Information processing device, method, and recording medium storing information processing program

Through biological information acquisition and machine learning models, the problem that the existing technology cannot predict the time of falling asleep is solved, and accurate prediction and notification before falling asleep is achieved, improving the timing and accuracy of care.

CN110772220BActive Publication Date: 2024-11-12PANASONIC INTELLECTUAL PROPERTY CORP OF AMERICA
View PDF 13 Cites 0 Cited by

Patent Information

Application Number
CN201910455015.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2018-07-25
Filing Date
2019-05-29
Publication Date
2024-11-12
Estimated Expiration
2039-05-29

AI Technical Summary

Technical Problem

The prior art cannot predict the sleep time before the subject falls asleep, making it difficult for caregivers to take care at the right time, which may awaken the occupants.

Method used

An information processing device is designed to obtain the biological information of the object (body movement data, breathing data, heart rate data) and use machine learning to construct a sleep prediction model. The object predicts that the object falls asleep after a specified time and predicts the value of the prediction. Notify the caregiver or subject when it is greater than the threshold.

Benefits of technology

It accurately predicts the sleep time of the subject before falling asleep, helps caregivers arrange appropriate care, and reduces the possibility of awakening the subject.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN110772220B_ABST
    Figure CN110772220B_ABST
Patent Text Reader

Abstract

The present invention provides an information processing device, an information processing method and a recording medium storing an information processing program that can predict when a subject will fall asleep before falling asleep. The sleep prediction device includes: a biological data acquisition unit that acquires biological information of the subject; a sleep prediction unit that uses the biological information to predict that the subject will fall asleep after a specified time; and a notification processing unit that notifies the subject of the fact that the subject falls asleep after the specified time.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to an information processing device, an information processing method, and a computer-readable recording medium storing an information processing program for predicting the onset of sleep of a subject before the subject falls asleep using biological information of the subject. Background Art

[0002] Conventionally, a method of estimating the sleep stage of a subject based on heart rate data is known.

[0003] For example, the sleep stage estimation device disclosed in International Application Publication No. 2016 / 148111 includes an acquisition device for acquiring biological data representing a wave of a heart rate, a parameter estimation device for estimating parameters using the biological data for a formula that defines an intermediate frequency component of a wave of a heart rate using parameters, and a sleep stage estimation device for estimating a sleep stage from the value of the intermediate frequency component at any time based on the formula to which the estimated parameters are applied. Furthermore, the parameter estimation device estimates the intermediate frequency component of the wave of a heart rate based on biological data acquired from an unconstrained sensor disposed between a bed and a mattress on which a subject sleeps. Furthermore, the sleep stage estimation device estimates the sleep stage based on the value of the intermediate frequency component.

[0004] In addition, International Application Publication No. 2016 / 148111 discloses monitoring the sleep of nursing facility residents using the estimated sleep stages. For example, by changing diapers when the resident enters deep sleep, the diapers can be changed without the resident noticing.

[0005] However, in the above-mentioned conventional technology, although the sleep stage of the subject after falling asleep can be estimated, it is not possible to predict when the subject will fall asleep before falling asleep, and further improvement is needed. Summary of the invention

[0006] The present invention is made to solve the above-mentioned problems, and an object of the present invention is to provide an information processing device, an information processing method, and a computer-readable recording medium storing an information processing program that can predict when a subject will fall asleep before falling asleep.

[0007] An information processing device according to one embodiment of the present invention includes: an acquisition unit for acquiring biological information of a subject; a prediction unit for predicting that the subject will fall asleep after a specified time using the biological information; and a notification unit for notifying that the subject has fallen asleep after the specified time.

[0008] According to the present invention, it is possible to predict when a subject will fall asleep before falling asleep. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1It is a schematic diagram showing the configuration of a sleep onset prediction system according to an embodiment of the present invention.

[0010] Figure 2 This is a first flowchart for explaining the overall operation of the sleep onset prediction device according to the present embodiment.

[0011] Figure 3 This is a second flowchart for explaining the overall operation of the sleep onset prediction device according to the present embodiment.

[0012] Figure 4 : is a schematic diagram showing an example of body motion data, respiratory data, and heart rate data in this embodiment.

[0013] Figure 5 1 is a schematic diagram showing an example of waveforms of body motion data, heart rate data, and respiratory data stored in the storage unit in the present embodiment.

[0014] Figure 6 This is a schematic diagram showing an example of changes in the prediction value output by the sleep onset prediction model for biological data after 18:00 according to the present embodiment.

[0015] Figure 7 It is used to illustrate Figure 2 Flowchart of the threshold value adjustment process performed in step S7.

[0016] Figure 8 Schematic diagram showing an example of the sleep onset time tendency distribution used in the present embodiment.

[0017] Fig. 9 It is used to illustrate Figure 3 Flowchart of the notification process performed in step S11. DETAILED DESCRIPTION

[0018] (Basic knowledge of the present invention)

[0019] The sleep stage estimation device disclosed in International Application Publication No. 2016 / 148111 is based on the premise of changing the diaper while the resident is sleeping. However, as other care for the resident, there is care that is not performed during sleep but just before falling asleep, such as brushing teeth. In the present invention, care that is performed just before falling asleep, such as brushing teeth, changing clothes, and taking medicine, is referred to as sleep care. In fact, there are cases where the caregiver does not know the timing of the resident's sleep and wakes up the resident because of care after the resident falls asleep. In International Application Publication No. 2016 / 148111, because the sleep stage is estimated after the subject falls asleep, it is not possible to predict when the subject will fall asleep before falling asleep.

[0020] In order to solve the above-mentioned problem, an information processing device involved in one embodiment of the present invention includes: an acquisition unit for acquiring biological information of a subject; a prediction unit for predicting that the subject will fall asleep after a specified time using the biological information; and a notification unit for notifying the subject of the fact that the subject has fallen asleep after the specified time.

