JP7420682B2 - Gait measurement system - Google Patents

Description

Article 30, Paragraph 2 of the Patent Act applies Event date: August 27, 2019 (Event period: August 27, 2019 to August 30, 2019) Meeting name and venue: Dynamics and Design Conference 2019 Kyushu University Ito Campus (744 Motooka, Nishi-ku, Fukuoka City, Fukuoka Prefecture)

The present invention relates to a gait measurement system.

Walking training is being carried out at welfare-related facilities as part of health promotion and care prevention projects associated with the aging of the population. A timed up and go (TUG) test is one of the exercise tests that can evaluate functional mobility, which includes walking ability, dynamic balance, agility, and the like. The TUG test is a test in which a subject (subject) stands up from a chair, turns around a marker 3 meters ahead, and sits back down on the chair.

In recent years, a gait measurement system described in Patent Document 1 has been developed as a system for tracking the leg position of a walking subject in such a gait test. The walking measurement system includes a distance information acquisition unit that emits a detection wave so as to scan along the horizontal direction and acquires distance information regarding the distance to an object over time based on the reflection state of the detection wave; and a data analysis unit that obtains a movement trajectory of the subject's legs based on at least the following. The data analysis unit detects observed values that are position candidates for the legs from the distance information, and correlates the observed values with the legs (correlation processing).

JP2016-137226A

In the above-mentioned technique, if an erroneous observed value is detected due to the influence of sensor noise, for example, there is a possibility that the erroneous observed value will be associated with the leg of the subject. Once you make a mistake in mapping in this way, the error will be carried over in subsequent tracking of leg positions, for example, the left and right leg positions may be tracked in the opposite direction (switching the left and right leg positions). There is a risk of incorrect tracking.

An object of the present invention is to provide a gait measurement system that can accurately track the position of a subject's legs.

The gait measurement system according to the present invention is a gait measurement system that tracks the leg position of a walking subject, and emits a detection wave so as to scan along the horizontal direction, and based on the reflection state of the detection wave. a distance information acquisition unit that acquires distance information over time regarding the distance to the object that reflected the detection wave; a data analysis unit that acquires a movement trajectory of the leg, the data analysis unit detects one or more observation values that are position candidates of the leg based on the distance information, and calculates the observation value for each of the plurality of particles. The predicted positions of both legs of the subject are randomly associated with either the value or the observation result of no observed value, and the weight is calculated from the transition of the walking state based on the predicted positions of the legs. Then, a particle with any correspondence is selected from among the plurality of particles based on the weight, and a movement trajectory is obtained based on the predicted positions of both legs of the selected particle.

In this walking measurement system, each predicted position of both legs of a subject is associated with an observed value for each of a plurality of particles (hereinafter also simply referred to as "association"). This allows for the possibility of multiple associations. Then, one of the associated particles is selected from among the plurality of particles based on the weight, the predicted positions of both legs are determined, and the movement trajectory is obtained. This weight is calculated from the transition of the walking state, which is a unique feature found in the legs of a walking subject. This makes it possible to obtain a movement trajectory based on any one of the possible associations that is plausible for the walking subject. Therefore, it is possible to prevent a movement trajectory from being acquired based on an erroneous association, and it is possible to accurately track the position of the subject's legs.

In the gait measurement system according to the present invention, the transition in the walking state of the subject based at least on the positions of both legs may be a transition in the walking phase. In this case, the weight of each particle can be calculated taking into account that the walking phase changes periodically.

In the gait measurement system according to the present invention, the transition of the subject's walking state based at least on the positions of both legs may be an observation pattern to which both legs correspond. In this case, the weight of each particle can be calculated by taking into account that the observed pattern of both legs changes periodically.

In the gait measurement system according to the present invention, the data analysis section includes an observation value detection section that detects an observation value from distance information using an observation pattern, and an observation value detection section that detects an observation value from distance information using an observation pattern. A predicted position calculation unit that calculates based on the positions of both legs of the particle, and a gate that spreads over a predetermined range based on the predicted position for each plurality of particles, and an observation that determines whether there are observed values or no observed values within the gate. It may also include a correlation processing unit that randomly associates predicted positions of both legs of the subject with any of the results. In this case, each particle can be associated using a gate set based on the predicted position.

In the gait measurement system according to the present invention, the data analysis section includes an update processing section that updates state quantities of the legs of the subject after each of the plurality of particles is associated with the correlation processing section; , and a weight calculation unit that calculates the weight from the transition of the walking state of the subject after the update by the update processing unit. In this case, in tracking the leg position using a particle filter, the weight of each particle can be specifically calculated.

In the gait measurement system according to the present invention, the predicted position calculation section calculates the predicted positions of both legs by predicting the positions of both legs using a Kalman filter, and the update processing section performs an update process of the Kalman filter. Good too. In this case, the position of the subject's legs can be tracked using a Kalman filter.

In the walking measurement system according to the present invention, the data analysis section may include a resampling section that resamples the plurality of particles so that they are selected and duplicated based on weights. In this case, the number of particles required can be reduced in tracking the leg position using the particle filter.

According to the present invention, it is possible to provide a gait measurement system that can accurately track the position of a subject's legs.

FIG. 1 is a block diagram showing the configuration of a gait measurement system according to an embodiment. FIG. 2 is a schematic diagram showing a TUG test to which the gait measurement system of FIG. 1 is applied. FIG. 3 is a plan view illustrating the laser range sensor of the gait measurement system of FIG. 1. FIG. 4A is a diagram illustrating an example of an observation pattern used for detecting observed values. FIG. 4(b) is a diagram illustrating an example of an observation pattern used for detecting observed values. FIG. 4(c) is a diagram illustrating an example of an observation pattern used for detecting observed values. FIG. 5A is a diagram illustrating an example of an observation pattern used for detecting observed values. FIG. 5(b) is a diagram illustrating an example of an observation pattern used for detecting observed values. FIG. 6A is a diagram illustrating an example of an observation pattern used for detecting observed values. FIG. 6(b) is a diagram illustrating an example of an observation pattern used for detecting observed values. FIG. 7A is a diagram illustrating an example of an observation pattern used for detecting observed values. FIG. 7(b) is a diagram illustrating an example of an observation pattern used for detecting observed values. FIG. 8 is a diagram illustrating a walking model. FIG. 9 is a diagram illustrating the predicted position of the leg. FIG. 10 is a diagram illustrating a gate set at a predicted position of a leg. FIG. 11 is a flowchart showing processing in the gait measurement system of FIG. FIG. 12A is a diagram illustrating association using a particle filter. FIG. 12(b) is a diagram illustrating association using a particle filter. FIG. 12(c) is a diagram illustrating association using a particle filter. FIG. 13(a) is a diagram illustrating association using a particle filter. FIG. 13(b) is a diagram illustrating association using a particle filter. FIG. 13(c) is a diagram illustrating association using a particle filter. FIG. 14(a) is a diagram showing the tracking results of the TUG test according to the comparative example. FIG. 14(b) is a diagram showing the tracking results of the TUG test according to the example.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings. In the following description, the same or equivalent elements will be denoted by the same reference numerals, and overlapping description will be omitted.

FIG. 1 is a block diagram showing the configuration of a gait measurement system according to an embodiment. FIG. 2 is a schematic diagram illustrating a TUG test to which the gait measurement system of FIG. 1 is applied. As shown in FIGS. 1 and 2, the gait measurement system 100 includes a chair 2, a marker 3, a laser range sensor (distance information acquisition unit) 10, a pressure sensor 15, and an electronic control device 20. , a monitor 30, and a pole 40. The gait measurement system 100 can be suitably applied to, for example, the TUG test, which is one of the gait tests that evaluate functional locomotor ability that comprehensively includes walking ability, dynamic balance, agility, and the like. The gait measurement system 100 is a system that tracks the leg position of a walking subject (subject) 1, and can perform quantitative gait measurement and walking ability evaluation.

