JP2007190306A - Life watching system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a life watching system for collecting biological information with the use of a sensor mounted on a resident person, discriminating the abnormality of a life rhythm, and reporting it. <P>SOLUTION: The life watching system comprises: a device to be mounted on a living body, including a body action sensor; a part for discriminating a life state, based on the detection signal of the device; a life rhythm discrimination part for calculating the featured value of the life rhythm, based on the time sequential detection signal of a life state; a knowledge database to be used for discriminating the abnormality of the featured value; an alert discrimination part for discriminating the abnormality of the life rhythm; and a part for outputting alert information, etc. The knowledge database stores a reference hour of rising and the reference value of the featured value of an action pattern corresponding to a difference from the reference hour of rising. The life rhythm discrimination part calculates an hour of rising and the featured value of the action pattern after the hour of rising. The alert discrimination part changes the reference value in response to a difference between the reference hour of rising and the hour of rising and discriminates the abnormality of the action pattern after getting up. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a system that collects biometric information of residents in elderly facilities such as care apartments, pay nursing homes, health care facilities for the elderly, etc., determines abnormal physical condition, and notifies a staff who watches the lives of the residents.

  In recent years, with the declining birthrate, aging, longevity, and the progress of the nuclear family, the number of elderly households consisting of only couples or singles over 60 years of age is increasing rapidly. For this reason, elderly facilities such as care condominiums and paid nursing homes where facility staff (hereinafter referred to as staff), such as doctors, nurses, care workers and life counselors, take care of residents and take care of their lives Has increased. In such elderly facilities, many residents live in the facility, but accidents such as falling from the bed, poor physical condition such as cold and malaise, dementia, disuse syndrome, depression There is a need for watching support that enables the monitoring staff to immediately identify and quickly respond to residents who have some physical condition such as symptoms, stress, and early signs of autonomic abnormalities.

Conventionally, there is a system for detecting resident information with various sensors installed in the facility for such watching support.
For example, in the “sleeping monitoring device” described in Japanese Patent Laid-Open No. 2000-214, a signal such as a heart rate while sleeping is extracted using a pressure sensor that detects the internal pressure of the air mat, and an abnormal alarm is notified.
Further, for example, in the “health management system” described in Japanese Patent Application Laid-Open No. 2002-245168, long-term bed leaving, long-term bed staying, and the like are determined and emergency notification is made using portable and fixed health state measuring means.
Further, for example, in the “life watching system” described in Japanese Patent Application Laid-Open No. 2003-109159, life movement information is received by a human sensor or the like, and a life pattern is displayed.
Further, for example, in the “monitoring device” described in Japanese Patent Application Laid-Open No. 2003-235813, a determination criterion that matches a biological clock for each user is set based on time-series data of a biological signal, and a notification signal is generated.

JP 2000-214

JP 2002-245168 JP 2003-109159 A JP 2003-235813 A

However, the conventional technology has a problem in that abnormal physical condition according to the life rhythm of each individual, which is important for watching the resident's life, cannot be determined. In particular, there was a problem that the abnormal behavior pattern after waking up was unknown when various physical conditions were abnormal, such as colds and malaise.
An object of the present invention is to provide a system that collects biological information using a sensor attached to a resident and determines and notifies an abnormal state of a life rhythm.

  The above-described problems include a wearable device that includes a body motion sensor that calculates a body movement amount and is worn on a living body, and a life state that determines a life state at a specific date and time including a sleep state based on a detection signal of the wearable device. A determination unit; a life rhythm determination unit that calculates a life rhythm feature value from the life state time series and a detection signal; a knowledge database that stores knowledge data that determines whether there is an abnormality in the feature value; A life monitoring system comprising an alert determination unit that determines an abnormality in a life rhythm using the feature amount, and an output unit that outputs a determination result of the alert determination unit as alert information, the knowledge database The reference value of the feature quantity of the action pattern according to the difference between the reference wake-up time and the reference wake-up time is stored as knowledge data, and the life rhythm determination unit The wake-up time, which is the time to transition to another living state, and the feature quantity of the behavior pattern after the wake-up time are calculated, and the alert determination unit determines the reference value according to the difference between the reference wake-up time and the wake-up time This can be solved by a lifestyle monitoring system characterized by changing behavior and determining abnormal behavior patterns after waking up.

With the life monitoring system according to the present invention, the staff can immediately grasp the abnormal physical condition of the resident and can quickly respond.
Furthermore, since the user can get a sense of being watched by the staff, it is possible to live with peace of mind in the facility.

  FIG. 1 shows a configuration diagram of a life watching system that is an embodiment of the present invention. This system includes a data server 100, a staff terminal 120, a network 140, a wearable device 150, and a wireless repeater 170.

  In the present embodiment, the staff terminal 120 uses one or more personal computers having an input unit such as a keyboard and a mouse, an output unit such as a display, and a communication unit communicating with the data server 100 and the like. A portable terminal such as a PDA, PHS, or mobile phone having an input unit, an output unit, and a communication unit can also be used.

  In this system, the data server 100 is installed in a management section where system management staff can enter and leave, and the staff terminal 120 is installed in a staff room where staff can enter and leave. The staff terminal 120 is used by staff.

  The wearable device 150 is used by a resident (hereinafter referred to as a user) of a facility wearing the body. Further, the wireless repeater 170 is installed in a resident's room or a common area where the resident can enter and exit.

  The data server 100 may be installed in a data center managed by a customer center who is an operator of the system. As a result, privacy information such as user personal information and biometric data collected from the user can be centrally managed, and security management such as information leakage prevention can be simplified.

