Estimation device, estimation method, estimation program, and learning model generation device

Abstract

To estimate a posture state accompanied by a person becoming moist using electric characteristics of bedding including a flexible material having conductivity without using a special detection device.SOLUTION: A posture state estimation device (1) detects, in a detection unit, electric characteristics between a plurality of detection points in bedding (2) including a flexible material having conductivity whose electric characteristic change according to a change in an applied pressure and moisture. An estimation unit (5) estimates a posture state accompanied by perspiration of a person from the electric characteristics of the bedding (2) using a learning model (51). Using time-series electric characteristics when applying a pressure and moisture to the bedding (2) and posture state information indicating the posture state accompanied by the perspiration of the person applying the pressure and moisture to the bedding (2) as learning data, the learning model (51) inputs the electric characteristics to the learning model that outputs the posture state information with the electric characteristics as an input, and learning is executed so as to output the posture state information indicating the posture state accompanied by the perspiration of the person corresponding to the input electric characteristics.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to an estimation device, an estimation method, an estimation program, and a learning model generation device.

Background Art

[0002] Conventionally, in order to identify the posture of a person on a bedding, a shape change occurring in the bedding is detected, and the posture of the person is estimated using the detection result. In terms of detecting the shape change occurring in the bedding, it is difficult to detect the deformation without inhibiting the deformation of the bedding. Further, since a strain sensor used for detecting a rigid body such as metal deformation is difficult to use for a bedding, a special detection device is required to detect the deformation of the bedding. For example, a technique of measuring the displacement and vibration of an object by a camera, acquiring a deformed image, and extracting a deformation amount is known (see, for example, Patent Document 1). Further, a technique related to a flexible tactile sensor for estimating a deformation amount from a light transmission amount is also known (see, for example, Patent Document 2).

[0003] Further, in terms of estimating a sleep state in which a person sleeps in various postures during sleep, a technique of detecting vibrations caused by the heartbeat and breathing of a bedridden person with a sensor such as a piezoelectric element and estimating the sleep state of the bedridden person based on the detected vibrations is known (see, for example, Patent Document 3).

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Patent Document 3

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in the above prior art, it is impossible to estimate the posture state accompanied by water content due to sweating or the like of a person on the bedding. When estimating the posture state accompanied by water content due to sweating or the like of a person on the bedding, it is necessary to independently detect the shape change and the moisture, and a special detection device that associates the two is required.

[0006] An object of the present disclosure is to provide an estimation device, an estimation method, an estimation program, and a learning model generation device that can estimate posture state information indicating a posture state accompanied by water content of a person by using the electrical characteristics of a bedding provided with a flexible material having conductivity without using a special detection device.

Means for Solving the Problems

[0007] To achieve the above object, a first aspect is a detection unit that detects the electrical characteristics between a plurality of predetermined detection points on the flexible material of a bedding provided with a flexible material having conductivity and whose electrical characteristics change in response to changes in the applied pressure and moisture; using the time-series electrical characteristics when pressure and moisture are applied to the flexible material and the posture state information indicating the posture state accompanied by water content of the person applying pressure and moisture to the flexible material as learning data, inputting the time-series electrical characteristics, and inputting the time-series electrical characteristics detected by the detection unit to a learning model trained to output the posture state information, and estimating the posture state information indicating the posture state accompanied by water content of the person corresponding to the input time-series electrical characteristics. An estimation unit; An estimation device including

[0008] A second aspect is the estimation device according to the first aspect, The posture state accompanied by water content of the person includes a posture state accompanied by sweating in the sleep state of the person.

[0009] A third aspect is the estimation device according to the first aspect or the second aspect, The electrical characteristic is volume resistivity, The bedding includes a mattress, The flexible material is a material in which conductivity is imparted to at least a part of a urethane material having a structure with at least one of a fibrous and a mesh-like skeleton, or a structure in which a plurality of minute air bubbles are scattered inside.

[0010] A fourth aspect is, in the estimation device of the third aspect, the urethane material is impregnated with a conductive material.

[0011] A fifth aspect is, in the estimation device of any one of the first aspect to the fourth aspect, the learning model includes a model generated by learning using a network by reservoir computing using the flexible material as a reservoir.

[0012] A sixth aspect is a computer acquires the electrical characteristics from a detection unit that detects the electrical characteristics between a plurality of predetermined detection points on the flexible material of a bedding including a flexible material having conductivity and whose electrical characteristics change in response to changes in applied pressure and moisture, using, as learning data, the time-series electrical characteristics when pressure and moisture are applied to the flexible material and the posture state information indicating the posture state of a person with moisture content who applies pressure and moisture to the flexible material, inputs the time-series electrical characteristics, and estimates the posture state information indicating the posture state of the person with moisture content corresponding to the input time-series electrical characteristics with respect to a learning model learned to output the posture state information with the time-series electrical characteristics as an input estimation method.

[0013] A seventh aspect is to a computer acquires the electrical characteristics from a detection unit that detects the electrical characteristics between a plurality of predetermined detection points on the flexible material of a bedding including a flexible material having conductivity and whose electrical characteristics change in response to changes in applied pressure and moisture, Using, as learning data, the time-series electrical characteristics when pressure and moisture are applied to the flexible material and the posture state information indicating the posture state of a person with moisture content who applies pressure and moisture to the flexible material, for a learning model that is trained to take the time-series electrical characteristics as input and output the posture state information, the time-series electrical characteristics detected by the detection unit are input, and the posture state information indicating the posture state of a person with moisture content corresponding to the input time-series electrical characteristics is estimated It is an estimation program for executing the process.

