Health risk estimation system

Abstract

To provide a health risk estimation system capable of accurately estimating a health risk of a subject.SOLUTION: Provided are a behavior estimation unit (target person detection sensor 110, server 130) capable of estimating a behavior of a target person P, a motion estimation unit (server 130) capable of estimating a degree of shaking of a body when the target person P changes a posture, and a health risk estimation unit (server 130) capable of estimating a health risk of the target person P according to an estimation result of the degree of shaking of the body during the behavior estimated by the behavior estimation unit.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a technology for a health risk estimation system. [Background technology]

[0002] Conventionally, technology for a system for estimating a subject's health risk (health risk) has been publicly known, as described in Patent Document 1, for example.

[0003] In the technology described in Patent Document 1, a data measurement unit measures center of gravity sway while a subject is standing over a predetermined measurement period. Based on the body sway data measured by the data measurement unit, an evaluation value calculation unit calculates a balance age that evaluates the subject's musculoskeletal ability, a sensory response score that evaluates the subject's sensory ability, and a standing age that evaluates the subject's abilities based on both the musculoskeletal and sensory systems. The fall risk assessment unit then assesses the subject's fall risk (health risk) based on the balance age, sensory response score, and standing age calculated by the evaluation value calculation unit.

[0004] However, the technology described in Patent Document 1 acquires data from a subject who is standing. Here, health risks to the subject may occur, for example, when the subject moves around in a room, so the technology described in Patent Document 1 is insufficient to accurately estimate health risks of a subject who is living a normal life in a living environment. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2023-135502 Summary of the Invention [Problem to be solved by the invention]

[0006] The present invention has been made in consideration of the above-mentioned situation, and the problem it aims to solve is to provide a health risk estimation system that can accurately estimate the health risk of a subject. [Means for solving the problem]

[0007] The problem to be solved by the present invention is as described above, and the means for solving this problem will now be described.

[0008] That is, in claim 1, the device comprises a behavior estimation unit capable of estimating the behavior of a subject, a sway estimation unit capable of estimating the degree of body sway when the subject changes posture, and a health risk estimation unit capable of estimating the health risk of the subject based on the estimation result of the degree of body sway during the behavior estimated by the behavior estimation unit.

[0009] In claim 2, the degree of swaying of the body includes the degree of swaying of the center of gravity.

[0010] In claim 3, the sway estimation unit estimates the degree of swaying of the body when the posture of the subject changes from a sitting posture to a standing posture.

[0011] In claim 4, the motion estimation unit regards the degree of motion of the trunk of the subject as the degree of motion of the center of gravity.

[0012] In claim 5, the sway estimation unit acquires a movement trajectory of the torso when the subject changes posture, and estimates the degree of swaying of the body based on at least one of the number of times the torso sways or the amplitude of swaying of the torso obtained from the movement trajectory.

[0013] In claim 6, the health risk estimation unit learns in advance data that quantifies the degree of body shaking, and estimates the health risk based on the learning result.

[0014] In claim 7, the quantified data is created based on information obtained from the subject. [Effects of the Invention]

[0015] The present invention has the following effects.

[0016] In the present invention, the health risk of a subject can be estimated with high accuracy. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is a schematic diagram showing the overall configuration of a health risk estimation system according to one embodiment of the present invention and a room to which the health risk estimation system is applied. [Figure 2] A table showing the relationship between the areas set up in the room and the subject's behavior. [Figure 3] 10 is a flowchart showing processing in a measurement phase. [Figure 4] FIG. 10 is an explanatory diagram showing a stay time for identifying the type of standing-up motion. [Figure 5] (a) A diagram showing the torso movement trajectory when a healthy person stands up, (b) A diagram showing the torso movement trajectory when a person with a low level of lower limb muscle strength stands up, and (c) A diagram showing the torso movement trajectory when a person at risk of falling stands up. [Figure 6] FIG. 10 is an explanatory diagram regarding the number of times the center of gravity sways and the amplitude of the center of gravity sway. [Figure 7] (a) A graph plotting the center of gravity sway index when using the toilet, (b) A graph plotting the center of gravity sway index when eating, and (c) A graph plotting the center of gravity sway index when getting up. [Figure 8] 10 is a flowchart showing the processing of the learning phase. [Figure 9] FIG. 10 is a diagram showing an example of learning data. [Figure 10] 10 is a flowchart showing a first estimation flow. [Figure 11]1A is a graph plotting the most recent body sway index, and FIG. 1B is a graph plotting the most recent body sway index and the regression line calculated in the learning phase. [Figure 12] 10 is a flowchart showing a second estimation flow. [Figure 13] (a) A diagram showing a graph on which past measurement results of the center of gravity sway index are plotted. (b) A diagram showing how to determine the change in the intercept of the regression line. [Figure 14] 10 is a flowchart showing a third estimation flow. [Figure 15] FIG. 10 is a diagram showing a state in which the center of gravity sway index is on the rise. DETAILED DESCRIPTION OF THE INVENTION

[0018] A health risk estimation system 100 according to one embodiment of the present invention will be described below with reference to FIG.

[0019] The health risk estimation system 100 is capable of estimating the health risk (health risk) of the subject P. In this embodiment, the "health risk" refers to a health risk that can be estimated from, for example, the risk of the subject P falling (fall risk).

[0020] In this embodiment, it is assumed that the health risk estimation system 100 is applied to a room 1 of a subject P (see FIG. 4) in a facility for the elderly. The health risk estimation system 100 also estimates the risk of falls of elderly people. The health risk estimation system 100 may also estimate the health risk of people other than elderly people. FIG. 1 shows an example of a room 1 of a subject P (elderly person).

[0021] The living room 1 is equipped with a bed 10, a dining table 20, dining chairs 30, a toilet 40, a bathtub 50, a kitchen 60, and an entrance / exit 70. The entrance / exit 70 is the entrance to the living room 1. The space in which the toilet 40 and the bathtub 50 are arranged is separated from other spaces by a partition wall 41. A sink (not shown) is also arranged in the space in which the bathtub 50 is arranged. The partition wall 41 is provided with an entrance / exit for accessing the toilet 40 and the bathtub 50. Although the illustration does not show a wall separating the space in which the toilet 40 is arranged from the space in which the bathtub 50 is arranged, a wall separating each of the above spaces may be provided. The configuration of the living room 1 is not limited to the one described above, and various spaces such as a study or a sofa area (a space in which a sofa or the like is arranged) can be arranged.

