Walking speed estimation method and device

Abstract

According to one embodiment, a walking speed estimation method is provided that includes the steps of determining N subjects to be measured for walking speed measurement and acquiring physical data of the subjects, conducting a clinical evaluation on the subjects, conducting a gait evaluation on the subjects to acquire measured gait data, and developing a walking speed estimation model having a plurality of walking parameters based on the measured gait data, wherein the walking speed estimation model includes, as walking parameters, a roll angle, a yaw angle, and body weight measured using an accelerometer worn at the center of gravity of the body of the subjects.

Description

[Technical field]

[0001] The following embodiments relate to a walking speed estimation method and device. [Background technology]

[0002] Human gait changes due to muscle and sensory degeneration. Older adults may have various aging-related problems that change their gait to some degree. In previous studies, human body kinematics and muscle weakening were the main concerns that change stride length. People with more degenerative changes may have shorter leg length because their walking efficiency decreases the longer they walk. Research has also shown that both stride length and body weight show a curvilinear relationship with walking speed.

[0003] Gait speed has been associated with aging, disability, falls, cognitive decline and mortality. However, in most previous studies, walking speed was measured manually using a stopwatch, which is subject to human error and limited accuracy. This has led to the development of laboratory-based motion capture systems and instrumented walkways to provide more accurate and reliable estimates of walking speed than manual measurements using a stopwatch. However, laboratory-based motion capture systems and instrumented walkways have the disadvantages of only being able to measure a limited number of steps due to spatial constraints and of high initial cost.

[0004] To overcome these spatial and cost limitations, walking speed can be estimated using wearable acceleration sensors, which are cheaper and less spatially constrained than motion capture systems and instrumented walkways, and less sensitive to human error than stopwatches. However, wearable acceleration sensors indirectly estimate walking speed by direct integration of acceleration, kinematic modeling and walking motion correction, data regression modeling, or hybrid methods without directly measuring walking speed, which results in a problem that the accuracy of walking speed estimated by inertial sensors is not high. Summary of the Invention [Problem to be solved by the invention]

[0005] To provide a walking speed estimation method and device capable of precisely estimating walking speed and diagnosing aging and associated risks.

[0006] The problems to be solved by the present invention are not limited to those described above, and other problems and advantages of the present invention not mentioned can be understood from the following description and will be more clearly understood by the embodiments of the present invention. Furthermore, it will be understood that the problems and advantages to be solved by the present invention can be realized by the means and combinations thereof shown in the claims. [Means for solving the problem]

[0007] According to one embodiment of the present invention, a walking speed estimation method is provided, which includes the steps of determining N subjects to be measured for walking speed measurement and acquiring physical data of the subjects, acquiring result data of a clinical evaluation performed on the subjects, performing a gait evaluation on the subjects to acquire measured gait data, and developing a walking speed estimation model having a plurality of walking parameters based on the measured gait data, wherein the walking speed estimation model includes, as walking parameters, a roll angle, a yaw angle, and body weight measured using an accelerometer worn at the center of gravity of the subject's body.

[0008] In the present invention, determining the N subjects can be made by excluding people in the population who have a disease that may affect walking or whose balance ability assessment is below a predetermined score.

[0009] In the present invention, the clinical assessment may be to diagnose a psychiatric disorder according to pre-established diagnostic criteria, determine the overall severity of cognitive impairment, and evaluate walking and balance ability using a Performance Oriented Mobility Assessment (POMA).

[0010] In the present invention, the step of acquiring the measured gait data can acquire the gait characteristic values ​​of the subject using an accelerometer attached to the body's center of gravity and an existing gait measuring device.

[0011] In the present invention, the accelerometer may be an inertial measurement unit (IMU) that is attached to the waist of the subject with a predetermined adhesive material and is capable of measuring acceleration in three axial directions.

[0012] In the present invention, the step of developing the walking speed estimation model may include a first model that performs a linear regression analysis including a roll angle, a yaw angle, and a body weight as walking parameters, and a second model that separates the subject into a plurality of subgroups based on NSL and performs the linear regression analysis for each of the subgroups.

[0013] In the present invention, the NSL is calculated by the following formula: JPEG2025501161000002.jpg1281 In the above formula, l l is the leg length, l s is the stride length, α1 is the person's leg / height ratio, α1 is the leg bending coefficient, and a is the angular displacement.

[0014] The present invention may further include a step of statistically analyzing the developed walking speed estimation model.

[0015] The present invention may further include a step of calculating the walking speed of the person whose walking speed is to be estimated using the walking speed estimation model that has been developed.

[0016] In the present invention, the step of calculating the walking speed includes inputting and acquiring age, weight, and leg length data of the person whose walking speed is to be estimated, acquiring cadence, VHD (Vertical height displacement), roll angle, and yaw angle from an inertial measurement unit (IMU), and substituting the acquired values ​​into the walking speed estimation model to calculate the walking speed.

[0017] According to another embodiment of the present invention, there is provided a walking speed estimation device including a processor and a memory, the processor determining N subjects to be measured for walking speed measurement, acquiring physical data of the subjects, conducting a clinical evaluation on the subjects, conducting a gait evaluation on the subjects to acquire measured gait data, and developing a walking speed estimation model having a plurality of gait parameters based on the measured gait data, the walking speed estimation model including, as gait parameters, a roll angle, a yaw angle and a body weight measured using an accelerometer worn at the center of gravity of the body of the subjects.