[0021] According to this configuration, since the subject's biological information is used to predict that the subject will fall asleep after a specified time, and the subject is notified of the fact that the subject will fall asleep after the specified time, it is possible to predict when the subject will fall asleep before falling asleep. For example, when the subject is a person being cared for, a caregiver who is notified that the subject will fall asleep after a specified time can provide the subject with prescribed care before the subject falls asleep.

[0022] Furthermore, in the information processing device, the acquisition unit may acquire the biological information from a sensor that measures the biological information from the subject person in a non-contact manner.

[0023] According to this configuration, since the biological information is acquired from the sensor that measures the biological information from the subject in a non-contact manner, the subject does not need to wear the sensor on his body, and the biological information can be acquired without hindering the subject's behavior.

[0024] Furthermore, in the information processing device, the sensor may be a radio wave Doppler sensor. According to this configuration, biological information can be measured in a non-contact manner using the radio wave Doppler sensor.

[0025] Furthermore, in the information processing device, the biological information may include body movement data indicating body movement of the subject, breathing data indicating breathing of the subject, and heart rate data indicating a heart rate of the subject.

[0026] According to this configuration, it is possible to more accurately predict whether the subject will fall asleep after a predetermined time using body movement data indicating the subject's body movement, breathing data indicating the subject's breathing, and heart rate data indicating the subject's heart rate.

[0027] Moreover, the information processing device may further include a storage unit for storing a sleep prediction model constructed through machine learning, wherein the sleep prediction model takes the biological information as an input value and takes a predicted value representing the probability of the subject falling asleep after a specified time as an output value, wherein the prediction unit determines whether the predicted value obtained by inputting the biological information into the sleep prediction model is greater than a threshold value, and when it is determined that the predicted value is greater than the threshold value, it is determined that the subject falls asleep after the specified time.

[0028] According to this configuration, by inputting biological information into the sleep prediction model constructed by machine learning, a prediction value representing the probability of the subject falling asleep after a specified time can be easily obtained, and by comparing the prediction value and the threshold, it is possible to easily predict whether the subject will fall asleep after the specified time.

[0029] Furthermore, the information processing device may further include a threshold value adjustment unit configured to adjust the threshold value based on a tendency distribution of the subject's past sleeping times.

[0030] According to this configuration, since the threshold is adjusted based on the tendency distribution of the subject's past sleeping times, by adjusting the threshold to a lower value when the time after the specified time is a time when the probability of the subject falling asleep in the past is higher, and by adjusting the threshold to a higher value when the time after the specified time is a time when the probability of the subject falling asleep in the past is lower, it is possible to predict with higher accuracy whether the subject will fall asleep after the specified time.

[0031] Moreover, in the information processing device, the notification unit determines whether the deviation of the subject's past sleeping time is below a specified value. If it is determined that the deviation is below the specified value, the subject can also be notified that he or she has fallen asleep after the specified time.

[0032] According to this configuration, when it is determined that the magnitude of the deviation in the subject's past sleeping time is less than a predetermined value, the subject can be notified that he or she has fallen asleep after a predetermined time. Therefore, the subject can be notified that he or she has fallen asleep after a predetermined time with higher accuracy.

[0033] Moreover, in the information processing device, the subject person is a care recipient who receives prescribed care before falling asleep; and the notification unit, when determining that the size of the deviation is greater than the prescribed value, sends confirmation information to the subject person to confirm whether the prescribed care is received.

[0034] According to this configuration, the subject is a care receiver who receives prescribed care before falling asleep. When the magnitude of the deviation is determined to be greater than a prescribed value, confirmation information for confirming whether the subject receives the prescribed care is sent to the subject.

[0035] When the time of falling asleep in the past is not constant, the possibility of the prediction result that the subject falls asleep after the prescribed time being wrong increases. However, when the time of falling asleep in the past is not constant, by confirming whether the subject receives the prescribed care, the subject can be accurately notified of the situation of falling asleep after the prescribed time.

[0036] Another embodiment of the present invention relates to an information processing method, which allows a computer to execute the following steps: obtaining biological information of a subject; using the biological information, predicting that the subject will fall asleep after a specified time; and notifying the subject of the fact that the subject has fallen asleep after the specified time.

[0037] According to this configuration, since the subject's biological information is used to predict that the subject will fall asleep after a specified time, and the subject is notified of the fact that the subject will fall asleep after the specified time, it is possible to predict when the subject will fall asleep before falling asleep. For example, when the subject is a person being cared for, a caregiver who is notified that the subject will fall asleep after a specified time can provide the subject with prescribed care before the subject falls asleep.

[0038] Another embodiment of the present invention relates to a computer-readable recording medium storing an information processing program that enables a computer to have the following functions: an acquisition unit for acquiring biological information of a subject; a prediction unit for predicting whether the subject will fall asleep after a specified time using the biological information; and a notification unit for notifying the subject of the fact that the subject has fallen asleep after the specified time.

[0039] According to this configuration, since the subject's biological information is used to predict that the subject will fall asleep after a specified time, and the subject is notified of the fact that the subject will fall asleep after the specified time, it is possible to predict when the subject will fall asleep before falling asleep. For example, when the subject is a person being cared for, a caregiver who is notified that the subject will fall asleep after a specified time can provide the subject with prescribed care before the subject falls asleep.

[0040] Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings. In addition, the following embodiment is only an example of the embodiment of the present invention and is not intended to limit the technical protection scope of the present invention.

[0041] (Implementation Method)

[0042] Figure 1 It is a schematic diagram showing the configuration of a sleep onset prediction system according to an embodiment of the present invention.

[0043] Figure 1 The sleep onset prediction system shown includes a sleep onset prediction device 1 , a measurement device 2 , and a care notification device 3 .

