阅读说明：本技术 一种三轴磁传感器的转向差校正方法 (Steering difference correction method of three-axis magnetic sensor ) 是由 章雪挺 李崇贝 何欢 黄浩楠 廖章锦 徐航 于 2019-09-02 设计创作，主要内容包括：本发明公开了三轴磁传感器的转向差校正方法,包括以下步骤,S10,选择一片空旷且没周围没有磁性物质的区域,在一个固定点上,转动磁传感器,并采集磁传感器所得到三轴磁场数值Hx,Hy,Hz；S20,设X’Y’Z’轴为磁传感器的实际坐标轴,XYZ轴为标准坐标轴,构建地磁场矢量的测量输出模型；S30,构造神经网络辨识系统,采用迭代原理更新误差参数,从而得到误差参数的各个估计值；S40,利用误差参数,计算补偿后的磁传感器数据。本发明适用性广,且不要求巨大的数据量,对参数估计的精度更高,减少转向差的效果较好,有效提高了测量磁场值时的准确度。(the invention discloses a steering difference correction method of a triaxial magnetic sensor, which comprises the following steps of S10, selecting a region which is open and has no magnetic substances around, rotating the magnetic sensor on a fixed point, and collecting triaxial magnetic field values Hx, Hy and Hz obtained by the magnetic sensor; s20, setting an X ' Y ' Z ' axis as an actual coordinate axis of the magnetic sensor, and setting an XYZ axis as a standard coordinate axis, and constructing a measurement output model of the geomagnetic field vector; s30, constructing a neural network identification system, and updating error parameters by adopting an iteration principle so as to obtain each estimation value of the error parameters; s40, the compensated magnetic sensor data is calculated using the error parameter. The method has wide applicability, does not require huge data volume, has higher precision of parameter estimation, has better effect of reducing steering difference, and effectively improves the accuracy when measuring the magnetic field value.)

1. a steering difference correction method of a three-axis magnetic sensor is characterized by comprising the following steps:

s10, selecting a region which is open and has no magnetic substances around, rotating the magnetic sensor on a fixed point, and collecting triaxial magnetic field values Hx, Hy and Hz obtained by the magnetic sensor;

s20, setting an X ' Y ' Z ' axis as an actual coordinate axis of the magnetic sensor, and setting an XYZ axis as a standard coordinate axis, and constructing a measurement output model of the geomagnetic field vector;

S30, constructing a neural network identification system, and updating error parameters by adopting an iteration principle so as to obtain each estimation value of the error parameters;

S40, the compensated magnetic sensor data is calculated using the error parameter.

2. the method according to claim 1, wherein the X ' Y ' Z ' axis is an actual coordinate axis of the magnetic sensor, the XYZ axis is a standard coordinate axis, and a measurement output model of the geomagnetic field vector is constructed, expressed by the following formula,

Wherein the content of the first and second substances,to an asymmetric proportionality coefficient, K_xand K_ysensitivity scales in the direction of an actual coordinate axis OX ' of the magnetic sensor and sensitivity scales in the direction of an OY ' axis are respectively set to be 1 by taking the sensitivity scales in the direction of the OZ ' axis as a standard, namely the sensitivity scales are the same as the standard coordinate axis OZ of the magnetic sensor;

The space transformation matrix is adopted, beta represents an included angle between a Y ' axis and a Y axis, alpha represents an included angle between an X ' axis and an XOZ plane, and gamma represents an included angle between the X ' axis and the XOY plane; the cos alpha is approximately equal to 1, the cos beta is approximately equal to 1, the sin alpha is approximately equal to alpha, the sin beta is approximately equal to beta, sin gamma is approximately equal to gamma to obtain

B＝[B_x,B_y,B_z]^TIs a zero offset error matrix, B_x，B_y，B_znamely zero offset values of the actual X axis, Y axis and Z axis of the magnetic sensor;

for the actual output component values of the magnetic sensors,the magnetic field values on the X axis, the Y axis and the Z axis which are actually obtained by the magnetic sensor are obtained;

setting an error matrixObtained by the formula (1)

and the further calculation is carried out to obtain,Wherein E is an identity matrix;

