JP2013072643A - Magnetic measurement data calibration device and azimuth angle measurement device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a magnetic measurement data calibration device and an azimuth angle measurement device for improving followability to variation of an environmental offset.SOLUTION: A device for obtaining an offset of magnetic measurement data for each measurement axis outputted from a magnetic sensor comprising an axis sensor having three measurement axes orthogonal to each other and measuring earth magnetism in the measuring axis direction and correcting the magnetic measurement data includes: an error function calculation part for obtaining an error function from a difference between magnetic data for each measurement axis of the measured magnetic measurement data and a previous offset; an offset residual calculation part for calculating an offset residual from the error function and a previous covariance matrix; an offset updating part for adding the offset residual to the offset when the previous magnetic measurement data is measured, and calculating a new offset; and a covariance matrix updating part for updating the covariance matrix of the magnetic measurement data for which the previously measured magnetic measurement data is a parent population using the measured magnetic measurement data.

Description

  The present invention relates to a magnetic measurement data calibration apparatus and an azimuth angle measurement apparatus for calibrating a magnetic sensor mounted on a portable device.

In recent years, a three-axis magnetic sensor (hereinafter simply referred to as a magnetic sensor) is mounted on a mobile device such as a mobile phone and is used for orientation measurement in the mobile device.
The magnetic sensor used for this azimuth measurement is usually adjusted with a zero magnetic field output value for each axis (hereinafter referred to as offset adjustment) using a triaxial Helmholtz coil before being mounted on a portable device.
In this specific adjustment procedure, geomagnetism is measured with a magnetic sensor calibrated in advance, and magnetic data for each of the three axes output from the magnetic sensor calibrated in advance is used as reference magnetic measurement data for azimuth measurement.

Next, a magnetic field in the direction opposite to the reference magnetic measurement data is applied to the target magnetic sensor that is the target of offset adjustment by the three-axis Helmholtz coil.
Here, when the target magnetic sensor has no offset, the magnetic measurement data to be output is 0 for all three axes, and when it has the offset, the magnetic measurement data to be output is not 0.
Here, the magnetic measurement data output from the target magnetic sensor is measured as an offset of the target magnetic sensor itself, and the offset of the magnetic measurement data output from the magnetic sensor is adjusted based on the magnetic measurement data.
That is, the offset measured by the 3-axis Helmholtz coil is subtracted from each magnetic measurement data from the magnetic measurement data output from the magnetic sensor, and the direction is detected as the measurement result of the magnetic sensor.

  However, even if the offset adjustment is performed with the magnetic sensor alone, many other parts are used in the portable device in addition to the magnetic sensor that detects the orientation, and the magnetic field generated by these other parts is applied to the magnetic sensor. It may have an effect on it. For example, a magnetic field is generated from a magnet mounted on a speaker, a magnet provided on an open / close switch, or nickel plating applied to a component of a portable device.

For this reason, an environmental offset due to the surrounding environment occurs with respect to the magnetic sensor provided inside the portable device, and magnetic measurement data in which the environmental offset is superimposed is output from the magnetic sensor alone that has performed the offset adjustment.
Therefore, after performing offset adjustment with the magnetic sensor alone, the actual orientation that the portable device is facing and the orientation that is obtained from the magnetic data output by the magnetic sensor due to the magnetic field generated by other components of the mounted portable device In addition, a difference due to the environmental offset occurs.

In addition, it is conceivable to perform offset adjustment including environmental offset with a three-axis Helmholtz coil in a state where the magnetic sensor is mounted on a portable device together with other components.
However, the value of the environmental offset always fluctuates due to changes in the magnetic force of each component over time, changes in the temperature around the portable device, and the presence of an object that generates a strong magnetic field around the portable device. .

Therefore, even if the offset adjustment of the magnetic sensor including the environmental offset is performed after being mounted on the portable device by the three-axis Helmholtz coil, the environmental offset fluctuates every moment depending on the surrounding conditions. For this reason, the magnetic measurement data output from the magnetic sensor shifts from the true value due to the fluctuation of the environmental offset superimposed on the magnetic measurement data of the magnetic sensor.
For this reason, although it is necessary to perform offset again, the offset adjustment of the magnetic sensor mounted on the portable device cannot always be performed by the three-axis Helmholtz coil. Therefore, it is necessary to give the portable device the function of adjusting the offset of the magnetic sensor.

For this reason, the offset adjustment unit mounted on the portable device measures a number of magnetic field strengths of each axis by the magnetic sensors of each of the three axes, and temporarily stores the measured magnetic measurement data in an internal data buffer.
There is a technique in which the offset adjustment unit obtains an offset that minimizes the distance (magnetic field strength) with respect to the magnetic measurement data stored in the data buffer (for example, see Patent Document 1). In this method, as shown in the following equation, magnetic measurement data is substituted into a sphere equation, and the offset Sx is obtained by the least square method.
Sx = Σ {(X _i −X ₀ ) ² + (Y _i −Y ₀ ) ² + (Z _i −Z ₀ ) ² −R ² } = 0 (i)
X _i , Y _i and Z _i are magnetic measurement data, X ₀ , Y ₀ and Z ₀ are offset coordinates, and R is a constant.

In addition, the magnetic measurement data of each of the three-axis magnetic sensors is set as one set, and four sets of magnetic measurement data are stored in the data buffer. Similarly to Patent Document 1 described above, the offset is determined according to the distance from the magnetic measurement data. (For example, refer to Patent Document 2).
In Patent Document 2, when four sets of magnetic measurement data are extracted from measurement data measured in time series, extraction conditions for extraction are set as follows. That is, the following extraction conditions are set with the magnetic measurement data as coordinate points in a three-dimensional space constituted by three axes of the magnetic sensor.
When drawing (arranging) three-axis magnetic measurement data in a three-dimensional space (three-dimensional coordinate system),
a. The distance between the first set of magnetic measurement data and the second set of magnetic measurement data is sufficiently large b. The third set of magnetic measurement data is the apex of the obtuse angle of the obtuse triangle having the first and second sets of magnetic measurement data. C. The fourth set of magnetic measurement data is sufficiently separated from the plane formed by the first to third sets of magnetic measurement data.

And in patent document 2, the offset is calculated by solving simultaneous equations using the following spherical equations.
^{_{(X i -X 0) 2 +}} (Y i -Y 0) 2 + (Z i -Z 0) 2 -R 2 = 0 ... (ii)
X _i , Y _i and Z _i are magnetic measurement data, X ₀ , Y ₀ and Z ₀ are offset coordinates, and R is a constant.

Japanese Patent No. 4391416 Japanese Patent No. 4590511

  However, in Patent Document 1, a plurality of magnetic measurement data is accumulated in a data buffer, and an offset is calculated by the least square method using the equation (i) using the accumulated magnetic measurement data. Here, in Patent Document 1, after a large number of magnetic measurement data for the data buffer is measured, the offset is calculated. For this reason, Patent Document 1 requires time for accumulating a large number of magnetic measurement data, and the follow-up performance of the offset adjustment with respect to the change in the environmental offset is deteriorated.

Further, Patent Document 2 discloses that when the magnetic noise due to the temporal variation factor of the characteristics of the magnetic sensor itself and the temporal variation factor of the external magnetic field is superimposed on the magnetic measurement data output from the magnetic sensor, the calculated offset is accurate. Cannot be calculated.
That is, when magnetic noise is superimposed on any of the four sets of magnetic measurement data, the offset calculated using these magnetic measurement data includes an error due to magnetic noise, and the actual offset and It will shift.

Furthermore, in both Patent Document 1 and Patent Document 2, it is assumed that the measurement point of the magnetic measurement data for obtaining the offset is measured on the surface of a certain sphere.
For this reason, the user needs to move the portable device so as to draw a circle, that is, along a certain spherical surface at the time of offset adjustment.
When this movement is performed, a large number of magnetic measurement data is acquired and stored in the buffer, and the magnetic measurement data used for the calculation described above is extracted.
Therefore, there are cases where the environmental offset cannot follow the change, and new magnetic measurement data must be acquired again and the offset adjustment must be repeated.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide a magnetic measurement data calibration apparatus and an azimuth angle measurement apparatus that can improve the follow-up performance with respect to fluctuations in environmental offsets.

