JP4599502B1 - Magnetic gyro - Google Patents

Abstract

To provide a magnetic gyro capable of accurately measuring a rotation angle during both high-speed rotation and low-speed rotation.
A three-axis magnetic sensor 21, a three-axis acceleration sensor 22, a memory 3, a first rotation axis determination unit 41, a second rotation axis determination unit 42, and a rotation angle of an object to be measured are represented by magnetic vectors. First rotation angle calculation means 51 that calculates based on the data, second rotation angle calculation means 52 that calculates the rotation angle of the measured object based on the magnetic vector data and the acceleration vector data, and the rotational speed determination means 6 And an output means 8 for outputting a calculation result by the first rotation angle calculation means in the high speed mode and outputting a calculation result by the second rotation angle calculation means in the low speed mode.
[Selection] Figure 1

Description

  The present invention continuously measures the geomagnetic vector with a magnetic sensor every minute time even when the object to be measured is rotating while changing the rotation axis and the rotation speed with time like a coma or the like. Thus, the present invention relates to a magnetic gyro capable of measuring an instantaneous rotation axis and rotation angle that change every moment.

In recent years, for example, when a portable electronic device such as a mobile phone or a PDA is tilted or swung, the rotation angle in the pitch, yaw and roll directions (device posture) and the rotation angular velocity are detected, and these are detected. A technology for input signals has been developed.
In addition, a technique has been developed in which a camera rotation angle and rotation angular velocity are detected and used as a correction signal for preventing camera shake of a photograph (photographed image).
In these techniques, a gyro is used as means for detecting the rotation angle and rotation angular velocity of the device.

As the above-mentioned gyro, for example, one using a kinematic principle such as a directional gyro (Patent Document 1) or a rate gyro (Patent Document 2) is generally used.
However, these gyros may react to mechanical vibrations or impacts other than the rotational motion that is the object of measurement, and noise may be superimposed on the output signal for accurate measurement. There is a problem that it may become difficult.

Moreover, since the conventional gyro described above requires a complicated mechanism, it is difficult to reduce cost and size.
Therefore, these gyros have a problem that it is difficult to incorporate them in, for example, portable electronic devices whose size and density are increasing.

Therefore, a magnetic gyro that measures the rotation angle based on an arbitrary posture using geomagnetism has been proposed (Patent Document 3).
The magnetic gyro disclosed in Patent Document 3 can measure the rotation angle of the measured object based on the temporal change of the geomagnetic vector detected by the three-axis magnetic sensor.

US Pat. No. 3,143,892 Japanese Patent Laid-Open No. 7-139951 International Publication No. 2007/099599

  However, when measuring the rotation angle using the magnetic gyro described in Patent Document 3, it is necessary to calculate the rotation axis based on the change of the geomagnetic vector with respect to the measured object. In the case of low-speed rotation, since the change in the geomagnetic vector is small (slow), it takes time to calculate the rotation axis. For this reason, there is a problem that it is difficult to quickly and accurately measure the rotation angle of the object to be measured during low-speed rotation.

  The present invention has been made in view of such problems, and an object of the present invention is to provide a magnetic gyro capable of accurately measuring a rotation angle even at high speed rotation and low speed rotation.

The present invention includes a three-axis magnetic sensor that detects geomagnetism as a magnetic vector in a three-axis orthogonal coordinate system fixed to a measurement object;
A triaxial acceleration sensor for detecting gravitational acceleration as an acceleration vector in the triaxial orthogonal coordinate system;
A memory for storing data of the magnetic vector detected in time series by the triaxial magnetic sensor and data of the acceleration vector detected in time series by the triaxial acceleration sensor;
First rotation axis determination means for determining a rotation axis as a reference for the rotational motion of the object to be measured, based on data of the magnetic vector at two or more different time points stored in the memory;
Second rotation axis determination means for determining a rotation axis as a reference for the rotational movement of the measured object based on the magnetic vector data and the acceleration vector data at two or more different points in time stored in the memory;
First rotation angle calculation means for calculating a rotation angle of the measurement object around the rotation axis determined by the first rotation axis determination means based on the magnetic vector data;
Second rotation angle calculation means for calculating the rotation angle of the measured object around the rotation axis determined by the second rotation axis determination means based on the magnetic vector data and the acceleration vector data; ,
Based on the magnetic vector data detected in time series by the three-axis magnetic sensor, the measured object is rotating at a speed higher than the reference rotation speed or less than the reference rotation speed. A rotational speed discriminating means for discriminating whether the low-speed rotation is performed,
When the rotational speed discriminating means determines that the object to be measured is rotating at the high speed mode, the rotation angle calculation means outputs the calculation result of the rotation angle of the object to be measured by the first rotation angle calculating means, An output for outputting the calculation result of the rotation angle of the measurement object by the second rotation angle calculation means when the rotation speed determination means determines that the measurement object is rotating at the low speed mode. And a magnetic gyroscope.

  The magnetic gyro has the rotational speed determining means. The output means outputs the calculation result of the rotation angle (posture change amount) of the measured object by the first rotation angle calculation means when the rotation speed determination means determines that the high speed mode is selected, and the low speed When the mode is determined, the calculation result of the rotation angle of the object to be measured by the second rotation angle calculation means is output.

  Here, the first rotation angle calculation means is configured to calculate the rotation angle of the measurement object around the rotation axis based on the magnetic vector data. In other words, since the acceleration vector data is not used in calculating the rotation angle, even if the measured object rotates at high speed, the accompanying acceleration does not become a noise when calculating the rotation angle, and the accurate rotation angle. Can be calculated. Further, since the first rotation angle calculation means is used in the high speed mode, the first rotation axis determination means can quickly determine the rotation axis only from the magnetic vector data, and the rotation axis can be determined. The center rotation angle can be quickly calculated by the first rotation angle calculation means.

On the other hand, the second rotation angle calculation means is configured to calculate the rotation angle of the measurement object based on the magnetic vector data and the acceleration vector data. That is, since the acceleration vector data is used in addition to the magnetic vector data in calculating the rotation angle, the magnetic vector data and the acceleration vector data are also obtained by the second rotation axis determination means in the low speed mode. Can be used to quickly determine the axis of rotation. Then, the second rotation angle calculation means can quickly calculate the rotation angle of the object to be measured based on the rotation axis determined thereby.
In the low speed mode in which the object to be measured is rotating at a low speed, the acceleration accompanying the rotational motion, that is, the acceleration other than the gravitational acceleration is sufficiently small and can be ignored. Therefore, this acceleration does not cause noise when calculating the rotation angle.

  As described above, by properly using the rotation angle calculation means in the high-speed mode and the low-speed mode, the magnetic gyro described above can be applied to everything from low-speed rotation to high-speed rotation of the measured object that could not be realized by the conventional gyro. Accurate rotation angle detection in an exercise situation can be performed.

  As described above, according to the present invention, it is possible to provide a magnetic gyro capable of accurately measuring a rotation angle even at high speed and at low speed.

1 is a conceptual diagram of a magnetic gyroscope in Embodiment 1. FIG. FIG. 3 is a flowchart of measurement of the rotation angle and the rotation angular velocity by the magnetic gyroscope in the first embodiment. 3 is an explanatory diagram of a three-axis orthogonal coordinate system, a rotation axis, a magnetic vector, and the like in Embodiment 1. FIG. FIG. 3 is an auxiliary explanatory diagram of a calculation method of rotation center coordinates in the first embodiment. FIG. 3 is an auxiliary explanatory diagram of a rotation angle calculation method in the first embodiment. 1 is a perspective view of a triaxial magnetic sensor in Embodiment 1. FIG. The performance comparison figure of a magneto impedance sensor element and other magnetic sensors. FIG. 10 is an auxiliary explanatory diagram of a method for calculating a rotation center coordinate when there is an offset error in the second embodiment. 9 is an auxiliary explanatory diagram of a calculation method of a rotation angle in Embodiment 2. FIG.

In the present invention, the magnetic gyro can be mounted on various measured objects such as portable electronic devices such as mobile phones and PDAs, cameras, vehicles, robots, aircrafts, ships, and the like.
The magnetic vector is a vector parallel to the geomagnetism starting from the origin in the three-axis orthogonal coordinate system. The magnitude may vary depending on the location, but the magnitude is constant at the same position. is there. The acceleration vector is a vertically downward vector (gravity direction) starting from the origin in the three-axis orthogonal coordinate system, and the magnitude thereof is constant.

