JP2024167937A - Information processing device, information processing program, and machine learning device - Google Patents

Abstract

To provide an information processing device, an information processing program, and a machine learning device capable of reducing a burden of a patient and surrounding people caused by onset of a dementia symptom.SOLUTION: An information processing device includes: an acquisition unit 211 that acquires index information including at least one of a state index related to a life state of a target dementia patient in a first period and an environment index related to a life environment of the target dementia patient in the first period; and a prediction unit 212 that predicts onset information related to onset of a dementia symptom of the target dementia patient in a second period later than the first period on the basis of the acquired index information.SELECTED DRAWING: Figure 10

Description

The present invention relates to an information processing device, an information processing program, and a machine learning device.

In Japan, life expectancy has increased significantly due to improvements in living standards, sanitary conditions, and medical standards that have accompanied the rapid economic growth since the end of the war. This, combined with a declining birth rate, has resulted in an aging society with a high aging rate. Dementia is one of the problems that an aging society faces (for example, Patent Document 1).

JP 2022-139993 A

In care facilities and other facilities where elderly people live, when a dementia patient shows symptoms, it places a heavy burden not only on the patient themselves, but also on those around them, such as their caregivers.

The present invention has been made in consideration of these circumstances. That is, the objective of the present invention is to provide an information processing device, an information processing program, and a machine learning device that can reduce the burden on patients and those around them caused by the onset of dementia symptoms.

The above-mentioned problems of the present invention are solved by the following means.

(1) An information processing device including: an acquisition unit that acquires index information including at least one of a condition index related to the living condition of a target dementia patient in a first period and an environmental index related to the living environment of the target dementia patient in the first period; and a prediction unit that predicts occurrence information related to the occurrence of dementia symptoms of the target dementia patient in a second period after the first period based on the acquired index information.

(2) The information processing device according to (1) above, in which the acquisition unit acquires the index information including the state index and the environmental index.

(3) The information processing device according to (1) above, in which the state index includes at least one of an index representing the movement of the target dementia patient and an index representing the sleep state of the target dementia patient.

(4) The information processing device described in (3) above, in which the motion includes at least one of the following: movement, stopping, repetitive behavior, and unsteadiness of the target dementia patient.

(5) The information processing device according to (3) above, in which the sleep state includes at least one of the sleep time, sleep rate, and sleep stability of the target dementia patient.

(6) The information processing device according to (1) above, wherein the environmental index includes at least one of an index representing the temperature of the living environment, an index representing the air pressure of the living environment, and an index representing the humidity of the living environment.

(7) The information processing device according to (1) above, wherein the occurrence information includes information regarding the occurrence or nonoccurrence of the dementia symptom in the target dementia patient and information regarding the type of the dementia symptom that occurs.

(8) The information processing device according to claim 7, wherein the generation information may include information relating to a plurality of the types.

(9) The information processing device according to (1) above, in which the prediction unit predicts the occurrence information using a machine learning model.

(10) The information processing device described in (9) above, in which the machine learning model is trained using learning index information including at least one of a learning state index relating to the living state of a dementia patient in a first learning period and a learning environment index relating to the living environment of the dementia patient in the first learning period as explanatory variables, and learning occurrence information relating to the occurrence of dementia symptoms of the dementia patient in a second learning period that is later than the first learning period as objective variables.

(11) The information processing device according to (1) above, further comprising an output unit that outputs the predicted occurrence information.

(12) The information processing device according to (1) above, in which the first period is longer than the second period.

(13) The information processing device according to (1) above, wherein the first period is one week or longer.

(14) The information processing device according to (1) above, wherein the index information includes at least one of information regarding a change in the state index during the first period and information regarding a change in the environmental index during the first period.

(15) An information processing program for causing a computer to execute a process including acquiring index information including at least one of a condition index relating to the living condition of a target dementia patient in a first period and an environmental index relating to the living environment of the target dementia patient in the first period, and predicting occurrence information relating to the occurrence of dementia symptoms of the target dementia patient in a second period subsequent to the first period based on the acquired index information.

(16) A machine learning device comprising: a storage unit that stores learning data including, as explanatory variables, learning indicator information including at least one of a learning condition indicator related to the living condition of a dementia patient in a first learning period and a learning environment indicator related to the living environment of the dementia patient in the first learning period, and learning occurrence information related to the occurrence of dementia symptoms of the dementia patient in a second learning period that is later than the first learning period, as a target variable; and a model generation unit that performs machine learning on the relationship between the learning indicator information and the learning occurrence information based on the stored learning data, to generate a machine learning model.

In the information processing device and information processing program of the present invention, occurrence information regarding the onset of dementia symptoms is predicted based on index information of a target dementia patient. This enables caregivers and others to take necessary measures in advance when it is predicted that a dementia symptom will occur in the target dementia patient. This makes it possible to reduce the burden on the patient and those around them caused by the onset of dementia symptoms. In the machine learning device of the present invention, a machine learning model capable of predicting occurrence information is generated based on index information.