In the TUG test, subject 1 stands up from sitting on chair 2, walks toward a marker 3 such as a cone a predetermined distance (for example, 3 m) ahead, turns at the marker 3, and then returns to chair 2. Take a seat again in chair 2. The TUG test is one of the Ministry of Health, Labor and Welfare's physical fitness measurement manuals. The TUG test requires walking at maximum walking speed without running. As the marker 3, various markers can be used as long as they serve as a mark for the turn position. In the TUG test, in order to align the marker 3 and the laser range sensor 10, a plurality of (here, two) poles 40 are used as reference members. The pole 40 has a cylindrical outer shape with a predetermined height. The predetermined height is a height at which the upper end of the pole 40 is located above the height position of the laser light emitted from the laser range sensor 10 (the height of the optical window portion of the laser range sensor 10). The shape of the pole 40 is not particularly limited, and various poles can be used. The gait measurement system 100 measures and evaluates the movement history and gait parameters of both legs of the subject 1 in such a TUG test.

By detecting the two poles 40 with the laser range sensor 10, the gait measurement system 100 can define a plane coordinate system having one axis as a straight line connecting the centers of the two poles 40. The floor surface on which the device is installed can be used as this planar coordinate system. In this case, if the positional relationship between the two poles 40 and the marker 3 is strictly adjusted (specifically, from the midpoint of the line segment connecting the two poles 40 to the vertical line of this line segment, Even if the installation position and/or orientation of the laser range sensor 10 is roughly aligned, it is possible to accurately grasp the relative positional relationship between the laser range sensor 10 and the marker 3. It becomes possible.

The gait measurement system 100 determines a leg observation position (hereinafter referred to as an "observation value") as an observation value that is a leg position candidate based on a characteristic pattern seen when both legs of the subject 1 are scanned by the laser range sensor 10. ” or “observation position”), tracking is performed using a Kalman filter, and walking parameters are calculated based on the acquired movement trajectory of both legs.

Specifically, in the gait measurement system 100, in order to reduce loss of sight and misidentification of the legs in situations where the legs of the subject 1 are close together and in situations where the legs cannot be observed from the laser range sensor 10 temporarily. Obtained values based on five types of observation patterns. Furthermore, using an effective region (gate) that takes into account the state of the leg, the observed values and predicted values are associated with each other. In the following, a "leg part" is also simply referred to as a "leg," a "standing leg part" is also simply referred to as a "standing leg," and a "swivel leg part" is also simply referred to as a "swivel leg." The "left leg" is also simply referred to as the "left leg," and the "right leg" is also simply referred to as the "right leg." "One leg" is also simply referred to as "one leg," and "both legs" is also simply referred to as "both legs."

An actual person's gait (speed of both legs) changes periodically. In particular, in the TUG test, since subject 1 is instructed to perform at the maximum walking speed without running, there is a large change in the speed of both legs. Therefore, in the gait measurement system 100, an acceleration/deceleration model that takes the gait phase into consideration is applied as the gait model. When humans walk, they perform periodic movements such as pivoting one leg (stance leg) and swinging the other leg (swing leg). "Walking phase" is a classification of periodic movements of both legs during a person's walking based on the state of swing between the stance leg and the swing leg (see, for example, FIG. 8).

Furthermore, in the gait measurement system 100, a particle filter is used to correlate observed values and predicted values, and a Kalman filter is used to update the state quantities of both legs for each hypothesis (correspondence). At this time, the weight as the likelihood of each particle is calculated, taking into consideration the transition of the walking phase and the transition of the observed pattern.

FIG. 3 is a plan view illustrating the laser range sensor 10 of the gait measurement system 100 of FIG. 1. The laser range sensor 10 is a sensor for acquiring the locus of movement of both legs. The laser range sensor 10 acquires two-dimensional plane distance information (hereinafter also simply referred to as "distance information"), which is the distance to an object around the sensor on a two-dimensional plane at a certain height. As shown in FIG. 3, the laser range sensor 10 emits a laser beam (detection wave) L so as to scan along the horizontal direction, and also reflects the laser beam L based on the reflection state of the laser beam L. Two-dimensional plane distance information regarding the distance to the legs F (right leg F _R and left leg F _L ) of the subject 1 is acquired over time.

Specifically, the laser range sensor 10 emits the laser beam L and reflects the laser beam L with a rotating mirror, thereby horizontally scanning the laser beam L in a fan shape in the measurement area. More specifically, the laser range sensor 10 scans the laser beam L in a fan shape in the horizontal direction so that the test subject 1 and the poles 40 installed on both sides of the test subject 1 are included in the measurement area. Then, for example, the reflected light of the laser beam L reflected by the leg F of the subject 1 is received, and the detection angle (scanning angle) of the reflected light and the time from emission to reception of the laser beam L (propagation time) are measured. Then, distance information including information regarding the angle and distance to the leg F is detected.

The laser range sensor 10 scans the laser beam L onto the subject 1 before or after the TUG test trial, for example, and obtains leg information regarding the width of the leg F of the subject 1 (length corresponding to the diameter of the shin). To detect. The laser range sensor 10 outputs detected distance information and leg information to the electronic control device 20.

The laser range sensor 10 is configured such that the height of the optical window portion that emits the laser beam L can be adjusted. The laser range sensor 10 is installed so that the laser beam L is emitted at a height position corresponding to the height of the shin of the subject 1 (that is, the height from the ankle to below the knee). The height of the optical window of the laser range sensor 10 is set higher than the height of the marker 3 so that the laser beam L is not reflected by the marker 3. The height of the optical window of the laser range sensor 10 is such that the leg F of the subject 1 can be observed even when the leg F leaves the bed during the swing phase, and the width of the leg F is at its maximum. Taking into consideration the average height, the height is, for example, 0.15 m to 0.27 m (here, 0.27 m) from the floor (ground).

The laser range sensor 10 is arranged in a space formed below the seat of the chair 2 so that the emission direction of the laser beam L is directed toward the marker 3 side. The laser range sensor 10 emits a laser beam L to the subject 1 from the rear side of the subject 1 approaching the marker 3 (moving away from the chair 2), that is, from the front side of the subject 1 moving away from the marker 3 (approaching the chair 2). irradiate. The laser range sensor 10 is installed so that the emission direction of the laser beam L is horizontal to the floor surface.

As the laser range sensor 10, a sensor with a distance measurement range of 0.1 to 30 m, a range measurement angle of 270 degrees, a distance measurement accuracy of ±30 mm, an angular resolution of 0.25 degrees, and a sampling period of 25 ms/scan is used. There is. Incidentally, the laser range sensor 10 is not limited in its type or specifications, and various types can be used depending on the measurement environment, for example. Note that the laser range sensor 10 may be surrounded by a housing in order to suppress the adverse effects of EMI (electromagnetic noise), for example.

The pressure sensor 15 is a sensor that detects an event in which the subject 1 stands up from the chair 2 and an event in which the subject 1 takes a seat on the chair 2. The pressure sensor 15 is provided so as to be placed on the seat surface of the chair 2. The pressure sensor 15 detects pressure information regarding the applied pressure and outputs the pressure information to the electronic control device 20. The pressure information of the pressure sensor 15 here is a voltage value, and the voltage value is A/D converted via a microcomputer board and input to the electronic control unit 20. The electronic control unit 20 associates and stores a stand-up event and its time, and associates and stores a seating event and its time.

The electronic control device 20 performs calculations based on the distance information acquired by the laser range sensor 10, specifies the position and speed of the leg F of the subject 1 over time, and acquires the walking characteristics of the subject 1. The electronic control device 20 outputs at least one of a sound (voice) and a display start signal (described later) to the subject 1. The start signal is output from the electronic control device 20 in response to the measurement operator's operation. The electronic control unit 20 stores the time of the start signal. Note that the start signal may be given by a device other than the electronic control device 20 or by a measurer. In this case, the time of the start signal is input to the electronic control device 20 by a device different from the electronic control device 20 or by a measurer.

As the electronic control device 20, for example, a personal computer or a dedicated control computer is used, and is configured to include a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. The electronic control device 20 includes a sensor data acquisition section 21, a storage section 22, a seating determination section 23, and a data analysis section 24 as functional elements.