[0044] The sleep prediction device 1 predicts the falling asleep of a person existing in a target area. In the present embodiment, the target area is, for example, an area in a single room used by a predicted target person in a facility such as a nursing facility or a residence with ancillary services for the elderly. If the target area is an area in a single room of a nursing facility or a residence with ancillary services for the elderly, the predicted target person is the person being cared for living in the single room. Furthermore, if the target area is an area in a hospital ward, the predicted target person is a patient who is admitted to the ward for treatment, etc. For example, in a nursing facility or a residence with ancillary services for the elderly, there are cared for persons who need care such as brushing teeth, changing clothes, and taking medicine before falling asleep. In order to smoothly perform sleep care without waking up the cared for who has already fallen asleep, there is a demand from the caregiver to know the sleeping time of the cared for in advance. The sleep prediction device 1 of the present embodiment predicts the falling asleep of the target person and notifies the caregiver in advance.

[0045] The sleep onset prediction device 1 is, for example, a personal computer or a tablet computer, and is used by a caregiver of a care recipient.

[0046] The care notification device 3 is, for example, a personal computer, a smartphone, or a tablet computer, and is used by the subject.

[0047] The sleep onset prediction device 1 is connected to the care notification device 3 via a network so as to be communicable with each other. The network is, for example, the Internet.

[0048] The measuring device 2 measures the physical quantity of the change in value corresponding to the presence or absence of a person in the target area and the action. In this embodiment, the measuring device 2 measures the physical quantity of the change in value corresponding to the presence or absence of a person in the target area and the action without contact. The measuring device 2 includes a sensor for measuring biological data (biological information) from the subject without contact. The measuring device 2 transmits the measured biological data to the sleep prediction device 1.

[0049] The sleep onset prediction device 1 includes a biological data acquisition unit 11 , a storage unit 12 , a processor 13 , a display unit 14 , and a communication unit 15 .

[0050] The processor 13 includes a sleep onset prediction unit 101 , a threshold value adjustment unit 102 , and a notification processing unit 103 .

[0051] The biological data acquisition unit 11 acquires biological data (biological information) of the subject. The biological data acquisition unit 11 acquires biological data related to the subject's body movement, heart rate, and respiration from the measurement device 2. The biological data includes body movement data representing the subject's body movement, respiration data representing the subject's respiration, and heart rate data representing the subject's heart rate.

[0052] The storage unit 12 stores the biological data acquired by the biological data acquisition unit 11. Furthermore, the storage unit 12 also stores data such as programs and parameters as necessary.

[0053] The sleep prediction unit 101 predicts the subject's sleep using the biometric data stored in the storage unit 12. The sleep prediction unit 101 predicts whether the subject will fall asleep after a predetermined time using the biometric data acquired by the biometric data acquisition unit 11. The predetermined time is, for example, 30 minutes. That is, the sleep prediction unit 101 predicts whether the subject will fall asleep after 30 minutes using the biometric data acquired by the biometric data acquisition unit 11. The predetermined time is not limited to 30 minutes.

[0054] Furthermore, the storage unit 12 stores the sleep prediction model constructed by machine learning, using the biological data as an input value and the predicted value indicating the probability of the subject falling asleep after a specified time as an output value. The sleep prediction unit 101 determines whether the predicted value obtained by inputting the biological data into the sleep prediction model is greater than a threshold value, and if it is determined that the predicted value is greater than the threshold value, it is determined that the subject falls asleep after a specified time. In addition, in the present embodiment, the subject is a care recipient who receives prescribed care just before falling asleep.

[0055] The threshold value adjustment unit 102 adjusts the threshold value used by the sleep onset prediction unit 101 based on the tendency distribution of the subject's past sleep onset time.

[0056] The notification processing unit 103 performs notification processing to the subject or the caregiver when the subject falls asleep after a predetermined time is detected by the sleep prediction unit 101. The notification processing unit 103 notifies the subject that the subject falls asleep after a predetermined time.

[0057] Specifically, the notification processing unit 103 determines whether the deviation of the subject's past sleeping time is less than a specified value. When the notification processing unit 103 determines that the deviation is less than the specified value, it generates notification information for notifying that the subject fell asleep after the specified time, and outputs the generated notification information to the display unit 14. Moreover, when the notification processing unit 103 determines that the deviation is greater than the specified value, it generates confirmation information for confirming whether the subject accepts the specified care, and sends the generated confirmation information to the care notification device 3.

[0058] The display unit 14 is, for example, a liquid crystal display, and displays the notification information generated by the notification processing unit 103. The caregiver recognizes that the subject falls asleep after a predetermined time by observing the notification information displayed on the display unit 14. Then, the caregiver performs predetermined care before the subject falls asleep.

[0059] The communication unit 15 transmits confirmation information for confirming whether the subject has received the prescribed care to the care notification device 3. Furthermore, the communication unit 15 receives response information from the care notification device 3 indicating whether the subject has received the prescribed care.

[0060] That is, the sleep onset prediction device 1 plays a role of assisting a caregiver in performing sleep onset care by detecting and notifying a subject of a sign of falling asleep based on the biological data acquired from the measurement device 2 .

[0061] In the present embodiment, the display unit 14 displays that the subject has fallen asleep after the predetermined time. However, the present invention is not particularly limited to this, and the sleep prediction device 1 may include a speaker for notifying the subject that the subject has fallen asleep after the predetermined time by sound.

[0062] The care notification device 3 includes a communication unit 31 , a processor 32 , a display unit 33 , and an input unit 34 .

[0063] The communication unit 31 receives confirmation information for confirming whether the subject has received predetermined care from the sleep prediction device 1 .

[0064] The processor 32 includes a care confirmation processing unit 301. The care confirmation processing unit 301 generates a confirmation screen for confirming whether the subject has received a predetermined care, and causes the display unit 33 to display the generated confirmation screen.

[0065] The display unit 33 is, for example, a liquid crystal display, and displays a confirmation screen for confirming whether the subject has received predetermined care.

