KR20050097181A - Walking pattern analysis apparatus and method using inertial sensor - Google Patents

Abstract

The present invention relates to an apparatus and method for analyzing gait patterns, and more particularly, inertia useful for rehabilitation treatment and development of medical products by analyzing gait patterns by measuring a user's three-dimensional posture using an inertial sensor. An apparatus and method for analyzing walking patterns using a sensor are provided.

The present invention for this purpose is to measure the angular rotation speed and acceleration of the pedestrian based on the sensor axis (S10); Calculating the measured rotational angular velocity and converting the rotational rotation rate in the absolute coordinate system into Φg, θg and ψg (S20); Calculating rotation rate of change (Φg, θg, ψg) in the gravitational direction in the absolute coordinate system (S30); Integrating the rotation rate (Φg, θg, ψg) and subtracting the rotational change rate (Φa, θa, ψa) in the gravity direction from the integrated rotational change rate to correct the error so that the left and right angles when walking (Φm) (S40) Wow; Integrating the rotation rate (Φg, θg, ψg) and subtracting the rotational change rate (Φa, θa, ψa) in the gravity direction from the integrated rotational change rate to correct the error to calculate the forward and backward angles (θm) during walking Step S50; And outputting a walking pattern as a data signal in the time domain through the calculated left and right angles Φ m and the front and rear angles θ m.

Description

Walking pattern analysis apparatus and method using inertial sensor

The present invention relates to an apparatus and method for analyzing gait patterns, and more particularly, inertia useful for rehabilitation treatment and development of medical products by analyzing gait patterns by measuring a user's three-dimensional posture using an inertial sensor. An apparatus and method for analyzing walking patterns using a sensor are provided.

Research on gait analysis for the development of rehabilitation treatments and medical products for people with functional or pathological defects in the lower extremities has been underway for a long time in Korea. However, the existing gait analysis work can be performed only in a dedicated laboratory where numerous CCD cameras and equipment are installed, and there is a limitation of the place, and the inconvenience of patients having to take off their clothes and attach several markers to the body. There are many difficulties due to the lack of basic analysis data on characteristics.

In addition, the evaluation process of rehabilitation of various procedures performed in the lower extremities remains relatively inferior compared to the technical level in the medical field, and many doctors themselves point out the subjective tendency of the existing evaluation techniques. There is an increasing need to introduce more scientific and quantitative evaluation techniques.

Therefore, in recent years, the research on the inertial navigation system which calculates the user's posture, speed and moving distance by using inertial sensor including accelerometer and gyroscope has been accelerated. The gyro and accelerometer included in the inertial sensor used are the inertial sensors that measure the rotational angular velocity and the linear angular velocity, respectively. At this time, the user's posture is calculated by integrating the gyro output, and the user's speed and position are calculated by integrating the accelerometer output. However, since the inertial sensor includes an error such as a bias, the navigation information is emitted as the time is accumulated and the error is accumulated in the integral calculation process. In addition, in case of using the existing advanced inertial sensor, the error divergence with time is small because the error of the sensor is small, but the advanced inertial sensor used in the inertial navigation system is not only expensive but also difficult to purchase and bulky. Therefore, the field using this is being limited to military use, nautical use and the like.

Accordingly, with the recent development of technology, low-cost and small-sized semiconductor inertial sensors have been developed, and their application fields have also been expanded to private use. However, the performance is significantly lower than that of the existing advanced inertial sensor. Therefore, the navigation information with the desired accuracy cannot be obtained by the existing navigation method alone.In particular, when the inertial sensor is mounted on the human body and the posture is measured, the accuracy due to mounting error There is a problem that fall further.

Therefore, an object of the present invention devised to solve the above problems is to calculate the data on the walking characteristics when the inertial sensor is attached to the human body, and by calculating the pedestrian pattern based on the calculated data It is an object of the present invention to provide an analysis device for pedestrians using inertial sensors and a method for distinguishing abnormal walking and normal walking by numerically analyzing them.

