JP2012245211A - Walking state estimating device and walking assist device using the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device and the like which highly accurately estimates a walking state of a human being regardless of degree of asymmetry between his or her right and left body motions.SOLUTION: A "difference" between a right and a left hip joint angles or shoulder joint angles of a human being is sampled, thereby a waveform signal (differential oscillator) is obtained. Also, a power spectrum is obtained from the waveform signal. As a result, regardless of the degree of asymmetry between the right and left body motions of the human being, a walking state of the human being is highly accurately estimated according to a certain reference of referring to a basic frequency fhaving height equal to or larger than threshold in the power spectrum and indicative of a single peak position positioned in a lowest frequency band.

Description

  The present invention relates to a device for estimating a human walking state and a device for assisting the human walking motion using the estimation result.

  In order to detect acceleration in the human vertical direction, etc., an output waveform of an acceleration sensor attached to the human body is obtained, and the human walking cycle is detected based on the power spectrum of the output waveform. A technique has been proposed (see Patent Document 1).

JP 2005-342254 A

  However, when the movements of the left and right legs of the human are asymmetrical, the human walking cycle is calculated based on the frequencies indicating the multiple peak positions that appear in the power spectrum according to multiple criteria according to the number of peaks. Is done. For this reason, depending on the degree of asymmetry of the human leg motion, an optimal criterion for estimating the human walking cycle may not be selected from the plurality of criteria, and as a result, the estimation accuracy of the walking cycle may be reduced. There is.

  Accordingly, an object of the present invention is to provide a device and the like that can estimate the walking state of a human with high accuracy regardless of the degree of asymmetry between the left and right body movements of the human.

  The walking state estimation device according to the present invention performs window processing for applying a window to a differential transducer, which is a waveform signal obtained by sampling over a specified period of a difference between left and right hip joint angles or shoulder joint angles of a human being walking. A configured window processing element, a frequency analysis processing element configured to acquire a power spectrum by performing frequency analysis on the differential transducer to which the window is multiplied, and a high value equal to or higher than a threshold in the power spectrum. And a spectral analysis processing element configured to determine a fundamental frequency indicating a peak located in the lowest frequency band, wherein the window processing element is determined by the previous spectral analysis processing element. Based on the fundamental frequency, the current window width is set according to the decreasing function with the fundamental frequency as a variable. Characterized in that it is configured.

  According to the walking state estimation apparatus of the present invention, a difference transducer as a waveform signal is obtained by sampling a “difference” between the human hip joint angle or shoulder joint angle, and a power spectrum is obtained from the difference transducer. can get. For this reason, the “fundamental frequency” representing the position of a single peak located in the lowest frequency band having a height equal to or higher than a threshold value in the power spectrum is referred to regardless of the level of asymmetry between the left and right body movements of the human. According to a certain standard, the human walking state can be estimated with high accuracy.

  As the previous fundamental frequency is higher, the width of the current window to be applied to the differential transducer in the current specified period is set narrower, while as the previous fundamental frequency is lower, the width of the current window is set wider. For this reason, from the viewpoint of estimating the fundamental frequency this time, a differential transducer having an appropriate width can be extracted as a frequency analysis target.

  The window processing element is configured to perform the window processing after removing a high frequency component exceeding a first designated frequency from the differential transducer by down-sampling the differential transducer. Is preferred.

  The window processing element is configured to execute the window processing after removing the low frequency component equal to or lower than a second specified frequency from the differential transducer by passing the differential transducer through a high-pass filter. It is preferable.

  The present invention includes a first orthosis attached to the upper body of a human, a pair of second orthosis attached to each of the left and right thighs of the human, a pair of actuators, and the hip joint angles of the human left and right A left hip joint angle sensor and a right hip joint angle sensor configured to output a signal corresponding to each of the signals, and an operation of each of the pair of actuators based on at least output signals of the left hip joint angle sensor and the right hip joint angle sensor And a control device configured to control the left and right of the upper body by moving each of the pair of second devices relative to the first device by movement of each of the pair of actuators. The present invention relates to a device for assisting the human walking movement with relative periodic movement of the thigh.