Further, the rotation angle of the measured object between two different time points calculated by the first rotation angle calculating means or the second rotation angle calculating means, and the difference between the sampling times of the magnetic vector data at the two time points It is preferable to have an angular velocity calculating means for calculating the rotational angular velocity of the object to be measured around the rotation axis.
In the present invention, magnetic vectors are continuously measured at minute time intervals (approximately several milliseconds or less), and the instantaneous rotation axis and rotation angle at each time are obtained from the data. Therefore, the instantaneous rotation angular velocity (attitude change rate) of the measured object at each time can be easily detected by using the measurement time data of the magnetic vector stored in the memory for the calculated rotation angle. A magnetic gyro can be obtained. Therefore, not only the posture change amount of the measured object but also the posture change speed can be detected.

Further, the first rotation axis determination means rotates an arbitrary rotation axis that is not limited to the three axes of the three-axis orthogonal coordinate system based on the magnetic vector data stored in the memory at three different time points or more. It is preferable to be configured to determine the axis.
In this case, the rotation axis can be made substantially coincident with the actual rotation axis of the measured object, and a more accurate rotation angle can be measured.

In addition, the first rotation axis determination means calculates two difference vectors that are differences between two magnetic vectors among the magnetic vectors at three or more different time points, and calculates the outer product of these two difference vectors, It is preferable to calculate a rotation axis vector in the same direction as the rotation axis.
In this case, the rotation axis can be calculated easily and accurately.

  Further, the first rotation angle calculation means calculates the center coordinates of a trajectory circle passing through the coordinate points of the three or more magnetic vectors based on the magnetic vector data at three or more different points in the three-axis orthogonal coordinate system. Rotation center coordinate calculation means for calculating, and radius calculation means for calculating a radius of the trajectory circle by calculating a distance between the center coordinate calculated by the rotation center coordinate calculation means and a coordinate point of the magnetic vector. The rotation angle between the measurement times of the magnetic vector data used in the calculation is calculated based on the radius of the locus circle calculated by the radius calculation means and the coordinate points of the magnetic vector at two different time points. It is preferable to be configured to calculate (claim 2).

In this case, even if there is an offset error in the three-axis magnetic sensor, the rotation angle of the measurement object can be measured easily and accurately.
That is, when there is a magnetized part in the vicinity of the magnetic sensor of the object to be measured, a value obtained by adding the magnetic vector generated by the part to the geomagnetic vector is measured. However, in the present invention, an accurate rotation angle can be directly measured regardless of the presence or absence of an offset error by appropriate processing in the calculation process of the rotation angle of the object to be measured even if measurement is performed in an environment where this addition (offset) exists. can do. In addition, since the rotation angle of the measured object is calculated by directly obtaining the rotation center coordinates of the magnetic vector without calculating the offset origin, no special correction calculation is required.
Therefore, with a simple configuration, even if an offset error occurs in the three-axis magnetic sensor, the rotation angle of the measured object can be accurately measured. That is, it is possible to prevent a phenomenon in which a measurement error occurs due to a drift of the measurement origin with use time as in a conventional vibration gyroscope with a simple configuration.

The first rotation axis determination means may be configured to determine, as the rotation axis, an axis in a direction in which the change in the magnetic vector component is smallest among the three axes of the three-axis orthogonal coordinate system. preferable.
In this case, the calculation performed in the first rotation axis determination means can be simplified, and the rotation axis can be easily determined. In the magnetic gyro of this example, basically, the user needs to accurately determine the rotation axis regardless of what the user performs on the object to be measured. Therefore, the rotation axis to be obtained is an arbitrary axis that changes with time. . However, when the rotation axis substantially coincides with any rotation axis of the three-axis orthogonal coordinate system or when it can be determined that the rotation axis is very close, the accuracy can be lowered so much without performing an operation for obtaining an arbitrary axis. In this case, any one of the three axis directions can be set as the rotation axis.
Thereby, calculation can be simplified.
In other words, the direction of rotation given to a mobile phone or the like is often the direction of yaw, roll, or pitch, based on the case, so if you can identify which of these rotations, you can input by rotational movement. Therefore, it is effective when a mobile phone or the like is used as a measurement object.

The calculation result of the rotation angle of the measured object calculated immediately after the high speed mode and the low speed mode are switched uses at least one or a plurality of data immediately before the high speed mode and the low speed mode are switched. It is preferable that the smooth continuity is ensured between the calculation result of the rotation angle of the measured object calculated immediately before the high speed mode and the low speed mode are switched ( Claim 3).
In this case, a measurement error associated with switching between the high speed mode and the low speed mode can be reduced. That is, when the high-speed mode and the low-speed mode are switched, the rotation angle calculation means to be used is switched between the first rotation angle calculation means and the second rotation angle calculation means. The axis determination means switches between the first rotation axis determination means and the second rotation axis determination means. Therefore, before and after the switching, the rotation axis used for the calculation changes, and the calculation result also changes. By smoothing and continuing this change, the calculation error of the rotation angle can be reduced.

Further, when the direction of the actual rotation axis of the object to be measured coincides with the direction of the magnetic vector, the measured object around the rotation axis instead of the first rotation angle calculation means or the second rotation angle calculation means. It is preferable to have a third rotation angle calculation means for calculating the rotation angle of the measuring body based on the acceleration vector data.
In this case, since the magnetic vector does not change with the rotation of the measured object, the rotation angle of the measured object cannot be calculated using the magnetic vector data. That is, neither the first rotation angle calculation means nor the second rotation angle calculation means can be used. Therefore, the third rotation angle calculation means calculates the rotation angle using only the acceleration vector data. Thereby, even when the direction of the actual rotation axis of the measured object coincides with the direction of the magnetic vector, the rotation angle of the measured object can be calculated.
Here, “the direction of the actual rotation axis of the object to be measured coincides with the direction of the magnetic vector” is not only the case where these directions are completely coincident, but also the first rotation axis determination means and the second rotation axis. The case where the rotation axis is substantially coincident to the extent that it is difficult to determine the rotation axis is also included.

The three-axis magnetic sensor is preferably constituted by a magneto-impedance sensor element.
In this case, a magnetic gyro with higher accuracy, higher sensitivity, higher response, and smaller size can be obtained. In other words, in order to be able to accurately measure the rotational motion of the measured object whose rotational axis, rotational angle, and rotational angular velocity are changing rapidly, every very short time interval (approximately several milliseconds or less). The magnetic vector must be measured continuously. That is, in order to accurately determine the rotational angular velocity of an object whose rotational state changes, the definition of the angular velocity, that is, the value obtained by differentiating the rotational angle with time (the rotational angle when Δt is infinitely small) It is necessary to obtain a value that substantially matches the value obtained by dividing the above by time. For this purpose, a magnetic sensor that can be measured accurately in a very short time is required.
The magneto-impedance sensor element (MI element) is an excellent element that can cope with the above-described measurement conditions and has high sensitivity, and therefore can detect weak geomagnetism with high accuracy. Furthermore, since the magneto-impedance sensor element is small, a small three-axis magnetic sensor can be obtained. This also makes it possible to fit the magnetic gyro in the IC chip.
The three-axis magnetic sensor can be formed by disposing the three magneto-impedance sensor elements so that their magnetic sensitive directions are in the three-axis directions orthogonal to each other.

As described above, the triaxial magnetic sensor needs to be able to measure in a very short time, and if it is the performance of the current sensor, a magneto-impedance sensor element is optimal, There is no problem even if other sensors are used as long as they can realize sensitivity and high response. For example, even if various elements for magnetic detection such as Hall elements, magnetoresistive elements, flux gates, etc. can be selected, elements that can be measured with high accuracy in a short time can be used to construct a magnetic gyro with the same accuracy. You can also.
In addition, the triaxial acceleration sensor can be configured by, for example, a capacitive acceleration sensor.

In addition, the magnetic gyro measures the geomagnetic vector continuously when the measured object is in a rotational motion state, and the instantaneous rotational axis at the time measured from the measured geomagnetic vector and the rotational axis as a center. The memory is configured to store the magnetic vector data and the acceleration vector data together with the time information of the measured moment, and the first rotation axis. The determination means and the second rotation axis determination means are configured to determine a rotation axis used as a reference for the instantaneous rotation motion of the measured object within the measurement time of the magnetic vector at the two or more time points. The first rotation angle calculation means and the second rotation angle calculation means are configured to calculate an instantaneous rotation angle of the measured object around the rotation axis. Preferred (claim 6).
In this case, the geomagnetic vector is continuously measured by the magnetic sensor every minute time even when the object to be measured is rotating, changing the rotation axis and the rotation speed with time like a top. Thus, the instantaneous rotation axis and rotation angle that change every moment can be measured.