[0014] The eighth aspect is the electrical characteristics from a detection unit that detects the electrical characteristics between a plurality of predetermined detection points on the flexible material of a bedding including a flexible material having conductivity and whose electrical characteristics change in response to changes in the applied pressure and moisture, and the posture state information indicating the posture state of a person with moisture content who applies pressure and moisture to the flexible material, an acquisition unit that acquires a learning model generation unit that generates a learning model that inputs the time-series electrical characteristics when pressure and moisture are applied to the flexible material and outputs the posture state information indicating the posture state of a person with moisture content who applies pressure and moisture to the flexible material, based on the acquisition result of the acquisition unit It is a learning model generation device including.

Advantages of the Invention

[0015] According to the present disclosure, without using a special detection device, it is possible to estimate the posture state information indicating the posture state of a person with moisture content by utilizing the electrical characteristics of a bedding including a flexible material having conductivity, which has the effect of

Brief Description of the Drawings

[0016]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Mode for Carrying Out the Invention

[0017] Hereinafter, embodiments for realizing the technology of the present disclosure will be described in detail with reference to the drawings. Note that components and processes having the same functions may be given the same reference numerals throughout the drawings, and redundant descriptions may be omitted as appropriate. Further, the present disclosure is not limited to the following embodiments, and can be implemented with appropriate modifications within the scope of the object of the present disclosure. In the present disclosure, mainly the estimation of physical quantities for members that deform non-linearly is described, but it goes without saying that it is applicable to the estimation of physical quantities for members that deform linearly.

[0018] In the present disclosure, the term "bedding" refers to a concept representing an article used when a person sleeps, which includes a flexible material that can be deformed, such as being bent at least in part when an external force is applied. An example of bedding is a mattress. An example of an external force is pressure as a stimulus applied to the bedding. The term "posture state of a person accompanied by water content" refers to a concept including the state of a person who applies pressure to a flexible material and also provides moisture such as sweating. Examples of water content of a person include any moisture such as sweating, vomiting, hemoptysis, and excretion. The posture state includes, for example, a state where a person is lying on the back on the bedding, a state where a person is lying face down on the bedding, a state where a person is lying on the side on the bedding, and a state where a person is sitting on the bedding. Also, each of the posture states includes the sleep state of a sleeper or the like at bedtime.

[0019] In the present disclosure, the term "flexible material" refers to a concept including a material that can be deformed, such as being bent at least in part when an external force is applied, and includes a soft elastic body such as a rubber material, a structure having at least one of a fibrous and a net-like skeleton, and a structure in which a plurality of minute air bubbles are scattered inside. An example of an external force is pressure. Examples of a structure having at least one of a fibrous and a net-like skeleton and a structure in which a plurality of minute air bubbles are scattered inside include polymer materials such as urethane materials. The term "flexible material provided with conductivity" refers to a concept including a material having conductivity, and includes a material in which a conductive material is provided to a flexible material to impart conductivity and a material in which the flexible material has conductivity. Also, a flexible material provided with conductivity has a function in which its electrical characteristics change according to deformation. Note that an example of a physical quantity that causes a function in which electrical characteristics change according to deformation is a pressure value indicating a stimulus by pressure (hereinafter referred to as pressure stimulus) applied to the flexible material. The flexible material deforms according to an external force generated in the posture state of a person, for example, the distribution of the pressure stimulus. Also, an example of a physical quantity representing electrical characteristics that change according to deformation is an electrical resistance value. Also, other examples include a voltage value or a current value. The electrical resistance value can be regarded as the volume resistance value of the flexible material.

[0020] By imparting conductivity, the soft material exhibits electrical characteristics corresponding to deformation caused by pressure. That is, in the soft material with imparted conductivity, the electrical paths are intricately coordinated, and the electrical paths expand and contract or expand and shrink in response to deformation. Also, there may be cases where the electrical paths are temporarily cut off and where connections different from before occur. Therefore, the soft material exhibits behaviors with different electrical characteristics between positions separated by a predetermined distance (for example, the positions of detection points where electrodes are arranged) according to the magnitude and distribution of the applied force (for example, pressure stimulus). For this reason, the electrical characteristics change according to the magnitude and distribution of the force (for example, pressure stimulus) applied to the soft material. Since a soft material with imparted conductivity is used, it is not necessary to provide detection points such as electrodes at all locations where pressure is applied to the soft material by an object such as a human. It is sufficient if detection points such as electrodes are provided at any at least two locations sandwiching the location where pressure is applied to the soft material.

[0021] In addition, the electrical characteristics that change according to the deformation of the soft material are affected by moisture (water content). The physical quantity (electrical resistance value) representing the electrical characteristics corresponding to the posture state accompanied by water content of a person changes from the physical quantity (electrical resistance value) representing the electrical characteristics corresponding to the posture state without water content of the person. That is, since the soft material has different electrical characteristics according to the water content even when the person is in the same posture state, it is possible to discriminate the presence or absence of water content of the person.