  The walking assist device according to the present invention is the differential oscillator, which is a waveform signal obtained by sampling over a specified period of a difference between the left and right hip joint angles or shoulder joint angles of the human through the left joint angle sensor and the right hip joint angle sensor. The walking state estimation device for determining the fundamental frequency using the control device, wherein the control device performs a periodic operation of the actuator according to a cycle determined according to the fundamental frequency determined by the walking state estimation device. It is comprised so that it may control.

  According to the walking assist device of the present invention, the periodic motion of each human leg is assisted by the periodic motion of the actuator. At this time, the operation cycle of the actuator is determined based on the fundamental frequency representing the estimated human walking state by the walking state estimating device of the present invention. For this reason, in consideration of the human walking state estimated with high accuracy based on the fundamental frequency as described above, the human walking motion can be assisted at an appropriate period.

BRIEF DESCRIPTION OF THE DRAWINGS Configuration explanatory drawing of the walking assistance apparatus of this invention. Structure explanatory drawing of the control apparatus of a walk auxiliary device. Explanatory drawing regarding the control method of a walking assistance apparatus. Explanatory drawing regarding the setting method of a time constant. Explanatory drawing regarding the fundamental frequency estimation method. Explanatory drawing regarding a waveform signal, window processing, and frequency analysis.

(Configuration of walking assist device)
Hereinafter, the symbols “L” and “R” are used to distinguish the left and right of the leg and the like, but the symbols are omitted when it is not necessary to distinguish the left and right or when the vector having the left and right components is expressed. In addition, symbols “+” and “−” are used to distinguish the flexion motion (forward motion) and the extension motion (backward motion) of each thigh with respect to the upper body.

  A walking assist device 10 shown in FIG. 1 includes a first brace 11, a pair of left and right second braces 12, a pair of left and right actuators 14, a battery 16, a control device 20, and a hip joint angle sensor 202. I have.

  The first brace 11 is mounted so as to be wrapped around the upper body or waist (first body part) of a human or a user. The back part which contacts at least a human back among the 1st braces 11 is comprised by rigid materials, such as a lightweight alloy, hard resin, or carbon fiber, and the other part is comprised by flexible materials, such as a fiber.

  The second brace 12 is made of a flexible material such as fiber, and is worn so as to be wound around a human thigh (second body part). The second appliance 12 may be provided only on one of the left and right sides instead of the left and right sides.

  The actuator 14 is configured by an electric motor, and is configured by one or both of a speed reducer and a compliance mechanism in addition to the motor as necessary. The actuator 14 is connected to the first appliance 11 so as to be arranged on both the left and right sides of the upper body when the first appliance 11 is attached to the upper body. The actuator 14 is connected to the second orthosis 12 attached to the thigh via a connecting member 15 formed of a lightweight material such as a lightweight alloy, hard resin, or carbon fiber.

  Thereby, when the actuator 14 operates, a force acts on the upper body and each thigh so that the relative motion of the upper body and each thigh is assisted. The relative motion of the upper body and each thigh includes the longitudinal motion of the leg that is leaving the floor with respect to the upper body of the thigh, and includes the longitudinal motion of the upper body with respect to the leg that is landing.

  The battery 16 is housed in a case 13 attached to the back of the first appliance 11 together with the control device 20, and supplies power to the actuator 14, the control device 20, and the like. In addition, the arrangement | positioning location of each of the battery 16 and the control apparatus 20, or the case 13 which accommodates these may be changed suitably.

  The hip joint angle sensor 202 is composed of rotary encoders arranged on the left and right sides of the human waist, and outputs a signal corresponding to the hip joint angle. The hip joint angle is defined to be a positive value when the thigh is in front of the basic front face and to a negative value when the thigh is behind the basic front face.

  The control device 20 includes a computer (configured by a CPU, ROM, RAM, signal input circuit, signal output circuit, etc.) and software stored in the memory or storage device of the computer. The control device 20 controls the operation of the actuator 14 in addition to adjusting the power supplied from the battery 16 to the actuator 14.