The sensor data acquisition unit 21 acquires distance information detected by the laser range sensor 10 and pressure information detected by the pressure sensor 15 in synchronization. Further, the sensor data acquisition unit 21 may acquire the distance information in synchronization with the time of the start signal stored in the electronic control device 20. The sensor data acquisition unit 21 acquires leg information detected by the laser range sensor 10. The sensor data acquisition unit 21 outputs the acquired distance information, leg information, and pressure information to the storage unit 22. The sensor data acquisition section 21 outputs the acquired pressure information to the seating determination section 23 .

The storage unit 22 stores the distance information and pressure information acquired by the sensor data acquisition unit 21 in association with the processing time (sampling time). The storage unit 22 stores leg information of the subject 1 acquired by the sensor data acquisition unit 21. The seating determination unit 23 determines whether the subject 1 is seated based on the pressure information of the pressure sensor 15 acquired by the sensor data acquisition unit 21. For example, the seating determination unit 23 determines that the person is seated when the voltage value from the pressure sensor 15 rises to a value greater than a threshold value. The seating determination unit 23 determines that the person is not seated when the voltage value from the pressure sensor 15 is less than or equal to the threshold value. The seating determination unit 23 outputs the determination result to the storage unit 22, and the storage unit 22 stores the determination result.

The data analysis unit 24 acquires the walking parameters of the subject 1 by analyzing the stored data stored in the storage unit 22 after the TUG test trial. The data analysis unit 24 uses a particle filter to acquire the movement trajectory of the leg F of the subject 1 based on at least the distance information acquired by the laser range sensor 10. The data analysis section 24 includes a leg detection section (observed value detection section) 26, a leg tracking section 27, and a walking parameter calculation section 28. Gait parameters include general gait parameters and gait parameters related to the TUG test.

General gait parameters are gait parameters for evaluating fall risk in a general gait test, and include landing position, walking speed [m/sec], number of steps [steps], gait rate [steps/sec], It includes step length [m], stride length [m], step distance [m], and stance time (at least one of one-leg stance time [s] and double-leg stance time [s]). Gait parameters related to the TUG test are parameters for evaluating the risk of falling in the TUG test, and include reaction time (rise time) [seconds], TUG execution time [seconds], turn direction, and first step time (first step time). time) [seconds] and the shortest distance to marker 3 [m].

The gait parameters include the forward phase after starting the movement and before turning around marker 3, the turning phase while turning around marker 3, and the turning phase after turning around marker 3 until sitting down. It is classified and calculated as a return phase. In other words, gait parameters are calculated for each of these phases. At this time, the average, standard deviation, and coefficient of variation of the stride length, stride length, step interval, one-leg stance time, and double-leg stance time are calculated for each phase. Note that the gait parameters are not limited to being calculated by dividing them into the forward phase, turning phase, and return phase, but may be calculated as a whole (not divided), or in the first half and the second half (left and right). The calculation may be divided into two parts.

The landing position is the position where the legs F land on the floor. Walking rate is the number of steps per unit time. Single-leg stance time is the time from when the same leg touches the floor to when it leaves the floor. The double-leg stance time is the time from one leg landing on the floor to the other leg leaving the floor. The stride length is the distance from the landing position of one leg (for example, the heel position) to the landing position of the other leg. Stride length is the distance between adjacent landing positions for the same leg. The step distance is the distance between the heels of both legs in the frontal plane (when subject 1 is viewed from the front).

The reaction time is the time from the start time of the walking test (the time of the start signal) until the subject 1 stands up (the buttocks leave the seat of the chair 2). The TUG execution time is the execution time of the TUG test, and is the time from the start time to the end of the walking test (from when the subject 1 who is seated on the chair 2 begins to move until he is seated again). The first step time is the time from the start time of the walking test until the subject 1 raises the leg F (leaves the bed) for the first step. The shortest distance to the marker 3 is the shortest distance between the center of the marker 3 and the movement trajectory of the leg F.

In this embodiment, the data analysis unit 24 detects one or more observed values based on distance information acquired by the laser range sensor 10. The data analysis unit 24 randomly associates each of the predicted positions of both legs of the subject 1 with either the observed value or the observation result of no observed value for each of the plurality of particles, and Get the position of in relation to time. Along with this, the data analysis unit 24 calculates weights from the transition of the walking phase and the transition of the observation pattern to which both legs correspond. The data analysis unit 24 selects particles with any correspondence from among the plurality of particles based on the weight, and calculates the movement trajectory of the legs F of the subject 1 based on the predicted positions of both legs in the selected particles. get. The processing of the data analysis unit 24 will be described in detail below.

[Leg detection]
4 to 7 are diagrams illustrating examples of observation patterns used for detecting observed values. FIGS. 4A and 6A are diagrams illustrating examples of SL (Single Leg) patterns. FIG. 4(b) and FIG. 6(b) are diagrams illustrating an example of an LT (twoLegs Together) pattern. FIG. 4C and FIG. 7A are diagrams illustrating examples of FS_O (Forward Straddle Observable) patterns. FIGS. 5A and 7B are diagrams illustrating examples of FS_U (Forward Straddle Unobservable) patterns. FIG. 5B is a diagram illustrating an example of a UO (Unobservable) pattern.

The leg detection unit 26 uses a preset observation pattern to determine observed values (observed value vector y _k ^j :=(x _k ^j , y _k ^j ), (j=1...J)) is detected. Specifically, considering that the two-dimensional plane distance information of the leg F has a characteristic shape, the leg detection unit 26 performs edge detection processing on the two-dimensional plane distance information to detect the edge E of the leg F. Detect location. The leg detection unit 26 detects an observation value from the position of the edge E by pattern recognition using a plurality of observation patterns. At this time, the leg detection unit 26 uses the leg information regarding the width _wl of the leg F of the subject 1 to determine the width _wl of the leg F as known. Leg detection by the leg detection section 26 will be explained in detail below. Note that hereinafter, the "observed value vector" is also simply referred to as "observed value."

When the subject 1 is within the measurement range of the laser range sensor 10 and there is no obstacle between the laser range sensor 10 and the subject 1, the two-dimensional plane distance information acquired by the laser range sensor 10 is as shown in FIG. It has a characteristic shape as shown in a) to FIG. 5(b). The leg detection unit 26 calculates observed values using a leg detection method that takes into account five types of observation patterns.

The leg detection unit 26 first sets the i-th distance information from the right acquired by the laser range sensor 10 as l _i , and the two-dimensional plane distance information of the adjacent laser range sensor 10 is the width w of the leg F of the subject 1 _. On the other hand, if Equation 1 below is satisfied, an edge e ^h _m (h=B, F) is detected. At this time, if l _i >l _i+1 , edge e ^B _m =i and edge e ^F _m+1 = i+1. If l _i <l _i+1 , the edge e ^F _m =i and the edge e ^B _m+1 = i+1. The superscript B indicates the edge E on the back side with respect to the laser range sensor 10, and the superscript F indicates the edge E on the near side. The subscript m (m=1, . . . , M _k ) indicates the mth detected edge E. M _k is the total number of edges E detected at time k.

Then, based on the positional relationship of the four consecutive edges e ^h _n , e ^h _n+1 , e ^h _n+2 , e ^h _n+3 and the width w _e between the edges e ^h _n+1 and e ^h _n+2 , it is classified into five types of observation patterns. Then, the observed value y _k is calculated. However, where n=1, . . . , M _k -3, the subscript k indicates the observation position at time k.

As shown in FIGS. 4A and 6A, the observation pattern O1 as the SL pattern is an observation pattern when the leg F can be completely observed by the laser range sensor 10 alone. In observation pattern O1, the combination of edges E satisfies (e ^B _n , e ^F _n+1 , e ^F _n+2 , e ^B _n+3 ), and the width between edges E satisfies 0.5w _l <w _e ≦1.5w _l Fulfill. At this time, the observed value _yk is calculated as shown in FIG. 4(a), taking into consideration the width _wl of the leg F of the subject 1. For example, the observed value y _k ^j is determined from a line passing through (e ^F _n+1 +e ^F _n+2 )/2 and 1/2 of the width w _l of the leg F, as shown in the figure.