[0022] Next, we will explain the configuration of the health risk estimation system 100. The health risk estimation system 100 mainly comprises a subject detection sensor 110, a staff terminal 120, and a server .

[0023] The subject detection sensor 110 is for detecting the subject P. The subject detection sensor 110 is a non-contact sensor. In this embodiment, a millimeter wave sensor is used as the subject detection sensor 110. The subject detection sensor 110 irradiates a subject with millimeter waves (radio waves with a wavelength of 1 to 10 mm and a frequency of 30 to 300 GHz) and observes the reflected radio waves, thereby measuring the position, speed, angle, movement, posture, etc. of the subject.

[0024] In this embodiment, the subject detection sensor 110 is installed in a position (for example, on a ceiling or wall) in the living room 1 where radio waves can be irradiated to the entire living room 1. Note that although this embodiment shows an example in which one subject detection sensor 110 detects the entire living room 1, multiple subject detection sensors 110 may be used as necessary.

[0025] The staff terminal 120 is used to notify the estimation results, etc., obtained by the health risk estimation system 100. For example, a stationary personal computer can be used as the staff terminal 120. The staff terminal 120 is installed in a location where it can be seen by people who provide assistance (caregiving, nursing care, etc.) to the subject P in their daily lives. In this embodiment, it can be placed, for example, in a staff station at an elderly care facility. By displaying various information on the monitor of the staff terminal 120, the information can be notified to the staff. It is to be noted that various other devices (for example, a portable terminal, etc.) can also be used as the staff terminal 120.

[0026] The server 130 is used to perform processing for estimating the risk of falling using the health risk estimation system 100. The server 130 is configured, for example, by a cloud server. The server 130 can exchange information with the subject detection sensor 110 and the staff terminal 120.

[0027] The server 130 can determine the position and posture of the subject P based on the data obtained by the subject detection sensor 110. Specifically, the server 130 can determine that data with movement among the point cloud data obtained based on the reflected radio waves is a person (in this embodiment, the subject P). In particular, since a millimeter wave sensor is used in this embodiment, even the slightest movement of the subject P can be detected with high accuracy. The server 130 can extract only the data with movement and determine that the position of the data in the room 1 is the position of the subject P (each area described below). Furthermore, the server 130 can determine the posture of the subject P, such as whether the subject P is standing (standing position), sitting (sitting position), or lying down (supine position), by extracting only the data with movement and checking its distribution.

[0028] Furthermore, the server 130 can determine the position of each part of the subject P's body based on the data obtained by the subject detection sensor 110. For example, the server 130 can identify a specific area corresponding to the subject P's torso from the above-mentioned moving data (data group), and determine this as the position of the subject P's torso.

[0029] The server 130 stores in advance location information of areas obtained by dividing the living room 1 into a plurality of areas. In Fig. 1, as an example, a bed area A, a dining area B, a toilet area C, a bathroom area D, an other area E, and an outdoor area F are set.

[0030] The bed area A is an area (bedroom) where the bed 10 is placed. The bed area A is set to be, for example, slightly larger than the bed 10 in a plan view. Similarly, the dining area B is set in the area where the dining table 20 and dining chairs 30 are placed. The toilet area C and bathroom area D (washroom) are set within the area surrounded by the partition wall 41, where the toilet 40 and bathtub 50 are placed. Areas of the living room 1 other than the bed area A, dining area B, toilet area C, and bathroom area D are set as other areas E. In addition, the area outside the entrance 70 is set as the outdoor area F.

[0031] It should be noted that the above areas are merely examples, and any area other than areas A to F can be set. However, in this embodiment, as will be described later, the degree of body swaying when subject P stands up is measured, so it is desirable that the area be one in which subject P stands up. For example, a study area with a chair and desk, or a sofa area with a sofa may be set. Furthermore, for example, if subject P can cook while sitting, the area where kitchen 60 is located may be set as the kitchen area.

[0032] The server 130 estimates the behavior based on information such as information on the area where the subject P is located (location information), posture data of the subject P, and time (time when the subject P is located in a predetermined area).

[0033] FIG. 2 is a table showing the relationship between the area where the subject P is located and the behavior of the subject P. In the server 130, several types of behaviors (e.g., "sleeping") shown in the table are set according to behaviors (e.g., lying down in bed area A) that are expected from information such as the position information and posture data of the subject P. When the server 130 detects any of the above behaviors, it infers that an action corresponding to the behavior has been performed. An example of a method by which the server 130 infers the behavior of the subject P using the table will be described below.

[0034] When the subject P is located in the bed area A (bedroom), the server 130 estimates that the behavior of the subject P is one of "sleeping," "sleeping (asleep)," and "wake-up." For example, when the server 130 detects the subject P in the bed area A at a bedtime in a preset schedule and the subject P is in a lying position, the server 130 estimates that the subject P is about to go to sleep (the behavior of the subject P is "sleeping"). Furthermore, for example, when the server 130 detects that the subject P is in a lying position for a predetermined period or longer after "sleeping," the server 130 estimates that the behavior of the subject P is "sleeping." Furthermore, for example, when the server 130 detects that the subject P has changed position from a lying position to a sitting position and then to a standing position after "sleeping" (i.e., when the server 130 detects that the subject P has stood up after sleeping), the server 130 estimates that the behavior of the subject P is "wake-up."

[0035] Furthermore, when the subject P is located in the toilet area C, the server 130 estimates that the behavior of the subject P is either "excretion (small)" or "excretion (large)." More specifically, for example, if the time that the subject P stays seated on the valve seat of the toilet 40 is less than a preset period, the server 130 estimates that the behavior of the subject P is "excretion (small)," and if the time that the subject P stays in the toilet 40 is equal to or longer than the above period, the server 130 estimates that the behavior of the subject P is "excretion (large)." Note that the server 130 can also estimate that the behavior of the subject P is "excretion (small)" or "excretion (large)" based only on the posture of the subject P detected by the subject detection sensor 110.