  As shown in FIG. 2, the control device 20 performs a calculation process described later or performs a function, a motion oscillator output element 21, a fundamental frequency estimation element 23, a state oscillator generation element 24, And a control oscillator generating element 25. The fundamental frequency estimation element 23 constitutes the “walking state estimation device” of the present invention, and includes a differential transducer output element 231, a window processing element 232, a frequency analysis processing element 233, and a spectrum analysis processing element 234. Yes.

  Each component of the control device 20 is configured by an arithmetic processing device that reads a program stored in a storage device and executes arithmetic processing in charge according to the program. Each component may be constituted by a common arithmetic processing device or may be constituted by a physically separate arithmetic processing device. For example, the fundamental frequency estimation element 23 may be configured by an arithmetic processing device that is separate from the other components.

  When the operation switch (not shown) is operated and power is supplied from the battery 16 to the control device 20, the control device 20 can exhibit various functions.

(Function of walking assist device and walking state estimation device)
Based on the output of the hip joint angle sensor 202, the motion oscillator output element 21 generates a waveform signal representing a time change of the human left and right hip joint angles as a motion oscillator φ = (φ _L , φ _R ) (FIG. 3 / STEP 1). “Φ _L ” represents the left motion oscillator, and “φ _R ” represents the right motion oscillator.

Further, the fundamental frequency estimation element 23 performs a calculation process described later, thereby estimating a fundamental frequency f ₀ corresponding to the reciprocal of the human walking cycle (FIG. 3 / STEP 3).

Subsequently, the state oscillator generation element 24 inputs the motion oscillator output from the motion oscillator output element 21 as an input waveform signal to the state oscillator model defined by the simultaneous differential equations (010). Thus, the state oscillator ξ is generated as an output waveform signal (FIG. 3 / STEP 5). The state oscillator ξ includes a left bending oscillator ξ _{L +} , a left extension oscillator ξ _L− , a right bending oscillator ξ _{R +} and a right extension oscillator ξ _R− .

The “state oscillator model” includes a plurality of state variables u _i (i = L +, L−, R +, R−) representing the bending motion state and the stretching motion state of each thigh, and the bending motion state and stretching of each thigh. It is defined by simultaneous differential equations (010) of self-inhibiting factors v _i for expressing the adaptability of each motion state.

τ _1L (du _{L +} / dt) = c _{L +} −u _{L +} + w _{L + / L-} ξ _L- + w _{L + / R +} ξ _{R +} −λ _L v _{L +} + f ₁ (ω _L ) + f ₂ (ω _L ) Kφ _L ,
τ _1L (du _L- / dt) = c _L- -u _L- + w _{L- / L +} ξ _{L +} + w _{L- / R-} ξ _R- -λ _L v _L- + f ₁ (ω _L ) + f ₂ (ω _L ) Kφ _L ,
-τ _1R (du _{R +} / dt) = c _{R +} −u _{R +} + w _{R + / L +} ξ _{L +} + w _{R + / R-} ξ _{R +} −λ _R v _{R +} + f ₁ (ω _R ) + f ₂ (ω _R ) Kφ _R ,
τ _1R (du _R- / dt) = c _R- -u _R- + w _{R- / L-} ξ _L- + w _{R- / R +} ξ _{R +} -λ _R v _R- + f ₁ (ω _R ) + f ₂ (ω _R ) Kφ _R ,
τ _2i (dv _i / dt) = − v _i + ξ _i (i = L +, L-, R +, R-),
ξ _i = H (u _i −u _th ) = 0 (u _i <u _th ) .. (010).

“C _{L +} ” is a “left bending coefficient” that determines the magnitude of the amplitude of the state variable u _{L +} representing the bending operation state of the left thigh. “C _L− ” is a “left extension coefficient” that determines the magnitude of the amplitude of the state variable u _L− representing the extension operation state of the left thigh. “C _{R +} ” is a “right bending coefficient” that determines the magnitude of the amplitude of the state variable u _{R +} representing the bending operation state of the right thigh. “C _R− ” is a “right extension coefficient” that determines the magnitude of the amplitude of the state variable u _R− representing the extension operation state of the right thigh.

