JP2014176461A - Walking measurement system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a walking measurement system capable of tracing leg positions with high accuracy.SOLUTION: A walking measurement system 100 includes: a laser range sensor 10 acquiring two-dimensional plane distance information; and an electronic controller 20 specifying leg positions of a subject with time on the basis of the two-dimensional plane distance information. The electronic controller 20 has: an observation position detection part 21 detecting leg observation positions; a prediction positions calculation part 22 calculating leg prediction positions; and a correlation processing part 23 performing correlation processing between the leg observation positions and the leg prediction positions to specify the leg positions. The correlation processing part 23 includes: a gate setting part 23a setting gates each spreading in a prescribed range with each leg prediction position as a reference; and an association part 23b associating the leg observation position as the leg position when the leg observation position is present inside the gate. The gate setting part 23a changes the range of the gate according to a state of a leg of the subject.

Description

  The present invention relates to a walking measurement system.

  2. Description of the Related Art Conventionally, in walking training such as Multi-target stepping test (MTST), a walking measurement system that measures walking characteristics (leg transport) during walking of a subject is used. As this kind of technology, for example, Patent Document 1 has attempted to record walking training data, quantitatively grasp the training situation, and apply the training data to a diagnostic device by comparison with past data. A footprint analyzer is described.

  The footprint analysis apparatus described in Patent Document 1 includes an image capturing device, an image calculation device, and a display device. From the image captured by the image capturing device, the floor surface contact position and the floor surface contact position of each foot are measured. The temporal change for each foot is calculated by the image calculation device, and the temporal change for each foot of the floor contact position is displayed on the display device.

JP 2002-345785 A

  By the way, generally, in the walking measurement system, the reliability (observation accuracy) for the observation value varies depending on the state of the subject's leg, and therefore it is difficult to specify the position of the subject's leg. There is a possibility that the leg position cannot be traced accurately.

  This invention is made | formed in view of such a situation, and makes it a subject to provide the walk measurement system which can trace a leg part position accurately.

  In order to solve the above problem, the walking measurement system according to the present invention emits a detection wave so as to scan along the horizontal direction, and the distance from the object that reflects the detection wave based on the reflection state of the detection wave Distance information acquisition means for acquiring distance information over time, and leg position specifying means for specifying a subject's leg position over time based at least on the distance information acquired by the distance information acquisition means ing. The leg position specifying means includes an observation position detection unit that detects a leg observation position that is an observation value of the leg position based at least on the distance information, and a predicted value of the leg position based on at least a leg position at a past time. A prediction position calculation unit that calculates a certain leg predicted position; and a correlation processing unit that performs a correlation process between the leg observation position and the leg predicted position to identify the leg position. The correlation processing unit is configured to associate a gate setting unit that sets a gate extending in a predetermined range with respect to the predicted leg position and a leg observation position as a leg position when the leg observation position exists in the gate. Part. The gate setting unit changes the range of the gate according to the state of the leg of the subject.

  In this walking measurement system, the range of the gate set in the correlation process when specifying the leg position can be changed according to the state of the leg of the subject. Thereby, it can be considered that the reliability of the detected leg observation position differs depending on the leg state, and it is possible to prevent the leg position from being accurately identified due to the difference in the state of the leg. As a result, the leg position can be traced with high accuracy.

  The leg position specifying means further includes a standing leg swing leg determining unit that determines whether the leg part of the subject is in a free leg state or a standing leg state. It is preferable to change the range of the gate when the state is determined to be larger than the range of the gate when the leg is determined to be in the standing state. When the leg is in the free leg state, it is highly necessary to cope with sudden acceleration of the leg, while when the leg is in the standing state, the necessity is low and the reliability of the leg observation position is high. Is found. Thus, when the leg is determined to be in the swinging state, the movement of the leg in the stance phase and the swinging phase is changed by changing the range of the gate so as to be larger than that in the standing state. It is possible to accurately trace the position of the leg by making use of this feature.

  Moreover, it is preferable that a gate setting part changes the range of a gate so that it may become large as the speed of a test subject's leg part becomes large. There is a finding that the reliability of the detected leg observation position varies depending on the speed of the leg, and decreases as the speed increases. Therefore, in this way, by changing the gate range so as to increase as the leg speed increases, the leg position can be accurately traced in consideration of the knowledge related to the leg speed. It becomes possible.

  In addition, when one leg of the subject is hidden as viewed from the detection wave emission direction, the gate setting unit determines the range of the gate related to the one leg as the hidden state continues. It is preferable to change so that it may become large. When one leg is hidden, this one leg may move while it is hidden, and may have a high speed when it appears. Therefore, in this way, by changing the range of the gate to increase as the hidden state continues for one leg, the position of the leg is preferably determined in consideration of the case where one leg is hidden. It becomes possible to trace with high accuracy.