FIG. 1 is a diagram showing the overall configuration of a monitoring system. FIG. 13 is a diagram showing an example of a detection unit installed in a subject's room. FIG. 2 is a block diagram showing a hardware configuration of a detection unit. FIG. 2 is a block diagram showing a hardware configuration of a server. FIG. 4 is a diagram illustrating an example of a status indicator stored in a storage unit of a server. FIG. 6 is a diagram showing an example of a status indicator following FIG. 5 . FIG. 4 is a diagram illustrating an example of an environmental index stored in a storage unit of the server. FIG. 2 is a block diagram showing a hardware configuration of an administrator terminal. FIG. 2 is a block diagram showing a hardware configuration of a staff terminal. 2 is a block diagram showing a functional configuration of a control unit of the server. 11A and 11B are diagrams illustrating an example of index information and generation information. 10 is a flowchart illustrating an example of an operation of the server. 1 is a flowchart illustrating an example of a learning process of a machine learning model.

Hereinafter, an embodiment of the present invention will be described with reference to the attached drawings. Note that in the description of the drawings, the same elements are given the same reference numerals, and duplicate descriptions will be omitted. Also, the dimensional ratios in the drawings have been exaggerated for the convenience of explanation, and may differ from the actual ratios.

<Embodiment>
[Configuration of monitoring system 1]
(Overall composition)
FIG. 1 is a diagram showing an overall configuration of a monitoring system 1 according to an embodiment, and FIG. 2 is a diagram showing an example of a detection unit 10 installed in a room of a subject 70. As shown in FIG.

As shown in FIG. 1, the monitoring system 1 has multiple detection units 10, a server 20, an administrator terminal 30, and one or more staff terminals 40. These are connected to each other via a network 50 such as a LAN (Local Area Network), a telephone network, or a data communication network, either wired or wirelessly, so that they can communicate with each other. The network 50 may include relays such as repeaters, bridges, routers, or cross-connects that relay communication signals. In the example shown in FIG. 1, the staff terminal 40 is connected to the detection units 10, the server 20, and the administrator terminal 30 via a network 50 such as a wireless LAN (e.g., a LAN conforming to the IEEE 802.11 standard) that includes an access point 51, so that they can communicate with each other.

The monitoring system 1 is disposed at an appropriate location according to the subject 70. The subject 70 is, for example, a patient who needs nursing due to illness or injury, a person who needs nursing due to a decline in physical ability due to aging, a person living alone, or a patient hospitalized in a hospital facility. In particular, from the viewpoint of enabling early detection and early response, the subject 70 may be a person who needs to be detected when a predetermined inconvenient event such as an abnormal state occurs to the subject. In this embodiment, the subject 70 is a dementia patient. The monitoring system 1 is suitably disposed in buildings such as elderly welfare facilities, hospitals, and residences according to the type of subject 70. In the example shown in FIG. 1, the monitoring system 1 is disposed in a building of a facility that has multiple rooms (living rooms) in which multiple subjects 70 live and multiple rooms including a nurse's station.

The detection unit 10 is arranged, for example, in each room, which is the observation area of the subject 70. In the example shown in FIG. 1, four detection units 10 are arranged in each room of the subjects 70, Mr. A, Mr. B, Mr. C, and Mr. D. The observation area of the detection unit 10 includes a bed 60. Staff members 80 who provide nursing care or other such services (e.g., care) to the subject 70 each carry a staff terminal 40, which is a mobile terminal. However, the location and number of each component of the monitoring system 1 are not limited to the example shown in FIG. 1. For example, the server 20 does not have to be arranged in a nurse's station, and may be an external server unit connected to the network 50.

(Detection Unit 10)
Fig. 3 is a block diagram showing a hardware configuration of the detection unit 10. As shown in Fig. 3, the detection unit 10 has a control unit 11, a communication unit 12, a camera 13, a care call unit 14, and a voice input/output unit 15, which are connected to each other by a bus.

The control unit 11 is composed of a CPU (Central Processing Unit) and memories such as a RAM (Random Access Memory) and a ROM (Read Only Memory), and controls each part of the detection unit 10 and performs calculation processing according to a program. The control unit 11 may further include a HDD (Hard Disk Drive) as a memory.

The communication unit 12 is an interface circuit (e.g., a LAN card) for communicating with other devices, such as the server 20, the administrator terminal 30, or the staff terminal 40, via the network 50.

The camera 13 is placed, for example, on the ceiling or upper part of a wall of a room, captures an area including the subject 70's bed 60 as an observation area, and outputs a captured image (image data). Hereinafter, the image captured by the camera 13 will also be referred to simply as a "captured image." The captured image includes an image including the subject 70. The captured image includes still images and videos. The camera 13 is a near-infrared camera, but a visible light camera may be used instead, or both may be used in combination.