[0018] Furthermore, there is provided a computer-readable recording medium having a program recorded thereon for executing the method of the present invention on a computer. Effect of the Invention

[0019] According to the above-mentioned means for solving the problems of the present disclosure, walking speed can be precisely estimated to diagnose aging and related risks.

[0020] Furthermore, according to the means for solving the problems of the present disclosure, weight, roll angle and yaw angle can be used as additional walking parameters to reduce training error, and normalized subgroups can be used to develop a walking speed estimation model.

[0021] Furthermore, according to the means for solving the problems of the present disclosure, the model of the present invention can minimize the demographic and anthropometric influences on walking speed estimation by using NSL instead of the original stride length value. [Brief description of the drawings]

[0022] [Figure 1] FIG. 1 is a system diagram including a user terminal and an estimation server according to one embodiment. [Diagram 2] FIG. 2 is a flow chart illustrating a walking speed estimation method according to one embodiment. [Diagram 3] FIG. 3 is a table showing characteristics of subjects according to one embodiment. [Figure 4] FIG. 4 is a table illustrating a comparison between the 0th model and the 1st model according to one embodiment. [Diagram 5] FIG. 5 is an embodiment showing a table of subgroups sorted by normalized stride length according to one embodiment. [Figure 6] FIG. 6 is a table illustrating a method of estimating walking speed for subgroups classified by normalized stride length according to one embodiment. [Figure 7] FIG. 7 is a comparative statistical analysis example of a walking speed estimation model according to one embodiment compared to other models. [Figure 8] FIG. 8 is a walking model of a walking speed estimation model according to one embodiment. [Figure 9] FIG. 9 is a block diagram of a walking speed estimation device according to an embodiment. BEST MODE FOR CARRYING OUT THEINVENTION

[0023] According to one embodiment of the present invention, a walking speed estimation method is provided, which includes the steps of determining N subjects to be measured for walking speed measurement and acquiring physical data of the subjects, acquiring result data of a clinical evaluation performed on the subjects, performing a gait evaluation on the subjects to acquire measured gait data, and developing a walking speed estimation model having a plurality of walking parameters based on the measured gait data, wherein the walking speed estimation model includes, as walking parameters, a roll angle, a yaw angle, and body weight measured using an accelerometer worn at the center of gravity of the subject's body. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0024] The advantages and features of the present invention, as well as the methods of achieving them, will become apparent from the detailed description of the embodiments in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments presented below, and can be embodied in different forms, and should be understood to include all modifications, equivalents, or alternatives falling within the spirit and technical scope of the present invention.

[0025] The terms used in this application are merely used to describe certain embodiments and are not intended to limit the present invention. A singular expression includes a plural expression unless the context clearly indicates otherwise. In this application, the terms "include" or "have" are intended to specify the presence of features, numbers, operations, components, parts, or combinations thereof described herein, and are understood not to preclude the presence or additional possibility of one or more other features, numbers, operations, components, parts, or combinations thereof.

[0026] Some embodiments of the present disclosure may be represented by functional block configurations and various processing operations. Some or all of such functional blocks may be implemented in various numbers of hardware and / or software configurations that perform specific functions. For example, the functional blocks of the present disclosure may be implemented by one or more microprocessors, or may be implemented by circuit configurations for a given function. Also, for example, the functional blocks of the present disclosure may be implemented in various programming or scripting languages. The functional blocks may be implemented by algorithms executed by one or more processors. Also, the present disclosure may employ conventional techniques for electronic configuration, signal processing, and / or data processing, etc. Terms such as "mechanism," "element," "means," and "configuration" may be used broadly and are not limited to mechanical and physical configurations.

[0027] Furthermore, the connecting lines or members between components shown in the figures are merely illustrative of functional and / or physical or circuit connections, and in an actual device the connections between components may be represented by various alternative or additional functional, physical or circuit connections.

[0028] Hereinafter, an action performed by a user may refer to an action performed by a user through a user terminal. As an example, a command corresponding to an action performed by a user may be input to the user terminal through an input device (e.g., a keyboard, a mouse, etc.) embedded in or additionally connected to the user terminal. As another example, a command corresponding to an operation performed by a user may be input to the user terminal through a touch screen of the user terminal. In this case, the operation performed by the user may include a predetermined gesture. For example, the gesture may include a tap, a touch & hold, a double tap, a drag, a pan, a flick, a drag & drop, etc.

[0029] The present disclosure will now be described in detail with reference to the accompanying drawings.

[0030] FIG. 1 is a system diagram including a user terminal and an estimation server according to one embodiment.

[0031] A system according to one embodiment may include multiple user terminals 1000, an estimation server 2000, and an IMU 10.

[0032] The user terminal 1000 may be a smartphone, a tablet PC, a PC, a smart TV, a mobile phone, a laptop, and other mobile or non-mobile computing devices. The user terminal 1000 may also be a wearable device such as glasses, a headband, etc., with communication and data processing capabilities. The user terminal 1000 may include any kind of device that can communicate with other devices via a network.

[0033] The estimation server 2000 may be implemented as a computing device or multiple computing devices that communicate over a network to provide instructions, code, files, content, services, and the like.

[0034] The user terminal 1000 and the estimation server 2000 can communicate with each other using a network. The estimation server 2000 can communicate with the user terminal 1000 via the network and provide the walking speed estimation model and the estimation result to the user terminal 1000. In one embodiment, the estimation server 2000 can estimate the walking speed of the user using values ​​obtained from an IMU (inertial measurement unit) 10.

