State estimation device, state estimation system, and control method for state estimation device

Abstract

To provide a state estimation device that can accurately estimate the state of a person to be observed.SOLUTION: A state estimation device comprises: an input unit that receives input of estimation material information used for estimating the state of a person to be observed; a control unit that generates estimation information obtained by estimating the state of the person to be observed by using an estimation model on the basis of the estimation material information received by the input unit; and an output unit that outputs the estimation information generated by the control unit. The control unit acquires difference information indicating the difference between estimation history information obtained through estimation in advance by using the estimation model on the basis of the past estimation material information of the person to be observed and actual information indicating the actual state of the person to be observed, and when determining that the difference information falls out of a predetermined range, changes the estimation model to a corrected estimation model obtained by correcting the estimation model according to the difference information before generating the estimation information.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a state estimation device, a state estimation system, and a control method for a state estimation device.

Background Art

[0002] An abnormality detection device that determines the presence or absence of an abnormality of an occupant based on an estimated value is disclosed (Patent Document 1). The abnormality detection device disclosed in Patent Document 1 can correct the determination condition for determining the presence or absence of an abnormality using the estimated value based on the measured value.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Patent Document 1 does not consider the deviation between the estimated result and the actual information indicating the actual result that occurs for each observer such as an occupant, and does not estimate the state of the observer. The object of the present disclosure is to provide a state estimation device and a state estimation system capable of accurately estimating the state of an observer.

Means for Solving the Problems

[0005] A state estimation device according to one aspect of the present disclosure includes an input unit that receives input of estimation material information used to estimate the state of a person being observed, a control unit that generates estimation information by estimating the state of the person being observed using an estimation model based on the estimation material information received by the input unit, and an output unit that outputs the estimation information generated by the control unit. The control unit acquires difference information indicating the difference between estimation history information obtained in advance using the estimation model based on the past estimation material information of the person being observed and actual information indicating the actual state of the person being observed, and if it determines that the difference information is outside a predetermined range, it changes the estimation model to a modified estimation model that has been modified according to the difference information and generates the estimation information.

[0006] A state estimation system according to one aspect of the present disclosure includes: an input unit that receives input of estimation material information used to estimate the state of a person being observed; a control unit that generates estimation information by estimating the state of the person being observed using an estimation model based on the estimation material information received by the input unit; and an output unit that outputs the estimation information generated by the control unit. The control unit acquires difference information indicating the difference between estimation history information obtained in advance using the estimation model based on the past estimation material information of the person being observed and actual information indicating the actual state of the person being observed. If it is determined that the difference information falls outside a predetermined range, the control unit changes the estimation model to a modified estimation model that has been modified according to the difference information and generates the estimation information. The difference information is associated with the identification information of each of a plurality of persons being observed. The control unit also includes: a state estimation device that receives the identification information of each of the persons being observed and acquires the difference information corresponding to the identification information; and a personal authentication server that authenticates the person being observed and transmits the identification information of the person being authenticated to the state estimation device.

[0007] A control method for a state estimation device according to one aspect of the present disclosure includes the steps of: receiving input of estimation material information used to estimate the state of a person being observed; generating estimation information by estimating the state of the person being observed using an estimation model based on the estimation material information received in the step of receiving input of estimation material information; and outputting the estimation information generated in the step of generating the estimation information. The method further includes, in the step of generating the estimation information, acquiring difference information indicating the difference between estimation history information obtained in advance using the estimation model based on the past estimation material information of the person being observed and actual information indicating the actual state of the person being observed, and if it is determined that the difference information falls outside a predetermined range, changing the estimation model to a modified estimation model that has been modified according to the difference information. [Brief explanation of the drawing]

[0008] [Figure 1] This is a block diagram showing an example of the main components of a state estimation device according to the first embodiment of this disclosure. [Figure 2] This flowchart shows an example of a state estimation method using a state estimation device according to the first embodiment of this disclosure. [Figure 3] This block diagram shows the main components of the state estimation system according to the second embodiment of this disclosure. [Figure 4] This is a table showing an example of differential information stored in the memory unit of the state estimation device according to the second embodiment of this disclosure. [Figure 5] This is a table showing an example of differential information stored in the memory unit of the state estimation device according to the second embodiment of this disclosure. [Figure 6] This is a table showing an example of differential information stored in the memory unit of the state estimation device according to the second embodiment of this disclosure. [Figure 7] This graph shows an example of pulse information obtained by separating blood flow information from a subject into individual beats. [Figure 8] Figure 7 is a graph showing an example of first-derived pulse information obtained by first-derived differentiation of the pulse information shown. [Figure 9]Figure 7 is a graph showing an example of second-derived pulse information obtained by taking the second derivative of the pulse information shown. [Figure 10] This figure shows an example of an estimated blood glucose level output by a state estimation device according to the second embodiment of this disclosure. [Figure 11] This figure shows an example of an estimated blood glucose level output by a state estimation device according to the second embodiment of this disclosure. [Figure 12] This figure shows an example of an estimated blood glucose level output by a state estimation device according to the second embodiment of this disclosure. [Figure 13] This flowchart shows an example of a state estimation method using a state estimation device according to a second embodiment of this disclosure. [Figure 14] This is a block diagram showing an example of the main components of a state estimation system according to a first modified example of a second embodiment of the present disclosure. [Figure 15] This figure shows an example of an input screen for inputting update differential information to a state estimation device in a state estimation system according to a first modification of the second embodiment of the present disclosure. [Figure 16] This figure shows an example of an input screen for inputting update differential information to a state estimation device in a state estimation system according to a first modification of the second embodiment of the present disclosure. [Figure 17] This is a block diagram showing an example of the main components of a state estimation system according to the third embodiment of this disclosure. [Figure 18] This flowchart shows an example of a state estimation method using a state estimation device according to the third embodiment of this disclosure. [Figure 19] This is a block diagram showing an example of the main components of a state estimation system according to the fourth embodiment of this disclosure. [Figure 20] This graph shows an example of pulse information obtained by separating blood flow information from a subject into individual beats. [Figure 21] This flowchart shows an example of a state estimation method using a state estimation device according to the fourth embodiment of this disclosure. [Modes for carrying out the invention]

[0009] Hereinafter, embodiments and modifications of the present disclosure will be described while referring to the drawings. In the following, the same or corresponding members are denoted by the same reference numerals throughout all the drawings, and redundant descriptions thereof are omitted. In addition, the embodiments and modifications described below are merely examples of the present disclosure, and the present disclosure is not limited to the embodiments and modifications. Even outside these embodiments and modifications, various changes can be made according to the design and the like as long as the technical idea of the present disclosure is not deviated from.