The control unit 11 recognizes the behavior of the subject 70 from the images captured by the camera 13. The recognized behaviors include "getting up" to get up from the bed 60, "getting out" to leave the bed 60, "falling" to fall off the bed 60, and "falling" to fall onto the floor or the like.

The control unit 11 detects image silhouettes (hereinafter referred to as "human silhouettes") from multiple captured images (moving images). Human silhouettes can be detected, for example, by extracting a range of pixels with a relatively large difference using a time difference method that extracts the difference between images captured before and after the time of capture. Human silhouettes may also be detected using a background difference method that extracts the difference between a captured image and a background image. Whether the subject 70 is getting up, getting out of bed, falling over, or falling down is recognized from the detected human silhouette based on the subject 70's posture (e.g., standing, sitting, lying down, etc.) and the relative position to objects installed in the room such as the bed 60. These recognitions may be performed by a program processed by the CPU of the control unit 11, or may be performed by an embedded processing circuit. In addition, without being limited to this, all or most of these recognition processes may be performed on the server 20 side, and the control unit 11 may only transmit the captured images to the server 20. When the control unit 11 recognizes any of the actions, it transmits a notification that an event has occurred to the server 20, etc.

The care call unit 14 includes a push button switch, and detects a care call when the subject 70 presses the switch. A care call is also called a nurse call. Instead of a push button switch, a care call may be detected by an audio microphone. When the switch of the care call unit 14 is pressed, i.e., when a care call is detected, the control unit 11 transmits a notification that a care call has been made (care call notification) to the server 20, etc., via the communication unit 12 and the network 50.

The voice input/output unit 15 is, for example, a speaker and a microphone, and enables voice calls by transmitting and receiving voice signals between the staff terminal 40 and the like via the communication unit 12. The voice input/output unit 15 may be connected to the detection unit 10 via the communication unit 12 as an external device of the detection unit 10.

The detection unit 10 may further include a Doppler shift type body movement sensor that transmits and receives microwaves in the direction of the bed 60 and detects the Doppler shift of the microwaves caused by body movements (e.g., breathing movements) of the subject 70. This body movement sensor detects chest body movements (up and down movement of the chest) caused by the breathing movements of the subject 70, and recognizes that there is an abnormality in the micro-body movement when it detects a disruption in the period of the chest body movement or an amplitude of the chest body movement that is below a preset threshold.

The detection unit 10 may further include a sensor that detects the living environment of the subject 70, such as the temperature, air pressure, and humidity in the subject's 70 room. The detection unit 10 may include a sensor that detects the temperature, air pressure, and humidity in the facility other than the subject's 70 room, and may include a sensor that detects the temperature, air pressure, and humidity outside the facility.

The staff 80 responds to various events according to the task. The tasks may include medical tasks and nursing care tasks. Here, the response content for each event will be described when the task of the staff 80 is nursing care tasks for the subject 70. The event is determined as "wake up", and if the determination is within a predetermined time (wake up time set by the facility (e.g., 7 to 8 a.m.)), morning care is performed. This morning care includes face washing, assistance with brushing teeth, wearing dentures, assistance with changing clothes, etc. In addition, if the event is "getting out of bed", wheelchair transfer and walking assistance may be required. In addition, periodic (fixed time) events other than the events determined by the detection unit 10 include water intake, meal assistance, excretion assistance, wheelchair transfer, walking assistance, and position change (prevention of bedsores). For these periodic events, an alert may be generated by the care call unit 14 or the like at the fixed time.

Events broadly include occurrences for which the staff 80 should respond to the subject 70. For example, events include an occurrence for which the staff 80 should respond by being notified (alert) to the staff 80 (referred to as a first occurrence), an occurrence for which the staff 80 should respond at a fixed time (referred to as a second occurrence), and an occurrence for which the staff 80 should respond when they notice (referred to as a third occurrence). The first occurrences include, for example, a care call, as well as getting up, getting out of bed, falling, falling, and abnormal micro-movements. The second occurrences include, for example, eating, bathing, changing clothes, and transferring to a wheelchair for a walk, etc. The third occurrences include, for example, excretion care at the verbal request of the subject 70, temperature measurement when a fever is suspected, a response when vomiting, and a response when dementia symptoms occur.

The detection unit 10 transmits (outputs) information on the detected event (hereinafter also simply referred to as "event information"), such as Doppler shift, living environment, and captured images, to the server 20.

(Server 20)
FIG. 4 is a block diagram showing the hardware configuration of the server 20.

The server 20 includes a control unit 21, a communication unit 22, and a storage unit 23. Here, the server 20 corresponds to a specific example of an information processing device of the present invention. The server 20 predicts the occurrence or non-occurrence of dementia symptoms of the subject 70 in a second period following the first period, for example, using information such as the Doppler shift, living environment, and captured images detected by the detection unit 10 in a first period. The second period is, for example, about one day from the time the subject wakes up on the day until the subject wakes up the next day, and the first period is, for example, about one week prior to the time the subject wakes up on the day. The dementia symptoms are so-called BPSD (Behavioral and Psychological Symptoms of Dementia).