Example 1
A magnetic gyro according to an embodiment of the present invention will be described with reference to FIGS.
As shown in FIG. 1, the magnetic gyro 1 of this example includes a three-axis magnetic sensor 21, a three-axis acceleration sensor 22, a memory 3, a first rotation axis determination unit 41, and a second rotation axis determination unit 42. , First rotation angle calculation means 51, second rotation angle calculation means 52, rotation speed determination means 6, angular speed calculation means 7, and output means 8.

The triaxial magnetic sensor 21 detects geomagnetism as a magnetic vector in the triaxial orthogonal coordinate system 10 fixed to the measurement object shown in FIG.
The triaxial acceleration sensor 22 detects gravitational acceleration as an acceleration vector in the triaxial orthogonal coordinate system 10.

  The memory 3 stores the magnetic vector data detected in time series by the triaxial magnetic sensor 21 and the time information of the moment when the acceleration vector data detected in time series by the triaxial acceleration sensor 22 is measured. To do.

The first rotation axis determining means 41 rotates based on the magnetic vector data stored in the memory 3 at two or more different points in time and passes through the origin of the three-axis orthogonal coordinate system 10 and is used as a reference for the rotational movement of the measured object. The axis is determined (see symbol K in FIG. 3).
The second rotation axis determination means 42 passes through the origin of the three-axis orthogonal coordinate system 10 and rotates the measured object based on the magnetic vector data and acceleration vector data stored at two or more different points in time stored in the memory 3. Determine the axis of rotation.

The first rotation angle calculation means 51 calculates the rotation angle of the measured object around the rotation axis determined by the first rotation axis determination means 41 based on the magnetic vector data.
The second rotation angle calculation unit 52 calculates the rotation angle of the measured object around the rotation axis determined by the second rotation axis determination unit 42 based on the magnetic vector data and the acceleration vector data.

  Based on the magnetic vector data detected by the three-axis magnetic sensor 21 in time series, the rotational speed discriminating means 6 is rotating at a speed higher than the reference rotational speed or the reference It is determined whether or not a low-speed rotation that is less than the rotation speed is performed.

  The angular velocity calculation means 7 includes the rotation angle of the measured object between two different time points calculated by the first rotation angle calculation means 51 or the second rotation angle calculation means 52, and the sampling time of the magnetic vector data at the two time points. Based on the difference, the rotational angular velocity of the measured object around the rotational axis is calculated.

  The output means 8 outputs the calculation result of the rotation angle of the measurement object by the first rotation angle calculation means 51 when the rotation speed determination means 6 determines that the measurement object is rotating at high speed. When the rotation speed discriminating means 6 determines that the measured object is rotating at a low speed, the second rotation angle calculating means 52 outputs the calculation result of the rotation angle of the measured object.

  The triaxial magnetic sensor 2 is constituted by a magneto-impedance sensor element 20 as shown in FIG. That is, the three-axis magnetic sensor 2 causes the three magneto-impedance sensor elements 20 to be in three-axis directions (X-axis direction, Y-axis direction, and Z-axis direction) in which the respective magnetic sensing directions are orthogonal to each other. It is formed by arranging. The three-axis acceleration sensor is a so-called MEMS acceleration sensor that combines two parts obtained by finely processing silicon in a comb shape and detects a change in the gap between parts caused by the application of acceleration as a change in capacitance. Combined in the axial direction. In FIG. 6, electronic components and wiring other than the magneto-impedance sensor element 20 are omitted.

A flow of a series of processes for detecting the rotation angle (posture change amount) and the rotation speed (posture change speed) of the measurement object by the magnetic gyro 1 of this example will be described with reference to FIG.
First, after the start of detection (step S1), the magnetic vector data measured by the triaxial magnetic sensor 21 and the triaxial acceleration sensor 22 and the acceleration vector data are sequentially input to the memory 3 together with the instantaneous time information of the measurement. To do.

Using these data, the first rotation axis determination means 41, the second rotation axis determination means 42, the first rotation angle calculation means 51, and the second rotation angle calculation means 52 respectively determine the rotation axis and rotate the rotation. The rotation angle of the measured object about the axis is calculated (steps S3 to S6).
As a result, as long as the direction of the actual rotation axis of the measured object does not coincide with or substantially coincides with the direction of the geomagnetic vector, at least one of the first rotation angle calculation means 51 and the second rotation angle calculation means 52 is accurate. A rotation angle is calculated. Note that the rotation axis determination means in the case where the direction of the actual rotation axis of the object to be measured matches or substantially matches the direction of the geomagnetic vector will be described later as a third rotation axis determination means in the fourth embodiment.

Next, the rotational speed discriminating means 6 discriminates whether the movement state of the object to be measured is the high speed mode of high speed rotation or the low speed mode of low speed rotation (step S7).
That is, based on the magnetic vector data detected in time series by the three-axis magnetic sensor 21, whether the measured object rotates at a higher speed than the reference (high speed mode) with a predetermined rotation speed as a reference, It is determined whether the motor is rotating at a speed lower than the standard (low speed mode). Specific means will be described later.

Here, when the high-speed mode is determined, it can be determined that the former is more accurate among the calculation results by the first rotation angle calculation unit 51 and the calculation results by the second rotation angle calculation unit 52. This is adopted (step S8).
On the other hand, when the low-speed mode is determined, it can be determined that the latter is more accurate among the calculation results by the first rotation angle calculation means 51 and the calculation results by the second rotation angle calculation means 52. Is adopted (step S9).

  Next, the obtained calculation result of the rotation angle is smoothed and made continuous by the low-pass filter (step S10). That is, when the obtained rotation angle data is smoothed and continuous using the rotation angle data obtained immediately before, the low speed mode and the high speed mode are switched on the way. However, it is possible to smooth and continuous discontinuities when outputting across the angle calculated by the first rotation angle calculation means and the angle calculated by the second rotation angle calculation means, resulting in mode switching. If not, the data can be smoothed and noise can be removed.

And the calculation result of the corrected rotation angle is output by the output means 8 (step S11).
Further, using the rotation angle calculation result, the angular velocity calculation means 7 calculates the rotation angular velocity (step S12), and the result is output by the output means 8 (step S13).

Next, a method of detecting the posture change amount and posture change speed of the measurement object using the magnetic gyro 1 of this example will be specifically described.
As shown in FIG. 1, the magnetic gyro 1 includes magnetic vector data and acceleration vector data detected by the three-axis magnetic sensor 21 and the three-axis acceleration sensor 222, and the three-axis magnetic sensor 21 and the three-axis acceleration sensor 22, respectively. A computer 11 that performs calculation to calculate the posture change amount and the posture change speed of the measurement object based on the time information of the measured moment, and output means such as a monitor that outputs the calculation result calculated by the computer 11 8. That is, the computer 11 includes the memory 3, the first rotation axis determination unit 41 and the second rotation axis determination unit 42, the first rotation angle calculation unit 51 and the second rotation angle calculation unit 52. The rotational speed discriminating means 6 and the angular speed calculating means 7 are provided.

The memory 3 is composed of hardware, and includes the first rotation axis determination means 41 and the second rotation axis determination means 42, the first rotation angle calculation means 51, and the second rotation angle calculation means 52, The rotational speed discriminating means 6 and the angular velocity calculating means 7 are constructed as calculation programs in software.
The first rotation axis determination means 41 and the first rotation angle calculation means 51, or the second rotation axis determination means 42 and the second rotation angle calculation means 52 do not necessarily need to be clearly distinguished from each other, and a series of operations by them is not necessarily required. The rotation axis and the rotation angle may be calculated by the above.

The triaxial magnetic sensor 21 is fixed to a part of the measured object, and detects the geomagnetism as magnetic vectors m ₁ , m ₂ , m ₃ in time series. The magnetic vectors m ₁ , m ₂ , and m ₃ are vectors starting from the origin O of the three-axis orthogonal coordinate system 10 fixed to the measured object. At this time, when the measured object is moving and changing its posture, the plurality of magnetic vectors m ₁ , m ₂ , and m ₃ detected in time series are different from each other.