  The present invention has been made to solve the above-described problems, and a magnetic measurement data calibration apparatus according to the present invention has three measurement axes orthogonal to each other, and includes a magnetic sensor comprising an axis sensor that measures the geomagnetism in the measurement axis direction. An apparatus for obtaining an offset of magnetic measurement data that is a geomagnetic measurement result for each measurement axis output from a sensor, and correcting the magnetic measurement data, wherein the magnetic data for each measurement axis of the measured magnetic measurement data and An error function calculation unit for obtaining an error function from a difference from the previously obtained offset, an offset residual calculation unit for calculating an offset residual from the error function and the previously obtained covariance matrix, and the previous magnetic measurement An offset update unit that calculates the new offset by adding the offset residual to the offset calculated at the time of data measurement, and the measured magnetic field Using the measurement data, update the covariance matrix of the magnetic measurement data with the previously measured magnetic measurement data as a population, and add the measured magnetic measurement data to the population to obtain a new covariance matrix. And a covariance matrix updating unit to be generated.

The magnetic measurement data calibration apparatus of the present invention is based on the measurement axis of the magnetic sensor between the magnetic measurement data used when obtaining the previously calculated offset and the new magnetic measurement data newly input from the magnetic sensor. A first distance in the three-dimensional magnetic field space is obtained, and the first distance is compared with a preset threshold value. When the first distance exceeds the threshold value, a new value is added to the offset update unit. And a magnetic measurement data determination unit that performs control not to update a new offset when the first distance is less than or equal to the threshold value.

  In the magnetic measurement data calibration apparatus according to the present invention, the threshold is determined as a distance between a coordinate point of measurement noise in each measurement axis direction of the magnetic sensor in the three-dimensional magnetic field space and an origin of the three-dimensional magnetic field space. It is characterized by being.

  The magnetic measurement data calibration apparatus of the present invention includes a buffer for storing the magnetic measurement data used when the offset has been calculated up to the present time in time series, each of the magnetic measurement data stored in the buffer, and the previous time Obtain a second distance from the offset, obtain a first difference between each of the second distances and the total magnetic force obtained from the magnetic measurement data when obtaining the previous offset, and correspond to all the second distances. A first evaluation value is calculated by adding the first difference, a third distance between each of the magnetic measurement data stored in the buffer and the currently calculated offset is obtained, and each of the third distances and this time are calculated. The second difference from the total magnetic force obtained from the magnetic measurement data when the offset is obtained is obtained, the second difference corresponding to all the third distances is added to calculate a second evaluation value, and the second 1 evaluation value and previous When the second evaluation value is less than the first evaluation value when compared with a second evaluation value, the newly obtained offset is used as an offset used for calibration of the magnetic measurement data, while the second evaluation value is An offset validity determination unit that uses the previously obtained offset as an offset to be used for calibration of the magnetic measurement data when it is equal to or greater than the first evaluation value.

  In the magnetic measurement data calibration apparatus of the present invention, the error function calculation unit is configured to measure the data coordinates of the measured magnetic measurement data and the previous measurement of the magnetic measurement data in a three-dimensional coordinate system including the three measurement axes. The total magnetic force is obtained from the square of the distance of the offset calculated from time to time, the previous total magnetic force calculated last time is subtracted from the total magnetic force, and the subtraction result is used as the error function.

  The azimuth measuring device of the present invention has three measurement axes orthogonal to each other, and includes a magnetic sensor composed of an axis sensor for measuring geomagnetism in the measurement axis direction, and a geomagnetic field for each measurement axis output from the magnetic sensor. A magnetic measurement data calibration unit for obtaining an offset of magnetic measurement data as a measurement result and correcting the magnetic measurement data, and an azimuth angle from the magnetic calibration data calibrated from the magnetic measurement data output from the magnetic measurement data calibration device An error function for obtaining an error function from the difference between the magnetic data for each measurement axis of the measured magnetic measurement data and the previously obtained offset. A calculation unit, an offset residual calculation unit for calculating an offset residual from the error function and the previously obtained covariance matrix, and measurement of the previous magnetic measurement data. The offset update unit that calculates the new offset by adding the offset residual to the offset that is sometimes calculated, and the measured magnetic measurement data, and the previously measured magnetic measurement data as a population A covariance matrix updating unit configured to update a covariance matrix of magnetic measurement data and add the measured magnetic measurement data to the population to generate a new covariance matrix;

  According to the present invention, an offset residual with respect to the previously calculated offset is obtained from the previously calculated offset and newly obtained magnetic measurement data, and this offset residual is added to the previously calculated offset to obtain a new offset. Since there is no delay time for calculating the offset after obtaining a plurality of magnetic measurement data as before, it is possible to calculate the offset in real time, that is, at high speed, and follow the fluctuation of the environmental offset. The property can be improved as compared with the prior art.

It is a schematic block diagram which shows the structural example of the azimuth angle measuring apparatus using the magnetic measurement data calibration apparatus by the 1st Embodiment of this invention. It is a flowchart which shows the operation example of the process which calibrates the magnetic measurement data M which the magnetic sensor 2 detected in the magnetic measurement data calibration apparatus 1 of 1st Embodiment. Compared to the magnetic measurement data calibration apparatus 1 and the calibration apparatuses # 1 and # 2 according to the present embodiment, a magnetic field opposite to magnetism, a specific moving magnetic field (8-shaped movement and oblique movement), and an offset magnetic field It is a table which shows the offset estimation error measured by applying a synthetic magnetic field. It is a schematic block diagram which shows the structural example of the azimuth angle measuring apparatus using the magnetic measurement data calibration apparatus by the 2nd Embodiment of this invention. It is a flowchart which shows the operation example of the process which calibrates the magnetic measurement data M which the magnetic sensor 2 in the magnetic measurement data calibration apparatus 1 of 2nd Embodiment detected. It is a schematic block diagram which shows the structural example of the azimuth angle measuring apparatus using the magnetic measurement data calibration apparatus 1 by the 3rd Embodiment of this invention. It is a flowchart which shows the operation example of the process which calibrates the magnetic measurement data M which the magnetic sensor 2 in 3rd Embodiment detected. It is a figure explaining the shift | offset | difference of each axis | shaft of the absolute coordinate system calculated | required by a triaxial acceleration sensor, and the sensor coordinate system which the magnetic measuring device which is a triaxial magnetic sensor detects.

<First Embodiment>
Hereinafter, a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a schematic block diagram showing a configuration example of an azimuth measuring device using the magnetic measurement data calibration device according to the first embodiment of the present invention. Each azimuth measuring device is mounted on a portable device or the like.
The azimuth measuring apparatus according to this embodiment includes a magnetic measurement data calibration apparatus 1, a magnetic sensor 2, and an azimuth measuring unit 3.
The magnetic sensor 2, for example, X-axis direction of the X-axis direction magnetic detector unit 21 for measuring the magnetic data M _x indicating the strength of the magnetic field in the measurement axis, Y-axis magnetic data M _y indicating the strength of the magnetic field in the measurement axis X-axis, Y-axis, and Z-axis comprising a Y-axis direction magnetic detection unit 22 that measures the magnetic data M _z indicating the magnetic field strength on the measurement axis in the Z-axis direction, and a Z-axis direction magnetic detection unit 23 that measures the magnetic data M _z This is a magnetic sensor having three measurement axes. Here, as the X-axis direction magnetic detection unit 21, the Y-axis direction magnetic detection unit 22, and the Z-axis direction magnetic detection unit 23, for example, a Hall element, a magnetoresistive element, a flux gate type magnetic element, or the like is used.

Magnetic measurement data correcting device 1, calibration of the offset of the supplied from the magnetic sensor 2, the magnetic data _{M x,} magnetic measurements made of the magnetic data _{M y} and the magnetic data _{M z} data _M (M _x, M y, _{M z)} The magnetic calibration data M _f (M _fx , M _fy , M _fz ) is output.
The azimuth measuring unit 3 uses the magnetic calibration data M _f (M _fx , M _fy , M _fz ) calibrated by the magnetic measurement data calibration device 1 and outputs the azimuth that the azimuth measuring device is facing.

