WO2016088842A1 - Gait analysis method and gait analysis system - Google Patents

Gait analysis method and gait analysis system Download PDF

Info

Publication number
WO2016088842A1
WO2016088842A1 PCT/JP2015/084034 JP2015084034W WO2016088842A1 WO 2016088842 A1 WO2016088842 A1 WO 2016088842A1 JP 2015084034 W JP2015084034 W JP 2015084034W WO 2016088842 A1 WO2016088842 A1 WO 2016088842A1
Authority
WO
WIPO (PCT)
Prior art keywords
sensor
subject
angular velocity
axis
body
Prior art date
Application number
PCT/JP2015/084034
Other languages
French (fr)
Japanese (ja)
Inventor
茂 但野
量 武田
晴一 遠山
佐野 嘉彦
証英 原田
Original Assignee
国立大学法人北海道大学
ニプロ株式会社
原田電子工業株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority to JP2014-244753 priority Critical
Priority to JP2014244753A priority patent/JP2018015017A/en
Application filed by 国立大学法人北海道大学, ニプロ株式会社, 原田電子工業株式会社 filed Critical 国立大学法人北海道大学
Publication of WO2016088842A1 publication Critical patent/WO2016088842A1/en

Links

Images

Classifications

    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Detecting, measuring or recording for diagnostic purposes; Identification of persons
    • A61B5/103Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb

Abstract

[Problem] To reduce the drift error that accumulates from integration. [Solution] A gait analysis method for analyzing the gait of a test subject using a wearable sensor, said method comprising executing a drift removal protocol for finding the orientation angle of each axis of a three-axis angular velocity sensor by first performing double time differentiation of the orientation angle of the sensor along each axis obtained from measurements by the sensor and, after removing the linear drift error, performing double time integration.

Description

Walking analysis method and walking analysis system

The present invention relates to a gait analysis method and a gait analysis system using a body-mounted sensor (also referred to as “wearable sensor”), and more particularly to a gait analysis method and gait for reducing drift of gait data measured by a body-mounted sensor. It relates to an analysis system.

Conventionally, there are walking analysis methods that use body-mounted sensors. It is theoretically possible to calculate the three-dimensional posture of the wearing part of the body-mounted sensor by calculating the angular displacement / position displacement by integrating the angular velocity / acceleration data measured by the body-worn sensor with time. It is. However, signal noise and errors accumulated in small amounts with time cause drift of the integrated value from the true value, and the three-dimensional posture (posture) of each body element of the subject who is walking cannot be correctly derived.

Many researchers have proposed a method for removing this noise (such as a calibration method using a geomagnetic sensor for a body-mounted sensor such as MTx (product name of Xsens Technologies, Enschede, The Netherlands)). However, since the influence of this noise appears when measurement is performed for a long time, a three-dimensional walking analysis method using a body-mounted sensor has not been established yet.

Incidentally, as a gait analysis method using a body-mounted sensor, for example, a method using the H-Gait system previously proposed by the present inventors in Non-Patent Document 1 has been known. In medical practice, a 10 m gait test is usually performed to evaluate patients with gait irregularities associated with stroke, spinal cord injury (SCI), osteoarthritis (OA), multiple sclerosis (MS), and the like. There are few reports of using body-worn sensors for 10-meter walking tests that provide clinicians with three-dimensional kinematic and spatiotemporal walking parameters. For this reason, the H-Gait system was developed with the intention of measuring walking for a short time, such as in a 10-meter walking test, for use in evaluating the walking ability and rehabilitation effects of patients with walking disorders.

The H-Gait system disclosed in Non-Patent Document 1 does not use a geomagnetic sensor that is affected by a surrounding magnetic field, but has a stretchable band that is in close contact with the subject's body, which is the pelvis (PE), left and right legs Three-axis acceleration sensors and 3 are attached to the thighs (RT, LT), the left and right leg shins (RS, LS), and the left and right leg feet (RF, LF), respectively, in small pockets provided in these bands. Using a body-mounted sensor configured to house a sensor unit having an axial angular velocity sensor (three-axis gyro sensor), acceleration and angular velocity are measured by each sensor unit while the subject walks on a flat road.

The initial posture of these sensor units is estimated using the gravitational acceleration component obtained from the acceleration data when the subject is standing and sitting, and the angle change is estimated using the angular velocity data when the subject is walking. . An algorithm based on quaternion calculation is executed to estimate the attitude of the sensor unit. The estimated posture of the sensor unit is converted into the posture of the body element by the rotation matrix obtained from the trial calculation, and the posture of the body element is determined based on the known motion trajectory analysis protocol for obtaining the motion trajectory of each lower limb part of the subject. It is used to construct a 3D wireframe model animation of the subject inside. The gait analysis was performed on five subjects, and the results were compared with those from a camera-based motion capture system. As a result, 10.14 °, 7.88 ° and 9 ° for hip joint angle, knee joint angle and ankle joint angle, respectively. An average rms error (RMSE) of .75 ° was obtained.

