Falling tendency estimation device, falling tendency estimation system, falling tendency estimation method, and program

Abstract

In order to suitably estimate fall probability in everyday life, the present invention provides a fall probability estimation device comprising: a data acquisition unit for acquiring feature amount data that includes a feature amount used to estimate a user's fall probability and that is extracted from sensor data relating to movement of the user's feet; a storage unit for storing an estimation model that outputs a fall probability index that corresponds to an input of feature amount data; an estimation unit for inputting the acquired feature amount data into the estimation model, and estimating the user's fall probability according to the fall probability index that is output from the estimation model; and an output unit for outputting information relating to the user's fall probability that has been estimated.

Description

[Technical Field]

[0001] The present disclosure relates to a fall tendency estimation device and the like that estimates fall tendency using sensor data related to foot movement. [Background technology]

[0002] With growing interest in healthcare, attention is being focused on services that provide information based on the characteristics contained in walking patterns (also called gait). For example, technology is being developed that analyzes gait based on sensor data measured by sensors mounted on footwear such as shoes. Time-series data from sensor data reveals characteristics of gait events (also called walking events) related to physical conditions. For example, if fall susceptibility, one indicator of fall risk, could be estimated based on features extracted from sensor data, it may be possible to avoid unexpected falls.

[0003] Patent Document 1 discloses a device for detecting foot abnormalities based on the walking characteristics of a walker. The device in Patent Document 1 uses data acquired from a sensor attached to the footwear to extract gait features that are characteristic of the gait of a walker wearing the footwear. The device in Patent Document 1 detects abnormalities in a walker wearing the footwear based on the extracted gait features. For example, the device in Patent Document 1 extracts characteristic parts related to hallux valgus from gait waveform data for one stride cycle. The device in Patent Document 1 estimates the progression of hallux valgus using the gait features of the extracted characteristic parts.

[0004] Patent Document 2 discloses a system for assessing the risk of falls of elderly persons under management based on images taken during daily life. The system of Patent Document 2 authenticates the person under management who is photographed by a stereo camera that outputs two-dimensional images and three-dimensional information. The system of Patent Document 2 tracks the authenticated person under management and calculates the feature quantities of the person's gait. The system of Patent Document 2 calculates a fall index value for the person under management based on integrated data obtained by integrating the obtained data. The system of Patent Document 2 evaluates the risk of falls of the person under management according to the relationship between the calculated fall index value and a threshold value. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] International Publication No. 2021 / 140658 [Patent Document 2] International Publication No. 2021 / 186655 Summary of the Invention [Problem to be solved by the invention]

[0006] The method of Patent Document 1 estimates the progression of hallux valgus using gait feature quantities of characteristic parts extracted from data acquired by a sensor attached to footwear. Patent Document 1 does not disclose estimating fall susceptibility using gait feature quantities of characteristic parts extracted from data acquired by a sensor attached to footwear.

[0007] The method of Patent Document 2 evaluates the risk of a person being managed falling using two-dimensional images and three-dimensional information output from a stereo camera. The method of Patent Document 2 can evaluate the risk of a person being managed falling as long as the environment is one that can be photographed by a stereo camera. However, the method of Patent Document 2 cannot appropriately estimate the likelihood of a person being managed falling in environments that cannot be covered by a stereo camera, such as places with many obstacles or outdoors.

[0008] An object of the present disclosure is to provide a fall tendency estimation device, etc. that can appropriately estimate fall tendency in daily life. [Means for solving the problem]

[0009] A fall proneness estimation device according to one aspect of the present disclosure includes a data acquisition unit that acquires feature data including features used to estimate a user's fall proneness, extracted from sensor data relating to the user's foot movements; a memory unit that stores an estimation model that outputs a fall proneness index according to input of the feature data; an estimation unit that inputs the acquired feature data into the estimation model and estimates the user's fall proneness according to the fall proneness index output from the estimation model; and an output unit that outputs information relating to the estimated user's fall proneness.

[0010] In one aspect of the fall proneness estimation method of the present disclosure, feature data including features used to estimate a user's fall proneness is acquired from sensor data related to the movement of the user's feet, the acquired feature data is input into an estimation model that outputs a fall proneness index according to the input feature data, the user's fall proneness is estimated according to the fall proneness index output from the estimation model, and information related to the estimated user's fall proneness is output.

[0011] A program according to one aspect of the present disclosure causes a computer to perform the following processes: acquiring feature data including features used to estimate a user's tendency to fall, extracted from sensor data relating to the movement of the user's feet; inputting the acquired feature data into an estimation model that outputs a fall tendency index according to the input of the feature data; estimating the user's tendency to fall according to the fall tendency index output from the estimation model; and outputting information relating to the estimated user's tendency to fall. [Effects of the Invention]

[0012] According to the present disclosure, it is possible to provide a fall tendency estimation device, etc., that can appropriately estimate fall tendency in daily life. [Brief explanation of the drawings]

[0013] [Figure 1] 1 is a block diagram showing an example of the configuration of a fall likelihood estimation system according to a first embodiment. [Figure 2] 1 is a block diagram showing an example of the configuration of a gait measurement device included in a fall tendency estimation system according to a first embodiment. FIG. [Figure 3] FIG. 1 is a conceptual diagram showing an example of the arrangement of a gait measurement device according to a first embodiment. [Figure 4] FIG. 2 is a conceptual diagram for explaining an example of the relationship between a local coordinate system and a world coordinate system set in the gait measurement device according to the first embodiment. [Figure 5] FIG. 1 is a conceptual diagram for explaining a human body surface used in explaining the gait measurement device according to the first embodiment. [Figure 6] FIG. 1 is a conceptual diagram for explaining a walking cycle used in explaining the gait measurement device according to the first embodiment. [Figure 7] FIG. 1 is a conceptual diagram for explaining gait parameters used in explaining the gait measurement device according to the first embodiment. [Figure 8] 3 is a graph for explaining an example of time-series data of sensor data measured by the gait measurement device according to the first embodiment. [Figure 9] FIG. 3 is a diagram for explaining an example of normalization of gait waveform data extracted from time-series data of sensor data measured by the gait measurement device according to the first embodiment. [Figure 10] FIG. 2 is a conceptual diagram for explaining an example of a walking phase cluster from which a feature amount data generation unit of the gait measurement device according to the first embodiment extracts feature amounts. [Figure 11] 1 is a block diagram showing an example of the configuration of a fall tendency estimation device included in a fall tendency estimation system according to a first embodiment. [Figure 12] FIG. 2 is a conceptual diagram for explaining items related to fall tendency that are the subject of estimation by the fall tendency estimation system according to the first embodiment. [Figure 13]1 is a table summarizing feature amounts related to the total muscle strength (grip strength) of the whole body, which is related to the tendency to fall, which is an estimation target of the fall tendency estimation system according to the first embodiment. [Figure 14] 1 is a table summarizing feature quantities related to dynamic balance associated with fall tendency, which is an estimation target of the fall tendency estimation system according to the first embodiment. [Figure 15] 1 is a table summarizing feature amounts related to lower limb muscle strength associated with fall tendency, which is an estimation target of the fall tendency estimation system according to the first embodiment. [Figure 16] 1 is a table summarizing feature quantities related to mobility associated with fall tendency, which is an estimation target of the fall tendency estimation system according to the first embodiment. [Figure 17] 1 is a table summarizing feature quantities related to static balance, which is related to fallability, which is an estimation target of the fallability estimation system according to the first embodiment. [Figure 18] 1 is a table summarizing feature quantities related to fall tendency, which is an estimation target of the fall tendency estimation system according to the first embodiment. [Figure 19] 1 is a conceptual diagram showing an example of an estimation of a fall tendency score (fall tendency index) by a fall tendency estimation device included in a fall tendency estimation system according to the first embodiment. FIG. [Figure 20] 1 is a conceptual diagram showing an example of an estimation of a fall tendency score (fall tendency index) by a fall tendency estimation device included in a fall tendency estimation system according to the first embodiment. FIG. [Figure 21] 5 is a flowchart for explaining an example of the operation of the gait measurement device included in the fall tendency estimation system according to the first embodiment. [Figure 22] 5 is a flowchart for explaining an example of the operation of the fall tendency estimating device included in the fall tendency estimating system according to the first embodiment. [Figure 23] 5 is a flowchart for explaining an example of the operation of the gait measurement device included in the fall tendency estimation system according to the first embodiment. [Figure 24] 5 is a flowchart for explaining an example of the operation of the fall tendency estimating device included in the fall tendency estimating system according to the first embodiment. [Figure 25] FIG. 2 is a conceptual diagram for explaining an application example of the fall tendency estimation system according to the first embodiment. [Figure 26] FIG. 2 is a conceptual diagram for explaining an application example of the fall tendency estimation system according to the first embodiment. [Figure 27] FIG. 10 is a block diagram showing an example of the configuration of a learning system according to a second embodiment. [Figure 28] FIG. 10 is a block diagram showing an example of the configuration of a learning device included in a learning system according to a second embodiment. [Figure 29] FIG. 10 is a conceptual diagram for explaining an example of learning by a learning device included in a learning system according to a second embodiment. [Figure 30] FIG. 10 is a conceptual diagram for explaining another example of learning by a learning device included in a learning system according to the second embodiment. [Figure 31] FIG. 10 is a block diagram showing an example of the configuration of a fall tendency estimation device according to a third embodiment. [Figure 32] FIG. 2 is a block diagram showing an example of a hardware configuration for executing control and processing in each embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0014] Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. However, the embodiments described below are limited in a manner that is technically preferable for carrying out the present invention, but the scope of the invention is not limited to the following. In all drawings used to describe the following embodiments, the same reference numerals are used for similar parts unless otherwise specified. Furthermore, in the following embodiments, repeated explanations of similar configurations and operations may be omitted.

[0015] (First embodiment) First, a fall susceptibility estimation system according to a first embodiment will be described with reference to the drawings. The fall susceptibility estimation system of this embodiment measures sensor data related to foot movements according to a user's walking. The fall susceptibility estimation system of this embodiment uses the measured sensor data to estimate the user's fall susceptibility.

[0016] In this embodiment, an example is given in which fall susceptibility is estimated according to the correlation between features included in walking patterns (also referred to as gait) and fall risk. Fall susceptibility, which is one index of fall risk, can be evaluated based on the variability of gait parameters. In this embodiment, fall susceptibility is estimated using five related items (also referred to as five items) related to gait. The five items relate to total muscle strength of the whole body (grip strength), dynamic balance, lower limb muscle strength, mobility, and static balance. These five items are correlated with fall susceptibility. Although the five items are correlated to each other to some extent, they are basically considered to be independent. In this embodiment, an example is given in which fall susceptibility is estimated based on all five items, but fall susceptibility can also be estimated based on at least one of the five items.

[0017] (composition) FIG. 1 is a block diagram showing an example of the configuration of a fall liability estimation system 1 according to this embodiment. The fall liability estimation system 1 includes a gait measurement device 10 and a fall liability estimation device 13. In this embodiment, an example will be described in which the gait measurement device 10 and the fall liability estimation device 13 are configured as separate pieces of hardware. For example, the gait measurement device 10 is attached to the footwear of a subject (user) whose fall liability is to be estimated. For example, the functions of the fall liability estimation device 13 are installed in a mobile device carried by the subject (user). Below, the configurations of the gait measurement device 10 and the fall liability estimation device 13 will be described separately.

[0018] [Gait measurement device] 2 is a block diagram showing an example of the configuration of gait measurement device 10. Gait measurement device 10 has sensor 11 and feature amount data generation unit 12. In this embodiment, an example is given in which sensor 11 and feature amount data generation unit 12 are integrated. Sensor 11 and feature amount data generation unit 12 may also be provided as separate devices.

[0019] 2, the sensor 11 has an acceleration sensor 111 and an angular velocity sensor 112. Fig. 2 shows an example in which the acceleration sensor 111 and the angular velocity sensor 112 are included in the sensor 11. The sensor 11 may include sensors other than the acceleration sensor 111 and the angular velocity sensor 112. Description of sensors other than the acceleration sensor 111 and the angular velocity sensor 112 that may be included in the sensor 11 will be omitted.

[0020] The acceleration sensor 111 is a sensor that measures acceleration in three axial directions (also called spatial acceleration). The acceleration sensor 111 measures acceleration (also called spatial acceleration) as a physical quantity related to foot movement. The acceleration sensor 111 outputs the measured acceleration to the feature data generation unit 12. For example, a piezoelectric, piezo-resistive, or capacitive sensor can be used as the acceleration sensor 111. There is no limitation on the measurement method of the sensor used as the acceleration sensor 111 as long as it can measure acceleration.

[0021] The angular velocity sensor 112 is a sensor that measures angular velocity (also called spatial angular velocity) around three axes. The angular velocity sensor 112 measures angular velocity (also called spatial angular velocity) as a physical quantity related to foot movement. The angular velocity sensor 112 outputs the measured angular velocity to the feature data generation unit 12. For example, a vibration type or capacitance type sensor can be used as the angular velocity sensor 112. There is no limitation on the measurement method of the sensor used as the angular velocity sensor 112 as long as it can measure angular velocity.

[0022] The sensor 11 is realized by, for example, an inertial measurement unit that measures acceleration and angular velocity. An example of an inertial measurement unit is an IMU (Inertial Measurement Unit). The IMU includes an acceleration sensor 111 that measures acceleration in three axial directions and an angular velocity sensor 112 that measures angular velocity around three axes. The sensor 11 may be realized by an inertial measurement unit such as a VG (Vertical Gyro) or an AHRS (Attitude Heading). The sensor 11 may also be realized by a GPS / INS (Global Positioning System / Inertial Navigation System). The sensor 11 may be realized by a device other than an inertial measurement unit as long as it can measure physical quantities related to foot movement.

[0023] FIG. 3 is a conceptual diagram showing an example in which gait measurement device 10 is placed inside shoe 100 of a right foot. In the example of FIG. 3, gait measurement device 10 is installed at a position corresponding to the back of the arch of the foot. For example, gait measurement device 10 is placed in an insole inserted into shoe 100. For example, gait measurement device 10 may be placed on the bottom of shoe 100. For example, gait measurement device 10 may be embedded in the body of shoe 100. gait measurement device 10 may or may not be detachable from shoe 100. gait measurement device 10 may be installed at a position other than the back of the arch of the foot as long as it can measure sensor data related to foot movement. gait measurement device 10 may also be installed in socks worn by the user or in an accessory such as an anklet worn by the user. gait measurement device 10 may also be attached directly to the foot or embedded in the foot. 3 shows an example in which the gait measurement device 10 is installed in the right shoe 100. The gait measurement device 10 may also be installed in the shoes 100 on both feet.