In the present embodiment, an offset to be corrected is an offset O (O _x , O _y , O _z ). Here, O _x is an offset in the X-axis direction, O _y is an offset in the Y-axis direction, and O _z is an offset in the Z-axis direction.
The magnetic measurement data M _n (M _xn , M _yn , M _zn ) is the nth magnetic data. The magnitude of the geomagnetic vector (hereinafter referred to as total magnetic force) is R.
Here, the total magnetic force is obtained by the following equation (1) using the magnetic measurement data M _n (M _xn , M _yn , M _zn ) and the offset O (O _x , O _y , O _z ).

This relationship of formula (1) always holds. That is, when the user carrying the portable device moves, the magnetic data in each measurement axis direction in the magnetic measurement data changes as the portable device rotates. However, as long as the total magnetic force detected from the magnetic measurement data does not change, when the magnetic measurement data is arranged as coordinate points in a three-dimensional space consisting of the X axis, the Y axis, and the Z axis, the magnetic measurement data consists of 3 measurement axes. It exists at any position on the same sphere in the dimensional space (this point is assumed to move on the same sphere as in the conventional example).
Therefore, if the total magnetic force is the same, the relationship of the above-described equation (1) always holds even if the magnetic data in each measurement axis direction in the magnetic measurement data changes.

In the above equation (1), by using the four pieces of magnetic measurement data M, the offset components O _x , O _y , O _z and the total magnetic force R in each measurement axis of the offset O can be calculated.
Here, in order to modify the equation (1), a constant Q shown in the following equation (2) is defined.

  By using the above equation (2) and changing the equation (1) into a matrix configuration, the following equation (3) is transformed.

By solving the above equation (3), the offset components O _x , O _y , O _z and the total magnetic force R can be obtained. At this time, it is necessary that an inverse matrix on the left side of the equation (3) exists.
However, magnetic noise from each part of the portable device or from the outside is superimposed on the actual magnetic measurement data M.
For this reason, the offset components O _x , O _y , O _z and the total magnetic force R calculated from the equation (3) may not be able to obtain accurate values.
Therefore, by using a statistical method, that is, RLS (Recursive Least Square) method, by reducing magnetic noise, an estimate of the offset components O _x , O _y , O _z , total magnetic force R ( Hereinafter, the maximum likelihood estimated value) is calculated. The description that the maximum likelihood estimated value can be calculated will be described later. Hereinafter, calibration of magnetic measurement data using the RLS method in the present embodiment will be described.

Next, the magnetic measurement data calibration apparatus 1 includes a measurement data input unit 11, a covariance matrix calculation unit 12, a covariance matrix update unit 13, a storage unit 14, an error function calculation unit 15, an offset residual calculation unit 16, an offset update. Unit 17 and magnetic measurement data processing unit 18.
The measurement data input unit 11 converts an analog axis sensor measurement value input from each of the X-axis direction magnetic detection unit 21, the Y-axis direction magnetic detection unit 22, and the Z-axis direction magnetic detection unit 23 into a digital value. Output. Here, each of the X-axis direction magnetic detection unit 21, the Y-axis direction magnetic detection unit 22, and the Z-axis direction magnetic detection unit 23 magnetizes a voltage value that is an analog value corresponding to the magnetic field detected in its own sensing direction. Output as data.
The measurement data input unit 11 also converts analog magnetic data M (M _x , M _y , M _z ) into digital magnetic measurement data M (M _x , M _y , M _z ). In order to remove the offset value included in the magnetic data of the value, the offset adjustment with the magnetic sensor alone may be performed.

  The covariance matrix calculation unit 12 sets the initial value of the covariance matrix P used in the offset residual calculation unit 16 described later as shown in the following equation (4).

In the equation (4), the covariance matrix calculation unit 12 sets the constant α in a range from 1.000 to 100.000, for example.
In addition, a column matrix indicating magnetic measurement data in a matrix of 4 rows × 4 columns on the left side in the equation (4) is defined as a vector z _{n in} the following equation (5).

Further, a matrix of 4 rows × 1 column on the left side in the equation (4) is defined as an offset vector x _n in the following equation (6).

The above equation (6) is a matrix indicating a vector composed of an offset O _n (O _xn , O _yn , O _zn ) and constant Q _n calculated from the total magnetic force when updated using the n-th magnetic measurement data. It is.

Error function calculation unit 15, using the following equation (7), calculates the error function e _n.

That is, the error function calculator 15 calculates the n-th measured magnetic measurement data M _n and the previously calculated offset On _-1 (the offset obtained using the (n-1) th measured magnetic data M _n-1 ). If, because the _(total magnetic force were determined using _n-1 th magnetic measurement data M _n-1) total force R _n-1 previously calculated, to calculate the error function e _n.
Here, the error function calculation unit 15, the distance between the square of the distance between the magnetic measurement data M _n and the offset O _n, calculates the square of the total magnetic R _n-1, a magnetic measurement data M _n and the offset O _n from square, subtracting the square of the total magnetic R _n-1, the subtraction result and the error function e _n. Here, the distance between the magnetic measurement data M _n and the offset O _n, X-axis is the measurement axis, in 3-dimensional space consisting of Y-axis and Z-axis, and the offset O _n-1 magnetometric data M _n This is the distance between the coordinate points of the magnetic measurement data M _n and the offset On _-1 when arranged.

Offset residual calculating section 16, and the error function e _n obtained error function calculation unit 15, using the covariance matrix P _n-1 obtained at the previous measurement, the following equation (8), offset residual η Is calculated.

  In this equation (8), the forgetting factor ρ is a constant used in the RLS method. Generally, it is set in the range of 0.95 <ρ <1.

The covariance matrix updating unit 13 includes the covariance matrix P _n when the magnetic measurement data M _n is included in the covariance matrix population, the matrix z _{n in the} equation (1), and the transposed matrix z in the equation (1). and _n ^T, the covariance matrix _{P n-1} to a population of _{M n-1} from the magnetic measurement data _{M 0,} using the forgetting factor [rho, calculated by shown below formula (9).

The offset updating unit 17 obtains an offset vector x _n corresponding to the magnetic measurement data M _n from the offset vector x _n−1 obtained from the immediately preceding magnetic measurement data M _n−1 and the offset residual η as follows ( 10) Calculated by the equation.

The magnetic measurement data processing unit 18 subtracts O _xn , O _yn , and O _zn in the offset vector X _n from the magnetic data M _xn , M _yn , and M _zn of the magnetic measurement data M _n , respectively, and magnetic calibration data M _f ( _Mfx , _Mfy , _Mfz ) is calculated.
The storage unit 14 stores a covariance matrix P, an offset vector x, and a total magnetic force R. Each of the covariance matrix P, the offset vector x, and the total magnetic force R is the data immediately before used in the next calculation when written, respectively, the covariance matrix P _n−1 , the offset vector _n−1, and the total magnetic force R Written as Rn _-1 .
In addition, the storage unit 14 includes initial values of the covariance matrix P, a forgetting factor ρ, equations (5), (6), (7), (8), (9), and (10). Pre-written and stored.

Next, a process for calibrating magnetic measurement data in the magnetic measurement data calibration apparatus 1 of the present embodiment will be described with reference to FIGS. FIG. 2 is a flowchart showing an operation example of processing for calibrating the magnetic measurement data M detected by the magnetic sensor 2.
Step S1:
The covariance matrix calculator 12 initializes the offset O and the total magnetic force R.
At this time, for example, as the initialization value of the offset O, the offset value measured by the triaxial Helmholtz coil of the single magnetic sensor 2 is written and stored in the storage unit 14 in advance. In addition, as a value for initializing the total magnetic force R, geomagnetic reference values in a plurality of regions are written and stored in the storage unit 14 in advance.
Then, after powering on the portable device, the user selects a region including his / her residence from a plurality of regions displayed on the display unit of the portable device.
Thereby, the covariance matrix calculation unit 12 reads the initial value of the offset O from the storage unit 14 and reads the geomagnetic reference value stored corresponding to the area selected by the user as the initial value of the total magnetic force R. .
In addition, when a GPS (Global Positioning System) function is installed in a portable device, the covariance matrix calculation unit 12 includes an area including latitude and longitude information obtained from GPS (an area defined by a range of latitude and longitude information). Alternatively, the geomagnetic reference value stored corresponding to may be read as the initial value of the total magnetic force R.
Then, the covariance matrix calculation unit 12 advances the process to step S2.