The configuration of the present invention for achieving the above object, the inertial sensor unit for measuring the angular rotation speed and acceleration of the pedestrian; A transmitter for transmitting a measurement signal output from the inertial sensor unit; A receiver which receives a signal transmitted from the transmitter; An AC / DC converter converting the AC signal received by the receiver into a DC signal; The integral error is calculated by converting the measurement signal of the inertial sensor unit applied from the AC / DC converter into an absolute coordinate system signal, integrating the change rate of the converted rotational angular velocity, and subtracting the change rate of the acceleration from the integrated change rate. Comprising a calculation unit for calculating the walking pattern of the pedestrian.

Here, the inertial sensor is characterized in that at least one of a gyro sensor or an accelerometer.

In addition, the calculating unit multiplies the calculated integral error by the proportional gain (Kp) and the integral gain (Ki), so as to reduce the result from the rotation rate of change of the angular velocity, characterized in that for correcting the integral error.

Or, in the walking pattern analysis method for analyzing the walking pattern of the pedestrian, including a gyro sensor 11 for measuring the angular rotation speed of the pedestrian, and an accelerometer 12 for measuring the acceleration of the pedestrian, the pedestrian Measuring rotational angular velocity and acceleration of the pedestrian (S10); Calculating a rotation angular velocity measured in the step S10 and converting the rotation angular velocity into a rotation change rate Φg, θg, ψg in an absolute coordinate system (S20); Calculating a rate of change of rotation (Φg, θg, ψg) in the gravitational direction in the absolute coordinate system through the acceleration measured in the step (S10) (S30); Integrate the rotational change rate Φg, θg, ψg calculated in the step S20, and correct the error by subtracting the rotational change rate Φa, θa, ψa in the gravity direction from the integrated rotational change rate. Left and right angle Φ m (S40); Integrate the rotational change rate Φg, θg, ψg calculated in the step S20, and correct the error by subtracting the rotational change rate Φa, θa, ψa in the gravity direction from the integrated rotational change rate. Calculating a front and rear angle θm; And outputting a walking pattern as a data signal in a time domain through the left and right angles Φ m and the front and rear angles θ m calculated in the left and right angle calculation step S40 and the front and rear angle calculation step S50. do.

The method further includes the step of obtaining the walking pattern data by converting the extracted left and right and front and rear angle data into a signal in the frequency domain (S60), and outputting the obtained data (S70).

In addition, the step of calculating the left and right angles is integrated with the rotational change rate, and calculating the integral error by subtracting the rotational change rate component in the gravity direction from the integrated result; Multiplying the integral error calculated in the above step by the integral gain Ki and integrating the resultant value; The step of multiplying the integrated error and the proportional gain (Kp), calculating the accumulated error value of the combined operation of the integrated error and the integral gain (Ki) calculated at the step and that the resultant value; And calculating the left and right angles or the front and rear angles by subtracting the accumulated error calculated in the step from the rotational change rate and integrating it.

The step of calculating the front and rear angles may include: integrating a rotational change rate and calculating an integration error by subtracting the rotational change rate component in the gravity direction from the integrated result; Multiplying the integral error calculated in the above step by the integral gain Ki and integrating the resultant value; The step of multiplying the integrated error and the proportional gain (Kp), calculating the accumulated error value of the combined operation of the integrated error and the integral gain (Ki) calculated at the step and that the resultant value; And calculating the left and right angles or the front and rear angles by subtracting the accumulated error calculated in the step from the rotational change rate and integrating it.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram showing a walking pattern analysis apparatus using an inertial sensor according to the present invention.

As shown, the gait pattern analysis apparatus according to the present invention is an inertial sensor 10 including a gyro sensor 11 for measuring the rotational angular velocity of the pedestrian walking, and an accelerometer 12 for measuring the acceleration of the pedestrian; In addition, the transmitter 20 transmits the measurement value measured by the inertial sensor 10, the receiver 30 receives the signal measured by the transmitter 20, and the analog signal transmitted from the receiver 30 digital An arithmetic unit that calculates a walking pattern of a pedestrian based on the measured signals of the A / D converter 40 and the gyro sensor 11 and the accelerometer 12 transmitted from the A / D converter 40. It consists of 50.