As shown in FIGS. 4(b) and 6(b), the observation pattern O2 as the LT pattern is an observation pattern when all the legs F can be completely observed from the laser range sensor 10. Observation pattern O2 has a combination of edges E such as (e ^B _n , e ^F _n+1 , e ^F _n+2 , e ^B _n+3 ), (e ^F _n , e ^B _n+1 , e ^F _n+2 , e ^B _n+3 ), or (e ^B _n , e ^F _n+1 , e ^B _n+2 , e ^F _n+3 ), and the width between the edges E satisfies 1.5 w _l <w _e <3.0 w _l . At this time, the observed value _yk is calculated as shown in FIG. 4(b), taking into account that both legs of subject 1 are aligned. For example, the observed value y _k ^j+1 is obtained from a line passing through (e ^F _n+1 +3e ^F _n+2 )/4 and 1/2 of the width w _l of the leg F, as shown in the figure. For example, the observed value y _k ^j is obtained from a line passing through (3e ^F _n+1 +e ^F _n+2 )/4 and 1/2 of the width w _l of the leg F, as shown in the figure.

As shown in FIGS. 4(c) and 7(a), the observed pattern O3 as an FS_O pattern has a step-like pattern due to the influence of one leg F, a cane, etc., and the width between the edge E positions is This is an observation pattern when the width w _l of the leg F of the subject 1 is 1/2 or more, and the center position of the leg F can be observed from the laser range sensor 10. Observation pattern O3 is such that the combination of edges E satisfies (e ^F _n , e ^B _n+1 , e ^F _n+2 , e ^B _n+3 ) or (e ^B _n , e ^F _n+1 , e ^B _n+2 , e ^F _n+3 ), and the edges The width between E satisfies 0.5 w _l ≦w _e <1.5 w _l . At this time, the observed value _yk is determined as shown in FIG. 4(c), taking into account that the center position of the leg F of the subject 1 can be observed from the laser range sensor 10. For example, the observed value y _k ^j is obtained from a line passing through (e ^B _n+1 +e ^F _n+2 )/2 and 1/2 of the width w _l of the leg F, as shown in the figure.

As shown in FIGS. 5(a) and 7(b), the observation pattern O4 as the FS_U pattern has a width between the edges E that is the width of the leg F of the subject 1 compared to the observation pattern O3 as the FS_O pattern. This is an observation pattern when w _l is smaller than 1/2 and the center position of the leg F cannot be directly observed from the laser range sensor 10. Observation pattern O3 is such that the combination of edges E satisfies (e ^F _n , e ^B _n+1 , e ^F _n+2 , e ^B _n+3 ) or (e ^B _n , e ^F _n+1 , e ^B _n+2 , e ^F _n+3 ), and the edges The width between E satisfies 0.3 w _l <w _e <0.5 w _l . At this time, unlike the observation patterns O1 to O3, the center position of the leg F cannot be directly observed by the laser range sensor 10, so the observed value _yk is virtually calculated as shown in FIG. 5(a). For example, the observed value y _k ^j is determined from a line passing through e ^B _n+1 , a line passing through e ^F _n+2 , and 1/2 of the width w _l of the leg F, as shown in the figure. The hollow squares in FIG. 7(b) indicate virtual edges.

As shown in FIG. 5(b), the observation pattern O5 as a UO pattern is observed when one leg F is hidden behind the other leg F or when the leg F is raised higher than the laser range sensor 10. , is an observation pattern when it cannot be observed.

Note that even in a situation where the leg F cannot be directly observed from the laser range sensor 10, the measurement accuracy of the position of the leg F can be improved by defining the observation pattern O4, which is the FS_U pattern, and virtually obtaining the observed value _yk . Improvement is expected. In detecting the observed value _yk , it can be determined that the position of the edge E that does not satisfy the observed patterns O1 to O4 is not the leg F. Furthermore, in order to avoid unnecessary correlation processing, a region wider by a predetermined length than the area of the walking path is set as the measurement region, and observed values _yk outside the measurement region and the position of the edge E corresponding thereto are excluded.

In detecting the observed value _yk , the observed value _yk is not detected at the center between the positions of the edge E, but for example, a position located inside the position of the edge E by 1/2 of the width _wl of the leg F. Ru. As described above, by detecting the observed value y _k of the leg F according to the plurality of observation patterns O1 to O5, even when one leg F is overlapped and hidden from the laser range sensor 10, the leg It becomes possible to suitably detect portion F.

[Leg tracking]
The leg tracking unit 27 performs prediction processing using a walking model that takes into account the walking phase, and estimates the position and speed of both legs. The prediction process uses a Kalman filter.

FIG. 8 is a diagram illustrating a walking model. When a person walks, he or she moves by using one leg as the pivot foot (stance leg) and swinging the other leg (swing leg). Both legs move periodically with acceleration and deceleration, changing between swing legs and stance legs upon landing on the floor. As shown in FIG. 8, the leg tracking unit 27 focuses on changes in the moving speed v and acceleration a of the leg F, and applies a walking model including the following six walking phases, including when standing still. .
・Phase 0: At rest, both legs are standing ・Phase 1: Left leg is swinging (acceleration, swinging motion), right leg is stance ・Phase 2: Left leg is swinging (deceleration, movement toward landing) , right leg is stance leg ・Phase 3: left leg is stance leg, right leg is swing leg (acceleration, swinging motion)
・Phase 4: Left leg is stance, right leg is swinging (deceleration, movement towards landing)
・Phase 5: Both legs are swinging (not seen when walking)

In addition, in the TUG test, subject 1 is instructed to perform at the maximum walking speed without running, so there is a large change in the speed of both legs. Therefore, a model that performs acceleration and deceleration is applied as a walking model. Additionally, when tracking a person's center of gravity, a nonlinear model that includes angle and angular velocity as state variables is often used. , it is desirable to apply a linear movement model that does not have directivity in the movement direction. Therefore, as a walking model, a linear movement model that has no directivity in the direction of movement is applied.

The leg tracking unit 27 sets a state equation for each of the left and right legs F using the position and velocity as a state quantity vector, and performs tracking processing for each leg F. The leg tracking unit 27 defines three motion models: stance leg, swing leg (acceleration), and swing leg (deceleration). A discrete time system equation for each leg F when such a walking model involving acceleration and deceleration is considered is given by Equation 2 below. The superscript f indicates left leg and right leg, where f=L and R, respectively.

The state quantity vector is shown in Equation 3 below. Equation 4 below represents the position and velocity of each leg F of the subject 1 at time k.

The vector u ^f,m _k is an unknown external input (acceleration input) given according to the walking phase, and is shown in Equation 5 below. The superscript m indicates the mth motion model. Model 1 is a stance leg (stationary) motion model, Model 2 is a swing leg (acceleration) motion model, and Model 3 is a swing leg (deceleration) motion model. It is given as 0 in kinematic model 1 during stance, a positive value in the moving direction in kinematic model 2 that accelerates with the free leg, and a negative value in the moving direction in kinematic model 3 that decelerates with the free leg. The magnitude of the acceleration input is calculated as the average value of the norm of the acceleration vector (see Equation 6 below) during the free leg for the past N _ac time steps.

In addition, since the movement of the leg F fluctuates including the direction of movement, the fluctuation from the motion model is treated as an acceleration disturbance and included in the system noise (see Equation 7 below), which is normal white noise with a mean of 0 and a covariance matrix Q ^f . shall be given. The state transition matrix A ^f , the input matrix B ^f _u , and the driving matrix B ^f of the state equation are each expressed by Equation 8 below.

Further, if the observed value y ^f _k of each leg F obtained using the laser range sensor 10 is given by the following equation 9, the observation equation is given by the following equation 10. Here, it is assumed that Δy ^f _k is normal white noise with a mean of 0 and a covariance R ^f . At this time, the matrix C ^f is expressed by Equation 11 below.