[0036] Furthermore, when the subject P is located in a study (not shown), the server 130 estimates that the behavior of the subject P is either "work" or "hobby." For example, if the time that the subject P is seated in the study (stay time) is less than a preset period, the server 130 estimates that the behavior of the subject P is "hobby," and if the time that the subject P is staying in the study is equal to or longer than the above period, the server 130 estimates that the behavior of the subject P is "work." Note that the server 130 can also estimate that the behavior of the subject P is "work" or "hobby" based only on the posture of the subject P detected by the subject detection sensor 110.

[0037] Furthermore, when subject P is located in bathroom area D, server 130 estimates that subject P's behavior is either "bathing" or "grooming." More specifically, server 130 estimates, for example, that subject P's behavior of sitting on a chair in the bathroom (washroom) is "grooming," and that subject P's behavior of staying in the bathroom is "bathing." In this embodiment, it is estimated that subject P is sitting on a bath stool while staying in the bathroom. Server 130 can also estimate that subject P's behavior is "grooming" or "bathing" based solely on the posture of subject P detected by subject detection sensor 110.

[0038] Furthermore, when the subject P is located in the dining area B, the server 130 estimates that the behavior of the subject P is one of "eating," "resting," and "watching television." More specifically, the server 130 estimates that the behavior of the subject P, who is seated in a dining chair in the dining area B during a preset mealtime, is "eating." Furthermore, when the subject P is seated in a dining chair outside of a mealtime and the time spent sitting (staying time) is less than a preset period, the server 130 estimates that the behavior is "resting." Furthermore, when the subject P is seated in a dining chair outside of a mealtime and the time spent sitting (staying time) is equal to or longer than a preset period, the server 130 estimates that the behavior is "watching television." The server 130 can also estimate that the behavior of the subject P is "eating," "resting," or "watching television" based on the posture of the subject P detected by the subject detection sensor 110, rather than the time of day or the time spent sitting.

[0039] Furthermore, even when the target person P is located in the sofa area (not shown), the server 130 can estimate that the target person P's activity is one of "eating," "resting," and "watching TV." The method for estimating "eating," etc. may be the same as when the target person P is located in the dining area B, except for the area.

[0040] Furthermore, when the target person P is located in the outdoor area F, the server 130 estimates that the target person P's behavior is "going out" or "recreation." More specifically, when the server 130 no longer detects the target person P at the entrance / exit 70, the server 130 estimates that the target person P is located in the outdoor area F (the target person P's behavior is "going out" or the like).

[0041] The above-described method for estimating the behavior of the subject P is an example, and the method for estimating the behavior of the subject P is not limited to the above-described example. Specifically, the behavior of the subject may be estimated using any of information on the stay time in an area, the current time, and calendar (schedule). In addition to the above-described example, the server 130 can estimate the behavior of the subject P using various methods that use data obtained by the subject detection sensor 110.

[0042] Here, when a person stands up, they may experience dizziness or lightheadedness and become unsteady (increased body sway). For example, orthostatic hypotension is a symptom that strongly manifests dizziness and other symptoms. If subject P develops orthostatic hypotension, subject P is likely to experience increased body sway when standing up, which increases the risk of falling. Furthermore, the applicant's findings indicate that orthostatic hypotension (i.e., various health-related factors that tend to increase body sway when standing up) is strongly correlated with events (behaviors) that increase or decrease blood pressure, such as after excretion, eating, bathing, and going to bed. Therefore, the health risk estimation system 100 according to this embodiment effectively estimates (occurrence and tendency of) the risk of falling by observing the degree of body sway for each type of behavior of subject P.

[0043] Furthermore, body movement is linked to the movement of the body's center of gravity. Therefore, in this embodiment, the degree of body swaying (degree of unsteadiness) of subject P is determined based on the subject P's body, particularly the center of gravity. Specifically, the degree of swaying of subject P's center of gravity is considered to be the degree of body swaying of subject P. Note that the body's center of gravity is the center of mass distribution of the body, and is generally located near the navel (the trunk, which has the greatest mass among all body parts) in a standing position, but its position changes depending on the subject P's posture. Therefore, in this embodiment, for the convenience of system processing, the position of the trunk among all body parts of subject P is considered to be the position of the subject P's center of gravity, regardless of the subject P's posture.

[0044] Below, an outline of the process (fall risk estimation process) executed in the health risk estimation system 100 will be described. The fall risk estimation process is a process that estimates the risk of a subject P falling for each (type of) movement of the subject P based on the degree of sway of the center of gravity when standing up. In this embodiment, the fall risk estimation process is executed by the server 130. The fall risk estimation process can be divided into multiple steps (three in this embodiment) depending on the respective purposes. Specifically, the fall risk estimation process includes a measurement phase, a learning phase, and a prediction phase.

[0045] In this embodiment, the "behavior" of subject P for estimating the risk of falling refers to the behavior accompanying the standing-up action from a sitting position to a standing position, and the standing-up action itself. For example, the "behavior" of the standing-up action itself includes "getting up" (which is the standing-up action itself) in bed area A. Furthermore, for example, the "behavior" accompanying the standing-up action includes "excretion (small)" and "excretion (large)" (which occur before the standing-up action) in toilet area C.

[0046] First, the specific processing of the measurement phase will be described with reference to FIG.

[0047] The processing in the measurement phase is a step (processing) of measuring the degree of swaying of the center of gravity when the subject P stands up for each action of the subject P. The processing in the measurement phase is started, for example, after the health risk estimation system 100 is introduced into an aging facility, by a caregiver performing a predetermined operation on the staff terminal 120. Once the processing in the measurement phase is started, it is repeatedly and continuously executed at predetermined time intervals (for example, every few seconds).

[0048] In step S11, the server 130 detects the movement of the subject P based on the position of the subject P detected by the subject detection sensor 110 and the position information of each of the areas A to F. At this time, the server 130 detects the movement of the subject P between the areas. When the server 130 detects the movement of the subject P, the process proceeds to step S12.

[0049] In step S12, the server 130 detects the destination of the subject P based on the position of the subject P detected by the subject detection sensor 110 and the position information of each of the areas A to F. After performing the process of step S12, the server 130 proceeds to step S13.