Preferably, the operation unit 50 is also configured to install a walking pattern control method according to the present invention on a personal PC, more preferably, the receiving unit 30 and the A / D converter in the personal PC Connecting 40 to the personal PC also corresponds to the gist of the present invention.

The gyro sensor 11 and the accelerometer 12 are attached to a part of the body of the pedestrian to measure each rotational angular velocity and acceleration when the pedestrian is walking (see FIG. 4). That is, the gyro sensor 11 measures the rotational angular velocity during walking on the basis of the sensor axis, and the accelerometer 12 measures the acceleration on the basis of the sensor axis.

The measurement signal measured by the inertial sensor unit 10 is transmitted to the transmitter 20, and the transmitter 20 transmits it to the receiver 30. The receiver 30 applies the received signal to the A / D converter 40, and the A / D converter 40 converts the measured signal, which is an analog signal, into a digital signal and applies it to the calculator 50.

Therefore, the calculation unit 50 converts the rotational angular velocity rotated on the basis of the sensor axis measured by the gyro sensor 11 and the acceleration based on the sensor axis measured by the accelerometer 12 into an absolute coordinate system according to the walking pattern method. The gait pattern is extracted by converting the data into an angular dimension and calculating data of left, right, and front and rear rotational angles during walking. Here, since the signals of the accelerometer 12 operate on the absolute coordinate system, data on the walking pattern of the pedestrian is calculated immediately after calculating the rotational change rate of the acceleration through a predetermined calculation process.

However, the gyro sensor 11 detects the rotational angle by integrating the measured rotational angular velocity as the rotational angular velocity is detected, but an integration error is generated at the time of integration, and the walking pattern data of the accelerometer 12 is the gyro sensor 11. Compared with this, the error is not emitted but the response is slow.

Therefore, the calculation unit 50 of the present invention corrects the integration error of the gyro sensor 11 by calculating the mixed angular velocity measured by the gyro sensor 11 and the measurement signal of the accelerometer 12 to calculate the walking pattern quickly. Data can be converted. Details of this will be described in detail in the following description of the walking pattern analysis method.

2 is a block diagram showing a walking pattern analysis method using an inertial sensor according to the present invention.

As shown, the gait pattern analysis method according to the present invention calculates the rotational change rate in the absolute coordinate system by calculating the rotational angular velocity measured by the gyro sensor 11 and the accelerometer 12 The cumulative error is calculated using the result of the gravity direction rotation change rate calculating means 52 and the rotation direction change rate calculating means 52 and the rotation change rate calculating means 51 which calculate the rotation change rate in the gravity direction by calculating the measurement signal. Calculate and correct the cumulative error by using the roll filter 53 for calculating the left and right angles by calculating and correcting, and the resultant values of the rotation direction change rate calculating means 52 and the rotation rate change calculating means 51. A pitch filter 54 for calculating the angle.

Detailed description thereof will be described using a flowchart illustrating a method of analyzing a walking pattern using the inertial sensor of FIG. 3 and a coordinate system of FIG. 4.

First, in the gyro sensor, as shown in FIGS. 4A and 4B, the sensor axis is set as the X-axis, the right-side Y-axis, and the lower-direction Z-axis direction of the pedestrian, and the origin P is the center of gravity of the human body. . Therefore, the gyro sensor 11 measures the rotational angular velocity during walking based on the X-axis, Y-axis, and Z-axis when the pedestrian walks. The accelerometer 12 measures the acceleration of the pedestrian based on the sensor axes of the X and Y axes (S10).