Step S2:
Next, the covariance matrix calculation unit 12 initializes the covariance matrix P. That is, the covariance matrix calculation unit 12 reads the equation (4) previously written and stored in the storage unit 14, and sets the read equation (4) as the initial value of the covariance matrix.
Then, the covariance matrix calculation unit 12 advances the process to step S3.

Step S3:
The measurement data input unit 11 includes magnetic data M _xn , M as magnetic measurement data M _n from each of the X-axis direction magnetic detection unit 21, the Y-axis direction magnetic detection unit 22, and the Z-axis direction magnetic detection unit 23 of the magnetic sensor 2. _{Read yn} and _Mzn .
Then, the measurement data input unit 11 writes and stores the read magnetic measurement data M _n in the storage unit 14.
After reading the magnetic measurement data M _n , the measurement data input unit 11 advances the process to step S4.

Step S4:
Then, the error function calculation unit 15, read from the storage unit 14 with respect to (7), magnetic data _M xn read from the magnetic sensor _2, and _{M yn} and _{M zn,} offset _{O n} read from the storage unit 14 _{-1 (O xn-1, O} yn-1, O zn-1) substituted and, to calculate the error function _{e n.}
That is, the error function calculation unit 15 squares the result of subtracting the offset O _xn-1 from the magnetic data M _xn , squares the result of subtracting the offset O _yn-1 from the magnetic data M _yn , and calculates the magnetic data M _zn from the magnetic data M _zn. The result of subtracting the offset O _zn−1 is squared.
The error function calculation unit 15 adds the results of the respective subtracted from the addition result, by subtracting the square of the total magnetic R _n-1, obtaining an error function e _n.
After calculating the error function e _n, error function calculation unit 15, the process proceeds to step S5.
Further, in the loop (repetitive processing) from step S3 to step S7 in the flowchart of FIG. 2, in the first loop after the portable device is turned on, the offset On _-1 is the numerical value of the offset in initialization. Similarly, the numerical value of the total magnetic force in initialization is used as the total magnetic force R _n−1 .

Step S5:
The offset residual calculation unit 16 reads the expression (5), the expression (8), the forgetting factor ρ, and the covariance matrix P _n−1 obtained last time from the storage unit 14.
Next, the offset residual calculation unit 16 generates a matrix of the vector z _n by substituting the magnetic data M _n into the equation (5), and generates a transposed matrix z _n ^T of the generated vector z _n matrix. .
Then, the offset residual calculation unit 16 substitutes each of the forgetting coefficient ρ, the covariance matrix P _n−1 , the vector z _n and the transposed matrix z _n ^{T into} the equation (8) to calculate the offset residual η, This offset residual η is output to the offset update unit 17.
After calculating the offset residual η, the offset residual calculation unit 16 advances the process to step S6.

Step S6:
The covariance matrix updating unit 13 reads the expression (5), the expression (9), the forgetting coefficient ρ, and the covariance matrix P _n−1 from the storage unit 14.
Next, the covariance matrix update unit 13 generates a matrix of the vector z _n by substituting the magnetic data M _n into the equation (5), and generates a transposed matrix z _n ^T of the generated matrix of the vector z _n. .
Then, the covariance matrix updating unit 13 substitutes the forgetting coefficient ρ, the covariance matrix P _n−1 , the vector z _n, and the transposed matrix z _n ^T into the equation (9), and creates a new covariance matrix. _Pn is generated and written into the storage unit 14 for storage. The covariance matrix P _n written by the covariance matrix update unit 13 in the storage unit 14 is used as the covariance matrix P _n−1 in the magnetic field loop.
Including the current magnetic data M _n, after all of the magnetic data M that were used to generate the covariance matrix P to generate the covariance matrix P _n to the population ever, the covariance matrix updater 13 Then, the process proceeds to step S7.

Step S7:
The offset update unit 17 reads the expressions (6) and (10) from the storage unit 14.
Next, the offset updating unit 17, (6) with respect to formula, and substituting the offset _{O n-1} calculated last time and the total force _{R n-1,} obtaining the vector _{x n-1} of the offset.
Then, the offset update unit 17 substitutes the offset residual η supplied from the offset residual calculation unit 16 and the obtained vector into the equation (10), and measures the magnetic measurement data M _n (that is, the current time). An offset vector x _n indicating the offset at is calculated. That is, the offset update unit 17 calculates a vector x _n indicating a new offset by adding the offset residual η to the vector x _n−1 .
After calculating the vector _xn , the offset updating unit 17 advances the processing to step S3.
As described above, the magnetic measurement data correcting device 1 according to this embodiment, in accordance with the flowchart of FIG. 2 performs the update processing of the offset O _n.

The magnetic measurement data processing unit 18, when the azimuth measuring unit 3 for calculating the azimuth angle, the offset O _n corresponding to the magnetic measurement data Mn measured data input unit 11 inputs, from the magnetic measurement data M _n Subtract.
The magnetic measurement data processing unit 18, the result of subtracting the offset _{O n} from magnetometric data _{M n,} calibrated magnetic calibration data _{M f (M fx, M fy} , M fz) as, the azimuth measuring part 3 Output.

According to the embodiment described above, the offset O _n-1 previously obtained, using a magnetic measurement data M _n measured at the present time, so that the error function e _n is minimized, and calculates an offset residual η the order to calculate the offset O _n and offset residual η using the offset O _n-1 previously obtained, it is possible to also minimize magnetic noise superimposed on the magnetic measurement data M _n, as compared to conventional higher precision Te, it is possible to obtain the offset O _n of the magnetic measurement data M _n being measured.

Further, according to this embodiment, since the current of the magnetic measurement data M _n seek offset O _n, when determining the offset residual η to be added to the offset O _n-1 previously obtained, to what extent reflects the error function Is set by the covariance matrix P _n−1 having the previous measured values as the population, the current magnetic measurement data M _n, and the forgetting factor ρ, the move because there is no need to measure a plurality of magnetometric data M, each time measuring the new magnetic measurement data M _n, to obtain the offset O _n-1 easily offset O _n the current from up to the previous time It becomes possible to improve the compliance with the fluctuation of the environmental offset.

Next, compared with the magnetic measurement data correcting device 1 of this embodiment, a first method of Patent Document 1, the error between the true value of the offset O _n estimated by each of the second technique of Patent Document 2 The results will be described.
A magnetic measurement apparatus equipped with each of the magnetic measurement data calibration apparatus 1 according to the present embodiment, the calibration apparatus # 1 using the algorithm of the first technique, and the calibration apparatus # 2 using the algorithm of the second technique. Prepared.
And each offset of a magnetic measuring device is measured with a triaxial Helmholtz coil, and it is set as the state which performed offset adjustment of the magnetic sensor provided with each calibration device.

Next, each of the magnetic measurement data calibration apparatus 1, the calibration apparatus # 1 using the algorithm of the first technique, and the calibration apparatus # 2 using the algorithm of the second technique are each mounted. An offset calculation process was performed.
First, the magnetic measuring device to be measured is set stationary on the stage of the 3-axis Helmholtz coil with the offset calibrated.
Then, a magnetic field opposite to the geomagnetism is generated so that the magnetic field becomes zero, and a specific moving magnetic field that is a magnetic field applied to the magnetic sensor when the magnetic measuring device is moved in a specific direction in this state is triaxial. Generate in Helmholtz coils.
That is, a magnetic field estimated to be detected by the magnetic sensor when the magnetic measurement device is moved is applied to the magnetic measurement device set stationary on the stage of the three-axis Helmholtz coil by the three-axis Helmholtz coil. Thereby, the state which moved the portable equipment carrying a magnetic measuring device is simulated.

In this embodiment, as a simulation of the movement of the portable device using the above-described three-axis Helmholtz coil, for example, a first simulation in which a movement for drawing a figure 8 is performed, and a second simulation in which the movement is performed in an oblique direction. And went. The oblique direction refers to a direction having an angle with respect to an axis perpendicular to the ground surface.
In this simulation, not only the magnetic field opposite to the geomagnetism and the specific moving magnetic field, but also a magnetic field having a specific direction of a certain direction in a certain direction as a virtual offset (as if an environmental offset exists) A combined magnetic field obtained by combining a magnetic field opposite to the geomagnetic field and a specific moving magnetic field.
And the synthetic magnetic field mentioned above was applied and the output of each magnetic measuring device was measured. At this time, when the magnetic field change of the synthetic magnetic field is detected and the calibration is normally performed, the calibration device in the magnetic measurement device determines the offset magnetic field as the offset inside the portable device. For this reason, the magnetic measurement device subtracts the offset magnetic field and the offset inside the portable device from the magnetic measurement data and outputs the result.