[0050] In step S13, the server 130 determines whether the destination of the subject P is within the living room 1. Specifically, the server 130 determines whether the destination is other than the outdoor area F (bed area A, dining area B, etc.). If the destination is other than the outdoor area F ("YES" in step S13), the server 130 proceeds to step S15. On the other hand, if the destination is the outdoor area F ("NO" in step S13), the server 130 proceeds to step S14 and ends the processing of the measurement phase. In this way, the server 130 does not estimate the risk of falling when the subject P goes out and the subject detection sensor 110 cannot detect the subject P.

[0051] In step S15, the server 130 determines whether the posture of the subject P has changed to a sitting posture based on the posture of the subject P detected by the subject detection sensor 110. If the server 130 determines that the posture has changed to a sitting posture (YES in step S15), the server 130 proceeds to step S17. On the other hand, if the server 130 determines that the posture has not changed to a sitting posture (NO in step S15), the server 130 proceeds to step S16 and ends the processing of the measurement phase.

[0052] In step S17, the server 130 starts measuring the staying time T of the subject P at the destination detected in step S12. Note that the staying time T in this embodiment is the time from when the subject P sits down at the destination to when he or she stands up (the time from when the subject P's posture changes to a sitting posture to when he or she changes to a standing state), as shown in FIG. 4, for example. Note that the staying time T is a time mainly used to identify the type of standing-up movement of the subject P. In addition to the staying time T, the server 130 can measure various times, such as the time from when the subject P arrives at the destination to when he or she leaves (staying time at the destination) and the time according to the subject P's behavior at the destination. As shown in FIG. 3, the server 130 performs the process of step S17 and then proceeds to step S18.

[0053] In step S18, if the posture detected by the subject detection sensor 110 has changed to a standing posture ("YES" in step S18), the server 130 proceeds to step S19. On the other hand, if the posture has not changed to a standing posture ("NO" in step S18), the server 130 proceeds to step S18 again.

[0054] In step S19, the server 130 acquires the fluctuation of the center of gravity of the subject P during the standing-up motion from a sitting position to a standing position. The fluctuation of the center of gravity is a fluctuation in the position of the center of gravity. In this embodiment, as described above, regardless of the posture of the subject P, the position of the trunk of the subject P, among the various body parts, is regarded as the position of the center of gravity of the subject P. That is, in step S19, the server 130 regards the position of the trunk of the subject P detected by the subject detection sensor 110 as the position of the center of gravity, and acquires the fluctuation in the position of the center of gravity during the standing-up motion.

[0055] Specifically, the server 130 acquires the position of the torso of the subject P detected by the subject detection sensor 110 when the subject P stands up. The server 130 can determine, from the data group acquired by the subject detection sensor 110, a specific point within a specific area corresponding to the torso of the subject P as the position of the torso. For example, the server 130 can determine a point located in the center of the specific area as the position of the torso. The acquisition of the torso position by the server 130 is not limited to this. The server 130 then connects the acquired detection results in a chronological order to calculate a movement trajectory M of the torso of the subject P when the subject P stands up. The server 130 determines the calculated result as the fluctuation of the center of gravity. FIG. 5 shows an example of the movement trajectory M.

[0056] The movement trajectory M shown in Figure 5 is a plan view of the movement trajectory of the torso of the subject P when standing up in the toilet area C. As shown in Figure 5(a) and other figures, when the subject P sits on the valve seat of the toilet 40 and stands up, the torso of the subject P gradually moves generally forward (Y direction) (based on the orientation of the body) as the subject P stands up from the seated state. Furthermore, as the torso of the subject P moves forward, the body sways depending on the health condition (risk of falling) of the subject P, and the position of the torso fluctuates left and right.

[0057] Specifically, Fig. 5(a) shows a movement trajectory M1 of a healthy person standing up. Fig. 5(b) shows a movement trajectory M2 of a person whose lower limb muscle strength has begun to decline. Fig. 5(c) shows a movement trajectory M3 of a person at risk of falling due to the progression of the decline in lower limb muscle strength from Fig. 5(b). Note that people at risk of falling include not only people whose lower limb muscle strength has progressed, but also people who are prone to falling due to unsteadiness when standing up, such as people who have developed orthostatic hypotension.

[0058] 3, when the server 130 acquires the movement trajectory M (sway of the center of gravity), the process proceeds to step S20. In step S20, the server 130 ends the measurement of the staying time T (see FIG. 4). After performing the process of step S20, the server 130 proceeds to step S21.

[0059] In step S21, server 130 calculates a center of gravity sway index based on movement trajectory M acquired in step S19. In this embodiment, the center of gravity sway index is a numerical value representing the degree of sway of the center of gravity during a standing-up movement.

[0060] Here, since a person at risk of falling tends to stagger when standing up, the movement trajectory M3 of the person at risk of falling will meander and have a more complex shape than the movement trajectories M1 and M2 of other people (see FIG. 5). Therefore, the server 130 quantifies the complexity (sway component) of the movement trajectory M according to the number of times the center of gravity sways and the sway amplitude W of the center of gravity. The number of times the center of gravity sways is the number of times the direction of movement of the center of gravity of the subject P changes on the movement trajectory M. In this embodiment, the sway amplitude W indicates the magnitude of lateral sway when the subject P stands up.

[0061] First, an example of a process for calculating the number of times the center of gravity has swayed (the number of swaying motions) will be described. As shown in FIG. 6, the server 130 identifies a portion Ma where the direction of the curve (swaying of the center of gravity) forming the movement trajectory M has changed in a planar view. Here, the identification of the portion Ma focuses on the movement direction of a point moving on the movement trajectory M from a start point S (the center of gravity sway index acquired first in the standing-up movement) to an end point F (the center of gravity sway index acquired last in the standing-up movement). That is, the server 130 resolves the movement direction of the point into two vectors, the X direction and the Y direction, and recognizes a point where the direction of at least one of the X direction and the Y direction vector has changed to the opposite direction from the most recent point as the portion Ma. The server 130 counts the number of identified portions Ma and calculates the result as the number of swaying motions.

[0062] Next, an example of the process for calculating the shaking width W will be described. The server 130 estimates the lateral direction X based on the subject P. For example, the server 130 estimates the facial orientation of the subject P in a sitting position based on the detection result of the subject detection sensor 110. The server 130 then determines the direction perpendicular to the estimated facial orientation in a plan view as the lateral direction X. Note that the Y direction shown in FIG. 6 is the direction parallel to the facial orientation.