[0035] Meanwhile, in another embodiment, the user terminal 1000 may be a walking speed estimation device that has downloaded an application capable of executing the walking speed estimation method of the present invention from the estimation server 2000 or another external device not shown. That is, the walking speed estimation method of the present disclosure may be executed by the user terminal 1000 or may be executed by the estimation server 2000. Hereinafter, for convenience of explanation, the user terminal 1000 or the estimation server 2000 will be referred to as a walking speed estimation device.

[0036] The IMU 10 may be an acceleration measuring device, more specifically, a three-axis acceleration measuring device. According to an embodiment of the present invention, the IMU 10 may be attached to the center of gravity of a person's body as shown in Fig. 1. More specifically, the IMU 10 may be a three-axis accelerometer attached to the waist using a predetermined adhesive material.

[0037] Meanwhile, the network includes a local area network (LAN), a wide area network (WAN), a value-added network (VAN), a mobile radio communication network, a satellite communication network, and a combination thereof, and is a comprehensive data communication network that enables each network component shown in FIG. 1 to communicate smoothly with each other, and may include wired Internet, wireless Internet, and mobile radio communication networks. In addition, the wireless communication may include, but is not limited to, wireless LAN (wi-fi), Bluetooth (registered trademark), Bluetooth (registered trademark) low energy, ZigBee (registered trademark), WFD (wi-fi Direct), UWB (ultrawideband), infrared communication (IrDA, infrared Data Association), NFC (Near Field Communication), etc.

[0038] FIG. 2 is a flow chart illustrating a walking speed estimation method according to one embodiment.

[0039] Referring to FIG. 2, first, N subjects who are to be measured for walking speed measurement are determined, and physical data of the subjects are acquired (S10).

[0040] Next, a clinical evaluation is performed on the determined subject (S20).

[0041] Next, a gait evaluation is performed based on the clinical evaluation (S30).

[0042] Next, a walking speed estimation model having a plurality of walking parameters is developed based on the walking evaluation (S40). At this time, the walking speed estimation model may include a roll angle, a yaw angle, and a body weight measured using an accelerometer attached to the center of gravity of the subject's body as walking parameters.

[0043] In the following, the walking speed estimation method of the present invention will be considered in more detail.

[0044] First, the subject to be measured in the present invention can be specified as follows.

[0045] FIG. 3 is a table showing characteristics of subjects according to one embodiment.

[0046] As summarized in the table of FIG. 3, the present invention can determine a portion of the total N participants of two existing cohort studies on aging and senility, namely the first study and the second study, as subjects. Furthermore, the present invention obtains physical characteristic data of the subjects. For example, the physical characteristic data of the subjects can be sex, age, MiniMentalStateExamination (MMSE), balance ability assessment (PerformanceOrientedMobilityAssessment, POMA), height, weight, BMI including overweight and underweight ratios, foot length, vertical height displacement (VHD), cadence, walking speed, step length, roll angle, yaw angle, etc., which can be described in the first column of the table of FIG. 3. More specifically, the VHD can be the average difference between the maximum and minimum vertical heights of the center of gravity of the body in one walking cycle. The walking speed and stride length listed in the table of FIG. 3 are measured using an existing gait analysis device (for example, GAITRite TM ) and the roll angle may be measured using a waist-worn inertial measurement unit.

[0047] More specifically, according to one embodiment of the present invention, the first study and the second study may be population-based exploratory cohort studies on elderly Koreans. According to one embodiment, the first study is "Overview of the Korean longitudinal study on cognitive aging and dementia" (JW Han, TH Kim, KP Kwak, K. Kim, BJ Kim, SG Kim, JL Kim, TH Kim, SW Moon, and J.Y. Park, Psychiatry investigation, vol. 15, no. 8, pp. 767, 2018.), and the second study is "The Korean version of the FRAIL scale: clinical The subject of the present invention may be, but is not necessarily limited to, "feasibility and validity of assessing the frailty status of Korean drivers" (H.-W. Jung, H.-J. Yoo, S.-Y. Park, S.-W. Kim, J.-Y. Choi, S.-J. Yoon, C.-H. Kim, and K.-i. Kim, The Korean journal of internal medicine, vol. 31, no. 3, pp. 594, 2016.), and the subject of the present invention may be freely modified.

[0048] More specifically, in the first study, a group of elderly people selected from a specific region may be tracked at a preset interval. More specifically, in the first study, 6,818 Koreans aged 60 or older randomly sampled from 13 regions across Korea may be tracked every two years from 2009. In the second study, follow-up observations may be conducted on subjects in a specific region at a preset interval. More specifically, in the second study, follow-up observations may be conducted on 3,000 Korean volunteers aged 70 to 84 every two years from 2016 to 2020.

[0049] Meanwhile, the walking speed measurement model of the present invention can be established by using N healthy elderly people living in a community as subjects, excluding participants with neurological diseases such as major psychiatric diseases and neurocognitive disorders, among the M participants in the first and second studies. That is, according to one embodiment of the present invention, the N subjects can be determined from a population of participants in existing studies, excluding those with psychiatric or neurological diseases.

[0050] Referring to the embodiment of Fig. 3, a total of 759 people can be included as subjects, including 338 men aged 74.8 ± 5.0 years and 421 women aged 73.3 ± 4.5 years. According to one embodiment, participants with stroke and musculoskeletal disease that may affect walking or balance in baseline or subsequent evaluations, and participants with a preset score (e.g., less than 25 points) in the balance ability assessment (Tinetti Performance Oriented Mobility Assessment, POMA) are excluded from the subjects.

