KR101503046B1 - inertial measurement unit and method for calibrating the same - Google Patents

Abstract

The present invention relates to a method of calibrating a multi-axis sensing device. The calibration method of the multi-axis sensing device of the present invention automatically performs calibration with respect to an acceleration sensor or a geomagnetic sensor of a multi-axis sensing device and uses the measured values of the calibrated acceleration sensor or geomagnetic sensor to calibrate the gyro sensor The calibration of the multi-axis sensing device can be performed automatically without a separate calibration device.

Description

[0001] The present invention relates to a multi-axis sensing device and a calibrating method thereof,

The present invention relates to a multi-axis sensing device.

Recently, an inertial measurement unit (IMU) has been adopted in various fields such as smart phones, portable game machines, remote controllers, and unmanned reconnaissance machines. The multi-axis sensing device is for measuring a position or direction in an arbitrary direction, and may include an acceleration sensor, a geomagnetic sensor, and a gyroscope sensor.

FIG. 1 is a view schematically showing an example of a multi-axis sensing device. 1, the multi-axis sensing device 1 includes a printed circuit board 40 and an acceleration sensor 10, a geomagnetic sensor 20, and a gyro sensor 30 in the form of an integrated circuit mounted thereon.

In the case of the multi-axis sensing device 1 including various kinds of sensors as described above, it is necessary to calibrate a plurality of sensors in order to measure an accurate attitude value. In the multi-axis sensing device 1, As a method for calibrating, a special calibration device for arbitrarily changing the orientation of the multi-axis sensing device 1 may be used.

2 is a view schematically showing a calibration apparatus for arbitrarily changing the orientation of the multi-axis sensing apparatus.

Referring to FIG. 2, the multi-axis sensing device 1 is mounted on the calibration device 5 and the orientation of the multi-axis sensing device 1 is arbitrarily changed by the calibration device 5. The calibration device 5 outputs information on the attitude of the multi-axis sensing device 1 mounted thereon, so that it can calibrate each sensor of the multi-axis sensing device 1 using the true value.

However, since such a method requires a separate calibration device, there is a problem that it is difficult to apply to situations requiring calibration from time to time or requiring calibration in real time.

Also, there is a method of calibrating by taking into consideration the characteristics of each sensor without a separate calibration device. Such a calibration method is vulnerable to noise and perturbation, and the calibration method of the geomagnetic sensor and the acceleration sensor is different The computation of the algorithm becomes complicated. Further, there is a problem that a method of using the measurement information of the acceleration sensor and the geomagnetic sensor can not be provided for the gyro sensor.

In order to solve the above problems, one aspect of the present invention is to provide a method for integrally and automatically correcting a plurality of sensors included in a multi-axis sensing apparatus without a separate calibration apparatus, and a multi-axis sensing apparatus having such a calibration function .

An automatic calibration method of a multi-axis sensing device according to an embodiment of the present invention is an automatic calibration method of an multi-axis sensing device including an acceleration sensor or a geomagnetic sensor and a gyro sensor,

(a) acquiring a measured value of the acceleration sensor or the geodetic sensor as the multi-axis sensing device rotates in an arbitrary direction;

(b) calculating an internal calibration value of the acceleration sensor or the geomagnetic sensor;

(c) calibrating a measurement value of the acceleration sensor or the geomagnetic sensor using an internal calibration value of the acceleration sensor or the geomagnetic sensor;

(d) calculating an internal calibration value of the gyro sensor using the measured value of the acceleration sensor or the geomagnetic sensor.

Also, the step (b) may include performing ellipsoid fitting using the measured values of the acceleration sensor.

The step (b) may further include performing the ellipsoid fitting again, excluding the outlier at the time of fitting the ellipsoid.

Also, the step (d) may include calculating a scale factor of the gyro sensor using the gravity vector measured by the acceleration sensor in accordance with the rotation of the multi-axis sensing device.

The step (d) may further include the step of calculating a bias value of the gyro sensor using the measured value of the calibrated acceleration sensor for a predetermined time and the frequency domain analysis of the measured value of the gyro sensor .

The automatic calibration method of the multi-axis sensing device may further include an external calibration step of calculating a conversion relationship between the coordinate system of the acceleration sensor or the coordinate system of the gyro sensor.

Also, the step (b) may include performing ellipsoid fitting using the measured values of the geomagnetic sensor.

The step (b) may further include performing the ellipsoid fitting again, excluding the outlier at the time of fitting the ellipsoid.