Here, it is usually sufficient to detect the cross step in which the legs of the subject cross each other based on the landing positions of the legs. However, when the subject moves the body direction (trunk) itself in a direction different from the traveling direction, it may be difficult to accurately detect the cross step only with the landing positions of both legs. There is. In this regard, as a result of repeated studies by the present inventors, it has been found that when a cross step is detected, the leg positions and trajectories of both legs have a characteristic tendency. That is, when the landing position of one leg is located at a predetermined position, it has been found that there is a tendency to move while avoiding the one leg so that the other leg draws an arc.
Therefore, in the walking measurement system of the present invention, a cross step for detecting a cross step in which one and the other leg of the subject cross each other based on at least the leg position specified by the leg position specifying means. The cross-step detection means further includes a plurality of landing positions from among the plurality of leg positions, and a pair of landing positions that are closest to each other among the plurality of landing positions on one leg. A first landing line that is parallel to the landing line and that is separated from the landing line by a first predetermined distance from the landing line, and is parallel to the landing line, and A second reference line that is separated by a second predetermined distance from the floor line to the outside of one leg is set, and the landing position on the other leg is positioned outside the one leg from the first reference line. If a part of the trajectory of the leg position between the pair of landing positions for one leg is located outside the one leg from the second reference line, a cross step is detected. Is preferred.
This makes it possible to detect the cross step with high accuracy by using the characteristic tendency at the time of the cross step.

  The leg position specifying unit further includes a storage unit that stores leg information related to the thickness of the leg of the subject, and the observation position detection unit detects the edge position of the leg from the distance information, and It is preferable to detect the leg observation position from the position and leg information. In this case, the leg observation position can be detected with high accuracy.

  Moreover, it is preferable that the distance information acquisition means is installed so that a detection wave is emitted at a height position corresponding to the height of the subject's shin. In this case, the leg position can be specified with reference to the subject's shin. Thereby, since the speed of a shin part is stable compared with a toe part, for example, a leg part position can be traced stably.

  Moreover, it is preferable that the distance information acquisition means is installed so that a detection wave is emitted from the front side of the subject. As a result, for example, when one leg is hidden by the other leg as viewed from the direction in which the detection wave is emitted, the one leg tends to be in a standing state, and thus the leg observation position of the hidden one leg is Reliability can be increased.

  According to the present invention, it is possible to accurately trace the leg position.

It is a block diagram which shows the structure of the walking measurement system which concerns on one Embodiment. It is the schematic which shows MTST to which the walking measurement system of FIG. 1 was applied. It is a top view explaining the laser range sensor of the walking measurement system of FIG. It is a flowchart which shows the process in the walking measurement system of FIG. It is a figure for demonstrating specification of the leg part position by the walking measurement system of FIG. It is a figure explaining the example of the observation pattern used for the detection of the leg observation position by the walking measurement system of FIG. It is a figure explaining the other example of the observation pattern used for the detection of the leg part observation position by the walking measurement system of FIG. It is a graph which shows the speed of the leg part for demonstrating standing leg swing leg determination by the walking measurement system of FIG. It is a figure for demonstrating the step determination by the walking measurement system of FIG. It is a figure for demonstrating an example of the cross step detection by the walking measurement system of FIG. It is a figure which shows the example of the trace result of the leg part position measured by the walking measurement system of FIG. It is an enlarged view in the round frame in FIG. It is a figure which shows the other example of the trace result of the leg part position measured by the walking measurement system of FIG.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings. In the following description, the same reference numerals are used for the same or corresponding elements, and duplicate descriptions are omitted.

  FIG. 1 is a block diagram illustrating a configuration of a walking measurement system according to an embodiment, and FIG. 2 is a schematic diagram illustrating an MTST to which the walking measurement system of FIG. 1 is applied. As shown in FIG. 1 and FIG. 2, the walking measurement system 100 acquires the walking characteristics of the subject 1 by detecting and specifying the leg position and the movement of the leg of the walking subject 1. A range sensor (Laser Range Sensor, distance information acquisition means) 10, an electronic control device (leg position specifying means) 20, and a monitor 30 are included.

  The gait measurement system 100 here can be applied to the Multi-Target Stepping Test (MTST), which is one of the gait trainings of the subject 1 such as the elderly and the people with reduced mobility. The effect on prevention can be evaluated, and quantitative walking measurement and walking ability can be evaluated.

  MTST is a walking training for simultaneously strengthening the motor function and cognitive function, and prepares a walking path 3 such as a mat in which a plurality of three color targets 2 are randomly arranged. This is a walking exercise in which the subject 1 walks along the traveling direction 4 so as to step on only the single color target 2 that has been made. In addition, in MTST, in order to avoid the fall of the test subject 1, etc., attendant 5 such as a walking trainer is accompanied by the test subject 1 next to or behind. In FIG. 2, the back side of the paper is the start side, and the front side of the paper is the goal side.

  FIG. 3 is a plan view for explaining a laser range sensor of the walking measurement system of FIG. The laser range sensor 10 acquires two-dimensional plane distance information (distance information) that is a distance to an object around the sensor in 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 based on the reflected state of the laser beam L, the laser beam L is emitted. Two-dimensional plane distance information regarding the distance to the reflected object is acquired over time.

  Specifically, as shown in FIGS. 2 and 3, the laser range sensor 10 emits a laser beam L and reflects the laser beam L with a rotating mirror, thereby measuring in a measurement region including the walking path 3. The laser beam L is scanned in a fan shape in the horizontal direction. Then, for example, the reflected light of the laser light 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 the emission to reception of the laser light L (propagation time) are measured. Then, two-dimensional plane distance information including information related to the angle and distance from the leg F is detected.