[0010] (First Embodiment) Referring to FIG. 1, the configuration of the state estimation device 1 according to the first embodiment of the present disclosure will be described. FIG. 1 is a block diagram showing an example of the main configuration of the state estimation device 1 according to the first embodiment of the present disclosure.

[0011] The state estimation device 1 is a device that outputs estimation result information 21 for estimating the state of the observer 2 based on the estimation material information 20 acquired from the observer 2. As shown in FIG. 1, the state estimation device 1 has a configuration including an input unit 10, a control unit 11 (estimation unit), an output unit 12, and a storage unit 13.

[0012] The estimation material information 20 is information used to estimate the state of the observer 2, and is information that can be used to estimate the state of the observer 2 without directly measuring the state of the observer 2. For example, when estimating the pulse rate as the state of the observer 2, the estimation material information 20 can be a face image of the observer 2 or an image of a fingertip or palm captured by an imaging device such as a camera. When estimating blood pressure or body temperature as the state of the observer 2, the estimation material information 20 can be a face image of the observer 2 captured by an imaging device. When estimating the blood glucose level as the state of the observer 2, the estimation material information 20 can be a blood flow video of the finger of the observer 2 captured by an imaging device. When estimating the emotion (joy, anger, sorrow, happiness) as the state of the observer 2, the estimation material information 20 can be a face image of the observer 2 captured by an imaging device, the loudness and pitch of the voice of the observer 2, the speaking speed of the observer 2, and the like. In this specification, the term "image" includes not only still images but also moving images.

[0013] Estimated result information 21 (estimated information) is information that shows the result of estimating the state of the observed person 2. Examples include estimated values ​​such as blood pressure, body temperature, and blood glucose level, or information that shows the estimated emotions of the observed person 2 (e.g., joy, anger, sadness, etc.).

[0014] The input unit 10 receives the estimated material information 20, and the estimated material information 20 received by the input unit 10 is transmitted to the control unit 11. If the estimated material information 20 is, for example, a facial image or finger image of the subject 2, the input unit 10 can be an input interface that receives video signals from an imaging device (not shown). If the estimated material information 20 is, for example, the voice of the subject 2, the input unit 10 can be an input interface that receives audio signals from a microphone (not shown). Furthermore, if the state estimation device 1 is configured to include an imaging device or microphone as an integrated unit, the input unit 10 can be, for example, the imaging device or microphone.

[0015] The control unit 11 performs various controls on each part of the state estimation device 1, and is composed of a processor such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit).

[0016] The control unit 11 includes an estimation unit 30 as a functional block. If the control unit 11 is, for example, a CPU, the estimation unit 30 can be realized by the CPU reading a program stored in the storage unit 13 into a memory (not shown) and executing it.

[0017] The estimation unit 30 generates estimation result information 21 by estimating the state of the observed person 2 using the estimation model 23, based on the estimation material information 20 received by the input unit 10. The generated estimation result information 21 is output to the outside of the state estimation device 1 via the output unit 12. If the estimation unit 30 determines that the difference information 22 is outside a predetermined range, it changes the estimation model 23 to a modified estimation model 24 that has been modified according to the difference information and generates estimation result information 21. Details of how the estimation unit 30 generates estimation result information 21 will be described later.

[0018] The difference information 22 is information that shows the difference between the estimation result (estimated history information) obtained in advance using the estimation model 23 based on the past estimated material information 20 obtained from the observed person 2, and the actual information that shows the actual state of the observed person 2. The actual information of the observed person 2 can be, for example, actual measured values ​​such as blood pressure, body temperature, and blood glucose level obtained from the observed person 2 using measuring instruments, etc. Alternatively, it can be information that shows the emotions of the observed person 2 obtained by directly observing the observed person 2.

[0019] The estimation model 23 is a standard model used when the state estimation device 1 estimates the state of the observed person 2. Estimation model 23 is used when the difference information 22 of the observed person 2 falls within a predetermined range. On the other hand, the first modified estimation model 24a is used when the difference information 22 of the observed person 2 is greater than the predetermined range, and the second modified estimation model 24b is used when the difference information 22 of the observed person 2 falls below the predetermined range. The modified estimation model 24 includes the first modified estimation model 24a and the second modified estimation model 24b.

[0020] The difference information 22 is a difference value that shows the difference between the measured value and the estimated value when the estimated historical information is obtained as an estimated value and the actual information is obtained as an actual measured value. Alternatively, the difference information 22 can also be information that represents the trend of the relationship between the magnitude of the estimated value and the actual value, derived from the difference between the measured value and the estimated value.

[0021] Furthermore, if the estimated historical information and actual information are obtained as information indicating emotion, the difference information 22 will be the difference between the standard threshold for discriminating between different emotions based on the estimated material information 20 (e.g., volume of voice) and the threshold at which the observed person 2 can actually discriminate between different emotions.

[0022] The storage unit 13 is a recording medium capable of storing various data and programs. The storage unit 13 includes an area for storing various programs, an area for storing data used in the various programs, an area where the various programs are loaded, and an area used when the various programs are executed. The storage unit 13 can be composed of, for example, semiconductor memory such as ROM (Read Only Memory) or RAM (Random Access Memory), an HDD (Hard Disk Drive), an SDD (Solid State Drive), or a programmable logic circuit. The storage unit 13 stores the differential information 22, the estimation model 23, and the modified estimation model 24 described above.

[0023] The difference information 22 may be obtained from an external server (not shown). Alternatively, the storage unit 13 may store estimated history information, which shows the estimated state of the observed person 2 generated using an estimation model 23 based on the observed person 2's past estimated material information 20, and actual information of the observed person 2, and the control unit 11 may calculate the difference information 22 from the estimated history information and actual information.

[0024] The output unit 12 can be an output interface that outputs a video signal to a display device (not shown) when outputting the estimated result information 21 as video. The output unit 12 can also be an output interface that outputs an audio signal to a speaker (not shown) when outputting the estimated result information 21 as audio. Furthermore, the output unit 12 can be an output interface that outputs print data to a printing device (not shown) when outputting the estimated result information 21 as printed data. In the case where the state estimation device 1 integrates a display device, speaker, or printing device, the output unit 12 can be, for example, a display device, speaker, or printing device.