“Τ _1L ” is a first left time constant that determines the level of the frequency of the state variables u _{L +} and u _L− . “Τ _1R ” is a first right time constant that determines the frequency level of the state variables u _{R +} and u _R− . The first left time constant τ _1L and the first right time constant τ _1R may be the same value or different values. “Τ _2i ” is a second time constant that determines the frequency of the suppression factor v _i . The second time constant τ _2i may be the same value or a different value. For example, the second left bending time constant τ _{2L +} and the second left extension time constant τ _2L− are equivalent, and the second right bending time constant τ _{2R +} and the second right extension time constant τ _2R− are equivalent. The second left bending time constant τ _{2L +} and the second right bending time constant τ _{2R +} may be different values.

“W _{i / j} ” is a negative correlation coefficient representing the correlation between the state variables u _i and u _j . “Λ _L ” and “λ _R ” are habituation coefficients. “K” is a feedback coefficient corresponding to the motion oscillator φ.

“Ω” is a natural angular velocity determined according to the length of the human walking cycle. For example, the natural angular velocity ω is set to a value obtained by multiplying the fundamental frequency f ₀ (or the corrected fundamental frequency f _{0 — d} ) by 2π. The natural angular velocity ω includes a left component ω _L and a right component ω _R, but ω _L and ω _R have the same value because the fundamental frequency f ₀ has no difference between left and right.

“F ₁ ” is a linear function of the intrinsic angular velocity ω defined by the equation (011) using the positive coefficient c.

f ₁ (ω) = cω.. (011).

“F ₂ ” is a quadratic function of the intrinsic angular velocity ω defined by the equation (012) using the coefficients c ₀ , c ₁ and c ₂ .

f ₂ (ω) = c ₀ + c ₁ ω + c ₂ ω ² .. (012).

Based on the corrected fundamental frequency f _{0 — d} and the coefficients α ₀ (> 0) and β ₀ (> 0), the state oscillator generation element 24 sets the value of the first time constant τ ₁ according to the relational expression (023). Adjust or set.

τ ₁ = α ₀ / f _{0_d} + β ₀ .. (023).

The state oscillator generating element 24 adjusts the value of the second time constant τ ₂ according to the relational expression (024) based on the first time constant τ ₁ and the coefficient γ ₀ (> 0, for example, “2”). Set.

τ ₂ = γ ₀ τ ₁ .. (024).

The function defined by the relational expression (023) exhibits a change characteristic as shown in FIG. For this reason, the value of the first time constant τ ₁ and the second time constant τ ₂ is adjusted to decrease as the corrected basic frequency f 0 — _d becomes higher (the walking cycle becomes shorter). Further, the value of the first time constant τ ₁ and the second time constant τ ₂ is adjusted to increase as the corrected basic frequency f 0 — _d becomes lower (the walking cycle becomes longer).

Each coefficient c _i (i = L +, L−, R +, R−) is determined in advance during the period from the start of the operation of the walking assist device 10 until the motion oscillator necessary for evaluating the asymmetry s is obtained. Is set to the initial value. Similarly, the natural angular velocity ω is set to a predetermined initial value during a period from the start of the operation of the walking assist device 10 until a differential vibrator (described later) necessary for estimating the fundamental frequency f ₀ is obtained. .

  As the motion oscillator φ input to the state oscillator model, instead of the waveform signal representing the temporal change mode of the human left and right hip joint angles, the left and right hip joint angular velocities, the left and right shoulder joint angles, the left and right shoulder joint angular velocities, etc. A waveform signal representing a time change mode of a pair of left and right variables that change in a cycle closely related to the human walking cycle may be employed. Similarly to the hip joint angle, the shoulder joint angle can be measured based on the output of the shoulder joint angle sensor, in which a pair of shoulder joint angle sensors configured by a rotary encoder is arranged outside the left and right shoulders of a human.

State oscillator xi] _i, if the value of the state variable u _i is smaller than the threshold u _th 0, if the value of the state variable u _i is the threshold value u _th or more takes the value of the u _i (simultaneous differential equations (Refer to (010)). As a result, when the state variable u _{L +} representing the bending motion state of the left thigh increases, the amplitude of the left bending oscillator ξ _{L +} becomes larger than that of the left extension oscillator ξ _L− . Further, when the state variable u _{R +} representing the bending motion state of the right thigh increases, the amplitude of the right bending oscillator ξ _{R +} becomes larger than the amplitude of the right extension oscillator ξ _R− .