[0022] The estimation device of the present disclosure estimates the posture state accompanied by water content of a person from the electrical characteristics of a flexible material having conductivity provided in a bedding using a learned learning model. The posture state accompanied by water content includes, for example, a posture state accompanied by sweating, a posture state accompanied by water content due to at least one of a person's body and clothes being wet, and the like. The flexible material can be arranged on the bedding. The learning model uses, as learning data, the time-series electrical characteristics when pressure and moisture are applied to the flexible material having conductivity and the posture state information indicating the posture state accompanied by water content of the person applying the pressure and moisture to the flexible material. The learning model is learned to take the time-series electrical characteristics as input and output the posture state information indicating the posture state accompanied by water content of the person corresponding to the time-series electrical characteristics. Regarding the influence of moisture, since the electrical characteristics change even without deformation due to pressure, the water content state of the bedding may be estimated from the electrical characteristics in a state where there is no pressure applied by a person.

[0023] In the following description, as an example of the bedding, a case where a mattress in which a sheet member (hereinafter referred to as conductive urethane) in which at least a part of a urethane member is impregnated (also referred to as infiltrated) with a conductive material is arranged is applied will be described. Further, as the physical quantity for deforming the conductive urethane, a value indicating a pressure stimulus containing moisture (hereinafter referred to as a moisture pressure value) applied to the bedding to which the mattress is applied is applied. The moisture pressure value in this case is generated by the posture state accompanied by water content of a person sleeping on the mattress. Note that the posture state accompanied by water content of a person includes a posture state accompanied by sweating in the person's sleep state. Further, as an example of the moisture, moisture due to a person's sweating is taken, and a case where the electrical resistance value of the conductive urethane is applied as the physical quantity that changes according to the pressure stimulus containing moisture will be described.

[0024] FIG. 1 shows an example of the configuration of a posture state estimation device 1 as the estimation device of the present disclosure.

[0025] As shown in FIG. 1, in the posture state estimation device 1, the estimation process uses the learned learning model 51 to estimate, as the posture state accompanied by sweating of an unknown person, the posture state accompanied by sweating of the person OP on the bedding 2, and outputs it as posture state information. Thereby, it becomes possible to identify the posture state accompanied by sweating of the person sleeping on the bedding 2 without using a special device or a large device, or directly measuring the deformation and water content of a flexible material such as a mattress included in the bedding. The learning model 51 is learned with the posture state (for example, the posture state value accompanied by sweating) of the person accompanied by sweating in the bedding 2 as a label and the electrical characteristics of the bedding 2 (that is, the electrical resistance value of the conductive urethane arranged on the bedding 2) in the posture state accompanied by sweating as an input. The learning of the learning model 51 will be described later.

[0026] As shown in FIG. 2, the bedding 2 according to the present embodiment is configured by arranging conductive urethane 22 on a mattress 21. The bedding 2 constituted by the mattress 21 on which the conductive urethane 22 is arranged is connected to an electrical characteristic detection unit 76 which is an example of a detection unit. The conductive urethane 22 is assumed to be formed by either blending or impregnating a conductive material, but impregnation is more desirable because it has higher conductivity than blending. As shown in FIG. 2, the conductive urethane 22 may be arranged on at least a part of the mattress 21, and may be arranged inside or outside. Specifically, as shown by taking the A-A cross section of the bedding as the bedding cross section 2-1, the entire inside of the mattress 21 may be constituted by the conductive urethane 22. Also, as shown in the bedding cross section 2-2, the conductive urethane 22 may be formed on the person side (surface side) inside the mattress 21, and as shown in the bedding cross section 2-3, the conductive urethane 22 may be formed on the side opposite to the person side (back side) inside the mattress 21. Further, as shown in the bedding cross section 2-4, the conductive urethane 22 may be formed near the center inside the mattress 21.

[0027] Also, as shown in the bedding cross-section 2-5, the conductive urethane 22 may be disposed outside the person side (front side) of the mattress 21, or as shown in the bedding cross-section 2-6, the conductive urethane 22 may be disposed outside the side opposite to the person side (back side). When the conductive urethane 22 is disposed outside the mattress 21, the conductive urethane 22 and the mattress 21 may be merely laminated, or the conductive urethane 22 and the mattress 21 may be integrated by adhesion or the like. Note that even when the conductive urethane 22 is disposed outside the mattress 21, since the conductive urethane 22 is a urethane member having conductivity, the flexibility of the mattress 21 is not impaired.

[0028] Hereinafter, for simplicity of explanation, an example of forming a bedding by disposing the conductive urethane 22 outside the person side (front side) of the mattress 21 will be described (bedding cross-section 2-5).

[0029] In the present embodiment, as shown in FIG. 3, it is possible to detect the electrical characteristics (that is, the volume resistance value which is the electrical resistance value) of the conductive urethane 22 by signals from a plurality (two in FIG. 3) of detection points 75 arranged at a distance. In the example shown in FIG. 3, a first detection set #1 for detecting the electrical resistance value by signals from a plurality of detection points 75 arranged at diagonal positions at a distance on the conductive urethane 22 is shown. Note that the arrangement of the plurality of detection points 75 is not limited to the positions shown in FIG. 3, and any position may be used as long as it is a position where the electrical characteristics of the conductive urethane 22 can be detected. Note that for the electrical characteristics of the conductive urethane 22, an electrical characteristic detection unit 76 for detecting the electrical characteristics (that is, the volume resistance value which is the electrical resistance value) may be connected to the detection point 75, and the output thereof may be used.