[0024] In the example of FIG. 3 , a local coordinate system is set with the gait measurement device 10 (sensor 11) as the reference, and includes an x-axis in the left-right direction, a y-axis in the front-back direction, and a z-axis in the up-down direction. The x-axis is positive to the left, the y-axis is positive backward, and the z-axis is positive upward. The directions of the axes set in the sensor 11 may be the same for the left and right feet, or may be different for the left and right feet. For example, when sensors 11 manufactured to the same specifications are placed in left and right shoes 100, the up-down directions (directions of the Z-axis) of the sensors 11 placed in the left and right shoes 100 are the same. In this case, the three axes of the local coordinate system set for the sensor data derived from the left foot and the three axes of the local coordinate system set for the sensor data derived from the right foot are the same for the left and right feet.

[0025] FIG. 4 is a conceptual diagram illustrating a local coordinate system (x-axis, y-axis, z-axis) set in gait measurement device 10 (sensor 11) installed on the backside of the arch of the foot, and a world coordinate system (x-axis, y-axis, z-axis) set with respect to the ground. In the world coordinate system (x-axis, y-axis, z-axis), when a user is standing upright and facing the direction of travel, the user's sideways direction is set as the x-axis direction (positive for left), the direction of the user's back is set as the y-axis direction (positive for backward), and the direction of gravity is set as the z-axis direction (positive for vertical upward). Note that the example in FIG. 4 conceptually illustrates the relationship between the local coordinate system (x-axis, y-axis, z-axis) and the world coordinate system (x-axis, y-axis, z-axis), and does not accurately illustrate the relationship between the local coordinate system and the world coordinate system, which changes according to the user's walking.

[0026] FIG. 5 is a conceptual diagram illustrating planes (also called human body planes) set for the human body. In this embodiment, a sagittal plane that divides the body into left and right, a coronal plane that divides the body into front and back, and a horizontal plane that divides the body horizontally are defined. As shown in FIG. 5, when the user is standing upright with the center line of the feet facing the direction of travel, the world coordinate system and the local coordinate system coincide. In this embodiment, a rotation in the sagittal plane about the x-axis as the rotation axis is defined as roll, a rotation in the coronal plane about the y-axis as the rotation axis is defined as pitch, and a rotation in the horizontal plane about the z-axis as the rotation axis is defined as yaw. Furthermore, a rotation angle in the sagittal plane about the x-axis as the rotation axis is defined as roll angle, a rotation angle in the coronal plane about the y-axis as the rotation axis is defined as pitch angle, and a rotation angle in the horizontal plane about the z-axis as the rotation axis is defined as yaw angle.

[0027] As shown in FIG. 2 , feature data generation unit 12 (also referred to as a feature data generation device) includes an acquisition unit 121, a normalization unit 122, an extraction unit 123, a generation unit 125, and a feature data output unit 127. For example, feature data generation unit 12 is implemented by a microcomputer or microcontroller that performs overall control of gait measurement device 10 and data processing. For example, feature data generation unit 12 includes a CPU (Central Processing Unit), RAM (Random Access Memory), ROM (Read Only Memory), flash memory, etc. The feature data generation unit 12 controls an acceleration sensor 111 and an angular velocity sensor 112 to measure angular velocity and acceleration. For example, feature data generation unit 12 may be implemented on the side of a mobile terminal (not shown) carried by the subject (user).

[0028] The acquiring unit 121 acquires acceleration in three axial directions from the acceleration sensor 111. The acquiring unit 121 also acquires angular velocities around three axes from the angular velocity sensor 112. For example, the acquiring unit 121 performs analog-to-digital conversion (AD conversion) on physical quantities (analog data) such as the acquired angular velocities and accelerations. The physical quantities (analog data) measured by the acceleration sensor 111 and the angular velocity sensor 112 may be converted into digital data by the acceleration sensor 111 and the angular velocity sensor 112, respectively. The acquiring unit 121 outputs the converted digital data (also referred to as sensor data) to the normalizing unit 122. The acquiring unit 121 may be configured to store the sensor data in a storage unit (not shown). The sensor data includes at least acceleration data converted into digital data and angular velocity data converted into digital data. The acceleration data includes acceleration vectors in the three axial directions. The angular velocity data includes angular velocity vectors around three axes. The acceleration data and angular velocity data are associated with the time when the data was acquired. The acquiring unit 121 may also apply corrections to the acceleration data and angular velocity data, such as correction for mounting error, temperature correction, and linearity correction.

[0029] The normalization unit 122 acquires sensor data from the acquisition unit 121. The normalization unit 122 extracts time series data for one walking cycle (also referred to as walking waveform data) from the time series data of accelerations in three axial directions and angular velocities around three axes included in the sensor data. The normalization unit 122 normalizes the time of the extracted walking waveform data for one walking cycle to a walking cycle of 0 to 100% (percent) (also referred to as first normalization). Timings such as 1% and 10% included in the 0 to 100% walking cycle are also referred to as walking phases. The normalization unit 122 also normalizes the first normalized walking waveform data for one walking cycle so that the stance phase is 60% and the swing phase is 40% (also referred to as second normalization). The stance phase is a period when at least a part of the sole of the foot is in contact with the ground. The swing phase is a period when the sole of the foot is off the ground. By subjecting the walking waveform data to second normalization, it is possible to suppress fluctuations in the walking phase from which feature values are extracted due to the influence of disturbances.

[0030] FIG. 6 is a conceptual diagram illustrating a step cycle based on the right foot. The step cycle based on the left foot is similar to that of the right foot. The horizontal axis of FIG. 6 represents one step cycle of the right foot, starting from the point when the heel of the right foot hits the ground and ending from the point when the heel of the right foot hits the ground again. The horizontal axis of FIG. 6 is first normalized so that the step cycle is 100%. The horizontal axis of FIG. 6 is also second normalized so that the stance phase is 60% and the swing phase is 40%. One step cycle of one leg is roughly divided into a stance phase in which at least a portion of the sole of the foot is in contact with the ground, and a swing phase in which the sole of the foot is off the ground. The stance phase is further divided into a load response period T1, a mid-stance period T2, a final stance period T3, and an early swing period T4. The swing phase is further divided into an early swing period T5, a mid-swing period T6, and a final swing period T7. Note that FIG. 6 is an example and does not limit the periods that make up a walking cycle or the names of these periods.

[0031] As shown in Figure 6, multiple events (also called walking events) occur during walking. P1 represents the event in which the heel of the right foot touches the ground (heel contact: HC). P2 represents the event in which the toe of the left foot leaves the ground (opposite toe off: OTO) while the sole of the right foot is in contact with the ground. P3 represents the event in which the heel of the right foot rises (heel rise: HR) while the sole of the right foot is in contact with the ground (opposite heel strike: OHS). P5 represents the event in which the toe of the right foot leaves the ground (toe off: TO) while the sole of the left foot is in contact with the ground (toe off: TO). P6 represents the event in which the left and right feet cross (foot crossing: FA) while the sole of the left foot is in contact with the ground (foot adjacent). P7 represents the event where the tibia of the right foot is nearly perpendicular to the ground (Tibia Vertical: TV) with the sole of the left foot touching the ground. P8 represents the event where the heel of the right foot touches the ground (Heel Contact: HC). P8 corresponds to the end point of the walking cycle that begins with P1 and also corresponds to the starting point of the next walking cycle. Note that Figure 6 is an example and does not limit the events that occur during walking or the names of those events.

[0032] FIG. 7 is a conceptual diagram for explaining an example of gait parameters. R , left foot step length S L , stride length T, step width W, foot angle F, and rotational distance DI are shown. Also, FIG. 7 shows the axis of travel PA, which is parallel to the axis of travel (Y-axis) and corresponds to the trajectory connecting the midpoints of the left and right feet. Right foot step length S R is the difference in Y coordinate between the heel of the right foot and the heel of the left foot when the state transitions from the state where the sole of the left foot is on the ground to the state where the heel of the right foot, which is swung in the direction of travel, is on the ground. Left foot step length S L is the difference in Y coordinate between the heel of the left foot and the heel of the right foot when the state transitions from the state where the sole of the right foot is on the ground to the state where the heel of the left foot, which is swung in the direction of travel, lands on the ground. The stride length T is the difference in Y coordinate between the right foot step length S R and left foot step length S LThe step width W is the distance between the right foot and the left foot. In FIG. 7, the step width W is the difference between the center line (X coordinate) of the heel of the right foot when in contact with the ground and the center line (X coordinate) of the heel of the left foot when in contact with the ground. The foot angle F is the angle formed between the center line of the foot and the direction of travel (Y axis) when the sole of the foot is in contact with the ground. In this embodiment, the foot angle is evaluated when the foot is in contact with the ground during the stance phase. The amount of rotation DI is the distance between the traveling axis PA and the foot at the time when the center axis of the foot is farthest from the traveling axis PA during the swing phase. In this embodiment, the amount of rotation DI is normalized by height because it is affected by the length of the lower limbs.

[0033] FIG. 8 is a diagram illustrating an example of detecting heel strike HC and toe lift TO from time series data (solid line) of forward acceleration (Y-direction acceleration). The timing of heel strike HC is the timing of the minimum peak immediately after the maximum peak that appears in the time series data of forward acceleration (Y-direction acceleration). The maximum peak that marks the timing of heel strike HC corresponds to the maximum peak of the gait waveform data for one step cycle. The section between consecutive heel strikes HC is one step cycle. The timing of toe lift TO is the timing of the rise of the maximum peak that appears after the stance phase period during which no fluctuations appear in the time series data of forward acceleration (Y-direction acceleration). FIG. 8 also shows time series data (dashed line) of roll angle (angular velocity about the X-axis). The midpoint between the timing of the minimum roll angle and the timing of the maximum roll angle corresponds to the mid-stance phase. For example, parameters such as walking speed, stride length, circumduction, internal rotation / external rotation, and plantar flexion / dorsiflexion (also called gait parameters) can be determined using the mid-stance phase as a reference.

[0034] FIG. 9 is a diagram illustrating an example of gait waveform data normalized by the normalization unit 122. The normalization unit 122 detects heel strikes HC and toe lifts TO from time-series data of forward acceleration (Y-direction acceleration). The normalization unit 122 extracts the interval between successive heel strikes HC as gait waveform data for one step cycle. The normalization unit 122 converts the horizontal axis (time axis) of the gait waveform data for one step cycle into a gait cycle of 0 to 100% by first normalization. In FIG. 9, the gait waveform data after the first normalization is shown by a dashed line. In the gait waveform data (dashed line) after the first normalization, the timing of toe lifts TO is shifted from 60%.

[0035] In the example of FIG. 9, the normalization unit 122 normalizes the section from heel-strike HC, where the walking phase is 0%, to toe-off TO, which follows the heel-strike HC, to 0-60%. The normalization unit 122 also normalizes the section from toe-off TO to heel-strike HC, where the walking phase is 100%, to 60-100%. As a result, the gait waveform data for one step cycle is normalized into a section where the gait cycle is 0-60% (stance phase) and a section where the gait cycle is 60-100% (swing phase). In FIG. 9, the gait waveform data after the second normalization is shown by the solid line. In the gait waveform data after the second normalization (solid line), the timing of toe-off TO coincides with 60%.

[0036] 8 and 9 show an example in which gait waveform data for one step cycle is extracted and normalized based on the traveling acceleration (Y-direction acceleration). With respect to accelerations / angular velocities other than the traveling acceleration (Y-direction acceleration), the normalization unit 122 extracts and normalizes gait waveform data for one step cycle in accordance with the traveling acceleration (Y-direction acceleration) gait cycle. The normalization unit 122 may also generate time series data of angles around three axes by integrating time series data of angular velocities around three axes. In this case, the normalization unit 122 extracts and normalizes gait waveform data for one step cycle in accordance with the traveling acceleration (Y-direction acceleration) gait cycle, also with respect to angles around three axes.

[0037] The normalization unit 122 may extract / normalize the gait waveform data for one step gait cycle based on acceleration / angular velocity other than the forward acceleration (Y-direction acceleration) (not shown in the drawings). For example, the normalization unit 122 may detect heel strike HC and toe lift TO from time series data of vertical acceleration (Z-direction acceleration). The timing of heel strike HC is the timing of a steep minimum peak that appears in the time series data of vertical acceleration (Z-direction acceleration). At the timing of the steep minimum peak, the value of vertical acceleration (Z-direction acceleration) becomes almost zero. The minimum peak that marks the timing of heel strike HC corresponds to the minimum peak of the gait waveform data for one step gait cycle. The section between consecutive heel strikes HC is one step gait cycle. The timing of toe-off TO is the timing of an inflection point in the time-series data of vertical acceleration (Z-direction acceleration) where the vertical acceleration (Z-direction acceleration) gradually increases after passing through a section of small fluctuations following the maximum peak immediately after heel-contact HC. The normalization unit 122 may also extract / normalize the gait waveform data for one walking cycle based on both the forward acceleration (Y-direction acceleration) and the vertical acceleration (Z-direction acceleration). The normalization unit 122 may also extract / normalize the gait waveform data for one walking cycle based on acceleration, angular velocity, angle, etc. other than the forward acceleration (Y-direction acceleration) and the vertical acceleration (Z-direction acceleration).

[0038] The extraction unit 123 acquires walking waveform data for one step cycle normalized by the normalization unit 122. The extraction unit 123 extracts feature amounts used to estimate fall susceptibility from the walking waveform data for one step cycle. The extraction unit 123 extracts feature amounts for each walking phase cluster from walking phase clusters that integrate temporally consecutive walking phases based on preset conditions. A walking phase cluster includes at least one walking phase. A walking phase cluster may also include a single walking phase. The walking waveform data and walking phases from which feature amounts used to estimate fall susceptibility are extracted will be described later.