[0063] When the server 130 identifies the lateral direction X, it calculates the width of the lateral sway when the center of gravity sways once. For example, the server 130 calculates the width along the lateral direction X between two adjacent portions Ma on the movement trajectory M as the width of one lateral sway. Furthermore, if the center of gravity sways two or more times (if there are three or more portions Ma), the server 130 calculates the width of the lateral sway for all swaying. The server 130 then sets the maximum value of the calculated widths as the sway width W. Note that the sway width W is not limited to the sway when the subject P stands up. For example, the sway width W can also be the length along the movement trajectory M between two adjacent portions Ma on the movement trajectory M.

[0064] The server 130 calculates the center of gravity sway index by substituting the calculated sway amplitude W and number of swayings into a predetermined formula, for example. In this embodiment, the center of gravity sway index becomes higher as the sway amplitude W increases. Furthermore, the center of gravity sway index becomes higher as the number of swayings increases. The center of gravity sway index is calculated once for each standing-up movement of the subject P. As shown in FIG. 3, after performing the process of step S21, the server 130 proceeds to step S22.

[0065] In step S22, the server 130 estimates the behavior of the subject P. Specifically, the server 130 estimates the behavior of the subject P when standing up based on the destination detected in step S12 and the staying time T measured in steps S17 to S20, etc. For example, if the destination is toilet area C and the staying time T is less than a preset threshold, the behavior of the subject P is estimated to be "excretion (small)", and if the staying time T is equal to or greater than the threshold, the behavior of the subject P is estimated to be "excretion (large)". After performing the process of step S22, the server 130 proceeds to step S23.

[0066] In step S23, server 130 plots the center of gravity sway index calculated in step S21 on a graph for each activity. Specifically, server 130 stores graphs showing the center of gravity sway index calculated in step S21 in chronological order for each activity. An example is shown in FIG. 7. FIG. 7 shows a graph with the center of gravity sway index on the vertical axis and time (date and time) on the horizontal axis. FIG. 7(a) is a graph for toileting (e.g., "excretion (large)"). FIG. 7(b) is a graph for "meal." FIG. 7(c) is a graph for "waking up."

[0067] In step S23, server 130 plots the calculation result of the center of gravity sway index in step S21 on the graph corresponding to the behavior estimated in step S22, among the graphs for each behavior. For example, if server 130 estimates in step S22 that the behavior of subject P is "eating," it plots the calculation result of the center of gravity sway index on the graph shown in Figure 7(b). As shown in Figure 3, after performing the process of step S23, server 130 proceeds to step S24.

[0068] In step S24, the server 130 proceeds to a fall risk estimation flow (processing of the estimation phase). As will be described later, in the fall risk estimation flow, if there is a risk of the subject P falling, information is notified to a caregiver or the like. After performing the processing of step S24, the server 130 ends the processing in the measurement phase.

[0069] In this embodiment, since the measurement does not end when the subject P passes through a lying position from a sitting position ("NO" in step S18), the server 130 acquires the fluctuation of the center of gravity of the subject P even when the subject P stands up after passing through a lying position. This makes it possible to acquire the fluctuation of the center of gravity when the subject P stands up, for example, even if the subject P sitting on a sofa lies down halfway through.

[0070] Next, specific processing in the learning phase will be described with reference to FIGS.

[0071] The learning phase process is a process (processing) for learning data such as a center of gravity sway index, which quantifies the degree of swaying of the subject P's center of gravity. Part of the learning phase process overlaps with, for example, part of the measurement phase process. That is, in this embodiment, part of the measurement phase process also serves as part of the learning phase process. The learning phase process is started together with the measurement phase process and then executed for a predetermined period, for example, for each "action," so that learning is performed under the assumption that the physical condition does not change. Specifically, in the learning phase process, the calculation results of the center of gravity sway index are plotted as described below, and the learning phase process is executed for a period (for example, from several weeks to one or two months) during which a regression line can be statistically determined and a quantity of plots can be obtained to determine a normal distribution, etc.

[0072] The processing of steps S31 to S33 corresponds to the processing of steps S11 to S23 in the measurement phase. That is, the processing of steps S31 to S33 shown in Fig. 8 is a simplified description of the processing of steps S11 to S23 in the measurement phase (see Fig. 3). After performing the processing of step S33, the server 130 proceeds to step S34.

[0073] In step S34, if the number of body sway indices plotted on the graphs for each behavior is equal to or greater than the specified number for all graphs for each behavior ("YES" in step S34), server 130 proceeds to step S35. On the other hand, if there is at least one graph in which the number of body sway indices is less than the specified number ("NO" in step S34), server 130 proceeds to step S31. Note that it is also possible to perform the learning phase processing for each behavior, and then proceed to step S35 after determining the number of body sway indices plotted for each behavior.

[0074] 8 and 9, in step S35, the server 130 calculates a regression line L0 for the graphs for each behavior. For example, the server 130 calculates the regression line L0 using the least squares method or the like. The server 130 also calculates the regression line L0 for each of the graphs for each behavior (not shown). After performing the process of step S35, the server 130 proceeds to step S36.

[0075] In step S36, server 130 obtains (creates) a normal distribution (mean value, standard deviation, etc.) of the calculation results of the center of gravity sway index for each behavior. Note that hereinafter, the created normal distribution (learning result) may be referred to as "learning data." When the processing of step S36 ends, learning by server 130 is completed (step S37), and server 130 proceeds to step S38.

[0076] In step S38, the server 130 saves the learning data for each behavior. For example, the server 130 saves, as the learning data, data associating the center of gravity sway index calculated in step S31 with the date and time of the standing-up movement, data showing the regression line L0 calculated in step S35, data related to the normal distribution (average value, standard deviation, etc.) calculated in step S36, etc.

[0077] The learning content in the learning phase is not limited to that of the present embodiment and can be changed as appropriate. For example, the server 130 may learn the range R of the center of gravity sway index (see FIG. 9) plotted on the graph for each behavior. The range R may be, for example, the range from the lower limit to the upper limit of the center of gravity sway index in the graph for each behavior. After processing step S38, the server 130 ends the processing of the learning phase.

[0078] Next, specific processing in the estimation phase will be described with reference to FIGS.