[0025] (Method for estimating the state) Next, a state estimation method using the state estimation device 1 according to the first embodiment will be described with reference to Figure 2. Figure 2 is a flowchart showing an example of a state estimation method using the state estimation device 1 according to the first embodiment of this disclosure.

[0026] First, when the power to the state estimation device 1 is turned ON and the output of the estimated result information 21 of the observed person 2 is instructed, the estimation unit 30 of the state estimation device 1 acquires the estimated material information 20 of the observed person 2 via the input unit 10 (step S11). Furthermore, the estimation unit 30 acquires the difference information 22 from the storage unit 13 (step S12). Note that step S12 may be executed before step S11, and the order in which steps S11 and S12 are executed is arbitrary.

[0027] Next, the estimation unit 30 determines whether the difference information 22 obtained in step S12 is within a predetermined range (step S13). The predetermined range is, for example, a range where the magnitude of the difference information exceeds ±20 mg / dl for blood glucose levels, or a range where it exceeds ±20 mmHg for blood pressure.

[0028] If it is determined that the difference information 22 is within a predetermined range (Yes in step S13), the state of the observed person 2 is estimated using the estimation model 23 (step S14).

[0029] On the other hand, if it is determined that the difference information 22 falls outside a predetermined range (No in step S13), the state of the observed person 2 is estimated using the modified estimation model 24 selected based on the difference information 22 (step S15). The modified estimation model 24 selected in step S15 may be the first modified estimation model 24a or the second modified estimation model 24b.

[0030] If the value of the actual information indicating the actual state of the observed person 2 is greater than the value of the estimated result estimated by estimation model 23, then the first modified estimation model 24a is selected, which yields a value smaller by the difference than the estimated result obtained when estimation model 23 is estimated based on the estimated material information 20.

[0031] Furthermore, if the value of the actual information representing the actual state of the observed person 2 is smaller than the estimation result estimated by estimation model 23, a second modified estimation model 24b is selected, which yields a value that is larger by the difference than the estimation result obtained when estimation model 23 is estimated based on the estimated material information 20.

[0032] The estimation unit 30 generates estimation result information 21 using the estimation model 23 or the modified estimation model 24 based on the estimated material information 20, and then outputs the generated estimation result information 21 to the outside via the output unit 12 (step S16).

[0033] As described above, the state estimation device 1, with its estimation unit 30 in the control unit 11, can generate estimation result information 21 based on the estimated material information 20 acquired from the observed person 2, taking into account the difference information 22.

[0034] Therefore, when the state estimation device 1 uses the estimation model 23 to estimate the state of the observed person 2, it is possible to generate corrected estimation result information 21 that is appropriate even if the estimation result is always larger or smaller than the actual information that represents the actual state. Therefore, even if there are individual differences in the estimation results using the estimation model 23 for each observed person 2, accurate estimation result information 21 can be obtained.

[0035] Therefore, the state estimation device 1 can accurately estimate the state of the observed person 2.

[0036] (Second Embodiment) The state estimation system 100 according to the second embodiment of this disclosure will be described with reference to Figure 3. Figure 3 is a block diagram showing the main components of the state estimation system 100 according to the second embodiment of this disclosure. In the state estimation system 100 according to the second embodiment, blood glucose levels will be used as an example to explain the information indicating the state of the observed person 2.

[0037] The state estimation system 100 according to the second embodiment comprises a state estimation device 1 and a personal authentication server 40.

[0038] The state estimation device 1 according to the second embodiment differs from the state estimation device 1 according to the first embodiment in that it acquires the observed person identification information 25 of the observed person 2 from an external personal authentication server 40. In addition, in the state estimation device 1 according to the second embodiment, the estimated material information 20 is the blood flow information of the observed person 2, and the estimated result information 21 is the estimated value of the blood glucose level. In all other respects, the state estimation device 1 according to the second embodiment is the same as the state estimation device 1 according to the first embodiment, so the same reference numerals are used for similar components, and their descriptions are omitted.

[0039] The state estimation device 1 according to the first embodiment is a device used exclusively for the observed person 2, whereas the state estimation device 1 according to the second embodiment is a device shared by multiple observed people 2. Therefore, the state estimation device 1 according to the second embodiment needs to identify the observed person 2 whose state is to be estimated and generate estimation result information 21 using difference information 22 provided according to the identified observed person 2.

[0040] Therefore, in the state estimation device 1 according to the second embodiment, the estimation unit 30 obtains observed person identification information 25 from the personal authentication server 40 to identify the observed person 2, and generates estimation result information 21 using difference information 22 corresponding to the observed person 2 identified by the observed person identification information 25. The observed person identification information 25 can be, for example, an ID that can identify the observed person 2.

[0041] The personal authentication server 40 is a server that authenticates the observed person 2 whose state is estimated by the state estimation device 1, and transmits the observed person identification information 25 (identification information) of the observed person 2 who has successfully been authenticated to the state estimation device 1. The personal authentication server 40 may accept input such as a password from the observed person 2 and perform the authentication process, or it may acquire a facial image of the observed person 2 and perform the authentication process based on the acquired facial image.

[0042] In the state estimation device 1, the estimation unit 30 receives blood flow information (estimated material information 20) of the observed person 2 via the input unit 10. The estimation unit 30 also receives observed person identification information 25 from the personal authentication server 40.

[0043] The estimation unit 30 then reads the difference information 22 corresponding to the observed person identification information 25 from the storage unit 13 and selects either the estimation model 23 or the modified estimation model 24 according to the difference information 22. Then, using the selected estimation model 23 or modified estimation model 24, the estimation unit 30 generates an estimated value of blood glucose level as estimation result information 21 based on the received blood flow information.

[0044] The difference information 22 stored in the memory unit 13 may be, for example, a table showing the correspondence between the ID of the observed person 2 and the difference, as shown in Figure 4. Figure 4 is a table showing an example of difference information 22 stored in the memory unit 13 of the state estimation device 1 according to the second embodiment of this disclosure. As shown in Figure 4, for example, the difference information 22 records the ID of observed person A and the difference (-22 mg / dl) between the estimated blood glucose level generated using the estimation model 23 based on the past blood flow information of observed person A and the actual blood glucose level measured from observed person A. Furthermore, the ID of observed person B and the difference (0 mg / dl) between the estimated blood glucose level generated using the estimation model 23 based on the past blood flow information of observed person B and the actual blood glucose level measured from observed person B are recorded in association.