  The laser range sensor 10 is configured such that the height of the light window portion that emits the laser light L can be adjusted, and corresponds to the height of the shin portion of the subject 1 (that is, the height from the ankle to the lower knee). It is installed so that the laser beam L is emitted at the height position. The height of the optical window portion of the laser range sensor 10 is such that the leg F leaves the floor during the swing phase (the time when the leg F is in the free leg state) and the subject 1 moves away from the laser range sensor 10. Considering that the leg portion F can be detected even in the case of being present, the width is set based on the average height at which the width of the leg portion F becomes maximum, for example, 0.27 m from the floor surface.

  Further, the laser range sensor 10 is arranged so that the emission direction of the laser light L is directed toward the walking path 3 at the tip of the traveling direction 4 of the walking path 3, from the front side of the subject 1 toward the subject 1. Laser light L is emitted. The laser range sensor 10 is installed such that the emission direction of the laser light L is horizontal with respect to the floor surface.

  As the laser range sensor 10, a laser range sensor having a range of 0.1 to 30 m, a range angle of 270 deg, a range accuracy of ± 30 mm, an angle resolution of 0.25 deg, and a sampling period of 25 ms / scan is used. It has been. Incidentally, the laser range sensor 10 is not limited in its type and specification (spec), and various types can be used according to the measurement environment, for example.

  The electronic control unit 20 performs calculation based on the two-dimensional plane distance information acquired by the laser range sensor 10 to identify the leg position that is the position of the leg F of the subject 1 over time, and determines the walking characteristics of the subject 1. To get. For example, a personal computer or a dedicated control computer is used as the electronic control device 20 and includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like.

  The electronic control unit 20 includes, as functional elements, an observation position detection unit 21 that detects a leg observation position that is an observation value of the leg position of the subject 1 and a leg that is a predicted value of the leg position of the subject 1. A predicted position calculating unit 22 for calculating a predicted position, a correlation processing unit 23 for performing correlation processing between the leg observation position and the predicted leg position and specifying (assigning) the leg position, and a post-processing of the specified leg position And a post-leg processing unit 24 for determining whether the leg F is in the stance state or the free leg state.

  The observation position detection unit 21 detects the leg observation position based on the two-dimensional plane distance information. Specifically, the observation position detection unit 21 detects the edge position of the leg F by performing edge detection processing from the two-dimensional plane distance information in view of the characteristic shape of the two-dimensional plane distance information of the leg F. The leg observation position is detected as a candidate point of the leg F from the edge position by pattern recognition using a plurality of observation patterns. At this time, in this embodiment, the leg information regarding the thickness of the leg F of the subject 1 is used, and the thickness of the leg F is known.

  The predicted position calculation unit 22 calculates the predicted leg position based on the leg position at the past time. Specifically, the predicted position position is calculated from the past leg position and the speed of the leg F. Is calculated. The correlation processing unit 23 specifies the leg position from the relative positional relationship between the leg observation position and the predicted leg position, and includes a gate setting unit 23a and an association unit 23b.

  The gate setting unit 23a sets a gate that is a determination range that extends to a predetermined range centering on the predicted leg position as a part for limiting the leg observation position that can be correlated with the target in a crowded environment. In addition, the gate setting unit 23a changes the range of the gate according to the state of the leg (the standing leg, the free leg, the speed and the hiding) which is the state of the leg F (details will be described later).

  When the leg observation position exists in the gate, the associating unit 23b associates only the leg observation position in the gate with the target, for example, specifies the leg observation position as the leg position. In addition, when the leg observation position does not exist in the gate, the associating unit 23b associates the predicted leg position with the target, for example, specifies the predicted leg position as the leg position.

  The post-processing unit 24 removes the influence of the observation noise from the specified leg position. For example, the post-processing unit 24 performs a calculation based on a human behavior model on the leg position, thereby calculating the laser range sensor from the leg position. The effect of 10 noises is removed. Based on the speed of the leg F, the standing leg free leg determination unit 25 determines whether the leg F is in the standing state or the free leg state.

  The free leg state means a state of the leg that does not apply weight to the body, and includes a state where the leg F is separated from the floor (ground). The standing state means the state of the leg part where the weight of the body is applied, and includes the state where the leg part F is in contact with the floor, and is also referred to as a supporting leg state.

  Further, the electronic control unit 20 has, as functional elements, a landing position acquisition unit 26 that acquires a landing position, which is a position where the leg F has landed on the floor surface, and a designated target 2 (see FIG. 2). A step determination unit 27 that determines whether or not the subject 1 has stepped correctly, a cross step detection unit 28 that detects a cross step in which one and the other of the legs F cross each other, and each piece of information related to walking characteristics Is further stored.

  The landing position acquisition unit 26 acquires the landing position based on the speed of the leg F. Here, the time at which the speed of the leg F becomes the smallest during the stance period (the time when the leg F is in the stance state) is set as the landing time, and the leg position at that time is set as the landing position. The step determination unit 27 sets a step determination area at a position corresponding to the target 2, and determines that the target 2 has been correctly stepped on when the landing position is included in the step determination area.

  The cross step detection unit 28 sets the landing line L0, the first reference line L1, and the second reference line L2 from the pair of landing positions that are closest to each other among the plurality of landing positions in the one leg F, and these A cross step is detected with reference to the lines L0 to L3 (details will be described later).