Here, the rotational motion about the X axis measured by the gyro sensor 11 and the accelerometer 12 is roll (φ), the rotational motion about the Y axis is rotated about the pitch (θ), Z axis This motion is called yaw (Ψ) and the direction defines the rotation direction shown in FIG. 4 in the + direction. The roll (φ), pitch (θ) and yaw (Ψ) are represented by the dimension of the angle, and the unit is deg or rad. The angles φ, θ, and Ψ are measured in total, and the measured values output from the gyro sensor 11 are the rates of change (p, q, r, and time with respect to the time of roll, pitch, yaw (φ, θ, Ψ). Equation 1).

Here, the rotational change rate (p, q, r, Equation 1) with respect to the time of the pedestrian roll, pitch, yaw (φ, θ, Ψ) measured by the gyro sensor 11 is measured based on the sensor axis. You can measure the change in how many degrees are wrong compared to the previous posture, but it is not known how many degrees your current posture is. This is because there is no absolute reference value because it is measured using a sensor axis set arbitrarily. Therefore, the rotation change rate calculating means 51 converts the measurement value of the gyro sensor 11 based on the sensor axis as described above into an absolute coordinate system having an absolute reference.

That is, the measurement value based on the sensor axis measured by the inertial sensor 10 is applied to the calculation unit 50 through the transmitter 20, the receiver 30, and the A / D converter 40. In 50), the rotation change rate is calculated by converting the measured value based on the sensor axis into an absolute coordinate system (S20).

First, the rotational angular velocities of the X, Y, and Z axes around the sensor axis measured by the gyro sensor 11 are transmitted to the rotation change rate calculating means 51. Therefore, the rotation change rate calculating means 51 measures the rotation change rate in the absolute coordinate system by Equation 1 below.

p, q and r are the rotational angular velocities of the roll, pitch and yaw relative to the sensor axis, Φ g, θ g and ψ g are the rate of change of the roll, pitch and yaw in the absolute coordinate system, and Φ m and θ m are the rotation of the sensor axis in the absolute coordinate system. As an angle, it means the angle between the roll and the pitch in the previous position.

Equation 1 is the rate of change of the rotation angle of the absolute coordinate system (Φg, θg, using the rotation angle velocity (p, q, r) around the X, Y, Z axis with respect to the sensor axis and the rotation angle in the absolute coordinate system ψg) is obtained. That is, the gyro sensor 11 mounted on a part of the body measures rotational angular velocities (p, q, r) of X, Y, and Z axes which are changed when the pedestrian walks with the sensor as the central axis, and the rotation The change rate calculating means 51 converts the rotational angular velocities p, q and r around the sensor axis into the rotation change rates Φg, θg and ψg in the absolute coordinate system through Equation 1 above.

In this case, the rotation change rate calculating means 51 is an angle rotated about the X and Y axes of the previous posture calculated by the roll filter 53 and the pitch filter 54, that is, the roll and the pitch φm and θm. Is applied to Equation 1 to calculate the rate of change of rotation in the absolute coordinate system, and calculates the rate of change of the rotation angle (Φg, θg, ψg) as described below. Pitch Filter). Here, the yaw (Ψm), which is the angle of the Z axis in the absolute coordinate system in the previous posture, becomes 0 as the pedestrian assumes the linear movement without the vertical movement.

In addition, the gravity direction rotation change rate calculating unit 52 calculates the rotation change rate in the gravity direction by the following equation 2 when the measurement value of the linear acceleration of the pedestrian is transmitted from the accelerometer 12 (S30).

Where Φa and θa are the rotational rate of change in the direction of gravity in the X and Y axes in the absolute coordinate system, ax and ay are the accelerometer accelerations, fx and fy are the acceleration components of the X and Y axes relative to the sensor axis, and g is Gravity acceleration.

In detail, the accelerometer 12 is a sensor for measuring linear acceleration of an object. The accelerometer 12 includes acceleration components due to the movement of the object, gravitational acceleration of the earth, acceleration components due to centrifugal force, and the Coriolis effect. By correcting each component by the output only the pure acceleration component by the motion of the object.