Tadano, Takeda and Miyagawa, 3D walking analysis using wearable acceleration and angular velocity sensors based on quaternion calculation, Sensors 2013; 13; 9321-9343 (http://www.mdpi.com/1424-8220/13 / 7/9321)

However, in the above conventional H-Gait system, the drift of the attempt was made by subtracting the difference between the first value and the last value of each attitude angle (roll, pitch and yaw) of the sensor unit from the entire signal. It was not completely eliminated, and drift appeared even with a measurement time of 14 seconds, affecting the final results both in terms of joint kinematics and the position and posture of the wireframe model in space. Therefore, an object of the present invention is to provide a gait analysis method and a gait analysis system that improve the gait analysis method described in Non-Patent Document 1 and reduce the drift of gait data measured by a body-mounted sensor.

The present invention advantageously solves the problems of the conventional walking analysis method described above, and the walking analysis method according to the present invention has a three-axis angular velocity sensor mounted on a plurality of body parts including the lower limbs of a subject. When performing a gait analysis of a subject using a type sensor,
After always estimating that the drift error increases linearly with time, the attitude angle of each axis of the sensor obtained from the measurement value of each triaxial angular velocity sensor is first time-differentiated twice to remove the linear drift error The drift removal protocol for obtaining the attitude angle of each axis of the sensor from which the drift error has been removed by executing time integration twice is executed.

The gait analysis system of the present invention includes a body-mounted sensor having a triaxial angular velocity sensor that is mounted on each of a plurality of body parts including the lower limbs of the subject, and inputs measurement data from the body-mounted sensor to the subject. In the gait analysis system that performs gait analysis of
Each attitude angle of each sensor axis obtained from the measured values of the three-axis angular velocity sensors is first time-differentiated twice to remove the linear drift error, and then integrated twice for each sensor. A drift removing means for obtaining a posture angle of the shaft is provided.

According to the walking analysis method and the walking analysis system of the present invention, it is always estimated that the drift error increases linearly with time, and the posture angle of each sensor axis obtained from the measured value of each triaxial angular velocity sensor is first twice. By performing a drift removal protocol that obtains the posture angle of each axis of the sensor by integrating the time twice after removing the linear drift error by time differentiation, from the posture angle displacement of each axis of the sensor as the subject walks The drift error can be effectively reduced, and the motion trajectory of each lower limb part of the subject can be obtained with high accuracy, and consequently, the three-dimensional wire frame model animation of the subject during walking can be constructed with high positional accuracy. it can.

In the gait analysis method of the present invention, the body-mounted sensor is mounted on each of a plurality of body parts including a lower limb of a subject, and a plurality of sensor units each having a triaxial acceleration sensor together with the triaxial angular velocity sensor. When the subject's gait analysis is performed using the body-mounted sensor, the body part is used by using the gravitational acceleration vector obtained from the measurement value of the acceleration sensor of each lower limb part in at least two kinds of postures of the subject. A calibration protocol for reducing the sensor mounting error may be executed.

Furthermore, in the gait analysis method of the present invention, when analyzing the gait of the subject using the body-mounted sensor, high-frequency noise is removed from the raw measurement data of each triaxial angular velocity sensor using a low-pass filter. A filtering protocol may be executed, and an offset removal protocol is executed to remove the offset value from the measurement data of each angular velocity sensor by subtracting the mode value of the measurement data from the measurement data of each of the three-axis angular velocity sensors. Also good.

In the walking analysis system of the present invention, the body-mounted sensor is mounted on a plurality of body parts including a lower limb of a subject, and a plurality of sensor units each having a triaxial acceleration sensor together with the triaxial angular velocity sensor. The gait analysis system includes a gravitational acceleration vector obtained from a measurement value of the three-axis acceleration sensor of each lower limb portion in at least two types of postures of the subject, and an error in mounting the sensor unit to the body portion Calibration means may be provided to reduce the.

Furthermore, the gait analysis system of the present invention may include filtering means for removing high-frequency noise from the raw measurement data of each angular velocity sensor using a low-pass filter, and each of the three-axis angular velocity sensors. You may provide the offset removal means which removes an offset value from the measurement data of each triaxial angular velocity sensor by subtracting the mode value of the measurement data from the measurement data.

In the walking analysis system of the present invention, each of the body-mounted sensors is mounted at positions corresponding to a plurality of body parts including the subject's lower limbs of a stretchable exercise clothing worn by the subject. It is preferable to provide a plurality of sensor units having an axial acceleration sensor and a triaxial angular velocity sensor because the position of each sensor unit relative to the body part does not shift during the walking motion of the subject.