[0045] In the difference information 22, the difference associated with subject 2 may be configured to be updated each time an estimated blood glucose level is calculated based on subject 2's blood flow information. The updated difference value may also be the difference between the estimated blood glucose level calculated using the estimation model 23 based on subject 2's blood flow information and the measured blood glucose level. Alternatively, it may be the average of the difference between the previously estimated blood glucose level and the measured blood glucose level using the estimation model 23, and the difference between the currently estimated blood glucose level and the measured blood glucose level using the estimation model 23. For example, if the previously calculated difference value was -21 mg / dl and the currently calculated difference value is, for example, -23 mg / dl, the value updated as difference information 22 may be the average of the two, which is -22 mg / dl. Alternatively, the value updated as difference information 22 may be the smallest of the two, which is -23 mg / dl.

[0046] Furthermore, the difference information 22 can also be a table of information that associates the ID of the observed person 2, an estimated value obtained by estimating the blood flow information of the observed person 2 using the estimation model 23, the estimated time when this estimated value was obtained, and the actual measured blood glucose value obtained from the observed person 2, as shown in Figure 5. The actual measured value can be the blood glucose value measured from the observed person 2 at the time closest in time before or after the estimated time. Figure 5 is a table showing an example of the difference information 22 stored in the storage unit 13 of the state estimation device 1 according to the second embodiment of this disclosure.

[0047] Thus, the difference information 22 may not be the difference itself, but rather a table showing the correspondence between the estimated value and the measured value associated with each observed person 2.

[0048] The measured values ​​included in the difference information 22 may be measured values ​​obtained at the same time as the estimated value was calculated, or measured values ​​obtained at the time closest to the estimated time when the estimated value was calculated.

[0049] Furthermore, the difference information 22 may also be a table showing the correspondence between the ID of the observed person 2, the estimated blood glucose level estimated using a standard estimation model 23 based on the blood flow information of the observed person 2, and information representing the trend of the magnitude relationship between the estimated value and the measured value, derived from the difference between the estimated blood glucose level and the actual blood glucose level obtained by measurement from the observed person 2. Figure 6 is a table showing an example of the difference information 22 stored in the storage unit 13 of the state estimation device 1 according to the second embodiment of this disclosure.

[0050] Information that describes the trend in the relationship between estimated values ​​and measured values ​​includes, for example, information that indicates a trend, such as "estimated to be low" when the estimated blood glucose level is smaller than a predetermined range based on the measured value, "reasonable" when the estimated blood glucose level falls within a predetermined range based on the measured value, or "estimated to be high" when the estimated blood glucose level is larger than a predetermined range based on the measured value.

[0051] The state estimation device 1 was configured to perform the authentication process for the observed person 2 using the personal authentication server 40, and to transmit the observed person identification information 25 of the observed person 2 who successfully authenticated to the state estimation device 1. However, the state estimation device 1 may also be configured to have an authentication unit instead of the personal authentication server 40, and to perform the authentication process for the observed person 2. However, the configuration in which the authentication process for the observed person 2 is performed by a personal authentication server 40 located outside the state estimation device 1 is more advantageous in terms of security, as it can prevent the leakage of personal information.

[0052] Furthermore, the estimated model 23 stored in the memory unit 13 can be created, for example, as follows: The blood flow information obtained from the observed person 2 is divided into pulse information (pulse waveforms) for each beat, as shown in Figure 7. Each of the divided pulse information is further differentiated to obtain the first derivative pulse information shown in Figure 8 and the second derivative pulse information shown in Figure 9, respectively.

[0053] Figure 7 is a graph showing an example of pulse information obtained by dividing the blood flow information from subject 2 into individual beats. Figure 8 is a graph showing an example of first-derived pulse information obtained by first-derived differentiation of the pulse information shown in Figure 7. Figure 9 is a graph showing an example of second-derived pulse information obtained by second-derived differentiation of the pulse information shown in Figure 7. In the graphs shown in Figures 7, 8, and 9, the horizontal axis represents time, and the vertical axis represents the detected amount detected from subject 2.

[0054] Then, feature quantities are obtained for the feature points of the waveforms of the pulse information shown in Figure 7, the first derivative pulse information shown in Figure 8, and the second derivative pulse information shown in Figure 9. Examples of feature quantities for the feature points include the first maximum value, the difference between the first minimum value and the first maximum value, and the time from the starting point to the first maximum value for each of the pulse information, first derivative pulse information, and second derivative pulse information. By preparing a large amount of data that associates the feature quantities of multiple feature points with measured blood glucose levels, and using this prepared data to perform machine learning, an estimation model 23 can be created.

[0055] The first modified estimation model 24a can be created in the same manner as the estimation model 23 described above, for example, by learning more pulse information obtained from each of the observed subjects 2 whose blood glucose levels are slightly higher or higher than the normal range.

[0056] The second modified estimation model 24b can be created in the same manner as the estimation model 23 described above, for example, by learning more pulse information obtained from each of the observed subjects 2 whose blood glucose levels are slightly lower or lower than the normal range.

[0057] In the above examples, the first modified estimation model 24a and the second modified estimation model 24b are generated by training them using machine learning based on the feature quantities of the feature points of the waveforms of pulse information, first derivative pulse information, and second derivative pulse information, respectively, but the model is not limited to this. For example, the first modified estimation model 24a may be an estimation model that can obtain an estimated value obtained by subtracting a predetermined value from the estimated blood glucose level estimated using estimation model 23, and the second modified estimation model 24b may be an estimation model that can obtain an estimated value obtained by adding a predetermined value to the estimated blood glucose level estimated using estimation model 23.

[0058] The above-described estimation model 23 and modified estimation model 24 may be created by the control unit 11 of the state estimation device 1, or they may be created by an external server (not shown), and the state estimation device 1 may acquire them from the external server and store them in the storage unit 13.

[0059] As described above, in the configuration where estimation model 23 and modified estimation model 24 are created from the feature quantities of the feature points of the blood flow information, the state estimation device 1, when the estimation unit 30 receives blood flow information via the input unit 10, calculates the feature quantities of the feature points of the waveforms of pulse information, first derivative pulse information, and second derivative pulse information from the blood flow information. Then, the estimation unit 30 uses the estimation model 23 or modified estimation model 24 selected based on the difference information 22 to obtain an estimated value of blood glucose level from the calculated feature quantities. The estimation unit 30 outputs the obtained estimated value of blood glucose level to the outside via the output unit 12.