[0030] In the bedding 2 composed of the mattress 2 provided with the above-described conductive urethane 22, the electrical resistance value detected changes at least before and after a pressure stimulus containing moisture is applied to the bedding 2, depending on the deformation and water content of the conductive urethane 22. Therefore, the electrical resistance value changes before and after a posture state accompanied by sweating of a person, which is accompanied by a pressure stimulus containing moisture to the bedding 2. Thus, by detecting the electrical resistance value in time series, that is, detecting the change in the electrical resistance value from a state where no pressure stimulus containing moisture is applied to the bedding 2 (for example, detecting an electrical resistance value exceeding a predetermined threshold value), it becomes possible to detect a posture state accompanied by sweating of a person with respect to the bedding 2. Specifically, since the posture state accompanied by sweating of a person with respect to the bedding 2 is accompanied by a pressure stimulus containing moisture due to the contact of the person with the bedding 2, it includes the contact state. Therefore, by disposing the conductive urethane 22 on the bedding 2, it becomes possible to detect the contact of a person with respect to the bedding 2. Further, the electrical resistance value changes even if any one of the position, distribution, and magnitude of the pressure stimulus containing moisture applied to the bedding 2 changes. Therefore, it is also possible to detect a contact state accompanied by sweating including the contact position of a person with respect to the bedding 2 from the electrical resistance value that changes in time series.

[0031] Note that, in order to improve the detection accuracy of the electrical characteristics of the conductive urethane 22, more detection points than the detection points (two) shown in FIG. 3 may be used.

[0032] As an example, a conductive urethane 22 may be formed by arranging one or a plurality of rows of columns each composed of a plurality of conductive urethane pieces with detection points arranged thereon, and the electrical characteristics may be detected for each of the plurality of conductive urethane pieces. For example, the conductive urethane pieces 23 shown in FIG. 4 may be arranged to form the conductive urethane 22 (FIGS. 5 and 6). The example shown in FIG. 4 shows a first detection set #1 that detects the electrical resistance value based on the signal from the detection point 75A arranged at a diagonal position with a distance therebetween, and a second detection set #2 that detects the electrical resistance value based on the signal from the detection point 75B arranged at the other diagonal position. Also, in the example shown in FIG. 5, the conductive urethane pieces 23 (FIG. 4) are arranged (4×1) in the longitudinal direction of the mattress 21 (not shown) that constitutes the bedding 2 to form the conductive urethane 22, and it is shown that the first detection set #1 to the eighth detection set #8 are formed in order. Further, in the example shown in FIG. 6, the first detection set #1 is adopted for each of the conductive urethane pieces 23 (FIG. 4), and they are arranged (4×2) in the longitudinal direction and the width direction of the mattress 21 (not shown) that constitutes the bedding 2 to form the conductive urethane 22, and it is shown that the first detection set #1 to the eighth detection set #8 are formed.

[0033] Also, as another example, the detection range on the conductive urethane 22 may be divided, detection points may be provided for each divided detection range, and the electrical characteristics may be detected for each detection range. For example, a region corresponding to the size of the conductive urethane piece 23 shown in FIGS. 5 and 6 may be set as the detection range for the conductive urethane 22, detection points may be arranged for each set detection range, and the electrical characteristics may be detected for each detection range.

[0034] As shown in FIG. 1, the posture state estimation device 1 includes an estimation unit 5. Time-series input data 4 representing the magnitude of the electrical resistance (electrical resistance value) in the conductive urethane 22 is input to the estimation unit 5. The input data 4 corresponds to the posture state 3 of a person with sweating on the bedding 2 indicating the posture state of the person with sweating. Also, the estimation unit 5 outputs output data 6 representing a physical quantity (posture state value with sweating) indicating the posture state of the person with sweating on the bedding 2 as an estimation result. Note that the estimation unit 5 includes a learned learning model 51.

[0035] The learning model 51 is a model that has completed learning to derive posture state information (output data 6) indicating the posture state of a person with sweating on a bedding from the electrical resistance (input data 4) of the conductive urethane 22 that changes due to a pressure stimulus containing moisture according to the posture state 3 of the person with sweating. The learning model 51 is, for example, a model that defines a learned neural network and is expressed as a set of information on the weights (intensities) of the connections between the nodes (neurons) that make up the neural network.

[0036] The learning model 51 is generated by the learning process of the learning processing unit 52 shown in FIG. 7. The learning processing unit 52 performs a learning process using the electrical characteristics (input data 4) in the conductive urethane 22 that change due to a pressure stimulus containing moisture, which is caused by the posture state 3 of the person with sweating. That is, a large amount of data obtained by measuring the electrical resistance in the conductive urethane 22 over time with the posture state 3 of the person with sweating as a label is used as learning data. Specifically, the learning data contains a large number of sets of input data including the electrical resistance value (input data 4) and information (output data 6) indicating the posture state 3 of the person with sweating corresponding to the input data. Here, time-series information is associated by attaching information indicating the measurement time to each of the electrical resistance values (input data 4) of the conductive urethane 22. In this case, for the period determined as the posture state 3 of the person with sweating, time-series information may be associated by attaching information indicating the measurement time to the set of time-series electrical resistance values in the conductive urethane 22.

[0037] Next, the learning processing unit 52 will be described. In the learning process performed by the learning processing unit 52, the bedding 2 composed of the mattress 21 on which the conductive urethane 22 described above is disposed is applied as a detection unit, and the posture state 3 accompanied by sweating of a person and the electrical resistance value (input data 4) by the conductive urethane 22 are used as learning data. For example, the person OP is instructed to assume a posture state accompanied by predetermined sweating on the bedding 2, and the electrical resistance value at that time is detected and used as learning data in association with the posture state accompanied by sweating. Note that at the time of learning, various sweating patterns may be reproduced by spraying moisture or the like. Note that, for the posture state, for example, a state in which the person OP is lying on his / her back, a state in which the person OP is lying on his / her side, etc. are applicable. Also, the electrical characteristics (that is, the volume resistance value which is the electrical resistance value) may be detected by connecting the electrical characteristic detection unit 76 (FIG. 3) to the detection point 75.