  In this system, the data server 100 is installed in the data center, the network 140 is the Internet, the wireless repeater 170 is installed in a detached house (hereinafter referred to as a living room), and the wearable device 150 is installed. The resident of the living room (hereinafter referred to as user) can also wear it on his body. As a result, this system can also be used in detached houses.

  The data server 100 includes a DS control unit 101, a DS memory 102, a life state determination unit 105, a life rhythm determination unit 106, an alert determination unit 107, a DS communication unit 108, a knowledge base generation unit 109, A user information database 111, a data storage database 112, a knowledge base 113, and an alert history database 114 are included.

  The wearable device 150 includes a WD control unit 151, a WD memory 152, a WD_ID storage unit 153 that stores a WD_ID that is a unique ID for identifying the wearable device, a WD calculation unit 154, and a WD wireless communication unit 155. , A pulse wave sensor 156 for detecting a change in blood flow in a blood vessel accompanying a heart beat of a living body, a temperature sensor 158 for detecting a heat energy amount generated from the living body, and a body for detecting a body activity as a body motion. And a motion sensor 159.

In the present embodiment, “wd01” is stored in the WD_ID storage unit 153.
The wearable device 150 performs wireless communication with the wireless repeater 170.

  The wearable device 150 is incorporated in a wristband, and a user wears it on his wrist. It can also be incorporated into things worn on the body, such as watches, rocker keys, pendants and rings, clothing, shoes, hats, and glasses. It can also be used by directly sticking it to the skin, such as a banquet or a compress.

  FIG. 14 shows a configuration diagram when the wearable device 150 is incorporated in a wristband. The pulse wave sensor 156 and the temperature sensor 158 are configured to be close to or in contact with the user's skin. Thereby, it becomes possible to detect a pulse wave and a body temperature. A body motion sensor 159 is incorporated in the wristband so as to detect movement in a direction parallel to the back of the hand. Thereby, it becomes possible to detect a user's body movement efficiently.

  The body motion sensor 159 may use an impact sensor. Thereby, power consumption can be reduced. The body motion sensor 159 may use an acceleration sensor. As a result, body movement can be detected with high accuracy.

  In the network 140, the data server 100, the staff terminal 120, and the wireless repeater 170 are connected. The data server 100 communicates with the staff terminal 120 and the wireless repeater 170 via the network 140.

  The network 140 uses wired communication using a LAN (Local Area Network) cable. In addition to power line communication (PLC, Power Line Communication), other wired communication, or wireless communication such as IEEE802.11b, a unique communication method can also be used.

  In addition, the network 140 may use other local networks such as a local PHS, or other wide area networks such as the Internet, VPN, mobile phone communication network, and PHS communication network.

  Although this system is described as a hardware configuration, some of the functions of this system may be configured by software.

  FIG. 2 shows an example 200 of the user information database 111. The database example 200 includes a user personal information table 210 that manages user personal information. The user personal information table 210 includes a field 211 for storing a user ID for identifying a user, a field 212 for storing a user password used when the user performs user authentication when using the system, and personal information such as a name. A field 213 for storing and a field 215 for storing a WD_ID for identifying a wearable device used by the user. For example, in the example 210A of the user personal information record in the table example 210, the user with the user ID “user0001” and the name “F. Kuri” uses the wearable device identified by the WD_ID “wd01”. Show.

  FIG. 3 shows an example 300 of the data storage database 112. The data storage database 112 includes a user data history table 310. The user data history table 310 is calculated from a field 311 for storing the user ID, a field 312 for storing SD data including user biometric data transmitted from the wearable device 150, and the user biometric data. A field 313 for storing a life state ID for identifying a life state and a field 314 for storing the date and time when the SD data is received are configured.

  In this embodiment, the life state ID “S_SL” identifies the life state “sleeping”, the life state ID “S_ST” identifies the life state “resting”, and the life state ID “S_AC” identifies the life state “active”. Defined as ID.

  FIG. 4 shows an example 400 of the knowledge base 113. The knowledge base 114 includes a life state determination knowledge table 420 that stores knowledge for determining a life state, and a life rhythm abnormality determination knowledge table 440 that stores knowledge for determining the presence / absence of an abnormality from the feature amount of the life rhythm. Is done.

  The living state determination knowledge table 420 includes a field 421 for storing the user ID, a field 422 for storing a living state ID for identifying a living state, and a determination condition for determining a living state identified by the living state ID. Is composed of a field 423 for storing the information and a remarks field 424.

  For example, in the case of the life state determination knowledge record 420C, when the body motion Ac is 50 or more for the user identified by the user ID “user0001”, the life state ID “S_AC”, that is, the life state “active” is set. It indicates that it is determined.

  The life rhythm abnormality determination knowledge table 440 stores a field 441 for storing the user ID, a field 442 for storing a feature amount ID for identifying a feature amount of the life rhythm, and various parameters for determining abnormality of the feature amount. Fields 443 to 445 and a remarks field 446.

  The field 443 stores an average value, and the field 444 stores a condition for determining abnormality of the feature amount.

  In this embodiment, the feature amount ID “SL_S” is the feature amount “sleep start time (sleeping time)”, the feature amount ID “SL_E” is the feature amount “sleep end time (wake-up time)”, and the feature amount ID “SL_L” is The feature amount “sleep time”, the feature amount ID “ST_L” is the feature amount “rest time”, the feature amount ID “AC_L” is the feature amount “activity time”, and the feature amount ID “SL_Ps” is the feature amount “average pulse during sleep”. The number “,” the feature amount ID “ST_Ps” is defined as an ID for identifying the feature amount “average pulse rate at rest”, and the feature amount ID “AC_Ps” is the feature amount “average pulse rate during activity”.