[0066] The input unit 34 is, for example, a touch panel, and receives an input from the subject whether to accept the prescribed care. For example, on the confirmation screen, a text is displayed to confirm whether the subject accepts the prescribed care, and a first button to be pressed when accepting the prescribed care and a second button to be pressed when not accepting the prescribed care are displayed. The subject presses the first button when accepting the prescribed care, and presses the second button when not accepting the prescribed care.

[0067] The care confirmation processing unit 301 generates response information indicating whether the subject has accepted the prescribed care based on the information indicating whether the prescribed care has been accepted input through the input unit 34, and transmits the generated response information to the sleep prediction device 1. The communication unit 31 transmits the response information indicating whether the subject has accepted the prescribed care to the sleep prediction device 1.

[0068] Hereinafter, the sleep onset prediction system according to the present embodiment will be described in detail.

[0069] The following description will take as an example a case where the target space is set in a single room of a senior citizen housing providing services. The target person in this case is a resident of the single room.

[0070] The sleep onset prediction device 1 predicts the onset of sleep of a subject and notifies a caregiver or the like.

[0071] The measuring device 2, for example, includes a radio wave Doppler sensor and a signal processing unit. The Doppler sensor, for example, transmits radio waves in the microwave band to the target area at a predetermined time interval (for example, 1 second interval). The Doppler sensor receives reflected waves reflected by people and the like existing in the target area. The signal processing unit of the measuring device 2 performs signal processing on the reflected waves received by the Doppler sensor to generate body motion data representing body motion of the people existing in the target area. Furthermore, the signal processing unit of the measuring device 2 filters the body motion data representing the body motion of the people, and generates heart rate data representing the body motion caused by the heart rate by extracting the frequency components of the body motion caused by the heart rate. Furthermore, the signal processing unit of the measuring device 2 filters the body motion data representing the body motion of the people, and generates breathing data representing the body motion caused by breathing by extracting the frequency components of the body motion caused by breathing.

[0072] The measuring device 2 outputs body motion data indicating body motion of a person existing in the target area, heart rate data indicating body motion due to heart rate, and respiration data indicating body motion due to respiration to the sleep onset prediction device 1. In addition, the measuring device 2 includes a wireless communication unit compliant with Bluetooth (registered trademark) or the like, for example, and wirelessly transmits the measured biological data to the sleep onset prediction device 1.

[0073] Furthermore, the sensor provided in the measuring device 2 is not limited to the radio wave Doppler sensor, but may be an ultrasonic Doppler sensor that transmits ultrasonic waves. Furthermore, the communication between the measuring device 2 and the sleep prediction device 1 is not limited to wireless communication, but may be wired communication. Furthermore, since the measuring device 2 measures the biological data such as body movement data in a non-contact manner, it does not interfere with the movement of a person.

[0074] In addition, the measurement device 2 may include, for example, a wearable sensor that can be worn on a wrist, and measure biological data in a state of contact with the subject.

[0075] In the present embodiment, the sleep onset prediction device 1 is realized by, for example, a personal computer installed in an office where caregivers and the like gather in facilities such as a nursing facility or a home for the elderly that provides services.

[0076] The biological data acquisition unit 11 includes, for example, a wireless communication unit conforming to Bluetooth (registered trademark) or the like. The biological data acquisition unit 11 acquires body motion data, heart rate data, and respiratory data from the measurement device 2 by performing wireless communication with the measurement device 2 periodically or irregularly. When the biological data acquisition unit 11 acquires body motion data, heart rate data, and respiratory data from the measurement device 2, it outputs the acquired body motion data, heart rate data, and respiratory data to the storage unit 12.

[0077] The storage unit 12 includes, for example, an electrically rewritable nonvolatile memory such as EEPROM (Electrically Erasable Programmable Read Only Memory) and a volatile memory such as RAM (Random Access Memory). The storage unit 12 stores the body movement data, heart rate data, and breathing data acquired by the biological data acquisition unit 11.

[0078] The functions of the sleep prediction unit 101, the threshold adjustment unit 102, and the notification processing unit 103 are realized by causing the processor 13 to execute a predetermined program. The program may be stored in advance in the memory or the storage unit 12, or may be stored and provided in a non-temporary recording medium such as a memory card via a telecommunication line such as the Internet.

[0079] The storage unit 12 stores a sleep prediction model. The sleep prediction model outputs a prediction value using the biological data obtained before a predetermined time. In the present embodiment, as an example, the prediction value is represented by a real number greater than 0 and less than 1, and is an indicator representing the probability that the subject will fall asleep within 30 minutes from the present time. Moreover, in the present embodiment, the sleep prediction model is pre-constructed by machine learning using biological data from 30 minutes before the sleep time recorded in the past to the time of falling asleep and biological data from 18:00 in the afternoon to 30 minutes before the sleep time as input data. In the present embodiment, input data at a time within 30 minutes of the subject falling asleep is referred to as positive data with a sign of falling asleep, and input data at a time before 30 minutes before the subject falling asleep is referred to as negative data without a sign of falling asleep.

[0080] The sleep onset prediction unit 101 inputs the biological data acquired by the biological data acquisition unit 11 into the sleep onset prediction model, and obtains a prediction value output from the sleep onset prediction model.

[0081] The threshold adjustment unit 102 calculates a threshold for comparison with the predicted value output by the sleep onset prediction model. The threshold adjustment unit 102 calculates a threshold used in a sleep onset prediction process described later using the first threshold TH1, the second threshold TH2, and the sleep onset time tendency distribution.

[0082] In this embodiment, as the judgment result of the sleep prediction unit 101, the case where the prediction value exceeds the threshold is called positive, and the case where the prediction value is lower than the threshold is called negative. In addition, the case where the positive data is input and the case is misjudged as negative is called false negative, and the case where the negative data is input and the case is misjudged as positive is called false positive. The false negative rate is the proportion of data judged as negative in all positive data, and the false positive rate is the proportion of data judged as positive in all negative data.

