State inference device and state inference method

Abstract

The present invention relates to a state inference device for inferring a relaxation level indicating the degree of relaxation of a user, comprising: an indicator value calculation unit (102) for calculating multiple types of indicator values ​​using biological information of the user; and an inference unit (109) for inferring the user's relaxation level using multiple relaxation levels specified based on the multiple types of indicator values ​​and the change trends of the multiple types of indicator values.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application is based on Japanese Patent Application No. 2019-112954 filed on June 18, 2019, the contents of which are incorporated herein by reference. Technical Field

[0003] The present disclosure relates to a state inference device and a state inference method for inferring a user's degree of relaxation. Background Art

[0004] Technology has been developed to estimate the user's relaxation level using the user's biological information. For example, Patent Document 1 discloses a relaxation level determination method that measures the time required for the user's heart rate to decrease to a predetermined heart rate and determines the user's relaxation level based on the time.

[0005] Patent Document 1: Japanese Patent Application Laid-Open No. 9-70399

[0006] In the relaxation degree determination method described in Patent Document 1, only the time required for the heart rate to decrease to a predetermined heart rate is used as an indicator to determine the user's relaxation degree, and therefore, it may not be possible to estimate the degree of relaxation in detail. Summary of the invention

[0007] An object of the present disclosure is to provide a state estimating device and a state estimating method that can estimate the degree of relaxation of a user in detail.

[0008] A state inference device involved in one embodiment of the present disclosure infers a relaxation level indicating the degree of relaxation of a user, and comprises: an indicator value calculation unit, which uses the user's biological information to calculate multiple types of indicator values; and an inference unit, which infers the user's relaxation level using multiple relaxation levels specified based on the multiple types of indicator values ​​and the change trends of the multiple types of indicator values.

[0009] A state inference method involved in one embodiment of the present disclosure is a method performed by a state inference device for inferring a relaxation level representing the degree of relaxation of a user, comprising: using the user's biological information to calculate multiple types of indicator values; and inferring the user's relaxation level using multiple relaxation levels specified based on the multiple types of indicator values ​​and the change trends of the multiple types of indicator values.

[0010] In the state estimation device and state estimation method of the present disclosure, the relaxation level of the user can be estimated using a plurality of relaxation levels defined based on a plurality of types of index values ​​and their changing tendencies, and thus the degree of relaxation of the user can be estimated in detail. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The objects, features and advantages of the present invention will become more apparent from the following detailed description with reference to the accompanying drawings.

[0012] Figure 1 is a diagram showing a state estimation system according to a first embodiment of the present disclosure.

[0013] Figure 2 This is a graph showing the heart rate, HF integral value, and α wave time content calculated from the electrocardiographic signal and α wave acquired in 1 minute.

[0014] Figure 3 is a diagram showing an example of a model for clustering index values ​​and differential values ​​based on the index values.

[0015] Figure 4 is a diagram showing an example of a relaxation level correspondence table.

[0016] Figure 5 is a diagram showing an example of a user's relaxation level estimated by the state estimation device according to the first embodiment of the present disclosure,

[0017] Figure 6 is a flowchart showing the processing performed by the state estimation system of the first embodiment.

[0018] Figure 7 is a flowchart showing the processing performed by the state estimation system of the first embodiment.

[0019] Figure 8 is a diagram showing an example of a relaxation level estimation table according to the second embodiment.

[0020] Fig. 9 is a diagram showing another example of a model for clustering index values ​​and differential values ​​based on the index values.

[0021] Fig.10 is a diagram showing another example of the relaxation level correspondence table,

[0022] Fig.11 is a diagram showing another example of a model for clustering index values ​​and differential values ​​based on the index values.

[0023] Fig.12 is a diagram showing another example of the relaxation level correspondence table.

[0024] Fig.13 is a diagram showing another example of a model for clustering index values ​​and differential values ​​based on the index values.

[0025] Fig.14 This is a diagram showing still another example of the relaxation level correspondence table. DETAILED DESCRIPTION

[0026] (First embodiment)

[0027] The first embodiment of the present disclosure will be described with reference to the accompanying drawings. Figure 1 As shown, the state estimation system 1 according to the first embodiment includes a state estimation device 10 , an identification information acquisition device 20 , an electrocardiogram acquisition device 30 , and an electroencephalogram acquisition device 40 .

[0028] The state estimation device 10 is a device that estimates a relaxation level indicating the degree of relaxation of the user. The state estimation device 10 includes a microcontroller, various electronic circuits, and a communication interface.

[0029] The microcontroller is a device that controls the operation of the state estimation device 10, and includes a computing device, a volatile storage device, and a non-volatile storage device. The computing device is a computing device such as a CPU that can execute various programs. The computing device executes the state estimation method of the present disclosure by executing the program stored in the non-volatile storage device.

[0030] The communication interface is an interface for transmitting and receiving various data between the state estimation device 10 and the identification information acquisition device 20 , the electrocardiogram acquisition device 30 , the electroencephalogram acquisition device 40 , and other devices (not shown).