Therefore, the gravity direction rotation change rate calculating means 52 assumes that the earth's gravity acceleration is perpendicular to the Z axis in the absolute coordinate system, and the rotational change rates (Φa, θa) of the gravity accelerations are based on the sensor axis with reference to the sensor axis. Calculate from the acceleration components (ax, ay) of the X and Y axes. Here, θm means the front and rear angle in the previous posture, that is, the pitch angle, and the front and rear angle θm of the previous posture measured by the pitch filter 54 is fed back so that the gravitational direction rotation change rate calculating means 52 Is applied to.

As described above, the rotation change rates Φg, θg, ψg and the rotation change rates Φa, θa calculated by the rotation change rate calculating means 51 and the gravity direction rotation change rate calculating means 52 are respectively roll filters ( 53) and the pitch filter 54, and the roll filter 53 and the pitch filter 54 mix the applied rotation rate of change (Φg, θg, ψg) with the rotational rate of change (Φa, θa) of gravity acceleration, respectively. It calculates the left and right angles and the front and rear angles (S40, S50), a detailed description thereof will be described in detail with reference to Figs.

5 is a block diagram illustrating a method of calculating a roll filter or a pitch filter, and FIG. 6 is a flowchart illustrating a method of calculating a roll filter or a pitch filter.

As shown, the rotational change rate (Φg, θg, ψg) output from the rotational change rate calculating means 51 is integrated (S41, S42), and the integrated rotational change rate is output from the gravity direction rotational change rate calculating means 52. It is mixed with the rotation rate of change (Φa, θa) in the gravity direction (S43). At this time, the roll filter 53 calculates the roll angle, that is, the left and right angles Φ m, and the pitch filter 54 calculates the pitch, that is, the front and rear angles θ m. Therefore, the components of the rotational change rate (Φg, θg, ψg) and the gravity direction rotational change rate (Φa, θa) input to the roll filter 53 input only the rotational change rate around the X axis, that is, the roll (φg, φa,) components. The components input to the pitch filter 54 are input only the components of the pitch (θg, θa) around the Y axis, and the rotation rate of change (Ψg, Ψa) around the Z axis is considered to be a linear movement of the pedestrian. Becomes zero.

Then, the rotational change rate (Φg, θg, ψg) in the absolute coordinate system of the gyro sensor 11 integrated in the step (S42) is reduced by the rotational change rate (Φa, θa, ψa) in the gravity direction of the accelerometer 12. Compute the error (S44).

The calculated integral error is multiplied by the proportional gain Kp and the integral gain Ki, respectively (S45 and S47). In addition, the multiplication result of the integration error and the integral gain Ki is integrated again to transfer the result value to the next step S48 (S46).

Subsequently, a cumulative error is calculated by adding the multiplication value of the error and the integral gain Ki calculated in the steps S45 and S47 and the multiplication value of the proportional gain Kp and the integration error (S48). The cumulative error is a value close to zero that is output even when there is no physical quantity input to the sensor as a bias.

The gyro sensor 11 obtains an angular component through integration as the rotational angular velocity is measured, but when the measured value is integrated, the integrated result is infinitely diverged due to the error value caused by the bias, thus making it impossible to construct a precise sensor. do. However, the present invention calculates the error due to the integral gain (Ki) and the proportional gain (Kp) by calculating the measurement component of the gyro and the measurement component of the accelerometer, and corrects the error through the following correction step (S49) to precise walking pattern Can be measured.

In other words, the cumulative error calculated in the step S48 is corrected by reducing the error of the gyro sensor 11 by reducing the rotation change rates Φg, θg, ψg of the gyro sensor 11 (S49). And the rotational change rate (Φg, θg, ψg) of the corrected gyro sensor 11 is integrated again to calculate the left and right angles (Φm) or the front and rear angles (θm) when walking the pedestrian by changing the components of the angular velocity to output the angle. And outputs it (S51, S52, S53).