Therefore, after the calibration process described above is performed on the magnetic measurement device, the magnetic measurement data M (Mx, My, and M) output from the magnetic measurement device in a state where a combined magnetic field of the magnetic field opposite to the geomagnetism and the offset magnetic field is applied. Mz) is zero.
In the experiment described above, 10 μT (micro Tesla) was used as the offset magnetic field for each measurement axis in the X, Y, and Z axis directions.
In addition, the square of the magnetic data of each measurement axis output from the magnetic measurement device in a state where a magnetic field opposite to the geomagnetism and the offset magnetic field is applied is added, and the square root of the addition result is calculated as an offset estimation error. The closer this offset estimation error is to 0, the closer the offset calculated by the calibration apparatus is to the true value.
FIG. 3 shows the result of the experiment in this embodiment. FIG. 3 shows a magnetic field opposite to magnetism and a specific moving magnetic field (eight-shaped movement and diagonal movement) with respect to the magnetic measurement data calibration apparatus 1, the calibration apparatus # 1, and the calibration apparatus # 2 according to the present embodiment. ) And an offset magnetic field applied to the offset magnetic field, and measured offset estimation error.

As shown in FIG. 3, the magnetic measurement data calibration in this embodiment is performed in either the first state where the portable device is moved in the shape of figure 8 or the second state where the portable device is moved in an oblique direction. The result that the offset by the apparatus 1 is close to the true value, that is, the calculated offset O is close to the actual offset was obtained.
In the case of the second state that is moved in the oblique direction, there is no difference from the case of the first state in the present embodiment, but in the calibration devices # 1 and # 2, the calculated offset is compared with the first state. Calculation accuracy is low. In particular, the calibration apparatus # 2 is 17.32 = (10 ² +10 ² +10 ² ) ^1/2 , and thus it can be understood that the offset estimation itself is not performed.
From the result of FIG. 3 described above, according to the present embodiment, the movement of the portable device necessary for the calibration of the magnetic measurement data can be minimized, and the offset estimation accuracy is improved as compared with the conventional method. Can be made.

<Second Embodiment>
The second embodiment of the present invention will be described below with reference to the drawings. FIG. 4 is a schematic block diagram showing a configuration example of an azimuth measuring device using the magnetic measurement data calibration device 1 according to the second embodiment of the present invention. Each azimuth measuring device is mounted on a portable device or the like as in the first embodiment.
The azimuth measuring apparatus according to this embodiment includes a magnetic measurement data calibration apparatus 1, a magnetic sensor 2, and an azimuth measuring unit 3. In FIG. 4, the same components as those in the first embodiment are denoted by the same reference numerals. The difference from the first embodiment is that a magnetic measurement data determination unit 19 is newly provided for the magnetic measurement data calibration apparatus 1. Only the configuration and operation different from those of the first embodiment will be described below.

The magnetic measurement data determination unit 19 calculates the distance d between the magnetic measurement data M _{n -1 for} which the previous offset On _-1 was calculated and the magnetic measurement data M _n measured at the present time. As already described, the distance d is obtained by measuring the magnetic measurement data M _n and M in a three-dimensional space (magnetic field three-dimensional space) constituted by the measurement axes in the X-axis, Y-axis, and Z-axis directions of the magnetic sensor 2. It is the distance of each coordinate value when _n-1 is arranged.
The magnetic measurement data determining section 19, if the calculated distance d does not exceed the threshold value d _s set in advance, the magnetic measurement data correcting device 1 does not perform the calculation of the offset O _n by new magnetometric data M _n Take control. In this case, the data stored in the storage unit 14 is stored as it is without rewriting the numerical value calculated in the magnetic measurement data M _n−1 .

Here, the threshold value d _s is set as follows. That is, the threshold value d _s is determined by a numerical value corresponding to the measurement noise in the measurement of the magnetic sensor 2. When the measurement noise of each measurement axis in the X-axis, Y-axis, and Z-axis directions for measuring magnetic data is measurement noise v _x , v _y , z _v , the magnetic sensor 2 measures in time series even in a stationary state. The distance between the magnetic measurement data to be performed is not zero.
With the azimuth measuring device in a stationary state, the fluctuation width of the magnetic measurement data of each measurement axis in the X-axis, Y-axis, and Z-axis directions is measured in advance, and the range including this fluctuation width is measured as noises v _x , v _y , z _v And

Therefore, the threshold value d _s is obtained by the following equation (11). This threshold value d _s is written and stored in the storage unit 14 in advance. Here, the threshold value d _s is the length of the measurement noise vector (v _x , v _y , z _v ), that is, the origin of the three-dimensional space constituted by the measurement axes of the magnetic sensor 2, and the measurement noise v _x , v _y, is set at a distance between the coordinate values of z _v.
For example, when a magnetic sensor having a standard deviation of measurement noise of 1 μT on each measurement axis in the X-axis, Y-axis, and Z-axis directions is used, the fluctuation range obtained by equation (11) is 4.5 μT at the maximum. In this case, the threshold value ds is set as 5 μT.
The measurement data input unit 11, after the offset O _n is determined, and stores write magnetic measurement data M _n in the storage unit 14.

Next, processing for calibrating magnetic measurement data in the magnetic measurement data calibration apparatus 1 of the present embodiment will be described with reference to FIGS. FIG. 5 is a flowchart showing an operation example of processing for calibrating the magnetic measurement data M detected by the magnetic sensor 2 in the present embodiment.
Steps S1 to S7 in FIG. 5 are the same as steps S1 to S7 in FIG. Only differences from the flowchart in FIG. 2 of the first embodiment will be described below.
In step S3, after reading the magnetic measurement data M _n , the measurement data input unit 11 advances the processing to step S13.

Step S13:
Magnetic measurement data determining section 19, the magnetic measurement data M _n is supplied from the measurement data input unit 11, a threshold d _s, and a magnetic measurement data M _n-1 used to calculate the previous offset O _n-1 Read from the storage unit 14.
Next, the magnetic measurement data determination unit 19 calculates a distance d between the supplied magnetic measurement data M _n and the magnetic measurement data M _n−1 read from the storage unit 14.
Then, the magnetic measurement data determination unit 19 compares the calculated distance d with the threshold value d _s read from the storage unit 14.
At this time, when the distance d is greater than or equal to the threshold value d _s , the magnetic measurement data determination unit 19 proceeds to step S4. On the other hand, when the distance d is less than the threshold value d _s , the process proceeds to step S14.

Step S14:
Magnetic measurement data determination unit 19 does not perform the processing for calculating the offset O _n on the magnetic measurement data correcting device 1, the values stored in the storage unit 14 as it is, the process proceeds to step S3.

  With the configuration described above, in the present embodiment, the same or similar magnetic measurement data, that is, magnetic measurement data with a short distance, is repeatedly estimated to calculate the offset by repeating the calculation to calculate the offset, that is, all the magnetic measurement data is It is possible to eliminate the disadvantage of the first embodiment that determines that the offset is present.

<Third Embodiment>
The third embodiment of the present invention will be described below with reference to the drawings. FIG. 6 is a schematic block diagram showing a configuration example of an azimuth measuring device using the magnetic measurement data calibration device 1 according to the third embodiment of the present invention. Each azimuth measuring device is mounted on a portable device or the like as in the second embodiment.
The azimuth measuring apparatus according to this embodiment includes a magnetic measurement data calibration apparatus 1, a magnetic sensor 2, and an azimuth measuring unit 3. In FIG. 6, the same components as those in the second embodiment are denoted by the same reference numerals. The difference from the second embodiment is that an offset validity determination unit 10 is newly provided for the magnetic measurement data calibration apparatus 1. Only the configuration and operation different from those of the second embodiment will be described below.

Offset validity judging unit 10, a determination is offset O _n determined from magnetic measurement data M _n read from the magnetic sensor 2 is whether valid or, if it is determined that the offset O _n is not valid, the offset previously obtained the O _n-1 is used as a new offset _{O n.}
The storage unit 14 is provided with a data buffer, and 16 to 32 sets of magnetic measurement data M are written and stored in the data buffer. For example, when 16 sets of magnetic measurement data are stored, the magnetic measurement data from the magnetic measurement data M ₁ to M ₁₆ are stored in the data buffer.