[0039] FIG. 10 is a conceptual diagram illustrating the extraction of feature quantities for estimating fall likelihood from gait waveform data for one step cycle. For example, the extraction unit 123 extracts temporally consecutive walking phases i to i+m as a walking phase cluster CL (i and m are natural numbers). The walking phase cluster CL includes m walking phases (components). That is, the number of walking phases (components) constituting the walking phase cluster CL (also referred to as the number of components) is m. While FIG. 10 illustrates an example in which the walking phases are integer values, the walking phases may be subdivided to the nearest decimal point. When the walking phases are subdivided to the nearest decimal point, the number of components of the walking phase cluster CL corresponds to the number of data points in the section of the walking phase cluster. The extraction unit 123 extracts feature quantities from each of the walking phases i to i+m. When the walking phase cluster CL is composed of a single walking phase j, the extraction unit 123 extracts feature quantities from the single walking phase j (j is a natural number).

[0040] The generation unit 125 applies a feature composition formula to a feature (first feature) extracted from each of the walking phases constituting the walking phase cluster to generate a feature (second feature) of the walking phase cluster. The feature composition formula is a calculation formula set in advance for generating a feature of the walking phase cluster. For example, the feature composition formula is a calculation formula related to four arithmetic operations. For example, the second feature calculated using the feature composition formula is an integral average value, arithmetic mean value, slope, variance, etc. of the first feature in each walking phase included in the walking phase cluster. For example, the generation unit 125 applies, as the feature composition formula, a calculation formula for calculating the slope or variance of the first feature extracted from each of the walking phases constituting the walking phase cluster. For example, when a walking phase cluster is composed of a single walking phase, it is not possible to calculate the slope or variance, so a feature composition formula for calculating an integral average value, arithmetic mean value, etc. may be used.

[0041] The feature data output unit 127 outputs the feature data for each gait phase cluster generated by the generation unit 125. The feature data output unit 127 outputs the feature data of the generated gait phase cluster to the fall tendency estimation device 13, which uses the feature data.

[0042] [Easy to fall property estimation device] 11 is a block diagram showing an example of the configuration of the fall tendency estimating device 13. The fall tendency estimating device 13 has a data acquiring unit 131, a storage unit 132, an estimating unit 133, and an output unit 135.

[0043] Data acquisition unit 131 acquires feature amount data from gait measurement device 10. Data acquisition unit 131 outputs the received feature amount data to estimation unit 133. Data acquisition unit 131 may receive the feature amount data from gait measurement device 10 via a wired connection such as a cable, or may receive the feature amount data from gait measurement device 10 via wireless communication. For example, data acquisition unit 131 is configured to receive the feature amount data from gait measurement device 10 via a wireless communication function (not shown) that complies with standards such as Bluetooth (registered trademark) or WiFi (registered trademark). Note that the communication function of data acquisition unit 131 may be compliant with standards other than Bluetooth (registered trademark) or WiFi (registered trademark).

[0044] The storage unit 132 stores an estimation model for estimating fall proneness using feature amount data extracted from the gait waveform data. The storage unit 132 stores estimation models for estimating fall proneness that have been trained on a plurality of subjects. For example, the storage unit 132 stores an estimation model that outputs a fall proneness index (also referred to as a fall proneness score) in response to input of feature amount data extracted from the gait waveform data.

[0045] 12 is a conceptual diagram for explaining five related items (also referred to as the five items) related to fall susceptibility. Five items related to fall susceptibility are: total whole-body muscle strength (grip strength), dynamic balance, lower limb muscle strength, mobility, and static balance. Details of the five items related to fall susceptibility, total whole-body muscle strength (grip strength), dynamic balance, lower limb muscle strength, mobility, and static balance, will be described later.

[0046] For example, the storage unit 132 stores an estimation model (also referred to as a first estimation model) that outputs a fall tendency index (fall tendency score) in response to input of feature amount data common to the estimation of the five items. For example, the storage unit 132 stores an estimation model (also referred to as a pre-estimation model) that outputs a score for each of the five items in response to input of feature amount data used to estimate the score for each of the five items. For example, the storage unit 132 stores an estimation model (also referred to as a second estimation model) that outputs a fall tendency index (fall tendency score) in response to input of the scores for the five items.

[0047] The estimation model may be stored in storage unit 132 at the time of product shipment from the factory or at the time of calibration before a user uses fall likelihood estimation system 1. For example, the system may be configured to use an estimation model stored in a storage device such as an external server. In this case, the estimation model may be used via an interface (not shown) connected to the storage device.

[0048] The estimation unit 133 acquires feature data from the data acquisition unit 131. The estimation unit 133 uses the acquired feature data to estimate the tendency to fall. The estimation unit 133 inputs the feature data to an estimation model stored in the storage unit 132. The estimation unit 133 outputs an estimation result corresponding to the tendency to fall output from the estimation model. When using an estimation model stored in an external storage device constructed on a cloud, a server, or the like, the estimation unit 133 is configured to use the estimation model via an interface (not shown) connected to the storage device.

[0049] The output unit 135 outputs the fall tendency estimation result obtained by the estimation unit 133. For example, the output unit 135 displays the fall tendency estimation result on the screen of the subject (user)'s mobile terminal. For example, the output unit 135 outputs the estimation result to an external system or the like that uses the estimation result. There are no particular limitations on how the fall tendency output from the fall tendency estimation device 13 is used.

[0050] For example, the fall susceptibility estimation device 13 is connected to an external system, such as a cloud or a server, via a mobile device (not shown) carried by the subject (user). The mobile device (not shown) is a portable communication device. For example, the mobile device is a mobile communication device with a communication function, such as a smartphone, a smart watch, or a mobile phone. For example, the fall susceptibility estimation device 13 is connected to the mobile device via a wired connection, such as a cable. For example, the fall susceptibility estimation device 13 is connected to the mobile device via wireless communication. For example, the fall susceptibility estimation device 13 is connected to the mobile device via a wireless communication function (not shown) conforming to standards such as Bluetooth (registered trademark) or WiFi (registered trademark). Note that the communication function of the fall susceptibility estimation device 13 may conform to standards other than Bluetooth (registered trademark) or WiFi (registered trademark). The fall susceptibility estimation result may be used by an application installed on the mobile device. In this case, the mobile device executes processing using the estimation result using application software, etc., installed on the mobile device.

[0051] Next, we will explain each of the five items related to fall susceptibility shown in Figure 12: total whole-body muscle strength (grip strength), dynamic balance, lower limb muscle strength, mobility, and static balance. After explaining each of the five items, we will explain the feature amounts used to estimate fall susceptibility. Fall susceptibility is estimated using feature amounts common to the five items.

[0052] <Related Item A> Related item A relates to total muscle strength of the whole body. There is a correlation between total muscle strength and grip strength. Grip strength is also correlated with knee extension strength. One indicator of total muscle strength related to related item A is grip strength. For example, an estimated value of grip strength is an indicator of total muscle strength. For example, a score based on the estimated value of grip strength (also called a total muscle strength score) is an indicator of total muscle strength. The total muscle strength score is a value obtained by scoring grip strength, which is an indicator of total muscle strength, using a predetermined standard. Grip strength is affected by attributes such as gender, age, and height. Therefore, the total muscle strength score may be scored using a standard for each attribute. In particular, grip strength is affected by gender. Therefore, the total muscle strength score may be scored using different standards depending on gender. Note that the indicator of total muscle strength is not limited to grip strength, as long as it is possible to score total muscle strength.

[0053] FIG. 13 is a correspondence table summarizing the feature values used to estimate total muscle strength (grip strength). The correspondence table in FIG. 13 associates the feature value number with the gait waveform data from which the feature value is extracted, the gait phase (%) from which the gait phase cluster is extracted, and the associated muscles. The gait phase from which the feature values used to estimate grip strength are extracted differs depending on gender. For men, there is a correlation between quadriceps activity and grip strength. Therefore, to estimate men's grip strength, feature values AM1 to AM4 extracted from gait phases in which the characteristics of quadriceps activity are apparent are used. For women, there is a correlation between grip strength and the activities of the quadriceps muscles, vastus lateralis, vastus intermedius, and vastus medialis. Therefore, to estimate women's grip strength, feature values AF1 to AF3 extracted from gait phases in which the characteristics of vastus lateralis, vastus intermedius, and vastus medialis are apparent are used.

[0054] The feature AM1 is extracted from the 3% walking phase of the walking waveform data Ay, which is related to the time-series data of forward acceleration (Y-direction acceleration). The 3% walking phase is included in the load response period T1. The feature AM1 mainly includes features related to the movements of the vastus lateralis, vastus intermedius, and vastus medialis muscles of the quadriceps femoris.

[0055] Feature AM2 is extracted from the section of the walking phase 59-62% of the walking waveform data Ay related to the time-series data of forward acceleration (Y-direction acceleration). The walking phase 59-62% is included in the early swing phase T4. Feature AM2 mainly includes features related to the movement of the rectus femoris muscle, which is one of the quadriceps muscles.

[0056] Feature AM3 is extracted from the 59% to 62% walking phase of the walking waveform data Az, which is related to the time-series data of vertical acceleration (Z-direction acceleration). The 59% to 62% walking phase is included in the early swing phase T4. Feature AM3 mainly includes features related to the movement of the rectus femoris, which is one of the quadriceps muscles.

[0057] Feature AM4 is the ratio of the period from heel-contact to toe-off of the opposite foot to the period when both feet are simultaneously on the ground (DST1: Double Support Time). DST1 is the ratio of the period from heel-contact to toe-off of the opposite foot in a gait cycle. Feature AM4 mainly includes features attributable to the quadriceps femoris.

[0058] The feature AF1 is extracted from the 13% section of the walking phase of the walking waveform data Ax, which is related to the time-series data of lateral acceleration (X-direction acceleration). The 13% walking phase is included in the mid-stance phase T2. The feature AF1 mainly includes features related to the movement of the vastus lateralis, vastus intermedius, and vastus medialis muscles of the quadriceps femoris.

[0059] The feature AF2 is extracted from the 7-10% section of the walking phase of the walking waveform data Gy, which is related to the time series data of angular velocity (pitch angular velocity) in the coronal plane (around the Y-axis). The 7-10% walking phase is included in the load response period T1. The feature AF2 mainly includes features related to the movements of the vastus lateralis, vastus intermedius, and vastus medialis muscles.

[0060] Feature AF3 is the proportion of the period from heel contact to toe-off of the opposite foot (DST2) to the period during which both feet are simultaneously in contact with the ground (DST: Double Support Time). DST2 is the proportion of the period during a gait cycle from heel contact to toe-off of the opposite foot. The sum of DST1 and DST2 corresponds to the period during which both feet are simultaneously in contact with the ground during a gait cycle. Feature AF3 mainly includes features related to the movements of the vastus lateralis, vastus intermedius, and vastus medialis.

[0061] <Related Item B> Related item B relates to dynamic balance. Dynamic balance can be evaluated by the performance of the Functional Reach Test (FRT). In this embodiment, the performance of the FRT is evaluated based on the distance between the fingertips (also referred to as the functional reach distance) when the subject stands with both hands raised 90 degrees to the horizontal and then moves the upper limbs as far forward as possible. The functional reach distance (hereinafter referred to as the FR distance) is the performance value of the FRT. The larger the FR distance, the higher the performance of the FRT. Related item B may also be evaluated using methods other than the FRT performed with both hands. For example, related item B may be evaluated based on the performance of the FRT performed with one hand or other variations of the FRT.

[0062] The dynamic balance index for related item B is the FR distance. For example, an estimated value of the FR distance is the dynamic balance index. For example, a score according to the estimated value of the FR distance (also called the dynamic balance score) is the dynamic balance index. The dynamic balance score is a value obtained by scoring the FR distance, which is an index of dynamic balance, based on a preset standard. Dynamic balance is affected by attributes such as height. Therefore, the dynamic balance score may be scored based on a standard for each attribute. Note that the dynamic balance index is not limited to the FR distance as long as it is possible to score dynamic balance.

[0063] FIG. 14 is a correspondence table summarizing the feature quantities used in estimating dynamic balance. The correspondence table in FIG. 14 associates the feature quantity number, the gait waveform data from which the feature quantity is extracted, the gait phase (%) from which the gait phase cluster is extracted, and the associated muscles. The FR distance is correlated with the activity of the gluteus medius, iliacus, hamstrings (long head of biceps femoris), tibialis anterior, etc., as well as the magnitude of the compensatory movement of turning the toes outward. Therefore, feature quantities B1 to B5 extracted from the gait phases in which these features appear are used to estimate the FR distance.

[0064] Feature B1 is extracted from the 75-79% gait phase of the gait waveform data Ay, which is related to the time-series data of forward acceleration (Y-direction acceleration). The 75-79% gait phase is included in the mid-swing phase T6. Feature B1 mainly includes features related to the movement of the tibialis anterior muscle and the short head of the biceps femoris muscle.

[0065] Feature B2 is extracted from the 62% walking phase of the walking waveform data Az, which is related to the time-series data of vertical acceleration (Z-direction acceleration). The 62% walking phase is included in the initial swing phase T5. Feature B2 mainly includes features related to the movement of the iliacus muscle.

[0066] Feature B3 is extracted from the 7-8% section of the walking phase of the walking waveform data Gy, which is related to the time series data of angular velocity in the coronal plane (around the Y-axis). The 7-8% walking phase is included in the load response period T1. Feature B3 mainly includes features related to the movement of the gluteus medius muscle.

[0067] Feature B4 is extracted from the 57-58% section of the walking phase of the walking waveform data Ez, which is related to time-series data of angles (postural angles) in the horizontal plane (around the Z axis). The 57-58% walking phase is included in the early swing phase T4. Feature B4 mainly includes features related to compensatory movements. Compensatory movements are movements that change the foot angle to gain stability in order to compensate for the decline in balance ability and muscle function that occurs with aging.

[0068] The feature quantity B5 is the average value of the foot angle in the horizontal plane during the swing phase. For example, the feature quantity B5 is the average value of the gait waveform data Ez during the swing phase. In other words, the feature quantity B5 is the integral value of the gait waveform data Gz related to the time-series data of angular velocity in the horizontal plane (around the Z-axis). The feature quantity B5 mainly includes features related to compensatory movements.