[0051] Next, according to an embodiment of the present invention, a clinical evaluation is performed on the determined subject (S20). More specifically, a geriatric psychiatrist or neurologist with expertise in dementia research may meet with the subject and perform a standardized diagnostic interview, physical and neurological examinations, complete blood counts, chemistry profiles, serological tests for syphilis, cardiac ultrasound, chest X-ray, and other laboratory tests.

[0052] In addition, the present invention can diagnose dementia and other mental disorders according to preset diagnostic criteria, and use clinical dementia scale to determine the overall severity of cognitive impairment.Also, gait and balance can be evaluated using POMA as described above.In one embodiment, higher POMA score indicates better gait and balance, and the maximum score of POMA can be 28 points.

[0053] Next, the present invention performs a gait assessment to obtain measured gait data (S30). More specifically, the present invention can obtain results of measuring the gait of each subject using an IMU attached to the center-of-body-mass (CoM) and an existing gait measurement device. In one example, the IMU can be FITMET (registered trademark) FitLife Inc., but is not necessarily limited thereto, and can be replaced with ActiGraph (registered trademark). Furthermore, in one embodiment, the existing gait measurement device can be GAITite TM (CISystems, HAVERT), but is not necessarily limited thereto, and the device that generates the walking data required for developing the walking speed estimation model of the present invention described later can be an existing gait measurement device without any restrictions. In the following specification, for convenience of explanation, the existing gait measurement device is assumed to be GAITRite.

[0054] According to one embodiment, the IMU is a smooth-edged hexagon (35x35x13mm [14g] or 30x40x10mm [17g]) and an example of an IMU can use a similar digital 3-axis accelerometer (e.g., BMA255, BOSH, Germany) and gyroscope (e.g., BMX055, BOSH, Germany). The example 3-axis accelerometer can measure 3-axis acceleration up to ±8g (resolution 0.004g) and 3-axis acceleration up to ±1,000μ / s (resolution 0.03μ / s) at 250Hz.

[0055] In a measurement method according to an embodiment of the present invention, an IMU can be fixed at the third to fourth lumbar vertebrae of a subject using a predetermined adhesive (e.g., Hypafix). In addition, during the experiment, a subject can walk three times back and forth along a 14m flat straight walkway at a speed of their own choosing, and a movement line can be designed so that the subject turns after crossing the 14m line. In addition, according to one embodiment, a GAITRite electronic mat can be placed in the center of the walkway to measure steady-state walking.

[0056] In one embodiment, the GAITRite can be a portable gait analysis walking device that measures temporal and spatial gait parameters through a 100 Hz electronic walkway connected to a computer's USB port. The walkway size of the GAITRite's electronic mat can be 520 (L) x 90 (W) x 0.6 cm (H) and the active sensing area can be 427 (L) x 61 cm (W). Furthermore, the electronic mat can include 16,128 sensors arranged with a spatial accuracy of 1.27 cm.

[0057] Next, the present invention develops a walking speed estimation model (S40). In one embodiment, the present invention can remove a walking section of a preset length before and after the data generated in the above-mentioned walking data acquisition step in order to measure the steady-state walking of the subject. More specifically, the walking speed estimation device of the present invention can analyze the remaining intermediate 10 m data after removing a walking section of 2 m length before the reference point and before each rotation in order to measure the steady-state walking of the subject. Furthermore, the present invention can identify each step after pre-processing the IMU signal and selecting features. In this case, the present invention can use a comparative standard walking speed obtained from another existing gait measurement device (e.g., GAITRite).

[0058] The present invention can establish a 0th model based on the obtained comparative standard walking speed, and estimate the walking speed using the IMU signal according to the 0th model. In the 0th model, the walking speed can be estimated as a regression model including age, sex, foot length, cadence, and vertical height displacement (hereinafter referred to as VHD) and cadence (gait rate).

[0059] More specifically, according to one embodiment, the 0th model may be a regression model for estimating walking speed of healthy older adults using accelerations and angles of IMU sensors. In the 0th model, two demographic features (age and sex), an anthropometric feature (foot length), and two features measured by the IMU (VHD and cadence) may be used as gait parameters to estimate walking speed.

[0060] Meanwhile, walking speed is the product of cadence and step length. Regarding walking speed estimation, the cadence measured by the IMU is reliable, so if the walking speed estimation model includes features related to step length, the accuracy of the model can be improved.

[0061] The following formula (1) is a mathematical linear regression formula of the 0th model developed based on the physical data of the subject acquired in the embodiment of FIG. 2 described above. From formula (1), it can be seen that the linear regression formula of the 0th model includes age, sex, cadence, VHD, and foot length as walking parameters for estimating the walking speed (Gaitspeed), and can have a coefficient corresponding to each walking parameter. In addition, the foot length is the average length of both feet, and the gender can be set to 1 for men and 2 for women.

[0062] [Number 1] Walking speed [0th model] =-113.2-0.388*age+3.06*gender+1.17*cadence+12.1*VHD+3.25*foot length

[0063] However, the 0th model showed low accuracy for people with a walking speed of 100 cm per second or less, and there was a problem that the estimation error gradually increased as the speed decreased. Therefore, in the present invention, a feature related to step length is added to complement the 0th model to estimate walking speed, and the walking speed estimation model of the present invention can be optimized in subgroups stratified by normalized step length (NSL). Furthermore, elderly people may have different walking patterns depending on their health conditions, and linear regression models are vulnerable to non-uniform samples. To compensate for this, a regression model can be developed in subsequent subgroups of walking speed to further improve the accuracy of the model. Furthermore, the accuracy of the walking speed estimation model of the present invention was compared with that of other existing gait analysis systems.