[0031] The ECG acquisition device 30 is a device that acquires an ECG signal representing the action potential generated by the user's cardiomyocytes. Specific examples of the ECG acquisition device 30 include electrodes provided on the handrails of a vehicle, wearable devices that can be worn by a user, and the like. The ECG acquisition device 30 can acquire an ECG signal from the user while in contact with the user's body surface, and send the ECG signal to the state estimation device 10. In addition, the ECG acquisition device 30 can also acquire an ECG signal without being in contact with the user's body surface.

[0032] The brain wave acquisition device 40 is a device that acquires the electrical signal of the brain wave generated by the nerve cells of the user's brain (hereinafter referred to as "brain wave signal"). As a specific example of the brain wave acquisition device 40, for example, an electrode provided on the headrest of the seat of the vehicle in which the user rides, a helmet-type brain wave measuring device, etc. can be cited. The brain wave acquisition device 40 can acquire the brain wave signal in a state of contact with the user's scalp and provide it to the state estimation device 10. In addition, the brain wave acquisition device 40 can also acquire the brain wave signal in a state of not contacting the user's scalp.

[0033] The identification information acquisition device 20 is a device that acquires the identification information of the user. A specific example of the identification information acquisition device 20 is a touch panel input device. After acquiring the identification information of the user, the identification information acquisition device 20 provides the identification information to the state estimation device 10.

[0034] like Figure 1 As shown, the state inference device 10 includes: a system control unit 100, an α wave detection unit 101, an index value calculation unit 102, a standardization processing unit 103, a standardization database (DB) 104, a differential value calculation unit 105, a discretization processing unit 106, a cluster processing unit 107, a relaxation level correspondence table database (DB) 108, an inference unit 109, a notification unit 110 and a relaxation guidance processing unit 111.

[0035] The system control unit 100 is a functional unit that controls the activation and termination of the state estimation system 1. The α wave detection unit 101 is a semiconductor circuit that detects α waves from the electroencephalogram signal supplied from the electroencephalogram acquisition device 40.

[0036] The index value calculation unit 102 is a functional unit that calculates the index value based on the user's electrocardiogram signal and α wave. The index values ​​include: (1) the user's heart rate, (2) the integral value of the high-frequency component obtained by frequency analysis of the user's heartbeat fluctuations (hereinafter referred to as "HF integral value") (msec 2 ), and (3) an α wave time content ratio which represents the ratio of the α wave generation time to the unit time.

[0037] The index value calculation unit 102 can calculate the heart rate (bpm) by counting the electrocardiographic waveforms included in the electrocardiographic signal of a unit time (eg, 10 seconds) and converting the count into a heart rate per minute.

[0038] In addition, the index value calculation unit 102 performs a fast Fourier transform on the time intervals of the R waves of a plurality of electrocardiographic waveforms included in a unit time (e.g., 10 seconds, etc.) Furthermore, the index value calculation unit 102 integrates the high-frequency components (e.g., 0.15 Hz to 0.4 Hz, etc.) of the frequency data obtained by the fast Fourier transform, thereby being able to calculate the HF integral value.

[0039] Furthermore, the index value calculation unit 102 can calculate the α wave time content rate by measuring the α wave generation time within a unit time (eg, 10 seconds) and calculating the ratio of the α wave generation time to the unit time.

[0040] The normalization processing unit 103 is a functional unit that normalizes the heart rate, HF integrated value, and α wave time content rate, which are index values, using the following formula 1.

[0041] [Formula 1]

[0042]

[0043] Here, x represents each index value, A represents the average value of each index value, and σ represents the standard deviation of each index value.

[0044] When the average value and standard deviation of each index value of the user calculated in advance are stored in the standardization DB 104, the standardization processing unit 103 uses the average value and standard deviation to standardize the heart rate, HF integral value, and α wave time content rate. When the average value and standard deviation of each index value of the user are not stored in the standardization DB 104, the standardization processing unit 103 uses the average value and standard deviation of each common index value to standardize the heart rate, HF integral value, and α wave time content rate.

[0045] The differential value calculation unit 105 is a functional unit that calculates a differential value indicating a change tendency of each index value. Figure 2 , the heart rate, HF integral value, and α wave time content ratio calculated from the electrocardiographic signal and α wave acquired in 1 minute are shown in FIG. In this example, the heart rate, HF integral value, and α wave time content ratio are calculated with the unit time being 10 seconds. The differential value calculation unit 105 can calculate the differential value of the index value by differentiating the formula representing the straight line defined by two index values ​​continuous in time.

[0046] The discretization processing unit 106 is a functional unit that discretizes each normalized index value and a differential value based on the index value. The discretization processing unit 106 can discretize each index value and a differential value based on the index value using unsupervised morphological analysis based on a Bayesian hierarchical language model.

[0047] For example, in Figure 2 In the example shown, the heart rate calculated from the electrocardiogram signal acquired during the period from 0 to 30 seconds is constant. The heart rate calculated from the electrocardiogram signal acquired during the period from 31 to 60 seconds decreases. Figure 2 In the example shown, the heart rate calculated from the electrocardiographic signal acquired during the period of 31 seconds to 40 seconds corresponds to the change point. The change point is a point at which the normalized value or differential value of at least one index value changes.

