measures human information (hereinafter, biological information) such as a pulse and movement, It is transmitted wirelessly as data. When you want to keep a special event that you want to record and keep a time of action, select an icon displayed on the liquid crystal display device (LCD) with the button operation, event information It has a function to record the event type and time. This not only saves the trouble of inputting action contents later, but also prevents forgetting the exact contents and time. In addition, a storage unit such as a nonvolatile memory that records the measured data therein is provided. Each sensor node 1 is assigned and stored with a Mac (Media Access Control: MAC) address, which is a unique identification ID. Further, an identification ID (short address) unique only within a network constituted by one base station 6 can be stored. By adding these addresses to the data and wirelessly transmitting, it is possible to identify which sensor node 1 is the transmitted data.

The base station 6 receives a radio wave from the antenna 7 and transmits the contents of the sensor data transmitted from the sensor terminal in accordance with a request from the connected PC 2. When the base station 6 configures one wireless network, the base station 6 communicates with the sensor node 1 to communicate a network joining procedure called association, and one short address for one MAC address. Is allocated. The base station 6 holds a correspondence table of MAC addresses and short addresses in an internal memory, converts all addresses added to the data received from the sensor node 1 into MAC addresses, and creates a personal computer (Personal Computer: PC) 2. As a result, the PC 2 can easily identify the sensor node 1 that is the transmission source of the received data.

The PC 2 has, for example, a normal computer configuration shown in FIG. That is, in FIG. 24, a PC indicated by 2401 is connected via an internal bus 2406 to a central processing unit (CPU) 2403 that is a processing unit, a main storage unit 2404 that is a storage unit, an auxiliary storage unit 2405, and A network device such as the Internet 9 and a network device serving as an interface such as a USB (Universal Serial Bus: USB) connected to the base station 6 necessary for capturing the data of the sensor data 1 into the PC are connected to each other. Further, a display device 3 such as an LCD shown in FIG. 1 and an input device 4 such as a keyboard and a mouse are connected by an appropriate interface.

This PC 2 includes various programs executed by the CPU, such as a sensor data receiving program 220 that receives sensor data from the base station 6 and stores it in the database, sensor data 280 that is data measured by the sensor node 1, Based on the sensor data 280, various data such as behavior definition data 290, which is a definition of a threshold necessary for determining human behavior, for example, sleep, walking, rest, exercise, etc., is stored in the storage unit. .
Further, the PC 2 uses the behavior definition data 290 to read and compare the sensor data 280 in time series, discriminates the behavior, stores the behavior data in the database, and the sensor data to simplify the behavior discrimination. The amount of exercise, which is intermediate data to be calculated, the number of steps data 320, the individual behavior determined by the behavior determination program 230, the behavior element data 300 which is a database for storing behavior elements, and the behavior of the same behavior definition in the same time zone among the behavior elements Habit behavior data 330 in which elements are grouped, target index data 310 that is a database for storing an index that the user 8 wants to improve, analysis of changes in past behavior element data 300 and target index data 310, How you should change your daily behavior from your behavior, for example early or late An action guideline generation program 240 that generates information indicating that the information is long and slow and displays the information on the display device 3 that is a display unit, action guideline data that is a result of the action guideline generated by the action guideline generation program 240, An action suggestion program 250 that searches the display device 3 for a specific action history of a past person or other person who matches, and displays it on the display device 3. An action history or comment such as a diary entered by the user 8 using the input device 4. The action history input program 210 for storing the action history in the database, the action history data 270 that is a database storing the action history and comments input by the user 8, and the database stored in the PC 2 via a network such as an intranet or the Internet, Receive and update data from computers such as other servers Equipped with a personal data update program 260 of the eye.
On the other hand, the server 5 is connected to the PC 2 via a network 9 such as an intranet or the Internet, and has a processing unit and a storage unit having a normal computer configuration as shown in FIG. Data 506, target index data 510, and action history data 511 are provided. The action definition data 506 and the target index data 510 are the same as or newer than the action definition data 290 and the target index data 310 of the PC 2. For example, when the score of the target index measured by another sensor or the like cannot be directly added to the behavior index data 310, it is stored once in the target index data 510 of the server 5, and can be copied to the PC 2 and updated. it can. Further, when the contents of the action definition data 506 are added, the update work can be facilitated by copying the data of the server 5 to the action definition data 290 by each user.

The sensor data receiving program 220 in the PC 2 functions as a biological information receiving unit by program processing of the processing unit, and first recognizes an interface with the base station 6 connected to the PC 2 through dedicated driver software or the like. The sensor data receiving program 220 is a rewritable ID for identifying one user 8 or a plurality of users using the PC 2 and the sensor node 1 worn by each user in association with each other. Holds the table. Therefore, an ID or name for identifying the user 8 can be added to the received data and stored in the sensor data 280 as the biological information of the user. Further, when the sensor node 1 is provided with an input means for key operation or button operation, in order for the user 8 to record his / her actions and events, that is, event information without fail, the sensor node 1 is operated and input, It is efficient to receive the same as the sensor data. In this case, the sensor data receiving program 220 stores the received event information in the action history data 270.

Now, the behavior determination program 230 in FIG. 1 uses the processing definition program stored in the behavior definition data 290 and a list of determination / detection conditions such as threshold values corresponding thereto by program processing in the processing unit. 280 data are compared in time series. As a result, the behavior name or the identification ID of the behavior, the time, and the time information are stored in the behavior element data 300 as a behavior element that matches the determination / detection condition. Thus, the sensor data 280 is merely a collection of waveform data measured by an acceleration sensor or the like, but the waveform data can be converted into information that can be seen and understood by humans in the form of behavior. At the same time, information such as time and time can be added to the action. That is, by recording numerical information such as an action and the time and time associated with the action over a long period of time, it is possible to quantitatively evaluate the change of the action and its transition.