[0069] <Related Item C> Related item C relates to lower limb muscle strength. Lower limb muscle strength can be evaluated by the results of a chair stand test. In this embodiment, the results of the 5-chair stand test, in which the subject stands up and sits down from a chair five times, are evaluated. The 5-chair stand test is also called the SS-5 (Sit to Stand-5) test. The results of the 5-chair stand test are evaluated based on the time it takes to stand up and sit down from a chair five times (also called the sit-to-stand time). The sit-to-stand time is the score value of the SS-5 test. The shorter the sit-to-stand time, the higher the score of the SS-5 test. The lower limb muscle strength may also be evaluated by the results of a 30-second chair stand (CS-30) test, which measures the number of times the subject stands up and sits down from a chair in 30 seconds.

[0070] An index of lower limb muscle strength related to related item C is the standing-sitting time. For example, an estimated value of the standing-sitting time five times is an index of lower limb muscle strength. For example, a score according to the estimated value of the standing-sitting time (also called a lower limb muscle strength score) is an index of lower limb muscle strength. The lower limb muscle strength score is a value obtained by scoring the standing-sitting time, which is an index of lower limb muscle strength, using a preset standard. Lower limb muscle strength is affected by attributes such as age. Therefore, the lower limb muscle strength score may be scored using a standard for each attribute. Note that the index of lower limb muscle strength is not limited to the standing-sitting time, as long as lower limb muscle strength can be scored.

[0071] FIG. 15 is a correspondence table summarizing the feature values used to estimate lower limb muscle strength. The correspondence table in FIG. 15 associates the feature value number, the gait waveform data from which the feature value is extracted, the gait phase (%) from which the gait phase cluster is extracted, and the associated muscles. The standing-sitting time is correlated with the quadriceps, hamstrings, tibialis anterior, and gastrocnemius. Therefore, feature values C1 to C4 extracted from the gait phases in which these features appear are used to estimate the standing-sitting time.

[0072] The feature C1 is extracted from the section of the walking phase 42 to 54% of the walking waveform data Gx related to the time series data of angular velocity in the sagittal plane (around the X axis). The walking phase 42 to 54% is the section from the end of stance phase T3 to the early swing phase T4. The feature C1 mainly includes features related to the movement of the gastrocnemius muscles.

[0073] Feature C2 is extracted from the 99% to 100% gait phase section of the walking waveform data Gy, which is related to the time series data of angular velocity in the coronal plane (around the Y-axis). The 99% to 100% gait phase corresponds to the end of the swing phase T7. Feature C2 mainly includes features related to the movements of the quadriceps, hamstrings, and tibialis anterior.

[0074] Feature C3 is extracted from the gait phase 10-12% section of the gait waveform data Gy, which is related to the time series data of angular velocity in the coronal plane (around the Y axis). The gait phase 10-12% corresponds to the beginning of mid-stance phase T2. Feature C3 mainly includes features related to the movements of the quadriceps, hamstrings, and gastrocnemius muscles.

[0075] Feature C4 is extracted from the 99% walking phase section of the walking waveform data Ez, which is related to time-series data of angles (postural angles) in the horizontal plane (around the Z axis). The 99% walking phase corresponds to the end of the swing phase T7. Feature C4 mainly includes features related to the movements of the quadriceps, hamstrings, and tibialis anterior.

[0076] <Related Item D> Related item D relates to mobility. Mobility can be evaluated by the results of a TUG (Time Up and Go) test. In this embodiment, the results of the TUG test are evaluated based on the time it takes to stand up from a chair, walk to a landmark 3 meters away, turn around, and sit back down in the chair (also called the TUG time). The TUG time is the score value of the TUG test. The shorter the TUG time, the higher the score on the TUG test. Related item D may also be evaluated by the results of a test related to mobility other than the TUG test.

[0077] The mobility index for related item D is the time required to complete the TUG. For example, an estimated value of the time required to complete the TUG is an index of mobility. For example, a score (also called a mobility score) according to the estimated value of the time required to complete the TUG is an index of mobility. The mobility score is a value obtained by scoring the time required to complete the TUG, which is an index of mobility, based on a preset standard. Mobility is affected by attributes such as age. Therefore, the mobility score may be scored based on a standard for each attribute. Note that the mobility index is not limited to the time required to complete the TUG, as long as mobility can be scored.

[0078] Figure 16 is a correspondence table summarizing the features used to estimate mobility. The correspondence table in Figure 16 associates feature numbers, gait waveform data from which the features are extracted, gait phases (%) from which gait phase clusters are extracted, and related muscles. The TUG duration is correlated with the quadriceps, gluteus medius, and tibialis anterior muscles. Therefore, feature values D1 to D6 extracted from gait phases in which these features appear are used to estimate the TUG duration. The characteristics of the tensor fasciae latae muscle appear in gait phases 0 to 45% and 85 to 100%. The characteristics of the gluteus medius muscle appear in gait phases 0 to 25%. The characteristics of the tibialis anterior muscle appear in gait phases 0 to 10% and 57 to 100%.

[0079] The feature value D1 is extracted from the section of the walking phase 64 to 65% of the walking waveform data Ax related to the time-series data of the lateral acceleration (X-direction acceleration). The walking phase 64 to 65% is included in the initial swing phase T5. The feature value D1 mainly includes features related to the movement of the quadriceps femoris during the standing-to-sitting motion.

[0080] Feature D2 is extracted from the 57% to 58% walking phase of the walking waveform data Gx, which is related to time-series data of angular velocity in the sagittal plane (around the X-axis). The 57% to 58% walking phase is included in the early swing phase T4. Feature D2 mainly includes features related to the movement of the quadriceps, which is related to the kick-off speed of the foot.

[0081] Feature D3 is extracted from the gait phase 19-20% section of the gait waveform data Gy, which is related to the time series data of angular velocity in the coronal plane (around the Y axis). The gait phase 19-20% is included in the mid-stance phase T2. Feature D3 mainly includes features related to the movement of the gluteus medius muscle during changes of direction.

[0082] Feature D4 is extracted from the gait phase 12-13% section of the gait waveform data Ez, which is related to time-series data of angular velocity in the horizontal plane (around the Z axis). The gait phase 12-13% corresponds to the beginning of mid-stance phase T2. Feature D4 mainly includes features related to the movement of the gluteus medius muscle during changes of direction.

[0083] Feature D5 is extracted from the gait phase 74-75% section of the walking waveform data Ez, which is related to the time series data of angular velocity in the horizontal plane (around the Z axis). The gait phase 74-75% corresponds to the beginning of the mid-swing phase T6. Feature D5 mainly includes features related to the movement of the tibialis anterior muscle when standing up, sitting down, and changing direction.

[0084] Feature D6 is extracted from the 76% to 80% gait phase of the gait waveform data Ey, which is related to the time series data of angles (postural angles) in the coronal plane (around the Y-axis). The 76% to 80% gait phase is included in the mid-swing phase T6. Feature D6 mainly includes features related to the movement of the tibialis anterior muscle when standing up, sitting down, and changing direction.

[0085] <Related Items E> Related item E relates to static balance. Static balance can be evaluated by the performance of a one-leg standing test. In this embodiment, the performance of the one-leg standing test is evaluated based on the time spent with one leg raised 5 cm (centimeters) from the ground with the eyes closed (also referred to as the one-leg standing time). The one-leg standing time is a static balance performance value. The longer the one-leg standing time, the higher the static balance performance. Related item E may also be evaluated by performance other than the one-leg standing test with eyes closed. For example, related item E may be evaluated by a one-leg standing test with the eyes open (one-leg standing test with eyes open) or other variations of the one-leg standing test.

[0086] The static balance index for related item E is the one-leg standing time. For example, an estimated value of the one-leg standing time is the static balance index. For example, a score (also called a static balance score) according to the estimated value of the one-leg standing time is the static balance index. The static balance score is a value obtained by scoring the one-leg standing time, which is an index of static balance, based on a preset standard. Static balance is affected by attributes such as age and height. Therefore, the static balance score may be scored based on a standard for each attribute. Note that the static balance index is not limited to the one-leg standing time, as long as static balance can be scored.

[0087] FIG. 17 is a correspondence table summarizing the feature quantities used in estimating static balance. The correspondence table in FIG. 17 associates the feature quantity number, the gait waveform data from which the feature quantity is extracted, the gait phase (%) from which the gait phase cluster is extracted, and the related muscles. The single-leg standing time is correlated with the gluteus medius, adductor longus, sartorius, and abductor and adductor muscles. Therefore, feature quantities E1 to E7 extracted from the gait phases in which these features appear are used to estimate the single-leg standing time.

[0088] The feature E1 is extracted from the gait phase 13-19% section of the gait waveform data Ax related to the time series data of lateral acceleration (X-direction acceleration). The gait phase 13-19% is included in the mid-stance phase T2. The feature E1 mainly includes features related to the movement of the gluteus medius muscle.

[0089] The feature E2 is extracted from the 95% walking phase of the walking waveform data Az related to the time-series data of vertical acceleration (Z-direction acceleration). The 95% walking phase corresponds to the end of the swing phase T7. The feature E2 mainly includes features related to the movement of the gluteus medius muscle.

[0090] Feature E3 is extracted from the gait phase 64-65% section of the gait waveform data Gy, which is related to the time series data of angular velocity in the coronal plane (around the Y-axis). The gait phase 64-65% is included in the initial swing phase T5. Feature E3 mainly includes features related to the movement of the adductor longus and sartorius muscles.

[0091] Feature E4 is extracted from the gait phase 11-16% section of the gait waveform data Gz, which is related to the time series data of angular velocity in the horizontal plane (around the Z axis). The gait phase 11-16% is included in the mid-stance phase T2. Feature E4 mainly includes features related to the movement of the gluteus medius muscle.

[0092] Feature E5 is extracted from the 57-58% section of the walking phase of the walking waveform data Gz, which is related to the time series data of angular velocity in the horizontal plane (around the Z axis). The 57-58% walking phase is included in the early swing phase T4. Feature E5 mainly includes features related to the movement of the adductor longus and sartorius muscles.

[0093] The feature quantity E6 is extracted from the 100% walking phase section of the walking waveform data Ez, which is related to time-series data of angles (postural angles) in the horizontal plane (around the Z-axis). The 100% walking phase corresponds to the timing of heel contact, when the phase switches from the end-swing phase T7 to the load response phase T1. The feature quantity of the walking waveform data Ez in the 100% walking phase corresponds to the foot angle when the sole of the foot is in contact with the ground. The feature quantity E6 mainly includes features related to the movement of the gluteus medius muscle.

[0094] Feature E7 is the distance (minutes of rotation) between the axis of forward movement and the foot at the time when the central axis of the foot is farthest from the axis of forward movement during the swing phase. Feature E7 is the minute of rotation normalized by the subject's height. Feature E7 mainly includes features related to the movement of the abductor and adductor muscles.

[0095] <Easy to fall over> FIG. 18 is a correspondence table summarizing the feature quantities used to estimate fall susceptibility. The correspondence table in FIG. 18 associates the feature quantity number, the gait waveform data from which the feature quantity is extracted, the gait phase (%) from which the gait phase cluster is extracted, and the associated muscles. Fall susceptibility can be estimated based on the features that appear in the gait phases common to all five of the above-mentioned five gait phases in which the features appear. The gait phases common to all five are the period from just before heel contact to just after foot contact, the period before and after toe-off, and the period before and after minimum toe clearance in the mid-swing phase. Therefore, feature quantities F1 to F15 extracted from these gait phases are used to estimate fall susceptibility.

[0096] Features F1 to F5 (also referred to as the first feature group) are extracted from the 0-13 and 95-100% gait phases. These periods correspond to the period from 95% of the preceding gait cycle to 13% of the following gait cycle (the 95-13% gait phase period). The 95-13% gait phase is the period from just before heel contact to just after foot contact (also referred to as the first period). The 95-13% gait phase spans from the end of swing phase T7 to the beginning of mid-stance phase T2. Features F1 to F5 include features related to the movements of the quadriceps (vastus lateralis, vastus intermedius, vastus medialis), hamstrings (semimembranosus, semitendinosus), tibialis anterior, tibialis posterior, and gluteus medius.

[0097] Feature F1 is extracted from gait waveform data Ax, which is related to time series data of lateral acceleration (X-direction acceleration). Feature F2 is extracted from gait waveform data Az, which is related to time series data of vertical acceleration (Z-direction acceleration). Feature F3 is extracted from gait waveform data Gy, which is related to time series data of angular velocity in the coronal plane (around the Y-axis). Feature F4 is extracted from gait waveform data Gz, which is related to time series data of angular velocity in the horizontal plane (around the Z-axis). Feature F5 is extracted from gait waveform data Ez, which is related to time series data of angles (posture angles) in the horizontal plane (around the Z-axis).

[0098] Features F6 to F12 (also referred to as the second feature group) are extracted from the 57-65% walking phase. The 57-65% walking phase is the period before and after toe-off (also referred to as the second period). The 57-65% walking phase spans from the early swing phase T4 to the early swing phase T5. Features F6 to F12 include features related to the movements of the iliopsoas muscle, quadriceps (rectus femoris), adductor longus, gracilis, sartorius, and tibialis anterior.

[0099] Feature F6 is extracted from gait waveform data Ax, which is related to time series data of lateral acceleration (X-direction acceleration). Feature F7 is extracted from gait waveform data Ay, which is related to time series data of forward acceleration (Y-direction acceleration). Feature F8 is extracted from gait waveform data Az, which is related to time series data of vertical acceleration (Z-direction acceleration). Feature F9 is extracted from gait waveform data Gx, which is related to time series data of angular velocity in the sagittal plane (around the X-axis). Feature F10 is extracted from gait waveform data Gy, which is related to time series data of angular velocity in the coronal plane (around the Y-axis). Feature F11 is extracted from gait waveform data Gz, which is related to time series data of angular velocity in the horizontal plane (around the Z-axis). Feature F12 is extracted from gait waveform data Ez, which is related to time series data of angles (posture angles) in the horizontal plane (around the Z-axis).

[0100] Features F13 to F15 (also called the third feature group) are extracted from the 74-80% gait phase. The 74-80% gait phase is the period (also called the third period) before and after the minimum toe clearance movement in the mid-swing phase. The 74-80% gait phase is included in mid-swing phase T6. Features F13 to F15 include features related to the movement of the hamstrings (short head of the biceps femoris), tibialis anterior, and gracilis.