The server 20 may be provided in the same building as the room for the subject 70, or may be provided in a remote location and connectable via a network. For example, the server 20 may be a cloud server that is virtually constructed by multiple servers located on a network such as the Internet. Each component is connected to each other via a bus so that they can communicate with each other. The control unit 21 and the communication unit 22 have the same functions as the control unit 11 and the communication unit 22 of the detection unit 10, respectively, and therefore detailed description thereof will be omitted. The specific functions of the control unit 21 will be described later.

The memory unit 23 stores an information processing program according to this embodiment. The above components of the server 20 are controlled by the control unit 21 according to the program. The memory unit 23 stores various information such as an event list, captured images, Doppler shift of the subject 70, the living environment of the subject 70, a care recipient list, a care list, and index information. The care recipient list includes information such as the subject 70's name, nursing care certification level, whether or not they have dementia, and if so, the level of dementia, and whether or not they use a wheelchair. The care list also includes the care history of each subject 70, such as the amount of food intake, amount of water intake, and excretion status. The care list also includes the care history when each subject 70 developed dementia symptoms.

The index information is information including at least one of a state index and an environmental index of the subject 70. The state index is an index related to the living state of the subject 70 in the first period. The environmental index is an index related to the living environment of the subject 70 in the first period. It is preferable that the index information includes a state index and an environmental index. This makes it possible to predict with higher accuracy the occurrence or nonoccurrence of dementia symptoms in the subject 70.

5 and 6 show examples of status indicators, and FIG. 7 shows examples of environmental indicators. The status indicators are, for example, indicators related to ADL (Activities of Daily Living), and preferably include at least one of an indicator representing the movements of the subject 70 and an indicator representing the sleeping state of the subject 70. This is because the movements and sleeping state of the subject 70 are thought to be closely related to the occurrence of dementia symptoms in the subject 70.

The motion of the subject 70 includes, for example, at least one of the following: movement, stopping, repetitive behavior, and unsteadiness of the subject 70.

For example, the indices representing the movement of the subject 70 are index 1 to index 3 for the daytime movement range, index 4 to index 6 for the nighttime movement range, index 15 for the overflow area, index 16 for the movement speed, and indexes 21 to 24 for the movement time and movement distance. Index 1 to index 6 are calculated based on the area of the visited area of the subject 70 outside the bed 60. Index 15 is calculated based on the overflow area from the high frequency passing route outside the bed 60 at night. Index 16 is calculated based on the movement speed of the subject 70 outside the bed 60 at night. Index 21 to index 24 are calculated based on the stay time and movement distance of the subject 70 outside the bed 60.

For example, the indices representing the subject 70 stopping are stopping indicators 17 to 20. These indicators 17 to 20 are calculated based on the time and number of times the subject 70 stops, other than for daily activities outside the bed 60.

For example, the indices representing the repetitive behavior of the subject 70 are Indicators 7 to 10 for daytime repetitive behavior and Indicators 11 to 14 for nighttime repetitive behavior. Indicators 7 to 14 are calculated based on the number of times and duration of the repetitive behavior of the subject 70.

For example, the index representing the subject's 70 unsteadiness is the unsteadiness index 25. The index 25 is calculated based on the degree of unsteadiness of the subject's 70 trajectory outside the bed 60 at night.

The sleep state of the subject 70 includes, for example, at least one of the sleep time, sleep rate, and sleep stability of the subject 70. For example, the indices representing the sleep time and sleep rate of the subject 70 are indexes 26 to 30 of nighttime sleep time/sleep rate. Indices 26 to 30 are calculated based on the nighttime sleep time and nighttime wakefulness time of the subject 70. For example, the indices representing the sleep stability of the subject 70 are indexes 31 and 32 of nighttime sleep stability. Indices 31 and 32 are calculated based on the sleep pattern stability of the subject 70 from midnight to 5 am, etc.

The environmental index includes at least one of an index representing the temperature of the living environment of the subject 70, an index representing the air pressure of the living environment of the subject 70, and an index representing the humidity of the living environment of the subject 70. The temperature, air pressure, and humidity are, for example, the temperature, air pressure, and humidity of the subject 70's room. For example, the index representing the temperature of the living environment of the subject 70 is index 35, the index representing the air pressure of the living environment of the subject 70 is index 33, and the index representing the humidity of the living environment of the subject 70 is index 34.