[0079] The processing of the estimation phase is a step (processing) of estimating the risk of a fall of the subject P. Part of the processing of the estimation phase overlaps with, for example, part of the processing of the measurement phase. That is, in this embodiment, part of the processing of the measurement phase also serves as part of the processing of the estimation phase. In this embodiment, the fall risk of the subject P estimated by the server 130 is divided into three types, as will be described later. Furthermore, in order to perform estimation for each of the three types of fall risk, three types of flowcharts (hereinafter referred to as a "first estimation flow," a "second estimation flow," and a "third estimation flow") corresponding to each fall risk are used. In this way, the server 130 executes the first estimation flow, the second estimation flow, and the third estimation flow independently in the processing of the estimation phase.

[0080] First, the processing of the first estimation flow will be described with reference to the flowchart shown in FIG. 10 and FIG.

[0081] The process of the first estimation flow shown in FIG. 10 is for estimating a temporary increase in the risk of falling, for example, when the subject P becomes unsteady when standing up due to a sudden illness.

[0082] The processing of steps S41 to S43 corresponds to the processing of steps S11 to S23 in the measurement phase. That is, the server 130 acquires the sway of the center of gravity of the subject P, calculates a center of gravity sway index, estimates the behavior of the subject P, and plots the results on graphs for each behavior. Note that FIG. 11(a) shows an example in which the calculation result of the center of gravity sway index is plotted at time TM1 on the graph for toilet ("excretion (large)") by the processing of step S43. Time TM1 is also the date and time when the subject P most recently stood up. As shown in FIG. 10, after performing the processing of step S43, the server 130 proceeds to step S44.

[0083] In step S44, server 130 acquires learning data (see FIG. 7) of the same behavior as the behavior estimated in step S42. At this time, server 130 acquires learning data (regression line L0, etc.) that matches the behavior estimated in step S42 from the learning data saved in step S38 (see FIG. 8) of the learning phase processing. After performing the processing of step S44, server 130 proceeds to step S45.

[0084] In step S45, server 130 determines whether the calculation result of the center of gravity sway index in step S41 (the most recent center of gravity behavior index) is out of sync with the learning data acquired in step S44. An example of the determination process will be described below.

[0085] The server 130 compares the body sway index of the regression line L0 at time TM1 (the date and time when the subject P most recently stood up) shown in Figure 11(b) with the body sway index calculated in step S41. If the calculation result is higher than the body sway index of the regression line L0 by a predetermined amount or more, the server 130 determines that the calculation result deviates from the learning data. On the other hand, if the calculation result is not higher than the body sway index of the regression line L0 by a predetermined amount or more, the server 130 determines that the calculation result does not deviate from the learning data.

[0086] The contents of the determination process in step S45 are not limited to those described above and can be changed as appropriate. For example, the server 130 may determine whether the calculation result of the center of gravity sway index deviates from the learning data using the normal distribution (mean value, standard deviation, etc.) for each behavior obtained in the learning phase. Furthermore, for example, the server 130 may determine that the calculation result deviates from the learning data when the calculation result is outside the range R of the center of gravity sway index plotted on the graph for each behavior (see FIG. 9).

[0087] 10, when the server 130 determines that the center of gravity sway index calculated in step S41 is out of sync with the learning data ("YES" in step S45), the server 130 proceeds to step S46 and issues a notification (alert) about the risk of falling. For example, the server 130 outputs a message indicating that the risk of falling has temporarily increased via the staff terminal 120. After performing the processing of step S46, the server 130 ends the processing of the first estimation flow.

[0088] On the other hand, if the server 130 determines that the center of gravity stability index calculated in step S41 is not out of sync with the learning data ("NO" in step S45), it ends the processing of the first estimation flow without notifying the user of the risk of falling (step S47).

[0089] According to the first estimation flow, if the most recent measured data (center of gravity sway index) deviates significantly from the learned data, information is notified to the caregiver, etc. For example, if the subject P becomes unsteady when standing up due to a sudden illness, information is notified to the caregiver, etc. This configuration allows the caregiver, etc. to understand that the subject P's risk of falling has temporarily increased. Furthermore, by the caregiver, etc. temporarily assisting the subject P, the subject P can be prevented from falling.

[0090] Next, the processing of the second estimation flow will be described with reference to the flowchart shown in FIG. 12 and FIG.

[0091] The processing of the second inference flow shown in Fig. 12 is for inferring a chronic increase in the risk of falling. More specifically, the processing of the second inference flow is for inferring whether a state in which the risk of falling is relatively high continues.

[0092] The processing of steps S51 to S53 corresponds to the processing of steps S11 to S23 in the measurement phase. That is, the server 130 acquires the sway of the center of gravity of the subject P, calculates a center of gravity sway index, estimates the behavior of the subject P, and plots the behavior-specific graphs. After performing the processing of step S53, the server 130 proceeds to step S54.

[0093] In step S54, the server 130 acquires data for a certain period of time in the past (for example, a period going back a predetermined period including the most recent time; in this embodiment, for the most recent month). More specifically, the server 130 acquires the data for the certain period of time in the past from a graph that matches the behavior estimated in step S52, among the graphs for each behavior obtained in the measurement phase (see FIG. 7). For example, if the server 130 estimates in step S52 that the behavior of the subject P is toilet (“excretion (large)”), it acquires the detection results of the center of gravity sway index for the most recent month from the graph shown in FIG. 7(a). In this way, as shown in FIG. 13(a), the server 130 acquires a data group (history) of the center of gravity sway index for the behavior estimated in step S52. As shown in FIG. 12, after performing the process of step S54, the server 130 proceeds to step S55.

[0094] In step S55, if the number of data groups acquired in step S54 (the number of plots in FIG. 13(a)) is equal to or greater than the specified number ("YES" in step S55), the server 130 proceeds to step S56. On the other hand, if the number of data groups is less than the specified number ("NO" in step S55), the server 130 proceeds to step S51.

[0095] 12 and 13(b), in step S56, the server 130 calculates a regression line L1 for the data group acquired in step S54. Hereinafter, this regression line L1 will be referred to as the "regression line L1 of the estimation phase." After performing the process of step S56, the server 130 proceeds to step S57.