In the gait analysis system of the present invention, it is preferable that the low-pass filter of the filtering means is an infinite impulse response digital Butterworth filter because there is almost no signal deterioration in a low frequency band that passes through.

Furthermore, in the gait analysis system of the present invention, when provided with motion trajectory analysis means for obtaining the motion trajectory of each lower limb part of the subject from the posture angle displacement of each axis of the sensor accompanying the walking of the subject, This is preferable because the motion trajectory can be obtained with high accuracy.

BRIEF DESCRIPTION OF THE DRAWINGS In the front view (left figure) and side view (right figure) which show the sensor mounting arrangement | positioning and walking wire frame model in the walking analysis system of one Embodiment of this invention and the walking analysis method of one Embodiment of this invention using the same is there. It is explanatory drawing which shows the method used for the detection of foot HC and TO timing in the said embodiment, and has shown the change with time progress of the angular velocity of an ankle joint, and the relative position of a toe. It is a perspective view which shows the test subject's wire frame model and its coordinate system in the said embodiment. It is a relationship diagram showing the change over time of the hip joint flexion angle showing the analysis result according to the above embodiment and the analysis result by the other two gait analysis methods with different signal drift reduction protocols, the dotted line is the raw data, A broken line indicates IIR + offset removal, and a solid line indicates the embodiment. It is a relationship diagram which shows the change with time progress of the knee joint bending angle which shows the analysis result by the said embodiment, and the analysis result by the other two gait analysis methods from which the signal drift reduction protocol differs, and a dotted line is raw data The broken line indicates IIR + offset removal, and the solid line indicates the above embodiment. It is a relationship diagram which shows the change with time progress of the ankle joint bending angle which shows the analysis result by the said embodiment, and the analysis result by the other two gait analysis methods from which the signal drift reduction protocol differs, and a dotted line is raw data The broken line indicates IIR + offset removal, and the solid line indicates the above embodiment. (A)-(E) are in the sagittal plane of the greater trochanter (GT), knee joint center (Knee) and ankle joint center (Ankle) of the right leg during one walking cycle of each of five subjects. It is explanatory drawing which shows the plot by. (A)-(E) are in the sagittal plane of the greater trochanter (GT), knee joint center (Knee), and ankle joint center (Ankle) of the left leg during three gait cycles of each of five subjects. It is explanatory drawing which shows the plot by. (A) to (E) are explanatory diagrams showing plots in the horizontal plane at the center of the knee joint (left figure) and ankle joint (right figure) during three walking cycles of each of five subjects.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the gait analysis method of this embodiment, gait analysis of five healthy subjects is measured using the H-Gait system (see Non-Patent Document 1) that constitutes the main part of the gait analysis system of this embodiment. It was. This H-Gait system does not require an external magnetic field for reference, and the measured values are only from the 3-axis acceleration sensor and 3-axis angular velocity sensor of the sensor unit which is arranged at right angles and fixed at 7 places on the lower limb of the subject. Collected.

As shown in FIG. 1, the body-mounted sensor used in the H-Gait system in this embodiment includes a plurality of body parts including the lower limbs of the subject, that is, the pelvis part, and the left and right large parts. Five sensor units each mounted with a three-axis acceleration sensor and a three-axis angular velocity sensor stored in pockets provided at positions corresponding to the thigh and the left and right shins, respectively, Three-axis acceleration sensors and three-axis angular velocities that are housed in pockets provided at positions corresponding to the insteps of the elastic ring bands attached to the left and right legs, respectively, and mounted at those positions. The sensor unit is equipped with two sensor units, and each sensor unit is thus attached to a stretchable exercise suit that is in close contact with the subject's body. The positional deviation of each sensor unit with respect to the body portion during walking motion of the subject is prevented. The data measured by each sensor unit is sent wirelessly during a walking test to a relay router installed at a fixed place, for example, as in the conventional H-Gait system, and from there, the data of the H-Gait system of this embodiment is preliminarily stored. It is transferred to a normal personal computer that stores the program, or it is sent to a storage medium such as a USB memory possessed by the subject wirelessly or by wire during a walking test and temporarily stored there, and from that storage medium after the walking test It is transferred to the personal computer.