  The storage unit 29 stores information on the leg position, the speed of the leg F, the standing leg swing leg determination result of the standing leg swing leg determination unit 25, the determination result of the step determination unit 27, and the detection result of the cross step detection unit 28, At least stored in association with (sampling time). Further, the storage unit 29 stores leg information regarding the thickness of the leg F of the subject 1 (the length corresponding to the diameter of the shin part). The thickness of the leg portion F can be acquired by detecting the laser beam using the laser range sensor 10 before or after the subject 1 walks, for example.

  The monitor 30 is for displaying the calculation result by the electronic control unit 20. The monitor 30 here displays each information stored in the storage unit 29 as, for example, the walking characteristics of the subject 1 or a movement result indicating whether or not the subject has moved correctly, and feeds back to the subject 1. The monitor 30 may be configured to output sound or the like instead of or in addition to display. Further, instead of or in addition to the monitor 30, an output unit that prints out the calculation result or the like on a paper medium may be further provided.

  Next, the case where the walking characteristic in MTST is measured using the walking measurement system 100 of this embodiment is demonstrated, referring the flowchart shown in FIG.

First, in the walking measurement system 100, the acquisition of the two-dimensional plane distance information by the laser range sensor 10 and the processing based on the acquired two-dimensional plane distance information are sequentially performed, and the leg positions of both legs F _R and F _L of the subject 1 are sequentially performed. And the velocity is estimated and tracked by the Kalman filter. Hereinafter, the process at the sampling time k will be described as an example.

[Detection of leg observation position]
FIG. 5 is a diagram for explaining the specification of the leg position by the walking measurement system of FIG. 1. As shown in FIG. 5A, the observation position detection unit 21 detects the leg observation position y _k of the subject 1 from the two-dimensional plane distance information acquired by the laser range sensor 10 (S1). Specifically, the edge position E of the leg F is detected from the two-dimensional plane distance information, and a plurality of observation patterns O1 to O5 (see FIGS. 6 and 7) are used based on the positional relationship of the detected edge positions. The leg observation position y _k is detected by detecting the detected pattern.

  6 and 7 are diagrams for explaining examples of observation patterns used for detection of leg observation positions. 6A is an example using an SL (Single Leg) pattern, FIG. 6B is an example using an LT (two Legs Together) pattern, and FIG. 6C is an example using an FSO (Forward Straddle Observable) pattern. It is. FIG. 7A shows an example using an FSU (Forward Straddle Unobservable) pattern, and FIG. 7B shows an example using a UO (Unobservable) pattern.

  As shown in FIGS. 6 and 7, the observation pattern O <b> 1 as the SL pattern is a state in which the leg portion F can be completely observed from the laser range sensor 10 alone. The observation pattern O <b> 2 as the LT pattern is a state in which the legs F are aligned and can be completely observed from the laser range sensor 10. The observation pattern O3 as the FSO pattern has a step shape due to the influence of one leg F or a cane, and the width between the edge positions E is ½ or more of the width of the leg F of the subject 1, In this state, the center position of the part F can be observed from the laser range sensor 10.

  The observation pattern O4 as the FSU pattern has a platform shape like the FSO pattern, but the width between the edge positions E is less than ½ of the width of the leg F of the subject 1 and the center position of the leg F is The laser range sensor 10 cannot be observed directly. The observation pattern O5 as the UO pattern is in a state where it cannot be observed due to the influence of one leg F or a cane.

Incidentally, in the detection of the legs observation position y _k, it does not satisfy the observed pattern O1~O4 edge position E is determined not to be a leg F. In addition, since the measurement region is known in advance in MTST, in order to avoid unnecessary correlation processing, a region wide by a predetermined length with respect to the region of the walking path 3 is set as the measurement region, and the leg observation position y _k outside the measurement region. And the corresponding edge position E are excluded.

Here, in the detection of the leg observation position y _k , the thickness (length corresponding to the diameter of the shin part) w of the leg F is used. For example, the observation pattern of the FSU pattern shown in FIG. in O4, the right leg F _R, the center between the edge position E is not being detected as a leg observed position y _k, a position that has entered the inside by half the thickness w, for example, from the edge position E It is detected as a leg observed position y _k. As described above, by detecting the leg observation position of the leg F according to the plurality of observation patterns O1 to O5, even when the leg F overlaps and is hidden when viewed from the laser range sensor 10, the leg F Can be suitably detected.

[Calculation of predicted leg position]
Subsequently, as shown in FIG. 5 (b), based on the leg position at the past time, the predicted position calculation section 22 calculates the predicted leg position y ^ _k (see the x mark in the figure) ( S2). Specifically, assuming that the movement model described by the linear discrete time equations shown in the following equations (1) and (2) is estimated using a Kalman filter, each state estimated value obtained at the previous sampling time k−1 is Based on the following formula (1), the prior state estimated value and the prior error covariance matrix are obtained by the following formulas (3) and (4), respectively.

Here, M is the target number to be tracked, v _k is the system noise vector of covariance matrix Q with mean vector 0, w _k is the observed noise vector of covariance matrix R with mean vector 0, and normal white that is independent of each other Assumes noise. Based on the following equation (5), the leg predicted position y ^ _k is calculated.