[0101] The feature F13 is extracted from the gait waveform data Ay, which is related to the time series data of the acceleration in the forward direction (acceleration in the Y direction). The feature F14 is extracted from the gait waveform data Gz, which is related to the time series data of the angular velocity in the horizontal plane (around the Z axis). The feature F15 is extracted from the gait waveform data Ey, which is related to the time series data of the angle (posture angle) in the coronal plane (around the Y axis).

[0102] [Estimation example 1] FIG. 19 is a conceptual diagram showing an example in which feature quantities F1 to F15 extracted from sensor data measured as a user walks are input to estimation model 151 (first estimation model) constructed in advance to estimate fallability. Estimation model 151 outputs an estimated value of a fallability index (fallability score SF) in response to the input of feature quantities F1 to F15. For example, estimation model 151 is generated by learning using training data in which feature quantities F1 to F15 used to estimate fallability are used as explanatory variables and fallability is used as a target variable. There are no limitations on the estimation result of estimation model 151, as long as an estimation result regarding the fallability index (fallability score SF) is output in response to the input of feature quantity data for estimating fallability. For example, estimation model 151 may be a model that estimates fallability using attribute data as explanatory variables in addition to feature quantities F1 to F15 used to estimate fallability.

[0103] For example, an estimation model for estimating the likelihood of falling using a multiple regression prediction method is stored in storage unit 132. For example, parameters for estimating the likelihood of falling are stored in storage unit 132 using the following equation 1. Easy to fall = f1×F1+f2×F2+···+f15×F15+f0···(1) In the above formula 1, F1, F2, ..., F15 are feature quantities for each gait phase cluster used to estimate fall likelihood, as shown in the correspondence table of FIG. 18. In the above formula 1, F3 to F14 are omitted. f1, f2, ..., f15 are coefficients (weights) by which F1, F2, ..., F15 are multiplied. f0 is a constant term. In the above formula 1, f3 to f14 are omitted. For example, f0, f1, ..., f15 are stored in the storage unit 132.

[0104] [Estimation example 2] 20 is a conceptual diagram showing an example in which the output of estimation model 152 (pre-estimation model) that estimates scores for five items is input to estimation model 153 (second estimation model) that is constructed in advance to estimate fall proneness. Estimation model 152 includes estimation model 152A, estimation model 152B, estimation model 152C, estimation model 152D, and estimation model 152E.

[0105] The estimation model 152A outputs a score (total muscle strength score SA) related to the total muscle strength (grip strength) of the whole body in response to input of feature amounts AM1 to AM4 or feature amounts AF1 to AF3 extracted from sensor data measured as the user walks. For example, the estimation model 152A may be a different model for men and a different model for women. There are no limitations on the estimation result of the estimation model 152A as long as an estimation result related to an index of total muscle strength is output in response to input of feature amount data for estimating total muscle strength. For example, the estimation model 152A may be a model that estimates dynamic balance using attribute data such as age and height as explanatory variables in addition to feature amounts AM1 to AM4 or feature amounts AF1 to AF3.

[0106] For example, an estimation model 152A that estimates the total muscle strength score SA using a multiple regression prediction method is stored in the storage unit 132. For example, parameters for estimating the total muscle strength score SA are stored in the storage unit 132 using the following Equation 2. SA=m(am1×AM1+am2×AM2+am3×AM3+am4×AM4+am0)+f(af1×AF1+af2×AF2+af3×AF3+af0)...(2) In the above formula 2, AM1, AM2, AM3, and AM4 are feature quantities for each walking phase cluster used to estimate the total muscle strength score SA of a male shown in the correspondence table of FIG. 13. am1, am2, am3, and am4 are coefficients (weights) by which AM1, AM2, AM3, and AM4 are multiplied. am0 is a constant term. AF1, AF2, and AF3 are feature quantities for each walking phase cluster used to estimate the total muscle strength score SA of a female shown in the correspondence table of FIG. 13. af1, af2, and af3 are coefficients (weights) by which AF1, AF2, and AF3 are multiplied. af0 is a constant term. m and f are flags according to gender. If the user is male, m is 1 and f is 0. If the user is female, m is 0 and f is 1. For example, am0, am1, am2, am3, am4, af0, af1, af2, and af3 are stored in the storage unit 132.

[0107] The estimation model 152B outputs a score related to dynamic balance (dynamic balance score SB) in response to input of feature amounts B1 to B5 extracted from sensor data measured as the user walks. There are no limitations on the estimation result of the estimation model 152B as long as an estimation result related to a dynamic balance index is output in response to input of feature amount data for estimating dynamic balance. For example, the estimation model 152B may be a model that estimates dynamic balance using attribute data such as height as explanatory variables in addition to the feature amounts B1 to B5.

[0108] For example, an estimation model for estimating the dynamic balance score SB using a multiple regression prediction method is stored in the storage unit 132. For example, parameters for estimating the dynamic balance score SB using the following Equation 3 are stored in the storage unit 132. SB=b1×B1+b2×B2+b3×B3+b4×B4+b5×B5+b0...(3) In the above formula 3, B1, B2, B3, B4, and B5 are feature quantities for each walking phase cluster used to estimate dynamic balance, as shown in the correspondence table of FIG. 14. b1, b2, b3, b4, and b5 are coefficients (weights) by which B1, B2, B3, B4, and B5 are multiplied. b0 is a constant term. For example, b0, b1, b2, b3, b4, and b5 are stored in the storage unit 132.

[0109] The estimation model 152C outputs a score related to lower limb muscle strength (lower limb muscle strength score SC) in response to input of feature amounts C1 to C4 extracted from sensor data measured as the user walks. There are no limitations on the estimation result of the estimation model 152C as long as an estimation result related to an index of lower limb muscle strength is output in response to input of feature amount data for estimating lower limb muscle strength. For example, the estimation model 152C may be a model that estimates dynamic balance using attribute data such as age as explanatory variables in addition to the feature amounts C1 to C4.

[0110] For example, an estimation model for estimating the lower limb muscle strength score SC using a multiple regression prediction method is stored in the storage unit 132. For example, parameters for estimating the lower limb muscle strength score SC using the following Equation 4 are stored in the storage unit 132. SC=c1×C1+c2×C2+c3×C3+c4×C4+c0...(4) In the above formula 4, C1, C2, C3, and C4 are feature quantities for each walking phase cluster used to estimate lower limb muscle strength, as shown in the correspondence table of FIG. 15. c1, c2, c3, and c4 are coefficients (weights) by which C1, C2, C3, and C4 are multiplied. c0 is a constant term. For example, c0, c1, c2, c3, and c4 are stored in the storage unit 132.

[0111] The estimation model 152D outputs a score related to mobility (mobility score SD) in response to input of feature amounts D1 to D6 extracted from sensor data measured as the user walks. There are no limitations on the estimation result of the estimation model 152D as long as an estimation result related to a mobility index is output in response to input of feature amount data for estimating mobility. For example, the estimation model 152D may be a model that estimates mobility using attribute data such as age as an explanatory variable in addition to the feature amounts D1 to D6.

[0112] For example, an estimation model for estimating the mobility score SD using a multiple regression prediction method is stored in the storage unit 132. For example, parameters for estimating the mobility score SD are stored in the storage unit 132 using the following Equation 5. SD=d1×D1+d2×D2+d3×D3+d4×D4+d5×D5+d6×D6+d0...(5) In the above formula 5, D1, D2, D3, D4, D5, and D6 are feature quantities for each walking phase cluster used to estimate mobility, as shown in the correspondence table of FIG. 16. d1, d2, d3, d4, d5, and d6 are coefficients (weights) by which D1, D2, D3, D4, D5, and D6 are multiplied. d0 is a constant term. For example, d0, d1, d2, d3, d4, d5, and d6 are stored in the storage unit 132.

[0113] The estimation model 152E outputs a score related to static balance (static balance score SE) in response to input of feature amounts E1 to E7 extracted from sensor data measured as the user walks. There are no limitations on the estimation result of the estimation model 152E as long as an estimation result related to a static balance index is output in response to input of feature amount data for estimating static balance. For example, the estimation model 152E may be a model that estimates static balance using attribute data such as age and height as explanatory variables in addition to the feature amounts E1 to E7.

[0114] For example, an estimation model for estimating static balance using a multiple regression prediction method is stored in the storage unit 132. For example, parameters for estimating static balance are stored in the storage unit 132 using the following Equation 6. Single leg standing time = e1 × E1 + e2 × E2 + e3 × E3 + e4 × E4 + e5 × E5 + e6 × E6 + e7 × E7 + e0 (6) In the above equation 6, E1, E2, E3, E4, E5, E6, and E7 are feature quantities for each walking phase cluster used to estimate static balance, as shown in the correspondence table of FIG. 17. e1, e2, e3, e4, e5, e6, and e7 are coefficients (weights) by which E1, E2, E3, E4, E5, E6, and E7 are multiplied. e0 is a constant term. For example, e0, e1, e2, e3, e4, e5, e6, and e7 are stored in the storage unit 132.

[0115] The estimation model 153 (second estimation model) outputs a score related to the tendency to fall (a falling tendency score SF) in response to input of the scores output from the estimation model 152. That is, the estimation model 153 outputs the falling tendency score SF in response to input of the scores of the five items. At least one of the scores output from the estimation model 152A, the estimation model 152B, the estimation model 152C, the estimation model 152D, and the estimation model 152E is input to the estimation model 153. That is, it is sufficient that at least one of the scores output from the estimation model 152A, the estimation model 152B, the estimation model 152C, the estimation model 152D, and the estimation model 152E is input to the estimation model 153. The more scores input to the estimation model 153, the more accurately the tendency to fall can be estimated. As long as an estimation result related to the tendency to fall is output in response to input of the scores output from the estimation model 152, there are no limitations on the estimation result of the estimation model 153. For example, the estimation model 153 may be a model that estimates the tendency to fall using attribute data as explanatory variables in addition to the score output from the estimation model 152.

[0116] For example, an estimation model for estimating the fall tendency score SF using a multiple regression prediction method is stored in the storage unit 132. For example, parameters for estimating the fall tendency score SF are stored in the storage unit 132 using the following Equation 7. SF=sa×SA+sb×SB+sc×SC+sd×SD+se×SE+sf...(7) In the above formula 7, SA, SB, SC, SD, and SE are scores output from estimation model 152A, estimation model 152B, estimation model 152C, estimation model 152D, and estimation model 152E included in the estimation model 152. sa, sb, sc, sd, and se are coefficients (weights) by which SA, SB, SC, SD, and SE are multiplied. sf is a constant term. For example, sa, sb, sc, sd, se, and sf are stored in the storage unit 132.

[0117] (operation) Next, the operation of fall tendency estimation system 1 will be described with reference to the drawings. Below, estimation example 1 (FIG. 19) and estimation example 2 (FIG. 20) will be described separately. Also below, gait measurement device 10 and fall tendency estimation device 13 included in fall tendency estimation system 1 will be described separately. With regard to gait measurement device 10, the operation of feature amount data generation unit 12 included in gait measurement device 10 will be described.

[0118] [Estimation example 1] Fig. 21 is a flowchart for explaining the operation of the feature amount data generation unit 12 included in the gait measurement device 10 in estimation example 1 (Fig. 19). In the explanation following the flowchart of Fig. 21, the feature amount data generation unit 12 will be described as the subject of the operation.

[0119] In FIG. 21, first, the feature amount data generator 12 acquires time-series data of sensor data relating to foot movements (step S101).

[0120] Next, the feature amount data generating unit 12 extracts gait waveform data for one gait cycle from the time series data of the sensor data (step S102). The feature amount data generating unit 12 detects heel strikes and toe lifts from the time series data of the sensor data. The feature amount data generating unit 12 extracts time series data of the sections between successive heel strikes as gait waveform data for one gait cycle.

[0121] Next, the feature amount data generating unit 12 normalizes the extracted walking waveform data for one step cycle (step S103). The feature amount data generating unit 12 normalizes the walking waveform data for one step cycle to a walking cycle of 0 to 100% (first normalization). Furthermore, the feature amount data generating unit 12 normalizes the ratio of the stance phase to the swing phase of the first normalized walking waveform data for one step cycle to 60:40 (second normalization).

[0122] Next, the feature data generating unit 12 extracts feature amounts from the normalized walking waveform from the walking phases used to estimate the likelihood of falling (step S104). For example, the feature data generating unit 12 extracts feature amounts to be input to a pre-constructed estimation model (first estimation model).

[0123] Next, the feature data generator 12 uses the extracted feature to generate a feature for each walking phase cluster (step S105).

[0124] Next, the feature amount data generator 12 integrates the feature amounts for each walking phase cluster to generate feature amount data for one walking cycle (step S106).

[0125] Next, the feature amount data generating unit 12 outputs the generated feature amount data to the fall tendency estimating device 13 (step S107).

[0126] Fig. 22 is a flowchart for explaining the operation of the fall tendency estimating device 13 in estimation example 1 (Fig. 19). In the explanation following the flowchart of Fig. 22, the fall tendency estimating device 13 will be described as the subject of the operation.

[0127] In FIG. 22, first, the fall tendency estimating device 13 acquires feature amount data generated using sensor data related to foot movements (step S111).

[0128] Next, the fall tendency estimation device 13 inputs the acquired feature amount data into an estimation model (first estimation model) that estimates the fall tendency (single-leg standing time) (step S112).

[0129] Next, the fall tendency estimation device 13 estimates the fall tendency of the user according to the output (estimated value) from the estimation model (first estimation model) (step S113). For example, the fall tendency estimation device 13 estimates the one-leg standing time of the user as the fall tendency.

[0130] Next, the fall tendency estimation device 13 outputs information about the estimated fall tendency (step S114). For example, the fall tendency is output to a terminal device (not shown) carried by the user. For example, the fall tendency is output to a system that executes processing using the fall tendency.

[0131] [Estimation example 2] Fig. 23 is a flowchart for explaining the operation of the feature amount data generation unit 12 included in the gait measurement device 10 in estimation example 2 (Fig. 20). In the explanation following the flowchart in Fig. 23, the feature amount data generation unit 12 will be described as the subject of the operation.

[0132] In FIG. 23, first, the feature amount data generator 12 acquires time-series data of sensor data relating to foot movements (step S121).