The target index data 310 records an index that the user 8 desires to improve over a long period of time. For example, the question of what action should be taken and what should be done when a person wants to improve work efficiency and life satisfaction in their daily life. It becomes. From the viewpoint of work efficiency and satisfaction, work results, goal achievement, objective evaluation, subjective evaluation, etc. can be used. However, in terms of achievement and goal achievement, depending on the type of work, the results may occur once every few months or half a year, and sufficient data necessary for analysis cannot be obtained. Another problem is not being able to capture minute changes every day. In terms of objective evaluation, it is not easy in the long run to ask other people for evaluation on a daily basis.
Therefore, subjective evaluation, for example, answering a questionnaire about once every day or several hours and obtaining the result can be most easily realized. The resulting score can be used as target index data. It is also possible to estimate the subjective evaluation score using the data of the sensor node 1. This depends on the sensor to be applied and the content of the evaluation, but can be made by generating a prediction formula using a known method such as multiple regression analysis. This prediction result can also be applied as the target index data 310. The target index data 310 is not limited to such a score such as satisfaction of work and life. For example, the target index can be a measurement result of biological information such as weight and blood pressure. In the case where the behavior index data 310 is the weight, it is possible to support life improvement for dieting. This can be applied to the same health care purpose in other health indicators such as those obtained in health examinations.

The behavior guideline generation program 240 analyzes the time series changes of the behavior element data 300 and the target index data 310 and the correlation by the program processing in the processing unit, and generates a behavior guideline. This action guideline is held in a storage unit (not shown) as action guideline data including action definition ID, date / time, time length, and the like of the action element. If the simplest method is applied, when the time of an action is early (for example, when the attendance time is early), if the job satisfaction score of the target index data 510 is always increased, the time of the action is always set. Guide the action that should be done early. Pearson's product-moment correlation coefficient and the like are generally well known as analysis methods for deriving the correlation between these two values, and these known correlation analysis methods can be applied. Further, when performing this analysis, it is necessary to recognize from the behavior element data 300 that the same behavior is exceeded across days and weeks. This technique is based on, for example, identifying morning and evening commuting, regular meetings, lunch, evening jogging, habitual behavior such as sleep, non-habitable event behavior, and the like, and evaluating daily time changes. In such a plurality of days, action elements having the same action definition in the same time zone are grouped, and one group is set as a habit action. Since this habit behavior includes a plurality of behavior elements, a quantitative evaluation such as correlation analysis can be performed using those times and times.

The action history input program 210 displays on the display device 3 a user interface for allowing the user to easily input past or current action contents via the input device 4 by program processing of the processing unit. Is recorded in the action history data 270 as an action history. Here, the action history input program 210 refers to the sensor data 280 and the action element data 300 and displays them, whereby the user 8 can be reminded of past action contents and can assist the user 8 in input. Moreover, the action history input program 210 can transmit the input action content to the server 5 connected via the network 9 and similarly store it in the action history data 511 that is a database. Thereby, not only data backup and input from a plurality of terminals are enabled, but other users can use the action history data 511 of the server 5 as knowledge of life improvement. In the action guideline generation program 240, the derived guideline is displayed on the display device 3 and notified to the user 8, whereby the life improvement of the user 8 can be promoted. Moreover, the user 8 can ensure privacy by enabling the user to select whether or not to transmit the input action content to the server 5.

The action suggestion program 250 receives the action guideline data generated and held in cooperation with the action guideline generation program 240 from the program processing of the processing unit, and receives past action definitions, time, and time, which are the contents of the action guideline. The action history data 270 is searched and the matching action content is displayed on the display device 3. The user 8 can recognize specific measures for life improvement based on his / her past behavior. Moreover, the action proposal program 250 can also refer to the action history data 511 of the server 5, that is, the action history of another user. Thereby, since effective knowledge may not be obtained from one's past behavior, it is possible to learn from the behavior of others and promote life improvement.

FIG. 25 is a flowchart showing in detail the process of the action suggestion program 250 executed by the processing unit of the PC 2 of FIG. 1. The action guideline data generated by the action guideline generation program 240, the action history data 270, and In order to achieve the target guideline from the action history data 511 of the server 5, a preferred example of a process for generating a more specific action proposal is shown.

When the action guideline generation program 240 is completed and the action guideline data is generated and held, the action proposal program 250 starts the action proposal process in process 251.

In process 252, the action guideline data generated by the action guideline generation program 240 is read. The action guideline data to be read is the action definition ID 302, the date and time 303, and the time length 306 of the action elements included in the habit action having a high contribution rate to the achievement of the target index.

In the process 253, the past action content of the user 8 that matches the action definition ID 302, the date and time 303, and the time length 306 of the action element of the action guide data read in the process 252 is retrieved from the action history data 270 and read. When the PC 2 is connected to the server 5 via the network 9 and the action history data 511 can be read, the action history that does not overlap is read.

In the process 254, similarly to the process 253, the PC 2 is connected to the server 5 via the network 9, can read the action history data 511, and if there is an action history input by a user other than the user 8, the action of the action guideline data An action history that matches the element's action definition ID 302, the date and time 303, and the time length 306 is read.

In the process 255, as shown in the action guideline display screen 431 shown in FIG. 14, for example, the action history read out in the processes 253 and 254 is displayed on the action suggestion display 432 and the actions of users other than the user 8 are displayed. The history is displayed separately on the action proposal display 433. The order displayed on the action proposal display 432 and the action proposal display 433 is rearranged in the order in which the target index is improved. In this case, depending on the screen size of the display device 3, it is not possible to display all of them, and therefore, the images are displayed in the sorted order. In addition, a user interface for scrolling operation is added to the screen. Next, FIG. 2 is a block diagram showing in detail the personal data update program 260 of the PC 2. The personal data update program 260 includes an action definition generation program 261, an action evaluation program 262, an action definition update program 263, and a target index update program 264.

In the figure, the action definition generation program 261 can correct the action definition data 290 or add a new definition with reference to the action history data 270 and the sensor data 280. An initial value such as a threshold value stored in the action definition data 290 is set so as to be commonly applied regardless of who the user 8 is. However, when estimating human behavior from sensor data such as movement, individual differences become a problem when attempting to increase accuracy. Therefore, based on the action history data 270 that is the actual action content, the corresponding sensor data 280 is searched and referenced, a new determination condition that can be discriminated with high probability is generated, and added to the action definition data 290b. Can do. In addition, an action definition that does not exist in the action definition data 290 in the initial state can be added by a similar method.

The behavior evaluation program 262 is a program that generates the target index data 310 based on the sensor data 280. As described above, when the target index data 310 is a subjective evaluation related to job satisfaction, an approximate expression for calculating the result of the subjective evaluation based on the sensor data 280 is generated by a technique such as multiple regression analysis. can do. In the simplest implementation method, the behavior evaluation program 262 generates the target index data 310 based on the subjective evaluation approximate expression.