Also, the step (d) may include calculating a scale factor of the gyro sensor using the gravity vector measured by the geomagnetic sensor according to the rotation of the multi-axis sensing device.

The step (d) may further include calculating a bias value of the gyro sensor using the measured value of the geomagnetic sensor and the frequency domain analysis of the measured value of the gyro sensor for a predetermined period of time .

The automatic calibration method of the multi-axis sensing device may further include an external calibration step of calculating a conversion relationship between the coordinate system of the geomagnetic sensor or the coordinate system of the gyro sensor.

The automatic calibration method of the multi-axis sensing apparatus may further include an external calibration step of calculating a conversion relationship between the coordinate system of the acceleration sensor, the geodetic sensor coordinate system, or the coordinate system of the gyro sensor.

Also, the step (b) may be a step of calculating the internal calibration value of the geomagnetic sensor using the geomagnetism vector value measured by rotating the multi-axis sensing device in one direction and its opposite direction.

An automatic calibration method for a multi-axis sensing apparatus according to another embodiment of the present invention is a method for automatically calibrating a multi-axis sensing apparatus including a gyro sensor, which tracks a reference object using a camera fixed in position relative to the multi- Calculating the internal calibration value of the gyro sensor using the relative positional information of the camera with respect to the reference object and the measurement value of the gyro sensor, calculating the relative position information of the camera with respect to the reference object, .

Also, the automatic calibration method of the multi-axis sensing apparatus may include a step of calculating a bias value of the gyro sensor by using the relative position information of the camera with respect to the reference object for a predetermined time and the frequency domain analysis of the measured value of the gyro sensor And a step of calculating

According to another embodiment of the present invention, there is provided a program for performing the method of any one of the above, and a computer readable medium may be provided.

According to another aspect of the present invention, there is provided a multi-axis sensing device including a body, an acceleration sensor or a geomagnetic sensor relatively fixed to the body, a gyro sensor relatively fixed to the body, An internal calibration unit for calculating an internal calibration value of the acceleration sensor or the geomagnetic sensor by using the measured value of the acceleration sensor or the calibration value of the geomagnetic sensor and the measured value of the gyro sensor And a gyro sensor calibration unit for calculating an internal calibration value of the gyro sensor using the gyro sensor.

The multi-axis sensing device may further include an external calibration unit for calculating a coordinate value between the coordinate system of the acceleration sensor or the coordinate system of the geomagnetic sensor and the coordinate system of the gyro sensor.

The internal calibration unit may include an ellipsoidal fitting unit for performing elliptical fitting using the measured values of the acceleration sensor or the geomagnetic sensor.

In addition, the multi-axis sensing device may further include a camera fixed relative to the body.

According to the calibration method of the multi-axis sensing device according to one aspect of the present invention, a plurality of sensors included in the multi-axis sensing device can be integrally and automatically calibrated without a separate calibration device.

FIG. 1 is a view schematically showing an example of a multi-axis sensing device.
2 is a schematic view of a conventional calibration device for calibrating each sensor of a multi-axis sensing device.
3A to 3C are views schematically showing a sensor coordinate system of an acceleration sensor, a ground magnetic sensor and a gyro sensor, respectively.
4 is a schematic block diagram of a multi-axis sensing apparatus according to an embodiment of the present invention.
5 is a schematic configuration diagram of an internal calibration unit of the multi-axis sensing apparatus of FIG.
6 is a view schematically showing a relationship between sensor coordinate systems of a multi-axis sensing device.
Fig. 7 is a flowchart schematically illustrating a method of automatically calibrating the inertia measurement unit of Fig. 4; Fig.
Fig. 8 is a view schematically showing performing elliptical fitting.
9 is a view schematically showing a result of performing ellipse fitting.
10 is a view schematically showing an image photographed using a camera.

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

The multi-axis sensing device includes an acceleration sensor and a gyro sensor for measuring the inertia, and may further include a geomagnetic sensor for sensing the geomagnetism and measuring the absolute orientation. Various types of sensors provided in the multi-axis sensing device have respective reference coordinate systems and output measurement values of their relative orientations with reference to their reference coordinate system.

3A to 3C show the coordinate system of the acceleration sensor 110, the geomagnetic sensor 120 and the gyro sensor 130, that is, the sensor coordinate system. As shown in FIGS. 3A to 3C, each sensor measures different physical quantities and coordinate axes can be set in different directions.