[0083] The first threshold TH1 is a threshold for suppressing the false negative rate of the sleep prediction model below a certain level. The second threshold TH2 is a threshold for suppressing the false positive rate of the sleep prediction model below a certain level, and is a value greater than the first threshold TH1. The first threshold TH1 and the second threshold TH2 are numerical values ​​that can be additionally obtained by performing cross-validation using various thresholds when constructing a sleep prediction model using machine learning. In this embodiment, the storage unit 12 stores the first threshold TH1 and the second threshold TH2 in advance.

[0084] The sleep time tendency distribution is a sleep time tendency distribution that indicates the probability of falling asleep in which time period and with what degree, and is produced based on the subject's past activity / sleep information or the personal opinion of the caregiver of the subject. The sleep time tendency distribution, for example, may be a frequency distribution or a probability distribution simply based on the distribution of sleep times during a specified period in the past, or may be a probability distribution obtained by inputting daytime body movement data, heart rate data, and breathing data into a sleep time tendency distribution prediction model that is different from the sleep prediction model. The sleep time tendency distribution prediction model is generated by machine learning and is pre-stored in the storage unit 12. In addition, in the present embodiment, the storage unit 12 may also pre-store the sleep time tendency distribution.

[0085] When the prediction value output by the sleep prediction model is greater than the threshold value calculated by the threshold value adjustment unit 102 , the sleep prediction unit 101 determines that a sign of the subject falling asleep has been detected, that is, the subject will fall asleep after a predetermined time.

[0086] The notification processing unit 103 performs a notification process to the subject or a caregiver when the sleep prediction unit 101 detects a sign of the subject falling asleep.

[0087] In addition, in the present embodiment, the sleep prediction device 1 may also include the measurement device 2. In this case, the sleep prediction device 1 is arranged in the room where the subject is, and the terminal device arranged in the room where the caregiver is has the function of the display unit 14. In addition, in the present embodiment, the sleep prediction device 1 may also be a server. In this case, the terminal device arranged in the room where the caregiver is has the function of the display unit 14.

[0088] Secondly, refer to Figure 2 The overall operation of the sleep onset prediction device 1 according to the present embodiment will be described using the flowchart.

[0089] Figure 2 is a first flowchart for explaining the overall operation of the sleep onset prediction device according to the present embodiment. Figure 3 This is a second flowchart for explaining the overall operation of the sleep onset prediction device according to the present embodiment.

[0090] First, the biological data acquisition unit 11 acquires biological data from the measurement device 2 regularly or irregularly (step S1). In this embodiment, as an example, the biological data acquisition unit 11 acquires body motion data, heart rate data, and breathing data from the measurement device 2 every second.

[0091] Next, the biological data acquisition unit 11 stores the acquired biological data in the storage unit 12 (step S2). In this embodiment, as an example, the biological data acquisition unit 11 stores the one-minute average values ​​of the body movement data, heart rate data, and respiratory data in the storage unit 12 every one minute.

[0092] The sleep prediction unit 101 determines whether the current time has reached the specified time (step S3). In addition, the specified time is, for example, the end time of dinner set by the facility. The specified time is, for example, 18:00. The above-mentioned specified time is an example and is set to an appropriate time according to the subject. Here, if it is determined that the current time has not reached the specified time (step S3 is "No"), the process returns to step S1.

[0093] On the other hand, when it is determined that the current time has reached the specified time (step S3 is "yes"), the threshold adjustment unit 102 creates a sleep time tendency distribution of the subject (step S4). The threshold adjustment unit 102, for example, inputs the body movement data, heart rate data, and breathing data of a day from 7:00 a.m. to 18:00 p.m. into the sleep time tendency distribution prediction model, and creates the sleep time tendency distribution by obtaining the sleep time tendency distribution output from the sleep time tendency distribution prediction model. The threshold adjustment unit 102 stores the created sleep time tendency distribution in the storage unit 12. In addition, when the storage unit 12 has pre-stored the sleep time tendency distribution, the processing of step S4 can be omitted.

[0094] Next, the biological data acquisition unit 11 acquires biological data from the measurement device 2 (step S1). In the present embodiment, as an example, the biological data acquisition unit 11 acquires body motion data, heart rate data, and breathing data from the measurement device 2 every one second.

[0095] Next, the biological data acquisition unit 11 stores the acquired biological data in the storage unit 12 (step S2). In this embodiment, as an example, the biological data acquisition unit 11 stores the one-minute average values ​​of the body movement data, heart rate data, and respiratory data in the storage unit 12 every one minute.

[0096] Next, the threshold adjustment unit 102 performs a threshold adjustment process for calculating a threshold used for the sleep onset prediction process at a predetermined timing (for example, every one minute) (step S7). The threshold adjustment process will be described in detail later.

[0097] Next, the sleep prediction unit 101 acquires biological data of a predetermined time period from the current time from the storage unit 12 (step S8 ). For example, the sleep prediction unit 101 acquires body movement data, heart rate data, and breathing data from the storage unit 12 for the past hour from the current time.

[0098] Next, the sleep onset prediction unit 101 inputs the acquired biological data into the sleep onset prediction model, obtains the prediction value output from the sleep onset prediction model, and calculates the prediction value (step S9 ).

[0099] In addition, the time interval for performing the threshold value setting process by the threshold value adjustment unit 102 may be the same as the time interval for calculating the prediction value by the sleep onset prediction unit 101, or may be longer than the time interval for calculating the prediction value. That is, for example, the time interval for performing the threshold value setting process may be one minute interval, which is the same as the time interval for calculating the prediction value, or may be 30 minutes interval, which is longer than the time interval for calculating the prediction value.

[0100] Moreover, the threshold adjustment unit 102 can also perform threshold setting processing at a specified time only once a day (for example, 18:00), calculate the threshold for each specified time after 18:00 (for example, every minute), and store the calculated threshold for each specified time in the storage unit 12.