  Further, the feature amount ID “SL_A1” identifies a feature amount “body motion when getting up”, which is the first body motion amount at the end of sleep (when waking up), which is a sign of physical condition such as falling from the bed. Defined as ID.

  In addition, the feature amount ID “SL_A60” is a sign of symptoms such as a physical condition such as a cold or malaise and depression, and a feature amount “post-wake body movement” after the end of sleep (after waking up), Is defined as an ID for identifying

  Further, the feature amount ID “ACR” is defined as an ID for identifying a feature amount “activity level” that is a degree of activity time during the day, which is a sign of abnormal physical condition such as dementia and disuse syndrome.

  The feature amount ID “PsR” identifies a feature amount “pulse fluctuation rate” which is a ratio of the average pulse rate at rest to the average pulse rate during sleep, which is a sign of abnormal physical condition such as an autonomic abnormality. Defined as ID.

  For example, in the case of the life rhythm abnormality determination knowledge record 440B, the user identified by the user ID “user0001” is the feature amount ID “SL_E”, that is, the parameter “average value” of the feature amount “sleep end time (wake-up time)” is “ 06:00 (6:00 am) ”, indicating that the parameter“ condition ”for judging abnormality is“ greater than 07:00 (7:00 am) ”.

  For example, in the case of the life rhythm abnormality determination knowledge record 440G, the user identified by the user ID “user0001” is the feature amount ID “SL_A60”, that is, the parameter “average value” of the feature amount “post-wake body movement amount” is “60”. ”, The standard body motion amount 60 when the parameter“ condition ”for determining abnormality is the transition of the standard body motion amount“..., T0 (reference wake time), T10 (reference body motion amount 70 after 10 minutes from the reference wake time, ..., T60 (reference body movement 80 after 60 minutes from the reference wake-up time, ...).

  In this embodiment, the transition of the standard body movement is set as the standard body movement for each difference with the difference from the standard wake-up time as 10 minutes, but is set at other intervals such as 5 minutes and 15 minutes. You can also. In addition, the transition of the reference body movement amount can be set as a function.

  FIG. 5 shows an example 500 of the implementation record database 114. The implementation record database 114 includes an implementation record history table 510. The implementation record history table 510 stores a field 511 for storing the user ID, a field 512 for storing the date and time when the alert is output, a field 513 for storing the type of alert, and an implementation record of the staff for the alert. Field 514 to 515.

  The field 514 stores the date and time when the staff responds, and the field 515 stores an abnormality flag indicating whether or not the user is abnormal as a result of the staff.

  For example, in the case of the execution record history record example 510C, the alert type “SL_E” (at the end of sleep (when waking up) is given to the user with the user ID “user0001” at 7:01 on April 3, 2004. This indicates that an alert indicating that the amount of movement is abnormal) is being output. In addition, as an execution record, it is shown that the staff responded to this alert at 7:03 on April 3, 2004 and actually input that the user had a physical condition abnormality.

  Next, the operation of this system will be described using a flowchart.

  FIG. 6 is a flowchart showing the operation of the wearable device 150. First, when the wearable device 150 starts operating, the WD control unit 151 executes step 601 for determining whether or not to end. If it is determined in step 601 that the operation is to be ended, the operation is ended.

  If it is determined in step 601 that the process does not end, the WD control unit 151 executes step 602 that waits for a waiting time Tw. In this embodiment, the standby time Tw is set to “60 seconds”, but any time may be used.

  Next, the WD control unit 151 executes Step 603 for determining whether or not the user wears the wearable device 150. For example, when the output value of the pulse wave sensor 156 is smaller than a predetermined value, it is determined that the device is mounted, and when the output value is greater than the predetermined value, it is determined that the device is not mounted.

  If it is determined in step 603 that it is not mounted, the WD control unit 151 executes step 601 and repeatedly executes the subsequent processing.

  If it is determined in step 603 that the WD control unit 151 is worn, the WD control unit 151 activates the WD calculation unit, the pulse rate Ps from the pulse wave sensor 156, the body temperature BT from the temperature sensor 158, and the body temperature BT. Step 604 of calculating the body motion Ac from the body motion sensor 159 as biological data is executed.

  For example, when calculating the pulse rate Ps, the peak value is calculated from the pulse wave output from the pulse wave sensor 156, the number of peak values in the past 60 seconds from the current time is counted, and calculated as the pulse rate Ps.

  In addition, for example, when calculating the body temperature BT, the body temperature BT is calculated using a conversion formula for converting the voltage value into the body temperature based on the voltage value output from the temperature sensor 158.

  For example, when calculating the body motion Ac, a value obtained by adding the scalar values corresponding to the magnitude of the body motion output from the body motion sensor 159 is calculated as the body motion Ac.

  Further, for example, when calculating the body motion Ac, the number of times that the scalar value corresponding to the magnitude of the body motion output from the body motion sensor 159 shows a value greater than a predetermined value from a value less than a predetermined value, and a predetermined value. A value obtained by adding up the number of times indicating a predetermined value or less from a larger value can also be calculated as the body movement Ac.

  Next, the WD control unit 151 activates the WD wireless communication unit 155 and executes Step 605 for transmitting the biometric data calculated in Step 604 as WD data together with the WD_ID.