[0133] Next, the feature amount data generating unit 12 extracts gait waveform data for one gait cycle from the time series data of the sensor data (step S122). The feature amount data generating unit 12 detects heel strikes and toe lifts from the time series data of the sensor data. The feature amount data generating unit 12 extracts time series data of the sections between successive heel strikes as gait waveform data for one gait cycle.

[0134] Next, the feature amount data generating unit 12 normalizes the extracted walking waveform data for one step cycle (step S123). The feature amount data generating unit 12 normalizes the walking waveform data for one step cycle to a walking cycle of 0 to 100% (first normalization). Furthermore, the feature amount data generating unit 12 normalizes the ratio of the stance phase to the swing phase of the first normalized walking waveform data for one step cycle to 60:40 (second normalization).

[0135] Next, the feature data generator 12 extracts, from the normalized walking waveform, feature amounts used to estimate scores for five items, namely, total muscle strength of the whole body, dynamic balance, lower limb muscle strength, mobility, and static balance (step S124). For example, the feature data generator 12 extracts feature amounts to be input to a pre-constructed estimation model (pre-estimation model).

[0136] Next, the feature data generator 12 uses the extracted feature to generate a feature for each walking phase cluster (step S125).

[0137] Next, the feature amount data generator 12 integrates the feature amounts for each walking phase cluster to generate feature amount data for one walking cycle (step S126).

[0138] Next, the feature amount data generating unit 12 outputs the generated feature amount data to the fall tendency estimating device 13 (step S127).

[0139] Fig. 24 is a flowchart for explaining the operation of the fall tendency estimating device 13 in estimation example 2 (Fig. 20). In the explanation following the flowchart of Fig. 24, the fall tendency estimating device 13 will be described as the subject of the operation.

[0140] 24, first, the fall tendency estimation device 13 acquires feature data generated using sensor data related to foot movement and used to estimate scores for five items related to fall tendency (step S131). The five items related to fall tendency are total muscle strength of the whole body, dynamic balance, lower limb muscle strength, mobility, and static balance.

[0141] Next, the fall tendency estimation device 13 inputs the acquired feature amount data into an estimation model (pre-estimation model) including an estimation model for each item (step S132).

[0142] Next, the fall tendency estimation device 13 inputs the scores output from the estimation model for each item included in the estimation model (pre-estimation model) into the fall tendency estimation model (second estimation model) (step S133 ).

[0143] Next, the fall tendency estimation device 13 estimates the user's fall tendency according to the output (estimated value) from the fall tendency estimation model (second estimation model) (step S134). For example, the fall tendency estimation device 13 estimates the user's fall tendency score using the sum of the scores output from the estimation model for each item. For example, the fall tendency estimation device 13 estimates the user's fall tendency score using the average value of the scores output from the estimation model for each item. For example, the fall tendency estimation device 13 estimates the user's fall tendency score by weighting the scores output from the estimation model for each item for each item.

[0144] Next, the fall tendency estimation device 13 outputs information about the estimated fall tendency (step S135). For example, the fall tendency is output to a terminal device (not shown) carried by the user. For example, the fall tendency is output to a system that executes processing using the fall tendency.

[0145] (Application example) Next, application examples according to this embodiment will be described with reference to the drawings. In the following application examples, a function of a fall tendency estimation device 13 installed on a mobile device carried by a user is used to estimate information related to fall tendency using feature amount data measured by a gait measurement device 10 placed on a shoe.

[0146] 25 and 26 are conceptual diagrams showing an example of displaying the estimation results by fall tendency estimation device 13 on the screen of mobile terminal 160 carried by a user walking while wearing shoes 100 on which gait measurement device 10 is placed. Figs. 25 and 26 show an example of displaying information corresponding to the estimation results of fall tendency using feature amount data corresponding to sensor data measured while the user is walking on the screen of mobile terminal 160.

[0147] FIG. 25 shows an example in which a fall susceptibility score SF corresponding to the fall susceptibility estimation result is displayed on the screen of mobile device 160. In the example of FIG. 25, a radar chart corresponding to the score value used to estimate fall susceptibility score SF is displayed on the screen of mobile device 160. A user who checks the value of fall susceptibility score SF displayed on the display unit of mobile device 160 can recognize their own fall susceptibility based on the value of fall susceptibility score SF. In the example of FIG. 25, the static balance score SE is low. A user who checks the radar chart displayed on the display unit of mobile device 160 can recognize that their static balance score SE is low. Information about the estimated fall susceptibility may be provided to a party other than the user. For example, the information about the estimated fall susceptibility may be output to a terminal device (not shown) used by a trainer who manages the user's physical condition, a family member of the user, or the like. For example, the information about the estimated fall susceptibility may be recorded in a database (not shown) constructed for purposes such as health management.

[0148] FIG. 26 illustrates another example in which information about fall susceptibility according to the fall susceptibility estimation result is displayed on the screen of mobile device 160. In the example of FIG. 26, information about fall susceptibility, such as "Your risk of falling is increasing," is displayed on the screen of mobile device 160 according to the fall susceptibility estimation result. Also, in the example of FIG. 26, information about static balance, which was the lowest score among the scores for the five items in which the fall susceptibility estimation was also used, is displayed on the display unit of mobile device 160. In the example of FIG. 26, information about static balance, such as "Your static balance is declining," is displayed on the display unit of mobile device 160 according to the estimated value of static balance score SE. Furthermore, in the example of FIG. 26, recommendation information according to the static balance estimation result, such as "Training Z is recommended. Please watch the video below," is displayed on the display unit of mobile device 160 according to the estimated value of static balance score SE. After checking the information displayed on the display unit of mobile device 160, the user can practice training that will improve their static balance by watching the video of Training Z and exercising in accordance with the recommendation information.

[0149] As described above, the fall liability estimation system of this embodiment includes a gait measurement device and a fall liability estimation device. The gait measurement device includes a sensor and a feature data generation unit. The sensor has an acceleration sensor and an angular velocity sensor. The sensor measures spatial acceleration using the acceleration sensor. The sensor measures spatial angular velocity using the angular velocity sensor. The sensor generates sensor data related to foot movement using the measured spatial acceleration and spatial angular velocity. The sensor outputs the generated sensor data to the feature data generation unit. The feature data generation unit acquires time-series data of the sensor data related to foot movement. The feature data generation unit extracts gait waveform data for one walking cycle from the time-series data of the sensor data. The feature data generation unit normalizes the extracted gait waveform data. From the normalized gait waveform data, the feature data generation unit extracts feature amounts used for estimating fall liability from a gait phase cluster composed of at least one gait phase that is consecutive in time. The feature data generation unit generates feature data including the extracted feature amounts. The feature data generation unit outputs the generated feature data.

[0150] The fall susceptibility estimation device includes a data acquisition unit, a memory unit, an estimation unit, and an output unit. The data acquisition unit acquires feature data including features used to estimate the user's fall susceptibility, extracted from sensor data related to the user's foot movements. The memory unit stores an estimation model that outputs a fall susceptibility index according to input of the feature data. The estimation unit inputs the acquired feature data to the estimation model. The estimation unit estimates the user's fall susceptibility according to the fall susceptibility index output from the estimation model. The output unit outputs information related to the estimated fall susceptibility.

[0151] The fall susceptibility estimation system of this embodiment estimates a user's fall susceptibility using feature amounts extracted from sensor data related to the user's foot movements. Therefore, the fall susceptibility estimation system of this embodiment can appropriately estimate a user's fall susceptibility in daily life without using any equipment for estimating fall susceptibility.

[0152] In one aspect of this embodiment, the data acquisition unit acquires feature data including feature amounts used to estimate a fall liability score as a fall liability index, the feature amount being extracted from gait waveform data generated using time-series data of sensor data related to foot movement. The storage unit stores an estimation model that outputs a fall liability score in response to input of the feature amount data. The estimation unit inputs the acquired feature amount data to the estimation model and estimates the user's fall liability in response to the fall liability score output from the estimation model. According to this aspect, fall liability can be appropriately estimated in response to the fall liability score estimated using the sensor data related to foot movement.

[0153] In one aspect of this embodiment, the storage unit stores an estimation model generated by learning using training data for a plurality of subjects, in which feature values used to estimate a fall susceptibility index are used as explanatory variables and the fall susceptibility indexes of the plurality of subjects are used as objective variables. The estimation unit inputs feature value data acquired about a user into the estimation model and estimates the user's fall susceptibility based on the user's fall susceptibility index output from the estimation model. According to this aspect, fall susceptibility can be appropriately estimated using the estimation model trained with the training data for the plurality of subjects.

[0154] In one aspect of this embodiment, the storage unit stores an estimation model trained using explanatory variables including attribute data of multiple subjects. The estimation unit inputs feature data and attribute data related to the user into the estimation model and estimates the user's fallability based on the user's fallability index output from the estimation model. In this aspect, fallability is estimated including attribute data that influences fallability. Therefore, according to this aspect, fallability can be measured with higher accuracy based on the user's attributes.

[0155] In one aspect of this embodiment, the storage unit stores a first estimation model generated by learning using training data for gait waveform data of multiple subjects, with muscle activity feature values extracted from specific sections as explanatory variables and a fall liability index as a response variable. The specific sections include at least one of a first section, a second section, and a third section. The first section extends from just before heel strike to just after foot contact. From the first section, feature values related to the activities of the quadriceps, hamstrings, tibialis anterior, tibialis posterior, and gluteus medius are extracted. The second section extends from the early swing phase to the early swing phase. From the second section, feature values related to the activities of the iliopsoas, quadriceps, adductor longus, gracilis, sartorius, and tibialis anterior are extracted. The third section extends before and after the minimum toe clearance movement in the mid-swing phase. From the third section, feature values related to the activities of the hamstrings, tibialis anterior, and gracilis are extracted. The estimation unit inputs feature data acquired according to the user's walking into a first estimation model, and estimates the user's fallability according to the user's fallability index output from the first estimation model. According to this aspect, by using an estimation model that has learned feature values according to muscle activity that affects the fallability, it is possible to estimate the fallability that is more suited to physical activity.

[0156] In one aspect of this embodiment, the storage unit stores a first estimation model generated by learning using training data in which at least one feature included in the first feature group, the second feature group, and the third feature group is used as an explanatory variable and a fall susceptibility index is used as a target variable. The first feature group is composed of at least one feature extracted from the first section. The second feature group is composed of at least one feature extracted from the second section. The third feature group is composed of at least one feature extracted from the third section. The data acquisition unit acquires feature data extracted according to the user's gait, including at least one feature included in the first feature group, the second feature group, and the third feature group. The estimation unit inputs the acquired feature data into the first estimation model and estimates the user's fall susceptibility based on the user's fall susceptibility index output from the first estimation model. According to this aspect, by using feature values extracted from a walking phase in which feature values corresponding to muscle activity that affect fall susceptibility are extracted, a fall susceptibility that is more closely matched to physical activity can be estimated.

[0157] In one aspect of this embodiment, the storage unit stores a first estimation model generated by learning using training data in which feature values extracted from gait phases common to five items related to fall susceptibility are used as explanatory variables and a fall susceptibility index of multiple subjects is used as a target variable. The five items are total body muscle strength, dynamic balance, lower limb muscle strength, mobility, and static balance. The data acquisition unit acquires feature data including at least one feature value included in a first feature group, a second feature group, and a third feature group extracted according to the user's gait. The estimation unit inputs the acquired feature data into the first estimation model and estimates the user's fall susceptibility based on the user's fall susceptibility index output from the first estimation model. In this aspect, an estimation model is used that has learned feature values extracted from gait waveform data that includes features corresponding to muscle activity that affects fall susceptibility. Therefore, this aspect enables a fall susceptibility estimation that is more suited to physical activity.

[0158] In one aspect of this embodiment, the storage unit stores a second estimation model generated by learning using training data in which at least one score of five items estimated using gait waveform data of multiple subjects is used as an explanatory variable and the fall susceptibility index is used as a target variable. The data acquisition unit acquires at least one of the scores related to the five items estimated according to the user's gait. The estimation unit inputs the acquired score into the second estimation model and estimates the user's fall susceptibility according to the user's fall susceptibility index output from the second estimation model. According to this aspect, the scores related to the five items can be used to appropriately estimate fall susceptibility.

[0159] In one aspect of this embodiment, the storage unit stores a pre-estimation model generated by learning using training data in which feature quantities related to at least one of the five items are used as explanatory variables and scores of the five items corresponding to the feature quantities used as explanatory variables are used as objective variables. The data acquisition unit acquires feature quantities related to at least one of the five items extracted according to the user's gait. The estimation unit inputs the acquired feature quantities into the pre-estimation model, inputs the scores output from the pre-estimation model into a second estimation model, and estimates the user's fall susceptibility based on the user's fall susceptibility index output from the second estimation model. According to this aspect, the scores related to the five items estimated using sensor data related to foot movement can be used to appropriately estimate fall susceptibility.

[0160] In one aspect of the present embodiment, the fall susceptibility estimation device is implemented in a terminal device having a screen viewable by a user. For example, the fall susceptibility estimation device displays information related to fall susceptibility estimated based on sensor data related to the user's foot movement on the screen of the terminal device. For example, the fall susceptibility estimation device displays recommendation information corresponding to the fall susceptibility estimated based on the sensor data related to the user's foot movement on the screen of the terminal device. For example, the fall susceptibility estimation device displays a video related to training for strengthening body parts related to fall susceptibility on the screen of the terminal device as recommendation information corresponding to the fall susceptibility estimated based on the sensor data related to the user's foot movement. According to this aspect, the fall susceptibility estimated based on features extracted from the sensor data related to the user's foot movement is displayed on a screen viewable by the user, allowing the user to check information corresponding to their own fall susceptibility.

[0161] (Second embodiment) Next, a learning system according to a second embodiment will be described with reference to the drawings. The learning system of this embodiment generates an estimation model for estimating fall tendency in response to input of feature values by learning using feature value data extracted from sensor data measured by a gait measurement device.