The triaxial acceleration sensor 22 is also fixed to a part of the object to be measured, and detects gravitational acceleration as acceleration vectors a ₁ , a ₂ , and a ₃ in time series. The acceleration vectors a ₁ , a ₂ , and a ₃ are also vectors starting from the origin O of the three-axis orthogonal coordinate system 10, and when the object to be measured moves and changes its posture, it is time-sequentially. The detected plurality of acceleration vectors a ₁ , a ₂ , and a ₃ are different from each other.

The data of the magnetic vectors m ₁ , m ₂ , m ₃ detected in time series and the data of the acceleration vectors a ₁ , a ₂ , a ₃ are stored in the memory 3 in the computer 11 together with the time information of the measured moment. The data is sent and stored as time-series data (S2 in FIG. 2). Further, at this time, the magnetic vector and acceleration vector data sent to the memory 3 may be regarded as m ₁ , m ₂ , m ₃ and a ₁ , a ₂ , a ₃ by applying a digital low-pass filter. As a result, noise in the measurement value can be reduced, and more accurate angles and angular velocities can be calculated.
Based on these data, the rotation of the measured object is performed by the first rotation axis determination means 41 and the first rotation angle calculation means 51, or by the second rotation axis determination means 42 and the second rotation angle calculation means 52. An angle (posture change amount) is calculated (S3 to S6 in FIG. 2).

First, a rotation angle calculation method by the first rotation axis determination unit 41 and the first rotation angle calculation unit 51 will be described.
In the first rotation axis determination means 41, the rotation axis K of the measured object is calculated based on the magnetic vector data at two or more different points accumulated in the memory 3. In this example, in order to determine an arbitrary rotation axis that is not limited to three axes of the three-axis orthogonal coordinate system 10 as the rotation axis K, data of magnetic vectors m ₁ , m ₂ , and m ₃ at three or more different time points are used. .

That is, for example, magnetic vector data at _three different time points (t ₁ , t ₂ , t ₃ ) are read from the memory 3. The magnetic vector at time t ₁ is m ₁ = (m _1x , m _1y , m _1z ), the magnetic vector at time t ₂ is m ₂ = (m _2x , m _2y , m _2z ), and at time t ₃ . Let the magnetic vector be m ₃ = (m _3x , m _3y , m _3z ). Here, the intervals between the times t ₁ , t ₂ , and t ₃ are several milliseconds or less.

The end points M ₁ , M ₂ and M ₃ of these magnetic vectors exist on one data plane S in the three-axis orthogonal coordinate system 10 as shown in FIG. Will exist. An axis orthogonal to the data plane S and passing through the center of the locus circle Q is the rotation axis K.
Although the number of magnetic vector data is three here, it is possible to draw an average orbit circle passing through four or more of them, and the more accurate the magnetic vector data, the more accurate the calculation. It becomes possible.

Therefore, first, as shown in FIG. 5, a difference vector n ₁ that is the difference between the magnetic vectors m ₁ and m ₂ and a difference vector n ₂ that is the difference between the magnetic vectors m ₃ and m ₂ are calculated. Then, the X, Y, and Z components of the difference vectors n ₁ and n ₂ can be arranged as in the following formulas (1) and (2).

_{n 1 = m 1 -m 2 =} (m 1x -m 2x, m 1y -m 2y, m 1z -m 2z) = (n 1x, n 1y, n 1z)
... (1)
_{n 2 = m 3 -m 2 =} (m 3x -m 2x, m 3y -m 2y, m 3z -m 2z) = (n 2x, n 2y, n 2z)
... (2)

Then, by taking the outer product n ₁ × n ₂ of the difference vector n ₁ and the difference vector n ₂ , a rotation axis vector is obtained as a vector perpendicular to the difference vectors n ₁ and n ₂ , that is, a vector perpendicular to the data plane S. k can be obtained as shown in equation (3) below.
_{k = n 1 × n 2 =} (n 1y n 2z -n 1z n 2y, n 1z n 2x -n 1x n 2z, n 1x n 2y -n 1y n 2x)
= (K ₁ , k ₂ , k ₃ ) (3)

  The rotation axis vector k calculated here is greatly influenced by the noise of the triaxial magnetic sensor 2. Moreover, generally the change of the rotating shaft of a to-be-measured body is continuous, and a sudden change cannot be considered. From the above, when the calculation accuracy of the rotation axis vector k does not reach practical use, a value obtained by applying a digital low-pass filter to k may be regarded as the rotation axis vector k. At that time, when the rotation direction is rapidly reversed or when the rotation speed is low, the directions of k at two or more different time points are reversed, and there is a possibility that a problem may occur during the smoothing process of the filter or the like. For this, an angle formed by k at the previous time and k at the current time is calculated using a vector inner product calculation or the like, and if it is 90 degrees or more, the current time k is reversed. Also good.

  When the triaxial magnetic sensor 2 has no offset error and the rotational axis K of the measured object passes through the origin O of the triaxial orthogonal coordinate system 10, it is parallel to the rotational axis vector k obtained as described above. A straight line passing through the origin O of the three-axis orthogonal coordinate system 10 can be set as the rotation axis K. The point where the rotation axis K and the data plane S intersect is the center coordinate C of the locus circle Q. Therefore, the first rotation angle calculation means 51 obtains the center coordinate C as the intersection of the rotation axis K and the data plane S as follows.

As shown in FIG. 4, the magnitude of the center coordinate vector OC is the magnetic vector m ₁ (or m ₂ or m ₃ ) having the end point M ₁ (or M ₂ or M ₃ ) on the rotation axis vector k and the locus circle Q. It can be obtained by the inner product. The direction of the center coordinate vector OC is the same as that of the rotation axis vector k. Therefore, when the center coordinate vector OC is set as a vector ak, the following equation (4) is established. Here, a is an arbitrary coefficient.
k · ak = k · m ₁ (4)

From this equation (4), the coefficient a can be obtained as in the following equation (5).
_{a = (m x k x +} m y k y + m z k z) / (k x 2 + k y 2 + k z 2) ··· (5)
Since the center coordinate vector OC in the data plane S is equal to ak, the center coordinate C (center coordinate vector OC) is obtained by (ak _x , aky _y , ak _z ).

The center coordinate C obtained in this way is the center of the locus circle Q. Then, the center coordinate vector OC with the center coordinate C as the end point and the magnetic vector m ₁ (or m ₂ or m _{3 with} the point M ₁ (or M ₂ or M ₃ ) on the circumference of the locus circle Q as the end point. ), The radius R of the locus circle Q is obtained by the following equation (6).
R ² = (m ₁ -OC) ² = (m ₂ -OC) ² = (m ₃ -OC) ² (6)

Here, only one magnetic vector may be used for the calculation, but more accurate calculation can be performed by taking an average using the _three data m ₁ , m ₂ , and m ₃ . Further, by using the data of four or more magnetic vectors and obtaining the average of the radii R calculated based on them, more accurate calculation can be performed.

Based on the radius R obtained as described above, the rotation angle is calculated as follows.
For example, as shown in FIG. 5, when the magnetic vector changes from m ₁ to m ₂ from time t ₁ to time t ₂ , the rotation angle θ in the locus circle Q is calculated by the following equation (7). The

As shown in FIG. 5, the above equation (7) is expressed as a sine in a right triangle M ₁ M ₂ G inscribed in the locus circle Q and having two vertices at the end points M ₁ and M ₂ of the magnetic vectors m ₁ and m _2. It can be obtained by applying the theorem. That is, in the triangle M ₁ M ₂ G, the following formula (8) is established. Here, the line segment M ₁ G corresponds to the diameter 2R of the locus circle Q. Since the angle M ₁ GM ₂ and the angle M ₁ CM ₂ have a relationship between the circumferential angle and the center angle, the angle M ₁ GM ₂ is half of the angle M ₁ CM ₂ (that is, the rotation angle θ). .

Therefore, the equation (7) is derived from the equation (8), and the rotation angle θ can be obtained. A signal of the rotation angle θ is output from the output means 8 (FIG. 1).
As described above, since the rotation axis K and the rotation angle θ around the rotation axis K can be obtained, the amount of change in posture of the measured object from time t ₁ to time t ₂ can be known.