  For example, when the biological data is a pulse rate Ps “80”, a body temperature BT “36.2”, and a body motion Ac “70”, the WD data is “WD_ID = wd01, Ps = 80, BT = 36.2, Ac = 70 "(referred to as WDDATA01).

  The wireless repeater 170 receives the WD data transmitted in step 605.

  FIG. 7 is a flowchart showing the operation of the wireless repeater 170 when the wireless repeater 170 receives the WD data.

  First, when the wireless repeater 170 starts operation, the wireless repeater 170 executes step 701 for determining whether or not to end. If it is determined in step 701 that the process is to be terminated, the operation is terminated.

  If it is determined in step 701 not to end, the wireless repeater 170 executes step 702 of receiving the WD data transmitted by the wearable device 150.

  Next, the wireless repeater 170 adds SD_ID (in this embodiment, SD_ID “td10”) for identifying the wireless repeater 170 to the received WD data, and transmits the SD data as step 703. Execute.

  For example, when the WDDATA01 is received, “WD_ID = wd01, SD_ID = td10, Ps = 80, BT = 36.2, Ac = 70” (referred to as SDDATA01) is transmitted as SD data.

  In the present embodiment, wireless communication with the WD wireless communication unit 155 of the wearable device 150 and the wireless repeater 170, and the WD wireless communication unit 155 of the wearable device 150 and the stationary device 160 are performed. The wireless communication with the SD wireless communication unit 165 is assumed to be IEEE 802.15.4, Zigbee. As a result, power consumption can be reduced.

  Further, the wireless communication may be other wireless communication such as weak wireless, specific low power wireless, Bluetooth, IEEE 802.11b, or the like. This makes it possible to construct a more versatile system.

  The data server 100 receives the SD data transmitted in the step 703 via the network 140.

  FIG. 8 is a flowchart showing the operation of the data server 100 when receiving data. First, when the data server 100 starts operation, the DS control unit 101 executes Step 801 for determining whether or not to end the operation at the time of reception. If it is determined in step 801 that the process is to be ended, the present operation is ended.

  If it is determined in step 801 that the process does not end, the DS control unit 101 executes step 802 of receiving the SD data transmitted from the wireless repeater 170.

  Next, the DS control unit 101 activates the living state determination unit 105 and uses the knowledge of the living state determination knowledge table 420 based on the SD data received in step 802 to determine a step 803 for determining a living state. Execute.

  For example, in the case of SDDATA01, the body movement “Ac” is 70 and matches the determination condition “Ac ≧ 50”, so the life state is “active state (active)” (life state ID is “S_AC”). .

  For example, when the SD data is “WD_ID = wd01, SD_ID = td10, Ps = 71, BT = 36.2, Ac = 20” (SDDATA02), the body motion “Ac” is 20, and the determination condition Since it matches “15 ≦ Ac <50”, the living state is “resting state (resting)” (the living state ID is “S_ST”).

  For example, when the SD data is “WD_ID = wd01, SD_ID = td10, Ps = 62, BT = 35.8, Ac = 10” (SDDATA03), the body movement “Ac” is 10, and the determination condition Since it matches with “Ac <15”, when the duration matching this determination condition “Ac <15” is 30 minutes or more (determination condition “L ≧ 30”), the life state is “sleep state (sleeping)”. (Life state ID is “S_SL”).

  Here, the duration is 30 minutes or more, but it can be set to any value.

  As described above, by using the living state determination knowledge base, it is possible to determine the living state of “resting”, “active”, and “sleeping” from the biological information collected by the wearable device 150.

  Also, by adding the duration to the condition for determining the sleep state, it is possible to prevent the sleep state from being erroneously determined when performing static daily operations such as newspapers, reading, and PC operations. , It becomes possible to determine the life state more accurately.

  In the living state knowledge table 420, the living state is determined only by the body movement Ac, but the pulse Ps and the body temperature BT can be combined. Thereby, it becomes possible to determine a life state more accurately.

  Next, the DS control unit 101 determines that the SD data received in Step 802, the life state calculated in Step 803, the date and time when the SD data was received, and the WD_ID included in the SD data are the individual user. Step 804 of storing in the data storage database 112 together with the user ID of the record that matches the WD_ID field 215 of the information table 210 is executed.

  For example, when the SDDATA01 is received at 6:32 on April 2, 2004, the record 310A is stored in the user data history table 310.

  In this way, all received SD data is stored in the user data history table 310 together with the received date and time.

  Next, the DS control unit 101 activates the life rhythm determination unit 106, calculates a life rhythm feature value based on the user data stored in the user data history table 310, and the DS memory 102 Step 805 is stored.

FIG. 9 is a flowchart showing the operation when the life rhythm is determined in step 805. First, the life rhythm determination unit 106 executes steps 901 to 902 in order to extract a sleep period that is a duration of a sleep state that is the basis of a life rhythm.
In step 901, the life rhythm determination unit 106 extracts a date from the current date and time, searches the user data history table 310 for user data after a predetermined time (for example, 0:00) of the date, and first sleeps. When a state changes to another life state, that is, when the life state ID is changed from “S_SL” to other than “S_SL”, a record is extracted, and the date / time in the date / time field 314 of the record is extracted as a feature quantity “sleep end time” (Feature amount ID “SL_E”).