[0162] (composition) FIG. 27 is a block diagram showing an example of the configuration of learning system 2 according to this embodiment. Learning system 2 includes gait measurement device 20 and learning device 25. Gait measurement device 20 and learning device 25 may be connected by wire or wirelessly. Gait measurement device 20 and learning device 25 may be configured as a single device. Alternatively, gait measurement device 20 may be removed from the configuration of learning system 2, and learning system 2 may be configured with only learning device 25. Although FIG. 27 shows only one gait measurement device 20, one gait measurement device 20 may be provided for each foot (two in total). Alternatively, learning device 25 may be configured to perform learning without being connected to gait measurement device 20, using feature data that has been generated in advance by gait measurement device 20 and stored in a database.

[0163] The gait measurement device 20 is installed on at least one of the left and right feet. The gait measurement device 20 has the same configuration as the gait measurement device 10 of the first embodiment. The gait measurement device 20 includes an acceleration sensor and an angular velocity sensor. The gait measurement device 20 converts measured physical quantities into digital data (also referred to as sensor data). The gait measurement device 20 generates gait waveform data for a normalized stride cycle from the time-series data of the sensor data. The gait measurement device 20 generates feature data used to estimate fall susceptibility. For example, the gait measurement device 20 generates feature data used to estimate a fall susceptibility score SF. For example, the gait measurement device 20 generates feature data used to estimate scores for five items: total whole-body muscle strength, dynamic balance, lower limb muscle strength, mobility, and static balance. The gait measurement device 20 transmits the generated feature data to the learning device 25. The gait measurement device 20 may be configured to transmit feature data to a database (not shown) accessed by the learning device 25. The feature data stored in the database is used for learning by the learning device 25.

[0164] The learning device 25 receives feature data from the gait measurement device 20. When feature data stored in a database (not shown) is used, the learning device 25 receives the feature data from the database. The learning device 25 performs learning using the received feature data. For example, the learning device 25 learns training data in which feature data extracted from gait waveform data of multiple subjects is used as an explanatory variable and a fall tendency score SF corresponding to the feature data is used as a response variable. For example, the learning device 25 learns training data in which feature data extracted from gait waveform data of multiple subjects is used as an explanatory variable and scores of five items corresponding to the feature data are used as response variables. For example, the learning device 25 learns training data in which at least one of the scores of the five items is used as an explanatory variable and a fall tendency score SF corresponding to the score is used as a response variable. There are no particular limitations on the learning algorithm executed by the learning device 25. The learning device 25 generates an estimation model trained using training data related to multiple subjects. The learning device 25 stores the generated estimation model. The estimation model learned by the learning device 25 may be stored in a storage device external to the learning device 25.

[0165] [Learning device] Next, details of the learning device 25 will be described with reference to the drawings. Fig. 28 is a block diagram showing an example of a detailed configuration of the learning device 25. The learning device 25 has a receiving unit 251, a learning unit 253, and a storage unit 255.

[0166] Receiving unit 251 receives feature amount data from gait measurement device 20. Receiving unit 251 outputs the received feature amount data to learning unit 253. Receiving unit 251 may receive feature amount data from gait measurement device 20 via a wired connection such as a cable, or may receive feature amount data from gait measurement device 20 via wireless communication. For example, receiving unit 251 is configured to receive feature amount data from gait measurement device 20 via a wireless communication function (not shown) that complies with standards such as Bluetooth (registered trademark) or WiFi (registered trademark). Note that the communication function of receiving unit 251 may be compliant with standards other than Bluetooth (registered trademark) or WiFi (registered trademark).

[0167] The learning unit 253 acquires feature data from the receiving unit 251. The learning unit 253 performs learning using the acquired feature data. For example, the learning unit 253 uses feature data extracted from sensor data measured according to the foot movement of the subject as explanatory variables and the subject's fall tendency score SF as a response variable to learn a data set as training data. For example, the learning unit 253 uses feature data extracted from sensor data measured according to the foot movement of the subject as explanatory variables and the subject's scores on five items as response variables to learn a data set as training data. For example, the learning unit 253 uses at least one of the scores on the five items as an explanatory variable and learns training data that uses the fall tendency score SF corresponding to the score as a response variable. For example, the learning unit 253 generates an estimation model according to the attribute data. For example, the learning unit 253 generates an estimation model that estimates the fall tendency score SF using feature data extracted from sensor data measured according to the foot movement of the subject and the subject's attribute data as explanatory variables. The learning unit 253 stores the estimation model learned for a plurality of subjects in the storage unit 255.

[0168] For example, the learning unit 253 performs learning using a linear regression algorithm. For example, the learning unit 253 performs learning using a support vector machine (SVM) algorithm. For example, the learning unit 253 performs learning using a Gaussian process regression (GPR) algorithm. For example, the learning unit 253 performs learning using a random forest (RF) algorithm. For example, the learning unit 253 may perform unsupervised learning that classifies the subjects that generated feature amount data according to the feature amount data. There are no particular limitations on the learning algorithm performed by the learning unit 253.

[0169] The learning unit 253 may perform learning using walking waveform data for one step gait cycle as explanatory variables. For example, the learning unit 253 performs supervised learning using walking waveform data of acceleration in three axial directions, angular velocity around three axes, and angle around three axes (posture angle) as explanatory variables, and the correct value of the fall tendency index as the objective variable. For example, if the walking phase is set in 1% increments in a walking cycle from 0 to 100%, the learning unit 253 performs learning using 909 explanatory variables.

[0170] Fig. 29 is a conceptual diagram for explaining learning for generating an estimation model (first estimation model). Fig. 29 is a conceptual diagram showing an example in which the learning unit 253 learns using, as training data, a data set of feature amounts F1 to F15, which are explanatory variables, and a fall tendency score SF (fall tendency index), which is a response variable. For example, the learning unit 253 learns data on a plurality of subjects and generates an estimation model (first estimation model) that outputs an output (estimated value) related to the fall tendency score SF (fall tendency index) in response to input of feature amounts extracted from sensor data.

[0171] FIG. 30 is a conceptual diagram illustrating learning for generating an estimation model (second estimation model). FIG. 30 is a conceptual diagram illustrating an example in which the learning unit 253 learns a data set of five item scores as explanatory variables and a fall susceptibility score SF (fall susceptibility index) as a response variable, as training data. In the example of FIG. 30, training data is used in which the whole-body total muscle strength score SA, the dynamic balance score SB, the lower limb muscle strength score SC, the mobility score SD, and the static balance score SE are used as explanatory variables, and the fall susceptibility score SF (fall susceptibility index) is used as a response variable. For example, the learning unit 253 learns data on multiple subjects and generates an estimation model (second estimation model) that outputs an estimated value for the fall susceptibility score SF (fall susceptibility index) in response to input of feature amounts extracted from sensor data. A detailed description of learning for generating a pre-estimation model will be omitted. In the learning for generating a pre-estimation model, multiple estimation models included in the pre-estimation model may be generated individually, or multiple estimation models may be generated collectively.

[0172] The storage unit 255 stores estimation models for estimating fall susceptibility that have been learned for a plurality of subjects. The estimation models stored in the storage unit 255 are used to estimate fall susceptibility by the fall susceptibility estimation device 13 of the first embodiment.

[0173] As described above, the learning system of this embodiment includes a gait measurement device and a learning device. The gait measurement device acquires time-series sensor data related to foot movement. The gait measurement device extracts gait waveform data for one step cycle from the time-series sensor data and normalizes the extracted gait waveform data. From the normalized gait waveform data, the gait measurement device extracts feature amounts used to estimate the user's fall tendency from a walking phase cluster composed of at least one temporally consecutive walking phase. The gait measurement device generates feature amount data including the extracted feature amounts. The gait measurement device outputs the generated feature amount data to the learning device.

[0174] The learning device has a receiving unit, a learning unit, and a storage unit. The receiving unit acquires feature data generated by the gait measurement device. The learning unit performs learning using the feature data. The learning unit generates an estimation model (first estimation model) that outputs fall likelihood in response to input of features (second feature values) of gait phase clusters extracted from time-series data of sensor data measured as the user walks. The estimation model (first estimation model) generated by the learning unit is stored in the storage unit.

[0175] The learning system of this embodiment generates an estimation model (first estimation model) using feature data measured by a gait measurement device. Therefore, according to this aspect, it is possible to generate an estimation model that allows appropriate estimation of fall proneness in daily life without using any equipment for estimating fall proneness.

[0176] In one aspect of this embodiment, the gait measurement device extracts feature quantities related to at least one of five items, namely, total body muscle strength, dynamic balance, lower limb muscle strength, mobility, and static balance, from the normalized gait waveform data. For example, the learning unit generates an estimation model (pre-estimation model) by learning using training data in which the feature quantities related to at least one of the five items are used as explanatory variables and the scores of the five items corresponding to the feature quantities used as explanatory variables are used as objective variables. The learning unit generates an estimation model (second estimation model) that outputs a fall tendency index in response to input of a score related to at least one of the five items. According to this aspect, it is possible to generate an estimation model that enables appropriate estimation of fall tendency in response to input of scores related to the five items.

[0177] (Third embodiment) Next, a fall tendency estimation device according to a third embodiment will be described with reference to the drawings. The fall tendency estimation device of this embodiment has a simplified configuration of the fall tendency estimation device included in the fall tendency estimation system of the first embodiment.

[0178] 31 is a block diagram showing an example of the configuration of a fall tendency estimation device 33 according to this embodiment. The fall tendency estimation device 33 includes a data acquisition unit 331, a storage unit 332, an estimation unit 333, and an output unit 335.

[0179] The data acquisition unit 331 acquires feature data including feature amounts used to estimate the user's fall susceptibility index, extracted from sensor data related to the user's foot movement. The storage unit 332 stores an estimation model that outputs a fall susceptibility index according to input of the feature data. The estimation unit 333 inputs the acquired feature data into the estimation model and estimates the user's fall susceptibility according to the fall susceptibility index output from the estimation model. The output unit 335 outputs information related to the estimated fall susceptibility.

[0180] As described above, in this embodiment, the fall likelihood of a user is estimated using feature amounts extracted from sensor data related to the user's foot movement. Therefore, according to this embodiment, the fall likelihood can be appropriately estimated in daily life without using any equipment for estimating the fall likelihood.

[0181] (Hardware) Here, a hardware configuration for executing control and processing according to each embodiment of the present disclosure will be described using an information processing device 90 in Fig. 32 as an example. Note that the information processing device 90 in Fig. 32 is an example configuration for executing control and processing according to each embodiment, and does not limit the scope of the present disclosure.

[0182] As shown in Fig. 32, an information processing device 90 includes a processor 91, a main storage device 92, an auxiliary storage device 93, an input / output interface 95, and a communication interface 96. In Fig. 32, interface is abbreviated as I / F (Interface). The processor 91, the main storage device 92, the auxiliary storage device 93, the input / output interface 95, and the communication interface 96 are connected to each other via a bus 98 so as to be able to communicate data with each other. The processor 91, the main storage device 92, the auxiliary storage device 93, and the input / output interface 95 are also connected to a network such as the Internet or an intranet via the communication interface 96.

[0183] The processor 91 loads a program stored in an auxiliary storage device 93 or the like into a main storage device 92. The processor 91 executes the program loaded into the main storage device 92. In this embodiment, a software program installed in the information processing device 90 may be used. The processor 91 executes control and processing according to each embodiment.

[0184] The main memory device 92 has an area in which programs are loaded. Programs stored in the auxiliary memory device 93 or the like are loaded into the main memory device 92 by the processor 91. The main memory device 92 is realized by a volatile memory such as a DRAM (Dynamic Random Access Memory). Furthermore, a non-volatile memory such as an MRAM (Magnetoresistive Random Access Memory) may be configured / added to the main memory device 92.

[0185] The auxiliary storage device 93 stores various data such as programs. The auxiliary storage device 93 is realized by a local disk such as a hard disk or flash memory. Note that it is also possible to configure the main storage device 92 to store various data, thereby omitting the auxiliary storage device 93.

[0186] The input / output interface 95 is an interface for connecting the information processing device 90 to peripheral devices based on standards and specifications. The communication interface 96 is an interface for connecting to external systems and devices via a network such as the Internet or an intranet based on standards and specifications. The input / output interface 95 and the communication interface 96 may be a common interface for connecting to external devices.

[0187] Input devices such as a keyboard, mouse, and touch panel may be connected to the information processing device 90 as needed. These input devices are used to input information and settings. When a touch panel is used as the input device, the display screen of the display device may also serve as the interface for the input device. Data communication between the processor 91 and the input devices may be mediated by an input / output interface 95.

[0188] The information processing device 90 may also be equipped with a display device for displaying information. When a display device is equipped, the information processing device 90 preferably includes a display control device (not shown) for controlling the display of the display device. The display device may be connected to the information processing device 90 via the input / output interface 95.

[0189] The information processing device 90 may also be equipped with a drive device. The drive device acts as an intermediary between the processor 91 and a recording medium (program recording medium) for reading data and programs from the recording medium, writing the processing results of the information processing device 90 to the recording medium, etc. The drive device may be connected to the information processing device 90 via an input / output interface 95.

[0190] The above is an example of a hardware configuration for enabling control and processing according to each embodiment of the present invention. Note that the hardware configuration in FIG. 32 is an example of a hardware configuration for executing control and processing according to each embodiment and does not limit the scope of the present invention. Furthermore, a program that causes a computer to execute control and processing according to each embodiment is also within the scope of the present invention. Furthermore, a program recording medium on which a program according to each embodiment is recorded is also within the scope of the present invention. The recording medium can be realized, for example, as an optical recording medium such as a CD (Compact Disc) or a DVD (Digital Versatile Disc). The recording medium may also be realized as a semiconductor recording medium such as a USB (Universal Serial Bus) memory or an SD (Secure Digital) card. The recording medium may also be realized as a magnetic recording medium such as a flexible disk or other recording medium. When a program executed by a processor is recorded on a recording medium, the recording medium corresponds to a program recording medium.

[0191] The components of each embodiment may be combined in any manner, and may be realized by software or by a circuit.