The behavior definition update program 263 is a program for communicating with the server 5 to read the behavior definition data 506 and updating the behavior definition data 290 of the PC 2. At this time, items corrected and generated by the action definition generation program 261 in the action definition data 290 are not overwritten. That is, while the determination conditions tailored to each individual are left as they are, the determination conditions common to the whole can reflect the latest contents.

Similarly, the target index update program 264 is a program that reads the target index data 510 of the server 5 and adds it to the target index data 310. When the server 5 stores another sensor or a summary result such as a questionnaire, the latest data can be reflected on the PC 2.

FIG. 3 is a configuration example of the sensor data 280, and shows a case where data measured by a triaxial acceleration sensor at intervals of 50 milliseconds are stored. Measurement of 10 times or more per second is suitable for accurately identifying human behavior and particularly the number of steps from acceleration data. The date 281 indicates the time at the sensor node 1 when the sensor measures. An acceleration x282, an acceleration y283, and an acceleration z284 each indicate a triaxial acceleration measurement value. When the resolution of the digital value of the acceleration sensor output or the digital value obtained by AD conversion of the analog value of the acceleration sensor output is 8 bits, the value range is 0 to 255. If the measurement range of the acceleration sensor is about ± 3G to ± 5G, it is sufficient to identify human movement. A temperature 285 indicates the temperature measured by the sensor node 1. The pulse wave 286 stores the output value of the pulse waveform only when the sensor node 1 includes a pulse sensor such as an optical type.

FIG. 4 shows a method of calculating the number of zero crossings indicating the amount of waveform fluctuation from the acceleration sensor data. Since the number of zero crossings is proportional to the amount of movement, it can be regarded as a momentum. Further, rough behavior such as sleep, exercise, and desk work can be determined from the level of the amount of exercise. The simplest method for obtaining the number of zero crosses is to convert triaxial acceleration into a uniaxial scalar value. When the posture is not taken into consideration, the momentum can be estimated only by the scalar value. At this time, it is preferable to apply a band-pass filter in order to remove a noise component unrelated to human movement. In addition, when it is desired to suppress detection of fine movement, it is preferable to set a threshold value. For example, in FIG. 4, only the point 12 where 0.03G and the acceleration scalar value 11 intersect is detected and counted, and when estimating human behavior from the amount of exercise, the most appropriate threshold value based on actual data It is.

FIG. 5 shows a configuration example of the exercise amount calculated from the sensor data 280 and the exercise amount and step number data 320 for storing the step number data. The amount of exercise data is intermediate data generated by the behavior determination program 230, and the definition and processing are facilitated by comparing the determination conditions of this data and the behavior definition data 290. The method of calculating the amount of exercise is as shown in FIG. The time range of one value is the minimum time range of the action to be recognized. However, assuming an action unit that can be recalled by a daily action, approximately 1 to 5 minutes is preferable. The date / time 321 indicates that data corresponding to one minute from the indicated value is stored in the same row. The amount of exercise 232 stores a value calculated by the method of FIG. If the simplest method is shown, the step count 323 can detect periodicity with respect to a scalar waveform of acceleration by a known method such as autocorrelation, and calculate the step count from the number of consecutive detected cycles.

FIG. 6 shows a configuration example of the action definition data 290. The behavior definition data 290 stores determination conditions for classifying a plurality of behaviors, and it is possible to increase the accuracy and cope with individual differences later by addition and correction. By determining the amount of exercise and the step count data 320, processing and conditions can be easily set. The action definition ID 291 is an identification ID unique to the action to be defined. The action definition name 292 indicates the name of the action determined by the conditions stored in the same row. The determination condition 293 can store a plurality of conditions and threshold values. For example, in addition to definitions relating to the amount of exercise and the number of steps, other detection conditions can also be specified, and the action determination program 230 performs a determination process by logical sum (or logical product can also be set) for each condition. . One determined behavior unit is called a behavior element. A configuration example of the determination condition is shown below. The momentum 294 indicates a condition corresponding to the momentum 322 stored in the momentum and step count data 320. The step count 295 indicates a condition corresponding to the step count 323. The continuous time 296 indicates how much time is matched with the conditions of the momentum 294 and the number of steps 295 to determine. In other words, it can be matched only when it continues for a certain period of time. This is because if the rest time with little movement is as short as 1 minute, it cannot be regarded as a characteristic action, but if it is resting for about 10 minutes, it is likely to be recalled as a characteristic action. .

FIG. 7 shows action element data 300 storing actions determined by the action determination program 230. The action element INDEX 301 is a unique identification ID for all the action elements. The action definition ID 302 stores an action definition ID 291 corresponding to the determined condition. The date and time 303 indicates the date and time when the action element exists, and includes a start date and time 304 and an end date and time 305. The time length 306 is the duration of the action element, and is the difference between the end date / time 305 and the start date / time 304. As described above, the behavior element is characterized by the amount of exercise, the number of steps, and the date and time (time), so that the user 8 can recall the actual memory when viewed later. Further, as described later, these times can be analyzed in the long term and quantitative evaluation can be performed.

FIG. 8 shows a method for performing a long-term behavior evaluation, that is, a quantitative evaluation of a person's habitual behavior, in the behavior guideline generation program 240 in this embodiment. For the behavior element data 300, if the neighboring time and the same behavior definition are used, they are grouped as similar behaviors. Since this is a similar behavior across multiple days, it can be called habit behavior or life rhythm. FIG. 8 shows an example in which the behavioral element data 300 for one week is analyzed. In the example of FIG. 8, groups 601 to 607 having the same action definition are classified at neighboring times. The neighborhood at this time is processed in the range of about 1 hour before and after.
Next, each group is expressed in a transition diagram as habit behaviors 608 to 615. Further, it is possible to know the transition relationship of habit actions from the context of the action elements constituting habit actions 608 to 615. Based on this transition relationship, an arrow can be added to the habit behaviors 608 to 615, making it easier to recognize a life pattern. Moreover, since the transition probability can be calculated from the number of times of each transition, it is possible to know a characteristic behavior transition pattern. The habit actions 608 to 615 are expressed by a figure such as a circle, and indicate the length of time of each action in size. The time length of the habit behavior can be easily calculated by averaging the time lengths of the behavior elements constituting the habit behavior. Further, the position where each habit behavior is expressed is the time point 616 when the central point of the constituent behavior elements and the behavior element 3 are taken as an example. By the above expression, not only can a person see and recall an actual action from the color and context, but also the temporal change of the constituting behavior element can be quantitatively evaluated for each habit action.