[0038] Specifically, the learning processing unit 52 can be configured to include a computer including a CPU (not shown), and executes a learning data collection process and a learning process. FIG. 8 shows an example of the learning data collection process executed by a CPU (not shown). In step S100, the learning processing unit 52 instructs the person OP to assume a posture state accompanied by sweating in the bedding 2 (conductive urethane 22). In step S102, the electrical resistance value that changes due to a pressure stimulus including moisture corresponding to the posture state accompanied by sweating is acquired in time series. In the next step S104, the acquired time-series electrical resistance value is labeled with the posture state 3 accompanied by sweating and stored. The learning processing unit 52 repeats the above processing until a predetermined number of sets of the posture state 3 accompanied by sweating of the person and the electrical resistance value of the conductive urethane 22 are reached, or until a predetermined time is reached (negatively judged until a positive judgment is made in step S106). Thereby, the learning processing unit 52 can acquire and store the electrical resistance value of the conductive urethane 22 in time series for each posture state accompanied by sweating of the person, and the set of the time-series electrical resistance values of the conductive urethane 22 for each posture state accompanied by sweating of the person stored becomes the learning data.

[0039] Incidentally, the postural state accompanied by sweating of the person OP can be identified by at least a part of the relative positional relationship of each part of the person OP with respect to the bedding 2, and the changes and maintenance of physical quantities such as the distribution, magnitude, and frequency of the pressure stimulus containing moisture by each part. Therefore, it is considered that some of these time-series physical quantities include features indicating the postural state accompanied by sweating of the person OP. In the present embodiment, by using the conductive urethane 22, it is possible to detect the electrical characteristics (volume resistance) in which these physical quantities are reflected in time series.

[0040] Fig. 9 shows an example of the electrical characteristics in the bedding 2 composed of the mattress 21 on which the conductive urethane 22 is arranged. The example in Fig. 9 shows how the electrical characteristics (resistance value) change in time series when water is sprayed on the conductive urethane 22.

[0041] Fig. 10 shows another example of the electrical characteristics in the bedding 2 composed of the mattress 21 on which the conductive urethane 22 is arranged. Similar to the example in Fig. 9, the example in Fig. 10 shows how the electrical characteristics (resistance value) change in time series when water is sprayed on the conductive urethane 22. 10(B) is an enlarged view of the X part of Fig. 10(A), and Fig. 10(C) is an enlarged view of the Y part of Fig. 10(A).

[0042] As shown in Figs. 9 and 10, since the electrical characteristics by the conductive urethane 22 change according to the water content of the conductive urethane 22 and the water content is reflected, it can be confirmed that it is possible to estimate the postural state accompanied by sweating of a person from the electrical resistance value that changes in time series according to the water content of the conductive urethane 22. That is, even for various postural states accompanied by sweating, by using the learning model 51 learned for each, the results regarding the postural state accompanied by sweating of a person can be separated, and it becomes possible to discriminate the postural state accompanied by sweating of a person.

[0043] Note that the above-described posture state accompanied by sweating includes the posture state accompanied by sweating during the sleep state of a person such as a sleeper. The sleep state includes, for example, the so-called REM sleep state and non-REM sleep state, the light sleep state and the deep sleep state. These states are accompanied by at least one of a rest state with a stable posture state at a predetermined time and an active state with a plurality of postures or posture states. Therefore, similar to the above-described posture state accompanied by sweating of the person OP, the sleep state can be identified by at least a part of the changes and maintenance of physical quantities such as the relative positional relationship of each part of the person OP with respect to the bedding 2, the distribution, magnitude, and frequency of the pressure stimulus by each part. Therefore, it is considered that a part of the time-series physical quantities includes the characteristics indicating the posture state accompanied by sweating during the sleep state, and by using the conductive urethane 22, it is possible to detect the electrical characteristics (volume resistance) in which these physical quantities in the sleep state are reflected in time series. Note that it is possible to discriminate whether a person is in a sleep state or an awake state by using a physical quantity that characteristically appears in either the sleep state or the awake state. An example of a physical quantity that characteristically appears in the sleep state includes a physical quantity indicating that the posture change is below a threshold value for a time exceeding a predetermined time. Further, by further inputting data indicating whether it is a sleep state or an awake state, it is possible to set whether the detected electrical characteristics correspond to either the sleep state or the awake state.

[0044] Next, an example of the above-described learning data is shown in a table. Table 1 is an example of data in which time-series electrical resistance value data (r) is associated with a posture state value accompanied by sweating as learning data regarding the posture state accompanied by sweating. Table 2 is an example of data in which time-series electrical resistance value data (R) is associated with a sleep state value accompanied by sweating as learning data regarding the sleep state accompanied by sweating.

Table 1

Table 2

[0045] Next, the learning process in the learning processing unit 52 will be described. FIG. 11 is a diagram showing the functions of a CPU (not shown) in the learning processing unit 52 in the learning process. The CPU (not shown) of the learning processing unit 52 includes functional units of a generator 54 and an arithmetic unit 56. The generator 54 has a function of generating an output in consideration of the context of the electrical resistance values acquired in the input time series.