[0060] For example, when the estimated blood glucose level is output to an external display device via the output unit 12, the estimated blood glucose level can be displayed on the display screen of the display device, for example, as shown in Figures 10 to 12. Figures 10 to 12 are diagrams showing examples of estimated blood glucose levels output by the state estimation device 1 according to the second embodiment of this disclosure. Figure 10 shows an example of display when the estimated blood glucose level is obtained using the second modified estimation model 24b, Figure 11 shows an example of display when the estimated blood glucose level is obtained using the estimation model 23, and Figure 12 shows an example of display when the estimated blood glucose level is obtained using the first modified estimation model 24a.

[0061] In this case, for subject A shown in Figure 10, the estimated blood glucose level obtained using estimation model 23 is lower than the actual measured value. Therefore, an upward-pointing arrow is displayed in the lower right corner of the display screen showing the estimated blood glucose level, indicating that the second modified estimation model 24b was used to obtain a higher estimated value than the one obtained using estimation model 23.

[0062] In Figure 11, when the blood glucose level of subject B is estimated using estimation model 23, the estimated value is the same as or close to the actual measured value. Therefore, no arrow is displayed in the lower right corner of the screen where the estimated blood glucose level is shown.

[0063] As shown in Figure 12, when the estimated blood glucose level of subject C is calculated using estimation model 23, the result is higher than the measured value. Therefore, a downward arrow is displayed in the lower right corner of the display screen showing the estimated blood glucose level, indicating that the first modified estimation model 24a was used to obtain a lower estimated value than the one obtained using estimation model 23.

[0064] Furthermore, as shown in Figures 10 to 12, instruction buttons for "Record," "Cancel," and "Reset" are displayed below the estimated blood glucose level. For example, if the "Record" instruction button is selected, the estimated blood glucose level displayed on the screen is recorded on the recording medium or the like. If the "Cancel" instruction button is selected, the estimated blood glucose level is canceled. If the "Reset" instruction button is selected, the state estimation device 1 is instructed to regenerate the estimated blood glucose level.

[0065] As described above, in the state estimation device 1 according to the second embodiment, the display screen showing the estimated blood glucose level indicates whether the estimated blood glucose level was generated using estimation model 23, first modified estimation model 24a, or second modified estimation model 24b.

[0066] (Method for estimating the state) The state estimation device 1 having the above-described configuration can perform a state estimation method for the observed person 2, as shown in Figure 13. Figure 13 is a flowchart showing an example of a state estimation method by the state estimation device 1 according to the second embodiment of this disclosure.

[0067] First, when the power to the state estimation device 1 is turned ON and the output of the estimated result information 21 of the state of the observed person 2 is instructed, the estimation unit 30 of the state estimation device 1 acquires the estimated material information 20 (blood flow information) of the observed person 2 via the input unit 10 (step S21). Furthermore, the estimation unit 30 acquires the observed person identification information 25 of the observed person 2 from the personal authentication server 40 (step S22). Based on the observed person identification information 25 acquired in step S22, the estimation unit 30 acquires the difference information 22 corresponding to the observed person 2 from the storage unit 13 (step S23). Note that steps S22 and S23 may be executed before step S21.

[0068] The processes from step S24 to step S27 shown in Figure 13 are the same as the processes from step S13 to step S16 shown in Figure 2, so their explanation will be omitted.

[0069] As described above, the state estimation device 1, with its estimation unit 30 in the control unit 11, can generate estimation result information 21 based on the estimated material information 20 acquired from the observed person 2, taking into account the difference information 22.

[0070] Therefore, when the state estimation device 1 uses the estimation model 23 to estimate the state of the observed person 2, it is possible to generate corrected estimation result information 21 that is appropriate even if the estimation result is always larger or smaller than the actual information that represents the actual state. Therefore, even if there are individual differences in the estimation results using the estimation model 23 for each observed person 2, accurate estimation result information 21 can be obtained.

[0071] Therefore, the state estimation device 1 can accurately estimate the state of the observed person 2.

[0072] (First variation) Next, the configuration of the state estimation system 100 according to the first modified example of the second embodiment of this disclosure will be described with reference to Figure 14. Figure 14 is a block diagram showing an example of the main components of the state estimation system 100 according to the first modified example of the second embodiment of this disclosure.

[0073] The state estimation system 100 according to the first modified example of the second embodiment differs from the state estimation system 100 according to the second embodiment in that the differential information 22 stored in the storage unit 13 can be changed to update differential information input from an external source.

[0074] More specifically, in the state estimation system 100 according to the first modification of the second embodiment, the state estimation device 1 further comprises a differential information input unit 14, and the control unit 11 further includes a differential information update unit 31 as a functional block. If the control unit 11 is, for example, a CPU, the differential information update unit 31 can be realized by the CPU reading a program stored in the storage unit 13 into a memory (not shown) and executing it.

[0075] Regarding the other components, the state estimation system 100 according to the first modified example of the second embodiment has the same configuration as the state estimation system 100 according to the second embodiment, so the same reference numerals are used for similar components, and their descriptions are omitted.

[0076] The differential information input unit 14 receives update instructions for differential information 22 along with update differential information from an external source, and the update instructions for differential information 22 received by the differential information input unit 14 are transmitted to the control unit 11. For example, the differential information input unit 14 can be an input interface that receives update instructions for differential information 22 and update differential information 22 from an input device (not shown). Also, if the state estimation device 1 is configured to have an integrated input device, the differential information input unit 14 can be an input device such as a touch panel.

[0077] More specifically, for example, if the external input device or the differential information input unit 14 is a touch panel, the input screen of the touch panel displays a table showing the differential information 22 stored in the storage unit 13, as shown in Figure 15. In this table, a change button is displayed in parallel with the differential information corresponding to each observed person 2. Figure 15 is a diagram showing an example of an input screen for inputting update differential information to the state estimation device 1 in the state estimation system 100 according to the first modified example of the second embodiment of this disclosure.

[0078] Here, if the operator of the state estimation device 1 wants to change the information showing the difference for observed person B, for example, they select the change button that is displayed in parallel with the difference corresponding to observed person B.