The acceleration and angular velocity data measured by each sensor unit during the horizontal walking of the subject is thus collected in the personal computer, and the measured data is then used using an algorithm based on the quaternion of the H-Gait system. These are converted into a three-dimensional posture of each sensor unit and the body part to which they are attached. In order to reduce errors caused by measurement data drift, a novel countermeasure is implemented in this embodiment. This new countermeasure includes a sensor mounting adjustment (calibration) protocol, a Butterworth filter design, removal of sensor offset values and a double differentiation / double integration method. By implementing these countermeasures, the drift is remarkably reduced as described later. As a result, the hip / knee and ankle flexion / extension (FE) angles, the hip / knee adduction-abduction (AA) angles, / Lower limb joint kinematics such as knee-ankle medial-lateral (IE) rotation angles are provided. In addition, a moving wireframe model is created to visually confirm the walking movement. Furthermore, spatio-temporal parameters such as walking cycle, pace, step length, stride length, stride length, stance ratio, and free leg ratio are calculated from the timing of foot contact (HC) and toe separation (TO). These methods are described in detail below.

(1) Body Part Posture Calculation Using Body-Worn Sensor In this embodiment, the personal computer performs the walking analysis method that is generally the same as in the H-Gait system. The walking posture is calculated by finding the gravitational acceleration direction from the gravitational acceleration component included in the output data of the acceleration sensor, and the initial three-dimensional posture of the body part to which the sensor is attached is calculated. A three-dimensional posture subsequent to the initial three-dimensional posture is estimated by integration of angular velocities measured by a three-axis angular velocity sensor. In order to show the three-dimensional rotational movement of each lower limb joint, the angular displacement is expressed using a posture expression based on a quaternion. Finally, the characteristic walking motion obtained from these angular displacements is expressed using a three-dimensional wire frame model. FIG. 1 shows a three-dimensional wire frame model obtained from the study of the H-Gait system. This wireframe model shows the posture of each body part, the subject's iliac crest width (bicristal breadth), superior anterior iliac spine width (iliospinal breadth), superior anterior iliac spine height (iliospinal height), tibial height ( Created with specific body dimensions such as tibial height and sphyrion height. The sensor unit is attached to seven body parts of the lower limbs.

(2) Countermeasures for signal noise One of the difficulties involved in using body-mounted sensors for gait analysis is sensor drift. Theoretically, it is possible to integrate the angular velocity to calculate the posture of the body-mounted sensor. However, signal noise and errors that accumulate in small amounts over time cause the integrated value to drift from the true value. The following measures are taken to reduce the effects of signal drift.

(2-1) Calibration protocol Calibration for reducing the mounting error of the sensor to the body part is executed. Execution of this calibration protocol by the personal computer constitutes a calibration means. In order to obtain a rotation matrix for converting the sensor coordinate system to the body part coordinate system, the procedure introduced in the walking analysis method of the H-Gait system is used. This procedure includes two simple steps: measuring the gravitational acceleration vector of each lower limb portion in two different postures, standing and sitting for the subject. Since acceleration data includes a gravitational acceleration component, the angle formed by the sensor with respect to the direction of gravity during standing and sitting, that is, a three-dimensional posture (posture) can be calculated. When standing and sitting, the body-mounted sensor is placed in a two-dimensional sagittal plane and assumes only rotational movement on the sagittal plane, and a rotation matrix that converts the dimensions of the sensor coordinate system to the global coordinate system Is guided. Execution of this protocol leads to minimization of wearing errors that occur with the use of body-worn sensors.

(2-2) Digital Filtering Protocol An IIR (infinite impulse response) digital fourth order Butterworth filter that constitutes a digital low pass filter is executed to remove high frequency noise from the raw data of the triaxial angular velocity sensor. Execution of this digital filtering protocol by the personal computer constitutes a filtering means. This low-pass filter is executed using a MATLAB (name of technical calculation language of Massworks, Natick, Massachusetts, USA) algorithm, where the cutoff frequency is set to, for example, 12 Hz by the Nyquist method. This low pass filter is applied to the acceleration data in the direction of travel and in the reverse direction to cancel out the phase lag caused by the Butterworth filter.

(2-3) Offset removal protocol The angular velocity data from the triaxial angular velocity sensor includes an offset value. This offset value is the mode value (mode of data) in the stationary state for each axis of the sensor unit. Value) and subtract it from the whole signal. Execution of this offset removal protocol by the personal computer constitutes an offset removal means.

(2-4) Drift Removal Protocol In the method of this embodiment, it is always estimated that the drift error of the joint angle increases linearly with time, and based on the estimation, the drift removal that is mathematically more robust (robust) than in the past Technology (DDI protocol) is implemented. That is, here, the attitude angle θout_i (t) of the sensor is obtained once along each axis (x, y, z). The true angle θi (t) and error ei (t) are estimated as follows:

Figure JPOXMLDOC01-appb-M000001

Then, the linear increase ei (t) is removed by two time differentiation operations. The drift error is estimated to increase linearly with time, and according to equations (2) and (3), once the drift error is differentiated with respect to time, it becomes a constant (const), and when it is differentiated with time again, the constant is removed.