[0101] Here, the process of calculating the prediction value by the sleep onset prediction unit 101 will be further described.

[0102] The sleep onset prediction unit 101 performs processing for determining a sleep onset sign of the subject based on the body movement data, heart rate data, and breathing data stored in the storage unit 12 .

[0103] Figure 4 is a schematic diagram showing an example of body motion data, breathing data, and heart rate data of this embodiment. In this embodiment, the body motion data, heart rate data, and breathing data used for input are as follows: Figure 4 As shown, each measurement data has a two-dimensional value, and six-dimensional biological data is stored every one minute. That is, the body motion signal output from the Doppler sensor of the measurement device 2 is divided into a body motion I signal and a body motion Q signal by I / Q detection as one of the signal processing, and by filtering the body motion I signal and the body motion Q signal, the respiration I signal, the respiration Q signal, the heart rate I signal, and the heart rate Q signal can be measured.

[0104] The sleep onset prediction unit 101 acquires biological data for one hour from the current time from the storage unit 12 at a predetermined timing, and inputs the biological data into the sleep onset prediction model to obtain a prediction value.

[0105] For example, when the current time is 19:00, the sleep prediction unit 101 obtains the biological data from 18:00 to 18:59 from the storage unit 12 and inputs it into the sleep prediction model. The sleep prediction model represents the input with a real value greater than or equal to 0 and less than or equal to 1, and outputs a predicted value representing the probability of falling asleep within 30 minutes from the current time.

[0106] Figure 5 1 is a schematic diagram showing an example of waveforms of body motion data, heart rate data, and respiratory data stored in the storage unit 12 in the present embodiment. Figure 6 FIG. 2 is a schematic diagram showing an example of changes in the predicted value outputted by the sleep prediction model for biological data after 18:00 in the present embodiment. Figure 5 In the figure, the body motion data represents the average value of the body motion I signal and the body motion Q signal, the heart rate data represents the average value of the heart rate I signal and the heart rate Q signal, and the respiration data represents the average value of the respiration I signal and the respiration Q signal.

[0107] according to Figure 5From the waveform shown, it can be generally assumed that the subject entered the room at around 19:00, his activities became stable from around 21:30, and he fell asleep between about 22:30 and 23:00. Figure 6 As shown, the predicted value output by the sleep onset prediction model starts to rise slowly from around 22:00 after the subject's activity becomes stable, and reaches a peak at around 22:30.

[0108] Return to Figure 3 Next, the sleep prediction unit 101 determines whether the calculated prediction value is greater than the threshold value calculated by the threshold value adjustment unit 102 (step S10). Here, if it is determined that the prediction value is less than the threshold value (step S10: No), the process returns to step S5.

[0109] On the other hand, when it is determined that the predicted value is greater than the threshold value (step S10 is "yes"), the notification processing unit 103 performs notification processing (step S11) to notify the caregiver that the subject has fallen asleep after a predetermined time, or to ask the subject to confirm whether to accept sleep care.

[0110] Next, the notification processing unit 103 determines whether the caregiver is notified that the subject has fallen asleep after the prescribed time (step S12). Here, if it is determined that the caregiver is not notified that the subject has fallen asleep after the prescribed time (step S12 is "No"), the process returns to step S5.

[0111] On the other hand, when it is determined that the caregiver has been notified that the subject has fallen asleep after the predetermined time (YES in step S12 ), the process returns to step S1 .

[0112] Secondly, refer to Figure 4 The following describes the threshold value adjustment process performed by the threshold value adjustment unit 102. In the threshold value adjustment process, the threshold value adjustment unit 102 sets a threshold value used in the sleep onset prediction process using the first threshold value TH1, the second threshold value TH2, and the sleep onset time tendency distribution.

[0113] Figure 7 It is used to illustrate Figure 2 This is a flowchart of the threshold adjustment process performed in step S7.

[0114] First, the threshold value adjustment unit 102 acquires the first threshold value TH1 and the second threshold value TH2 from the storage unit 12 (step S21 ).

[0115] Next, the threshold adjustment unit 102 acquires the sleep onset time tendency distribution from the storage unit 12 (step S22 ).

[0116] Figure 8: is a schematic diagram showing an example of the sleep time tendency distribution used in this embodiment. The sleep time tendency distribution is a histogram showing the distribution of the subject's past sleep time, or a probability density function f(t) generated by kernel density estimation based on the subject's past sleep time, and either one of them can be used according to the interval of time when the threshold adjustment process is performed. Figure 8 In the above, the range of the horizontal axis of the histogram is set to 18:00 to 24:00, but this range can also be changed to match the tendency of the subject's falling asleep. Moreover, as long as it is a distribution that can represent the tendency of the falling asleep time, a distribution generated by other methods can also be used.

[0117] Return to Figure 7 Next, the threshold value adjustment unit 102 calculates the threshold value TH corresponding to the current time t t (Step S23) The threshold adjustment unit 102 calculates the threshold TH1 so that the closer the current time t is to the time when there is a tendency to fall asleep, the closer it is to the first threshold TH1, and the closer the current time t is to the time when there is no tendency to fall asleep, the closer it is to the second threshold TH2. t The threshold adjustment unit 102 uses the tendency distribution of the falling asleep time Figure 8 In the case of the probability density function f(t), the threshold TH corresponding to each time t is calculated by the following formula (1): t In addition, in formula (1), max(f) represents the maximum value of function f.

[0118]

[0119] Here, the function is set to f(t+0.5) because the sleep prediction model is designed to output the probability of falling asleep within 30 minutes from the current time, and it can also be changed accordingly according to the design of the sleep prediction model. The denominator max(f) is used only when the function f is the maximum value. t =TH1 is an example and can be changed appropriately. By using the above formula (1) to calculate the threshold TH t If the time after 30 minutes is a time when the probability of the subject falling asleep is high in the past, the threshold TH t In order to obtain a lower value close to the first threshold value TH1, if the time after 30 minutes is a time when the probability of the subject falling asleep in the past is low, the threshold value TH t It is a high value close to the second threshold value TH2.