  Next, in step 902, the life rhythm determination unit 106 searches past user data from the feature amount “sleep end time” set in step 901, and first changes from a sleep state to a state other than a sleep state, that is, A record when the life state ID is changed from “S_SL” to other than “S_SL” is extracted, and the date / time in the date / time field 314 of the record is set as the feature amount “sleep start time” (feature amount ID “SL_S”) .

  Thereby, the sleep period from the sleep start time to the sleep end time can be extracted without the sleep period being dispersed on the day and the previous day. Moreover, even when the sleep start time is not the previous day but the current day, the sleep period can be accurately extracted.

  Next, the life rhythm determination unit 106 determines the difference between the feature value “sleep end time” set in step 901 and the feature value “sleep start time” set in step 902 as a feature value “sleep time”. Step 903 for setting as (feature amount ID “SL_L”) is executed.

  Next, the life rhythm determination unit 106 searches for user data from the feature amount “sleep start time” set in step 902 to the feature amount “sleep end time” set in step 901, and the SD data field Step 904 of calculating the average value of the pulse rate Ps stored in 312 and setting it as the feature value “average pulse rate during sleep” (feature value ID “SL_Ps”) is executed.

  In this way, by extracting the sleep period, that is, by setting the sleep end time and sleep start time, the sleep state during the day can be excluded, and the sleep time and sleep during as a characteristic amount of life rhythm The pulse rate can be accurately and easily calculated.

  Next, the life rhythm determination unit 106 detects the feature quantity “sleep start time” next from the period from the sleep period to the next sleep period, that is, the feature quantity “sleep end time” set in step 901. Step 905 is performed in which user data until the search is retrieved, a record whose life state ID is “S_ST” is extracted, and the combined time is set as a feature quantity “rest time” (feature quantity ID “ST_L”).

  Next, the life rhythm determination unit 106 searches for user data from the feature quantity “sleep end time” set in step 901 until the next feature quantity “sleep start time” is detected, and the life state ID is The record of “S_ST” is extracted, the average value of the pulse rate Ps stored in the SD data field 312 is calculated, and set as the feature quantity “average pulse rate during rest” (feature quantity ID “ST_Ps”) Step 906 is executed.

  Next, the life rhythm determination unit 106 searches for user data from the feature quantity “sleep end time” set in step 901 until the next feature quantity “sleep start time” is detected, and the life state ID is The record of “S_AC” is extracted, and step 907 is performed in which the combined time is set as the feature amount “activity time” (feature amount ID “AC_L”).

  Next, the life rhythm determination unit 106 searches for user data from the feature quantity “sleep end time” set in step 901 until the next feature quantity “sleep start time” is detected, and the life state ID is The record of “S_AC” is extracted, the average value of the pulse rate Ps stored in the SD data field 312 is calculated, and set as the feature value “average pulse rate during activity” (feature value ID “AC_Ps”) Step 908 is executed.

Thus, by extracting the sleep period, it is possible to accurately and easily calculate the activity time other than the sleep period, the pulse rate during the activity, the rest time, the pulse rate during the rest, and the like.

  Next, the life rhythm determination unit 106 searches for the first user data after the end of the sleep period, that is, from the feature amount “sleep end time” set in the step 901, and stores it in the SD data field 312. Step 909 is executed to extract the existing body motion Ac and set it as the feature amount “body motion when waking up” (feature amount ID “SL_A1”).

Next, the life rhythm determination unit 106 searches for user data from the feature amount “sleep end time (wake-up time t0)” set in step 901 to 60 minutes after the reference wake-up time (t60). As shown in FIG. 1, step 910 is executed in which the average value of the body movement Ac stored in the SD data field 312 is calculated and set as the feature quantity “post-wake body movement quantity” (feature quantity ID “SL_A60”). .

tn
SL_A60 = ΣAc (t) / (tn−t0) Equation 1
t = t0
However, if t0 <reference wake-up time, tn = t0 + 60
If t0 ≧ standard wake-up time, tn = t60

In this way, by setting tn, it is possible to determine the presence or absence of an abnormality within 60 minutes after wakeup and within 60 minutes from the reference wakeup time at the latest, so that the staff can take quick action when there is an abnormality. .

  Here, the time is set on the basis of 60 minutes, but can be set to an arbitrary time.

Next, the life rhythm determination unit 106 executes step 911 for calculating the feature amount “activity” (feature amount ID “ACR”) using Equation 2.

ACR = AC_L / (AC_L + ST_L) Equation 2

Next, the life rhythm determination unit 106 executes Step 912 of calculating the feature amount “pulse fluctuation rate” (feature amount ID “PsR”) using Equation 3.

PsR = ST_Ps / SL_Ps Formula 3

Next, the life rhythm determination unit 106 calculates the day of the week from the current date and time, and executes Step 913 to determine whether it is Friday. If it is determined NO in step 913, step 919 is executed.
If it is determined in step 913 that it is Friday, the life rhythm determination unit 106 calculates an average value of the feature amount “sleeping time” (feature amount ID “SL_L”) for the past one week, and the feature amount “average” Step 914 is set as “sleep time” (feature value ID “SL_L7”).

  Next, the life rhythm determination unit 106 calculates an average value of the feature amount “rest time” (feature amount ID “ST_L”) for the past one week, and calculates the feature amount “average rest time” (feature amount ID “ST_L7”). Step 915 is set.

Next, the life rhythm determination unit 106 calculates an average value of the feature amount “activity time” (feature amount ID “AC_L”) for the past one week, and the feature amount “average activity time” (feature amount ID “AC_L7”). ”) Is executed.
Next, the life rhythm determination unit 106 calculates an average value of the feature amount “activity” (feature amount ID “ACR”) for the past one week, and calculates the feature amount “average activity” (feature amount ID “ACR7”). Step 917 to be set as “)” is executed.