[0064] In the following, a first model and a second model of a walking speed estimation method capable of complementing the above-mentioned 0th model will be described as an embodiment of the present invention.

[0065] According to one embodiment of the present invention, the roll and yaw angles of the body center of gravity and the body weight can be added to the zeroth model as walking parameters to develop a first model of the walking speed estimation method of the present invention.

[0066] According to one embodiment, stride length is a product of trunk, knee and foot movements, and the roll and yaw angles of the body center of mass can reflect the angular motion of the trunk and limbs. More specifically, the roll and yaw angles of the body center of mass can be calculated from IMU sensor data, and the roll and yaw angle values ​​can be re-oriented to Cartesian coordinates. Stride length is also related to body weight, which is related to thrust power.

[0067] More specifically, the three features further selected in the linear regression model of the first model of the present invention are body weight, roll angle and yaw angle of CoM. Relatedly, body weight is the main gait parameter in that both weight loss and weight gain reduce walking speed by reducing muscle strength or inducing physical weakness. Also, in the sample of elderly adults, there are more overweight people with BMI of 25 or more than underweight people with BMI of less than 18.5, which can be the reason for adding body weight as a gait parameter in the walking speed estimation model of the present invention, and is also the reason why the coefficient of the regression model is negative.

[0068] Furthermore, body rotation is related to gait patterns. The roll angle and yaw angle represent the three-dimensional gait pattern. Viewed from above the body, yaw rotation is the addition of CoM motion to the forward ground motion of the stance leg when the subject rotates the hips in a gait where each leg moves forward. In the gait speed estimation model of the present invention, the yaw angle coefficient is positive, which may indicate that the yaw angle may increase as the gait speed increases. Thus, the present invention can define a relationship between the yaw angle of the CoM and gait speed that has not been studied before.

[0069] Furthermore, the roll angle indicates the behavior related to medio-lateral (ML) stability and is related to walking speed. In the walking speed estimation model of the present invention, the roll angle coefficient is positive, which indicates that the roll angle is likely to increase as the walking speed increases.

[0070] Meanwhile, gender (sex) may not be selected as a walking parameter in the first and second models of the walking speed estimation model of the present invention. This is because, among the three features added to the model of the present invention, the roll angle of the CoM can reflect the influence of gender on walking speed to some extent. That is, as seen in the table of FIG. 3 above, females have a larger roll angle than males. In contrast, the yaw angle does not statistically change between genders.

[0071] The following formula (2) is a mathematical linear regression formula of the first model developed based on the subject's physical data acquired in the above-mentioned embodiment of Fig. 2. From formula (2), it can be seen that the linear regression formula of the first model includes age, cadence, VHD, foot length, weight, roll angle, and yaw angle as walking parameters for estimating walking speed, and can have a coefficient corresponding to each walking parameter.

[0072] [Number 2] Walking speed [first model] =-106.0-0.328 x age + 1.10 x cadence + 10.1 x VHD + 3.29 x foot length - 0.115 x weight + 1.01 x roll angle + 0.647 x yaw angle

[0073] According to an embodiment of the present invention, the walking speed of a walking speed estimation subject can be accurately calculated using the developed first model. That is, the walking speed estimation device of the present invention can input and acquire the age, weight, and leg length data of a subject whose walking speed is to be calculated, or can acquire them from another terminal or server. Also, the walking speed estimation device acquires from the IMU the cadence, VHD, roll angle, and yaw angle, which are values ​​measured by the IMU while the walking speed estimation subject wears the IMU and walks. Also, the walking speed estimation device may calculate the walking speed by substituting the acquired walking parameter values ​​into the linear regression equation of the first model. According to an embodiment of the present invention, when estimating the walking speed using the developed first model, even if a gait measurement device (e.g., GAITRite) is not present, the walking speed can be easily and precisely estimated using only the IMU.

[0074] FIG. 4 is a table illustrating a comparison between the 0th model and the 1st model according to one embodiment.

[0075] 4, the first model of the present invention includes gender, roll angle, yaw angle, and weight as walking parameters, and the remaining walking parameters can use the walking parameters of the 0th model. In the table of FIG. 4, VIF is the variation inflation factor.

[0076] Furthermore, compared to the values ​​measured by Gaitrite, the 0th model has a high level of walking speed estimation (R = 0.935, R 2 =0.875, adjustR 2 = 0.874, F = 1051, p < 0.001), and the first model's estimation of walking speed is more accurate (R = 0.953, R 2 =0.908, adjustR 2 =0.908, F=10.001, p<0.001).

[0077] In addition, the present invention can set a second model, which is a subgroup-specific regression model having gait parameters included in the first model, within subgroups classified by normalized step length (NSL) using the K-means method. The regression model of the second model of the present invention can include only features with a variance inflation factor (VIF) of less than 2.5.

[0078] According to one embodiment of the present invention, the NSL is obtained by dividing the stride length by the height, and the stride length can be obtained by dividing the walking speed determined by the first model by the cadence. Hereinafter, the above-mentioned first model and second model can be collectively referred to as the walking speed estimation model of the present invention.

[0079] FIG. 5 is an embodiment showing a table of subgroups categorized by normalized stride length according to one embodiment.