[0120] Next, the threshold adjustment unit 102 determines whether there is an instruction from the subject that care is not required on the same day in the notification process described later (step S24). Here, if it is determined that there is no instruction that care is not required on the same day (step S24 is "No"), the threshold adjustment process is terminated.

[0121] On the other hand, when it is determined that there is an instruction that no care is required on the same day (step S24 is "yes"), the threshold adjustment unit 102 corrects the threshold TH based on the time from the time when the subject indicated that no care is required to the current time. t (Step S25) The threshold value adjustment unit 102 sets the threshold value TH to a value that is smaller as the time from the time when care is not needed to the current time is shorter. t As a correction method, for example, the threshold value adjustment unit 102 changes the value of the function f to 0 within 30 minutes from the time when the care is not needed, and reduces the threshold value TH by halving the function f from 30 minutes to 60 minutes after the time when the care is not needed. t Furthermore, the threshold value adjustment unit 102 increases the threshold value TH t Additionally, the sleep prediction unit 101 may not perform sleep prediction within a certain period of time (for example, 30 minutes) from the time when care is not required.

[0122] Next, refer to Fig. 9 The flowchart of will explain the notification processing performed by the notification processing unit 103.

[0123] Fig. 9 It is used to illustrate Figure 3 Flow chart of the notification process of step S11.

[0124] First, the notification processing unit 103 acquires the subject's sleeping time tendency distribution from the storage unit 12 (step S41 ).

[0125] Next, the notification processing unit 103 calculates the variance σ of the sleeping time based on the acquired sleeping time tendency distribution. 2 (Step S42).

[0126] Next, the notification processing unit 103 determines the variance σ 2 Is it below a specified value (step S43)? 2 represents the magnitude of the deviation in the time of falling asleep. Therefore, if the variance σ 2 is small, the time of falling asleep is fixed, and people fall asleep at roughly the same time every day. On the other hand, if the variance σ 2The larger the variance, the less likely it is that the time to fall asleep is fixed, and the more likely it is that the time to fall asleep will deviate. 2 If the subject's sleep time is less than the prescribed value (step S43 is "yes"), the notification processing unit 103 causes the display unit 14 to display notification information for notifying that the subject has fallen asleep after the prescribed time (step S44). By displaying the notification information, the caregiver is notified that the subject has fallen asleep after the prescribed time. The caregiver confirms the notification information displayed on the display unit 14 and performs prescribed care before the subject falls asleep.

[0127] On the other hand, when judging the variance σ 2 When it is greater than the prescribed value (step S43 is "No"), the notification processing unit 103 sends confirmation information for confirming whether the subject has accepted the prescribed care to the care notification device 3 (step S45). The communication unit 31 of the care notification device 3 receives the confirmation information. The care confirmation processing unit 301 generates a confirmation screen for confirming whether the subject has accepted the prescribed care, and displays the generated confirmation screen on the display unit 33. In addition, the care confirmation processing unit 301 can also output the confirmation information by sound. Then, the input unit 34 receives the input of whether the subject has accepted the prescribed care. The communication unit 31 sends the response information indicating whether the subject has accepted the prescribed care to the sleep prediction device 1.

[0128] Next, the communication unit 15 receives the response information transmitted from the care notification device 3 (step S46).

[0129] Next, the notification processing unit 103 determines whether the subject has accepted the prescribed care based on the received response information (step S47). If it is determined that the subject has accepted the prescribed care (step S47 is "yes"), the process transfers to step S44, and the notification processing unit 103 causes the display unit 14 to display notification information for notifying that the subject has fallen asleep after the prescribed time.

[0130] On the other hand, when it is determined that the subject does not accept the prescribed care, that is, when it is determined that there is an instruction that care is not required (step S47 is "No"), the notification processing unit 103 stores the current time as a care-not-required time indicating that care is not required in the storage unit 12 (step S48). The care-not-required time stored in the storage unit 12 is used for the correction of the threshold value by the threshold adjustment unit 102 as described above.

[0131] In addition, in step S46, when no response information is received even after the predetermined time has passed, that is, when there is no answer from the subject as to whether to accept the predetermined care, the process of step S48 may be performed.

[0132] In this embodiment, the variance σ of the sleep time is used. 2 The magnitude of the variation in the sleep onset time is represented by , but the present invention is not particularly limited to this, and the magnitude of the variation in the sleep onset time may be represented by the standard deviation of the sleep onset time.

[0133] Furthermore, for a subject who has difficulty or is unable to answer whether or not to accept a prescribed nursing care due to cognitive impairment or the like, confirmation information may not be sent and notification information may be displayed on the display unit 14 unconditionally by setting in advance.

[0134] Furthermore, before the confirmation information is displayed by the care notification device 3, the subject may actively input an operation to request the prescribed care into the input unit 34, the communication unit 31 may send information requesting the prescribed care to the sleep prediction device 1, and the display unit 14 of the sleep prediction device 1 may display the information requesting the prescribed care.

[0135] In the present invention, all or part of a unit, device, component or part, or all or part of a functional block of a block diagram shown in the accompanying drawings can be implemented by one or more electronic circuits including a semiconductor device, a semiconductor integrated circuit (IC) or an LSI (Large Scale Integration). LSI or IC can be integrated into one chip, or a structure combining multiple chips can be adopted. For example, functional modules other than storage elements can also be integrated into one chip. Although it is called LSI or IC here, its name changes according to the degree of integration, and it can also be called system LSI, VLSI (Very Large Scale Integration) or ULSI (Ultra Large Scale Integration). FPGA (Field Programmable Gate Array) that can be programmed after manufacturing LSI or reconfigurable Reconfigurable Logic Device that can reconstruct the connection relationship inside LSI or the circuit area inside LSI can also be used for the same purpose.