As shown in FIG. 1, the leg tracking unit 27 includes a stance leg swing determination unit 27a, a walking phase determination unit 27b, a predicted position calculation unit 27c, a correlation processing unit 27d, an update processing unit 27e, a weight calculation unit 27f, and It has a resampling section 27g.

The stance leg swing determining unit 27a performs stance leg swing determination to determine whether the leg F is a stance leg (support leg) or a swing leg, based on the speed of the leg F. The stance leg swing determining unit 27a determines that the right leg is a stance leg when the following formula 12 is satisfied, and determines that the right leg is a swing leg when the following formula 13 is satisfied. v _{st_th} indicates the velocity threshold of the stance leg, and v _{sw_th} indicates the velocity threshold of the swing leg. The stance leg swing determination unit 27a similarly performs stance leg swing determination for the left leg, that is, using the following equations 12 and 13, in which the right leg and the left leg are exchanged.

The walking phase determination unit 27b determines the walking phase based on the relative positional relationship and speed of both legs, and specifies which of the phases 0 to 5 the walking phase is. Specifically, first, when both legs are standing, phase 0 is specified. Next, when the left leg is a swing leg and the right leg is a stance leg, when the inner product of the position vector of the left leg with respect to the right leg and the velocity vector of the left leg (see Equation 14 below) takes a positive value, It is specified as phase 1, and when it takes a negative value, it is specified as phase 2. In addition, when the left leg is a stance leg and the right leg is a swing leg, similarly, if the inner product of the position vector of the right leg with respect to the left leg and the velocity vector of the right leg takes a positive value, it is identified as phase 3. However, when it takes a negative value, it is specified as phase 4. When both legs are free legs, it is specified as phase 5. When the walking phase determination unit 27b identifies the walking phase as phase 5, the data analysis unit 24 determines that the subject 1 is running.

Here, the leg tracking unit 27 of the present embodiment uses a Rao-Blackwellized particle filter (hereinafter also simply referred to as a "particle filter") to calculate multiple time points in consideration of the transition of the walking phase and the observation pattern of the leg F. We employ a two-leg tracking method based on a hypothesis of correspondence between the two legs. In tracking multiple objects, it is necessary to estimate the correspondence between the tracked object and the observed value y _k obtained by the sensor at time _k (c _k indicates the index of the correspondence) and the state quantity x _k of the tracked object. There is. Although the correspondence and the state of the tracking target can be obtained using only a particle filter, it is necessary to generate a large number of particles regarding the state amount for each combination of correspondences, which requires a large amount of calculation. Therefore, in the particle filter, the correspondence with the state quantity of the tracking target is determined using Equation 15 below. Note that, for example, the number of particles may be 100.

In the particle filter, when a certain observation is obtained in p(c _1:k |vector y _1:k ), only the correspondence is calculated by the particle filter. Next, in p (vector x _k |c _1:k , vector y _1:k ), the state quantity is updated using a Kalman filter according to the associated observed value y _k . Since the state quantity update process is performed using a Kalman filter, there is no need to generate a large number of particles for each association combination. The leg tracking unit 27 calculates a weight as the likelihood of each particle based on the walking phase and the state transition probability of the observed pattern, thereby improving the tracking performance of both legs during walking including turning movements of the subject 1, such as an elderly person. do.

The predicted position calculation unit 27c calculates the predicted positions of both legs by predicting the positions of both legs using a Kalman filter. The predicted position calculation unit 27c calculates a predicted leg position (hereinafter also simply "predicted position"), which is the predicted position of each leg F at time k, for each of the plurality of particles based on the positions of both legs at past times. (see "×" in FIG. 9). The superscript (n), (n=1, . . . , N) indicates the n-th particle. The predicted position vector y^ ^{f, (n)} _k/k-1 of each leg F is given by the following equation 16 based on the state equation shown in equation (2) above. The vector x^ ^{f, (n)} _k/k-1 indicates the prior state estimate of the n-th particle at time k. The vector x^ ^{f, (n)} _k-1/k-1 is the estimated posterior state of the n-th particle at time k-1. In FIG. 9, "◇" indicates the observed value _yk , the line indicates the movement locus of leg F, and "□" and "■" on the line indicate the positions of the left leg and right leg. (The same applies to subsequent figures). Note that hereinafter, the "predicted position vector" is also simply referred to as "predicted position."

The correlation processing unit 27d randomly correlates the predicted positions of both legs with the observed values _yk for each particle. However, in order to limit the observed values _yk to be correlated, the correlation processing unit 27d sets a gate GA (see FIG. 10) at the predicted position, ``observed value y <sub>k</sub> '' and ``(ii) Observation result with no observed value y ^f _k (when there is no observed value y _k )'' considering the case where observed value y _k includes false detection. The predicted positions of both legs are randomly associated with each leg. Whether or not the observed value y _k is included in the gate GA is determined using Equation 17 below. The gate GA is an effective area that extends over a predetermined range with the predicted position (see "X" in FIG. 10) as a reference (center). The area of the gate GA may be changed. For example, since the position of the swing leg changes more easily than that of the stance leg, the gate GA for the swing leg may be set to be wider than the gate GA for the stance leg.

d ^{(n) 2} _f,j is the Mahalanobis distance, and is given by Equation 18 below. S ^{f, (n)} _k is a covariance matrix of the observed prediction error (y ^j _k −y^ ^{f, (n)} _k/k−1 ). G is a gate, which is determined based on a chi-square (χ ² ) distribution with two degrees of freedom in this embodiment where the observed value y _k is a two-dimensional vector and uses a Kalman filter.

The update processing unit 27e updates the state quantity of the leg F after the correlation processing unit 27d associates each particle with each other. Specifically, the update processing unit 27e uses the observed values _yk that have been correlated by the correlation processing unit 27d for each of the plurality of particles to update the Kalman filter (for example, update the state quantities such as the covariance matrix). update). However, if there is no observed value _yk that has been correlated (for example, in the case of observation pattern O5 as a UO pattern), the update processing unit 27e replaces the prior state estimate and the prior covariance matrix with the posterior state estimate and the prior covariance matrix, respectively. Let it be the posterior covariance matrix. As a result, the positions and speeds of both legs are estimated for each of the plurality of particles. Furthermore, the update processing unit 27e determines the walking phase of each of the plurality of particles by the walking phase determining unit 27b based on the information on the estimated positions and speeds of both legs.

The weight calculation unit 27f calculates a weight w ⁽ⁿ⁾ _k for each of the plurality of particles from the transition of the walking state of the subject 1 after the update processing unit 27e updates. Specifically, the weight calculation unit 27f calculates the weight w ⁽ⁿ⁾ _k of each particle based on the walking phase of each leg from one time ago (phases 0 to 5) and the observation patterns of leg F (observation patterns O1 to O1). From the transition probability of O5), it is calculated (updated) as shown in Equation 19 below.

The second item in Equation 19 above is calculated as the product of the likelihood of the observed value y _k of the Kalman filter of each leg F and the state transition probability of the observed patterns O1 to O5. In this embodiment, the transition probability matrix from observation patterns O1 to O5 one time ago is defined as shown in Equation 20 below. The horizontal arrangement of Equation 20 below corresponds to each of observation patterns O1 to O5 (SL pattern, LT pattern, FS_O pattern, FS_U pattern, and UO pattern) before transition (one time ago) from the left. The values in the vertical column of Equation 20 below indicate the probability of transitioning from the observation pattern before transition to observation patterns O1 to O5 in order from the top. As expressed by Equation 20 below, the transition of observation patterns O1 to O5 for calculating the weight w ⁽ⁿ⁾ _k has the following characteristics. In the case of observation pattern O1, in which the leg is observed alone one time ago, the probability that the observation pattern O1 will be observed at the next time is high. It is set so that there is no transition to observation pattern O4, which is hidden. In the case of observation pattern O2, which is a state in which both legs are aligned one time ago, the probability of being in observation pattern O2 at the next time and the probability of transitioning to observation pattern O1, which is a state in which the legs are observed alone, are high. It is designed to be. In the case of observation pattern O3, in which both legs are observed in a stepped manner one time before, the probability that the next time will also be observation pattern O3, and the observation pattern O1 in which the legs are observed alone or the other leg. The design is such that the probability of transition to observation pattern O4, which is a state in which the observation pattern O4 is almost hidden, is high. If the observation pattern O4 is almost hidden behind the other leg one time ago, the probability that the same observation pattern O4 will be the same at the next time, and the observation pattern O3 that will be possible to observe after coming off the back of the other leg. Alternatively, the probability of the observation pattern O5 being completely hidden by one leg is set to be high. In the case of observation pattern O5 in which the leg was not observed one time ago, there is a high probability that it will be observation pattern O5 at the next time as well, and it is set so that it will transition to other observation patterns with almost the same probability. .