Next, a calculation method of the rotation angle by the second rotation axis determination means 42 and the second rotation angle calculation means 52 will be described.
In the second rotation axis determination means 42, based on the data of the magnetic vectors m ₁ and m _{2 and} the data of the acceleration vectors a ₁ and a ₂ at _two different time points stored in the memory 3, A rotation axis K that passes through the origin and serves as a reference for the rotational movement of the object to be measured is determined.

That is, for example, the magnetic vector and acceleration vector data at _two different time points (t ₁ , t ₂ ) are read from the memory 3. Then, the magnetic vector and the acceleration vector at time t ₁ are m ₁ = (m _1x , m _1y , m _1z ) and a ₁ = (a _1x , a _1y , a _1z ), respectively, and the magnetic vector at time t ₂ and The acceleration vectors are m ₂ = (m _2x , m _2y , m _2z ) and a ₂ = (a _2x , a _2y , a _2z ), respectively. Further, since the second rotation axis determination means 42 and the second rotation angle calculation means 52 are used in a low rotation area as will be described later, the m ₁ , a ₁ , m ₂ , and a ₂ obtained above are digitally displayed. the low-pass filter which is reacted with _{m 1, a 1, m 2} , may be regarded as a _2.

First, for example, the unit vector e _E in the east direction is calculated using the outer product of the magnetic vector and the acceleration vector at time t ₁ by the following equation (9).
Further, the unit vector e _U in the vertical direction can be calculated by the following equation (11) because the direction coincides with the acceleration vector a ₁ .
Further, the unit vector e _N in the north direction can be calculated by the following equation (10) from the outer product of the unit vector e _{E in} the east direction and the unit vector e _{U in the} vertical direction.
Then, the X, Y, and Z direction components of each unit vector are respectively set as in the following formulas (9) to (11).

By combining these unit vectors, the posture matrix P (t ₁ ) of the measured object (three-axis orthogonal coordinate system 10) at time t ₁ is obtained as in the following equation (12).

Similarly to the above, the posture matrix P (t ₂ ) of the measured object (three-axis orthogonal coordinate system 10) at time t ₂ is also calculated.
Then, these two orientation matrix are in a relationship that matches the rotation matrix R representing the rotation of the object to be measured (t ₂₎ between the two time (t _1, t _2). That is, it has the relationship of the following formula (13).

P (t ₁ ) R (t ₂ ) = P (t ₂ ) (13)
From this relationship, R is obtained by the following equation (14).
R (t ₂ ) = P ⁻¹ (t ₁ ) P (t ₂ ) (14)

Assuming that this rotation (posture change) is a rotation around an arbitrary rotation axis, the rotation matrix (relative posture matrix) R (t ₂ ) is a rotation axis vector k = (k) which is a unit vector in the direction of the rotation axis. _x , k _y , k _z ) and the rotation angle θ around the _x , k _y , k _z ) can be expressed as the following equation (15).

  Based on the above formula (15), the following formulas (16) to (18) are derived.

The rotation axis having the rotation axis vector k thus derived is the rotation axis K to be obtained.
Then, the rotation angle θ around the rotation axis K is calculated as follows.
That is, from the above equation (15), the following equations (19) and (20) can be derived, and the rotation angle θ of the measured object can be obtained as in equation (21).

  Next, the rotational speed discriminating means 6 discriminates whether the object to be measured is rotating at a speed higher than the reference (high speed mode) or rotating at a speed lower than the reference (low speed mode), and the determination result. Based on the above, it is determined which of the rotation angle θ calculated by the first rotation angle calculation means 51 and the rotation angle θ calculated by the second rotation angle calculation means 52 is to be adopted (S7 in FIG. 2).

Here, as a method for discriminating between the high speed mode and the low speed mode, for example, there are the following methods.
That is, the time Δt ₁ and Δt ₂ required for the change of the difference vectors n ₁ and n ₂ calculated above are calculated, and the time change rate vectors v ₁ and v ₂ of the difference are calculated using them.
Let L be the outer product of the difference time change rate vectors v ₁ and v ₂ obtained here.

Here, L points in the same direction as the rotation axis vector k, and its magnitude is proportional to the square of the rate of change per unit time of the magnetic vector used for calculating the rotation axis and the rotation angle. Since the rate of change of the magnetic vector per unit time increases as the rotation speed of the object to be measured increases, L is greater than or equal to a predetermined magnitude (for example, 5000 (mG / sec) ² [square of milligauss per second]). It can be determined depending on.
That is, for example, when L <5000 (mG / sec) ² , it is determined that the mode is the low speed mode, and when L ≧ 5000 (mG / sec) ² , it is determined that the mode is the high speed mode.

Alternatively, there is the following method as another method.
First, in the triaxial magnetic sensor 21, the difference between the detected magnetic vector m ₂ and the magnetic vector m ₁ collected immediately before (the magnitude of the difference vector n ₁ ) exceeds a predetermined magnitude (for example, 100 mG). Is stored (collected) in the memory 3 as the next data.
Therefore, it is possible to present sampling time t ₂ the data of the magnetic vector is from the previous sampling time t _1, the determination by whether or not a predetermined time has elapsed (e.g., 500m seconds) or more. That is, for example, when t ₂ −t ₁ = Δt <500 msec, the high speed mode is determined, and when Δt ≧ 500 msec, the low speed mode is determined.

Alternatively, the determination is made based on whether or not the difference (magnitude of the difference vector n ₁ ) between the newly collected magnetic vector m ₂ and the previously collected magnetic vector m ₁ is greater than or equal to a predetermined magnitude (for example, 250 mG [milli gauss]). You can also That is, for example, when the magnitude of the difference vector n ₁ is 250 mG or more, it is determined that the mode is the high speed mode, and when it is less than 250 mG, it is determined that the mode is the low speed mode.

Alternatively, only when the above two conditions, that is, the sampling time interval Δt is less than a predetermined time (eg, 500 msec) and the magnitude of the difference vector n ₁ is a predetermined size (eg, 250 mG) or more. It can also be determined that the mode is the high speed mode, and in other cases, it can be determined that the mode is the low speed mode.

  When the rotational speed discriminating means 6 determines that the mode is the high speed mode in this way, the rotational angle θ calculated by the first rotational angle calculating means 51 is adopted (S8 in FIG. 2). On the other hand, when the rotation speed determination means 6 determines that the low speed mode is selected, the rotation angle θ calculated by the second rotation angle calculation means 52 is adopted (S9 in FIG. 2).

Then, the adopted rotation angle θ is smoothed and made continuous by the low-pass filter (S10 in FIG. 2).
This low-pass filter can be, for example, a low-pass filter having a second-order IIR Butterworth characteristic with a cutoff of 5 Hz when the sampling rate is 50 Hz.

Also, the angular velocity ω can be obtained by dividing the corrected rotation angle θ by the interval Δt between the two time points using the measurement time information data stored in the memory as described above (see FIG. 2). S12). That is, ω = θ / Δt.
As described above, since the rotation axis K, the rotation angle θ and the rotation angular velocity ω around the rotation axis K can be obtained, the posture change amount and the posture change speed of the measured object can be known.
The signals of the rotation angle θ and the angular velocity ω are output from the output means 8 (S11 and S13 in FIG. 2).

Next, the function and effect of this example will be described.
The magnetic gyro 1 has a rotational speed discriminating means 6. The output unit 8 outputs the calculation result of the rotation angle of the measurement object by the first rotation angle calculation unit 51 when the rotation speed determination unit 6 determines that the mode is the high speed mode. The calculation result of the rotation angle of the measurement object by the second rotation angle calculation means 52 is output.

  Here, the first rotation angle calculation means 51 is configured to calculate the rotation angle of the measurement object around the rotation axis based on the magnetic vector data. In other words, since the acceleration vector data is not used in calculating the rotation angle, even if the measured object rotates at high speed, the accompanying acceleration does not become a noise when calculating the rotation angle, and the accurate rotation angle. Can be calculated. Further, since the first rotation angle calculation means 51 is used in the high speed mode, the first rotation axis determination means 41 can quickly determine the rotation axis only from the magnetic vector data, and the rotation axis is centered. The first rotation angle calculation means 51 can quickly calculate the rotation angle.