In addition, these sensors are mounted on the body of the multi-axis sensing device, mainly a printed circuit board, and the sensor coordinate system of each sensor can be configured differently with a predetermined conversion relationship not coinciding with the coordinate system of the body of the multi-

On the other hand, calibration work for sensors is divided into intrinsic calibration and extrinsic calibration.

Internal calibration refers to calibrating the output value of each sensor itself, and it mainly corresponds to the process of calculating parameters for calibrating the raw data.

If a sensor has n axes, the sensor can be mathematically modeled as follows.

는 교정 이전의 로우(raw) 데이터 측정값(measurment raw vector)이며,

은 교정 이후의 원하는 결과값(calibrated vector)이다. 또한 S는 n×n 스케일 팩터 행렬(n×n scale factor matrix)로서 측정값과 교정값 사이의 비율 관계를 정하는 값이며,

은 측정값과 교정값 사이의 편향의 정도를 나타내는 바이어스 벡터(bias vector)이다. here

Is a raw data measure before calibration,

Is the desired calibrated vector after calibration. Also, S is an n × n scale factor matrix (n × n scale factor matrix) which defines a ratio relationship between measured values and calibration values,

Is a bias vector indicating the degree of deviation between the measured value and the calibration value.

External calibration refers to calculating the relationship between a plurality of sensor coordinate systems, and is mainly a process of calculating a transformation matrix between sensor coordinate systems. Since the coordinate systems of the sensors do not exactly coincide with each other, an accurate conversion relationship between the sensor coordinate systems can be calculated through such external calibration. Therefore, reliability of each sensor can be further improved through external calibration.

The external calibration can be calculated using the following equation (2) used for the hand-eye calibration of the robot.

Where A and B are sensor measurements and can be a special orthogonal group or a special Euclidean group SE.

Based on the above technical background, a multi-axis sensing apparatus according to an embodiment of the present invention and a multi-axis sensing method using the same will be described.

4 is a block diagram schematically illustrating the configuration of a multi-axis sensing apparatus according to an embodiment of the present invention.

4, the multi-axis sensing device 2 of the present embodiment includes a body (not shown), a sensor unit 100, an internal calibration unit 200, a gyro sensor calibration unit 300, and an external calibration unit 400 And may be electrically connected to the camera 500 whose position is relatively fixed to the multi-axis sensing device 2. [ The internal calibration unit 200, the gyroscopic sensor calibration unit 300, and the external calibration unit 400 may be a plurality of physically divided calculation units, that is, a plurality of microprocessors. However, Or may be implemented in the form of software that performs the functions. The computing device corresponding to the internal calibration unit 200, the gyro sensor calibration unit 300 or the external calibration unit 400 may be built in the body of the multi-axis sensing device 2, 2 may be separately provided separately from the body. Even when the computing device is provided separately from the multi-axis sensing device 2 and the body, the computing device and the multi-axis sensing device 2 are connected by wire or wireless communication.

The sensor unit 100 includes an acceleration sensor 110 and a gyro sensor 130 corresponding to the inertial measurement sensor and a geomagnetic sensor 120 for measuring the orientation of the geomagnetism. And is rotated and moved integrally with the body of the multi-axis sensing device 2. The multi-axis sensing device 2 is mounted on the printed circuit board.

The internal calibration unit 200 is for calculating an internal calibration value for each of the acceleration sensor 110 and the geomagnetic sensor 120, that is, a scale factor and a bias vector. 5 is a view schematically showing the internal calibration unit 200 of the present embodiment. Referring to FIG. 5, the internal calibration unit 200 includes an ellipsoidal fitting unit 210, an outlier remover, and a calibration factor calculation unit 230.

The ellipsoidal fitting unit 210 receives the measured value of the sensor and performs an ellipsoid fitting algorithm on the received ellipsoid fitting algorithm. The ellipsoidal fitting unit 210 performs an ellipsoid fitting algorithm with measured values measured from the acceleration sensor 110 or the geomagnetic sensor 120 as the multiaxial sensing apparatus 2 of the present embodiment rotates in an arbitrary direction. A specific method for performing the ellipsoid fitting algorithm is disclosed in Korean Patent Registration No. 10-1209571 by the inventors of the present application and will be hereby incorporated by reference.

The outlier removing unit 220 recognizes a measurement value corresponding to an outlier largely deviated from the ellipsoid in the ellipsoid fitting and recognizes the error as an error, and removes the measured value. The outlier is removed from the ellipsoid by a predetermined distance or more, and the outlier removal unit 220 captures and excludes the outlier. When the outlier is removed from the measured value by the outlier removing unit 220, the ellipsoidal fitting unit 210 performs the ellipsoid fitting again on the measurement value excluding the outlier.