The third item in Equation 19 above relates to the state transition probability of the walking phase. In this embodiment, the transition probability matrix from phases 0 to 5 one time ago is defined as shown in Equation 21 below. The horizontal arrangement of Equation 21 below corresponds to phases 0 to 5 before the transition (one time before) from the left. The values in the vertical column of Equation 21 below indicate the probability of transition from the pre-transition phase to the post-transition phase 0 to 5, starting from the top. As expressed by Equation 21 below, the transition of phases 0 to 5 for calculating the weight w ⁽ⁿ⁾ _k has the following characteristics. If the phase is 0, which is a stationary state, one time ago, the probability of being in phase 0 at the next time and the probability of transitioning to phases 1 and 3 when starting to walk with either the left or right leg are set to be high. ing. If the left leg is in phase 1 and is accelerating with an idle leg one time ago, the probability of being in phase 1 at the next time and the probability of transitioning to phase 2 in which the left leg is decelerating are set to be high. If the left leg is in phase 2 and is decelerating in swing one time ago, the probability that the right leg will be in phase 2 at the next time as well as phase 3 where the right leg will start accelerating in swing or phase 0 where it will remain stationary. The probability of transition is set to be high. Phases 3 and 4, in which the right leg is free, are designed with the same transitions as phases 1 and 2 in mind. Furthermore, in the case of phases 0 to 4, the probability of transitioning to phase 5, which does not appear during walking, at the next time is designed to be extremely small. Furthermore, if the phase is 5 one time ago, it is designed to transition to phases 0 to 4 with equal probability at the next time.

The resampling unit 27g resamples the plurality of particles so that they are selected and duplicated based on the weight w ⁽ⁿ⁾ _k . For example, the resampling unit 27g resamples a plurality of particles such that the smaller the weight w ⁽ⁿ⁾ _k is, the more particles are culled, and the larger the weight w ⁽ⁿ⁾ _k is, the more the particles are duplicated. Specifically, the resampling unit 27g normalizes the weight w ⁽ⁿ⁾ _k of each particle and calculates the effective sample size using Equation 22 below. The resampling unit 27g performs resampling when N ^eff _k is smaller than the threshold value. For example, the resampling unit 27g performs resampling at equal intervals. Note that the specific resampling method performed by the resampling unit 27g is not particularly limited, and various known resampling methods may be used.

The leg tracking unit 27 estimates the positions and velocities of both legs of the plurality of particles at all times, and after the final weight w ⁽ⁿ⁾ _k has been updated and resampled, the leg tracking unit 27 calculates the Particles with any correspondence are selected from among them based on the weight w ⁽ⁿ⁾ _k . Here, the leg tracking unit 27 selects one of the plurality of particles having the largest weight w ⁽ⁿ⁾ _k . The leg tracking unit 27 acquires the movement trajectory of the leg F of the subject 1 based on the position and velocity of both legs in the selected particle.

[Calculation of walking parameters]
The walking parameter calculating unit 28 calculates walking parameters from the acquired movement trajectories of both legs. The gait parameter calculation section 28 includes a stance time calculation section 28a, a landing position calculation section 28b, a step length, step distance and stride length calculation section 28c, a first step time calculation section 28d, a reaction time calculation section 28e, and a TUG. It has an execution time calculation section 28f.

The stance time calculation unit 28a calculates the one-leg stance time and the double-leg stance time. In this embodiment, in order to determine the one-leg stance time and the double-leg stance time, the time of contact with the bed and the time of leaving the bed are determined based on the analysis results of a test subject in which pressure sensors were attached to the heels and metatarsals of both legs. is calculated strictly after acquiring the movement trajectory.

The landing position calculation unit 28b first calculates the landing position in order to calculate the stride length, stride length, and step distance. Since the laser range sensor 10 acquires the movement locus of the leg F at the height of the shin, it is difficult for the landing position calculation unit 28b to directly measure the position of the leg F. Therefore, the landing position calculation unit 28b takes into consideration that the leg F becomes perpendicular to the floor surface at the time when the speed of the leg F becomes the minimum during stance, and calculates the position of the leg F at that time. Calculate the position as the landing position.

The step length, step distance, and stride length calculation unit 28c calculates the step length, step distance, and stride length based on the calculated landing position. The first step time calculation unit 28d calculates the time difference from the time of the start signal to the time when the leg F leaves the floor for the first step, as the first step time. The reaction time calculation unit 28e calculates the time when the voltage value from the pressure sensor 15 installed on the chair 2 changes from the sitting state (the time when it falls below the threshold value) as the reaction time. The reaction time calculation unit 28e calculates the time difference from the time of the start signal to the reaction time as the reaction time.

The TUG execution time calculation unit 28f takes advantage of the fact that the voltage value of the pressure sensor 15 changes when the subject 1 sits on the chair 2 again, and sets the rise time of the voltage value as the end time. That is, based on the pressure information detected by the pressure sensor 15, the seating of the subject 1 on the chair 2 is detected. The TUG performance time calculation unit 28f calculates the time difference from the start signal to the end time as the test performance time of the TUG test.

Furthermore, the walking parameter calculation unit 28 calculates the shortest distance to the marker 3 based on the movement trajectory of both legs and the center position of the marker 3. The walking parameter calculation unit 28 calculates other general walking parameters from the acquired movement trajectory of both legs.

The monitor 30 is an output unit that outputs the analysis results analyzed by the data analysis unit 24. The monitor 30 displays the analysis results analyzed by the data analysis unit 24 and provides feedback to the subject 1. The monitor 30 can also output distance information, leg information, and pressure information acquired by the sensor data acquisition unit 21 or stored in the storage unit 22. Note that the monitor 30 may be configured to output audio or the like instead of or in addition to the display. Further, instead of or in addition to the monitor 30, a printer may be provided to print out calculation results and the like on a paper medium. The monitor 30 may be provided integrally with the electronic control device 20, and may be, for example, a display section of a personal computer, a dedicated control computer, or the like.

[Gait measurement by gait measurement system 100]
Next, the case where the gait characteristics in the TUG test are measured using the gait measurement system 100 of this embodiment will be described with reference to the flowchart shown in FIG. 11.

As shown in Figure 11, the measurement process in the TUG test consists of a calibration phase (calibration before the trial), a measurement phase (measurement during the TUG trial), and an analysis phase (data analysis after the trial). It can be divided into three phases: In the calibration phase, the two poles 40 are used to align the laser range sensor 10 and measure the width _wl of the leg F of the subject 1. After the calibration is completed, a start signal (instruction to start the TUG test) is notified to the subject 1 by sound, for example, from the electronic control device 20, and the process moves to the measurement phase. In the measurement phase, distance information from the laser range sensor 10 and pressure information from the pressure sensor 15 are acquired and stored in synchronization. Then, after the seating of subject 1 is detected from the pressure information, the process moves to the analysis phase. In the analysis phase, the data analysis unit 24 tracks the positions of both legs based on the stored data to obtain a movement trajectory, and calculates walking parameters based on the obtained movement trajectory. This will be explained in detail below.

First, before subject 1 starts a movement, calibration is performed. In the calibration, a scan is performed by the laser range sensor 10. That is, the laser beam L is emitted from the laser range sensor 10 so as to scan along the horizontal direction, and distance information is acquired based on the reflection state of the laser beam L (step S1). Thereby, the positions of the leg F of the subject 1 and the two poles 40 are observed.