Here, the offset validity determination unit 10 causes the magnetic measurement data determination unit 19 to write and store the magnetic measurement data M _{n in} which the distance d exceeds the preset threshold d _s in the above-described data buffer. At this time, the offset validity determination unit 10 overwrites the newest magnetic measurement data M _n on the oldest magnetic measurement data in the data buffer. That is, the offset validity determination unit 10 writes and stores only the magnetic measurement data determined to be updated by the magnetic measurement data determination unit 19 by calculating the offset. The distance in this embodiment is also a distance in a three-dimensional space constituted by the measurement axes in the X axis, Y axis, and Z axis directions of the magnetic sensor 2 shown in the second embodiment.

Further, the offset validity judging unit 10 from the storage unit 14 and the offset _{O n-1,} reads out the total magnetic _{R n-1,} vector _O 1 of the offset _{(O 1x, O 1y, O} 1z) and, Let R ₁ be the total magnetic force.
The offset validity determination unit 10 calculates the distance d _1n corresponding to each of the magnetic measurement data M ₁ to M _n stored in the data buffer by the following equation (12).

That is, for each magnetic measurement data M, the offset validity determination unit 10 squares the result of subtracting the offset O _1x from the magnetic data M _x and squares the result of subtracting the offset O _1y from the magnetic data _My. The result of subtracting the offset O _1z from M _z is squared.
And the offset effectiveness determination part 10 calculates _| requires the distance _d1n by adding the squared result, and calculating the square root of an addition result. Here, the offset validity judging section 10 corresponds to each of Mn from the magnetic measurement data M1, obtains each _{d 1n} from the distance _{d 11.}
Moreover, the offset validity determination part 10 calculates evaluation value G1 by the following (13) Formula.

That is, the offset validity determination unit 10 subtracts the total magnetic force R ₁ from each of the distances d ₁₁ to d _1n and squares each subtraction result to obtain the sum of the squares of the subtraction results. The results are added, and the sum of the squares as the addition result is set as the evaluation value G1.
Further, the offset validity determination unit 10 sets the offset vector of the offset On obtained from the magnetic measurement data Mn read from the magnetic sensor 2 as O ₂ (O _2x , O _2y , O _2z ), and this vector O ₂ (O _2x , O _2y , O _2z ) and the magnetic measurement data Mn, the total magnetic force R2 is newly obtained.
Then, the offset validity determination unit 10 calculates the distance d _2n corresponding to each of the magnetic measurement data M ₁ to M _n stored in the data buffer by the following equation (14).

That is, the offset validity judging unit 10, for each magnetic measuring data M, squares the result of subtracting the offset O _2x from the magnetic data M _x, squares the result of the offset O _2y subtracted from the magnetic data M _y, magnetic The result of subtracting the offset O _2z from the data M _z is squared.
Then, the offset validity determination unit 10 adds the squared results and calculates the square root of the addition result to obtain the distance d _2n . Here, the offset validity judging section 10 corresponds to each of Mn from the magnetic measurement data M1, obtains each _{d 2n} from the distance _{d 21.}
Further, the offset validity determination unit 10 calculates the evaluation value G2 by the following equation (15).

That is, the offset validity determination unit 10 subtracts the total magnetic force R ₂ from each of the distances d ₂₁ to d _2n and squares each subtraction result to obtain the sum of the squares of the subtraction results. The results are added, and the sum of squares as the addition result is set as an evaluation value G2.
When the evaluation values G1 and G2 are calculated, the offset validity determination unit 10 compares the evaluation value G1 and the evaluation value G2.
At this time, when the offset O ₁ and the total magnetic force R ₁ are true values, the evaluation value G 1 is a true value, and when the offset O ₂ and the total magnetic force R ₂ are true values, the evaluation value G2 is a true value. That is, it is considered that the evaluation value G1 becomes smaller as the offset O ₁ and the total magnetic force R ₁ become closer to the true value, and the evaluation value G2 becomes smaller as the offset O ₂ and the total magnetic force R ₂ become closer to the true value.

Therefore, the offset validity judging unit 10 compares the evaluation values G1 and G2, evaluation value G2 is smaller than the evaluation value G1 (G2 <G1) when the offset _{O n} and total magnetic _{R n} newly calculated It is determined that the value is closer to the true value than the previously calculated offset On _-1 and total magnetic force Rn _-1 .
On the other hand, when the evaluation value G2 is equal to or greater than the evaluation value G1 (G2 ≧ G1), the offset effectiveness determination unit 10 calculates the offset O _n−1 calculated previously and the offset O newly calculated from the total magnetic force R _n−1. determining from the _n and the total force R _n or close to the true value, or similar to.

Here, the offset validity judging unit 10, when the G2 <G1, the offset O _n and the total force R _n are newly calculated, and outputs to the magnetic measurement data processing unit 18, indicating that updating the offset The control signal is output to the offset update unit 17.
When the offset update unit 17 is supplied with control information indicating that the previously calculated offset On _-1 is updated, the offset update unit 17 newly calculates the offset On _-1 stored in the storage unit 14. to override the O _n. Thus, the latest offset O _n is held in the storage unit 14.
Further, when G2 ≧ G1, the offset validity determination unit 10 outputs the offset value and the total magnetic force R stored in the storage unit 14 to the magnetic measurement data processing unit 18 and does not update the offset. Is output to the offset update unit 17.
Offset updating unit 17, when the control information indicating that no update the offset is supplied, without overwriting the offset O _n-1 calculated last time is stored in the storage unit 14, an offset O _n-1 previously calculated Do not update the value of. Thus, the storage unit 14, as the latest offset, the offset O _n of the same value is maintained offset O _n-1 previously calculated.

Next, a process for calibrating magnetic measurement data in the magnetic measurement data calibration apparatus 1 of the present embodiment will be described with reference to FIGS. FIG. 7 is a flowchart showing an operation example of processing for calibrating the magnetic measurement data M detected by the magnetic sensor 2 in the present embodiment.
Steps S1 to S7 and steps S13 and S13 in FIG. 7 are the same as steps S1 to S7 and steps S13 and S14 in FIG. Only the differences from the flowchart in FIG. 5 of the second embodiment will be described below.
In step S7, the offset updating unit 17 calculates the vector _xn, and then advances the processing to step S15.

Step S15:
Next, the offset validity determination unit 10 reads the magnetic measurement data M ₁ to M _n stored in the storage unit 14 and the offset O _n−1 (O ₁ ) calculated and stored last time, and stores the magnetic data. For each M, the distance d _1n is calculated by the above-described equation (12).
Then, the offset validity judging unit 10 reads out the magnetic measurement data _{M n-1} in the memory 14 is stored with offset _{O n-1,} it calculates the total magnetic _{R 1} by (1).
Thereby, the offset validity determination unit 10 calculates the evaluation value G1 from the calculated distance d _1n for each magnetic data and the total magnetic force R _{1 according} to the equation (13), and advances the processing to step S16.

Step S16:
Next, the offset validity judging unit 10, a M _n from magnetometric data M ₁ stored in the storage unit 14, reads the offset O _n stored is calculated this time _{(O 2),} each magnetic data M Then, the distance d _2n is calculated by the above-described equation (14).
Then, the offset validity judging unit 10 reads the magnetic measurement data M _n stored in the storage unit 14 and the offset O _n, calculates the total magnetic R ₂ by (1).
Thereby, the offset validity determination unit 10 calculates the evaluation value G2 from the calculated distance d _2n for each magnetic data and the total magnetic force R ₂ by the equation (15), and advances the processing to step S17.

Step S17:
Next, the offset validity determination unit 10 compares the calculated evaluation values G1 and G2, and determines whether or not the evaluation value G2 is less than the evaluation value G1.
At this time, if the evaluation value G2 is less than the evaluation value G1, the offset validity determination unit 10 advances the process to step S18. On the other hand, if the evaluation value G2 is equal to or higher than the evaluation value G1, the process advances to step S19. .