[0080] Referring to FIG. 5, the characteristics of the three NSL subgroups are summarized in the table of FIG. 5. More specifically, the NSL subgroups may include short NSL, medium NSL, and long NSL. According to an embodiment of the present invention, the reference values ​​for distinguishing the short NSL, medium NSL, and long NSL subgroups may be set differently depending on the characteristics of the subject, and the number of subgroups may be set to be more or less than three. FIG. 5 shows body characteristic data of subjects classified into NSL subgroups according to preset reference values ​​according to an embodiment of the present invention.

[0081] More specifically, referring to FIG. 5, the present invention can obtain the sex, age, height, weight, foot length, gait speed, cadence, VHD, roll angle, yaw angle, and POMA value of the subject for each subgroup. The VHD and POMA are the same as those described in the description of FIG. 3, the roll angle and yaw angle are obtained from the IMU, and the foot length and gait speed can also be obtained from the GAITRite as described in FIG. 3. The remaining values ​​are also omitted from the description of FIG. 3. Also, according to one embodiment, the statistics are analyzed using variance analysis based on post-hoc comparison by Bonferron in FIG. 3, but the statistical analysis method may be changed in other embodiments.

[0082] 1 is a table illustrating a method for estimating walking speed by NSL subgroup according to one embodiment.

[0083] Referring to Figure 6 , when the second model, which is a walking speed estimation model for each NSL subgroup, was developed using the walking parameters included in the first model, it can be seen that all the walking parameters were selected as significant walking parameters for walking speed in all NSL subgroups and had low VIF.

[0084] Referring to FIG. 6, the NSL subgroup model (model 2) according to one embodiment has a walking speed that is more accurately estimated (R=0.955, R 2 = 0.912, p < 0.001), and the walking speed estimates in all NSL subgroups are good. More specifically, the statistics of walking speed estimated by the second model are by NSL subgroup, with R 2 = 0.828, F = 195.545, p < 0.001, and for the medium NSL subgroup (MediumNSL in Figure 6), the R 2 = 0.801, F = 217.135, p < 0.001, and for the long NSL (longNSL in Figure 6) subgroup, R 2 = 0.836, F = 140.971, p < 0.001.

[0085] The following formulas (3-1) to (3-3) are mathematical linear regression formulas of the second model shown in FIG. 6, which were developed based on the physical data of the subject obtained in the embodiment of FIG. 2 described above. From formulas (3-1) to (3-3), it can be seen that the linear regression formula for each NSL subgroup of the second model includes age, cadence, VHD, foot length, weight, roll angle, and yaw angle as walking parameters for estimating walking speed, and can have coefficients corresponding to each walking parameter. In addition, the short NSL subgroup can be set to a group with an NSL of 0.347 or less, the intermediate NSL subgroup to a group with an NSL of more than 0.347 and less than 0.394, and the long NSL subgroup to a group with an NSL of more than 0.394.

[0086] [Number 3-1] Walking speed [Second model, NSL≦0.347] =-95.6-0.462*age+1.06*cadence+13.3*VHD+3.24*foot length-0.154*weight+1.27*roll angle+0.550*yaw angle

[0087] [Number 3-2] Walking speed [second model, 0.347 <NSL≦0.394] =-117.8-0.327 x age + 1.14 x cadence + 11.8 x VHD + 3.26 x foot length - 0.122 x weight + 1.14 x roll angle + 0.828 x yaw angle

[0088] [Number 3-3] Walking speed [second model, 0.394 <NSL] =-121.4-0.272 x age + 1.15 x cadence + 9.90 x VHD + 3.68 x foot length -0.179 x weight + 1.04 x roll angle + 0.620 x yaw angle

[0089] According to an embodiment of the present invention, the walking speed of a walking speed estimation subject can be accurately calculated for each NSL subgroup using the developed second model. That is, the walking speed estimation device of the present invention can input and acquire age, weight, and leg length data of a subject whose walking speed is to be calculated, or can acquire the data from another terminal or server. Also, the walking speed estimation device acquires cadence, VHD, roll angle, and yaw angle, which are values ​​measured by the IMU while the walking speed estimation subject wears the IMU and walks, from the IMU. The walking speed estimation device can also use the walking parameter values ​​acquired in the above-mentioned first model use phase in the second model.

[0090] The walking speed estimation device may also determine an NSL subgroup of a walking speed estimation subject using the walking speed calculated by the first model. In this case, the NSL of a walking speed estimation subject is calculated by NSL=(walking speed of the first model) / (cadence / 60) / height, and the NSL subgroup may be determined using the calculated NSL. The walking speed estimation device may also calculate the walking speed by substituting the acquired walking parameter value into a second model linear regression equation (a regression equation of the corresponding NSL subgroup). According to an embodiment of the present invention, when estimating walking speed using the developed second model, even if a gait measurement device (e.g., GAITRite) is not present, it is possible to easily estimate the walking speed by NSL subgroup using only an IMU.

[0091] Next, the present invention can perform statistical analysis of the developed walking speed estimation model. According to one embodiment, the present invention can compare continuous variables using Student's t-test or analysis of variance (ANOVA) and compare categorical variables using a between-groups chi-square test. Using this statistical analysis method, the present invention can compare the statistical values ​​estimated by Gaitrite and the statistical values ​​estimated by the walking speed estimation model of the present invention. More specifically, the statistical values ​​compared may be intraclass correlation coefficient (ICC), mean error (ME), mean absolute error (MAE), and root mean square error (RMSE).