Figure JPOXMLDOC01-appb-M000002

¡Two time integrals are calculated in relation to the fact that attitude data is always required for further analysis. This requires the addition of appropriate integration constants (c1, c2) at each stage of the calculation as follows.

Figure JPOXMLDOC01-appb-M000003

Integral constant c1 is considered the initial angular velocity. In this embodiment, since any walking test is considered to start from a stationary state (stance phase), the initial angular velocity is 0 (zero). In addition, the integration constant c2 is considered to be the initial posture. Therefore, the initial posture calculated from the measurement data of the acceleration sensor in the stationary state (stance phase) is input to the integration constant c2. The drift error of each joint angle is removed by the above-described signal processing and calculation method used for reducing the noise and offset of the triaxial angular velocity sensor data. Execution of this drift removal protocol by the personal computer constitutes drift removal means.

(3) Derivation of spatiotemporal and joint kinematic gait parameters Gait is usually defined in terms of temporal and spatial factors that indicate both the time when a gait event occurs and the position and posture of the lower limb in space, respectively. The walking cycle is generally divided into a stance phase and a swing phase. The first begins with an initial foot contact called heel contact (HC) and the second begins with a toe-off (TO) event. Based on these main temporal events, the walking cycle (GC), pace (CD), stride length (SR), step length (SL), step width (SW), stance ratio (STR) that occur during one walking cycle A spatio-temporal walking parameter such as

In this embodiment, as listed in Table 1, we start by identifying both HC and TO timing events for each foot and take into account spatio-temporal parameters and gait phase (time). The timing of HC and TO can be detected by a tibial accelerometer. Here, the timing events are automatically and directly identified from the angular velocities measured and recorded by the sensor units arranged on both shins.

Figure JPOXMLDOC01-appb-T000004

FIG. 2 shows the method used for detecting the HC and TO timing of the foot, the vertical axis shows the angular velocity and the relative position of the toe, and the horizontal axis shows the time. HC timing is detected by a characteristic lateral angular velocity peak and is shown by a solid circle. The TO timing is detected by measuring the negative peak of the relative distance of the toe position relative to the origin of the pelvic (PE) coordinate system and is indicated by a dashed circle. As shown in FIG. 2, HC is detected by a characteristic lateral angular velocity peak, and TO timing is detected by measuring the negative peak of the relative distance of the toe position relative to the origin of the pelvis (PE) coordinate system. Based on the appropriate peak of angular velocity along the horizontal axis of each part, the walking cycle, stance ratio and free leg ratio are calculated, and by measuring the HC position of both legs, the pace, stride length, step length and step width are Calculated.

FIG. 3 shows a wire frame model representing the body of the subject. Xglobal, Yglobal, and Zglobal indicate coordinate axes of the global coordinate system, where the Xglobal axis is the walking direction, the Yglobal axis is the left lateral direction, and the Zglobal axis is the vertical direction. Xlocal, ylocal, zlocal and x'local, y'local, z'local are coordinate axes of a new foot partial coordinate system based on each step of walking. PE, RT, LT, RS, LS, RF, and LF indicate each body part.

FIG. 3 shows the relationship between the global coordinate system used in this embodiment and the new foot partial coordinate system created at each step. If RF (right foot) is on the ground between HC and foot flat (FF), foot partial coordinate systems xlocal, ylocal and zlocal are created from the heel position of RF. Between LF FF and HC, foot partial coordinate systems xlocal, ylocal and zlocal are created from the RF toe position. The 3D poses of other body parts in the global coordinate system are calculated based on their relative poses with respect to RF. Therefore, the three-dimensional posture of the body part in the global coordinate system is calculated in the order of RF → RS → RT → PE → LT → LS → LF. This order continues until the LF arrives on the ground, where new foot partial coordinate systems x'local, y'local and z'local are created at the landing location, and the order of posture calculation starts from LF and LF → LS → LT → PE → RT → RS → RF.

In this embodiment, the motion trajectory of each joint during one walking cycle of each subject is obtained from the joint angle obtained by the above-described procedure by executing the motion trajectory analysis protocol similar to that in the conventional H-Gait system. It is done. Execution of this motion trajectory analysis protocol by the personal computer constitutes motion trajectory analysis means. Analysis of joint kinematics regarding gait tendencies is generally performed during the clinical gait analysis period. According to the gait analysis method of this embodiment, the crotch / knee / ankle flexion / extension (FE) angle, the crotch / knee abduction-abduction (AA) angle, and the crotch / knee / ankle medial-lateral (IE) rotation angle, etc. Kinematic gait parameters can be calculated.