Step S18:
If the evaluation value G2 is smaller than the evaluation value G1, since the offset O _n the currently obtained is close to the true value than the offset O _n-1, which is stored is written as an offset to the previous memory section 14, an offset validity judging section 10 determines that you do not want to update the offset O _n found this time.
Then, the offset validity determination unit 10 outputs a control signal indicating that the offset value stored in the storage unit 14 is updated to the offset update unit 17.
When the control information indicating that performing this update is supplied, the offset updating unit 17, with respect to the offset stored in the storage unit 14 overwrites the offset O _n found this time to update the value of the offset .
And the offset update part 17 advances a process to step S3.

Step S19:
If the evaluation value G2 is rated value G1 or more, with respect to the offset O _n found this time, since towards the offset O _n-1 previously stored in the storage unit 14 is written as an offset is close to the true value, the offset The validity determination unit 10 outputs to the offset update unit 17 a control signal indicating that updating is performed while maintaining the previously written offset value.
When the control information indicating that this update is not performed is supplied, the offset updating unit 17 overwrites the offset O stored in the storage unit 14 while maintaining the previous offset On _-1 value. Update the offset value.
And the offset update part 17 advances a process to step S3.

  As described above, according to the present embodiment, every time the magnetic measurement data M determined to be usable for calculating the offset O is supplied, the offset On is calculated, and the storage unit 14 is determined to be valid last time. Are compared with the offset value On-1 stored in the evaluation value G1 and G2, which are closer to the true value, select an offset closer to the true value, and hold the selected offset in the storage unit 14. The offset O, which is always estimated to be closest to the true value, can be used for the calibration of the magnetic data M, and the magnetic data can be calibrated with high accuracy.

<Explanation that offset is obtained as maximum likelihood estimate by RLS method>
In the following, when the RLS method is used, the offset O is updated using the latest magnetic measurement data M using the covariance matrix, that is, the offset vector x is always updated, so that the maximum likelihood value of the offset O (maximum value) is always obtained. The likelihood that the likelihood estimate) can be obtained will be described.
First, the following expression (16) is defined from the magnetic measurement data M _n (M _xn , M _yn , M _zn ).

According to a general least square method, using the magnetic measurement data M ₁ to M _n measured up to the present time, the offset vector x that minimizes the evaluation function J ₁ of the following equation (17) is calculated. become.

Next, as shown in the equation (18), a column matrix Z _n whose elements are the transposed matrix z _n ^T of the vector z _n and a column matrix Y _n whose elements are the vectors y _n of the equation (16) are defined. To do.

By using the column matrices Z _n and Y _n of the above equation (18), the equation (17) can be transformed into the following equation (19).

As can be seen from the equation (18), the above equation (19) includes elements of all the magnetic measurement data measured from the magnetic measurement data M ₁ to M _n , that is, the present.

  Next, a matrix having the weighting coefficient Wn as an element, that is, a weighting matrix Wn is defined as the following equation (20).

Then, using this weighting matrix W _n and the column matrices Y _n and Z _n , the evaluation function J ₂ is shown as the following equation (21).

As shown in the equation (20), the weighting matrix Wn is introduced in order to increase the weight when using each magnetic measurement data M for calculation, that is, the degree of influence of the specific magnetic measurement data M. If this weighting matrix W _n is a unit matrix, the evaluation function J ₂ is a least square method for obtaining the same evaluation value (J ₂ = J ₁ ) as the evaluation function J ₁ according to the equation (17) already shown.

  Here, the offset vector x that minimizes the evaluation function J2 in the equation (21) is expressed by the following equation (22).

When the above equation (22) is substituted into the equation (21), the evaluation function J _{2 of the} equation (21) is minimum, that is, J ₂ = 0, so that the equation (22) indicates the vector x of the maximum likelihood estimated value. I know that.

Further, in order to represent the covariance matrix P _n used in the calculation, the matrix _Un is defined by the following equation (23).

  By substituting the equations (18) and (20) for the inverse matrix of the matrix Un of the equation (23), the following equation (24) is obtained.

  Here, in order to solve Equation (24) analytically, the inverse matrix theorem is used. That is, in the inverse matrix lemma, when the matrix A is an n × n matrix, the matrix b is an n × 1 vector, and the matrix c is an n × 1 vector, the relationship expressed by the following equation (25) is established. .

  Therefore, when the equation (24) is substituted into the equation (25), the following equation (26) is obtained.

  Thereby, (24) Formula can be expressed as a recurrence formula of the following (27) Formula by calculating | requiring the inverse matrix of (26) Formula.

  Here, the weighting matrix in the equation (20) is set as the following equation (28).

In the above (28), when the forgetting factor ρ is set in the range of 0 to 1 (0 <ρ <1), the degree of influence (reflection) of the magnetic measurement data acquired in the past gradually decreases.
Further, the covariance matrix _{P n,} expressed in weighting coefficient _{w n} and the matrix _{U n,} and becomes less (29), thereby coinciding with the (27) equation (9) below.

Further, for the convenience of subsequent calculations, defined by the following equation (30) to k _n.

  Next, by substituting the equations (27) and (30) into the equation (22), the following equation (31) is obtained.

Further, the estimated value of the offset vector x _n−1 obtained from the magnetic measurement data M ₁ to M _n−1 , that is, the estimated value of the n−1th vector x _n−1 is the following (32). It is calculated by the formula.

  Therefore, the estimated value of the nth vector xn is expressed as the following equation (33) by substituting the above equation (32) into the equation (31) and modifying the equation (31).

  And the following (34) Formula is obtained by transforming this (33) Formula.

  By using the above equation (34), the third term on the right side of the equation (33) can be simplified to kn × yn, so the recurrence equation of the vector x in the equation (35) shown below from the equation (33) Can lead.

  The above expression (35) is an expression obtained by integrating the expressions (8) and (10). Therefore, by updating the offset vector x using the latest magnetic measurement data M by the RLS method described above, the maximum likelihood estimated value of the vector x can always be obtained.

<Direction detection device>
Next, the operation of the azimuth measuring apparatus using the magnetic measurement data calibration apparatus 1 according to the first to third embodiments will be described below.
The azimuth angle measurement device mounted on the portable device includes a configuration in which the magnetic sensor 2 including the three-axis magnetic detection unit illustrated in FIG. 1 and a three-axis acceleration sensor (not shown) are combined.
That is, this azimuth measuring device detects gravity by a triaxial acceleration sensor, and calculates its own inclination angle from the degree of inclination of the detected gravity.
Then, in the magnetic measurement device, the magnetic measurement data measured by the magnetic sensor 2 generates magnetic calibration data M _f with the offset calibrated by the magnetic measurement data calibration device 1.
The azimuth angle measuring device calculates the azimuth angle based on the magnetic calibration data M _f output from the magnetic measurement data calibration device 1.

Hereinafter, the calculation of the azimuth angle by the magnetic calibration data _Mf performed by the azimuth angle measurement unit 3 will be described.
FIG. 8 is a diagram for explaining a shift of each axis between an absolute coordinate system obtained by a triaxial acceleration sensor and a sensor coordinate system detected by a magnetic measuring device which is a triaxial magnetic sensor.
In FIG. 8, an X axis, a Y axis, and a Z axis for calibrating a three-dimensional absolute coordinate system are defined. The rotation angle with the X axis as the rotation axis is the pitch angle (p), the rotation angle with the Y axis as the rotation axis is the roll angle (r), and the rotation angle with the Z axis as the rotation axis is the azimuth angle (θ ).
In the present embodiment, the absolute coordinate system is such that the plane formed by the X axis and the Y axis is perpendicular to the gravity vector, and the magnetic measurement data M (Mx, My, Mz) measured by the magnetic sensor 2 is used. ) Is a coordinate system with the origin O as the point where all of the points are zero. The sensor coordinate system is a coordinate system formed by magnetic data M (M _x , M _y , M _z ) measured by the magnetic sensor 2. For this reason, the sensor coordinate system is different for each measurement point where the magnetic sensor 2 measures the magnetic data M (M _x , M _y , M _z ).