Further, when the state variable u _L− representing the extension motion state of the left thigh increases, the amplitude of the left extension oscillator ξ _L− becomes larger than that of the left bending oscillator ξ _{L +} . Further, when the state variable u _R− representing the extension motion state of the right thigh increases, the amplitude of the right extension oscillator ξ _R− becomes larger than the amplitude of the right bending oscillator ξ _{R +} .

As a result, according to the amplitude according to each value of the left bending coefficient c _{L +} and the left extension coefficient c _L− included in the simultaneous differential equation (010) and the frequency according to the value of the first left time constant τ _1L. A changing left bending oscillator ξ _{L +} and left extending oscillator ξ _L- are generated. Moreover, it changes according to the amplitude according to each value of the right bending coefficient c _{R +} and the right extension coefficient c _R− included in the simultaneous differential equation (010) and the frequency according to the value of the first right time constant τ _1R. The right bending oscillator ξ _{R +} and the right extension oscillator ξ _R- are generated.

Then, the control vibrator generation element 25 sets the control vibrator η = (η _L , η _R ) according to the relational expression (040) based on the state vibrator ξ (FIG. 3 / STEP 6).

η _L = χ _{L +} ξ _{L +} −χ _L- ξ _L- , η _R = χ _{R +} ξ _{R +} −χ _R- ξ _R- .. (040).

The left control oscillator η _L is calculated as a difference between the product of the left bending oscillator ξ _{L +} and the coefficient χ _{L + and} the product of the left extension oscillator ξ _L− and the coefficient χ _L− . The right control oscillator η _R is calculated as the difference between the product of the right bending oscillator ξ _{R +} and the coefficient χ _{R + and} the product of the right extension oscillator ξ _R- and the coefficient χ _R- . The four coefficients χ _i may be set to the same value.

Then, the current I = (I _L , I _R ) supplied from the battery 16 to the left and right actuators 14L, 14R is adjusted by the control device 20 based on the control vibrator η. As a result, the force assisting the flexion and extension movements of the left and right thighs with respect to the upper body or the rotational force F around the hip joint F = (F _L , F _R ) is adjusted via the first brace 11 and the second brace 12. . The auxiliary force F is based on the current I and is expressed as, for example, F (t) = G · I (t) (G: proportional coefficient). The agent's walking movement may be performed on a treadmill.

  Thereafter, it is determined whether or not an operation end condition such as the operation switch being switched from ON to OFF or an operation abnormality is detected is satisfied (FIG. 3 / STEP 7). If the determination result is negative (FIG. 3 / STEP7... NO), the series of processes is repeated. On the other hand, if the determination result is positive (FIG. 3 / STEP7. The process ends.

(Estimation of fundamental frequency)
The fundamental frequency f ₀ estimation process (see FIG. 3 / STEP 3) by the fundamental frequency estimation element 23 will be described.

  First, the differential transducer output element 231 outputs a waveform signal representing a temporal change in the difference in hip joint angle as a differential transducer based on the outputs of the left and right hip joint angle sensors 202R and 202L (FIG. 5 / STEP 31). Thereby, a waveform signal as shown in FIG. 6A is obtained as a differential vibrator. In addition, instead of the difference in the hip joint angle, a waveform signal representing a temporal change in the difference in the shoulder joint angle may be output as the difference vibrator.

  Subsequently, the window processing element 232 applies a window to the differential transducer over a specified period (FIG. 5 / STEP 32). A Hann window is used as the window function, whereby the differential transducer over the specified period indicated by the broken line in FIG. 6B is deformed as indicated by the solid line. As the window function, various window functions such as a rectangular window, a Gaussian window, a Hamming window, a Blackman window, a Kaiser window, a Bartlett window, or an exponential window can be used in addition to a Hann window.