[0079] When the Change button is selected, a dropdown menu for updating the difference associated with the selected Change button appears on the input screen, as shown in Figure 16. In the examples shown in Figures 15 and 16, in addition to the information indicating the difference that was considered "reasonable" in the difference before the change, the dropdown menu also displays the options "estimated value is low" or "estimated value is high." Figure 16 is a diagram showing an example of an input screen for inputting update difference information to the state estimation device 1 in the state estimation system 100 according to the first modification of the second embodiment of this disclosure.

[0080] When the operator instructs the system to update the difference information 22 by selecting an appropriate option from the options displayed in the pull-down menu on the input screen, the state estimation device 1 changes the difference information contained in the difference information 22 to the information of the selected option. In other words, the state estimation device 1's difference information update unit 31 changes the difference information contained in the difference information 22 stored in the memory unit 13 to the new difference information selected on the input screen, based on the instruction to update the difference information 22.

[0081] Here, three types of difference information are set for each observed person 2: "Estimated value is low," "Valid," and "Estimated value is high." However, it is not limited to these three, and there may be more than three. Also, the difference information is not limited to the above-mentioned textual expressions, but may be numerical.

[0082] (Third embodiment) Next, with reference to Figure 17, a state estimation system 200 according to the third embodiment of this disclosure will be described. Figure 17 is a block diagram showing an example of the main components of the state estimation system 200 according to the third embodiment of this disclosure.

[0083] The state estimation system 100 according to the second embodiment had a configuration in which differential information 22 was stored in a storage unit 13 provided in the state estimation device 1. In contrast, the state estimation system 200 according to the third embodiment differs in that it further includes an information management server 60 that stores the differential information 22, and the state estimation device 1 is configured to acquire the differential information 22 from the information management server 60. In other respects, the state estimation system 200 according to the third embodiment has the same configuration as the state estimation system 100 according to the second embodiment, so the same reference numerals are used for similar components and their descriptions are omitted.

[0084] The information management server 60 is connected to the state estimation device 1 in a communicative manner and, as shown in Figure 17, comprises a server input unit 61, a server control unit 62, and a server storage unit 63.

[0085] The server input unit 61 receives differential information 22 from an external source, and the differential information 22 received by the server input unit 61 is transmitted to the server control unit 62. For example, the server input unit 61 can be an input interface that receives differential information 22 from an input device (not shown). Also, if the information management server 60 is configured to have an integrated input device, the server input unit 61 can be an input device such as a touch panel.

[0086] The server control unit 62 performs various controls on the various components of the information management server 60, and is composed of processors such as a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit).

[0087] The server control unit 62 includes a differential information recording unit 70 as a functional block. If the server control unit 62 is, for example, a CPU, the differential information recording unit 70 can be realized by the CPU reading a program stored in the server storage unit 63 into a memory (not shown) and executing it.

[0088] The differential information recording unit 70 records the differential information 22 received via the server input unit 61 in the server storage unit 63.

[0089] The server storage unit 63 is a recording medium capable of storing various data and programs. The server storage unit 63 includes an area for storing various programs, an area for storing data used in those programs, an area where those programs are loaded, and an area used when those programs are executed. The server storage unit 63 can be configured, for example, with semiconductor memory such as ROM (Read Only Memory) or RAM (Random Access Memory), an HDD (Hard Disk Drive), an SSD (Solid State Drive), or a programmable logic circuit. The server storage unit 63 stores the differential information 22 described above. The differential information 22 stored in the server storage unit 63 may be input from an external source via the server input unit 61 as described above.

[0090] Alternatively, the server storage unit 63 may store a history of the estimated state of the observed person 2 (estimated blood glucose level) estimated using the estimation model 23, and a history of the actual information of the observed person 2 corresponding to the estimated blood glucose level (actual measured blood glucose level), and the server control unit 62 may calculate the difference information 22 from the histories of both.

[0091] (Method for estimating the state) The state estimation device 1 of the state estimation system 200 according to the third embodiment of this disclosure can perform a state estimation method for the observed person 2, as shown in Figure 18. Figure 18 is a flowchart showing an example of a state estimation method by the state estimation device 1 according to the third embodiment of this disclosure.

[0092] First, when the power to the state estimation device 1 is turned ON and the output of the estimated result information 21 of the state of the person being observed 2 is instructed, the estimation unit 30 of the state estimation device 1 acquires the estimated material information 20 (blood flow information) of the person being observed 2 via the input unit 10 (step S31).

[0093] Furthermore, the estimation unit 30 obtains the observed person identification information 25 of the observed person 2 from the personal authentication server 40 (step S32). The estimation unit 30 transmits the observed person identification information 25 obtained in step S32 to the information management server 60 and requests differential information 22 corresponding to the observed person identification information 25. In response to this request from the state estimation device 1, the differential information 22 is transmitted from the information management server 60. In this way, the estimation unit 30 obtains the differential information 22 from the information management server 60 (step S33). Note that steps S32 and S33 may be executed before step S31.

[0094] The processes from step S34 to step S37 shown in Figure 18 are the same as the processes from step S13 to step S16 shown in Figure 2, so their explanation will be omitted.

[0095] As described above, the state estimation device 1, with its estimation unit 30 in the control unit 11, can generate estimation result information 21 based on the estimated material information 20 acquired from the observed person 2, taking into account the difference information 22.

[0096] Therefore, when the state estimation device 1 uses the estimation model 23 to estimate the state of the observed person 2, it is possible to generate corrected estimation result information 21 that is appropriate even if the estimation result is always larger or smaller than the actual information that represents the actual state. Therefore, even if there are individual differences in the estimation results using the estimation model 23 for each observed person 2, accurate estimation result information 21 can be obtained.

[0097] Therefore, the state estimation device 1 can accurately estimate the state of the observed person 2.

[0098] (Fourth Embodiment) Next, with reference to Figure 19, a state estimation system 300 according to the fourth embodiment of this disclosure will be described. Figure 19 is a block diagram showing an example of the main components of the state estimation system 300 according to the fourth embodiment of this disclosure.

[0099] The state estimation system 300 according to the fourth embodiment differs from the state estimation system 200 according to the third embodiment in that the server storage unit 63 further stores the estimated material information history 80, and the server control unit 62 further includes a differential information update determination unit 71 and a differential information update unit 72 as functional blocks. It also differs in that when the state estimation device 1 requests differential information 22 from the information management server 60, it transmits the estimated material information 20 received via the input unit 10 to the information management server 60. In all other respects, the state estimation system 300 according to the fourth embodiment is the same as the state estimation system 200 according to the third embodiment, so the same reference numerals are used for similar components, and their descriptions are omitted.