[0192] Although the present invention has been described above with reference to the embodiments, the present invention is not limited to the above embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

[0193] A part or all of the above-described embodiments can be described as, but not limited to, the following supplementary notes. (Appendix 1) a data acquisition unit that acquires feature data including feature amounts used to estimate the user's tendency to fall, the feature amount being extracted from sensor data related to the user's foot movements; a storage unit that stores an estimation model that outputs a fall likelihood index in response to input of the feature amount data; an estimation unit that inputs the acquired feature amount data into the estimation model and estimates the fall tendency of the user in accordance with the fall tendency index output from the estimation model; and an output unit that outputs information regarding the estimated fall tendency of the user. (Appendix 2) The data acquisition unit acquiring the feature amount data, which includes feature amounts used to estimate a fall tendency score as the fall tendency index, extracted from gait waveform data generated using time-series data of the sensor data relating to foot movement; The storage unit storing the estimation model that outputs the fall tendency score in response to input of the feature amount data; The estimation unit A fall proneness estimation device as described in Appendix 1, which inputs the acquired feature data into the estimation model and estimates the fall proneness of the user based on the fall proneness score output from the estimation model. (Appendix 3) The storage unit storing the estimation model generated by learning using training data for a plurality of subjects, the training data having feature quantities used to estimate the fall susceptibility index as explanatory variables and the fall susceptibility indexes of the plurality of subjects as objective variables; The estimation unit A fall proneness estimation device as described in Appendix 2, which inputs the feature data acquired about the user into the estimation model and estimates the fall proneness of the user based on the fall proneness index of the user output from the estimation model. (Appendix 4) The storage unit storing the estimation model trained using explanatory variables including attribute data of the plurality of subjects; The estimation unit A fall proneness estimation device as described in Appendix 3, which inputs the feature data and attribute data related to the user into the estimation model and estimates the fall proneness of the user based on the fall proneness index of the user output from the estimation model. (Appendix 5) The storage unit a first estimation model generated by learning using training data in which, with respect to the gait waveform data of the plurality of subjects, feature amounts related to activity of the quadriceps, hamstrings, tibialis anterior, tibialis posterior, and gluteus medius extracted from a first section spanning from immediately before heel contact to immediately after sole contact, feature amounts related to activity of the iliopsoas, quadriceps, adductor longus, gracilis, sartorius, and tibialis anterior extracted from a second section spanning from early swing to early swing phase, and feature amounts related to activity of the hamstrings, tibialis anterior, and gracilis extracted from a third section spanning before and after minimum toe clearance movement in mid-swing phase are used as explanatory variables, and the fall liability index of the plurality of subjects is used as a response variable; The estimation unit 5. A fall proneness estimation device as described in Appendix 3 or 4, which inputs the feature data acquired according to the user's walking into the first estimation model, and estimates the user's fall proneness according to the user's fall proneness index output from the first estimation model. (Appendix 6) The storage unit storing the first estimation model generated by learning using training data in which, with respect to the gait waveform data of the plurality of subjects, at least one feature included in a first feature group constituted by at least one feature extracted from the first section, a second feature group constituted by at least one feature extracted from the second section, and a third feature group constituted by at least one feature extracted from the third section is used as an explanatory variable, and the fall tendency index of the plurality of subjects is used as a response variable; The data acquisition unit acquiring the feature amount data including at least one feature amount included in the first feature amount group, the second feature amount group, and the third feature amount group, extracted in accordance with the user's walking; The estimation unit A fall proneness estimation device as described in Appendix 5, which inputs the acquired feature data into the first estimation model and estimates the fall proneness of the user based on the fall proneness index of the user output from the first estimation model. (Appendix 7) The storage unit storing the first estimation model generated by learning using teacher data in which, among walking phases in which feature quantities related to five items, namely, total muscle strength of the whole body, dynamic balance, lower limb muscle strength, mobility, and static balance, are extracted from the walking phase common to the five items, as an explanatory variable, and in which the fall susceptibility indexes of the plurality of subjects are used as a response variable, with respect to the walking waveform data of the plurality of subjects; The data acquisition unit acquiring the feature amount data including at least one feature amount included in the first feature amount group, the second feature amount group, and the third feature amount group, extracted in accordance with the user's walking; The estimation unit A fall proneness estimation device as described in Appendix 6, which inputs the acquired feature data into the first estimation model and estimates the fall proneness of the user based on the fall proneness index of the user output from the first estimation model. (Appendix 8) The storage unit a second estimation model generated by learning using training data in which scores of at least one of five items, namely, total muscle strength of the whole body, dynamic balance, lower limb muscle strength, mobility, and static balance, estimated using the walking waveform data of the plurality of subjects, are used as explanatory variables, and the fall susceptibility indexes of the plurality of subjects are used as objective variables; and The data acquisition unit acquiring at least one of the scores for the five items estimated according to the user's walking; The estimation unit A fall susceptibility estimation device as described in Appendix 3 or 4, which inputs the obtained score into the second estimation model and estimates the fall susceptibility of the user based on the fall susceptibility index of the user output from the second estimation model. (Appendix 9) The storage unit storing a pre-estimation model generated by learning using teacher data with respect to the walking waveform data of the plurality of subjects, the feature amount relating to at least one of the five items being an explanatory variable, and the score of the five items corresponding to the feature amount used as the explanatory variable being a response variable; The data acquisition unit acquiring a feature amount relating to at least one of the five items extracted in accordance with the user's walking; The estimation unit A fall susceptibility estimation device as described in Appendix 8, which inputs the acquired features into the pre-estimation model, inputs the score output from the pre-estimation model into the second estimation model, and estimates the fall susceptibility of the user based on the fall susceptibility index of the user output from the second estimation model. (Appendix 10) A fall likelihood estimation device according to any one of Supplementary Notes 1 to 9; a gait measurement device having a feature data generation unit that acquires time-series data of the sensor data including gait features, extracts gait waveform data for one step cycle from the time-series data of the sensor data, normalizes the extracted gait waveform data, extracts feature data used for estimating the fall liability from a gait phase cluster consisting of at least one gait phase that is consecutive in time, generates feature data including the extracted feature data, and outputs the generated feature data to the fall liability estimation device. (Appendix 11) The fall tendency estimation device includes: implemented in a terminal device having a screen viewable by the user, A fall tendency estimation system as described in Appendix 10, which displays information about the fall tendency estimated based on features extracted from the sensor data regarding the user's foot movement on the screen of the terminal device. (Appendix 12) The fall tendency estimation device includes: A fall proneness estimation system as described in Appendix 11, which displays recommendation information corresponding to the fall proneness estimated based on features extracted from the sensor data regarding the user's foot movements on the screen of the terminal device. (Appendix 13) The fall tendency estimation device includes: A fall proneness estimation system as described in Appendix 12, which displays on the screen of the terminal device a video related to training for strengthening body parts related to the fall proneness as the recommendation information according to the fall proneness estimated based on features extracted from the sensor data related to the movement of the user's feet. (Appendix 14) The computer acquiring feature amount data including feature amounts used to estimate the user's tendency to fall, the feature amount being extracted from sensor data relating to the user's foot movements; inputting the acquired feature amount data into an estimation model that outputs a fall susceptibility index in accordance with the input of the feature amount data; estimating the fall tendency of the user according to the fall tendency index output from the estimation model; A method for estimating a tendency to fall that outputs information relating to the estimated tendency to fall of the user. (Appendix 15) acquiring feature data including feature values used to estimate the user's likelihood of falling, the feature values being extracted from sensor data relating to the user's foot movements; inputting the acquired feature amount data into an estimation model that outputs a fall susceptibility index in response to input of the feature amount data; a process of estimating the fall tendency of the user according to the fall tendency index output from the estimation model; and outputting information relating to the estimated fall tendency of the user. [Explanation of symbols]

[0194] 1 Fall susceptibility estimation system 2. Learning System 10, 20 Gait measurement device 11 Sensors 12 Feature data generation unit 13 Easy-to-fall estimation device 25 Learning Device 111 Acceleration Sensor 112 Angular rate sensor 121 Acquisition Department 122 Normalization section 123 Extraction part 125 Generation part 127 Feature data output unit 131, 331 Data acquisition section 132, 332 storage section 133, 333 Estimation part 135, 335 output section 251 Receiving unit 253 Learning Department 255 Storage section

Claims

1. a data acquisition means for acquiring feature data including feature values used to estimate the user's tendency to fall, the feature values being extracted from sensor data relating to the user's foot movements; a storage means for storing an estimation model that outputs a fall tendency index in response to input of the feature amount data; an estimation means for inputting the acquired feature amount data into the estimation model and estimating the fall tendency of the user in accordance with the fall tendency index output from the estimation model; an output means for outputting information relating to the estimated fall tendency of the user; The output means The fall proneness estimation device outputs recommendation information according to the item with the lowest score among five items related to the fall proneness: total muscle strength of the whole body, dynamic balance, lower limb muscle strength, mobility, and static balance.

2. The data acquisition means acquiring the feature amount data, which includes feature amounts used to estimate a fall tendency score as the fall tendency index, extracted from gait waveform data generated using time-series data of the sensor data relating to foot movement; The storage means storing the estimation model, which is generated by learning using training data in which feature quantities used to estimate the fall susceptibility index for a plurality of subjects are used as explanatory variables and the fall susceptibility indexes of the plurality of subjects are used as objective variables, and which outputs the fall susceptibility score in response to input of the feature quantity data; The estimation means A fall proneness estimation device as described in claim 1, wherein the feature data acquired about the user is input into the estimation model, and the fall proneness of the user is estimated based on the fall proneness score of the user output from the estimation model.

3. The storage means a first estimation model generated by learning using training data in which, with respect to the gait waveform data of the plurality of subjects, feature amounts related to activity of the quadriceps, hamstrings, tibialis anterior, tibialis posterior, and gluteus medius extracted from a first section spanning from immediately before heel contact to immediately after sole contact, feature amounts related to activity of the iliopsoas, quadriceps, adductor longus, gracilis, sartorius, and tibialis anterior extracted from a second section spanning from early swing to early swing phase, and feature amounts related to activity of the hamstrings, tibialis anterior, and gracilis extracted from a third section spanning before and after minimum toe clearance movement in mid-swing phase are used as explanatory variables, and the fall liability index of the plurality of subjects is used as a response variable; The estimation means A fall susceptibility estimation device as described in claim 2, wherein the feature data acquired according to the user's walking is input into the first estimation model, and the fall susceptibility of the user is estimated according to the fall susceptibility index of the user output from the first estimation model.

4. The storage means storing the first estimation model generated by learning using training data in which, with respect to the gait waveform data of the plurality of subjects, at least one feature included in a first feature group constituted by at least one feature extracted from the first section, a second feature group constituted by at least one feature extracted from the second section, and a third feature group constituted by at least one feature extracted from the third section is used as an explanatory variable, and the fall tendency index of the plurality of subjects is used as a response variable; The data acquisition means acquiring the feature amount data including at least one feature amount included in the first feature amount group, the second feature amount group, and the third feature amount group, which are extracted in response to the user's walking; The estimation means A fall susceptibility estimation device as described in claim 3, wherein the acquired feature data is input into the first estimation model, and the fall susceptibility of the user is estimated based on the fall susceptibility index of the user output from the first estimation model.

5. The storage means storing the first estimation model generated by learning using teacher data in which, among walking phases in which feature quantities related to each of the five items, namely, total muscle strength of the whole body, dynamic balance, lower limb muscle strength, mobility, and static balance, are extracted from the walking phase common to the five items, as an explanatory variable, and in which the fall susceptibility indexes of the plurality of subjects are used as a response variable, with respect to the walking waveform data of the plurality of subjects; The data acquisition means acquiring the feature amount data including at least one feature amount included in the first feature amount group, the second feature amount group, and the third feature amount group, which are extracted in response to the user's walking; The estimation means A fall susceptibility estimation device as described in claim 4, wherein the acquired feature data is input into the first estimation model, and the fall susceptibility of the user is estimated based on the fall susceptibility index of the user output from the first estimation model.

6. The storage means a second estimation model generated by learning using training data in which scores of at least one of five items, namely, total muscle strength of the whole body, dynamic balance, lower limb muscle strength, mobility, and static balance, estimated using the walking waveform data of the plurality of subjects, are used as explanatory variables, and the fall susceptibility indexes of the plurality of subjects are used as objective variables; and The data acquisition means acquiring at least one of the scores for the five items estimated according to the user's walking; The estimation means A fall susceptibility estimation device as described in claim 2, wherein the obtained score is input into the second estimation model, and the fall susceptibility of the user is estimated based on the fall susceptibility index of the user output from the second estimation model.

7. The fall likelihood estimation device according to any one of claims 1 to 6, a gait measurement device having: a sensor that is attached to footwear of a user whose fall liability is to be estimated, measures spatial acceleration and spatial angular velocity, generates sensor data related to foot movement using the measured spatial acceleration and spatial angular velocity, and outputs the generated sensor data; and a feature data generation means that acquires time-series data of the sensor data including gait features, extracts gait waveform data for one step cycle from the time-series data of the sensor data, normalizes the extracted gait waveform data, extracts feature values used for estimating the fall liability from a gait phase cluster formed by at least one temporally consecutive gait phase, generates feature data including the extracted feature values, and outputs the generated feature data to the fall liability estimation device.

8. The fall tendency estimation device includes: implemented in a terminal device having a screen viewable by the user, A fall tendency estimation system as described in claim 7, wherein information regarding the fall tendency estimated based on features extracted from the sensor data regarding the user's foot movement is displayed on the screen of the terminal device.

9. The computer acquiring feature amount data including feature amounts used to estimate the user's tendency to fall, the feature amount being extracted from sensor data relating to the user's foot movements; inputting the acquired feature amount data into an estimation model that outputs a fall susceptibility index in accordance with the input of the feature amount data; estimating the fall tendency of the user according to the fall tendency index output from the estimation model; outputting information relating to the estimated fall tendency of the user; In the output, A method for estimating a tendency to fall that outputs recommendation information according to the item with the lowest score among five items related to the tendency to fall: total muscle strength of the whole body, dynamic balance, lower limb muscle strength, mobility, and static balance.

10. On the computer, acquiring feature data including feature values used to estimate the user's likelihood of falling, the feature values being extracted from sensor data relating to the user's foot movements; inputting the acquired feature amount data into an estimation model that outputs a fall susceptibility index in response to input of the feature amount data; a process of estimating the fall tendency of the user according to the fall tendency index output from the estimation model; outputting information about the estimated fall tendency of the user; In the output process, and outputting recommendation information according to the item with the lowest score among the five items related to the tendency to fall: total muscle strength of the whole body, dynamic balance, lower limb muscle strength, mobility, and static balance.

Images