[0048] In more detail, two differential values ​​that are continuous in time can be compared, and the point where the difference between these differential values ​​is greater than a predetermined threshold value can be regarded as a change point. The threshold value can be set for each type of index value. In addition, the difference between the first standardized value and the middle standardized value, and the difference between the middle standardized value and the last standardized value among three standardized values ​​that are continuous in time can be compared. If these differences are different, the last standardized value can be regarded as a change point. In addition, even if these differences are slightly different, that is, even if an increase or decrease in the standardized value that is close to monotonic occurs, the last standardized value can become a change point.

[0049] Therefore, the discretization processing unit 106 can discretize these index values ​​and the differential values ​​based on the index values ​​based on the change point of the heart rate, and divide them into two discretization groups. That is, the discretization based on the change point is to divide the front and back of the change point into respective groups. In the same discretization group, the change trends of the heart rate, the HF integral value, and the α wave time content rate are respectively the same. In addition, when there is no change point, the index value is not discretized.

[0050] The clustering processing unit 107 is a functional unit that clusters the normalized index values ​​and the differential values ​​of the index values ​​and outputs a value indicating the cluster after the data are classified as a clustering result. The clustering processing unit 107 can classify the data using a hidden Markov model.

[0051] Figure 3 The model in which each index value and the differential value of each index value after normalization are classified into seven clusters is shown. These clusters correspond to the user's relaxation levels 0 to 5. For relaxation levels 0 to 5, the higher the value, the stronger the degree of relaxation.

[0052] Specifically, relaxation level 0 indicates that the user is in an awake state. Relaxation levels 1 to 3 indicate that the user is in a shallow state of relaxation. In addition, the index value and differential value of relaxation level 3 (majority) indicate the results observed by the majority of subjects. The index value and differential value of relaxation level 3 (minority) indicate the results observed by a minority of subjects. Relaxation level 4 indicates that the user is in a deep state of relaxation (i.e., a drowsy state). Relaxation level 5 indicates that the user is in a state of falling asleep.

[0053] For example, it is assumed that a high heart rate, a low α wave time content rate and a low HF integral value, and a differential value indicating that there is no tendency of change in these index values ​​are input to the cluster processing unit 107. In this case, the cluster processing unit 107 classifies these data into a cluster corresponding to relaxation level 0, and outputs a value indicating the cluster as a cluster result.

[0054] The inference unit 109 is a functional unit for inferring the user's relaxation level. Figure 4 The relaxation level correspondence table shown in FIG. 1 is used to infer the relaxation level of the user using the clustering result output by the clustering processing unit 107. The clustering result is associated with the relaxation level and registered in the relaxation level correspondence table. For example, when the clustering result is a value indicating a cluster corresponding to relaxation level 0, the inference unit 109 infers that the relaxation level of the user is 0.

[0055] Figure 5 It represents the relaxation level of the user estimated by the inference unit 109 based on the clustering result. Figure 5In the example shown, the user's state is initially at relaxation level 0 (RL0), and then sequentially shifts to relaxation level 1 (RL1) to relaxation level 5 (RL5).

[0056] In relaxation level 1 (RL1), the heart rate decreases from the high area, and the α wave time content rate increases slightly from the low area. In relaxation level 2 (RL2), the heart rate decreases in the middle area, and the α wave time content rate increases slightly from the low area to the middle area. In relaxation level 3 (RL3), the heart rate decreases slightly in the low area, the α wave time content rate increases slightly in the middle area, and the HF integral value increases slightly in the low area. In relaxation level 4 (RL4), the heart rate decreases slightly in the low area, the α wave time content rate increases sharply in the middle area, and the HF integral value increases slightly in the low area. In relaxation level 5 (RL5), the heart rate decreases slightly in the low area, the α wave time content rate decreases sharply from the high area to the middle area, and the HF integral value increases sharply from the middle area to the high area. In this way, by adopting three types of index values, the user's relaxation level can be classified and estimated in detail.

[0057] In addition, the user's status is not necessarily limited to the order of transition. For example, there is a case where the transition is in the order of relaxation level 0, relaxation level 2, relaxation level 3 (majority), relaxation level 4, and relaxation level 5 (RL5). In addition, there is a case where the transition is in the order of relaxation level 0, relaxation level 1, relaxation level 3 (minority), relaxation level 4, and relaxation level 5 (RL5).

[0058] The notification unit 110 is a functional unit for notifying the user of the estimated relaxation level. For example, the notification unit 110 can notify the user of the relaxation level by causing an image representing the user's relaxation level to be displayed on a display device. In addition, the notification unit 110 can notify the user of the relaxation level by causing a sound reproduction device to reproduce a sound representing the user's relaxation level. Furthermore, the notification unit 110 can notify the user of the relaxation level by changing the illumination of the lighting device to an illumination corresponding to the user's relaxation level.

[0059] The relaxation guidance processing unit 111 is a functional unit that causes other devices to execute a process for improving the user's relaxation level (hereinafter referred to as “relaxation guidance process”). The relaxation guidance processing unit 111 includes a control unit 112 and a target level determination unit 113 .

[0060] The control unit 112 is a functional unit that controls other devices to perform relaxation guidance processing. Specifically, the control unit 112 can control air conditioning equipment and lighting devices and adjust indoor temperature, air flow, and illumination to improve the user's relaxation level. The control unit 112 can control the opening and closing of blackout curtains and curtains and adjust the indoor illumination to improve the user's relaxation level.