[0092] Furthermore, the present invention can compare the ICC, ME, MAE and RMSE of the estimated walking speed between the walking speed estimation models using repeated measures ANOVA. The present invention can perform all statistical analysis using the Statistical Package for Social Sciences version 25.0 (International Business Machines Corporation, Armonk, NY). In one embodiment of the present invention, the 0th model of the walking speed estimation model of the present invention can be set to the existing technology, and the 1st model and the 2nd model can be set to the developed technology to perform the statistical analysis.

[0093] FIG. 7 is a comparative statistical analysis example of a walking speed estimation model according to one embodiment compared to other models.

[0094] Referring to Fig. 7, the present invention can compare the accuracy of the walking speed estimation method of the present invention by using the walking speed measured by GAITRite as a comparison standard. Referring to Fig. 7, it can be seen that for all subjects, the first and second models show lower MAE, ME, and RMSE and better ICC than the 0th model. Furthermore, it can be seen that the second model shows lower MAE than the first model (p=0.007).

[0095] 7, subjects can be grouped into three subgroups (slow, medium, and fast) using the walking speed based on GAITRite according to the second model of the present invention. Slow walking was defined as a standard deviation below the mean (17.9 m / s) and fast walking was defined as a standard deviation above the mean (114.4 m / s).

[0096] Gait speeds estimated by the 0th model tended to be slower than those measured by GAITRite in the fast NSL subgroup (p<0.001), while they tended to be faster in the slow NSL subgroup (p<0.001). This was also true for gait speeds estimated by the 1st and 2nd models, but both IMU and GAITRite were significantly faster. TMIt can be seen that the difference in gait speed measurements between and was reduced in the fast NSL subgroup (p<0.001).

[0097] In summary, the linear regression walking speed estimation according to an embodiment of the present invention is essentially to calculate an approximation of the mathematical expression of human walking speed. The linear regression in the model of the present invention is a form of simplified Taylor expansion of the virtual true expression, which may include higher degrees and number of variables. Furthermore, the present invention can improve the accuracy of the IMU-based walking speed estimation by optimizing the linear regression model with the NSL subgroup using additional gait parameters in the linear regression model.

[0098] Furthermore, the walking speed estimation model of the present invention can use NSL to develop subgroup models. As stride length decreases with age, stride length can reflect the walking patterns of elderly people. As shown in Figure 5 above, anthropometric measurements differ among the three NSL subgroups, and each of the three subgroups can have different walking patterns associated with different anthropometric characteristics.

[0099] FIG. 8 is a walking model of a walking speed estimation model according to one embodiment.

[0100] Our model can minimize the demographic and anthropometric influences on walking speed estimation by using NSL instead of the original stride length value. Human gait during stance phase can be modeled as an inverted pendulum with a spring attached as shown in Figure 8, where CP is the Contact Point. The leg length during stance phase can be expressed as the following equation (4): l is the leg length and β1 is the leg bending coefficient.

[0101] [Number 4] JPEG2025501161000003.jpg719

[0102] Also, the stride length can be calculated. Taking into account the conservation of angular momentum and the assumption of a point mass, the energy conservation after contact (E con ) can be expressed as the following equation (5). In this case, m1 is the mass point, ω1 is the angular velocity before contact, and l l is the leg length, l s is the stride length.

[0103] [Number 5] JPEG2025501161000004.jpg1551

[0104] Therefore, taking into account the symmetry of movement, the stride length can be calculated using the following equation (6).

[0105] [Number 6] JPEG2025501161000005.jpg734

[0106] Therefore, the NSL can be calculated as follows: where α1 is the leg / height ratio of the person and a is the angular displacement in gait mechanics.

[0107] [Number 7] JPEG2025501161000006.jpg1281

[0108] In the first model of the present invention, the VHD coefficients are different from each other among the three NSL subgroups. More specifically, in the second model, the VHD coefficient is the highest in the short NSL subgroup and the lowest in the long NSL subgroup, but B is the smallest in the short NSL subgroup, so it can be seen that the short NSL subgroup has less effective dynamic walking and body rotation compared to the other NSL subgroups. In addition, the coefficient of age and B value are the lowest in the long NSL subgroup, which suggests that the walking of the long NSL subgroup may not be affected by age. In addition, the cadence is lower in the short NSL subgroup, which indicates that the short NSL subgroup may have a more conservative walking pattern compared to the other subgroups.

[0109] In summary, the present invention has generated a walking speed estimation model based on the results of analyzing the walking patterns of N elderly people. The present invention can improve the existing speed estimation model using additional walking parameters to reduce the training error and further advance the estimation equation using regularized subgroups.

[0110] According to one embodiment of the present invention, the additional gait parameters are roll angle, yaw angle and weight, which are related to the movement under the body center of mass in terms of angle and propulsion. The addition of the additional gait parameters can lead to better estimation results than the gender-based estimation of existing studies. Furthermore, since the NSL subgroups are generated by separating subjects with different gait mechanics, regardless of size, the walking speed estimation algorithm can be improved to a higher accuracy.

[0111] FIG. 9 is a block diagram of a walking speed estimation device according to an embodiment.

[0112] The walking speed estimation device 1100 in FIG. 9 may be the walking speed estimation device in FIG.

[0113] 9, the walking speed estimation device 1100 may include a communication unit 1110, a processor 1120, and a DB 1130. Only components related to the embodiment are shown in the walking speed estimation device 1100 in FIG. 9. Therefore, it may be understood by a person skilled in the art that the walking speed estimation device 1100 may further include other general-purpose components in addition to the components shown in FIG.