FIG. 9 shows an example of the configuration of the habit behavior data 330 in which the habit behaviors 608 to 615 in FIG. In the action guideline generation program 240, based on the habit action data 330, a correlation analysis with the target index data 310 is performed with respect to the time of action elements constituting each habit action. The habit action INDEX 331 is a unique identification number for each habit action. The behavior definition ID 332 indicates a behavior definition ID common to the behavior elements constituting the habit behavior. The start time 333 and the end time 334 are the average start and end times of the behavior elements constituting the habit behavior. Based on this, the display position of habit behavior is changed. The time length 335 is a difference between the end time 334 and the start time 333.
The behavior element INDEX 336 is an INDEX of the behavior element constituting the habit behavior. That is, by searching the behavior element data 300 from the behavior element INDEX 336, all information of the behavior elements constituting the habit behavior can be read out. Based on the read time information, correlation analysis with the target index data 310 is performed. The transition source INDEX 337 is information necessary to know the transition relationship of habit behavior. Based on the transition source INDEX 337, an arrow can be connected when displaying a habit action. The number of transitions 338 is obtained by counting the number of transitions from each transition source, and the transition probability can be known based on the number of transitions 338.

FIG. 10 shows a configuration example of the action history data 270. The action history data 270 stores results input by the user 8 via the action history input program 210. In addition, event information input by the user 8 by button operation in the sensor node 1 is also stored. The period of action is indicated by a start date 272 and an end date 273. The action content 274 stores content directly input by the user 8 using the input device 4 or content selected from a plurality of candidates. By leaving the data input by the user 8 himself / herself, the action definition data 290 can be used for correction or addition, and action proposal contents using the input contents can be generated.

FIG. 11 is a display example on the habit action display screen 400 that is generated by the action guideline generation program 240 in the present embodiment, transmitted to the display device 3, and displayed. In this display example, for June 2008, habit behaviors 403 to 428 are calculated by dividing weekday display 401 and holiday display 402. By separating holidays and weekdays, it is possible to make it easier for the user 8 to recognize his / her habits and lifestyle patterns even when, for example, the weekdays are work and the lifestyle patterns of the weekdays and holidays are large. Moreover, since the meaning may differ completely even if it is a similar action of the same time slot | zone on a holiday and a weekday, it will also lead to the improvement of the precision of analysis. Each of the habit actions 403 to 428 is color-coded by the action legend 429. Typically, it can be intuitively recognized that a warm color system is assigned to an action with a lot of movement and a cold color system is assigned to an action with a little movement. From the color classification, the time of action, and the transition relationship, the user 8 can recall the execution action indicated by (...). In the above display method, the habit actions 403 to 428 are arranged on the horizontal axis according to the time, but are not specified on the vertical axis, and are arranged so as to be optimized so that they do not overlap each other.

FIG. 12 is a graph 500 showing the result of analysis by the action guideline generation program 240 based on the data of one user in an actual year. Each plot shows the average data for one week. The vertical axis of the graph 500 applies an index indicating human work efficiency as the target index data 310. Here, work efficiency indicates how a person is immersed in work during work hours. Results obtained from multiple questionnaires (self-diagnosis) during past work and acceleration data before and after the questionnaire responses. By using multiple regression analysis based on the above, it is possible to calculate the state of being immersed in work from only acceleration data as an index. The horizontal axis of the graph 500 indicates the time length of the behavioral elements constituting the resting habitual behavior that appears around 10:00. As can be seen from FIG. 12, since the plots are arranged close to a straight line, it is suspected that there is a correlation. Further, when the correlation analysis is actually performed, the correlation coefficient is 0.59, and the significance probability is 0.0001 or less, which indicates that there is a significant correlation. Therefore, it can be seen that the longer this rest time, the more immersed in work and the higher the work efficiency. In this way, by analyzing all habit behaviors, a correlation between each person's goal index and habit behavior can be obtained. The action guideline generation program 240 displays and proposes an action guideline based on this correlation.

FIG. 13 shows a display example of the action guideline display screen 430 generated by the action proposal program 240, transmitted to the display device 3, and displayed. The weekday display 401 shows the same habit behavior as the habit action display 400, but the correlation coefficient between each habit action 403 to 416 and the target index (work efficiency in the example) is calculated by the method shown in FIG. The value on the vertical axis. As a result, it is possible to recognize at a glance which habit behavior and improvement of the target index are related. Furthermore, since a direction for improving the behavior, that is, a guideline is also shown, an arrow indicating a direction for improving the target index is added to the most relevant habit behavior. For example, using the results shown in FIG. 12, work efficiency improves when the time is extended. Therefore, an arrow that spreads the habit behavior 408 to the side is added, and the user recognizes his / her own life pattern and at the same time recognizes the improvement method at a glance. can do. If it is recognized that there is a significant correlation, it is also possible to display the same guideline in addition to the most correlated habit behavior.

FIG. 14 shows a display example of an action proposal display screen 431 that is generated by the action proposal program 250, transmitted to the display device 3, and displayed. The action proposal program 250 searches the action history data 270 based on the result of the action guideline on the action guideline display screen 430 generated by the action guideline generation program 240, and displays the past action history that matches the action guideline. For example, when a search is made for the habit behavior 408, the content actually input in the past is displayed as in the example shown in the behavior proposal display 432. Thereby, the optimal action proposal content based on one's action can be displayed. In addition, it is assumed that knowledge can be obtained from other than the own action history. In the action proposal display 433, it is possible to obtain knowledge by searching along the action guideline from the action history data 511 of the server 5, that is, from past actions of others. This is effective when the action history data 270 input by the user is small or when a sufficient improvement result cannot be obtained only from the past actions of the user.

FIG. 15 shows a display example of the action input screen 440 generated by the action history input program 210, transmitted to the display device 3, and displayed. In the action input screen 440, in order to allow the user 8 to input past actions efficiently, the action amount graph 442 and the time 444 of the date and time specified by the date and time selection 441 are displayed on the determined action element 443. By entering while observing the above momentum and behavior elements, it is possible to input easily even if you forget detailed time breaks, etc. Since rough actions have already been identified as behavior elements, more detailed behavior It is possible to recall contents and comments.

FIG. 16 shows in detail a flowchart of the process of the action determination program 230 executed by the processing unit, and shows a preferred example for realizing a process of determining a person's action from a person's movement and biological information. In the process of FIG. 3, if human behavior is classified into several types, using the fact that a feature appears in the amount of movement, sensor data obtained by measuring human movement such as acceleration is used as momentum data that represents the amount of movement once. Convert to In addition, the number of steps is detected from the periodicity of the waveform of the motion data, and is used to determine whether the user is moving. The calculated amount of exercise and the number of steps are compared with the determination condition of the action definition data 290 to determine the action. It is desirable that the amount of unloading and step count data 320 is non-volatile, and the data once calculated is retained even after the action determination program 230 ends.