(4) Experiment The experiment was conducted by five healthy subjects on a straight flat floor indoors. The body dimensions of subjects A to E were measured before the experiment and the results are shown in Table 2. Body dimensions are the body part dimensions between each anatomical feature, ie, the distance from the greater trochanter (GT) to the lateral tibial condyle (LCT), the distance from the LCT to the ankle joint (Ankle) (LCT-Ankle), ankle joint height (Ankle Height), right and left GT width (Right-Left GT Width). To simplify the calculations, the body dimensions of healthy subjects were estimated to be symmetrical, and only the right lower limb dimensions were measured.

Figure JPOXMLDOC01-appb-T000005

The subject was requested to perform a walking test consisting of a stationary state, a 10 m walking test, and a resting state again. The subject has a sports suit made of spandex that fits the body with five small pockets for holding the sensor unit of the body-mounted sensor of the H-Gait system at positions corresponding to each body part excluding both feet. Suit pants) and an elastic band that fits the shoe and has a small pocket for holding the sensor unit at a position corresponding to the upper part of each foot. Reflective markers were placed on 10 anatomical features of the lower limbs, and still images were taken from the front and both sides (see FIG. 1). The walking distance was approximately 10 m, which was equivalent to 10 steps (5 steps for each left and right leg).

(5) Results The joint angles (hip flexion angle, knee flexion angle and ankle flexion angle) obtained by different methods of the signal drift reduction protocol (raw data, IIR + offset removal, IIR + offset removal + DDI) for each lower limb joint. A comparison of the results is shown in FIGS. FIGS. 4-1 to 3 show joint angles (hip flexion angle, knee flexion angle, and ankle flexion angle) obtained by different methods of the signal drift reduction protocol (raw data, IIR + offset removal, IIR + offset removal + DDI) for the subjects. ), The hip joint flexion angle, FIG. 4-2 the knee joint flexion angle, and FIG. 4-3 the ankle joint flexion angle.

This result is an example of a subject walking for 10 seconds. The raw data shown in dotted lines is a joint angle that is calculated without performing any signal drift reduction protocol and is therefore highly error effective. Between 1.4 and 7.8 seconds, as the subject walked, the drift increased linearly. This is why the DDI method was a suitable method for eliminating linear drift effects during dynamic conditions such as walking. The IIR + offset removal protocol indicated by the broken line refers to the conventional H-Gait system (Non-Patent Document 1). The IIR + offset removal + DDI protocol indicated by the solid line shows the result of the gait analysis method of this embodiment.

FIGS. 5A to 5E show one gait of each subject obtained by executing the same motion trajectory analysis protocol as in the conventional H-Gait system from the joint angle obtained by IIR + offset removal + DDI described above. Figure 6 shows plots in the sagittal plane Zglobal-Xglobal of the greater trochanter (GT), knee joint center and ankle joint center of the right leg during the cycle. The vertical axis indicates the z axis in the global coordinate system, and the horizontal axis indicates the x axis in the global coordinate system. These trajectories are plotted at a sampling rate of 33 Hz.

6 (A) to 6 (E) show the three walks of each volunteer obtained by executing the same motion trajectory analysis protocol as in the conventional H-Gait system from the joint angles obtained by IIR + offset removal + DDI described above. FIG. 5 shows plots in the sagittal plane Zglobal-Xglobal of the greater trochanter (GT), knee joint center and ankle joint center of the left leg during the cycle. The vertical axis indicates the z axis in the global coordinate system, and the horizontal axis indicates the x axis in the global coordinate system. These trajectories are also plotted at a sampling rate of 33 Hz.

FIGS. 7A to 7E show three walks of each subject obtained by executing the same movement trajectory analysis protocol as in the conventional H-Gait system from the joint angle obtained by IIR + offset removal + DDI described above. Shown are plots projected on the horizontal plane Xglobal-Yglobal at the center of the knee joint (left) and ankle joint (right) during the cycle. The vertical axis indicates the x axis in the global coordinate system, and the horizontal axis indicates the y axis in the global coordinate system. The left leg is shown on the left side (shown in red), the right leg is shown on the right side (shown in blue), and the black line resulting from the locus of the ankle joint is the toe. Indicates the direction. These trajectories are also plotted at a sampling rate of 33 Hz.

Table 3 shows the results of spatiotemporal parameters for each subject. Here, the walking cycle, pace, step length, step width, stride length, asymmetric index, stance ratio, and free leg ratio are shown.

Figure JPOXMLDOC01-appb-T000006

The purpose of the gait analysis method and gait analysis system of this embodiment was to remove the drift effect in order to improve the measurement accuracy of gait using a body-mounted sensor. The above results show that the implementation of a combination of multiple measures, ie, sensor mounting error reduction protocol, IIR digital fourth order Butterworth filter, offset removal protocol, and DDI method resulted in a reduction in signal drift.