[0061] Furthermore, the control unit 112 can control the display device and the sound reproduction device to reproduce images and music that improve the user's relaxation level. For example, the control unit 112 can reproduce images and music that have a relaxing effect, or reproduce images and sounds that promote breathing techniques and mindfulness stress reduction techniques. Furthermore, the control unit 112 can control the sound reproduction device and adjust the volume of the sound reproduction device output so that the user's relaxation level is improved.

[0062] Furthermore, the control unit 112 can control a vibration device provided on a seat where the user sits to generate vibrations that improve the user's relaxation level, such as vibrations with a massage effect. Furthermore, the control unit 112 can control the seat where the user sits to change the angle to improve the user's relaxation level. Furthermore, the control unit 112 can control an aroma diffuser to release a fragrance that improves the user's relaxation level.

[0063] The target level determination unit 113 is a functional unit that calculates the user's relaxation level (hereinafter referred to as "target level") as a termination condition of the relaxation guidance process. The target level determination unit 113 can calculate the target level based on the user's schedule.

[0064] For example, when the user's next schedule is work or study, the target level can be set to any one of relaxation levels 1 to 3. When the user does not have a next schedule, the target level can be set to any one of relaxation levels 1 to 5. The target level determination unit 113 can determine the target level by acquiring the user's schedule stored in the user's smartphone, an external data server, or the like.

[0065] Next, refer to Figure 6 and Figure 7 , the processing performed by the state estimation system 1 is described. In S101, the system control unit 100 of the state estimation device 10 starts the state estimation system 1. In S102, the identification information acquisition device 20 acquires the identification information of the user whose relaxation level is to be estimated, and provides the identification information of the user to the state estimation device 10. In S103, the relaxation guidance processing unit 111 of the state estimation device 10 determines the target level.

[0066] In S104, the index value calculation unit 102 obtains the user's electrocardiographic signal and α wave for a predetermined period (e.g., 1 minute, etc.). In S105, the index value calculation unit 102 calculates the heart rate and HF integral value using the obtained electrocardiographic signal. In addition, the index value calculation unit 102 calculates the α wave time content rate using the obtained α wave.

[0067] In S106, the normalization processing unit 103 determines whether the average value and standard deviation of each index value calculated in advance for the user indicated by the user identification information are stored in the normalization database 104. If the average value and standard deviation of each index value of the user are stored (yes), the process branches to S107. In S107, the normalization processing unit 103 obtains the average value and standard deviation of each index value of the user from the normalization database 104.

[0068] If the average value and standard deviation of each index value of the user are not stored (No), the process branches to S108 . In S108 , the normalization processing unit 103 acquires the average value and standard deviation of each common index value from the normalization database 104 .

[0069] In S109, the standardization processing unit 103 uses the average value and standard deviation of each index value acquired in S107 or S108 to standardize the heart rate, HF integral value, and α wave time content rate. In S110, the differential value calculation unit 105 uses the heart rate, HF integral value, and α wave time content rate to calculate the respective differential values. In S111, the discretization processing unit 106 discretizes the standardized heart rate, HF integral value, and α wave time content rate, and their differential values. In addition, when the index value does not include a change point, these index values ​​are not discretized.

[0070] In S112, the clustering processing unit 107 clusters the index value and the differential value of the index value, and outputs the clustering result. In addition, when the index value and the differential value of the index value are discretized in S111, the clustering processing unit 107 clusters the index value and the differential value of the index value for each discretization group, and outputs the clustering result.

[0071] In S113, the inference unit 109 reads the relaxation level correspondence table from the relaxation level correspondence table DB 108. In S114, the inference unit 109 determines whether there is a cluster result in the relaxation level correspondence table. When there is no cluster result in the relaxation level correspondence table (No), the processing branch is S116. When there is a cluster result in the relaxation level correspondence table (Yes), the processing branch is S115.

[0072] In S115, the inference unit 109 determines whether the state transition of the user is abnormal based on the information indicating the abnormal state transition. For example, when the previous clustering result is relaxation level 0 (awake state), when the current clustering result is relaxation level 5 (sleeping state), the inference unit 109 can determine that the state transition of the user is abnormal.

[0073] If the user's state transition is abnormal (Yes), the process branches to S116. In S116, the inference unit 109 outputs a result indicating that the user's state inference is wrong. If the user's state transition is normal (No), the process branches to S117.

[0074] In S117 , the estimation unit 109 refers to the relaxation level correspondence table, identifies the relaxation level associated with the clustering result for each discretization group, and estimates the relaxation level of the user. In S118 , the notification unit 110 notifies the user of the estimated relaxation level.

[0075] In S119, the relaxation guidance processing unit 111 determines whether the inferred relaxation level is the same as the target level. In the case where the inferred relaxation level is different from the target level (no), the processing branch is S120. In S120, the relaxation guidance processing unit 111 uses other devices to perform relaxation guidance processing. In addition, in the case where the relaxation guidance processing has been performed, the relaxation guidance processing unit 111 continues the relaxation guidance processing.

[0076] If the estimated relaxation level is the same as the target level (Yes), the process branches to S121. In S121, the relaxation guidance processing unit 111 ends the relaxation guidance process. If the relaxation guidance process has not been started, the process of S121 is not executed.