[0046] Also, the learning processing unit 52 holds a large number of sets of the above-described input data 4 (electrical resistance values) and output data 6, which is a posture state 3 accompanied by sweating of a person who has been given a pressure stimulus containing moisture in the conductive urethane 22, as learning data.

[0047] The generator 54 includes an input layer 540, an intermediate layer 542, and an output layer 544 to constitute a known neural network (NN). Since the neural network itself is a known technology, a detailed description thereof will be omitted. The intermediate layer 542 includes a large number of node groups (neuron groups) having inter-node connections and feedback connections. Data from the input layer 540 is input to the intermediate layer 542, and the data of the calculation result of the intermediate layer 542 is output to the output layer 544.

[0048] The generator 54 is a neural network that generates generated output data 6A representing a postural state accompanied by sweating of a person from the input input data 4 (electrical resistance). The generated output data 6A is data obtained by estimating a postural state accompanied by sweating of a person to whom a pressure stimulus containing moisture is applied to the conductive urethane 22 from the input data 4 (electrical resistance). The generator 54 generates generated output data indicating a state close to the postural state accompanied by sweating of a person from the input data 4 (electrical resistance) input in time series. The generator 54 can generate generated output data 6A close to the postural state accompanied by sweating of a person to whom a pressure stimulus containing moisture is applied to the bedding 2, that is, the conductive urethane 22, by learning using a large number of input data 4 (electrical resistance). In another aspect, by capturing the electrical characteristics, which are the input data 4 input in time series, as a pattern and learning the pattern, it becomes possible to generate generated output data 6A close to the postural state accompanied by sweating of a person to whom a pressure stimulus containing moisture is applied to the bedding 2, that is, the conductive urethane 22.

[0049] The arithmetic unit 56 is an arithmetic unit that compares the generated output data 6A with the output data 6 of the learning data and calculates the error of the comparison result. The learning processing unit 52 inputs the generated output data 6A and the output data 6 of the learning data to the arithmetic unit 56. In response to this, the arithmetic unit 56 calculates the error between the generated output data 6A and the output data 6 of the learning data, and outputs a signal indicating the calculation result.

[0050] The learning processing unit 52 performs learning of the generator 54 that tunes the weight parameters of the connections between the nodes based on the error calculated by the arithmetic unit 56. Specifically, the weight parameters of the connections between the nodes of the input layer 540 and the intermediate layer 542, the weight parameters of the connections between the nodes within the intermediate layer 542, and the weight parameters of the connections between the intermediate layer 542 and the output layer 544 in the generator 54 are each fed back to the generator 54 using a method such as the gradient descent method or the error backpropagation method. That is, with the output data 6 of the learning data as the target, the connections between all the nodes are optimized so as to minimize the error between the generated output data 6A and the output data 6 of the learning data.

[0051] The learning model 51 is generated by the learning process of the learning processing unit 52. The learning model 51 is expressed as a set of information on the connection weight parameters (weights or strengths) between nodes of the learning result by the learning processing unit 52.

[0052] Fig. 12 shows an example of the flow of the learning process. In step S110, the learning processing unit 52 acquires input data 4 (electrical resistance) labeled with information indicating the postural state accompanied by sweating of the person OP, which is the learning data measured in time series. In step S112, the learning processing unit 52 generates the learning model 51 using the learning data measured in time series. That is, a set of information on the connection weight parameters (weights or strengths) between nodes of the learning result learned using a large number of learning data as described above is obtained. Then, in step S114, the data expressed as a set of information on the connection weight parameters (weights or strengths) between nodes of the learning result is stored as the learning model 51.

[0053] Note that the generator 54 may use a recurrent neural network having a function of generating an output in consideration of the context of the time-series input, or other methods may be used.

[0054] Then, in the postural state estimation device 1, the learned generator 54 (that is, the data expressed as a set of information on the connection weight parameters between nodes of the learning result) generated by the method exemplified above is used as the learning model 51. By using the sufficiently learned learning model 51, it is possible to identify the postural state accompanied by sweating of a person from the time-series electrical resistance values in the bedding 2, that is, the conductive urethane 22.

[0055] Note that the processing by the learning processing unit 52 is an example of the processing of the learning model generation device of the present disclosure. Also, the postural state estimation device 1 is an example of the estimation unit and the estimation device of the present disclosure. The output data 6, which is information indicating the postural state 3 accompanied by sweating, is an example of the postural state information of the present disclosure.

[0056] Incidentally, as described above, the conductive urethane 22 exhibits behaviors such as complex cooperation of electrical paths, elongation and contraction, expansion and contraction, temporary disconnection, and new connection of electrical paths according to deformation and water content. As a result, it exhibits behaviors with different electrical characteristics according to the applied force and moisture (for example, pressure stimulation containing moisture). This makes it possible to treat the conductive urethane 22 as a reservoir for storing data related to the deformation and water content of the conductive urethane 22. That is, the posture state estimation device 1 can apply the conductive urethane 22 to a network model called physical reservoir computing (PRC, hereinafter referred to as PRCN). Since PRC and PRCN themselves are known technologies, detailed descriptions are omitted. That is, PRC and PRCN are suitable for estimating information related to the deformation and water content of the conductive urethane 22.