Indicators 1 to 35 are calculated, for example, based on information detected by the detection unit 10. For example, indicators 1 to 35 calculated by the detection unit 10 are transmitted to the server 20 and stored in the storage unit 23. Alternatively, the server 20 may calculate indicators 1 to 35 based on information transmitted from the detection unit 10 and store them in the storage unit 23. The detection unit 10 may calculate a part of indicators 1 to 35, and the server 20 may calculate the other part. Alternatively, an external device may calculate a part or all of indicators 1 to 35 based on information transmitted from the detection unit 10, and indicators 1 to 35 transmitted from the external device to the server 20 may be stored in the storage unit 23.

The memory unit 23 further stores a machine learning model trained by machine learning. In this embodiment, the server 20 uses this machine learning model to predict the occurrence or non-occurrence of dementia symptoms in the subject 70. The memory unit 23 may also store learning data used for learning.

The above information may be stored in a location other than the memory unit 23 of the server 20, and may be stored in various storage devices such as a cloud server connected to the server 20 via a network. In addition, the above information is not limited to being constructed as a single database, and may be stored in a distributed manner in multiple databases that are associated with each other.

The server 20, by using the event information and captured images received from the detection unit 10 alone or in cooperation with the detection unit 10, determines (identifies) which subject 70 the event recognized by the detection unit 10, such as getting up, getting out of bed, falling, or calling the nurse, pertains to. This determination is made by determining the subject 70 (i.e., the occupant of the room) associated with the room number in which the detection unit 10 that recognized the event is installed. Note that, if the subject 70 carries an IC tag, the subject 70 may be determined by reading the IC tag with an RFID reader installed in each room. Note that, if there are multiple subjects 70 in one room, such as in a shared room, the subject 70 may be determined by placing a detection unit 10 for each bed 60.

When the server 20 receives event information from the detection unit 10, it notifies the staff member 80 of the occurrence of the event by sending an event notification including the name of the subject 70 who caused the event, the room number, and the details of the event to the staff terminal 40, and instructs the staff member 80 on how to respond to the event.

(Administrator terminal 30)
Fig. 8 is a block diagram showing the hardware configuration of the administrator terminal 30. The administrator terminal 30 is a so-called PC (Personal Computer) and has a control unit 31, a communication unit 32, a display unit 33, and an input unit 34, which are interconnected by a bus.

The control unit 31 has a similar configuration to the control unit 11 of the detection unit 10, and includes a CPU, RAM, ROM, etc.

The communication unit 32 is an interface for various local connections, such as a network interface for wired communication based on standards such as Ethernet (registered trademark), and an interface for wireless communication based on standards such as Bluetooth (registered trademark) and IEEE 802.11, and communicates with each terminal connected to the network 50.

The display unit 33 is, for example, a liquid crystal display, and displays various information.

The input unit 34 is equipped with a keyboard, numeric keypad, mouse, etc., and is used to input various information.

In this embodiment, the administrator terminal 30 is used as a terminal for the administrator 90. The administrator 90 is, for example, a manager who supervises the staff 80.

The administrator terminal 30 receives instructions from the administrator 90, etc., regarding the analysis and output of care information, and transmits the received instructions to the server 20.

(Staff terminal 40)
9 is a block diagram showing the hardware configuration of the staff terminal 40. The staff terminal 40 has a control unit 41, a wireless communication unit 42, a display unit 43, an input unit 44, and a voice input/output unit 45, which are interconnected by a bus.

The control unit 41 has the same configuration as the control unit 11 of the detection unit 10, and includes a CPU, RAM, ROM, etc.

The wireless communication unit 42 is capable of wireless communication using standards such as Wi-Fi and Bluetooth (registered trademark), and communicates wirelessly with each device either directly or via the access point 51.

The display unit 43 and the input unit 44 are, for example, touch panel displays, and a touch sensor serving as the input unit 44 is superimposed on the display surface of the display unit 43, which is made of liquid crystal or the like. Various instructions are notified to the staff 80 by the display unit 43 and the input unit 44. The display unit 43 and the input unit 44 also display an operation screen displaying an event notification, and accept various operations such as input of a response to accept a response to an event or input of a care record through the operation screen. When a dementia symptom occurs in a subject 70, for example, the staff 80 inputs to the staff terminal 40 via the input unit 44 what dementia symptom has occurred for which subject 70. The staff 80 may also input the date and time of the onset of the dementia symptom.

The voice input/output unit 45 is, for example, a speaker and a microphone, and enables the staff member 80 to make voice calls between other staff terminals 40 via the wireless communication unit 42. The staff terminal 40 can be configured as, for example, a portable communication terminal device such as a tablet computer, a smartphone, or a mobile phone.

[Function of control unit 21]
Next, specific functions of the control unit 21 of the server 20 will be described with reference to FIG. 10 and FIG.

Figure 10 is a block diagram showing the functional configuration of the control unit 21. The control unit 21 functions as an acquisition unit 211, a prediction unit 212, and an output unit 213 by reading a program stored in the storage unit 23 and executing processing.