The window processing element 232, based on the previous fundamental frequency f ₀ (k−1) determined by the spectrum analysis processing element 234, according to a decreasing function with the fundamental frequency f ₀ as a variable, the current window width w (k ) Is set. This reduction function is defined by, for example, the relational expression (050). The coefficients w ₀₁ and w ₀₂ are set so that at least two walking cycles are included. Note that the window processing element 232 employs a constant that includes at least two walking cycles as the initial window width w (0).

w (k) = w ₀₁ / f ₀ (k-1) + w ₀₂ .. (050).

  The window processing element 232 may perform window processing after removing a high frequency component exceeding the first designated frequency from the differential transducer by down-sampling the differential transducer (waveform signal). The window processing element 232 may perform the window processing after removing the low frequency component equal to or lower than the second designated frequency from the differential transducer by passing the differential transducer (waveform signal) through a high-pass filter.

  The frequency analysis processing element 233 creates a power spectrum by performing frequency analysis processing such as FFT on the differential transducer in the current designated period that has been windowed (FIG. 5 / STEP 33). As a result, as shown in FIG. 6C, a power spectrum is generated in which the frequency f is represented on the horizontal axis and the power as the frequency analysis result is represented on the vertical axis.

The spectrum analysis processing element 234 determines or estimates a frequency having a height equal to or higher than the threshold and representing the position of the peak in the lowest frequency band as the fundamental frequency f ₀ (FIG. 5 / STEP 34). The threshold is set to a value in the range of 0.05 to 0.20 times the peak maximum value in the power spectrum from the viewpoint of removing noise. Thus, as shown in FIG. 6C, the frequency representing the position of the peak P _{1 in} the lowest frequency band among the peaks P ₁ and P ₂ having a height _{equal to} or higher than the threshold value p _th is basically used. It is determined as the frequency f ₀ .

Furthermore, the corrected fundamental frequency f _{0 — d} is determined by correcting the fundamental frequency f ₀ according to the relational expression (060) representing the curve shown in FIG.

f _{0_d} = f ₀ (if f ₀ ≦ f _th ), f _th (if f ₀ > f _th ) .. (060).

(Operational effects of walking assist device and walking state estimation device)
According to the fundamental frequency estimation element 23 that performs the above function, a waveform signal (difference transducer) is obtained by sampling the “difference” between the human hip joint angle and shoulder joint angle (FIG. 5 / STEP 31, (See FIG. 6 (a)). Further, a power spectrum is obtained from this waveform signal (see FIG. 5 / STEP 33, FIG. 6C). For this reason, the fundamental frequency f ₀ representing the position of a single peak located in the lowest frequency band having a height equal to or higher than the threshold value in the power spectrum is referred to regardless of the level of asymmetry between the left and right body movements of the human. According to a certain standard, the human walking state can be estimated with high accuracy. Walking state, in addition to the estimated walking period which is the reciprocal of the fundamental frequency f _0, such as stride or walking rate, represented by any status value which can be calculated on the basis of the fundamental frequency f _0.

Further, as the previous fundamental frequency f ₀ (k−1) is higher, the width w (k) of the current window multiplied by the waveform signal in the current designated period is set narrower, while the previous fundamental frequency f ₀ (k−) is set. The lower the 1), the wider the window width w (k) is set this time. For this reason, from the viewpoint of estimating the current fundamental frequency f ₀ (k), a waveform signal having an appropriate width can be extracted as a frequency analysis target (see FIG. 5 / STEP 32, FIG. 6B). .

Furthermore, according to the walking assist device 10, the periodic motion of the human leg is assisted by the periodic motion of the actuator 14 (see FIG. 3 / STEP6). At this time, the operation period of the actuator 14 is determined based on the fundamental frequency f ₀ representing the estimated human walking state (see relational expressions (023) and (024)). For this reason, the human walking motion can be assisted with an appropriate period in view of the human walking state estimated with high accuracy based on the fundamental frequency f ₀ as described above.

(Other embodiments of the present invention)
The control device 20 performs the bending motion and the stretching motion of the human left thigh according to the magnitudes of the values of the left bending coefficient c _{L +} , the left stretching coefficient c _L− , the right bending coefficient c _{R +} and the right stretching coefficient c _R−. And if it is comprised so that the magnitude | size of the operation | movement amplitude of the actuator 14 for assisting each of bending | flexion movement and extension movement of a right thigh may be controlled, the structure is not limited to the structure of the said embodiment. The operation periods of the actuators 14L and 14R may be controlled to match 2π / ω _L and 2π / ω _R according to the natural angular velocities ω _L and ω _R.