The calibration factor calculating unit 230 calculates the internal correction factor for each sensor, i.e., a scale factor matrix and a bias vector, using the result of the ellipsoidal fitting unit 210. [ As a method of calculating the scale factor matrix, an ellipse scale factor value and a bias value are calculated by performing a semi-definite programming (SDP) algorithm, and an SVD (singular value decomposition) is performed on the ellipse scale factor value to calculate a scale factor matrix . ≪ / RTI > A specific method of calculating the correction factor is described in detail in the above-mentioned Korean Patent Registration No. 10-1209571, and therefore, it is hereby incorporated by reference.

As described above, the internal calibration unit 200 receives the measured values of the acceleration sensor 110 and the geomagnetic sensor 120 when the multi-axis sensing device 2 is arbitrarily rotated, and receives the measured values of the acceleration sensor 110 and the geomagnetic sensor 120, Lt; RTI ID = 0.0 > 120 < / RTI > When the internal calibration unit 200 calculates the internal calibration factor, the calibrated measurement value O can be obtained from the measured value I of the acceleration sensor 110 and the geomagnetic sensor 120 using the internal calibration factor.

In the above description, the ellipsoidal fitting is used for internal calibration of the geomagnetic sensor 120. However, internal calibration of the geomagnetic sensor 120 may be performed by rotating the multi-axis sensor device 2 in one direction and its opposite direction, A method of measuring a vector can be used. More specifically, the geomagnetic sensor 120 measures a geomagnetism vector having a predetermined magnitude and direction. When the orientation of the multi-axis sensor device 2, that is, the geomagnetic sensor 120, The geomagnetic sensor 120 measures a geomagnetism vector having the same magnitude but opposite direction only. Therefore, a correction factor including the scale factor matrix and the bias vector of the geomagnetic sensor 120 may be calculated so that the geomagnetism vector values measured by the geomagnetic sensor 120 have the same magnitude and only the directions are reversed.

The gyro sensor calibration unit 300 is a part that performs internal calibration of the gyro sensor 130. The gyro sensor 130 performs calibration using the calibrated output value of the acceleration sensor 110 or the geomagnetic sensor 120 since the gyro sensor 130 is difficult to automatically calibrate. That is, the gyro sensor calibration unit 300 measures the internal calibration of the gyro sensor 130 based on the calibrated measurement values of the acceleration sensor 110 or the geomagnetic sensor 120 obtained by arbitrarily rotating the multi- I.e., a scale factor matrix and a bias vector.

In order to calculate the calibration factor of the gyro sensor 130, the gyro sensor calibration unit 300 may calculate the calibration factor of the gyro sensor 130 based on the specific direction fixed to the space, for example, the gravity direction or the magnetic north direction, May be used. For example, when the gyro sensor calibration unit 300 calibrates the gyro sensor 130 using the calibrated output value of the acceleration sensor 110, the acceleration sensor 110 The scale factor matrix of the gyro sensor 130 can be calculated using the change amount of the measured gravity vector. When the gyro sensor calibration unit 300 calibrates the gyro sensor 130 using the calibrated measurement value of the geomagnetic sensor 120, the geomagnetic sensor 120 is rotated according to the rotation of the multi- The scale factor matrix of the gyro sensor 130 may be calculated using the change amount of the geomagnetism vector measured at the gyro sensor 130. [

On the other hand, the bias vector of the gyro sensor 130 is set such that the change of the calibrated measured value of the acceleration sensor 110 or the geomagnetic sensor 120 is small and the change of the measured value of the gyro sensor 130 is small , That is, when the movement of the multi-axis measuring device is recognized as being mild, the corrected measured value of the acceleration sensor 110 or the geomagnetic sensor 120 and the measured value of the gyro sensor can be calculated through analysis in the frequency domain . For example, the gyro sensor 130 recognizes that the change over time is slight using signal processing such as STFT (short time Fourier transform), wavelet transform, FFT (fast Fourier transform), or DTW (dynamic time warping) Can be obtained.