  Next, the life rhythm determination unit 106 calculates an average value of the feature amount “pulse fluctuation rate” (feature amount ID “PsR”) for the past one week, and the feature amount “average pulse fluctuation rate” (feature amount ID). Step 918 is set as “PsR7”).

  Next, the life rhythm determination unit 106 executes Step 919 for storing the feature amount set in Steps 901 to 918 in the DS memory 102, and ends this operation.

  In this embodiment, it is determined whether or not it is Friday in the step 911, but it can also be set for other days of the week.

Here, in order to examine the rehabilitation menu for one week this week on Friday, it is determined whether or not it is Friday in step 911. However, if the rehabilitation menu is considered on Monday, it is determined whether or not it is Monday in step 911. You can also
As a result, the rehabilitation menu can be examined while referring to the presence or absence of abnormalities in the daily rhythm of the past week.

  FIG. 10 shows an example 1000 of the DS memory 102 after execution of the step 805. It can be seen that the feature values set in steps 901 to 918 are stored.

  Next, the DS control unit 101 activates the alert determination unit 107, and executes step 806 of determining whether there is a life rhythm abnormality using the life rhythm abnormality determination knowledge table 440.

  For example, when a condition “> 23:00” is stored in the condition field 444 as in the life rhythm abnormality determination knowledge record 440A of the feature quantity ID “SL_S”, the value of the feature quantity ID “SL_S” is When it is larger than the upper limit value (here, 23:00), it is determined as abnormal.

  Further, for example, when the condition field 444 stores the condition “<1” as in the life rhythm abnormality determination knowledge record 440E of the characteristic amount ID “AC_L”, the value of the characteristic amount ID “AC_L” is When the value is smaller than the lower limit (here, 1), it is determined as abnormal.

  Further, for example, when the condition “<4 or> 10” is stored in the condition field 444 as in the life rhythm abnormality determination knowledge record 440C of the feature amount ID “SL_L”, the feature amount ID “SL_L” is stored. When the value of is smaller than the lower limit value (here 4) or larger than the upper limit value (here 10), it is determined as abnormal.

  Further, for example, in the case of the feature amount ID “SL_A60 (post-wake body motion amount)”, the post-wake body motion amount is compared with a reference body motion amount according to a difference between the reference wake-up time and the wake-up time to determine whether there is an abnormality. .

  FIG. 16 is a flowchart showing an operation when determining whether there is an abnormality in the amount of body movement after getting up.

  First, the alert determination unit 107 executes Step 1601 for setting the sleep end time SL_E to the wake-up time (WT).

  Next, the alert determination unit 107 searches the knowledge base 400 and searches the field of the feature amount “sleep end time (wake-up time)” (feature amount ID “SL_E”) from the life rhythm abnormality determination knowledge table 446. Step 1602 is executed to extract the average value from 443 and set it as the reference wake-up time BWT.

  Next, the alert determination unit 107 searches the knowledge base 400, calculates a difference ΔT between the reference wake-up time BWT and the wake-up time WT, and extracts a reference body movement amount BAC according to the difference ΔT 1603 Execute.

  Next, the alert determination unit 107 compares the body movement amount SL_A60 after waking up with the reference body movement amount BAC to determine whether there is an abnormality. If the body movement amount SL_A60 after getting up is smaller than the reference body movement amount BAC, it is determined that there is an abnormality (step 1605), and this operation is terminated. If the body motion amount SL_A60 after getting up is equal to or greater than the reference body motion amount BAC, it is determined that there is no abnormality (step 1606), and this operation is terminated.

  For example, when a user who always gets up at 6 o'clock (reference wake-up time BWT is “06:00”) gets up at 6 o'clock as usual (when the wake-up time WT is “06:00”), the reference body movement amount BAC Becomes “60” (T0 = 60), but when waking up at 6:10, since the difference ΔT is 10 minutes, the reference body movement amount BAC is “70” (T10 = 70).

  FIG. 15 shows a conceptual diagram of the abnormality determination of the behavior pattern after getting up using the amount of body movement after getting up.

  User A usually gets up at 6 o'clock and performs daily activities after waking up, such as toilets, washrooms, changing clothes, eating, relaxing, etc. When he gets up earlier than usual, there is little body movement after waking up, such as reading. There is a tendency. In addition, when waking up later than usual, daily activities are performed in a short time, so the amount of body movement after waking up tends to be large. Therefore, the amount of body movement corresponding to the difference between the usual wake-up time (reference wake-up time) and the wake-up time of the day is as shown in the transition graph shown by the solid line in FIG. By making this transition graph the transition of the reference body movement amount, for example, even if it is erroneously determined as abnormal only with the threshold value of the body movement amount, it can be determined as normal (bar graph 15A1 in FIG. 15). Further, for example, even when the body movement amount threshold value alone is erroneously determined as normal, it can be accurately determined as abnormal (bar graph 15A2 in FIG. 15).

  User B usually wakes up at 6 o'clock and performs daily activities after waking up, such as toilets, washrooms, changing clothes, eating, and relaxing. Will not change. However, when getting up later than usual, some movements such as changing clothes are omitted, so the amount of body movement after waking up tends to be small. Therefore, the amount of body movement corresponding to the difference between the usual wake-up time (reference wake-up time) and the wake-up time on the current day is as shown in the transition graph shown by the solid line in FIG. By using this transition graph as the transition of the reference body movement amount, for example, even if it is erroneously determined to be abnormal only with the threshold value of the body movement amount, it can be determined to be normal (bar graph 15B1 in FIG. 15).