[0057] Fig. 13 shows an example of a learning processing unit 52 that treats the bedding 2 containing the conductive urethane 22 as a reservoir for storing data related to the deformation of the bedding 2 containing the conductive urethane 22 and performs learning. The conductive urethane 22 has electrical characteristics (electrical resistance values) corresponding to each of various pressure stimulations containing moisture, functions as an input layer for inputting the electrical resistance values, and also functions as a reservoir layer for storing data related to the deformation and water content of the conductive urethane 22. Since the conductive urethane 22 outputs different electrical characteristics (input data 4) according to the pressure stimulation containing moisture given by the posture state 3 accompanied by a person's sweating, it is possible to estimate the pressure stimulation 3 (shape of the flexible material) containing moisture given from the electrical resistance value of the conductive urethane 22 in the estimation layer. Therefore, in the learning process, only the estimation layer needs to be learned.

[0058] The above-described posture state estimation device 1 can be realized, for example, by causing a computer to execute a program having each of the above functions.

[0059] Fig. 14 shows an example of a case where a computer is included as an execution device that executes a process for realizing various functions of the posture state estimation device 1.

[0060] The computer that functions as the posture state estimation device 1 includes a computer main body 100 shown in FIG. 14. The computer main body 100 includes a CPU 102, a RAM 104 such as a volatile memory, a ROM 106, an auxiliary storage device 108 such as a hard disk drive (HDD), and an input / output interface (I / O) 110. These CPU 102, RAM 104, ROM 106, auxiliary storage device 108, and input / output I / O 110 are configured to be connected via a bus 112 so as to be able to exchange data and commands with each other. Further, a communication unit 114, which is a communication interface (communication I / F) for communicating with an external device, and an operation display unit 116 such as a display and a keyboard are connected to the input / output I / O 110. The communication unit 114 has a function of acquiring input data 4 (electrical resistance) with the bedding 2 including the conductive urethane 22. That is, the communication unit 114 includes the bedding 2 in which the conductive urethane 22, which is a detection unit, is arranged, and can acquire the input data 4 (electrical resistance) from an electrical characteristic detection unit 76 connected to a detection point 75 in the conductive urethane 22.

[0061] The auxiliary storage device 108 stores a control program 108P for causing the computer main body 100 to function as the posture state estimation device 1 as an example of the estimation device of the present disclosure. The CPU 102 reads the control program 108P from the auxiliary storage device 108, expands it in the RAM 104, and executes the processing. As a result, the computer main body 100 that has executed the control program 108P operates as the posture state estimation device 1 as an example of the estimation device of the present disclosure.

[0062] Note that the auxiliary storage device 108 stores a learning model 108M including the learning model 51 and data 108D including various data. The control program 108P may be provided by a recording medium such as a CD-ROM.

[0063] Next, the estimation process in the posture state estimation device 1 realized by the computer will be described.

[0064] Fig. 15 shows an example of the flow of the estimation process by the control program 108P executed in the computer main body 100. The estimation process shown in Fig. 15 is executed by the CPU 102 when the power of the computer main body 100 is turned on. That is, the CPU 102 reads the control program 108P from the auxiliary storage device 108, expands it in the RAM 104, and executes the process.

[0065] First, in step S200, the CPU 102 reads the learning model 51 from the learning model 108M of the auxiliary storage device 108 and expands it in the RAM 104 to obtain the learning model 51. Specifically, a network model (see Figs. 11 and 13) that is the connection between nodes by the weight parameters expressed as the learning model 51 is expanded in the RAM 104. Thus, the learning model 51 in which the connection between nodes by the weight parameters is realized is constructed.

[0066] Next, in step S202, the CPU 102 acquires, in time series, unknown input data 4 (electrical resistance), which is the object for estimating the shape and water content of the flexible material by the pressure stimulus including the moisture given to the conductive urethane 22, via the communication unit 114.

[0067] Next, in step S204, the CPU 102 uses the learning model 51 acquired in step S200 to estimate output data 6 (posture state with unknown sweating) corresponding to the input data 4 (electrical resistance) acquired in step S202.

[0068] Then, in the next step S206, the output data 6 (posture state of a person with sweating) of the estimation result is output via the communication unit 114, and this processing routine is terminated.

[0069] Note that the estimation process shown in Fig. 15 is an example of the process executed by the estimation method of the present disclosure.

[0070] As described above, according to the present disclosure, based on the input data 4 (electrical resistance) that changes in response to the pressure stimulus including moisture given by the postural state 3 accompanied by sweating of a person with respect to the conductive urethane 22, it becomes possible to estimate the postural state accompanied by sweating of the person. That is, it becomes possible to estimate the postural state accompanied by sweating of an unknown person without using a special device or a large device, or directly measuring the deformation of a flexible material or the water content.

[0071] In addition, when the person OP is in a certain posture, the electrical characteristics by each detection set change due to the sweating state, and since the sweating state is reflected in the electrical characteristics (time-series electrical resistance), it is possible to estimate the postural state accompanied by sweating of the person from the electrical resistance value that changes in time series in the conductive urethane 22. That is, even for various postural states accompanied by sweating, by using the above-described learning model, it is possible to identify the postural state accompanied by sweating of the person, and it is possible to estimate the postural state accompanied by sweating of the person.

[0072] When estimating the postural state accompanied by sweating during the sleep state, by detecting the electrical characteristics (volume resistance) in which the physical quantities during the sleep state are reflected in time series, for example, the so-called REM sleep state and non-REM sleep state, the light sleep state and deep sleep state, the resting state, and the active state accompanied by a plurality of postures and postural states can be estimated. In addition, the estimated postural state accompanied by sweating of the person is suitably adapted to the sleeping posture accompanied by sweating in the sleep state of a person such as a sleeper, and it is possible to estimate the sleeping posture accompanied by sweating of the sleeper.