The gyro sensor calibration unit 300 may use an image captured by the camera 500 instead of using the calibration values of the acceleration sensor 110 or the geomagnetic sensor 120 to adjust the internal calibration factor of the gyro sensor 130 . That is, the camera 500 captures and tracks the stationary object on the photographed image, measures the relative orientation of the stationary object as the multiaxial sensing device 2 rotates, and then the gyro sensor calibration unit 300 measures the relative orientation of the stationary object, The internal correction factors of the sensor 130, in particular the scale factor matrix, may be calculated. Meanwhile, the bias vector of the gyro sensor 130 indicates that the change of the specific physical quantity vector obtained from the photographed image of the camera 500 is small for a certain period of time and the change of the gyro sensor 130 is slight, In this case, the bias of the gyro sensor 130 can be calculated through analysis of the frequency domain.

As a method for capturing and tracking a still object in the captured image of the camera 500, a parellel tracking and mapping (PTAM) algorithm can be utilized. The PTAM algorithm is described in G. Klein and D. Murray in "Parallel Tracking and Mapping for Small AR Workspaces" (Proceedings of the 6th IEEE and ACM International Symposium on Mixed and Augmented Reality (ISMAR '07), pp225-234, 2007) And http://www.robots.ox.ac.uk/~gk/PTAM, which are hereby incorporated by reference.

The external calibration unit 400 calculates a conversion value between the coordinate systems of the respective sensors using the measured values of the acceleration sensor 110, the geomagnetic sensor 120, and the gyro sensor 130. Accurate measurements of the acceleration sensor 110, the geomagnetic sensor 120 and the gyro sensor 130 can be obtained by arbitrarily rotating the multi-axis sensing device 2 of the present embodiment, And therefore, there is a mutual relationship therebetween. The process of obtaining the external calibration value using the calibrated measured value of each sensor corresponds to the process of obtaining X in Equation (2).

6 is a view schematically showing the relationship between the acceleration sensor 110 and the geomagnetic sensor 120. As shown in FIG. 6, there is no guarantee that the coordinate system of the acceleration sensor 110 and the geodetic sensor 120 are set to the same orientation. Therefore, the external calibration unit 400 can detect the angular position of the acceleration sensor 110, The relative orientation between the acceleration sensor 110 and the geomagnetic sensor 120, that is, the rotation transformation matrix T between the coordinate system of the acceleration sensor 110 and the coordinate system of the geomagnetic sensor 120 using the calibration measurement value output from the sensor, . When the external calibration unit 400 calculates the rotational transformation relation between the acceleration sensor 110 and the geomagnetic sensor 120, it can calculate the acceleration relative to the geomagnetism vector of the multi-axis sensing device 2, It is also possible to measure the acceleration situation of the multi-axis sensing device 2 in the east-south, north-south direction.

The external calibration unit 400 may be configured to determine the conversion relationship between the acceleration sensor 110 and the geomagnetic sensor 120 as well as the conversion relationship between the acceleration sensor 110 and the gyro sensor 130, 130 can be calculated in the same way. In this way, the external relationship between the sensor coordinate systems is calculated by the external calibration unit 400, so that the sensor coordinate system of each sensor can be calibrated. The external calibration unit 400 can also calculate the conversion relationship between the coordinate system of the camera 500 and the coordinate system of the acceleration sensor 110, the geomagnetic sensor 120 and the gyro sensor 130. [

Next, an example of an automatic calibration method using the multi-axis sensing device 2 of the present embodiment will be summarized.

7 is a flowchart schematically showing an automatic calibration method for a plurality of sensors using the multi-axis sensing device 2 of the present embodiment.

7, the automatic calibration method of the multiaxial sensing apparatus 2 of the present embodiment is a method of integrally calibrating a plurality of sensors of the multiaxial sensing apparatus 2, in which the multiaxial sensing apparatus 2 is rotated in an arbitrary direction (S30, S32) of acquiring the measured values of the acceleration sensor 110 or the geomagnetic sensor 120, performing the internal calibration on the acceleration sensor 110 or the geomagnetic sensor 120 Performing an internal calibration for the gyro sensor 130, and performing an external calibration for each sensor (S90).

The step S20 of rotating the multi-axis sensing device 2 in an arbitrary direction includes a plurality of sensors including an acceleration sensor 110, a geomagnetic sensor 120, a gyro sensor 130 and a camera 500 The multi-axis sensing device 2 is prepared (S10) and rotated in an arbitrary direction. When the multi-axis sensing device 2 is rotated, various sensors fixed thereto are also rotated integrally. Any rotation of the multi-axis sensing device 2 may be performed intentionally by the user, but it may be done naturally as part of the use of the device with the multi-axis sensing device 2. [

The steps S30 and S32 of acquiring the measured values of the acceleration sensor 110 or the geomagnetic sensor 120 may be performed by the accelerometer 110 or the geomagnetic sensor 120, respectively. Calibration of the acceleration sensor 110 and the geomagnetic sensor 120 is not necessary since it is not possible to guarantee that the measured values are accurate acceleration values or geomagnetic values.