Processing 231. The processing of the action determination program 230 is started. The behavior determination program 230 starts by detecting whether new data is added to the sensor data 280 or updated.

In the process 232, it is confirmed whether or not the exercise amount and the step count data 320 have already been calculated and recorded in the range (period) in which the action is determined for the sensor data. If the calculation has been completed, the process proceeds to process 235 without performing these calculation processes. If the calculation has not been completed, the process proceeds to process 233.

In the process 233, the amount of exercise is calculated from the sensor data 280 and recorded in the amount of exercise / step count data 320. In this processing, for example, if the sensor data 280 is waveform data of a three-axis acceleration sensor, the number of zero crosses obtained by counting how many times the waveform has shaken per unit time after converting to a scalar value that is an absolute value of a vector is obtained. Let it be momentum. The unit time here is preferably about 1 minute since a movement of a unit of several seconds is not regarded as an action. In this process, it is desirable to apply a band pass filter to the scalar value generated in the middle to remove frequency components unrelated to human movement.

In the process 234, the number of steps is calculated from the data obtained by measuring the movement of the person in the sensor data 280. Here, similarly to the processing 233, the scalar value is calculated once, and the periodicity is detected by a known means such as FFT (frequency analysis) or autocorrelation analysis. Count the number of steps followed by the number of steps. The calculated number of steps is recorded in the exercise amount / step count data 230.

Processing 235 refers to the amount of exercise and step count data 230, and compares the amount of exercise and the number of steps with the discrimination conditions included in the action definition data 290, respectively.

In the process 236, information such as the name and ID of the action definition, the start time, and the end time is recorded in the action element data 300 as a result of matching in the process 235.

In process 237, the process of the action determination program 231 is terminated. The behavior determination program 231 is executed every time sensor data 280 is added.

FIG. 17 shows the process of the action guideline generation program 240 in detail in a flowchart, and it is preferable that the action guideline optimal for improving the target index is generated from the action element data 300 and the target index data 310. An example is shown.

In process 241, the process of the action guideline generating program 240 is started.

In the process 242, necessary conditions are set for generating the action guideline. For example, the range to be analyzed and the day of the week are specified. This is a fixed value determined in advance by a program, or a value input by the user 8 via the input device 4 is used.

In the process 243, the data in the range set in the process 242 is read from the behavior element data 300 and referred to.

In process 244, habit behavior data 330 representing a human behavior pattern is generated. As the unit period for generating one habit behavior data, the period set in the processing 242 is applied. In the processing 244, as shown in FIG. 8, within the unit period, the behavior elements of the same behavior definition having the same time are grouped into one habit behavior data regardless of the date. A habit action ID that is an identification ID is added to these habit actions. In setting the closeness of this time, it depends on the granularity of the action pattern that the user 8 wants to see, but in either case, it is preferably about 1 to 2 hours. However, for habitual behaviors that appear only once or twice a day, such as sleep, it is appropriate to set this time to about 5 hours in order to prevent the connection of behaviors from being separated by turning over, for example. The average of the behavioral elements constituting the grouped habitual behavior is taken and the start time and end time of the habitual behavior are recorded. Moreover, ID of the action element which comprises is recorded. Further, the habit behavior immediately before reaching a certain habit behavior from the time series connection of the behavior elements is recorded as a transition source, and the number of transitions from each transition source is recorded. Information on the habit behavior is recorded in the habit behavior data 330.

In the process 245, all target index data within the analysis range is read and referenced, and further, habit behavior data corresponding to the date and time of the target index data is read. For example, if the target index data is the work efficiency of every weekday in 2008, the habit behavior in 2008 is similarly read from the habit behavior data 330 and compared. In order to analyze, each weekday behavior element corresponding to the target index is read from each habit behavior. Here, when there are a plurality of action elements on the same day, the average of the times is taken as one action element. Next, a correlation analysis is performed between the time and duration of each weekday action element included in each habit behavior, the transition relationship, and the like, and the target variable every weekday, and the correlation coefficient between the habit action and the target variable is calculated.

In the process 246, the habit behavior is displayed in association with the display device 3 so that the order of the correlation coefficient is high.

In the process 247, an action guideline for improving the target variable is displayed on the habit action having the highest correlation coefficient, for example, an instruction to make the time earlier, slower, longer or shorter.

Processing 248 completes the action guideline generation program 240 processing.

FIG. 18 shows an example of the configuration of the sensor node 1 in this embodiment. The sensor node 1 functions as a wireless communication unit 106 including an antenna 115, an acceleration sensor 102, a pulse sensor 103, a temperature sensor 104, a microcomputer 120, and a timer for triggering the microcomputer 120 at regular intervals. Furthermore, a real time clock (RTC) 105 that generates time information, a storage 140 that is a rewritable non-volatile storage medium, an EEPROM (Electrically Erasable and Programmable Read Memory EEPROM) 160, characters, Trigger the microcomputer 120 by detecting the USB connection from the external device to the LCD 101 for displaying the waveform, graph, etc., a plurality of switches 110, 111 that can trigger the microcomputer 120, and the terminal 112. Over An external power supply detection unit 108 that outputs the state to the microcomputer 120, a USB communication unit 107 that transfers data transmitted by serial communication with the microcomputer 120 to an external device connected via USB, a secondary battery 26, a personal computer A charging / power feeding circuit unit 109 that charges the secondary battery 113 with power supplied via a USB connection with an external device such as a computer, or supplies power to the sensor node 1 instead of the secondary battery 113, and a USB The terminal 25 is connected to a cable.

As shown in FIG. 18, the EEPROM 160 has an area 152 for recording the date and time of the event, an event recording table 150 including an area 153 for recording an event ID that is an identification code for classifying the event, and a source for selecting the event ID. The event list table 154 includes an area 155 for recording an event ID that is an item to be displayed, and an area 156 for recording an icon that is data for displaying a corresponding icon image on the LCD 101 when an event ID is selected. . Since the event list table 154 is rewritable, any event that is frequently used can be set and recorded by the user using the sensor node 1.

By adding not only the event occurrence date and time but also the event ID to the event record table 150, the event contents can be recorded by a simple operation, and the contents can be easily remembered later when viewing the event information.