  In this way, the amount of body movement after waking up is compared with the amount of body movement according to the difference between the reference wake-up time and the wake-up time, and the presence or absence of abnormality is determined, so that the determination according to the behavior pattern of the user's daily life can be made. Therefore, it is possible to more accurately determine abnormalities in life rhythm.

  In addition, as described above, by using the amount of body movement statistics within a predetermined time (for example, the average amount of body movement within 60 minutes) as the feature amount of the action pattern after getting up, regardless of the order of daily activities, The feature of the behavior pattern can be expressed as a feature amount.

  If it is determined in step 806 that the abnormality is “none”, the DS control unit 101 executes step 801 and repeats the subsequent processing.

  If it is determined in step 806 that the abnormality is “present”, the DS control unit 101 activates the DS communication unit 108 and executes step 807 for outputting an alert to the staff terminal 120. The staff terminal 120 receives the alert output in step 807 and displays it on the screen of the output unit of the staff terminal 120.

  Next, the DS control unit 101 executes step 808 of storing the alert output in step 807 in the execution history database 114.

  FIG. 11 shows a screen example 1100 of the staff terminal 120. The screen example 1100 includes a column 1101 for displaying a user, a column 1102 for displaying a date and time when an alert is output, a column 1103 for displaying an abnormal state in which an alert is output, and an operation record for inputting an execution record. A check box 1104, a selection input box 1105, and an update button 1106 that is selected when the staff updates the knowledge base.

  In the check box 1104, the staff inputs, as an execution record, that the response to the alert has been performed, such as going to check the state of the resident. The selection input box 1105 is used to input whether the resident's abnormality has been confirmed as a result of the staff responding to the alert.

  In the example of the screen example 1100, for the user “M. Ban”, at 6:01 on April 2, the feature quantity identified by the lifestyle rhythm feature quantity ID “SL_A1” “body movements when waking up” It can be seen that an alert example 1110A indicating that is in an abnormal state is output. Thereby, the staff knows that the user “M. Ban” had an abnormality in body movement after the end of sleep (after waking up) as a sign of an abnormal physical condition such as falling from the bed.

  Further, in the example of the screen example 1100, for the user “M. Ban”, the feature quantity “post-wake body movement amount” identified by the life rhythm feature quantity ID “SL_A60” at 7:01 on April 3rd. It can be seen that the alert example 1110B indicating that “is an abnormal state is output. As a result, the staff knows that an alert indicating abnormal body movement after waking up is output to the user “M. Ban” as a sign of poor physical condition such as a cold or malaise, or abnormal physical condition such as depression.

  In the example of the screen example 1100, the feature amount “sleeping time” identified by the life rhythm feature amount ID “SL_E” is given to the user “F. Kuri” at 7:01 on April 3rd. It can be seen that the alert example 1110C indicating that the state is abnormal is output. As a result, the staff knows that the alert “sleep time abnormality” is output to the user “F. kuri” as a sign of abnormal physical condition.

  Further, in the example of the screen example 1100, the feature amount “activity” identified by the life rhythm feature amount ID “ACR7” is given to the user “H. Shin” at 7:01 on April 9th. It can be seen that an alert example 1110D indicating an abnormal state is output. As a result, the staff knows that the alert “activity abnormality” is output to the user “F. kuri” as a sign of abnormal physical condition such as dementia or disuse syndrome.

  Further, in the example of the screen example 1100, for the user “H. Shin”, the feature quantity “pulse fluctuation rate” identified by the life rhythm feature quantity ID “PsR7” at 7:01 on April 16th. It can be seen that the alert example 1110E that is in an abnormal state is output. As a result, the staff knows that the alert “pulse fluctuation rate abnormality” is output to the user “H. shin” as a sign of abnormal physical condition such as an autonomic nerve abnormality.

  As described above, the staff can immediately recognize that the user has an abnormal physical condition by confirming the alert, so that it is possible to quickly respond such as going to confirm the state.

  Here, the staff operates the staff terminal 120 and inputs as an execution record whether he / she went to check the state and whether there was an abnormality as a result of checking the state. For example, if there is an abnormality as a result of checking the state, as shown in the alert example 1110A, “completed” indicating that the check has been performed is input, and there is an abnormality “present” indicating that there was an abnormality. ". If there is no abnormality, enter “None”.

  In this way, since the response status of alerts is displayed in a list with alerts, it can be easily understood whether other staff members have already responded to alerts as well as the corresponding staff members. Can be prevented.

  When a staff inputs an implementation record using the staff terminal 120, the DS control unit 101 updates the implementation record database 114 based on the input result. For example, in the case of alert example 1110C, record 510C is stored.

In the screen example 1100, when the staff selects the update button 1106,
The data server 100 starts an operation when updating the knowledge base.

  In the present embodiment, an example will be described in which the update button of the alert example 1110C is selected, that is, the knowledge base of the user “H.Kuri” with the user ID “user0001” is updated.

  FIG. 12 is a flowchart showing the operation of the data server 100 when updating the knowledge base. First, when the data server 100 starts an operation for updating the knowledge base, the DS control unit 101 reads out user data of the user ID “user0001” from the data storage database 112 and stores it in the DS memory 102 1202. Execute.

  Next, the DS control unit 101 reads out the execution record history of the user ID “user0001” from the execution record history table 510 and executes step 1203 for storing it in the DS memory 102.