[0114] The communication unit 1110 may include one or more components that enable wired / wireless communication with other nodes. For example, the communication unit 1110 may include at least one of a short-range communication unit (not shown), a mobile communication unit (not shown), and a broadcast receiving unit (not shown).

[0115] The DB 1130 is hardware that stores various data processed in the walking speed estimation device 1100, and can store programs for processing and control of the processor 1120. The DB 1130 can store payment information, user information, and the like.

[0116] DB1130 may include random access memory (RAM), such as dynamic random access memory (DRAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), CD-ROM, Blu-ray or other optical disk storage, hard disk drive (HDD), solid state drive (SSD), or flash memory.

[0117] The processor 1120 controls the overall operation of the walking speed estimation device 1100. For example, the processor 1120 can generally control the input unit (not shown), the display (not shown), the communication unit 1110, the DB 1130, etc. by executing a program stored in the DB 1130. The processor 1120 can generally control the operation of the walking speed estimation device 1100 by executing a program stored in the DB 1130.

[0118] The processor 1120 may be implemented using at least one of application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, micro-controllers, microprocessors, and electrical units for performing other functions.

[0119] The embodiments according to the present invention can be implemented in the form of a computer program that can be executed on a computer through various components, and such a computer program can be recorded on a computer-readable medium, which can include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program instructions, such as ROMs, RAMs, flash memories, and the like.

[0120] On the other hand, the computer programs may be those specially designed and constructed for the present invention, or they may be those well known and available to those skilled in the computer software art. Examples of computer programs may include not only machine code, such as produced by a compiler, but also high-level language code that can be executed by a computer using an interpreter, etc.

[0121] According to one embodiment, the methods according to various embodiments of the present disclosure may be provided in a computer program product. The computer program product may be traded between sellers and buyers as a commodity. The computer program product may be distributed in the form of a machine-readable storage medium (e.g., compact discreet only memory (CD-ROM)) or may be distributed through an application store (e.g., Play Store, etc.). TM ), or directly online between two user devices (e.g., downloaded or uploaded). In the case of online distribution, at least a portion of the computer program product may be at least temporarily stored or temporarily generated on a machine-readable storage medium, such as a memory of a manufacturer's server, an application store's server, or an intermediary server.

Claims

1. In the walking speed estimation method, determining N subjects to be measured for walking speed measurement and acquiring physical data of the subjects; acquiring data on the results of a clinical evaluation performed on the subject; A step of acquiring gait data measured by performing a gait evaluation on the subject; and developing a walking speed estimation model having a plurality of gait parameters based on the measured gait data; The walking speed estimation model includes, as walking parameters, a roll angle, a yaw angle, and a body weight measured using an accelerometer worn at the center of gravity of the subject's body.

2. Determining the N subjects includes: The walking speed estimation method according to claim 1 , wherein the N subjects are determined to be people in a population excluding people who have a disease that may affect walking or whose balance ability evaluation is below a preset score.

3. 2. The method of claim 1, wherein the clinical assessment is to diagnose a psychiatric disorder according to pre-defined diagnostic criteria, determine an overall severity of cognitive impairment, and evaluate walking and balance ability using a performance oriented mobility assessment (POMA).

4. The step of acquiring the measured gait data includes: The walking speed estimation method according to claim 1 , wherein the gait characteristic values ​​of the subject are obtained using an accelerometer attached to the center of gravity of the body and an existing gait measurement device.

5. 2. The walking speed estimation method according to claim 1, wherein the accelerometer is an inertial measurement unit (IMU) attached to the waist of the subject with a predetermined adhesive material, and is a device capable of measuring acceleration in three axial directions.

6. The step of developing the walking speed estimation model includes: A first model performing a linear regression analysis including roll angle, yaw angle and body weight as gait parameters; a second model for dividing the subjects into a plurality of subgroups based on NSL and performing the linear regression analysis for each of the subgroups; The walking speed estimation method according to claim 1 , comprising:

7. The NSL is calculated by the following formula: In the above formula, l l is the leg length, l s is the stride length, α 1 is the leg / height ratio of a person, β 1 The walking speed estimating method according to claim 6, wherein x is a leg bending coefficient and a is an angular displacement.

8. Statistically analyzing the developed walking speed estimation model. The walking speed estimation method according to claim 1 , further comprising:

9. A step of calculating the walking speed of a person to be estimated by using the walking speed estimation model developed. The walking speed estimation method according to claim 1 , further comprising:

10. The step of calculating the walking speed includes:

10. The walking speed estimation method according to claim 9, further comprising: inputting and acquiring age, weight, and leg length data of the person to be estimated for walking speed; acquiring cadence, VHD (Vertical height displacement), roll angle, and yaw angle from an inertial measurement unit (IMU); and substituting the acquired values ​​into the walking speed estimation model to calculate the walking speed.

11. As a walking speed estimation device, A processor; [0023] a memory; The processor, A walking speed estimation device comprising: determining N subjects to be measured for walking speed; acquiring physical data of the subjects; conducting a clinical evaluation on the subjects; conducting a gait evaluation on the subjects to acquire measured walking data; and developing a walking speed estimation model having a plurality of walking parameters based on the measured walking data, wherein the walking speed estimation model includes, as walking parameters, a roll angle, a yaw angle, and body weight measured using an accelerometer worn at the center of gravity of the body of the subjects.

12. A computer-readable recording medium having recorded thereon a program for executing the method according to claim 1 on a computer.

Images