As shown in the figure, the storage 140 includes a data table 141 for recording sensed data. Data recorded in the data table 141 is delimited by a variable length for each packet transmitted from the wireless communication unit 16. Thus, processing can be reduced by enabling data for one packet read from the storage 140 to be transmitted wirelessly or by wire without processing the data as it is.

One packet is recorded at the position of the corresponding address 142, and an area 143 for recording a flag is also included in the data table 141. The address 142 is uniquely assigned in the data table 141, and the storage 140 is read from the microcomputer 100 by serial communication, or refers to the address included in the write command, and is stored in an arbitrary position recorded in the storage 160. Data can be read or written.

In the flag 143, when the data is transmitted wirelessly, 1 is written if all the transmissions are successful, and 0 is written if the data includes transmission failure data. That is, when a packet in the storage 160 is read later, it can be determined whether or not untransmitted data is included, and only untransmitted data can be efficiently read without being left.

The microcomputer 120 includes a CPU 121 that executes arithmetic processing, a ROM 130 that records a program 131 executed by the CPU 121, a RAM 127 that temporarily records data, a real-time clock 105, a storage 140, an EEPROM 160, an LCD 11, and an acceleration sensor. 12, temperature sensor 14, pulse sensor 13, wireless communication unit 16, USB communication unit 17, serial communication units 122 and 126 that transmit and receive signals using digital signals, and I / O port 125 that inputs or outputs digital signals. And interrupt control units 123, 124, and 128 that interrupt the CPU 121 that is executing the program 131 by using an external signal as a trigger.

The programs 131 recorded in advance in the ROM 130 include a sensing program 132, a connection switching program 133, and an event recording program 134. The processing described in the event recording program 134 is executed by the CPU 121, starts processing using signals from the switches 110 and 111 as a trigger, and communicates with the real-time clock 15 via the serial communication unit 102 to acquire time information. , Recorded in the date and time 152 of the event recording table 150 of the EEPROM 160. Further, by operating the switches 110 and 111, an event ID 155 for classifying an event corresponding to the time information can be selected from the pre-programmed event list table 154 and recorded in the event ID 153. In order to assist the user's operation, an icon 156 corresponding to the event ID 155 is displayed on the LCD 101.

The processing described in the sensing program 131 is executed by the CPU 121, and the data sensed by the acceleration sensor 102, the temperature sensor 104, and the pulse sensor 103 is taken into the RAM 127 via the serial communication unit 122 using the signal of the real-time clock 105 as a trigger. Then, the wireless communication unit 106 is controlled to wirelessly transmit the data fetched into the RAM 127 to a predetermined gateway, and further written into the storage 160 and recorded.

The processing described in the connection switching program 133 is also executed by the CPU 121, starts wired communication with a signal from the external power supply detection unit 108 as a trigger, is recorded in the storage 140, and is transmitted to the outside by wireless or wired communication. Unread data (untransmitted data) is read out, transmitted to the USB communication unit 107, and wired from the USB communication unit 107 to the external device.

The real time clock 105 can generate an interrupt signal to the interrupt control unit 123 of the microcomputer 120 at a constant cycle. This period can be changed by a serial communication command. With this interrupt signal, the microcomputer 120 can start the sensing process described in the sensing program 131 at a constant cycle without being affected by the execution state of other processes.

The external power source detection unit 108 detects that the power source of the terminal 112 is connected. That is, it is possible to detect connection with an external device via a USB having a power source. When the connection is detected, the external power supply detection unit 108 generates an interrupt signal to the interrupt control unit 124 of the microcomputer 120 and outputs a digital signal 1 to the I / O port 125. When a disconnection is detected, an interrupt signal is generated and a digital signal of 1 is output to the I / O port 125. That is, the microcomputer 120 can immediately detect a change in the connection state to the terminal 112 and start or stop USB communication via the USB communication unit 107.

The USB communication unit 107 converts a serial communication signal with the microcomputer 120 into a USB standard signal via the data line (transmission and reception) of the USB terminal 112. Therefore, in the control of the microcomputer 120, only data to be transmitted to the external device can be automatically converted to USB standard data and transmitted to the external device only by transmitting to the USB communication unit 107 by serial communication. In addition, since the power of the USB communication unit 107 is supplied only from an external device via the power terminal 112, extra power is not consumed when the USB is not connected.

FIG. 19 shows the external appearance of the sensor node 1 to which the configuration of this embodiment is applied. In FIG. 19, since the USB connection terminal 112 of the sensor node 1 is on the side as viewed from the mounting surface where the pulse sensor 103 is located, the user can always wear it even during wired communication and charging, and sensing can be performed. It is characterized by not obstructing. Each will be described below.

A band 116 is attached to both ends of the sensor node 1 and can be worn on the arm like a typical wristwatch. As can be seen from the structure of the band 116 in FIG. 19, the surface having the pulse sensor 103 contacts the user's skin when worn. The pulse sensor 103 is a known optical measurement method, which irradiates the surface of the living body with infrared rays and makes it possible to estimate the pulse from the change in reflected light caused by the pulsation of the blood vessel. That is, it is essential that the pulse sensor 103 is in contact with the user's skin.

The terminal 112 is on a different side from the mounting surface on which the pulse sensor 103 is located. The terminal 112 corresponds to a USB cable, and can be connected to an external device compatible with USB through a power terminal and a data terminal (transmission and reception), and can communicate and supply power.

Also, the LCD 101 can display the time that the user can always see, the remaining battery level, the wireless communication status, and the like. By displaying the menu on the LCD 101 and operating the switches 110 and 111, the internal functions of the microcomputer 120 such as the CPU 121, the interrupt control unit 128, and various programs in the ROM 130 shown in FIG. The user can change the settings of the wireless channel and the like.

FIG. 20 shows the display on the LCD 101 of the sensor node 1. In FIG. 9, the user 8 can check the waveform display 166 of the sensed data, the time information 164, the radio wave state 161, the remaining battery level 162, and the remaining memory capacity 163 while wearing the sensor node 1.

The radio wave state 161 indicates the result of wireless transmission. It is not displayed when no wireless communication is used.

The battery remaining amount 162 displays the voltage of the secondary battery 113. Thereby, the user 8 can know the timing when charging is required by USB connection.

The remaining memory capacity 163 indicates the amount of untransmitted sensor data that can be recorded in the storage 140. In an environment in which the user 3 cannot perform wireless or wired communication, the time for which untransmitted sensor data can be recorded in the storage 140 is estimated. can do.