[0136] Furthermore, all or part of the functions or operations of a unit, device, component or part may also be implemented by software program processing. In this case, the software program is stored in one or more non-temporary recording media such as ROM, optical disk, hard disk drive, etc. When the software program is executed by a processing device (Processor), the functions determined by the software program are executed by the processing device (Processor) and peripheral devices. The system or device may also have one or more non-temporary recording media storing the software program, a processing device (Processor) and required hardware devices such as interfaces.

[0137] The present invention relates to an information processing device, an information processing method, and a computer-readable recording medium storing an information processing program. Since they can predict when a subject will fall asleep before falling asleep, the information processing device, the information processing method, and the computer-readable recording medium storing an information processing program have practical value.

[0138] This application is based on Japanese Patent Application No. 2018-139182 filed with the Japan Patent Office on July 25, 2018, the contents of which are incorporated herein by reference.

[0139] Although the present invention has been fully described with reference to the accompanying drawings, it should be noted that various modifications and variations are obvious to those skilled in the art. Therefore, unless such modifications and variations depart from the scope defined by the present invention, they are all included in the present invention.

[0140] Description of symbols

[0141] 1 Sleep prediction device

[0142] 2 Measurement device

[0143] 3. Nursing notice device

[0144] 11 Biological Data Acquisition Department

[0145] 12 Storage

[0146] 13 Processor

[0147] 14 Display

[0148] 15 Ministry of Communications

[0149] 31 Ministry of Communications

[0150] 32 processors

[0151] 33 Display

[0152] 34 Input section

[0153] 101 Sleep Prediction Department

[0154] 102 Threshold adjustment unit

[0155] 103 Notification Processing Department

[0156] 301 Nursing Confirmation Processing Department

Claims

1. An information processing device, characterized in that include: an acquisition unit that acquires biological information of a subject; a prediction unit that predicts that the subject will fall asleep after a predetermined time from the present time, using the biological information; a notification unit for notifying the subject that the subject has fallen asleep after a predetermined time from the present time; as well as A storage unit stores a sleep prediction model constructed by machine learning, the sleep prediction model taking the biological information as an input value and taking a predicted value representing the probability of the subject falling asleep after a predetermined time from the present time as an output value, wherein: The prediction unit determines whether the prediction value obtained by inputting the biological information into the sleep prediction model is greater than a threshold value, and determines that the subject will fall asleep after a predetermined time from a current time if the prediction value is determined to be greater than the threshold value.

2. The information processing device according to claim 1, characterized in that: The acquisition unit acquires the biological information from a sensor that measures the biological information from the subject person in a non-contact manner.

3. The information processing device according to claim 2, characterized in that: The sensor is an electric wave Doppler sensor.

4. The information processing device according to claim 1, characterized in that: The biological information includes body movement data indicating body movement of the subject, breathing data indicating breathing of the subject, and heart rate data indicating a heart rate of the subject.

5. The information processing device according to claim 1, characterized in that Also includes: The threshold value adjustment unit adjusts the threshold value based on the tendency distribution of the subject's past sleep onset times.

6. The information processing device according to claim 1, characterized in that: The notification unit determines whether the deviation of the subject's past sleeping times is less than a predetermined value, and notifies the subject that he or she has fallen asleep a predetermined time after a current time if the deviation is determined to be less than the predetermined value.

7. The information processing device according to claim 6, characterized in that: The subject person is a care recipient who receives prescribed care immediately before falling asleep; The notification unit transmits confirmation information for confirming whether or not to accept the prescribed care to the subject, when it is determined that the magnitude of the deviation is larger than the prescribed value.

8. An information processing method, characterized in that: Let the computer perform the following steps: Acquiring biological information of the subject; Using the biological information, predicting that the subject will fall asleep after a predetermined time from the present time; Notifying the subject that he / she will fall asleep after a predetermined time from the present time; A sleep prediction model constructed by machine learning is stored, the sleep prediction model taking the biological information as an input value and taking a predicted value representing the probability of the subject falling asleep after a predetermined time from the present time as an output value, wherein: In the prediction, it is determined whether the predicted value obtained by inputting the biological information into the sleep prediction model is greater than a threshold value, and if it is determined that the predicted value is greater than the threshold value, it is determined that the subject will fall asleep after a predetermined time from the present time.

9. A computer-readable recording medium, characterized in that: The computer stores information processing programs that enable the computer to: an acquisition unit that acquires biological information of a subject; a prediction unit that predicts that the subject will fall asleep after a predetermined time from the present time, using the biological information; a notification unit for notifying the subject that the subject has fallen asleep after a predetermined time from the present time; as well as A storage unit stores a sleep prediction model constructed by machine learning, the sleep prediction model taking the biological information as an input value and taking a predicted value representing the probability of the subject falling asleep after a predetermined time from the present time as an output value, wherein: The prediction unit determines whether the prediction value obtained by inputting the biological information into the sleep prediction model is greater than a threshold value, and determines that the subject will fall asleep after a predetermined time from a current time if the prediction value is determined to be greater than the threshold value.

10. A computer program product comprising a computer program, the program causing a computer to execute the steps of the information processing method according to claim 8.

Citation Information

Patent Citations

  • Solid electrolytic member, solid oxide type fuel cell, water electrolysis device, hydrogen pump and method for manufacturing solid electrolytic member

    JP2018139182A

  • Sleep stage estimation device, biological data estimation device, sleep stage estimation method, biological data estimation method, sleep stage estimation program, and biological data estimation program

    WO2016148111A1

  • Illumination and sleep evaluation system

    CN101227858A

  • Wearable device for sleep assistance

    CN107427665A

  • Sleep State Prediction Device

    CN107440680A