The step of internally calibrating the acceleration sensor 110 and the geomagnetic sensor 120 is a step of calculating an internal calibration value including a scale factor matrix and a bias vector of each of the acceleration sensor 110 and the geomagnetic sensor 120 . This step can be performed by the internal calibration unit 200 described above.

First, the step of internally calibrating the accelerometer sensor will be described. The step of internally calibrating the accelerometer sensor includes performing an ellipsoid fitting with the measured value of the acceleration sensor 110, removing the outlier at the time of fitting the ellipsoid, calculating the internal correction factor (S60). At this time, the step of performing the ellipsoid fitting (S40) and the step of removing the outlier (S50) may be repeatedly performed.

FIG. 8 is a view schematically showing the ellipsoid alignment performed on sensor measured values. FIG. As shown in FIG. 8, when the ellipsoid fitting is performed on the acceleration measurement value, it can be seen that the measurement value EL is located on the ellipsoid. At this time, some measured values are located far from the ellipsoid, which is classified as an outlier (OL) corresponding to the measurement error. The selection of the outlier (OL) can be based on whether the ellipsoid is separated by a predetermined distance or more.

The measurement values corresponding to the outliers OL are excluded and the ellipsoid alignment can be performed again using only the remaining measurement values (S40). If the ellipsoid fitting is performed again with the outlier corresponding to the error removed, the accuracy of the ellipsoid fitting can be further increased.

In step S60 of calculating the internal correction factor, the scale factor matrix and the bias vector are calculated by performing the SDP algorithm and the SVD algorithm on the result obtained by performing the ellipsoid fitting by the above-described method. The method of calculating the scale factor and the bias vector is described in the above-mentioned Korean Patent No. 10-1209571, and a detailed description thereof is omitted here.

Fig. 9 is a view schematically showing a calibrated measurement value using the calculated internal correction factor. As shown in Fig. 9, the calibrated measured value SL is located on the surface of the sphere relative to the origin. That is, the calibrated measurement value has no bias and the sensitivity is also constant depending on the direction of the sensor.

When the internal calibration factor for the acceleration is calculated, it is possible to effectively calibrate the measured value to be measured subsequently (S70). Therefore, the multi-axis sensing device 2 can output a reliable acceleration value.

The geomagnetic sensor 120 is also internally calibrated in the same manner as the acceleration sensor 110. That is, a method of performing ellipsoid fitting according to a measured value of the geomagnetic sensor 120, removing the outlier, performing ellipsoid fitting again, and calculating an internal correction factor based on the result, Internal calibration is possible. Accordingly, the geomagnetic sensor 120 performs internal calibration for the subsequent measurement values (S72), so that the multi-axis sensor device 2 equipped with the same can provide reliable geomagnetism direction information to the user.

As a method of calculating the internal calibration factor of the geomagnetic sensor 120, in addition to the calibration method using the ellipsoid fitting, the multiaxial sensing device 2 performs the symmetric rotation (S22) and obtains its measured value (S36) A method of calculating the internal correction factor (S66) may be used. That is, by using the fact that the geomagnetism vector is measured while rotating the geomagnetic sensor 120 in one direction and the opposite direction, the scale factor matrix of the geomagnetic sensor 120 and The bias vector can be calculated.

When the internal calibration for the acceleration sensor 110 and the geomagnetic sensor 120 is completed, a step of performing an internal calibration for the gyro sensor 130 is performed.

The step of performing internal calibration of the gyro sensor 130 includes a step of detecting a reference direction stopped in space and a step S80 of calculating an internal correction factor of the gyro sensor 130 using the reference direction.

The step of detecting a stationary reference direction on the space may include the step of detecting a gravity vector using the calibrated acceleration sensor 110 or the step of detecting the geomagnetism using the calibrated geomagnetic sensor 120 (S82).