[0096] In step S57, the server 130 determines whether or not there is a change in the intercept N0 of the regression line L0 obtained in the learning phase. More specifically, the server 130 determines whether or not the intercept N1 of the regression line L1 in the estimation phase is higher than the intercept N0 in the learning phase by a predetermined amount or more.

[0097] If the intercept N1 is higher than a predetermined value ("YES" in step S57), the server 130 proceeds to step S58 and notifies the user of the risk of falling. For example, the server 130 outputs a message indicating that the user is chronically at high risk of falling via the staff terminal 120.

[0098] On the other hand, if the intercept N1 is not higher than the predetermined value ("NO" in step S57), the server 130 does not notify the user of the risk of falling (step S59) and ends the processing of the second estimation flow.

[0099] According to the second estimation flow, when it is estimated that the subject P has a high chronic risk of falling (for example, when the increased risk of falling is not expected to decrease significantly in the future, such as when the subject P's lower limb muscle strength has decreased with age, making him or her more unsteady when standing up), information about the risk of falling is notified to the caregiver, etc. In this way, the caregiver, etc. can understand, based on the notification, that the subject P is in a state where the risk of falling is chronically high. Furthermore, by the caregiver, etc. providing more thorough assistance to the subject P than initially (before the notification in step S58), the subject P can be effectively prevented from falling.

[0100] The method for determining the risk of falling in the second estimation flow (the processing content of steps S56 and S57) is not limited to that of the present embodiment and can be modified as appropriate. For example, the server 130 may estimate the risk of falling based on the result of comparing the normal distribution (average value, standard deviation, etc.) of the data group acquired in step S54 with the learning data.

[0101] Next, the third estimation flow will be described with reference to the flowchart shown in FIG. 14 and FIG.

[0102] The processing of the third estimation flow shown in FIG. 14 is for estimating whether or not the risk of falling is gradually increasing.

[0103] In steps S61 to S66, the server 130 executes the same processing as in the second estimation flow (see FIG. 12) to calculate the regression line L2. An example of the regression line L2 calculated in step S66 is shown in FIG. 15. As shown in FIG. 14, after performing the processing of step S66, the server 130 proceeds to step S67.

[0104] In step S67, server 130 determines whether there is a change (abnormality) in the slope of regression line L2 calculated in step S66. Through this determination process, server 130 determines whether the center of gravity stabilization index is on the rise.

[0105] For example, if the regression line L2 in FIG. 15 slopes so that the center of gravity sway index increases as the current time (time TM2) approaches, the server 130 determines that there is a change in the slope of the regression line L2 ("YES" in step S45) and proceeds to step S68. In this case, the server 130 notifies the user of the risk of falling. For example, the server 130 outputs a message via the staff terminal 120 indicating that the risk of falling is increasing.

[0106] On the other hand, when the server 130 determines that there is no change in the slope of the regression line L1 ("NO" in step S45), it ends the processing of the third estimation flow without notifying the user of the risk of falling (step S69).

[0107] According to the third estimation flow, it is possible to evaluate whether the center of gravity sway index is on an increasing trend from the history of actual measurement data that matches the behavior estimated in step S62. As a result, for example, if the subject P's lower limb muscle strength decreases over time or the symptoms of orthostatic hypotension progress and the subject P becomes increasingly unsteady when standing up, information about the risk of falling is notified to the caregiver, etc. Based on this notification, the caregiver, etc. can understand that the subject P's risk of falling is on an increasing trend. Furthermore, by the caregiver, etc. providing more thorough assistance to the subject P than initially (before the notification in step S68), the subject P can be effectively prevented from falling.

[0108] In the estimation phase of this embodiment, the sway of the center of gravity of the subject P (center of gravity sway index) is evaluated for each behavior, and the risk of the subject P falling is estimated. This allows the server 130 to accurately estimate the risk of falling when standing up. More specifically, the subject P is relatively prone to staggering when standing up with elevated blood pressure. Furthermore, the increase in blood pressure is correlated with the behavior of the subject P. For example, behaviors such as "getting up," "eating," and "excretion (large)" are relatively likely to increase the subject P's blood pressure. On the other hand, behaviors such as "excretion (small)" are relatively unlikely to increase the subject P's blood pressure. For this reason, the center of gravity sway index calculated when "getting up" or the like is likely to be larger than the center of gravity sway index calculated when "excretion (small)" or the like is performed.

[0109] The server 130 of this embodiment evaluates the sway of the center of gravity of the subject P for each action and does not compare the degree of sway of the center of gravity when performing different actions, so it can accurately evaluate the sway of the center of gravity. Therefore, the server 130 can accurately estimate the risk of falling.

[0110] Furthermore, the likelihood of subject P unsteady when standing up may vary depending on the area. For example, in an area where chairs with relatively low seats are installed, subject P is likely to unsteady when standing up. Subject P is also likely to unsteady when standing up while carrying luggage in his / her hands. On the other hand, in an area where chairs with relatively high seats are installed or where assistive devices such as handrails are installed, subject P is less likely to unsteady when standing up.

[0111] In this embodiment, behavior is estimated depending on the area (see Figure 2), so the influence of the tendency to unsteady depending on the area (for example, different conditions such as chair height that are expected to affect the swaying of the subject P's center of gravity when standing up) can be eliminated, making it possible to accurately evaluate the swaying of the center of gravity.

[0112] Furthermore, by regarding the position of the subject P's torso as the position of the center of gravity, it is possible to both calculate the center of gravity sway index and estimate the subject P's behavior using a single type of sensor (millimeter wave sensor) that detects the position and posture of the subject P, thereby simplifying the configuration.

[0113] As described above, the health risk estimation system 100 according to this embodiment: A behavior estimation unit (subject detection sensor 110, server 130) capable of estimating the behavior of the subject P; a body sway estimation unit (server 130) capable of estimating the degree of body sway when the subject P changes posture; a health risk estimation unit (server 130) capable of estimating a health risk of the subject P based on an estimation result of the degree of body swaying during the behavior estimated by the behavior estimation unit; It is equipped with the following.

[0114] With this configuration, the health risk (risk of falling) of the subject P can be estimated with high accuracy. Furthermore, in this embodiment, attention is focused on the degree of body sway during the standing-up motion (i.e., the degree of body sway when the subject P's body generally becomes particularly unsteady), so it is possible to easily grasp an increasing tendency of the risk of falling, which is difficult to notice in normal daily life. In this way, it is possible to easily grasp the progression of health risks that are usually difficult to notice, making it easier to effectively maintain the health of the subject P.