[0077] In S122, the system control unit 100 determines whether to terminate the state estimation system 1. For example, when the vehicle equipped with the state estimation system 1 arrives at the destination, the system control unit 100 can determine that the state estimation system 1 should be terminated.

[0078] If the state estimation system 1 should not be terminated (No), the process returns to S104. If the state estimation system 1 should be terminated (Yes), the process branches to S123. In S123, the system control unit 100 terminates the state estimation system 1.

[0079] In this embodiment, the state estimation device 10 calculates the heart rate, HF integral value, and α wave time content rate as index values ​​using the electrocardiogram signal and α wave as the biological information of the user (S105). Furthermore, the state estimation device 10 estimates the relaxation level of the user using a plurality of relaxation levels defined based on the heart rate, HF integral value, α wave time content rate, and their changing trends (S117).

[0080] Thus, the state estimation device 10 can use Figure 3 By estimating the user's relaxation level based on the subdivided relaxation levels as shown, it is possible to estimate the user's relaxation level in detail.

[0081] Furthermore, experiments have confirmed that there are few individual differences in the changing trends of heart rate, HF integral value, and α wave time content. In this embodiment, by using such changing trends to estimate the user's relaxation level, it is possible to estimate the relaxation level with little influence of individual differences.

[0082] Then, the differential value calculation unit 105 of the state estimation device 10 calculates differential values ​​indicating the changing tendency of the heart rate, HF integral value, and α wave time content rate (S110). Next, the discretization processing unit 106 discretizes the heart rate, HF integral value, α wave time content rate, and their differential values ​​based on the changing points of the heart rate, HF integral value, and α wave time content rate (S111). Then, the clustering processing unit 107 clusters the discretized heart rate, HF integral value, α wave time content rate, and their differential values ​​for each discretization group (S112).

[0083] In the same discretization group, the change tendency of the index values ​​of the same type is the same. Figure 2 In the first discretization group shown, the heart rate, HF integral value, and α wave time content rate are constant. Therefore, their change tendency is no. In the subsequent discretization groups, the heart rate decreases, and the HF integral value and α wave time content rate are constant. Therefore, the change tendency of the heart rate is to decrease, and the change tendency of the HF integral value and α wave time content rate is no.

[0084] The clustering processing unit 107 performs clustering processing for each of the discretized groups. Figure 2 In the example shown, the cluster processing unit 107 clusters the leading discretization group and the subsequent discretization group separately. If the leading discretization group and the subsequent discretization group are clustered together, the standardized value of the heart rate with no change tendency and the standardized value of the heart rate with a decrease change tendency are used. That is, the standardized values ​​indicating different change tendencies are used for clustering for the same type of index value. Therefore, there is a possibility that the standardized values ​​that should be classified as different clusters are classified as the same cluster, and the reliability of the clustering result may be reduced.

[0085] In this embodiment, index values ​​of the same type are clustered using standardized values ​​that indicate the same tendency of change, so that standardized values ​​that should be classified into different clusters are not classified into the same cluster. Therefore, the reliability of the clustering result can be improved, and the accuracy of estimating the user's relaxation level can be improved.

[0086] Furthermore, when the average value and standard deviation of each index value of the user are stored in the standardized database 104, the standardization processing unit 103 uses the average value and standard deviation of each index value of the user to standardize the heart rate, HF integral value, and α wave time content rate (S109). Thus, the reliability of the standardized heart rate, HF integral value, and α wave time content rate can be improved. Furthermore, the clustering processing unit 107 clusters these heart rates, HF integral values, α wave time content rates, and their differential values. As a result, the reliability of the clustering result can be improved, and the accuracy of the estimation of the user's relaxation level can be improved.

[0087] Then, the relaxation guidance processing unit 111 causes other devices to execute a relaxation guidance process for improving the user's relaxation level (S120). Thus, the user's relaxation level can be improved.

[0088] Then, the notification unit 110 notifies the estimation unit 109 of the relaxation level of the user estimated (S118). Thus, the relaxation level of the user to be estimated can be grasped.

[0089] (Second embodiment)

[0090] Next, the second embodiment of the present disclosure will be described focusing on the differences from the first embodiment. In the second embodiment, the estimation unit 109 refers to the relaxation level estimation table and estimates the user's relaxation level using the normalized index value and the differential value based on the index value.

[0091] like Figure 8 As shown, in the relaxation level estimation table, each standardized index value and the differential value based on the index value are associated with the relaxation level and registered. For example, when the user's heart rate is high, the α wave time content rate and the HF integral value are low, and their differential values ​​are values ​​indicating that there is no tendency to change, the estimation unit 109 estimates that the user's relaxation level is 0.

[0092] In the second embodiment, the relaxation level of the user is estimated using the relaxation level estimation table, so that the relaxation level of the user can be estimated without performing discretization processing or clustering processing of the index value based on machine learning.

[0093] (Other Embodiments)

[0094] The present disclosure is not limited to the above-mentioned embodiments, and can be implemented with various modifications. For example, in the above-mentioned embodiments, the user's relaxation level is estimated using three index values, namely heart rate, α wave time content rate, and HF integral value, but in other embodiments, the relaxation level can be estimated using two of these index values.