[Correlation]
Subsequently, as shown in FIG. 5 (c), by performing the correlation processing of the legs observation position y _k and leg predicted position y ^ _k by the correlation processing unit 23, identifies the leg position P _k (S3) . Specifically, a circular gate G centered on the predicted leg position ＾ _k is set, and memory tracking is performed in which only the leg observation position y _{k in the} leg gate position P _k is set. Here, it is determined that based on the chi ² test, legs observation position y _k is the leg section observation position y _k when the following expression is satisfied (6) present in the gate G.

S _k is a residual covariance matrix calculated from the observation noise covariance matrix R of the laser range sensor 10 and the prediction error covariance matrix P _{k / k−1} of the Kalman filter, and is obtained by the following equation (7). . G is a parameter set based on the significance level of the test.

If the leg observation position y _k is not obtained in the gate G, the memory tracking is performed with the predicted leg position y ^ _k from the previous sampling time k−1 by the Kalman filter as the leg position P _k at the current time k. I do. When a plurality of leg observation positions y _k are obtained in the gate G, association using the NN (Nearest Neighbor) method or the like may be performed, and when the gates G overlap, A Global Nearest Neighbor (GNN) method may be used.

[post process]
Subsequently, the post-processing unit 24 performs an operation based on the human behavior model for the leg position _Pk , thereby removing the influence of noise of the laser range sensor 10 from the leg position _Pk (S4).

[Standing leg detection]
Subsequently, the standing leg swing leg determination unit 25 determines whether the leg part F is in the standing leg state or the free leg state based on the speed of the leg part F (S5). In human walking, it is found that the state of each leg F can be identified by using the speed of the leg F as an index. Therefore, in the S5, the speed ^v _{L k} of the right leg portion _F velocity of _R ^v _{R k} and the left leg _{F L} at time k, determined using the estimated value by the Kalman filter. Then, based on the magnitude relationship between these velocities v ^R _k and v ^L _k , the free leg state or the standing leg state is determined.

FIG. 8 is a graph showing the leg speed for explaining the standing leg swing leg determination. In FIG. 8, a small black circle indicates a value related to the leg F in the free leg state, and a double circle (including a large black circle) indicates a value related to the leg F in the standing state (FIG. 8). 5, 11-13). As shown in FIG. 8, in the stance swing determination here, considering that sometimes legs F _R, is F _L to ambulation is when there is a movement like stumbling in the subject 1, the stance _{Define the} maximum speed v _{st_max} and the minimum speed v _{sw_min} of the swing leg period. If the speed v of the leg F exceeds the maximum speed v _{st_max} , the swing leg state is assumed, and if it is slower than the minimum speed v _{sw_min, the stand} leg state is _assumed. Yes.

Preferably, the maximum speed v _{st_max} can be set to 1/3 of the average walking speed of a person 1.1 m / s, and the minimum speed v _{sw_min} can be set to 1/2 of the maximum speed v _{st_max.} . In the above-mentioned stance swing leg determination, for example, the leg F related to the larger one of the speeds v ^L _k and v ^R _k may be determined as the free leg state, and the leg F related to the smaller one may be determined as the stance state. .

Subsequently, the above S1~S5 are repeated until the acquisition of the two-dimensional plane distance information is completed, the legs of the subject 1 F _R, leg position P _k of F _L is to be traced (S6).

Here, in the present embodiment, the range of the gate G set in the correlation process (S3) is changed according to the state of the leg of the subject 1. That is, in consideration of the state of the leg F at the pre-sampling time k-1, the thickness w of the leg F of the subject 1, the speed v (k-1) of the leg F at the time k-1, and the laser range. The range of the gate G is set as shown in Table 1 below using the time that cannot be observed by the sensor 10. Here, H (k) is the number of times that the leg F could not be observed continuously at time k. In consideration of the analysis result and the average moving speed of humans, a = 0.75, b = 1.0, v _st = 1.1, and v _sw = v _st / 2.

  Specifically, as shown in FIG. 5 (c), the leg F sets the range of the gate G1 in the swing phase while the leg F sets the range of the gate G2 in the stance phase, The gate G1 is set small because the observation accuracy is high. That is, the range of the gate G1 when the leg F is determined to be in the free leg state is changed to be larger than the range of the gate G2 when the leg F is determined to be in the standing state.

  Here, the basic size a · w of the gate G2 in the stance phase and the basic size b · w of the gate G1 in the stance phase are set to satisfy a · w <b · w. As the basic size a · w of the gate G2 in the stance phase, a minimum value determined based on the performance error of the laser range sensor 10 can be used.

In addition, the reliability of the leg observation position y _k varies depending on the speed v of the leg F, and the observation accuracy decreases as the speed v increases. Therefore, for the range of the gate G, the product of the velocity v (k−1) at the sampling time k−1 and the sampling period Δt (the amount of movement between Δt) is added, and the range increases as the velocity v increases. To do.

  Furthermore, when the leg F is hidden and cannot be detected by the cross step or the like, the case where the leg F moves while it is hidden is also considered, and the hidden time H (k) The range of the gate G is increased according to the speed v during that time. That is, when one leg F is hidden when viewed from the emission direction of the laser beam L, the range of the gate G related to the one leg F increases as the hidden state continues. To change.