Afterwards, referring to FIG. 3 again, the operation is performed in the calculation steps (S40, S50, FIG. 6) of the left and right angles Φ m or the front and rear angles θm of the roll filter 53 or the pitch filter 54 as described above. The left and right angle Φ m and the front and rear angle θ m, that is, the data of the roll and the pitch, are converted into data in the frequency domain through the Fourier changing means 57 and output through the output means (not shown) (S60). (S70).

To describe the step (S60) in detail, the data output from the roll filter 53 and the pitch filter 54 is output through the output means (not shown) as the data of the time domain or according to the user's intention Frequency domain data obtained from the roll filter and the pitch filter by the Fourier transform means 57 is converted into a frequency domain through a Fast Fourier Transform (FFT) based on a Discrete Fourier Transform (DFT). Is achieved through Equation 3 below.

N is the number of discrete data, ｗ _N is the frequency when there are N discrete data.

A walking pattern of a pedestrian displayed as data of the frequency domain achieved through Equation 3 is as shown in FIG.

7A is a graph illustrating a walking pattern of a normal person in the time domain, and FIG. 7B is a graph illustrating a walking pattern of a normal person in the frequency domain. In addition, Figure 7c is a graph showing the walking pattern of the disabled in the time domain, Figure 7d is a graph showing the walking pattern of the disabled in the frequency domain.

As shown, the graph of the time domain linearly shows a change in rotation angle when the pedestrian walks for a certain time. That is, in the case of a normal person, the width of the maximum and minimum values of the angle rotated for a predetermined time is narrow, and in the case of an abnormal person, the width of the maximum and minimum values of the angle rotated for a certain time is large.

In addition, the graph of the frequency domain shows that the frequency component is clean in the case of a normal person, but in the case of an abnormal person, the frequency component is relatively much larger and its size is also very large.

Through the gait pattern graph in the time domain and the frequency domain, it can be inferred that a normal person has very little displacement of the center of gravity because the activity of each joint and muscle is active so that the center of gravity of the body is not shaken while walking. . On the other hand, if the lower limbs are defective, the joints of the lower limbs or the limbs can be interpreted as inflexible and inactive or inactive.

While the invention has been shown and described with respect to certain preferred embodiments, it will be appreciated that the invention can be modified and modified in various ways without departing from the spirit or scope of the invention as set forth in the claims below. One of ordinary skill in the art will readily know.

The pedestrian walking pattern analysis apparatus and method according to the present invention as described above, by properly mixing the advantages of the gyro sensor and the accelerometer, it is possible to correct the error of the conventional gyro sensor to accurately and accurately analyze the walking pattern, and also In the present invention, the present invention is performed only in a dedicated facility equipped with a CCD camera and equipment. However, the present invention is not limited to a place because it is equipped with a simple configuration using a wireless system. By analyzing, it is possible to establish a database on the type and extent of abnormal walking or disability, and through the established database, it is possible to objectively and quantitatively evaluate and diagnose the results of rehabilitation or the degree of disability. It works.

1 is a block diagram showing a walking pattern analysis apparatus using an inertial sensor according to the present invention;

Figure 2 is a block diagram showing the configuration of a walking pattern analysis method using an inertial sensor according to the present invention,

3 is a flowchart illustrating a method of analyzing a walking pattern using an inertial sensor according to the present invention;

4A to 4B are exemplary views showing coordinates based on a sensor axis of an inertial sensor;

5 is a block diagram showing a calculation method of a calculation filter according to the present invention;

6 is a flowchart illustrating a calculation method of an operation filter according to the present invention;

7A to 7D are graphs showing walking patterns through data obtained according to the present invention.