[0095] For example, when the heart rate and the α wave time content ratio are used as index values, the clustering processing unit 107 Fig. 9 The model shown in FIG. 1 clusters the heart rate and the differential value based on the heart rate, and the α wave time content rate and the differential value based on the α wave time content rate. Then, the cluster processing unit 107 outputs the value indicating the cluster after classifying these data as the cluster result. Then, the inference unit 109 refers to Fig.10 The relaxation level correspondence table shown in the figure estimates the relaxation level of the user using the clustering result output by the clustering processing unit 107 .

[0096] When the heart rate and HF integral value are used as index values, the cluster processing unit 107 performs the clustering based on Fig.11 The clustering processing unit 107 clusters the heart rate and the differential value based on the heart rate, and the HF integral value and the differential value based on the HF integral value. Then, the clustering processing unit 107 outputs the value indicating the cluster after classifying these data as the clustering result. Then, the inference unit 109 refers to Fig.12 The relaxation level correspondence table shown in the figure estimates the relaxation level of the user using the clustering result output by the clustering processing unit 107 .

[0097] When the α wave time content rate and the HF integral value are used as index values, the cluster processing unit 107 Fig.13 The clustering processing unit 107 clusters the α wave time content rate and the differential value based on the α wave time content rate, and the HF integral value and the differential value based on the HF integral value. Then, the clustering processing unit 107 outputs the value indicating the cluster after classifying these data as the clustering result. Then, the inference unit 109 refers to Fig.14 The relaxation level correspondence table shown in the figure estimates the relaxation level of the user using the clustering result output by the clustering processing unit 107 .

[0098] In addition, in the above-mentioned embodiment, when the average value and standard deviation of each index value of the user are not stored in the standardized database 104, the standardization processing unit 103 uses the average value and standard deviation of the common index value to standardize each index value of the user. In other embodiments, the standardization processing unit 103 may calculate the average value and standard deviation based on the heart rate, HF integrated value, and α wave of the user calculated by the index value calculation unit 102, and use the average value and standard deviation to standardize each index value of the user.

[0099] Furthermore, in the above-mentioned embodiment, when it is determined that the inferred relaxation level is the same as the target level, the relaxation guidance processing unit 111 ends the relaxation guidance processing. In other embodiments, when the inferred relaxation level is the same as the target level, the relaxation guidance processing unit 111 may also cause other devices to perform processing for maintaining the user's relaxation level.

[0100] Furthermore, in other embodiments, in addition to heart rate, α wave time content and HF integral value, other biological information such as cerebral blood flow, respiratory rate, galvanic skin response, action potential of facial muscles and peripheral blood flow may be used as index values.

[0101] The control unit and method described in the present disclosure can be implemented by a special-purpose computer, which is manufactured by constituting a processor programmed to execute one or more specific functions installed in a computer program. In addition, the device and method described in the present disclosure can also be implemented by a dedicated hardware logic circuit. Moreover, the device and method described in the present disclosure can also be implemented by a combination of one or more special-purpose computers manufactured by constituting a processor that executes a computer program and one or more hardware logic circuits. The computer program can be stored as a command executed by a computer in a non-transferable tangible recording medium that can be read by the computer.

[0102] The flowchart or the processing of the flowchart described in the present application is composed of a plurality of steps (or parts thereof), and each step is expressed as S101, for example. In addition, each step can be divided into a plurality of sub-steps, and a plurality of steps can be combined into one step.

[0103] The above is an example of the implementation, structure, and form of the state inference device and state inference method involved in one embodiment of the present disclosure, but the implementation, structure, and form involved in the present disclosure are not limited to the above-mentioned implementations, structures, and forms. For example, the implementation, structure, and form obtained by appropriately combining the technical parts disclosed in different implementations, structures, and forms are also included in the scope of the implementation, structure, and form involved in the present disclosure.

Claims

1. A state estimation device for estimating a relaxation level indicating a degree of relaxation of a user, wherein: The state estimation device comprises: an index value calculation unit that calculates a plurality of types of index values ​​using the biological information of the user; and an estimating unit that estimates the relaxation level of the user using a plurality of relaxation levels that are defined based on the plurality of types of index values ​​and the tendency of changes in the plurality of types of index values, The index value calculation unit calculates, as the index value, a heart rate of the user and an α wave time content ratio indicating a ratio of α wave generation time to unit time, the estimating unit estimates the relaxation level of the user using a plurality of relaxation levels defined based on the heart rate calculated as the index value, the tendency of change in the heart rate, the α wave time content rate calculated as the index value, and the tendency of change in the α wave time content rate, One of the relaxation levels is a shallow relaxation level defined based on the heart rate being in a high to medium range and the change tendency of the heart rate showing a sharp decrease, and the α wave time content being in a low range and the change tendency of the α wave time content showing a slight increase, Another of the relaxation levels is a deep relaxation level defined based on the heart rate being in a medium to low range and the change tendency of the heart rate indicating a slight decrease, while the α wave time content rate is in a high range and the change tendency of the α wave time content rate indicating a sharp increase, Another of the relaxation levels is a sleeping level defined based on the heart rate being in a low range and the change tendency of the heart rate showing a slight decrease, and the α wave time content being in an intermediate to low range and the change tendency of the α wave time content showing a sharp decrease.