Here, for example, when the UO pattern (see FIG. 7B) is obtained as a result of pattern detection in the detection of the leg observation position, the speed v is obtained when the leg F just before being hidden is in the standing state. On the other hand, it is assumed that it is moving at _st. On the other hand, if it is in the swinging state, it is assumed that it is moving at v _sw (where v _sw > v _st ), and the range of the gate G is expanded.

By the above processing, the leg F to be tracked is contained in the gate G, and the leg observation position y _k other than the leg F is prevented from being contained as much as possible. As a result, the left and right legs F _R , it is reduced to erroneously identify the legs of F _L and bridesmaid 5. In other words, by changing the range of the gate G, it is possible to reduce false detections by the staff, the attendant 5 and the like. In the example shown in FIG. 5 (c) for example, 'even if it is detected as _k, the leg observation position y' position of the wand subject 1 uses the legs observation position y identified as _k legs position P _k Can be avoided. Further, even when the leg F is temporarily unobservable, the leg F can be suitably traced when the leg F returns to an observable state by expanding the range of the gate G every moment. .

Next, after the above S1 to S6, the landing position acquisition unit 26 obtains the time when the earliest is the smallest during the stance phase as the landing time (see the large black circles in FIG. 8). The leg position _Pk at the floor time is acquired as the landing position (S7). Subsequently, for all of the plurality of acquired landing positions, the step determining unit 27 determines whether or not the subject 1 has correctly stepped on the designated target 2 (see FIG. 2) (S8).

  FIG. 9A is a diagram illustrating a leg model used for step determination, and FIG. 9B is a diagram illustrating a step determination region in step determination. In S8, as shown in FIG. 9, a leg model 40 based on the average value of the body data of the elderly is created, the leg model 40 is set at the acquired landing position, and the specified color is displayed. A step determination area 41 is set for the target 2. When the landing position where the leg model 40 is set is included in the step determination area 41, it is determined that the target 2 has been correctly stepped on.

  Subsequently, the cross step detection unit 28 detects the cross step of the subject 1 for all of the plurality of acquired landing positions (S9). Specifically, as described below, when the leg F in the free leg state moves while avoiding the leg F in the standing state so as to draw an arc, the cross step is detected.

FIG. 10 is a diagram for explaining an example of cross-step detection. In FIG. 10, the upper side is the traveling direction, the leg position _Pk is represented by a black circle, and the landing position of the leg position _Pk is represented by a large black circle. As shown in FIG. 10, if the right leg F _R is implanted in the landing zone 51, the origin implantation position 50 one time before the right leg F _R, take the landing zone 51 on the x-axis An xy coordinate system is defined, and a landing line 52 extending through the landing positions 50 and 51 is set.

Also, setting the first reference line 53 leaving the inner side of the right leg portion F _R (y-axis direction of the negative side, the left side) to a predetermined distance w _c and the implantation line 52 parallel to the implantation line 52. This together with the outer right leg F _R from and landing line 52 parallel to the implantation line 52 (y-axis direction of the positive side, the right side) to set the second reference line 54 leaving only a predetermined distance w _c. Note that w _c is a threshold value, and w _c = w / 2 in consideration of the thickness w of the leg F of the subject 1.

Then, as in the state illustrated, at the time when the landing position 55 in the left leg F _L were than the first reference line 53 positioned outside of the right leg portion F _R, implantation of the right leg portion F _R some of the trajectory 60 of the leg position _{P k} between positions 50 and 51, when located outside of the right leg portion _{F R} than the second reference line 54, left and right leg portions _F R, is _{F L} A cross step is detected as crossing.

  As described above, the processing by the electronic control unit 20 is completed, and thereafter, the measurement result of the walking characteristic, the step determination result, and the cross step detection result of the subject 1 are displayed on the monitor 30 and fed back to the subject 1.

By the way, when the range of the gate G is large, it is possible to cope with a rapid change in the speed of the leg F. In this case, the leg of the cane or the attendant 5 (see FIG. 2) is also included in the gate G. As a result, the correlation process may be complicated, and the risk of erroneous detection increases. On the other hand, when the range of the gate G is small, although the legs observation position y _k in the gate G is unlikely that false detection occurs in order to become difficult to contain, when it can not correspond to the rapid change in velocity of leg F There is.

In this respect, in the gait measurement system 100 of the present embodiment, as described above, the range of the gate G set in the correlation processing (the above S3) when specifying the leg position _Pk is set to the leg F of the subject 1. It can be changed according to the state. As a result, it can be considered that the reliability of the detected leg observation position y _k varies depending on the state of the leg F, and the leg position P _k can be accurately identified due to the difference in the state of the leg F. It is possible to suppress what cannot be done. That is, taking advantage of the features in walking, hidden or leg F due to cross step, etc., bridesmaid 5 and can wand or the like is suitably corresponds to be included in the two-dimensional plane distance information, the leg position P _k It becomes possible to trace with high accuracy.

Further, when the leg F is in the free leg state, it is highly necessary to make the range of the gate G correspond to the rapid acceleration of the leg F, and the reliability that the leg F exists at the leg observation position y _k is low. . On the other hand, when the leg F is in the standing state, it is less necessary to make the range of the gate G correspond to the rapid acceleration of the leg F, and the reliability of the leg observation position is increased. Therefore, in the present embodiment, as described above, it is determined whether the leg F is in the free leg state or the standing leg state, and the range of the gate G1 in which the leg F is in the free leg state is determined as the leg F. Is larger than the range of the gate G2 when it is determined to be in the standing state. Accordingly, it is possible to accurately trace the leg position P _k by making use of the characteristics of the movement of the leg F during the stance phase and the swing phase.