However, the mobile device user portable, and form planes of X-axis and Y-axis of the absolute coordinate system, the measurement axis of the magnetic sensor, that is, the X axis in the sensor coordinate system (measurement axis of the M _x) and Y-axis (M _y It is difficult to make the plane formed by the measurement axis) parallel to each other.
Therefore, when the azimuth angle is estimated by the magnetic sensor 2, the measurement axis (X-axis detection magnetic field) in the X direction in the sensor coordinate system of the magnetic sensor 2 is always relative to the plane formed by the X axis and the Y axis in the absolute coordinate system. If the plane formed by the measurement axis in the Y direction (T-axis detection magnetic field) is parallel, the triaxial acceleration sensor used as the tilt sensor is always measured even if there is an influence of changes in the magnetic field due to the surrounding environment. The inclination angle of the magnetic sensor 2, that is, the azimuth measuring device can be accurately measured by the gravity vector.

At this time, when the azimuth measuring device is in a horizontal state with respect to the ground surface, the acceleration measurement data output from the triaxial acceleration sensor is _{defined as} acceleration measurement data S (S _x , S _y , S _z ). On the other hand, when the azimuth measuring device is tilted with respect to the ground surface, the acceleration measurement data output from the triaxial acceleration sensor is defined as acceleration measurement data S ′ (S _x ′, S _y ′, S _z ′).
At this time, the relationship between the acceleration measurement data S (S _x , S _y , S _z ) and the acceleration measurement data S ′ (S _x ′, S _y ′, S _z ′) is expressed by the following equation (36). .

When the portable device carried by the user is in an inclined state, the output of the triaxial acceleration sensor normalized by the gravitational acceleration in the inclined state is assumed to be acceleration measurement data A (A _x , A _y , A _z ). When the azimuth measuring device is in a horizontal state with respect to the ground surface, since A (A _x , A _y , A _z ) = A (0, 0, 1), the above expression (36) is shown below ( 37).

Therefore, the azimuth measuring unit 3 uses the acceleration measurement data A (A _x , A _y , A _z ) output from the triaxial acceleration sensor, and the pitch angle (p) is expressed by the following equation (38): The roll angle (r) is calculated by the following equation (39).

Further, when the mobile device carried by the user is in an inclined state, the magnetic calibration data M _f output from the magnetic measurement data calibration device 1 is the magnetic measurement data M (M _tx , M _ty , M _tz ). Based on the pitch angle (p) and the roll angle (r) calculated from each of the equations (38) and (39), the magnetic measurement data M (H _x , H _y in the horizontal state is obtained by the following equation (40). , V).

Therefore, the azimuth measuring unit 3 calculates the azimuth angle (θ) from the following equation (41) using the magnetic measurement data M (H _x , H _y , V) calculated by the above equation (40).

  As described above, according to the present embodiment, the environmental offset is adjusted, and the magnetic measurement data M output from the magnetic sensor 2 is calibrated (that is, the magnetic measurement data in the zero magnetic field is adjusted). , And the azimuth angle can be obtained by correcting the tilt state with a three-axis magnetic sensor, and the plane formed by the X-axis direction and the Y-axis direction of the detection coordinate system of the azimuth angle measuring device is always the absolute coordinate system. Since the magnetic measurement data in which the ground surface is horizontal and the offset is calibrated can be used, it is possible to obtain the azimuth angle with high accuracy.

  Also, a program for realizing the functions of the magnetic measurement data calibration apparatus 1 in FIGS. 1, 4 and 6 is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into a computer system. The magnetic measurement data M may be calibrated by executing it. Here, the “computer system” includes an OS and hardware such as peripheral devices.

Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.
The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory in a computer system serving as a server or a client in that case, and a program that holds a program for a certain period of time are also included. The program may be a program for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes design and the like within a scope not departing from the gist of the present invention.

  DESCRIPTION OF SYMBOLS 1 ... Magnetic measurement data calibration apparatus 2 ... Magnetic sensor 3 ... Azimuth angle measurement part 10 ... Offset effectiveness determination part 11 ... Measurement data input part 12 ... Covariance matrix calculation part 13 ... Covariance matrix update part 14 ... Memory | storage part 15 ... Error function calculation unit 16 ... Offset residual calculation unit 17 ... Offset update unit 18 ... Magnetic measurement data processing unit 19 ... Magnetic measurement data determination unit 21 ... X-axis direction magnetic detection unit 22 ... Y-axis direction magnetic detection unit 23 ... Z-axis Direction magnetic sensor

Claims (6)

An offset of magnetic measurement data, which is a measurement result of the geomagnetism for each of the measurement axes, is output from a magnetic sensor having three measurement axes orthogonal to each other and measuring the geomagnetism in the measurement axis direction. It is a magnetic measurement data calibration device that corrects measurement data,
An error function calculation unit for obtaining an error function from the difference between the magnetic data for each measurement axis of the measured magnetic measurement data and the previously obtained offset;
An offset residual calculation unit for calculating an offset residual from the error function and the previously obtained covariance matrix;
An offset update unit that adds the offset residual to the offset calculated during the previous measurement of the magnetic measurement data and calculates a new offset;
Using the measured magnetic measurement data, update the covariance matrix of the magnetic measurement data with the previously measured magnetic measurement data as a population, and add the measured magnetic measurement data to the population to newly And a covariance matrix updating unit for generating a covariance matrix. The first distance in the three-dimensional magnetic field space formed by the measurement axis of the magnetic sensor between the magnetic measurement data used when obtaining the previously calculated offset and the new magnetic measurement data newly input from the magnetic sensor. Determining, comparing the first distance with a preset threshold, and if the first distance exceeds the threshold, causing the offset update unit to update a new offset, The magnetic measurement data calibration device according to claim 1, further comprising a magnetic measurement data determination unit that performs control so as not to update a new offset when the first distance is equal to or less than the threshold value. The threshold value is defined as a distance between a coordinate point of measurement noise in each measurement axis direction of the magnetic sensor in the three-dimensional magnetic field space and an origin of the three-dimensional magnetic field space. The magnetic measurement data calibration device described. A buffer for storing magnetic measurement data used in calculating the offset up to the present time in a time series;
A second distance between each of the magnetic measurement data stored in the buffer and the previous offset is obtained, and a total magnetic force obtained from the magnetic measurement data when each of the second distance and the previous offset is obtained; Each of the magnetic measurement data stored in the buffer and the offset calculated this time are calculated by adding the first differences corresponding to all the second distances and calculating the first evaluation value. The third distance is obtained, and the second difference between each of the third distances and the total magnetic force obtained from the magnetic measurement data when the current offset is obtained is obtained, and the third distances corresponding to all the third distances are obtained. A second evaluation value is calculated by adding two differences, the first evaluation value and the second evaluation value are compared, and if the second evaluation value is less than the first evaluation value, a new value is obtained. Use offset to calibrate the magnetic measurement data An offset validity determination unit that uses an offset used for calibration of the magnetic measurement data as an offset when the second evaluation value is equal to or greater than the first evaluation value. The magnetic measurement data calibration device according to claim 2 or 3. The error function calculator is
In a three-dimensional coordinate system composed of the three measurement axes, the total magnetic force is calculated from the square of the distance between the measured data coordinate of the magnetic measurement data and the offset coordinate of the offset calculated at the previous measurement of the magnetic measurement data. The magnetic measurement data calibration apparatus according to any one of claims 1 to 4, wherein the previous total magnetic force calculated last time is subtracted from the total magnetic force, and the subtraction result is used as the error function. . A magnetic sensor having three measurement axes orthogonal to each other and including an axis sensor for measuring geomagnetism in the measurement axis direction;
Obtaining a magnetic measurement data offset that is a measurement result of the geomagnetism for each measurement axis output from the magnetic sensor, and correcting the magnetic measurement data;
An azimuth measuring unit that calculates an azimuth from magnetic calibration data calibrated from the magnetic measurement data output from the magnetic measurement data calibration device, and
The magnetic measurement data calibration unit is
An error function calculation unit for obtaining an error function from the difference between the magnetic data for each measurement axis of the measured magnetic measurement data and the previously obtained offset;
An offset residual calculation unit for calculating an offset residual from the error function and the previously obtained covariance matrix;
An offset update unit that adds the offset residual to the offset calculated during the previous measurement of the magnetic measurement data and calculates a new offset;
Using the measured magnetic measurement data, update the covariance matrix of the magnetic measurement data with the previously measured magnetic measurement data as a population, and add the measured magnetic measurement data to the population to newly An azimuth angle measuring apparatus comprising: a covariance matrix updating unit that generates a covariance matrix.

Images