<Description of the symbols for the main parts of the drawings>

10: inertial sensor 11: gyro sensor

12 accelerometer 20 transmitter

30: receiver 40: A / D converter

50: calculation unit 51: rotation change rate calculation means

52: gravity direction rotation rate calculation means 53: roll filter

54: pitch filter 57: Fourier conversion means

Claims (7)

Inertial sensor unit 10 for measuring the angular rotation speed and acceleration of the pedestrian; A transmitter 20 for transmitting a measurement signal output from the inertial sensor unit; A receiver 30 for receiving a signal transmitted from the transmitter; An AC / DC converter 40 for converting an AC signal received by the receiver into a DC signal; The measurement signal of the inertial sensor unit 10 applied from the AC / DC converter 40 is converted into a signal of an absolute coordinate system, and the rate of change of the converted rotational angular velocity ((Φg, θg, ψg) is integrated and integrated. And a calculation unit (50) which calculates and corrects an integration error by subtracting the change rate (Φa, θa, ψa) of the acceleration from the changed change rate. The method of claim 1, wherein the inertial sensor 10 Gait pattern analysis device characterized in that at least one of the gyro sensor (11) or the accelerometer (12). The method of claim 1 or 2, wherein the calculation unit 50 Multiply the calculated integral error by the proportional gain (Kp) and the integral gain (Ki), and add the resultant value to reduce the rotation error rate (Φg, θg, ψg), thereby correcting the integral error. Pattern analysis device. In the walking pattern analysis method for analyzing the walking pattern of a pedestrian, including a gyro sensor 11 for measuring the angular rotation speed of the pedestrian and an accelerometer 12 for measuring the acceleration of the pedestrian, Measuring a rotation angle of the pedestrian and an acceleration of the pedestrian based on the sensor axis (S10); Calculating a rotation angular velocity measured in the step S10 and converting the rotation angular velocity into a rotation change rate Φg, θg, ψg in an absolute coordinate system (S20); Calculating a rate of change of rotation (Φg, θg, ψg) in the gravitational direction in the absolute coordinate system through the acceleration measured in the step (S10) (S30); Integrate the rotational change rate Φg, θg, ψg calculated in the step S20, and correct the error by subtracting the rotational change rate Φa, θa, ψa in the gravity direction from the integrated rotational change rate. Left and right angle Φ m (S40); Integrate the rotational change rate Φg, θg, ψg calculated in the step S20, and correct the error by subtracting the rotational change rate Φa, θa, ψa in the gravity direction from the integrated rotational change rate. Calculating a front and rear angle θm; And outputting a walking pattern as a data signal in a time domain through the left and right angles Φ m and the front and rear angles θ m calculated in the left and right angle calculation step S40 and the front and rear angle calculation step S50. Analysis method of walking pattern. The method of claim 4, wherein The method further includes a step (S60) of obtaining the walking pattern data by converting the extracted left and right and front and rear angle data into a signal in the frequency domain, and outputting the obtained data (S70). . The method of claim 4, wherein the right and left angle calculation step (S40) Integrating a rotation rate of change (S42) and calculating an integration error by subtracting the rotation rate of change component in the gravity direction from the integrated result value (S44); Multiplying the integral error (Kp) and the integral gain (Ki) calculated in the step (S44) (S45), and integrating the resultant value (S46); Multiplied by the integrated error and the proportional gain (Kp) (S47), the results and step (S48) for computing a combined accumulated error value operation of the integrated error and the integral gain (Ki) calculated at the step (S46) and; And subtracting the cumulative error calculated in the step S48 from the rotational change rates Φg, θg, ψg (S49), and integrating the calculated errors (S51, S52). Analysis method of walking pattern. The method of claim 4, wherein the front and rear angle calculation step (S50) Integrating a rotation rate of change (S42) and calculating an integration error by subtracting the rotation rate of change component in the gravity direction from the integrated result value (S44); Multiplying the integral error (Kp) and the integral gain (Ki) calculated in the step (S44) (S45), and integrating the resultant value (S46); Multiplied by the integrated error and the proportional gain (Kp) (S47), the results and step (S48) for computing a combined accumulated error value operation of the integrated error and the integral gain (Ki) calculated at the step (S46) and; And subtracting the cumulative error calculated in the step S48 from the rotational change rates Φg, θg, ψg (S49), and integrating the calculated errors (S51, S52). Analysis method of walking pattern.