The step of calculating the internal correction factor S80 includes a step S802 of calculating the scale factor of the gyro sensor 130 and a step S804 of calculating the bias of the gyro sensor 130. The step of calculating the scale factor of the gyro sensor 130 when using the gravity vector in the reference direction detects the gravity vector with the calibrated acceleration sensor 110 while slowly rotating the multiaxial sensing device 2 in an arbitrary direction Calculating a scale factor matrix of the gyro sensor 130 by calculating a relative rotation of the multi-axis sensing device 2 with respect to the rotation of the multi-axis sensing device 2, and then comparing the relative rotation with the measured value of the gyro sensor 130.

The step of calculating the bias of the gyro sensor 130 (S804) may include a short time Fourier transform (STFT), a wavelet transform, a fast Fourier transform (FFT) The gyro sensor 130 and the gyro sensor 130 recognize that the measurement values of the gyro sensor and the acceleration sensor 110 or the geomagnetic sensor 120 are mild by using a signal processing method such as Fourier transform or DTW, / RTI >

When the geomagnetism vector is used as the reference direction, the geomagnetism sensor 120 detects the geomagnetism vector by rotating the multi-axis sensing device 2 in an arbitrary direction (S82) May be used. In this case, in order to calculate the bias value, when the measured value of the acceleration sensor 110 or the geomagnetic sensor 120 or the measured measured value changes slightly for a predetermined time, the frequency of the gyro sensor 130 The bias vector can be calculated.

In the above description, the ellipsoidal fitting is used for internal calibration of the geomagnetic sensor 120. However, in the calculation of the internal calibration factor of the geomagnetic sensor 120, (S22), and the geomagnetism vector is measured (S36) in the opposite direction, as described above. If the internal calibration of the geomagnetic sensor 120 is performed in this way, the internal calibration values of the gyro sensor 130 can be calculated in the same manner using the calibrated measured values.

Meanwhile, as a method for internally calibrating the gyro sensor 130, a camera 500 may be used. The scale factor of the gyro sensor 130 can be calculated (S82) by capturing and tracking the stationary reference object as the camera 500 (S84) and calculating the orientation of the multi-axis sensing with respect to the reference object (S74). FIG. 10 is a simplified view of tracking objects that are stationary with the camera using PTAM software. The initial feature map DL is extracted using the outline of the object OB that is stopped on the image captured by the camera 500 as shown in FIG. 10, and the relative position of the camera is calculated from the initial feature map DL . The scale factor of the gyro sensor 130 is calculated by using the relationship between the positional fluctuation of the camera and the fluctuation of the measured value of the gyro sensor 130. [ The bias value of the gyro sensor 130 is a function of a frequency range of the rotation angle of the camera 500 and a measurement value of the gyro sensor 130 for the reference object when the movement of the multi- Lt; / RTI >

When the internal calibration value including the scale factor and the bias value for each sensor is calculated by the above method, the conversion relation for each sensor coordinate system, that is, the external calibration factor is calculated (S90). The external calibration factor can be calculated by substituting the measurement value measured at each sensor into equation (2), that is, by obtaining X.

As described above, according to the calibration method of the multi-axis sensing device 2 according to the present embodiment, a plurality of sensors of the multi-axis sensing device 2 can be internally calibrated And external calibration.

In addition, the above calibration method may be performed off-line before the multi-axis sensing device 2, but may be performed even when the multi-axis sensing device 2 is on-line. Therefore, the multi-axis sensing device 2 of the present embodiment can calibrate a plurality of sensors by itself during execution.

In addition, the multiaxial sensing device 2 of the present embodiment and its calibration method are robust against noise and disturbance by removing the outlier, and the accuracy of calibration is high. Therefore, by using such a multi-axis sensing device 2, various physical quantities such as acceleration, rotation, and azimuth can be measured with high accuracy.

The steps including the various arithmetic operations in the above-described steps may be performed by the arithmetic unit including the microprocessor. In addition, a program for performing this operation may be loaded in the arithmetic unit in order to perform the arithmetic operation of each step.

While the present invention has been described in connection with certain exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, is intended to cover various modifications and similarities.

For example, in the above embodiment, the multi-axis sensing device 2 is described as having the camera 500, but the multi-axis sensing device 2 does not necessarily include the camera 500.

Also, the multi-axis sensing device 2 of the above-described embodiment includes the geomagnetic sensor 120 and the acceleration sensor 110, but any one of them may be omitted.

Also, in the above embodiment, it is explained that the process of performing the ellipsoid fitting and then removing the outlier and performing the ellipsoid fitting is repeated, but the ellipsoid fitting may be performed only once, and the outlier removal may not be performed.