2. The state estimation device according to claim 1, wherein: The index value calculation unit calculates, as the index value, an integrated value of a high-frequency component obtained by frequency analyzing the heartbeat variation of the user in addition to the heart rate and the α wave time content rate. The inference unit infers the relaxation level of the user using a plurality of relaxation levels specified based on the heart rate calculated as the index value, the changing tendency of the heart rate, the α wave time content calculated as the index value, the changing tendency of the α wave time content, the integrated value of the high frequency component calculated as the index value, and the changing tendency of the integrated value of the high frequency component, One of the relaxation levels is a first shallow relaxation level defined based on the heart rate being in a high to medium range and the tendency of change of the heart rate showing a sharp decrease, the α wave time content being in a low range and the tendency of change of the α wave time content showing a slight increase, and the integrated value of the high frequency component being in a low range and the tendency of change of the integrated value of the high frequency component showing no change, Another of the relaxation levels is a second shallow relaxation level that is defined as a relaxation level deeper than the first shallow relaxation level, based on the heart rate being in the middle to low range and the tendency of the heart rate change indicating a slight decrease, the alpha wave time content being in the low to middle range and the tendency of the alpha wave time content changing indicating a slight increase, and the integral value of the high-frequency component being in the middle to high range and the tendency of the integral value of the high-frequency component changing indicating a slight increase, Another one of the relaxation levels is a deep relaxation level defined based on the heart rate being in the middle to low range and the change tendency of the heart rate indicating a slight decrease, the alpha wave time content being in the high range and the change tendency of the alpha wave time content indicating a sharp increase, and the integrated value of the high frequency component being in the middle to high range and the change tendency of the integrated value of the high frequency component indicating a slight increase, Another one of the relaxation levels is a sleeping level determined based on the heart rate being in a low range and the tendency of change of the heart rate indicating a slight decrease, the α wave time content being in a middle range to a low range and the tendency of change of the α wave time content indicating a sharp decrease, and the integral value of the high-frequency component being in a high range and the tendency of change of the integral value of the high-frequency component indicating a sharp increase.

3. The state estimation device according to claim 1 or 2, wherein: Also available: a differential value calculation unit that calculates a differential value indicating a change tendency of the index value; a discretization processing unit that discretizes the index value and the differential value based on the change point of the index value; as well as a clustering processing unit that clusters the index value and the differential value according to each of the discretized index values ​​and each of the differential values, and outputs a clustering result; The inference unit refers to a relaxation level correspondence table that associates a plurality of clustering results output by the clustering processing unit with a plurality of relaxation levels, and infers the relaxation level associated with the clustering result output by the clustering processing unit as the relaxation level of the user.

4. The state estimation device according to claim 3, wherein: The state estimation device further includes a normalization processing unit that normalizes the index value of the user using an average value and a standard deviation of the index value. The differential value calculation unit calculates a differential value indicating a change tendency of the normalized index value. The discretization processing unit discretizes the normalized index value and the differential value based on a change point of the normalized index value.

5. The state estimation device according to claim 1 or 2, wherein: The state estimation device further includes a normalization processing unit for normalizing the index value. The inference unit refers to a relaxation level inference table that associates the standardized indicator value output by the standardization processing unit and the change trend of the indicator value with the relaxation level of the user, and infers the relaxation level associated with the standardized indicator value output by the standardization processing unit as the relaxation level of the user.

6. The state estimation device according to claim 1 or 2, wherein: The state estimation device further includes a relaxation guidance processing unit that causes another device to execute a relaxation guidance process for improving the relaxation level of the user.

7. The state estimation device according to claim 6, wherein: The relaxation guidance processing unit causes another device to execute the relaxation guidance processing until the relaxation level of the user estimated by the estimation unit reaches a target relaxation level.

8. The state estimation device according to claim 1 or 2, wherein: The state estimation device further includes a notification unit configured to notify the user of the relaxation level estimated by the estimation unit.

9. A state estimation device for estimating a relaxation level indicating a degree of relaxation of a user, wherein: The state estimation device comprises: an index value calculation unit that calculates a plurality of types of index values ​​using the biological information of the user; and an estimating unit that estimates the relaxation level of the user using a plurality of relaxation levels that are defined based on the plurality of types of index values ​​and the tendency of changes in the plurality of types of index values, The index value calculation unit calculates, as the index value, an α wave time content rate indicating a ratio of α wave generation time to unit time and an integrated value of high frequency components obtained by frequency analysis of the user's heartbeat fluctuations, the estimating unit estimates the relaxation level of the user using a plurality of relaxation levels defined based on the α wave time content rate calculated as the index value, the change tendency of the α wave time content rate, the integrated value of the high frequency component calculated as the index value, and the change tendency of the integrated value of the high frequency component, One of the relaxation levels is a deep relaxation level defined based on the fact that the α wave time content is in a high region and the change tendency of the α wave time content shows a rapid increase, and the integrated value of the high frequency component is in a middle region to a high region and the change tendency of the integrated value of the high frequency component shows a slight increase, Another of the relaxation levels is a sleeping level defined based on the fact that the α wave time content rate is in the middle to low range and the change tendency of the α wave time content rate shows a sharp decrease, and the integrated value of the high frequency component is in the high range and the change tendency of the integrated value of the high frequency component shows a sharp increase.