For example, the control device 20 determines the values of the left flexion coefficient c _{L +} , the left stretch coefficient c _L− , the right flex coefficient c _{R +,} and the right stretch coefficient c _R− as the flexural target value of the left hip joint angle and the left hip joint angle. The left and right hip joint angles are controlled in accordance with a feedback control rule such as PID control as the extension side target value, the right hip joint angle flexion side target value, and the right hip joint angle extension side target value. Also good. In this case, the motion oscillator output element 21, the state oscillator generation element 24, and the control oscillator generation element 25 may be omitted.

As described in Japanese Patent No. 3930399, Japanese Patent No. 3950149, Japanese Patent No. 4008464, Japanese Patent No. 4271711, etc., the first vibrator according to the first model defined by Van der Pol equation etc. Is generated, and the natural angular velocity ω (see the simultaneous differential equation (010)) is set based on the first vibrator, and the time constant τ ₁ is set to be proportional to the inverse of the natural angular velocity ω. May be. The “second vibrator” in this publication corresponds to the “state vibrator” in the present invention.

  The control oscillator η may be generated so as to represent one or both of an elastic force due to a virtual elastic element and a damping force due to a virtual damping element, as described in Japanese Patent No. 4271711. Good.

DESCRIPTION OF SYMBOLS 10 ... Walking assistance device, 11 ... 1st equipment, 12 ... 2nd equipment, 14 ... Actuator, 20 ... Control device, 23 ... Fundamental frequency estimation element (walking state estimation device), 232 ... Window processing element, 233 ... Frequency analysis Processing element, 234. Spectrum analysis processing element.

Claims (4)

A window processing element configured to execute a window process for applying a window to a differential transducer, which is a waveform signal obtained by sampling over a specified period of a difference between the left and right hip joint angles or shoulder joint angles of a human being walking;
A frequency analysis processing element configured to obtain a power spectrum by performing frequency analysis on the differential transducer over which the window is hung, and
A spectral analysis processing element having a height equal to or higher than a threshold in the power spectrum and configured to determine a fundamental frequency indicating a peak located in a lowest frequency band,
The window processing element is configured to set the width of the current window based on the previous fundamental frequency determined by the spectrum analysis processing element according to a decreasing function having the fundamental frequency as a variable. A walking state estimation device as a feature. The walking state estimation device according to claim 1,
The window processing element is configured to perform the window processing after removing a high frequency component exceeding a first designated frequency from the differential transducer by down-sampling the differential transducer. A walking state estimation device characterized by the above. In the walking state estimation apparatus according to claim 1 or 2,
The window processing element is configured to execute the window processing after removing the low frequency component equal to or lower than a second specified frequency from the differential transducer by passing the differential transducer through a high-pass filter. A walking state estimation device characterized by that. A first orthosis attached to the upper body of a human, a pair of second orthosis attached to each of the left and right thighs of the human, a pair of actuators, and signals corresponding to the left and right hip joint angles of the human The left hip joint angle sensor and the right hip joint angle sensor configured to output the left hip joint angle sensor, and the operation of each of the pair of actuators based on at least the output signals of the left hip joint angle sensor and the right hip joint angle sensor. And a control device configured in
The human walking with relative periodic movement of the left and right thighs relative to the upper body by moving each of the pair of second braces with respect to the first brace by respective movements of the pair of actuators. A device for assisting exercise,
Using the differential vibrator, which is a waveform signal obtained by sampling over a specified period of the difference between the left and right hip joint angles or shoulder joint angles of the human through the left joint angle sensor and the right hip joint angle sensor, the fundamental frequency is obtained. It further comprises the walking state estimating device according to any one of claims 1 to 3 to be determined,
The walking assist device, wherein the control device is configured to control a periodic operation of the actuator according to a cycle determined according to the fundamental frequency determined by the walking state estimation device.

Images