ΔK_x＝K_x-1，ΔK_y＝K_y-1；

Conversion of formula (2) to

Neglecting the second order small quantity deltas deltak, we get,

By using the two-norm vector, the method can be used,

Neglecting second order fractional quantitiesso as to obtain the compound with the characteristics of,

Obtained from the formula (3) and the formula (4),

will be provided withΔ K, Δ S, B is taken into formula (5),

。

3. The method according to claim 2, wherein in S30, [ Δ K [ [ Δ K ]_x,ΔK_y,γ,α,β,B_x,B_y,B_z]The method for estimating the error parameters comprises the following steps:

S31, where t represents the number of iterations, and let t equal to 1, initialize a row matrix w with eight elements all being 1, and indicate eight error parameters to be estimated;

s32, setting iteration number M and learning efficiency μ, where M is 5000 and μ is 0.000015;

S33, according to the formula (6), the output of the magnetic sensor is adjustedthe development is that,

s34, calculating the total weight value z (t) ═ w × P^T；

S35, calculating an error vector

S36, according to WIDROW-HOFF learning rule, the parameter adjustment expression is w (t +1) ═ w (t) + μ e (t) P^T，

W is a matrix of parameters to be estimated, mu is learning efficiency, the speed of each parameter approaching a stable value can be adjusted, e (t) is an error vector, and P is an input matrix;

S37, t ═ t +1, return to S34;

And S38, when t is equal to the iteration times M, ending the iteration to obtain an error parameter estimation value.

4. the method according to claim 3, wherein in the step S40, the compensated magnetic sensor data is calculated according to equation (3) to obtain

Technical Field

the invention belongs to the field of measurement calibration and correction of electronic instruments, and relates to a steering difference correction method of a three-axis magnetic sensor.

background

magnetic field detection is an essential link in various scientific research and production activities related to magnetism, and the accuracy of surveying magnetic anomalies is very important. Among them, a three-axis magnetic sensor is often used as a tool for measuring three-component values of a magnetic field. However, since the processing processes are different and the level of the mounting process is also different, three axes of the magnetic sensor will not be strictly orthogonal, the sensitivities of the three axes will not be completely consistent, and zero drift may also occur, so that the magnetic field modulus and the actual value measured by the sensor in different forms at the same position have a large error, that is, a steering difference, and therefore, the correction of the steering difference is necessary.

The commonly used correction method of the three-axis magnetic sensor is ellipsoid fitting correction, namely, an ellipsoid equation is fitted by fully rotating the magnetic sensor and acquiring sufficient data, so that correction parameters are calculated. However, this method requires a large amount of data to be acquired, and it is necessary to ensure that the magnetic sensor can be rotated sufficiently to achieve a certain effect. But the effect has proved to be less than ideal for different magnetic sensors.

Disclosure of Invention

In order to solve the above problems, the present invention provides a method for correcting a steering difference of a three-axis magnetic sensor, which has wider applicability and better correction accuracy, and does not require a large amount of data.

in order to achieve the above object, the present invention provides a method for correcting a steering difference of a three-axis magnetic sensor, including the steps of:

s10, selecting a region which is open and has no magnetic substances around, rotating the magnetic sensor on a fixed point, and collecting triaxial magnetic field values Hx, Hy and Hz obtained by the magnetic sensor;

S20, setting an X ' Y ' Z ' axis as an actual coordinate axis of the magnetic sensor, and setting an XYZ axis as a standard coordinate axis, and constructing a measurement output model of the geomagnetic field vector;

s30, constructing a neural network identification system, and updating error parameters by adopting an iteration principle so as to obtain each estimation value of the error parameters;

s40, the compensated magnetic sensor data is calculated using the error parameter.