[0100] The estimated material information history 80 stored in the server storage unit 63 is information that shows the history of estimated material information 20 received from the state estimation device 1.

[0101] The differential information update determination unit 71 refers to the estimated material information history 80 stored in the server storage unit 63 to determine whether or not to update the differential information 22. If it determines that an update of the differential information 22 is necessary, it controls the differential information update unit 72 to update the differential information 22. The differential information update determination unit 71 and the differential information update unit 72 can be implemented, for example, if the control unit 11 is a CPU, by the CPU reading a program stored in the server storage unit 63 into a memory (not shown) and executing it.

[0102] For example, when the differential information update determination unit 71 receives pulse information having the pulse waveform shown in Figure 20 as estimated material information 20 from the state estimation device 1, it can determine whether or not to update the differential information 22 based on the received pulse information as follows. Figure 20 is a graph showing an example of pulse information obtained by dividing the blood flow information obtained from the observed person 2 into individual beats.

[0103] First, let us assume that the rising angle θ of the pulse waveform from the start point to the first maximum point in the pulse information shown in Figure 20 is an important feature for estimating the blood glucose level of subject 2. As shown in Figure 20, if the time interval from the start point to the maximum point is α and the difference in detected values ​​between the start point and the maximum point is β, then the rising angle θ can be evaluated by tanθ = β ÷ α. Here, let tanθ be the feature tθ.

[0104] The differential information update determination unit 71 stores the feature quantity tθ, which is obtained based on the pulse information of the blood flow information received from the state estimation device 1 along with the request for differential information 22, in the server storage unit 63 in association with the differential information 22.

[0105] In this manner, each time the state estimation device 1 estimates the blood glucose level of the observed person 2, the information management server 60 is configured to store the feature quantity tθ in the server storage unit 63. Therefore, the information management server 60 can grasp the time-series changes of the feature quantity tθ up to that point. Accordingly, when estimating the blood glucose level of the observed person 2, the differential information update determination unit 71 compares the feature quantity tθ obtained based on the blood flow information received from the state estimation device 1 with the time-series changes of past feature quantity tθ. The differential information update determination unit 71 then determines that the trend of the blood glucose level of the observed person 2 has changed, and that it is necessary to update the differential information 22 corresponding to the observed person 2. The predetermined range can be a range of ±10% of the average value of past feature quantity tθ. Furthermore, the predetermined number of times may be, for example, four or more times relative to the most recent five estimations (a probability of 80% or more in the most recent five estimations).

[0106] If the differential information update determination unit 71 determines that the differential information 22 needs to be updated, the differential information update unit 72 updates the differential information 22.

[0107] The update of the differential information 22 by the differential information update unit 72 may, for example, be performed by receiving input of differential information for update and updating the differential information 22 stored in the server storage unit 63, similar to the differential information update unit 31 included in the state estimation device 1 in the state estimation system 100 according to the first modified example of the second embodiment shown in Figure 14.

[0108] Alternatively, the differential information update unit 72 receives the estimation result estimated by the estimation model 23 based on the estimated material information 20 (blood flow information) received from the state estimation device 1, and calculates the difference between this received estimation result and the measured blood glucose level. The differential information update unit 72 may then be configured to update the differential information 22 with the newly calculated difference.

[0109] Furthermore, if the measured blood glucose values ​​stored in the server storage unit 63 have not been updated for a long period of time, the server control unit 62 may be configured to display a prompt to the operator of the state estimation system 300 to input the measured blood glucose values ​​when the differential information update determination unit 71 determines that the differential information 22 needs to be updated.

[0110] In the above explanation, the differential information update determination unit 71 used the rising angle θ of the pulse waveform of the blood flow information as an example of information for determining whether or not to update the differential information 22. However, the information for determining whether or not to update the differential information 22 is not limited to the rising angle θ of the pulse waveform. It may be another feature quantity obtained from the pulse information, or a feature quantity obtained from the first derivative pulse information obtained by first derivative of the pulse information, or a second derivative pulse information obtained by second derivative of the pulse information. Alternatively, it may be a combination of multiple types of feature quantities.

[0111] In the above explanation, the state estimation device 1 was described using an example configuration in which it receives blood flow information from an external source as estimated material information 20. However, the estimated material information 20 received from an external source may include other types of information in addition to blood flow information.

[0112] In this configuration, which receives multiple types of estimated material information 20 from an external source to estimate the state of the observed person 2, if the feature quantity of at least one piece of information included in the received estimated material information 20 differs significantly from the time-series changes of past feature quantities, the difference information update determination unit 71 determines that it is necessary to update the difference information 22 corresponding to the observed person 2.

[0113] (Method for estimating the state) The state estimation device 1 included in the state estimation system 300 according to the fourth embodiment of this disclosure can perform a state estimation method for the observed person 2, as shown in Figure 21. Figure 21 is a flowchart showing an example of a state estimation method by the state estimation device 1 according to the fourth embodiment of this disclosure.

[0114] First, when the power to the state estimation device 1 is turned ON and the output of the estimated result information 21 of the state of the observed person 2 is instructed, the estimation unit 30 of the state estimation device 1 acquires the estimated material information 20 (blood flow information) of the observed person 2 via the input unit 10 (step S41).

[0115] Furthermore, the estimation unit 30 obtains the subject identification information 25 of the subject 2 from the personal authentication server 40 (step S42). The estimation unit 30 transmits the estimated material information 20 (blood flow information) obtained in step S41, together with the subject identification information 25 obtained in step S42, to the information management server 60 (step S43).

[0116] Furthermore, the estimation unit 30 requests differential information 22 corresponding to the observed person identification information 25. In response to this request from the state estimation device 1, the differential information 22 is transmitted from the information management server 60. In this way, the estimation unit 30 obtains the differential information 22 from the information management server 60 (step S44). Note that step S42 may be executed before step S41, and the order of steps S41 and S42 is arbitrary.

[0117] The processes from step S45 to step S48 shown in Figure 21 are the same as the processes from step S13 to step S16 shown in Figure 2, so their explanation will be omitted.

[0118] As described above, the state estimation device 1, with its estimation unit 30 in the control unit 11, can generate estimation result information 21 based on the estimated material information 20 acquired from the observed person 2, taking into account the difference information 22.