[0115] In addition, in the health risk estimation system 100 according to this embodiment, The degree of body swaying includes the degree of center of gravity swaying.

[0116] With this configuration, the health risk (risk of falling) of the subject P can be estimated with high accuracy.

[0117] In addition, in the health risk estimation system 100 according to this embodiment, The motion estimation unit (server 130) The method estimates the degree of swaying of the subject P's body when the subject P changes his / her posture from a sitting position to a standing position.

[0118] With this configuration, the health risk of the subject P can be estimated based on the standing-up movement of the subject P.

[0119] In addition, in the health risk estimation system 100 according to this embodiment, The motion estimation unit (server 130) The degree of swaying of the trunk of the subject P is regarded as the degree of swaying of the center of gravity.

[0120] With this configuration, the center of gravity can be set in a relatively simple manner, and the health risk of the subject P can be estimated based on the standing-up movement of the subject P.

[0121] In addition, in the health risk estimation system 100 according to this embodiment, The motion estimation unit (server 130) Acquire a movement trajectory of the trunk when the subject P changes posture; The degree of shaking of the body is estimated based on at least one of the number of times the trunk sways or the amplitude of the trunk sway, which are obtained from the movement trajectory M.

[0122] This configuration allows for a quantitative assessment of health risks based on the sway amplitude W and the number of times the trunk sways. For example, if the sway amplitude W during a standing-up movement is large or if the number of times the trunk sways is large, it is possible to detect a risk of falling.

[0123] In addition, in the health risk estimation system 100 according to this embodiment, The health risk estimation unit (server 130) The system learns in advance data that quantifies the degree of body shaking, and estimates the health risk based on the results of the learning.

[0124] By configuring in this way, it is possible to estimate health risks based on the results of prior learning.

[0125] In addition, the quantified data is It is created based on information obtained from the subject P.

[0126] By configuring in this way, health risks can be estimated with high accuracy. Specifically, the degree of center of gravity sway can be appropriately learned according to the individual characteristics of the subject P (age, sex, habits, illnesses, etc.), so the risk of falling can be estimated with high accuracy.

[0127] The subject detection sensor 110 according to this embodiment is one embodiment of a subject detection unit according to the present invention. The server 130 according to this embodiment is an embodiment of the movement detection unit, the behavior estimation unit, and the health risk estimation unit according to the present invention.

[0128] <Note> Although the embodiment of the present invention has been described above, the present invention is not limited to the above configuration and various modifications are possible within the scope of the invention described in the claims. For example, the processes executed by the health risk estimation system 100 are not limited to those described above and can be modified as desired.

[0129] In addition, in this embodiment, an example has been shown in which the health risk estimation system 100 is applied to a room 1 of a facility for the elderly, but the present invention is not limited to this and can be applied to various other facilities, homes, etc.

[0130] Furthermore, in this embodiment, the subject P is detected using a millimeter wave sensor, but the present invention is not limited to this. It is also possible to detect the subject P using various devices, such as a camera or other sensors capable of detecting the subject P. Furthermore, in order to estimate the degree of shaking of the subject P's body, it is also possible to estimate the degree of shaking of a part other than the center of gravity of the body (for example, the head, etc.). Furthermore, in this embodiment, the position of the trunk is considered to be the position of the center of gravity of the body, but this is not limited to this. In other words, other parts of the body may also be considered to be the position of the center of gravity. Furthermore, the actual position of the center of gravity of the body may be identified using an appropriate sensor, calculation means, etc., and this may be used.

[0131] Furthermore, although the server 130 of this embodiment is configured to estimate the risk of falling when changing from a sitting position to a standing position, it is also possible to estimate the risk of falling when changing positions different from those of this embodiment. For example, the server 130 may estimate the risk of falling when changing from a lying position to a standing position. Furthermore, in this embodiment, the degree of body swaying of the subject P when standing up is estimated, but this is not limiting. In other words, if the subject P remains in a standing position for a predetermined period after standing up, the risk of falling can also be determined based on the degree of body swaying during this predetermined period.

[0132] Furthermore, although the server 130 is configured to calculate the center of gravity sway index according to both the number of times the torso sways and the sway amplitude W, the information used to calculate the center of gravity sway index is not limited to that in this embodiment as long as it is information that allows the degree of swaying of the center of gravity to be evaluated. For example, the server 130 may calculate the center of gravity sway index according to either the number of times the torso sways or the sway amplitude W. The server 130 may also calculate the center of gravity sway index according to information other than the number of times the torso sways and the sway amplitude W.

[0133] In addition, in this embodiment, an example is shown in which various processes are executed by a virtual server 130 constructed on a cloud server, but the present invention is not limited to this, and various processes can also be executed using a server constructed in an on-premises environment. [Explanation of symbols]

[0134] 100 Health Risk Estimation System 110 Target detection sensor 130 servers P Target Audience

Claims

1. a behavior estimation unit capable of estimating the behavior of a subject; a motion estimation unit capable of estimating a degree of motion of the body when the subject changes posture; a health risk estimation unit capable of estimating a health risk of the subject based on an estimation result of the degree of body swaying during the behavior estimated by the behavior estimation unit; A health risk estimation system comprising:

2. The degree of body swaying includes the degree of center of gravity swaying. The health risk estimation system according to claim 1 .

3. The motion estimation unit Estimating the degree of swaying of the body when the posture of the subject changes from a sitting posture to a standing posture. The health risk estimation system according to claim 1 .

4. The motion estimation unit The degree of swaying of the subject's trunk is regarded as the degree of swaying of the center of gravity. The health risk estimation system according to claim 2 .

5. The motion estimation unit acquiring a movement trajectory of the trunk when the subject changes posture; estimating a degree of swaying of the body based on at least one of the number of times the trunk swayed and the amplitude of the trunk swaying, which are obtained from the movement trajectory; The health risk estimation system according to claim 4.

6. The health risk estimation unit Numerical data of the degree of body shaking is learned in advance, and the health risk is estimated based on the learned results. The health risk estimation system according to claim 1 .

7. The quantified data is It is created based on information obtained from the subject. The health risk estimation system according to claim 6.

Images