Preferably, the X ' Y ' Z ' axis is an actual coordinate axis of the magnetic sensor, the XYZ axis is a standard coordinate axis, a measurement output model of the geomagnetic field vector is constructed, and is expressed by the following formula,

Wherein the content of the first and second substances,To an asymmetric proportionality coefficient, K_xAnd K_ySensitivity scales in the direction of an actual coordinate axis OX ' of the magnetic sensor and sensitivity scales in the direction of an OY ' axis are respectively set to be 1 by taking the sensitivity scales in the direction of the OZ ' axis as a standard, namely the sensitivity scales are the same as the standard coordinate axis OZ of the magnetic sensor;

the space transformation matrix is adopted, beta represents an included angle between a Y ' axis and a Y axis, alpha represents an included angle between an X ' axis and an XOZ plane, and gamma represents an included angle between the X ' axis and the XOY plane; obtaining the product with cos alpha being approximately equal to 1, cos beta being approximately equal to 1, sin alpha being approximately equal to alpha, sin beta being approximately equal to beta and sin gamma being approximately equal to gamma

B＝[B_x,B_y,B_z]^Tis a zero offset error matrix, B_x，B_y，B_znamely zero offset values of the actual X axis, Y axis and Z axis of the magnetic sensor;

For the actual output component values of the magnetic sensors,The magnetic field values on the X axis, the Y axis and the Z axis which are actually obtained by the magnetic sensor are obtained;

setting an error matrixObtained by the formula (1)

and the further calculation is carried out to obtain,wherein E is an identity matrix;

ΔK_x＝K_x-1，ΔK_y＝K_y-1；

Conversion of formula (2) to

neglecting the second order small quantity deltas deltak, we get,

by using the two-norm vector, the method can be used,

Neglecting second order fractional quantitiesso as to obtain the compound with the characteristics of,

Obtained from the formula (3) and the formula (4),

Will be provided withΔ K, Δ S, B is taken into formula (5),

Preferably, [ Δ K ] in S30_x,ΔK_y,γ,α,β,B_x,B_y,B_z]The method for estimating the error parameters comprises the following steps:

s31, where t represents the number of iterations, and let t equal to 1, initialize a row matrix w with eight elements all being 1, and indicate eight error parameters to be estimated;

s32, setting iteration number M and learning efficiency μ, where M is 5000 and μ is 0.000015;

s33, according to the formula (6), the output of the magnetic sensor is adjustedThe development is that,

S34, calculating the total weight value z (t) ═ w × P^T；

s35, calculating an error vector

s36, according to WIDROW-HOFF learning rule, the parameter adjustment expression is w (t +1) ═ w (t) + μ e (t) P^T，

w is a matrix of parameters to be estimated, mu is learning efficiency, the speed of each parameter approaching a stable value can be adjusted, e (t) is an error vector, and P is an input matrix;

s37, t ═ t +1, return to S34;

and S38, when t is equal to the iteration times M, ending the iteration to obtain an error parameter estimation value.

Preferably, in S40, the compensated magnetic sensor data is calculated according to equation (3) to obtain

The invention has the advantages of wider applicability, no requirement of huge data volume and more effective inhibition of the steering difference of the magnetic sensor. After triaxial magnetic field data are processed by the method, the total field value of the magnetic field is calculated, referring to fig. 3, the total field value before correction is shown, the difference between the maximum value and the minimum value of numerical fluctuation is about 600nT, namely, the steering difference is about 600nT, and as can be seen from fig. 4, the fluctuation after correction is reduced to be within 80nT, the steering difference is effectively inhibited, and the accuracy in measuring the magnetic field value is improved.

Drawings

FIG. 1 is a flow chart illustrating the steps of a method for correcting a steering differential of a three-axis magnetic sensor according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of an error angle between a standard coordinate axis and an actual coordinate axis of a three-axis magnetometer in the method for correcting a steering difference of a three-axis magnetic sensor according to an embodiment of the present invention;

Fig. 3 is a diagram of a total magnetic field value before correction of the steering difference correction method of the three-axis magnetic sensor according to the embodiment of the present invention;

fig. 4 is a corrected magnetic field total field value diagram of the steering difference correction method for the three-axis magnetic sensor according to the embodiment of the present invention.