[0119] Therefore, when the state estimation device 1 uses the estimation model 23 to estimate the state of the observed person 2, it is possible to generate corrected estimation result information 21 that is appropriate even if the estimation result is always larger or smaller than the actual information that represents the actual state. Therefore, even if there are individual differences in the estimation results using the estimation model 23 for each observed person 2, accurate estimation result information 21 can be obtained.

[0120] Therefore, the state estimation device 1 can accurately estimate the state of the observed person 2.

[0121] In the state estimation system 300 according to the fourth embodiment, the information management server 60 stores the differential information 22 and the estimated material information history 80 in the server storage unit 63, and the server control unit 62 determines whether or not to update the differential information 22. However, the control unit 11 of the state estimation device 1 may be configured to determine whether or not to update the differential information 22. In this configuration, the server control unit 62 of the information management server 60 does not include the differential information update determination unit 71 as a functional block, while the control unit 11 of the state estimation device 1 further includes the differential information update determination unit 71 as a functional block.

[0122] Then, in the state estimation device 1, the control unit 11 calculates the feature quantity tθ based on the blood flow information received via the input unit 10. The control unit 11 also obtains information on the time-series changes of past feature quantity tθ, which is stored in association with the difference information 22 of the observed person 2, from the information management server 60, and compares it with the feature quantity tθ. If the calculated feature quantity tθ falls outside a predetermined range based on the time-series changes of past feature quantity tθ, or if a feature quantity tθ outside the predetermined range is obtained a predetermined number of times, the control unit 11 determines that the trend of the observed person 2's blood glucose level has changed and determines that it is necessary to update the difference information 22 corresponding to the observed person 2.

[0123] When the control unit 11 determines that the differential information 22 needs to be updated, it controls the information management server 60 to update the differential information 22 by sending instruction information to the information management server 60 to instruct it to update the differential information 22. In the information management server 60, the differential information update unit 72 of the server control unit 62 updates the differential information 22 in accordance with the instruction information sent from the state estimation device 1. [Explanation of Symbols]

[0124] 1. State Estimation Device 2 Observed person 10 Input section 11 Control Unit 12 Output section 13 Storage section 14. Differential Information Input Section 20 Estimated Material Information 21 Estimated Result Information 22 Difference Information 23 Estimated Models 24. Modified Estimation Model 24a First Modified Estimation Model 24b Second Modified Estimation Model 25. Identification information of the person being observed 30 Estimation part 31 Difference information update section 40 Personal Authentication Server 60 Information Management Server 61 Server Input Section 62 Server Control Unit 63 Server Storage 70 Differential Information Recording Unit 71 Difference information update judgment unit 72 Difference information update section 80 Estimated material information history 100 State Estimation Systems 200 State Estimation Systems 300 State Estimation Systems

Claims

1. An input unit that accepts input of estimation material information used to estimate the state of the person being observed, A control unit that generates estimated information by estimating the state of the observed person using an estimation model based on the estimated material information received by the input unit, The system includes an output unit that outputs the estimated information generated by the control unit, The control unit acquires difference information that shows the trend of the magnitude relationship between the estimated history information and the actual information, which is derived from the difference between the estimated history information obtained in advance using the estimation model based on the past estimated material information of the observed person and the actual information showing the actual state of the observed person. If it is determined that the difference information tends to exceed a predetermined range, the estimation model is changed to a first modified estimation model that yields a smaller value than the estimation result estimated using the estimation model based on the estimated material information, and the estimation information is generated using the first modified estimation model. If it is determined that the difference information tends to fall below a predetermined range, the state estimation device changes the estimation model to a second modified estimation model that yields a larger value than the estimation result estimated using the estimation model based on the estimated material information, and generates the estimation information using the second modified estimation model.

2. The aforementioned difference information is associated with the identification information of each of the multiple observed individuals. The state estimation device according to claim 1, wherein the control unit receives identification information of the person being observed and acquires the difference information corresponding to the identification information.

3. A storage device that stores the aforementioned differential information, It includes a differential information input unit that accepts differential information for updates, The control unit updates the differential information stored in the storage device using the differential information for updating received by the differential information input unit. The state estimation device according to claim 1.

4. An input unit that accepts input of estimation material information used to estimate the state of the person being observed, A control unit that generates estimated information by estimating the state of the observed person using an estimation model based on the estimated material information received by the input unit, The system includes an output unit that outputs the estimated information generated by the control unit, The control unit acquires difference information indicating the difference between estimated history information obtained in advance using the estimation model based on the past estimated material information of the observed person and actual information indicating the actual state of the observed person, and if it is determined that the difference information is outside a predetermined range, it changes the estimation model to a modified estimation model that has been modified according to the difference information and generates the estimation information, the state estimation device The control unit extracts a first feature from the estimated material information received by the input unit, compares the extracted first feature with a second feature previously extracted from the estimated material information received by the input unit, and controls the unit to update the difference information if the first feature falls outside a predetermined range based on the second feature. State estimation device.

5. The state estimation device according to any one of claims 1 to 4, wherein the state of the observed person is the biological information of the observed person.

6. The state estimation device according to claim 2, Includes a personal authentication server that authenticates the observed person and transmits the identification information of the observed person who has successfully been authenticated to the state estimation device. State estimation system.

7. The information management server further includes a server storage device that stores the difference information in association with the identification information of each of the observed persons, and when it receives the identification information from the state estimation device, it transmits the difference information corresponding to the identification information to the state estimation device. The state estimation system according to claim 6.

8. A step of receiving input of estimation material information to be used to estimate the state of the person being observed, The step of receiving the input of the estimated material information includes a step of generating estimated information by using an estimation model to estimate the state of the observed person based on the received estimated material information, The step of generating the estimated information includes a step of outputting the generated estimated information, In the step of generating the estimation information, difference information is obtained that shows the trend of the magnitude relationship between the estimation history information and the actual information, derived from the difference between the estimation history information obtained in advance using the estimation model based on the past estimation material information of the observed person and the actual information showing the actual state of the observed person. If it is determined that the difference information tends to exceed a predetermined range, the estimation model is changed to a first modified estimation model that yields a smaller value than the estimation result estimated using the estimation model based on the estimated material information, and the estimation information is generated using the first modified estimation model. If it is determined that the difference information tends to fall below a predetermined range, the estimation model is changed to a second modified estimation model that yields a larger value than the estimation result estimated using the estimation model based on the estimated material information, and the estimation information is generated using the second modified estimation model, the process includes these steps. A control method for a state estimation device.

Images