FIG. 21 is a diagram showing an example of display transition of the LCD 101 when the process of selecting and recording an event by pressing the switches 110 and 111 is executed. The display example of FIG. 21 is characterized in that since selection and recording of events are performed only by selecting icons, even in busy daily life, selection and recording can be easily performed in a short time. Needless to say, this display switching control can also be realized by the internal function of the microcomputer 120 shown in FIG.

When the switch 110 is pressed from the display state of FIG. 20, the display is switched to the display 167. Time information is acquired from the real-time clock 105 using this time as the event occurrence time. When the event ID is 0 when the event ID is not selected, the display 167 is when the event ID is 0. This is recorded when there is no need to classify or when there is no time selected.

When the switch 110 is further pressed, an icon corresponding to the next event ID is displayed. For example, the display 168 is an icon indicating a meal, and the event ID is 1.

When the switch 110 is further pressed, the display 169 and the display 170 are sequentially switched. For example, the display 169 is an icon indicating that the vehicle is moving, and the display 170 is an icon indicating that the vehicle is running. By pressing the switch 111 while these displays 167 to 170 are displayed on the LCD 101, event IDs (0 to 3) corresponding to the displays can be recorded in the EEPROM 160.

FIG. 22 is a diagram showing a system configuration of the second embodiment. In the configuration of the present embodiment, unlike the embodiment of FIG. 1, the sensor data reception program 501, the action determination program 502, the sensor data 507, the action definition data 506, the action element data 509, and the target index data 510, action history data 511, action definition generation program 504, action guideline generation program 505, action proposal program 514, exercise amount / step count data 512, and habit action data 513 are provided in the server 5 instead of the PC 2. . The server 5 also includes a sensor data reception program 501 that receives data from a base station directly connected to the network 9 and a WEB display generation program 503 that has a WEB server function and generates a display screen of a WEB browser. These programs are executed by a processing unit such as a CPU in the server 5. The PC 2 includes an action guideline generation program 505 and an action guideline / suggestion display program 350 that receives the results generated by the action proposal program 514 via the network 9 and transmits them to the display device 3 for display. And The action proposal program 514 has the same function as the action proposal program 250 shown in FIG.

The PC 2 receives the data from the base station 6 and stores it in the database, and the action history input for storing the action history and comments such as a diary entered by the user 8 using the input device 4 in the database. The program 210, the action guideline generation program 505 of the server 5 and the result of the action proposal program 514 are received via the network 9, and the action guideline / suggestion display program for displaying on the display device 3 is provided. The sensor data reception program 220 stores the sensor data sent from the base station 6 in the sensor data 507 of the server 5 via the network 9. The action history input program 210 stores the result input by the user 8 via the input device 4 in the action history data 511 via the network 9. That is, the PC 2 does not hold a database. As a result, as long as the PC is connected to the network 9, the sensor data is collected from anywhere and analyzed by the server 5 without limitation, and the result can be received and displayed. Even when the processing of the behavior determination program 502, the behavior definition program 504, the behavior guideline generation program 505, and the behavior proposal program 514 is complicated and cannot be treated by the PC 2, the server 5 often has few restrictions on size and power consumption. It is possible to complete the processing in a short time with a high processing capacity at a low cost.

The behavior determination program 502 calculates the amount of exercise and the step count data 512 from the sensor data 507 in the same manner as the behavior determination program 230 of FIG. 1, and determines the behavior name and the corresponding threshold value stored in the behavior definition data 506. Using the list of conditions, the momentum and step count data 512 are discriminated in time series. As a result, the action name or the identification ID of the action, the time, and the time information are stored in the action element data 509 as a single action element that matches the determination condition.

The target index data 510 records an index that the user 8 desires to improve over a long period of time. Unlike the target index data 310 of FIG. 1, scores such as results measured by other sensors and work results input via other PCs can also be stored.

The action guideline generation program 505 generates habit action data 513 in the same manner as the action guideline generation program 240 of FIG. 1, and the time series changes of the action element data 509 and the target index data 310 included in each habit action and the correlation Analyze and generate an action guideline that is optimal for the user 8. Unlike the action guideline generation program 240, the action guideline generation program 505 can be started by a request from the action guideline / suggestion display program 350 of the PC2. The generated result is transmitted to the action guideline / suggestion display program 350, and the user 8 can view the result using the PC 2.

Similarly, the action proposal program 514 starts in accordance with the request of the action guideline / suggestion display program 350 or at the same time as the end of the action guideline generation program 505, and transmits the result to the action guideline / suggestion display program 350. Can be viewed at. Other processes are the same as those of the action proposal program 250 of FIG.

The behavior definition generation program 504 can correct the behavior definition data 290 or add a new definition with reference to the behavior history data 511 and the sensor data 507, similarly to the behavior definition generation program 261 of FIG. Further, unlike the behavior definition generation program 261, analysis processing can be performed using data of other users other than the user 8 stored in the sensor data 507 and the behavior history data 511. And diversification of definitions.

FIG. 23 is a diagram showing the system of the third embodiment. In this configuration, the general-purpose WEB browser 520 performs the same processing instead of requiring the action guideline / suggestion program 350 shown in FIG. Have According to the present embodiment, by using the WEB browser 520, it is not necessary to provide a dedicated program in the PC, and it is easy to act from many PCs or mobile phones or terminals having similar functions regardless of location. It is possible to view the guidelines / proposed results. Further, unlike the base station 6 shown in FIG. 1 or FIG. 22, the base station 16 can directly connect to the network 9 and send the received sensor data to the server 5.
The server 5 includes a sensor data program 501 similar to the sensor data reception program 220 in FIG. 1, so that data transmitted from the base station 16 via the network 9 can be stored in the sensor data 507. That is, if a plurality of base stations 16 are installed in an office or home and connected to the network 9, it is not necessary to connect and manage them with a personal PC or the like, and sensor data can be collected easily. The WEB input / output generation program of the server 5 has the function of the WEB server, and transmits display contents generated by the action guideline generation program 505 and the action proposal program 514 to the WEB browser 520 in accordance with a request from the WEB browser 520. The WEB browser 520 displays the received display content on the display device 3. The WEB input / output generation program 503 transmits an action input screen 440 generated by the action history input program 512 in accordance with a request from the WEB browser 520, and the WEB browser 520 displays the display on the display device 3. Further, the action history input by the user 8 via the input device 4 can be transmitted from the WEB browser 520 to the WEB input / output generation program 503 and stored in the action history data 511 via the action history input program 512.