In the above description, in order to calculate the internal calibration value of the gyro sensor 130, any one of the calibration value of the acceleration sensor 110, the calibration measurement value of the geomagnetic sensor 120, or the physical quantity vector acquired from the camera 500 The reliability of the calibration value of the gyro sensor 130 can be further increased by using them collectively.

It is needless to say that the present invention can be embodied in various forms.

1,2 ... multi-axis sensing device
110 ... acceleration sensor
120 ... Earth sensor
130 ... Gyro sensor
200 ... internal gating
300 ... gyro sensor calibration unit
400 ... external control unit
500 ... camera

Claims (20)

An automatic calibration method for a multi-axis sensing device including an acceleration sensor or a geomagnetic sensor and a gyro sensor,
(a) acquiring a measured value of the acceleration sensor or the geodetic sensor as the multi-axis sensing device rotates in an arbitrary direction;
(b) calculating an internal calibration value of the acceleration sensor or the geomagnetic sensor;
(c) calibrating a measurement value of the acceleration sensor or the geomagnetic sensor using an internal calibration value of the acceleration sensor or the geomagnetic sensor;
(d) calculating an internal calibration value of the gyro sensor using the measured value of the acceleration sensor or the geomagnetic sensor,
The step (b)
Performing ellipsoid fitting using the measured values of the acceleration sensor,
The step (d)
Calculating a scale factor of the gyro sensor using a gravity vector measured by the acceleration sensor in accordance with rotation of the multi-axis sensing device; And
Calculating a bias value of the gyro sensor using a measured value of the calibrated acceleration sensor for a predetermined time and a frequency domain analysis of the measured value of the gyro sensor, Calibration method. delete The method according to claim 1,
The step (b)
And performing the ellipsoid fitting again except for the outlier at the time of fitting the ellipsoid. delete delete The method according to claim 1,
And an external calibration step of calculating a conversion relationship between the coordinate system of the acceleration sensor or the coordinate system of the gyro sensor. The method according to claim 1,
The step (b)
And performing ellipsoid fitting using the measured values of the geomagnetic sensor. 8. The method of claim 7,
The step (b)
And performing the ellipsoid fitting again except for the outlier at the time of fitting the ellipsoid. 8. The method of claim 7,
The step (d)
And calculating a scale factor of the gyro sensor by using a gravity vector measured by the geomagnetic sensor according to the rotation of the multi-axis sensing device. 10. The method of claim 9,
The step (d)
Further comprising calculating a bias value of the gyro sensor using a measured value of the geomagnetic sensor and a frequency domain analysis of the measured value of the gyro sensor for a predetermined period of time Automatic calibration method. 8. The method of claim 7,
And an external calibration step of calculating a conversion relationship between the coordinate system of the geomagnetic sensor or the coordinate system of the gyro sensor. The method according to claim 1,
Further comprising an external calibration step of calculating a conversion relation between the coordinate system of the acceleration sensor, the geodetic sensor coordinate system or the coordinate system of the gyro sensor. The method according to claim 1,
The step (b)
Wherein the internal calibration value of the geomagnetic sensor is calculated using the geomagnetism vector value measured by rotating the multi-axis sensing device in one direction and its counterclockwise direction. delete delete A computer-readable medium containing a program for performing the method of any one of claims 1, 3, and 6 to 13. The body,
An acceleration sensor or a geomagnetic sensor relatively fixed to the body,
A gyro sensor relatively fixed to the body,
An internal calibration unit for calculating an internal calibration value of the acceleration sensor or the geomagnetic sensor using the measurement values of the acceleration sensor or the geomagnetic sensor,
A gyro sensor calibration unit for calculating an internal calibration value of the gyro sensor by using a calibrated measurement value of the acceleration sensor or a calibrated measurement value of the geomagnetic sensor and a measurement value of the gyro sensor,
And an external calibration section for calculating a conversion value between the coordinate system of the acceleration sensor or the coordinate system of the geomagnetic sensor and the coordinate system of the gyro sensor,
The internal calibrator,
And an ellipsoidal fitting unit for performing elliptical fitting using the measured values of the acceleration sensor or the geomagnetic sensor,
The gyro sensor calibration unit may calculate a scale factor of the gyro sensor using a gravity vector measured by the acceleration sensor in accordance with the rotation of the body and measure a measured value of the calibrated acceleration sensor for a predetermined period of time And a frequency domain analysis of the measured value of the gyro sensor to calculate a bias value of the gyro sensor. delete delete 18. The method of claim 17,
And a camera fixed relative to the body.

Images