The width (leg information) w _l of the leg F of the subject 1 is detected from the observed width between the edge positions of the leg F (step S2). The detected leg information is stored in the storage unit 22. By determining the positional relationship between the two poles 40 with the laser range sensor 10, a measurer etc. can confirm whether the laser range sensor 10 is installed at an appropriate position and direction with respect to the marker 3, and if necessary, Then, the positioning of the laser range sensor 10 is performed (step S3). Note that, as described above, as long as the relative positional relationship between the two poles 40 and the marker 3 is made strictly, the installation accuracy of the laser range sensor 10 may be rough.

The series of processes in steps S1 to S3 described above are repeated until the calibration is completed (step S4). After several seconds have elapsed since the end of the calibration, a start signal is automatically (or manually by the measurer) notified to the subject 1, and the TUG test is started (step S5).

During the trial of the TUG test, the laser range sensor 10 performs a scan, and the sensor data acquisition unit 21 acquires distance information regarding the distance of the leg F of the subject 1 over time (step S6). The sensor data acquisition unit 21 acquires pressure information from the pressure sensor 15 (step S7). The seating determination unit 23 determines whether the subject 1 has stood up based on the pressure information (step S8). If it is determined that the subject 1 is not standing up, the process returns to step S6. If it is determined that the subject 1 is standing up, the seating determination unit 23 determines whether the subject 1 is seated based on the pressure information (step S9). If it is determined that the person is not seated, the TUG test is still in progress, and the process returns to step S6. If it is determined that subject 1 is seated, it is assumed that the TUG test has ended, and the following data analysis is performed.

In the data analysis, a leg position tracking process is performed and a plurality of particles in a particle filter are initialized (step S10). At time k, the leg detection unit 26 detects the observed value _yk based on the stored distance information (step S11). At time k, a predicted position is calculated for each particle by the predicted position calculation unit 27c (step S12). At time k, the correlation processing unit 27d randomly correlates each particle (step S13). At time k, the state quantity is updated for each particle by the update processing unit 27e (step S14). As a result, the positions and speeds of both legs at time k are estimated. At time k, the weight calculation unit 27f calculates a weight w ⁽ⁿ⁾ _k for each particle (step S15). At time k, resampling is performed by the resampling unit 27g (step S16).

Subsequently, a data end determination is performed to determine whether each process related to the data analysis has been completed for all of the stored data (distance information stored in association with time) stored in the storage unit 22 (step S17). If No in step S17, the process proceeds to step S11 at the next time. On the other hand, if the answer in step S17 is Yes, it is determined that the leg position tracking process has ended.

Thereafter, the leg tracking unit 27 selects one particle having the largest weight w ⁽ⁿ⁾ _k from among the plurality of particles. The leg tracking unit 27 acquires the movement locus of both legs of the subject 1 in the TUG test based on the position and velocity of both legs in the selected particle. Then, the walking parameter calculation unit 28 calculates the walking parameters from the acquired movement trajectory of both legs (step S18). The analysis results and/or walking parameters of the saved data analysis are output to the monitor 30 (step S19).

As described above, according to the gait measurement system 100, each predicted position of both legs of the subject 1 is associated with an observed value for each of a plurality of particles. This allows for the possibility of multiple associations. Then, particles with any correspondence are selected from among the plurality of particles based on the weight w ⁽ⁿ⁾ _k , the predicted positions of both legs are determined, and the movement trajectory is obtained. This weight w ⁽ⁿ⁾ _k is calculated from the unique features found in the legs F of the walking subject 1, that is, the walking phase and the transition of the observed pattern. Therefore, it is possible to obtain a movement trajectory based on one of the plurality of association possibilities that is plausible for the subject 1. Therefore, it is possible to prevent a movement trajectory from being acquired based on an incorrect association, and it is possible to accurately track the leg position of the subject 1. This makes it possible to reduce erroneous tracking.

The gait measurement system 100 calculates the weight w ⁽ⁿ⁾ _k based on the transition of the gait phase. In this case, the weight w ⁽ⁿ⁾ _k of each particle can be calculated taking into account that the walking phase changes periodically.

The gait measurement system 100 calculates the weight w ⁽ⁿ⁾ _k based on the transition of the observation pattern that applies to both legs. In this case, the weight w ⁽ⁿ⁾ _k of each particle can be calculated by taking into account that the observed pattern of both legs changes periodically.

In the gait measurement system 100, the data analysis section 24 includes a leg detection section 26, a predicted position calculation section 27c, and a correlation processing section 27d. The leg detection unit 26 detects the observed value _yk from the distance information. The predicted position calculation unit 27c calculates the predicted positions of both legs for each of the plurality of particles based on the positions of both legs at past times. The correlation processing unit 27d sets a gate GA for each of the plurality of particles based on the predicted position, and calculates the value for either the observed value y _k existing within the gate GA or the observed value y _k absent. Then, the predicted positions of both legs of subject 1 are randomly associated with each other. In this case, it is possible to associate each particle using a gate GA set based on the predicted position.

In the gait measurement system 100, the data analysis section 24 includes an update processing section 27e and a weight calculation section 27f. The update processing unit 27e updates the state quantity of the leg F after the correlation processing unit 27d associates each particle with each other. The weight calculation unit 27f calculates a weight w ⁽ⁿ⁾ _k for each plurality of particles from the transition of the walking phase and the observed pattern after the update processing unit 27e updates the particles. In this case, in tracking the leg position using a particle filter, the weight w ⁽ⁿ⁾ _k of each particle can be specifically calculated.

In the gait measurement system 100, the predicted position calculation unit 27c calculates the predicted positions of both legs by predicting the positions of both legs using a Kalman filter. The update processing unit 27e performs Kalman filter update processing. In this case, the leg position of the subject 1 can be accurately tracked using the Kalman filter.

In the gait measurement system 100, the data analysis section 24 includes a resampling section 27g. The resampling unit 27g resamples the plurality of particles so that they are selected and duplicated based on the weight w ⁽ⁿ⁾ _k . In this case, the number of particles required can be reduced in tracking the leg position using the particle filter. It is possible to maintain the association hypothesis and estimate the state amount with a small number of particles.

FIGS. 12 and 13 are diagrams illustrating association using a particle filter. FIG. 12(a) shows the result of the association for the first particle PT1 at time k, FIG. 12(b) shows the result of association for the second particle PT2 at time k, and FIG. 12(c) shows the result of association for the second particle PT2 at time k. The results of the association for the third particle PT3 are shown. FIG. 13(a) shows the result of the association in the first particle PT1 at time k+1, FIG. 13(b) shows the result of association in the second particle PT2 at time k+1, and FIG. 13(c) shows the result of association in the second particle PT2 at time k+1. The results of the association for the third particle PT3 are shown.

In the examples of FIGS. 12(a) to 12(c), erroneously detected observed values y _k ¹ and y _k ² are detected. In the first particle P1, the predicted position of the left leg is associated with the observed value y _k ¹ , and the predicted position of the right leg is associated with the observed value y _k ² . In the second particle P2, the predicted position of the left leg is associated with the observed value y _k ² , and the predicted position of the right leg is associated with the observed value y _k ¹ . In the third particle P3, the predicted position of the left leg is associated with the observation result of no observed value, and the predicted position of the right leg is associated with the observed value y _k ² . For the third particle, since there is no observation value associated with the predicted position of the left leg, the predicted position of the left leg is acquired as the position of the left leg. In such an example, weight w ⁽¹⁾ _k > weight w ⁽²⁾ _k > weight w ⁽³⁾ _k .