Further, it is found that the reliability of the detected leg observation position y _k varies depending on the speed of the leg F, and decreases as the speed of the leg F increases. Therefore, in the present embodiment, as described above, the range of the gate G is changed so that the speed of the leg F increases as the speed of the leg F increases. Considering this, the leg observation position y _k can be traced with high accuracy.

Further, when one leg F is hidden behind the other leg F or the like with respect to the laser range sensor 10, this one leg F may move while hidden, and thereafter May appear at a point in time (a point at which observation by the laser range sensor 10 becomes possible). In this regard, in the present embodiment, as described above, the range of the gate G of the leg F is changed so as to increase as the hidden state of the leg F continues. Thereby, for example, even when the hidden leg F appears at a high speed, the leg position _Pk of the leg F can be accurately identified. That is, it is possible to trace the leg position P _k with high accuracy in consideration of the case where the leg F is hidden.

Further, usually, when performing the cross step detection of the walking training straight, if the subject 1 is moving toward a goal direction, left and right leg portions F _R, the determination by the coordinates of the landing position of F _L Just do it. However, when the body itself moves toward the direction of the next target 2 instead of the goal direction, it is difficult to detect the cross step only by the landing position.

Therefore, in the present embodiment, as described above, the landing line L0, the first and second reference lines L1 from the pair of landing positions 50 and 51 that are closest to each other among the plurality of landing positions in one leg F. , L2 are set. And when the landing position 55 in the other leg F is located outside the one leg F with respect to the first reference line L1, the leg between the landing positions 50 and 51 for the one leg F. When a part of the locus of the part position _Pk is located outside one leg part F from the second reference line L2, a cross step is detected.

  Thus, a characteristic tendency at the time of cross step, that is, at the time of cross step, the landing position of one leg F is located at a predetermined position, and the one leg F is arced by the other leg F. The tendency to move while avoiding drawing can be preferably used, and as a result, the cross step can be accurately detected.

In the present embodiment, as described above, the leg information related to the thickness w of the leg F of the subject 1 is stored in the storage unit 29, and the edge position E and the leg F detected from the two-dimensional distance information are stored. The leg observation position y _k is detected from the thickness of the leg. Thereby, for example, even if the center position of one leg F cannot be observed directly from the laser range sensor 10 due to the influence of the other leg F or the cane (the FSU pattern or the like), the leg observation is performed. The position y _k can be detected with high accuracy.

In the present embodiment, as described above, the laser range sensor 10 is installed so that the laser light L is emitted at a height position corresponding to the height of the shin part of the subject 1. Thereby, leg position _Pk can be specified on the basis of subject's 1 shin part. As a result, for example, the speed of the shin part is more stable than the toe part and the speed fluctuation is not steep, so that the leg position _Pk can be traced stably and suitably.

In the present embodiment, as described above, the laser range sensor 10 is installed so that the laser light L is emitted from the front side of the subject 1. Thereby, for example, when one leg F is hidden behind the other leg F when viewed from the emitting direction of the laser beam L, the one leg F is likely to be in a standing state. Thus, for the hidden leg F, can reduce the need to adapt to the rapid acceleration, it is possible to increase the reliability of the legs observation position y _k.

  FIG. 11 is a diagram showing an example of the trace result of the leg position measured by the walking measurement system, and FIG. 12 is an enlarged view in a round frame in FIG. 12A shows a state at one time, and FIG. 12B shows a state after a predetermined time has elapsed from the state shown in FIG. 12A.

As shown in FIG. 11 and FIG. 12 (a), the but legs observation position _{y k} of the right leg portion _{F R} is out of the gate G1 of the right leg portion _{F R,} the left leg _{F L} is the stance Therefore, the range of the gate G2 is small, the legs observation position y _k of the right leg portion F _R to the gate G2 are not included. Therefore, the left leg F _L can be associated with leg observation position y _k of the left leg portion F _L in the gate G2, the right leg portion F _R, memory tracking with legs predicted position y ^ _k expanding the range of the gate G1 while moving by, as shown in FIG. 12 (b), and after a predetermined time has elapsed, can detect the leg observation position y _k in the gate G1. As a result, it is possible to realize a precise trace.

  FIG. 13 is a diagram illustrating another example of the trace result of the leg position measured by the walking measurement system. In the result shown in FIG. 13, when the step determination of the target 2 is determined by the video camera, the number of times the target 2 has been correctly stepped is 414 times, but when the determination by the above step determination of the present embodiment is performed. Was 410 times. As a result, it was confirmed that step determination was possible with a high accuracy of 99% (410/414).

  Further, in the result shown in FIG. 13, when cross step detection is performed by analysis with a video camera, 19 cross steps are detected. In the cross step detection of the present embodiment, 15 cross steps are detected. Could be detected. In the present embodiment, no cross step is detected by mistake.

  The preferred embodiments of the present invention have been described above. However, the present invention is not limited to the above-described embodiments. The present invention can be modified without departing from the scope described in the claims or applied to other embodiments. May be.