On the other hand, the second rotation angle calculation means 52 is configured to calculate the rotation angle of the measured object based on the magnetic vector data and the acceleration vector data. That is, in calculating the rotation angle, the acceleration vector data is used in addition to the magnetic vector data. Therefore, even in the low speed mode, the second rotation axis determination means 42 uses the magnetic vector data and the acceleration vector data. Thus, the rotation axis can be determined quickly. Then, the second rotation angle calculation means 52 can quickly calculate the rotation angle of the measured object on the basis of the rotation axis determined thereby.
In the low speed mode in which the object to be measured is rotating at a low speed, the acceleration accompanying the rotational motion, that is, the acceleration other than the gravitational acceleration is sufficiently small and can be ignored. Therefore, this acceleration does not cause noise when calculating the rotation angle.

  In this way, by properly using the rotation angle calculation means in the high speed mode and the low speed mode, the magnetic gyro 1 can be used except for the case where the direction of the rotation axis, which will be described later, matches or substantially matches the direction of the magnetic vector. Thus, it is possible to accurately detect the rotation angle in any movement state of the measured object.

In addition, since the magnetic gyro 1 of this example includes the angular velocity calculation means 7, it is possible to easily detect the rotational angular velocity (posture change velocity) of the measured object.
Further, when the high speed mode is selected in this example, the first rotation axis determination means 51 uses the arbitrary rotation axis as the rotation axis K based on the magnetic vector data obtained at the moments of three or more different times. It is configured to be determined. In this example, when the low speed mode is selected, an arbitrary rotation axis is determined as the rotation axis K based on magnetic vector and acceleration vector data obtained at two or more different moments. It is. Therefore, even if the user rotates the object to be measured at any speed and at any speed, the actual instantaneous rotation axis of the object to be measured at each instant can be accurately determined. An accurate rotation angle θ can be measured.

The first rotation angle calculation means 51 calculates the center coordinate C by calculating the intersection of the data plane S and the rotation axis K, and the center coordinate C and the magnetic vector coordinate points M ₁ , M ₂ , M _3. By calculating the radius of the locus circle Q and calculating the rotation angle θ. Thereby, the rotation angle θ can be calculated easily and accurately.

  Further, as shown in FIG. 6, since the triaxial magnetic sensor 21 is constituted by the magneto-impedance sensor element 20, it is possible to obtain a magnetic gyro 1 with higher accuracy, higher sensitivity, higher response, and a smaller size. Can do. That is, as shown in FIG. 7, the magneto-impedance sensor element (MI) has a high sensitivity and extremely short measurement compared to other magnetic sensors such as a hall sensor (Hall) and a magnetoresistive element (MR). Since it is possible to measure with low noise and high accuracy in time (for example, 1 ms), for example, it is extremely rotated so that it can be intentionally swung and used for a user to enjoy a golf game stored in a mobile phone. Even when the speed is high and a plurality of magnetic vector data must be continuously measured at very short time intervals, the geomagnetic vector can be detected with high accuracy. Therefore, even in such a case, the instantaneous rotation axis, rotation angle, and rotation speed of the measurement object can be accurately obtained. Furthermore, since the magneto-impedance sensor element 20 is small, a small three-axis magnetic sensor 21 can be obtained. This also allows the magnetic gyro 1 to be housed in the IC chip.

The magnetic gyro 1 includes the low-pass filter. Thereby, even if the high speed mode and the low speed mode are switched during the measurement, the calculation result of the rotation angle of the measured object is also one or more immediately before the high speed mode and the low speed mode are switched. The data can be smoothed and continuous.
Therefore, a measurement error associated with switching between the high speed mode and the low speed mode can be reduced. That is, when the high-speed mode and the low-speed mode are switched, the rotation angle calculation means used is switched between the first rotation angle calculation means 51 and the second rotation angle calculation means 52, and accordingly, the rotation axis to be used is determined. The means is switched between the first rotation axis determination means 41 and the second rotation axis determination means 42. Therefore, before and after the switching, the rotation axis K used for the calculation changes, and the calculation result may change discontinuously. However, by smoothing and continuing this change, the calculation error of the rotation angle θ can be reduced.

  As described above, according to the present example, it is possible to provide a magnetic gyro capable of accurately measuring the rotation angle during both high-speed rotation and low-speed rotation.

(Example 2)
In this example, as shown in FIGS. 8 and 9, even when the triaxial magnetic sensor 2 has an offset error and the rotation axis K of the measured object does not pass through the origin O of the triaxial orthogonal coordinate system 10, it is accurate. It is an example of a magnetic gyro capable of measuring a rotation angle of a measurement object.
That is, in Example 1, there is no offset error in the triaxial magnetic sensor 2 or the offset error is negligible, and the rotation axis K of the measured object passes through the origin O of the triaxial orthogonal coordinate system 10, or We introduced a magnetic gyro that can easily measure the rotation angle of the object to be measured when there is no problem even if it passes.

  However, in actual use, the origin of the three-axis orthogonal coordinate system 10 depends on various factors such as the influence of a magnetic field generated by magnetizing a magnetic body attached to the object to be measured and the change in temperature characteristics of the sensor. It may deviate from the origin of the geomagnetic measurement by the sensor. It is conceivable that the deviation of the origin is individually corrected for each measurement result. However, in this case, since an operation called correction calculation is performed, the measurement may be complicated and the configuration may be complicated.

  Therefore, in this example, as shown below, the accurate rotation angle can be directly measured regardless of the presence or absence of an offset error by appropriate processing in the calculation process of the rotation angle of the object to be measured. is there.

That is, in this example, the first rotation angle calculation means 51 includes the following rotation center coordinate calculation means and radius calculation means.
Based on the magnetic vector data at three or more different points in the three-axis orthogonal coordinate system 10, the rotation center coordinate calculating means calculates the center coordinates of the locus circle Q passing through the coordinate points of the three or more magnetic vectors.
The radius calculation unit calculates the radius of the locus circle by calculating the distance between the center coordinate calculated by the rotation center coordinate calculation unit and the coordinate point of the magnetic vector.
Then, the rotation angle θ is calculated based on the radius of the locus circle Q calculated by the radius calculation means and the coordinate points of the magnetic vectors at two different time points.

This will be specifically described below.
First, there is an offset error in the triaxial magnetic sensor 2, that is, for example, a magnetic vector (for example, m ₁ in FIG. 8) detected by the triaxial magnetic sensor 2 is a geomagnetic vector (for example, m ₁ ′ in FIG. 8). Next, a case in which a magnetic vector that becomes noise is a combined vector will be described.

In this case, the geomagnetic vector (m ₁ ′) does not pass through the origin O of the three-axis orthogonal coordinate system 10 but passes through a point O ′ that deviates from the origin O. The point O ′ is also a point through which the rotation axis K passes. The intersection of the data plane S and the rotation axis K obtained in the first embodiment is the center coordinate C ′ of the locus circle Q as shown in FIG. That is, the center coordinate vector is OC ′. The rotation center coordinate calculation means obtains an accurate center coordinate C ′ by obtaining the center coordinate vector OC ′, and can obtain an accurate rotation angle of the end point of the magnetic vector around the center coordinate C ′. Therefore, it is possible to obtain an accurate rotation angle of the measurement object around the rotation axis K.

Here, after calculating the relative coordinate vector M ₂ C ′ having the origin (coordinate point) M ₂ of the magnetic vectors m ₁ , m ₂ , m ₃ obtained in the same manner as in the first embodiment, the central coordinate vector OC ′. = OM ₂ + M ₂ C ′ is calculated. A method for calculating the center coordinate vector OC ′ will be described below.
First, M ₂ C ′ = (a, b, c) is set, and the other coordinates M ₁ , M ₂ , M ₃ are also subjected to coordinate conversion as follows.