The results shown in FIGS. 4-1 to 3 show the difference in joint angle after walking for approximately 10 seconds. The difference between the average of all 5 subjects when raw data and IIR + offset removal + DDI after 10 seconds is 2.1 ° for the hip joint angle, the knee joint angle, and the ankle joint angle, respectively. 3 ° and 15.6 °. In addition, the difference between performing IIR + offset removal and IIR + offset removal + DDI was 6.2 °, 6.6 °, and 2.2 ° for hip, knee, and ankle angles, respectively. . This means that the proposed measure can eliminate drift errors by an average of 17 ° compared to the integration of raw angular velocity data and an average of 5 ° compared to the conventional H-Gait system (Non-patent Document 1). Indicates that

The results shown in FIGS. 5 and 6 indicate that according to the walking analysis method of this embodiment, the difference in joint trajectory between the left and right GTs, the knee joint, and the ankle joint is visualized using a wire frame model in the sagittal plane. It can be compared. These kinematic gait parameters allow comparison of knee stretch angles between different timings during the gait cycle.

The result shown in FIG. 7 enables comparison of relative trajectories in the horizontal plane between the knee joint center and the ankle joint center. According to this data, the left-right symmetry of the trajectories of the left and right knee joints and the ankle joint can be compared. In addition, the spatiotemporal parameters provided in Table 3 make it possible to quantify the difference between left and right walking events. By combining kinematic and spatio-temporal parameters, it is possible to detect differences in walking between the left and right lower limbs, and how these differences affect the posture of the body part and the joint position during walking. Can know.

While the present invention has been described above based on the embodiments, the present invention is not limited to the above-described embodiments, and can be appropriately changed within the scope of the claims. For example, the walking analysis method and the walking analysis of the present invention The system is not limited to application to the H-Gait system, and can be applied to other commercially available wearable sensor systems using a triaxial acceleration sensor and a triaxial angular velocity sensor. The data obtained with these systems is the data for follow-up diagnosis after total knee arthroplasty (TKA) by using gait parameters for gait recovery or quantifying the effects of special implants during gait. It would be useful in clinical settings.

Thus, according to the walking analysis method and the walking analysis system of the present invention, the drift error is effectively reduced from the posture angle displacement of each sensor axis accompanying the walking of the subject, and the motion trajectory of each lower limb portion of the subject is highly accurate. As a result, a three-dimensional wire frame model animation of a walking subject can be constructed with high positional accuracy.

PE Pelvis LT Left thigh RT Right thigh LS Left shin RS RS Right shin LF Left foot RF Right foot

Claims (9)

  1. When performing a gait analysis of a subject using a body-mounted sensor having a triaxial angular velocity sensor that is attached to each of a plurality of body parts including a lower limb of the subject,
    After always estimating that the drift error increases linearly with time, the attitude angle of each axis of the sensor obtained from the measurement value of each triaxial angular velocity sensor is first time-differentiated twice to remove the linear drift error A gait analysis method characterized by executing a drift removal protocol for obtaining a posture angle of each axis of a sensor from which a drift error has been removed by performing time integration twice.
  2. In a gait analysis system comprising a body-mounted sensor having a three-axis angular velocity sensor mounted on each of a plurality of body parts including a lower limb of a subject, and performing measurement analysis of the subject by inputting measurement data from the body-mounted sensor,
    Each attitude angle of each sensor axis obtained from the measured values of the three-axis angular velocity sensors is first time-differentiated twice to remove the linear drift error, and then integrated twice for each sensor. A gait analysis system comprising drift removing means for determining a posture angle of an axis.
  3. The body-mounted sensor has a plurality of sensor units each having at least the three-axis angular velocity sensor, each mounted at a position corresponding to a plurality of body parts including a lower limb of a subject in a stretchable exercise clothes worn by the subject. The gait analysis system according to claim 2, comprising:
  4. The walking according to claim 2, wherein the body-mounted sensor comprises a sensor unit having at least the triaxial angular velocity sensor mounted on a band fastened to a body part of a subject. Analysis system.
  5. The body-mounted sensor has a plurality of sensor units each having a three-axis acceleration sensor together with the three-axis angular velocity sensor, which are respectively mounted on a plurality of body parts including a lower limb of a subject.
    The gait analysis system uses a gravitational acceleration vector obtained from a measurement value of the three-axis acceleration sensor of each lower limb portion in at least two kinds of postures of the subject to perform calibration that reduces the mounting error of the sensor unit to the body portion. The gait analysis system according to any one of claims 2 to 4, further comprising an action means.
  6. The walking analysis system according to any one of claims 2 to 5, further comprising filtering means for removing high-frequency noise from the measurement data of each angular velocity sensor using a low-pass filter.
  7. The walking analysis system according to claim 6, wherein the low-pass filter of the filtering means is an infinite impulse response digital Butterworth filter.
  8. 3. The apparatus according to claim 2, further comprising offset removing means for removing an offset value from the measurement data of each angular velocity sensor by subtracting the mode value of the measurement data from the measurement data of each angular velocity sensor. The gait analysis system according to any one of 7 to 7.
  9. 9. A motion trajectory analyzing means for obtaining a motion trajectory of each lower limb portion of the subject from a posture angle displacement of each axis of the three-axis angular velocity sensor accompanying the walking of the subject is provided. The gait analysis system according to item 1.
PCT/JP2015/084034 2014-12-03 2015-12-03 Gait analysis method and gait analysis system WO2016088842A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
JP2014-244753 2014-12-03
JP2014244753A JP2018015017A (en) 2014-12-03 2014-12-03 Gait analysis method and gait analysis system