Figure 11 illustrates an example of index information acquired by the control unit 21 and occurrence information regarding the occurrence of dementia symptoms of subjects 70 (Mr. A and Mr. B) predicted based on this index information. Figure 11 illustrates an example of predicted occurrence of dementia symptoms of subject 70 from the time of waking up on January 2, 2023 to the time of waking up on January 3, 2023.

The acquisition unit 211 acquires the index information. As described above, the index information includes at least one of a state index related to the living condition of the subject 70 in the first period and an environmental index related to the living environment of the subject 70 in the first period. The acquisition unit 211 acquires the index information from, for example, the memory unit 23. The first period is, for example, a period of about one week from the time the subject 70 wakes up on December 26, 2022 to the time he wakes up on January 2, 2023. The first period is preferably longer than the second period, and more preferably a period of one week or more. This makes it possible to predict the onset of dementia symptoms with higher accuracy.

The index information includes, for example, index 33 to index 35 (a) from when the subject 70 wakes up on January 2, 2023. The index information includes index 1 to index 30 (d) from when the subject 70 wakes up on January 1, 2023 to when he wakes up on January 2, 2023. The index information includes index 31 and index 32 (g) from when the subject 70 wakes up on December 26, 2022 to when he wakes up on January 2, 2023.

It is preferable that the index information acquired by the acquisition unit 211 includes at least one of information regarding changes in the condition index of the subject 70 during the first period and information regarding changes in the environmental index of the subject 70 during the first period. This makes it possible to predict the occurrence information of the subject 70 with higher accuracy.

For example, the index information includes information (b) regarding the difference between indexes 33 to 35 when the subject 70 wakes up on January 2, 2023 and indexes 33 to 35 when the subject 70 wakes up on January 1, 2023. The index information includes information (c) regarding the variance of indexes 33 to 35 from the time the subject 70 wakes up on December 26, 2022 to the time the subject 70 wakes up on January 2, 2023. The variance of indexes 33 to 35 can be calculated, for example, using standard deviation.

For example, the index information includes information (f) regarding a comparison between indexes 26 to 30 from waking up on January 1, 2023 to waking up on January 2, 2023 and indexes 26 to 30 from waking up on December 31, 2022 to waking up on January 1, 2023. This comparison information is calculated, for example, by division.

For example, the index information includes information (e) regarding a comparison between indexes 1 to 25 from when the subject 70 wakes up on January 1, 2023 to when he wakes up on January 2, 2023 and the average of indexes 1 to 25 from when he wakes up on December 26, 2022 to when he wakes up on January 1, 2023. This information regarding the comparison is calculated, for example, by division.

The prediction unit 212 predicts occurrence information regarding the occurrence of dementia symptoms of the subject 70 in the second period based on the index information acquired by the acquisition unit 211. The second period is a period later than the first period, and is, for example, a period of about one day from the time when the subject 70 wakes up on January 2, 2023 to the time when he wakes up on January 3, 2023. The occurrence information preferably includes information regarding the occurrence or non-occurrence of dementia symptoms and information regarding the type of dementia symptoms that occurs. By including information regarding the type of dementia symptoms in the occurrence information, caregivers and the like can take more detailed measures in advance. The prediction unit 212 predicts the occurrence information using, for example, a machine learning model. This makes it possible to predict the occurrence or non-occurrence of dementia symptoms of the subject 70 with higher accuracy. The generation of the machine learning model will be described later.

For example, the prediction unit 212 predicts that Mr. A will develop the dementia symptom of wandering between waking up on January 2, 2023 and waking up on January 3, 2023. In addition to wandering, types of dementia symptoms include, for example, visual and auditory hallucinations, delusions, going out without permission, accompanying someone, irritability, depression, drowsiness, and shouting. The occurrence information predicted by the prediction unit 212 may include information on multiple types of dementia symptoms. For example, the prediction unit 212 may predict that Mr. A will develop the dementia symptoms of wandering and shouting.

The prediction unit 212 predicts that, for example, for Mr. B, dementia symptoms will not occur between waking up on January 2, 2023 and waking up on January 3, 2023.

The output unit 213 outputs the occurrence information of the subject 70 predicted by the prediction unit 212. The output unit 213 outputs the occurrence information, for example, by transmitting the occurrence information to the staff terminal 40 via the communication unit 22. The staff terminal 40 displays the received occurrence information on the display unit 43, for example. For example, the display unit 43 displays that there is a high possibility that Mr. A will develop the dementia symptom of wandering on January 2, 2023. The display unit 43 may display advance response instructions to the staff 80 in response to the development of the dementia symptom of the subject 70. When the prediction unit 212 predicts that the subject will not develop the dementia symptom in the second period, the output unit 213 may not output the occurrence information.

The detection unit 10, server 20, administrator terminal 30, and staff terminal 40 may include components other than those described above, or may not include some of the components described above.

[Operation of Server 20]
Next, the operation of the server 20 in the monitoring system 1 will be described.

Figure 12 is a flowchart showing the operation of the server. This flowchart can be executed by the control unit 21 according to a program stored in the memory unit 23 of the server 20.