Detailed Description

in order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

On the contrary, the invention is intended to cover alternatives, modifications, equivalents and alternatives which may be included within the spirit and scope of the invention as defined by the appended claims. Furthermore, in the following detailed description of the present invention, certain specific details are set forth in order to provide a better understanding of the present invention. It will be apparent to one skilled in the art that the present invention may be practiced without these specific details.

Referring to fig. 1, which is a flowchart illustrating steps of a steering difference correction method for a three-axis magnetic sensor according to an embodiment of the present invention, the method includes the following steps:

S10, selecting a region which is open and has no magnetic substances around, rotating the magnetic sensor on a fixed point, and collecting triaxial magnetic field values Hx, Hy and Hz obtained by the magnetic sensor;

s20, setting an X ' Y ' Z ' axis as an actual coordinate axis of the magnetic sensor, and setting an XYZ axis as a standard coordinate axis, and constructing a measurement output model of the geomagnetic field vector; with reference to figure 2 of the drawings,

s30, constructing a neural network identification system, and updating error parameters by adopting an iteration principle so as to obtain each estimation value of the error parameters;

s40, the compensated magnetic sensor data is calculated using the error parameter.

setting the X ' Y ' Z ' axis as the practical coordinate axis of the magnetic sensor and the XYZ axis as the standard coordinate axis, constructing the measurement output model of geomagnetic field vector, expressed by the following formula,

Wherein the content of the first and second substances,To an asymmetric proportionality coefficient, K_xAnd K_ysensitivity scales in the direction of an actual coordinate axis OX ' of the magnetic sensor and sensitivity scales in the direction of an OY ' axis are respectively set to be 1 by taking the sensitivity scales in the direction of the OZ ' axis as a standard, namely the sensitivity scales are the same as the standard coordinate axis OZ of the magnetic sensor;

The space transformation matrix is adopted, beta represents an included angle between a Y ' axis and a Y axis, alpha represents an included angle between an X ' axis and an XOZ plane, and gamma represents an included angle between the X ' axis and the XOY plane; obtaining the product with cos alpha being approximately equal to 1, cos beta being approximately equal to 1, sin alpha being approximately equal to alpha, sin beta being approximately equal to beta and sin gamma being approximately equal to gamma

B＝[B_x,B_y,B_z]^Tis a zero offset error matrix, B_x，B_y，B_zNamely zero offset values of the actual X axis, Y axis and Z axis of the magnetic sensor;

For the actual output component values of the magnetic sensors,the magnetic field values on the X axis, the Y axis and the Z axis which are actually obtained by the magnetic sensor are obtained;

Setting an error matrixobtained by the formula (1)

and the further calculation is carried out to obtain,wherein E is an identity matrix;

ΔK_x＝K_x-1，ΔK_y＝K_y-1；

Conversion of formula (2) to

Neglecting the second order small quantity deltas deltak, we get,

By using the two-norm vector, the method can be used,

Neglecting second order fractional quantitiesSo as to obtain the compound with the characteristics of,

obtained from the formula (3) and the formula (4),

Will be provided withΔ K, Δ S, B is taken into formula (5),

at S30, [ Delta K ]_x,ΔK_y,γ,α,α,B_x,B_y,B_z]the method for estimating the error parameters comprises the following steps:

s31, where t represents the number of iterations, and let t equal to 1, initialize a row matrix w with eight elements all being 1, and indicate eight error parameters to be estimated;

s32, setting iteration number M and learning efficiency μ, where M is 5000 and μ is 0.000015;

S33, according toequation (6) represents the output of the magnetic sensorthe development is that,

s34, calculating the total weight value z (t) ═ w × P^T；

s35, calculating an error vector

S36, according to WIDROW-HOFF learning rule, the parameter adjustment expression is w (t +1) ═ w (t) + μ e (t) P^T，

W is a matrix of parameters to be estimated, mu is learning efficiency, the speed of each parameter approaching a stable value can be adjusted, e (t) is an error vector, and P is an input matrix;

s37, t ═ t +1, return to S34;

and S38, when t is equal to the iteration times M, ending the iteration to obtain an error parameter estimation value.

in S40, the magnetic sensor data after compensation is calculated according to equation (3) to obtain

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.