Publications (1)

Publication Number Publication Date
WO2016088842A1 true WO2016088842A1 (en) 2016-06-09

Family

ID=56091782

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2015/084034 WO2016088842A1 (en) 2014-12-03 2015-12-03 Gait analysis method and gait analysis system

Country Status (2)

Country Link
JP (1) JP2018015017A (en)
WO (1) WO2016088842A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017014294A1 (en) * 2015-07-23 2017-01-26 国立大学法人北海道大学 Gait analysis method and gait analysis system

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006068312A (en) * 2004-09-02 2006-03-16 Tamagawa Seiki Co Ltd Posture monitoring method for rehabilitation and posture monitor for rehabilitation
JP2011092274A (en) * 2009-10-27 2011-05-12 Sumitomo Precision Prod Co Ltd Detector

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006068312A (en) * 2004-09-02 2006-03-16 Tamagawa Seiki Co Ltd Posture monitoring method for rehabilitation and posture monitor for rehabilitation
JP2011092274A (en) * 2009-10-27 2011-05-12 Sumitomo Precision Prod Co Ltd Detector

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
TAKEDA RYO ET AL.: "Drift Removal for Improving the Accuracy of Gait Parameters Using Wearable Sensor Systems", SENSORS(WEB, vol. 14, no. 12, pages 23230 - 23247 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017014294A1 (en) * 2015-07-23 2017-01-26 国立大学法人北海道大学 Gait analysis method and gait analysis system

Also Published As

Publication number Publication date
JP2018015017A (en) 2018-02-01

Similar Documents

Publication Publication Date Title
El-Gohary et al. Human joint angle estimation with inertial sensors and validation with a robot arm
Charlton et al. Reliability and concurrent validity of a Smartphone, bubble inclinometer and motion analysis system for measurement of hip joint range of motion
Clark et al. Validity of the Microsoft Kinect for assessment of postural control
Yeung et al. Evaluation of the Microsoft Kinect as a clinical assessment tool of body sway
Bonnechere et al. Validity and reliability of the Kinect within functional assessment activities: comparison with standard stereophotogrammetry
Mentiplay et al. Gait assessment using the Microsoft Xbox One Kinect: Concurrent validity and inter-day reliability of spatiotemporal and kinematic variables
Silder et al. Men and women adopt similar walking mechanics and muscle activation patterns during load carriage
Willson et al. Male and female gluteal muscle activity and lower extremity kinematics during running
Kristianslund et al. Effect of low pass filtering on joint moments from inverse dynamics: implications for injury prevention
Schepers et al. Ambulatory estimation of center of mass displacement during walking
Aminian et al. Capturing human motion using body‐fixed sensors: outdoor measurement and clinical applications
Tong et al. A practical gait analysis system using gyroscopes
Eltoukhy et al. Improved kinect-based spatiotemporal and kinematic treadmill gait assessment
Dejnabadi et al. A new approach to accurate measurement of uniaxial joint angles based on a combination of accelerometers and gyroscopes
Takeda et al. Gait posture estimation using wearable acceleration and gyro sensors
Mayagoitia et al. Accelerometer and rate gyroscope measurement of kinematics: an inexpensive alternative to optical motion analysis systems
Lin et al. Human pose recovery using wireless inertial measurement units
Frigo et al. Functionally oriented and clinically feasible quantitative gait analysis method
Gouwanda et al. Identifying gait asymmetry using gyroscopes—A cross-correlation and Normalized Symmetry Index approach
Lee et al. The use of a single inertial sensor to identify stride, step, and stance durations of running gait
Fluit et al. Prediction of ground reaction forces and moments during various activities of daily living
Manal et al. Comparison of surface mounted markers and attachment methods in estimating tibial rotations during walking: an in vivo study
Saber-Sheikh et al. Feasibility of using inertial sensors to assess human movement
Maynard et al. Intra-rater and inter-rater reliability of gait measurements with CODA mpx30 motion analysis system
Shorter et al. A new approach to detecting asymmetries in gait

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15865212

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase in:

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15865212

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase in:

Ref country code: JP