  Next, the DS control unit 101 executes the step 1204 of extracting update user data based on the user data stored in the DS memory 102 and the execution record history and storing it in the DS memory 102. To do.

  For example, in the case of alert example 1110C, the date of the abnormality flag “0” (no abnormality) in the abnormality flag field 515 of the implementation record history table 510 among the user data from April 3 (for example, four weeks ago) in the past. Extract user data with the same date as.

  Next, the DS control unit 101 activates the knowledge base generation unit 109 and executes Step 1205 of calculating life rhythm abnormality determination knowledge from the updating user data stored in the DS memory 102.

  For example, in the case of the feature quantity “wake-up time” (feature quantity ID “SL_E”), the average value m and the standard deviation σ are calculated and stored in the fields 443 to 444. For example, when the average value m is “07:00” and the standard deviation σ is “60 minutes”, “07: 0” is stored in the field 443, and “<08:00” is stored in the field 444 as the upper limit value m + σ. .

  Next, the DS controller 101 stores the knowledge calculated in step 1205 in the knowledge base and executes step 1206 for updating the knowledge base.

  FIG. 13 shows an example knowledge base 1300 updated in step 1106. As shown in the record example 1340B, it is understood that the parameter of the life rhythm feature quantity ID “SL_E” has been updated.

  As a result, the alert is not output from the next time even in the case of the same living state, for example, a living state such as the record 310C.

  In this way, by updating the daily rhythm features using the records entered by the staff, the accuracy of alert determination can be improved and unnecessary alerts can be reduced, improving work efficiency. Is possible.

  With the life monitoring system described above, the staff can immediately grasp the abnormal physical condition of the resident and can respond promptly.

  In addition, by updating the daily rhythm features using the implementation records entered by staff, the accuracy of alert determination can be improved and unnecessary alerts can be reduced, thus improving work efficiency. It becomes.

  Furthermore, since the user can get a sense of being watched by the staff, it is possible to live with peace of mind in the facility.

The block diagram of the life state notification system which is an Example of this invention. An example of a user information database. An example of a data storage database. Knowledge base example. An example of an alert history database. The flowchart showing operation | movement of a wearable device. The flowchart showing operation | movement of a radio repeater. The flowchart showing operation | movement of the data server at the time of data reception. The flowchart showing operation | movement of the data server at the time of life rhythm determination. An example of DS memory. Example of staff terminal screen. The flowchart showing operation | movement of the data server at the time of knowledge base update. An example of an updated knowledge base. The block diagram when a wearable device is built in the wristband. The conceptual diagram of abnormality determination of the action pattern after getting up using the amount of body movement after getting up. The flowchart showing operation | movement of the data server at the time of abnormality determination of the body movement amount after waking up.

Explanation of symbols

100 data server,
101 DS control unit, 102 DS memory, 105 living state determination unit, 106 life rhythm determination unit, 107 alert determination unit, 108 DS communication unit, 109 knowledge base generation unit, 111 user information database, 112 data storage database, 113 knowledge base , 114 Implementation record history database,
120 staff terminal,
140 network,
150 wearable device,
151 WD control unit, 152 WD memory, 153 WD_ID storage unit,
154 WD calculation unit, 155 WD wireless communication unit, 156 pulse wave sensor,
158 temperature sensor, 159 body motion sensor,
170 wireless repeater,
200 Example of user information database 111,
210 User personal information table,
300 Example of data storage database 112,
310 user data history table,
400 Example of knowledge base 113,
420 Living state determination knowledge table,
440 life rhythm abnormality determination knowledge table,
500 Example of implementation record history database 114,
510 implementation record history table,
601-605 each step of the flowchart representing the operation of the wearable device 150;
701 to 703 Steps of the flowchart representing the operation of the wireless repeater 170;
801 to 808 Each step of the flowchart showing the operation of the data server 100 when receiving data;
901-919 Each step of the flowchart showing the operation | movement when determining the life rhythm of the data server 100,
1601 to 1606 Steps of a flowchart representing an operation when determining an abnormality in the amount of body movement after getting up in the data server 100;
Example of 1000 DS memory 102,
1100 Example of staff terminal screen,
1201-1205 Each step of the flowchart showing the operation at the time of updating the knowledge base of the data server 100,
1300 An example of an updated knowledge base.

Claims (2)

A wearable device that includes a body motion sensor that calculates a body movement amount and is worn on a living body; a life state determination unit that determines a life state of a specific date and time including a sleep state based on a detection signal of the wearable device; A life rhythm determination unit that calculates a life rhythm feature quantity from a life state time series and a detection signal, a knowledge database that stores knowledge data that determines whether there is an abnormality in the feature quantity, the knowledge database, and the feature quantity A life monitoring system comprising an alert determination unit that determines an abnormality in a life rhythm and an output unit that outputs the determination result of the alert determination unit as alert information,
The knowledge database stores, as knowledge data, a reference value of a feature value of an action pattern according to a difference between the reference wake-up time and the reference wake-up time, and the life rhythm determination unit changes from a sleep state to another life state. The wake-up time, which is the transition time, and the feature quantity of the action pattern after the wake-up time are calculated, and the alert determination unit changes the reference value according to the difference between the reference wake-up time and the wake-up time, and Life monitoring system characterized by judging abnormal behavior patterns. The life monitoring system according to claim 1, wherein an average value of a body movement amount within a predetermined time after the wake-up time is used as a feature amount of the behavior pattern after wake-up.

Images