[0073] In the posture state estimation device 1 using the learning model 51 learned by the above-described learning process, it was confirmed that by inputting the electrical characteristics of the conductive urethane 22 in various unknown postural states accompanied by sweating, it is possible to estimate the postural state accompanied by sweating of the corresponding person.

[0074] As described above, in the present disclosure, the case where conductive urethane is applied as an example of the flexible material has been described, but it goes without saying that the flexible material is not limited to conductive urethane.

[0075] Moreover, the technical scope of the present disclosure is not limited to the scope described in the above embodiments. Various changes or improvements can be made to the above embodiments without departing from the gist, and the forms with such changes or improvements are also included in the technical scope of the present disclosure.

[0076] In addition, in the above embodiments, the case where the estimation process and the learning process are realized by a software configuration using a flowchart has been described, but it is not limited thereto. For example, each process may be realized by a hardware configuration.

[0077] Also, a part of the estimation device, for example, a neural network such as a learning model, may be configured as a hardware circuit.

Explanation of Reference Numerals

[0078] 1 Posture state estimation device 2 Bedding 3 Posture state accompanied by sweating 4 Input data 5 Estimation unit 6 Output data 6A Generated output data 21 Mattress 22 Conductive urethane 23 Conductive urethane piece 51 Learning model 52 Learning processing unit 54 Generator 56 Arithmetic unit 75 Detection point 76 Electrical property detection unit

Claims

1. A detection unit that detects the electrical characteristics between a plurality of predetermined detection points on the flexible material of a bedding provided with a flexible material having conductivity and whose electrical characteristics change according to changes in pressure and moisture given by a postural state accompanied by moisture content of a person; Using the time-series electrical characteristics when pressure and moisture are applied to the flexible material and the postural state information indicating the postural state accompanied by moisture content of the person who applies pressure and moisture to the flexible material as learning data, for a learning model learned to input the time-series electrical characteristics and output the postural state information, input the time-series electrical characteristics detected by the detection unit, and estimate the postural state information indicating the postural state accompanied by moisture content of the person corresponding to the input time-series electrical characteristics An estimation unit; including, An estimation device, wherein the postural state accompanied by moisture content of the person includes a postural state accompanied by sweating in the sleep state of the person.

2. The electrical characteristic is volume resistance, The bedding includes a mattress, The flexible material is a material in which conductivity is imparted to at least a part of a urethane material having a structure with at least one of a fibrous and a mesh-like skeleton, or a structure in which a plurality of minute air bubbles are scattered inside. The estimation device according to claim 1.

3. The urethane material is impregnated with a conductive material. The estimation device according to claim 2.

4. The learning model includes a model generated by treating the flexible material as a reservoir for storing data related to deformation of the flexible material in reservoir computing and learning using a network by the reservoir computing using the reservoir. The estimation device according to any one of claims 1 to 3.

5. A computer, acquires the electrical characteristics from a detection unit that detects the electrical characteristics between a plurality of predetermined detection points on the flexible material of a bedding provided with a flexible material having conductivity and whose electrical characteristics change according to changes in pressure and moisture given by a postural state accompanied by moisture content of a person; Using, as learning data, the time-series electrical characteristics when pressure and moisture are applied to the flexible material and the posture state information indicating the posture state of the person with moisture content who applies pressure and moisture to the flexible material, for a learning model trained to take the time-series electrical characteristics as input and output the posture state information, inputting the time-series electrical characteristics detected by the detection unit, and estimating the posture state information indicating the posture state of the person with moisture content corresponding to the input time-series electrical characteristics, which is an estimation method comprising: The posture state of the person with moisture content includes a posture state accompanied by sweating in the sleeping state of the person. Estimation method.

6. On a computer, Obtaining the electrical characteristics from a detection unit that detects the electrical characteristics between a plurality of predetermined detection points on the flexible material of a bedding having a flexible material that has conductivity and whose electrical characteristics change in response to changes in pressure and moisture given by the posture state of a person with moisture content, Using, as learning data, the time-series electrical characteristics when pressure and moisture are applied to the flexible material and the posture state information indicating the posture state of the person with moisture content who applies pressure and moisture to the flexible material, for a learning model trained to take the time-series electrical characteristics as input and output the posture state information, inputting the time-series electrical characteristics detected by the detection unit, and estimating the posture state information indicating the posture state of the person with moisture content corresponding to the input time-series electrical characteristics An estimation program for causing the execution of processing, The posture state of the person with moisture content includes a posture state accompanied by sweating in the sleeping state of the person. Estimation program.

7. An acquisition unit that acquires the electrical characteristics from a detection unit that detects the electrical characteristics between a plurality of predetermined detection points on the flexible material of a bedding having a flexible material that has conductivity and whose electrical characteristics change in response to changes in pressure and moisture given by the posture state of a person with moisture content, and the posture state information indicating the posture state of the person with moisture content who applies pressure and moisture to the flexible material, A learning model generation unit that generates a learning model that takes the time-series electrical characteristics when pressure and moisture are applied to the flexible material as input and outputs the posture state information indicating the posture state of the person with moisture content based on the acquisition result of the acquisition unit, Including, A learning model generation device in which the posture state of the person with moisture content includes a posture state accompanied by sweating in the sleeping state of the person.

Images