The control unit 21 first acquires index information for the subject 70 (step S101). For example, the control unit 21 reads and acquires the index information stored in the storage unit 23. This index information includes, for example, status information regarding the living condition of the subject 70 in the first period and environmental information regarding the living environment of the subject 70 in the first period.

Then, the control unit 21 predicts occurrence information regarding the occurrence of dementia symptoms of the subject 70 in a second period that is later than the first period, based on the index information acquired in the processing of step S101 (step S102). The control unit 21 predicts the occurrence information of the subject 70, for example, by inputting the index information acquired in the processing of step S101 into a machine learning model. The control unit 21 predicts the occurrence information of the subject 70 for the day, for example, when the subject 70 wakes up on that day. The control unit 21 predicts the occurrence information of the subject 70, for example, every day, when the subject 70 wakes up.

Then, the control unit 21 outputs the occurrence information predicted in the processing of step S102 (step S103) and ends the processing.

[Generating machine learning models]
Next, a method for generating a machine learning model used by the control unit 21 will be described.

Figure 13 is a flowchart showing a method for generating a machine learning model. For example, the control unit 21 functions as a model generation unit and generates a machine learning model. Here, the server 20 corresponds to a specific example of a machine learning device.

In the process of FIG. 13, machine learning is performed using training data prepared in advance. This training data includes many data sets of explanatory variables and objective variables. The explanatory variables are training index information of dementia patients, and the objective variables are training occurrence information of dementia patients. The training index information includes at least one of a training state index related to the living state of the dementia patient in the first training period and a training environment index related to the living environment of the dementia patient in the first training period. The training state index and the training environment index can be obtained, for example, based on information detected by the detection unit 10. The training occurrence information is information related to the occurrence of dementia symptoms of the dementia patient in the second training period after the first training period. The training occurrence information can be obtained, for example, from the care history of the dementia patient inputted by the staff 80 to the staff terminal 40. The training index information and the training occurrence information are, for example, past information of multiple subjects 70. The learning index information and the learning occurrence information may be stored in the storage unit 23.

The learning machine (not shown) may be, for example, a standalone high-performance computer using a CPU and GPU processor, or a cloud computer. Below, a learning method using a neural network formed by combining perceptrons such as deep learning in the learning machine will be described, but this is not limited to this, and various other methods may be applied. For example, random forests, decision trees, support vector machines (SVMs), logistic regression, k-nearest neighbors, topic models, etc. may be applied.

(Step S111)
The learning device reads learning data. If it is the first time, the first set of learning data is read, and if it is the i-th time, the i-th set of learning data is read.

(Step S112)
The learning device inputs the explanatory variables from the loaded learning data into the neural network.

(Step S113)
The learner compares the prediction results of the neural network with the objective variable.

(Step S114)
The learning device adjusts the parameters based on the comparison results, for example, by performing a process based on back-propagation, so as to reduce the difference in the comparison results.

(Step S115)
If the learning device has completed processing of all data from the first to the i-th set (YES), the process proceeds to step S116; if not (NO), the process returns to step S111, the next learning data is read, and the process from step S111 onwards is repeated.

(Step S116)
The learning device determines whether or not to continue learning. If learning is to continue (YES), the process returns to step S111 and executes the processing from the first set to the i-th set again in steps S111 to S115. If learning is not to continue (NO), the process proceeds to step S117.

(Step S117)
The learning device stores the machine learning model trained by machine learning in the processes up to this point and ends (END). Storage destinations include the memory unit 23 of the server 20. In the above-described operation of the server 20, the machine learning model thus generated is used to predict occurrence information regarding the onset of dementia symptoms in the subject 70. The learning device may learn the learning data for each facility, or may learn the learning data for each subject 70.

[Effects of the server 20 and the monitoring system 1]
In the server 20 and the monitoring system 1 according to the present embodiment, the future occurrence of dementia symptoms of the subject 70 is predicted based on the index information of the subject 70. This allows the staff 80 or a caregiver such as a family member to take necessary measures in advance when the occurrence of dementia symptoms is predicted. This makes it possible to reduce the burden on the subject 70 and those around him/her caused by the occurrence of dementia symptoms.

The configuration of the monitoring system 1 described above is the main configuration described in order to explain the features of the above-mentioned embodiment, but is not limited to the above-mentioned configuration and can be modified in various ways within the scope of the claims. Furthermore, the configuration of a general monitoring system is not excluded.

For example, in the above embodiment, an example was described in which the second period is immediately after the first period, but the second period and the first period may be separated from each other.

In the above embodiment, an example was described in which the second period is one day from when the subject wakes up on the day until when the subject wakes up on the next day, but the second period may be two days or more. Alternatively, it may be a period from when the subject wakes up on the day until when the subject goes to bed on the day.