For example, in the above-described embodiment, the walking measurement system 100 is adapted to the measurement of walking characteristics in MTST, but is not limited thereto, and can be applied to the measurement of walking characteristics in various other walking trainings. Moreover, in the said embodiment, although the gate G was set to the circle | round | _yen centering on the leg part predicted position y ^ _k , the gate G should just be set on the basis of the leg predicted position y ^ _k , and circular. It may be set to other shapes other than.

  In the above embodiment, the range of the gate G is varied according to the stance, swing leg, speed, and hiding of the leg F. However, the range of the gate G may be varied according to at least one of these. The range of the gate G may be varied according to the state of other legs F.

In the above embodiment, in the cross-step detection, although both set away by a predetermined distance w _c a first and second reference lines 53 and 54 from implantation line 52, but is not limited thereto. The first predetermined distance at which the first reference line 53 is separated from the landing line 52 and the second predetermined distance at which the second reference line 54 is separated from the landing line 52 may not be equal to each other, or other threshold values. May be.

  Moreover, in the said embodiment, when measuring a walk characteristic using the walking measurement system 100, after implementing said S1-S6 by real-time processing, it implements said S7-S9 by non-real-time processing, It is not limited to this. S1 to S6 may be performed by non-real time processing, and S7 to S9 may be performed by real time processing.

  In the above embodiment, the range of the gate G1 of the leg F in the free leg state is changed so as to be larger than the range of the gate G2 of the leg F in the standing leg state. The range of the gate G2 of the leg F may be changed to be smaller than the range of the gate G1 of the leg F in the free leg state. In the above, the landing position acquisition part 26 and the cross step detection part 28 comprise the cross step detection means of a claim.

DESCRIPTION OF SYMBOLS 10 ... Laser range sensor (distance information acquisition means), 20 ... Electronic control apparatus (leg position specifying means), 21 ... Observation position detection part, 22 ... Predicted position calculation part, 23 ... Correlation processing part, 23a ... Gate setting part , 23b ... association unit, 25 ... standing leg swing leg determination unit, 26 ... landing position acquisition unit (cross step detection unit), 28 ... cross step detection unit (cross step detection unit), 29 ... storage unit, 100 ... walking measurement system, F ... legs, F L _... left leg (legs), F R _... right leg (legs), L ... laser light (detection waves).

Claims (8)

A distance information acquisition unit that emits a detection wave so as to scan along a horizontal direction, and acquires distance information related to a distance from an object that reflects the detection wave based on a reflection state of the detection wave;
Leg position specifying means for specifying the position of the leg of the subject over time based on at least the distance information acquired by the distance information acquiring means,
The leg position specifying means includes:
An observation position detector that detects a leg observation position that is an observation value of the leg position based at least on the distance information;
A predicted position calculating unit that calculates a predicted leg position that is a predicted value of the leg position based at least on the leg position at a past time;
A correlation processing unit that performs correlation processing of the leg observation position and the predicted leg position to identify the leg position;
The correlation processing unit
A gate setting unit that sets a gate that extends into a predetermined range based on the predicted leg position;
An association unit that associates the leg observation position as the leg position when the leg observation position is present in the gate;
The said gate setting part changes the range of the said gate according to the state of the said test subject's leg part, The walk measurement system characterized by the above-mentioned. The leg position specifying means further includes a standing leg swing leg determining unit that determines whether the leg of the subject is in a free leg state or a standing leg state,
The gate setting unit changes the range of the gate when the leg is determined to be in a free leg state so as to be larger than the range of the gate when the leg is determined to be in a standing state. The walking measurement system according to claim 1, wherein:   The walking measurement system according to claim 1, wherein the gate setting unit changes the range of the gate so as to increase as the speed of the leg of the subject increases.   When the gate setting unit is in a state where one leg of the subject is hidden as viewed from the direction of emission of the detection wave, the hidden state continues in the range of the gate related to the one leg. The gait measurement system according to claim 1, wherein the gait measurement system is changed so as to increase with increasing speed. Cross step detecting means for detecting a cross step in which one and the other leg of the subject cross each other based on at least the leg position specified by the leg position specifying means;
The cross step detection means includes:
Identifying a plurality of landing positions from among the plurality of leg positions;
Setting a landing line that passes through a pair of landing positions closest to each other among the plurality of landing positions in the one leg,
A first reference line that is parallel to the landing line and that is separated from the landing line by a first predetermined distance inside the one leg, and
A second reference line that is parallel to the landing line and that is separated from the landing line by a second predetermined distance outside the one leg,
When the landing position of the other leg is positioned outside the one leg from the first reference line, the leg between the pair of landing positions of the one leg 5. The cross step is detected when a part of the locus of the position is located outside the one leg part with respect to the second reference line. 6. The walking measurement system according to the item. The leg position specifying means further includes a storage unit in which leg information related to the thickness of the subject's leg is stored,
The said observation position detection part detects the edge position of the said leg part from the said distance information, The said leg part observation position is detected from the said edge position and the said leg part information, It is characterized by the above-mentioned. The walking measurement system according to any one of claims.   The distance information acquisition means is installed so that the detection wave is emitted at a height position corresponding to the height of the shin part of the subject. The walking measurement system according to the item.   The gait measurement system according to any one of claims 1 to 7, wherein the distance information acquisition unit is installed such that the detection wave is emitted from the front side of the subject.

Images