In the examples shown in FIGS. 13(a) to 13(c), observed value y _k+1 ¹ and observed value y _k+1 ² are detected. In the first particle P1, the predicted position of the left leg is associated with the observed value y _k+1 ² , and the predicted position of the right leg is associated with the observed value y _k+1 ¹ . In the second particle P2, the predicted position of the left leg is associated with the observed value y _k+1 ² , and the predicted position of the right leg is associated with the observed value y _k+1 ¹ . In the third particle P3, the predicted position of the left leg is associated with the observed value y _k+1 ¹ , and the predicted position of the right leg is associated with the observed value y _k+1 ² . In such an example, weight w ⁽³⁾ _k+1 >> weight w ⁽²⁾ _k+1 > weight w ⁽¹⁾ _k+1 . As a result, by resampling, for example, the first particles PT1 are eliminated, and the third particles PT3 are duplicated.

FIG. 14(a) is a diagram showing the tracking results of the TUG test according to the comparative example. FIG. 14(b) is a diagram showing the tracking results of the TUG test according to the example. The tracking result according to the comparative example is a tracking result obtained by a conventional walking measurement system. The tracking results according to the embodiment are the tracking results obtained by the gait measurement system 100. As shown in FIG. 14(a), in the tracking results related to the comparative example, it can be confirmed that the left and right legs F are swapped at the start of the movement of turning the marker 3 (in the broken line circle frame in the figure). (see inside). On the other hand, as shown in FIG. 14(b), in the tracking results according to the example, both legs are It can be confirmed that tracking can be continued without being replaced.

The gait measurement system 100 uses a small laser range sensor 10 and a pressure sensor 15 that can acquire distance information on a two-dimensional plane to track the positions of both legs and improve measurement accuracy with less misidentification during TUG test measurements. can be realized. In the gait measurement system 100, the data analysis unit 24 performs analysis (post-analysis) after the gait test is completed, rather than performing analysis in real time during the gait test trial. This makes it possible to accurately calculate the movement trajectory and walking parameters. Note that in the gait measurement system 100, it is of course possible for the data analysis unit 24 to analyze the gait characteristics while attempting the gait test.

[Modified example]
Although one aspect of the present invention has been described above, one aspect of the present invention is not limited to the above embodiments, and may be modified or applied to other things without changing the gist of each claim. .

In the above embodiment, the weight w ⁽ⁿ⁾ _k is calculated from the transition of the walking phase of the subject 1 and the transition of the observation pattern, but the weight w ⁽ⁿ⁾ _k may be calculated from only one of them. In the above embodiment, the above formula 19 for calculating the weight w ⁽ⁿ⁾ _k is expressed as the product of the second item related to the transition of the observed pattern and the third item related to the transition of the walking phase, but the second item It may be expressed as the sum or difference between the item and the third item. In the embodiment described above, the transition of the walking state is not limited to the transition of the walking phase and the transition of the observed pattern, but may be a transition of the walking state based on the positions of other legs.

In the embodiment described above, a plurality of particles are resampled by the resampling unit 27g, but depending on the case, the resampling unit 27g may not be necessary. In the above embodiment, the movement locus of the leg F is obtained from the position and velocity of both legs of the particle with the largest weight w ⁽ⁿ⁾ _k , but the present invention is not limited thereto. For example, the movement trajectory may be acquired from any particle whose weight w ⁽ⁿ⁾ _k is a predetermined value or more. In short, it is sufficient to select a particle with any correspondence based on the weight w ⁽ⁿ⁾ _k and obtain a movement trajectory from the selected particle.

In the above embodiment, the gait measurement system 100 is applied to the measurement of gait characteristics in the TUG test, but the invention is not limited to this, and for example, gait characteristics in MTST, Rhythmic Stepping Exercise (RSE), and other various gait trainings. The gait measurement system 100 can be applied to the measurement of .

In the above embodiment, the laser range sensor 10 is used as the distance information acquisition unit, but various sensors and devices can be used as long as they can acquire distance information at a position at a predetermined height from the floor surface. For example, an infrared sensor or the like capable of acquiring the above-mentioned two-dimensional plane distance information may be used as the distance information acquisition section. Although the above embodiment includes one laser range sensor 10, it may include a plurality of laser range sensors. In the embodiment described above, the pole 40 is used as the reference member, but the present invention is not limited thereto. As the reference member, various members may be used as long as they can align the laser range sensor 10 with respect to the marker 3.

In the above embodiment, the laser range sensor 10 is arranged so that the laser beam L is irradiated from the rear side of the subject 1 approaching the marker 3 (away from the chair 2), but the laser range sensor 10 is irradiated from the front side of the subject 1 approaching the marker 3. The laser range sensor 10 may be arranged so as to be irradiated with the light L. In some cases, the laser range sensor 10 may be arranged so that the laser beam L is irradiated from the side of the subject 1 approaching the marker 3. In the above embodiment, the laser range sensor 10 is placed at the bottom of the chair 2, but the placement position of the laser range sensor 10 is not limited, and the laser range sensor 10 may be placed according to the irradiation mode of the laser beam L to the subject 1. Bye.

In the embodiment described above, the standing up (movement) of the subject 1 from the chair 2 and the sitting on the chair 2 are determined based on the pressure information of the pressure sensor 15. , based on the detection results of the laser range sensor 10, the rising and seating may be determined. In the embodiment described above, standing up and sitting down of the subject 1 was detected based on the pressure information, but only standing up may be detected based on the pressure information, or only sitting down may be detected based on the pressure information.

Note that the present invention can be understood as a gait measurement method implemented by the gait measurement system 100 described above. The present invention can be understood as a gait measurement program that causes a computer to execute the gait measurement method (each process of the gait measurement system 100 described above).

1... Subject (target person), 10... Laser range sensor (distance information acquisition section), 24... Data analysis section, 26... Leg detection section (observed value detection section), 27c... Predicted position calculation section, 27d... Correlation processing section , 27e...Update processing unit, 27f...Weight calculation unit, 27g...Resampling unit, 100...Walking measurement system, F...Leg, L...Laser light (detection wave).

Claims (7)

A gait measurement system that tracks the leg position of a walking subject,
a distance information acquisition unit that emits a detection wave so as to scan along a horizontal direction, and acquires distance information over time regarding a distance to an object that reflected the detection wave based on a reflection state of the detection wave;
a data analysis unit that uses a particle filter to acquire a movement trajectory of the leg of the subject based on at least the distance information acquired by the distance information acquisition unit;
The data analysis section includes:
Detecting one or more observed values that are position candidates for the leg based on the distance information,
For each of a plurality of particles, each of the predicted positions of both legs of the subject is randomly associated with either the observed value or the observation result of no observed value, and the predicted position of both legs of the subject is randomly associated. Calculate the weight from the transition of the walking state based on the position,
A gait measurement system that selects particles with any correspondence from among the plurality of particles based on the weights, and obtains the movement trajectory based on predicted positions of the legs in the selected particles. The gait measurement system according to claim 1, wherein the transition of the walking state of the subject based at least on the positions of both legs is a transition of a walking phase. The gait measurement system according to claim 1 or 2, wherein the transition of the walking state of the subject based at least on the positions of the legs is a transition of an observation pattern to which the legs correspond. The data analysis section includes:
an observed value detection unit that detects the observed value from the distance information using a preset observation pattern;
a predicted position calculation unit that calculates a predicted position of both the legs for each of the plurality of particles based on the positions of the legs at a past time;
For each of the plurality of particles, a gate that extends over a predetermined range based on the predicted position is set, and for either the observed value existing within the gate or the observation result in which there is no observed value, the The gait measurement system according to any one of claims 1 to 3, further comprising a correlation processing unit that randomly associates predicted positions of both legs of the subject. The data analysis section includes:
an update processing unit that updates the state quantity of the leg of the subject after being associated by the correlation processing unit for each of the plurality of particles;
The gait measurement system according to claim 4, further comprising: a weight calculation unit that calculates the weight based on a transition in the walking state of the subject after the update processing unit updates each of the plurality of particles. The predicted position calculation calculates the predicted position of both legs by predicting the positions of both legs using a Kalman filter,
The gait measurement system according to claim 5, wherein the update processing unit performs update processing of the Kalman filter. 7. The gait measurement system according to claim 1, wherein the data analysis section includes a resampling section that resamples the plurality of particles so that they are selected and duplicated based on the weights.

Images