In the above embodiment, an example was described in which the server 20 outputs the generated information by transmitting it to the staff terminal 40, but the server 20 may output the generated information by transmitting it to the administrator terminal 30, or may output the generated information by transmitting it to an external display device, etc.

In addition, in the above embodiment, an example was described in which the server 20 predicts the occurrence information of the subject 70 using a machine learning model, but the server 20 may predict the occurrence information of the subject 70 using other methods such as statistical processing.

In addition, some or all of the functions of the server 20 may be provided by the administrator terminal 30, the staff terminal 40, or the detection unit 10.

Furthermore, the detection unit 10, the server 20, the administrator terminal 30, and the staff terminal 40 may each be configured as multiple devices, or any one of the devices may be included in the other devices and configured as a single device.

In addition, some steps in the above-described flowchart may be omitted, and other steps may be added. Some of the steps may be executed simultaneously, or one step may be divided into multiple steps and executed. The order in which the steps are executed may also be changed.

The means and methods for performing various processes in the monitoring system 1 according to the above-described embodiment can be realized by either a dedicated hardware circuit or a programmed computer. The above program may be provided by a computer-readable recording medium such as a USB memory or a DVD (Digital Versatile Disc)-ROM, or may be provided online via a network such as the Internet. In this case, the program recorded on the computer-readable recording medium is usually transferred to and stored in a storage unit such as a hard disk. The above program may be provided as a standalone application software, or may be incorporated as a function into the software of the device such as the detection unit.

1. Monitoring system,
10 detection unit,
11 control unit,
12 Communications Department,
13 Camera,
14 Care Call Department,
15 Audio input/output unit,
20 servers,
21 control unit,
22 Communications Department,
23 memory unit,
30 Administrator terminal,
31 control unit,
32 Communications Department,
33 display unit,
34 input unit,
40 Staff terminal,
41 control unit,
42 wireless communication unit,
43 display unit,
44 input unit,
45 Audio input/output unit,
46 position acquisition unit,
50 networks,
51 access points,
60 beds,
70 Subjects,
80 staff,
90 Administrator.

Claims (16)

an acquisition unit that acquires index information including at least one of a condition index related to a living condition of a target dementia patient in a first period and an environmental index related to a living environment of the target dementia patient in the first period;
and a prediction unit that predicts occurrence information regarding the occurrence of dementia symptoms of the target dementia patient in a second period that is later than the first period based on the acquired index information. The information processing device according to claim 1, wherein the acquisition unit acquires the index information including the state index and the environmental index. The information processing device according to claim 1, wherein the state index includes at least one of an index representing the movement of the target dementia patient and an index representing the sleep state of the target dementia patient. The information processing device according to claim 3, wherein the motion includes at least one of the following: movement, stopping, repetitive behavior, and unsteadiness of the target dementia patient. The information processing device according to claim 3, wherein the sleep state includes at least one of the sleep time, sleep rate, and sleep stability of the target dementia patient. The information processing device according to claim 1, wherein the environmental index includes at least one of an index representing the temperature of the living environment, an index representing the air pressure of the living environment, and an index representing the humidity of the living environment. The information processing device according to claim 1, wherein the occurrence information includes information regarding the occurrence or nonoccurrence of the dementia symptom in the target dementia patient and information regarding the type of the dementia symptom that occurs. The information processing device according to claim 7, wherein the occurrence information may include information relating to a plurality of the types. The information processing device according to claim 1, wherein the prediction unit predicts the occurrence information using a machine learning model. The information processing device according to claim 9, wherein the machine learning model is trained using learning index information including at least one of a learning state index relating to the living state of a dementia patient in a first learning period and a learning environment index relating to the living environment of the dementia patient in the first learning period as explanatory variables, and learning occurrence information relating to the occurrence of dementia symptoms of the dementia patient in a second learning period after the first learning period as objective variables. The information processing device according to claim 1, further comprising an output unit that outputs the predicted occurrence information. The information processing device according to claim 1, wherein the first period is longer than the second period. The information processing device according to claim 1, wherein the first period is a period of one week or more. The information processing device according to claim 1, wherein the index information includes at least one of information regarding changes in the state index during the first period and information regarding changes in the environmental index during the first period. acquiring index information including at least one of a condition index related to the living condition of a target dementia patient during a first period and an environmental index related to the living environment of the target dementia patient during the first period;
and predicting occurrence information regarding the occurrence of dementia symptoms in the target dementia patient in a second period after the first period based on the acquired index information. a storage unit in which learning data is stored, the learning data including, as explanatory variables, learning index information including at least one of a learning state index relating to the living state of a dementia patient in a first learning period and a learning environment index relating to the living environment of the dementia patient in the first learning period, and learning occurrence information relating to the occurrence of dementia symptoms of the dementia patient in a second learning period after the first learning period, as a response variable;
and a model generation unit that performs machine learning on the relationship between the learning index information and the learning generation information based on the stored learning data, to generate a machine learning model.