M ₁ M ₂ = m ₁ −m ₂ = (m _1x −m _2x , m _1y −m _2y , m _1z −m _2z )
= N ₁ = (n _1x , n _1y , n _1z ) (22)
M ₂ M ₂ = m ₂ −m ₂ = (0, 0, 0) (23)
M ₃ M ₂ = m ₃ −m ₂ = (m _3x −m _2x , m _3y −m _2y , m _3z −m _2z )
= N ₂ = (n _2x , n _2y , n _2z ) (24)

At this time, if the radius of the locus circle Q is R, as shown in FIG. 9, the above three points are the circle equations (X-a) ² when the points on the circumference are (X, Y, Z). + (Y−b) ² + (Zc) ² = R ² is satisfied. That is, the following formulas (25) to (27) are established.
(N _1x −a) ² + (n _1y −b) ² + (n _1z −c) ² = R ² (25)
a ² + b ² + c ² = R ² (26)
(N _2x −a) ² + (n _2y −b) ² + (n _2z −c) ² = R ² (27)
Since the vector M ₂ C ′ exists in the data plane S, it is orthogonal to the rotation axis vector k, and the following equation (28) is established.
ak _x + bky _y + ck _z = 0 (28)

When the above is arranged, the following equations (29) to (31) are obtained.
an _1x + bn _1y + cn _1z = (n _1x ² + n _1y ² + n _1z ² ) / 2 (29)
an _2x + bn _2y + cn _2z = (n _2x ² + n _2y ² + n _2z ² ) / 2 (30)
ak _x + bky _y + ck _z = 0 (31)
By solving the simultaneous equations using linear algebra, a vector M ₂ C ′ = (a, b, c) is obtained as in the following equations (32) and (33).

From M ₂ C ′ = (a, b, c) calculated as described above, the center coordinate C ′ (center coordinate vector OC ′) becomes OC ′ = OM ₂ + M ₂ C ′ = (m _2x + a, m _2y + b, m _2z + c).
The radius calculation means calculates the radius R of the locus circle Q by calculating the distance between the center coordinate C ′ calculated as described above and the magnetic vector coordinate points M ₁ , M ₂ , M ₃ .
Since the central coordinate C ′ based on the central coordinate vector OC ′ obtained here can be calculated as C in the equation (6) shown in Embodiment 1, there is no problem. The radius R is obtained by substituting the data. Subsequent calculations are performed in the same manner as in the first embodiment by replacing the center coordinates C with the center coordinates C ′, so that the rotation angle of the measured object can be accurately calculated.

When the three-axis magnetic sensor 2 has no offset error and the rotation axis K of the measured object passes through the origin O of the three-axis orthogonal coordinate system 10, the above-described center coordinate C ′ is parallel to the rotation axis vector k. And the rotation axis K passing through the origin O of the three-axis orthogonal coordinate system 10 coincides with the central coordinate C (see FIGS. 3 to 5 and 9) intersecting the data plane S. Therefore, also in this case, the rotation angle of the object to be measured can be accurately calculated by performing a series of calculations similar to the case where there is the offset error described above.
Others are the same as in the first embodiment.

  In the case of this example, an accurate rotation angle can be directly measured regardless of the presence or absence of an offset error by an appropriate process in the calculation process of the rotation angle of the measured object. In addition, since the rotation angle of the measured object is calculated by directly obtaining the rotation center coordinates C ′ of the magnetic vector without calculating the offset origin O ′, no special correction calculation is required.

Therefore, even if an offset error occurs in the triaxial magnetic sensor 2 with a simple configuration, the rotation angle of the measured object can be accurately measured. That is, it is possible to prevent a phenomenon in which a measurement error occurs due to a drift of the measurement origin with use time as in a conventional vibration gyroscope with a simple configuration.
In addition, the same effects as those of the first embodiment are obtained.

When the center coordinates are calculated by the same method as in the first embodiment when there is an offset error, C shown in FIGS. 8 and 9 is obtained as the center coordinates, and M ₁ and M ₂ centered on this C are obtained. The center angle η is calculated as the rotation angle of the measured object. Therefore, a deviation occurs from the original rotation angle θ. On the other hand, according to this example, it is possible to directly measure the rotation angle θ without such a shift.

(Example 3)
In this example, the first rotation axis determination means 41 uses the axis in the direction in which the change of the magnetic vector component is the smallest among the three axes (X axis, Y axis, Z axis) of the three axis orthogonal coordinate system 10 as the rotation axis. Is an example of a magnetic gyro 1 configured to be determined as That is, in the rotation axis calculation means described in the first or second embodiment, magnetic vector data at three or more different points at least are necessary for calculating the rotation axis, and necessary calculations are complicated. Therefore, it is not always necessary when the rotation axis with such high accuracy is not required or when a certain level of high accuracy is required, but the actual rotation axis is close to any one of the three-axis orthogonal coordinate systems. There is a case where it is not necessary to employ a rotation axis determination method that is not limited to the three axes of the orthogonal coordinate system as described in Example 1 or 2. This example is effective as a rotation axis determination method that can simplify the calculation in such a case.

In the case of this example, for example, using the data of the geomagnetic vectors m ₁ and m ₂ collected at two different time points, the difference vector n ₁ between them is calculated. That is, the difference vector n ₁ shown in the expression (1) in the first embodiment is obtained.
The axis in the direction of the smallest component among the X, Y, and Z components (n _1x , n _1y , n _1z ) in the difference vector n ₁ is set as the rotation axis.

Specifically, for example, the following determination method can be considered.
That is, the magnitude of each component of the difference vector n ₁ is compared with a certain threshold value N (for example, 20 mG).
First, when all of the components of the difference vector n ₁ , n _1x , n _1y , n _1z , are larger than the threshold value N, it is determined as “determination impossible”. In this case, the rotation axis may be obtained by the rotation axis calculation method described in the first or second embodiment. Next, when the change in the measured object is static or at a very slow speed, all of n _1x , n _1y , and n _1z are smaller than the threshold value N, so that the rotation axis is calculated by the second rotation axis calculation means described above. Or all the axes of the 3-axis Cartesian coordinate system 10 are “determined to be stationary”, and the instantaneous rotation angle and rotation angular velocity at this point are determined to be 0. To do.
On the other hand, when any one or two components of n _1x , n _1y , and n _1z are smaller than the threshold value N, the rotation axis can be determined as follows.

Arbitrary axes of the X axis, the Y axis, and the Z axis constituting the orthogonal axis of the three-axis orthogonal coordinate system 10 are defined as an i axis, a j axis, and a k axis.
First, a case will be described in which any one component of n _1i , n _1j , and n _1k , which is each component of the difference vector n ₁ in the i-axis, j-axis, and k-axis directions, is smaller than the threshold value N.
For example, when the measured object rotates around the i-axis, the component in the i-axis direction of the difference vector n ₁ does not change, so n _1i is smaller than the threshold value N. The components in the other two axial directions, i.e., the j-axis direction and the k-axis direction, change greatly, so that n _1j and n _1k are larger than the threshold value N. Therefore, the i-axis that is the direction of the component smaller than the threshold value N is determined as the rotation axis.

Next, a case where any two components n _1i , n _1j , and n _1k that are components of the difference vector n ₁ are smaller than the threshold value N will be described.
Depending on the positional relationship that the triaxial magnetic sensor 21 makes with the geomagnetism, for example, the two components of the component n _1i in the i-axis direction and the component n _{1j in the} j-axis direction are smaller than the threshold value N, and 1 in the k-axis direction. Only the component n _1k may be larger than the threshold value N. At this time, the i axis and the j axis are listed as candidates for the rotation axis.

In this case, the rotation axis of rotation between the two different previous time points (t ₀ , t ₁ ) recorded in the memory 3 is selected as the current rotation axis from the i-axis and the j-axis. For example, when the previous rotation axis is the i-axis, the current rotation axis is also determined as the i-axis, and when the previous rotation axis is the j-axis, the current rotation axis is also determined as the j-axis. However, in the case where the rotation axes listed as candidates are the i-axis and the j-axis as described above, if the previous rotation axis is the k-axis or cannot be determined, the current rotation axis cannot be determined. In this case as well, the determination by the rotation axis calculation method is impossible, but the rotation axis may be calculated by switching to the rotation axis determination method described in the first or second embodiment.

Based on the rotation axis obtained as described above, the first rotation angle calculation means 51 obtains the rotation angle of the measurement object as follows.
For example, when the rotation axis is determined as the i-axis and the magnetic vector changes from m ₁ to m ₂ from time t ₁ to time t ₂ , the rotation angle θ around the i-axis is expressed by the following equation (34). Desired.
θ = tan ⁻¹ (m _2k / m _2j ) −tan ⁻¹ (m _1k / m _1j ) (34)

Here, when the rotation axis i is the Z axis,
θ = tan ⁻¹ (m _2y / m _2x ) −tan ⁻¹ (m _1y / m _1x ) (35)
Thus, the rotation angle θ with the Z axis as the rotation axis can be obtained. A signal of the rotation angle θ is output to the output means 8 (FIG. 1).