10. A state inference method, the state inference method being performed by a state inference device for inferring a relaxation level indicating a degree of relaxation of a user, wherein: The state inference method comprises: calculating a plurality of types of index values ​​using the biological information of the user; and inferring the user's relaxation level using a plurality of relaxation levels defined based on the plurality of types of index values ​​and the tendency of changes in the plurality of types of index values, In the step of calculating the index value, the heart rate of the user and an α wave time content ratio indicating a ratio of the α wave generation time to a unit time are calculated as the index value, In the step of inferring, the relaxation level of the user is estimated using a plurality of relaxation levels defined based on the heart rate calculated as the index value, the tendency of change in the heart rate, the α wave time content calculated as the index value, and the tendency of change in the α wave time content, One of the relaxation levels is a shallow relaxation level defined based on the heart rate being in a high to medium range and the change tendency of the heart rate showing a sharp decrease, and the α wave time content being in a low range and the change tendency of the α wave time content showing a slight increase, Another of the relaxation levels is a deep relaxation level defined based on the heart rate being in a medium to low range and the change tendency of the heart rate indicating a slight decrease, while the α wave time content rate is in a high range and the change tendency of the α wave time content rate indicating a sharp increase, Another of the relaxation levels is a sleeping level defined based on the heart rate being in a low range and the change tendency of the heart rate showing a slight decrease, and the α wave time content being in an intermediate to low range and the change tendency of the α wave time content showing a sharp decrease.

11. A state inference method, the state inference method being performed by a state inference device for inferring a relaxation level indicating a degree of relaxation of a user, wherein: The state inference method comprises: calculating a plurality of types of index values ​​using the biological information of the user; and inferring the user's relaxation level using a plurality of relaxation levels defined based on the plurality of types of index values ​​and the tendency of changes in the plurality of types of index values, In the step of calculating the index value, the heart rate of the user, the integrated value of the high-frequency component obtained by frequency analysis of the heartbeat variation of the user, and the α wave time content rate indicating the ratio of the generation time of the α wave of the user to the unit time are calculated as the index value, In the step of performing the inference, the relaxation level of the user is inferred using a plurality of relaxation levels specified based on the heart rate calculated as the index value, the tendency of change in the heart rate, the α wave time content calculated as the index value, the tendency of change in the α wave time content, the integrated value of the high-frequency component calculated as the index value, and the tendency of change in the integrated value of the high-frequency component, One of the relaxation levels is a first shallow relaxation level defined based on the heart rate being in a high to medium range and the tendency of change of the heart rate showing a sharp decrease, the α wave time content being in a low range and the tendency of change of the α wave time content showing a slight increase, and the integrated value of the high frequency component being in a low range and the tendency of change of the integrated value of the high frequency component showing no change, Another of the relaxation levels is a second shallow relaxation level that is defined as a relaxation level deeper than the first shallow relaxation level, based on the heart rate being in the middle to low range and the tendency of the heart rate change indicating a slight decrease, the alpha wave time content being in the low to middle range and the tendency of the alpha wave time content changing indicating a slight increase, and the integral value of the high-frequency component being in the middle to high range and the tendency of the integral value of the high-frequency component changing indicating a slight increase, Another one of the relaxation levels is a deep relaxation level defined based on the heart rate being in the middle to low range and the change tendency of the heart rate indicating a slight decrease, the alpha wave time content being in the high range and the change tendency of the alpha wave time content indicating a sharp increase, and the integrated value of the high frequency component being in the middle to high range and the change tendency of the integrated value of the high frequency component indicating a slight increase, Another one of the relaxation levels is a sleeping level determined based on the heart rate being in a low range and the tendency of change of the heart rate indicating a slight decrease, the α wave time content being in a middle range to a low range and the tendency of change of the α wave time content indicating a sharp decrease, and the integral value of the high-frequency component being in a high range and the tendency of change of the integral value of the high-frequency component indicating a sharp increase.

12. A state inference method, the method being performed by a state inference device for inferring a relaxation level indicating a degree of relaxation of a user, wherein: The state inference method comprises: calculating a plurality of types of index values ​​using the biological information of the user; and inferring the user's relaxation level using a plurality of relaxation levels defined based on the plurality of types of index values ​​and the tendency of changes in the plurality of types of index values, In the step of calculating the index value, an α wave time content rate indicating a ratio of the α wave generation time to a unit time and an integrated value of a high frequency component obtained by frequency analysis of the user's heartbeat fluctuations are calculated as the index value, In the step of inferring, the relaxation level of the user is estimated using a plurality of relaxation levels defined based on the α wave time content rate calculated as the index value, the change tendency of the α wave time content rate, the integrated value of the high frequency component calculated as the index value, and the change tendency of the integrated value of the high frequency component, One of the relaxation levels is a deep relaxation level defined based on the fact that the α wave time content is in a high region and the change tendency of the α wave time content shows a rapid increase, and the integrated value of the high frequency component is in a middle region to a high region and the change tendency of the integrated value of the high frequency component shows a slight increase, Another of the relaxation levels is a sleeping level defined based on the fact that the α wave time content rate is in the middle to low range and the change tendency of the α wave time content rate shows a sharp decrease, and the integrated value of the high frequency component is in the high range and the change tendency of the integrated value of the high frequency component shows a sharp increase.

Images