INDUSTRIAL APPLICABILITY The present invention is useful as a technique that can be worn on a human body, shows the characteristics of life by estimating human behavior from information on sensor terminals that measure biological information and behavioral states, and presents improvement means. high.

DESCRIPTION OF SYMBOLS 1 ... Sensor node 2 ... Personal computer 7 ... Antenna 8 ... User 9 ... Network 11 ... Acceleration scalar value 12 ... Intersection 105 ... Real time clock (RTC)
110, 111 ... switch 112 ... terminal 113 ... secondary battery 114 ... ground 115 ... antenna 120 ... microcomputer)
121. Calculation unit (CPU)
127 ... Volatile memory (RAM)
130 ... Nonvolatile storage (ROM)
132 ... Sensing program 133 ... Connection switching program 134 ... Event recording program 141 ... Packet recording table 153 ... Event recording table 154 ... Event list table 160 ... Electrically rewritable nonvolatile storage (EEPROM)
161 ... Wireless communication status display 162 ... Battery voltage display 163 ... Memory remaining amount display 164 ... Time display 165 ... Sensing status display 166 ... Sensor waveform display 167-170 ... Event display 231-237, 241-248 ... Processing 401 ... Weekdays Habit action display 402 ... holiday habit action display 429 ... action legend 442 ... amount of exercise 443 ... action element 444 ... time 445 ... action history 403 to 428, 608 to 615 ... habit action 601 to 607 ... groups of similar action elements.

Claims (20)

1. A behavior suggestion device comprising a processing unit and a storage unit, and performing personal behavior proposal from personal biometric information,
   The processor is
   Based on the collected biometric information of the individual, the behavior of the individual is determined as behavior element data,
   From the behavior element data of a predetermined period, generate the personal habit behavior data,
   Based on the behavior element data, the habit behavior data, and the individual target index data, to generate the individual behavior guidelines,
   Based on the generated action guideline, the action proposal for the individual is performed.
   An action suggestion device characterized by that.
2. The action proposing device according to claim 1,
   A display unit for displaying the action guideline and the action proposal;
   An action suggestion device characterized by that.
3. The action suggestion device according to claim 2,
   The processor is
   A sensor data receiving unit that collects the biological information from a sensor corresponding to the individual, and the sensor data receiving unit stores the behavior element data extracted from the biological information in the storage unit;
   An action suggestion device characterized by that.
4. The action proposing device according to claim 3,
   The processing unit includes an action guideline generating unit that generates the action guideline based on the target index data, the action element data, and the habit action data,
   The action guideline generating unit displays the generated action guideline on the display unit.
   An action suggestion device characterized by that.
5. The action proposing device according to claim 4,
   The processing unit has an action proposal unit that outputs the action proposal based on the action guideline, and displays the action proposal on the display unit.
   An action suggestion device characterized by that.
6. The action proposing device according to claim 5,
   The processing unit has an action history input unit for inputting the individual action history,
   The action history input unit accumulates the input action history in the storage unit,
   An action suggestion device characterized by that.
7. The action proposing device according to claim 6,
   The action suggesting unit outputs the action proposal based on the action guideline and the accumulated action history.
   An action suggestion device characterized by that.
8. The action proposing device according to claim 4,
   The behavior guideline generation unit generates the habit behavior data based on the behavior element data of a predetermined period, and holds the generated habit behavior data in the storage unit,
   An action suggestion device characterized by that.
9. The action proposing device according to claim 8,
   The behavior guideline generation unit performs a correlation analysis between the habitual behavior data held for a plurality of the predetermined periods and the target index data, and displays a correlation analysis result on the display unit.
   An action suggestion device characterized by that.
10. The action proposing device according to claim 9,
    The behavior guideline generation unit adds the habitual behavior data having high correlation in the correlation analysis result to the behavior guideline and displays it on the display unit.
    An action suggestion device characterized by that.
11. A behavior suggestion method in a system that includes a processing unit, a storage unit, and a display unit, and that proposes personal behavior from personal biometric information,
    The processor is
    Based on the collected biometric information of the individual, the behavior element data is extracted by determining the behavior of the individual,
    From the behavior element data of a predetermined period, generate the personal habit behavior data,
    Based on the behavior element data, the habit behavior data, and the individual target index data, to generate the individual behavior guidelines,
    Based on the generated action guideline, the action proposal for the individual is performed.
    An action suggestion method characterized by that.
12. The action suggestion method according to claim 11,
    The processing unit displays the action guideline and the action proposal on the display unit.
    An action suggestion method characterized by that.
13. The action suggestion method according to claim 12,
    The processor is
    Collecting the biological information from a sensor corresponding to the individual,
    Storing the behavior element data extracted from the collected biological information in the storage unit;
    An action suggestion method characterized by that.
14. The action suggestion method according to claim 13,
    The processor is
    Based on the target index data, the behavior element data and the habit behavior data, generate the behavior guideline,
    Displaying the generated action guideline on the display unit;
    An action suggestion method characterized by that.
15. 15. The action proposing method according to claim 14, wherein
    The processor is
    Storing the input action history data of the individual in the storage unit;
    Based on the stored action history data and the action guideline, the action proposal is output and displayed on the display unit.
    An action suggestion method characterized by that.
16. 15. The action proposing method according to claim 14, wherein
    The processing unit generates the habit behavior data based on the behavior element data of a predetermined period,
    Holding the generated habit behavior data in the storage unit,
    An action suggestion method characterized by that.
17. The action proposing method according to claim 16,
    The processing unit performs a correlation analysis between the habit behavior data and the target index data for a plurality of the predetermined periods, and displays a correlation analysis result on the display unit.
    An action suggestion method characterized by that.
18. The action proposing method according to claim 17,
    The processing unit adds the habit behavior data having a high correlation in the correlation analysis result to the behavior guideline and displays the data on the display unit.
    An action suggestion method characterized by that.
19. The action proposing method according to claim 15,
    The processor is
    A behavior suggestion method comprising: outputting the behavior suggestion based on the inputted behavior history data of another person, the stored behavior history data, and the behavior guideline.
20. The action proposing method according to claim 19,
    The processor is
    A behavior suggestion method, wherein the behavior history data of the other person that matches the behavior guideline or the stored behavior history data is searched, and the matched behavior history is displayed on the display unit as the behavior proposal.

Images