Also, using this result, the rotational angular velocity calculation means 7 can obtain the rotational angular velocity ω of the measured object around the rotational axis by dividing the rotational angle θ by Δt. That is, ω = θ / Δt. A signal of this rotational angular velocity ω is output to the output means 8 (FIG. 1).
Others are the same as in the first embodiment.

In the case of this example, the calculation performed in the first rotation axis determination means 41 can be simplified, and the rotation axis can be easily determined. Further, if any of the three axes (X axis, Y axis, Z axis) orthogonal to each other can be set as the rotation axis, it is possible to effectively measure the posture change of, for example, a mobile phone as the measured object. . In other words, the direction of rotation given to a mobile phone or the like is often the direction of yaw, roll, or pitch, based on the case, so if you can identify which of these rotations, you can input by rotational movement. Because it becomes.
In addition, the same effects as those of the first embodiment are obtained.

Example 4
In this example, when the direction of the actual rotation axis of the measured object coincides with the direction of the magnetic vector, the third rotation angle that calculates the rotation angle of the measured object around the rotation axis based on the acceleration vector data It is an example of the magnetic gyroscope 1 which has a calculation means.
That is, when the direction of the actual rotation axis of the measured object coincides with the direction of the magnetic vector, the magnetic vector in the three-axis orthogonal coordinate system 10 does not change with time. Therefore, the first rotation axis determination means 41 and the second rotation axis determination means 42 using the magnetic vector data cannot calculate the rotation axis, and the first rotation angle calculation means 51 and the second rotation angle calculation means. 52 cannot be used.

  Therefore, in this case, the third rotation angle calculation means calculates the rotation angle of the measured object using the acceleration vector data obtained by the triaxial acceleration sensor 22. Contrary to the above, if the time change of the acceleration vector data does not occur despite the change of the magnetic vector data, only some magnetic noise is added and the rotation is not performed. It may be determined that the rotation angle need not be calculated.

  First, whether or not the direction of the actual rotation axis of the object to be measured matches the direction of the magnetic vector satisfies the condition that the change in the magnetic vector is extremely small and the change in the acceleration vector is large to some extent. Do by. In other words, the only possible situation where the acceleration vector changes to some extent and the magnetic vector hardly changes is the situation where the measured object is rotating with the rotation axis coinciding with the magnetic vector (geomagnetism). Absent.

Therefore, for example, the magnitude of the difference vector n ₁ shown in the expression (1) of the first embodiment is used as the magnitude of the change in the magnetic vector. Further, as the magnitude of the change in the acceleration vector, the magnitude of the difference vector b ₁ = a ₁ −a ₂ between the acceleration vectors a ₁ and a ₂ at time t ₁ and time t ₂ is used. When the magnetic difference vector n ₁ is less than a predetermined value (for example, 20 mG) and the acceleration difference vector b ₁ is larger than a predetermined value (for example, 0.2 g [g is gravitational acceleration]), It is determined that the direction of the actual rotation axis coincides with the direction of the magnetic vector.

Since the direction of the rotation axis at this time substantially coincides with the geomagnetic vectors m ₁ and m ₂ , the rotation axis vector (unit vector) k is obtained by, for example, k = m ₁ / | m ₁ | be able to.
Then, based on the time change of the acceleration vector detected by the triaxial acceleration sensor, the rotation angle around the rotation axis that matches the geomagnetic vector is calculated.

Specifically, for example, in the rotation angle calculation method by the first rotation angle calculation means 51 described in the first embodiment, the rotation angle can be calculated based on the acceleration vector data by replacing the magnetic vector with the acceleration vector. It is.
The calculation accuracy decreases as the rotation speed of the object to be measured increases. However, if the rotation is slow to some extent, it is possible to obtain a calculation result with sufficient accuracy.
Others are the same as in the first embodiment.

In the case of this example, even when the magnetic vector does not change with the rotation of the measured object, the third rotation angle calculation means calculates the rotation angle using only the acceleration vector data. As a result, even when the direction of the actual rotation axis of the measured object coincides with the direction of the magnetic vector, the rotation angle of the measured object can be calculated.
In addition, the same effects as those of the first embodiment are obtained.

DESCRIPTION OF SYMBOLS 1 Magnetic gyro 21 3-axis magnetic sensor 22 3-axis acceleration sensor 3 Memory 41 1st rotation axis determination means 42 2nd rotation axis determination means 51 1st rotation angle calculation means 52 2nd rotation angle calculation means 6 Rotational speed discrimination means 7 Angular velocity calculation means 8 Output means

Claims (6)

A three-axis magnetic sensor for detecting geomagnetism as a magnetic vector in a three-axis orthogonal coordinate system fixed to the measurement object;
A triaxial acceleration sensor for detecting gravitational acceleration as an acceleration vector in the triaxial orthogonal coordinate system;
A memory for storing data of the magnetic vector detected in time series by the triaxial magnetic sensor and data of the acceleration vector detected in time series by the triaxial acceleration sensor;
First rotation axis determination means for determining a rotation axis as a reference for the rotational motion of the object to be measured, based on data of the magnetic vector at two or more different time points stored in the memory;
Second rotation axis determination means for determining a rotation axis as a reference for the rotational movement of the measured object based on the magnetic vector data and the acceleration vector data at two or more different points in time stored in the memory;
First rotation angle calculation means for calculating a rotation angle of the measurement object around the rotation axis determined by the first rotation axis determination means based on the magnetic vector data;
Second rotation angle calculation means for calculating the rotation angle of the measured object around the rotation axis determined by the second rotation axis determination means based on the magnetic vector data and the acceleration vector data; ,
Based on the magnetic vector data detected in time series by the three-axis magnetic sensor, the measured object is rotating at a speed higher than the reference rotation speed or less than the reference rotation speed. A rotational speed discriminating means for discriminating whether the low-speed rotation is performed,
When the rotational speed discriminating means determines that the object to be measured is rotating at the high speed mode, the rotation angle calculation means outputs the calculation result of the rotation angle of the object to be measured by the first rotation angle calculating means, An output for outputting the calculation result of the rotation angle of the measurement object by the second rotation angle calculation means when the rotation speed determination means determines that the measurement object is rotating at the low speed mode. And a magnetic gyro.   2. The magnetic gyro according to claim 1, wherein the first rotation angle calculation means is configured to coordinate the three or more magnetic vectors based on data of the magnetic vectors at three or more different time points in the three-axis orthogonal coordinate system. Rotation center coordinate calculation means for calculating the center coordinates of a trajectory circle passing through the point, and calculating the distance between the center coordinates calculated by the rotation center coordinate calculation means and the coordinate points of the magnetic vector, A radius calculating means for calculating the radius of the circle, and the magnetic vector used in the calculation based on the radius of the trajectory circle calculated by the radius calculating means and the coordinate points of the magnetic vector at two different time points. A magnetic gyro, which is configured to calculate a rotation angle between measurement times of data.   3. The magnetic gyro according to claim 1, wherein the rotation angle of the measured object calculated immediately after the high speed mode and the low speed mode are switched is at least the high speed mode and the low speed mode. Is corrected using one or more data immediately before switching, and smooth continuity between the rotation angle calculation result of the measured object calculated immediately before switching between the high speed mode and the low speed mode is obtained. A magnetic gyro characterized by being configured to ensure.   4. The magnetic gyro according to claim 1, wherein when the direction of the actual rotation axis of the measured object coincides with the direction of the magnetic vector, the first rotation angle calculation means or the second gyro. A magnetic gyro comprising third rotation angle calculation means for calculating the rotation angle of the measured object around the rotation axis based on the acceleration vector data instead of the rotation angle calculation means.   5. The magnetic gyro according to claim 1, wherein the three-axis magnetic sensor is configured by a magneto-impedance sensor element. 6.   The magnetic gyro according to any one of claims 1 to 5, wherein the magnetic gyro measures the geomagnetic vector continuously when the measured object is in a rotational motion state, and the measured geomagnetic vector. The instantaneous rotation axis at the time measured from the moment and the instantaneous rotation angle about the rotation axis are obtained, and the memory stores the magnetic vector data and the acceleration vector together with the measured time information. The first rotation axis determination means and the second rotation axis determination means are configured to instantaneously rotate the measured object within the measurement time of the magnetic vector at the two or more time points. The first rotation angle calculation means and the second rotation angle calculation means are configured to determine the measurement target with the rotation axis as a center. Magnetic gyro, characterized in that are configured to calculate the instantaneous angle of rotation of the.

Images