Fluxgate sensor for calibrating azimuth at slope and calibration method thereof

• US 20050114076A1
• Filed: 10/28/2004
• Published: 05/26/2005
• Est. Priority Date: 11/11/2003
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Lee, Woo-jong
• Granted Patent: US 7,363,185 B2
• Time in Patent Office: Days
• US Class Current: 702/151
• CPC Class Codes: G01C 17/30 Earth-inductor compasses
• CPC Class Codes: G01R 33/04 using the flux-gate principle
• CPC Class Codes: Y10T 29/49073 by assembling coil and core

Abstract

A two-axis fluxgate sensor has a driving pulse generating circuit which generates pulse signal and outputs as a driving signal, and X-axis and Y-axis fluxgates which are in proportional relation with each other. The two-axis fluxgate sensor generates voltage values of X-axis and Y-axis fluxgates corresponding to the magnetism which is generated from the driving signal, and a memory stores therein a neural network weight matrix. When the voltage values of the X-axis and Y-axis fluxgates are measured, a control unit compensates for the voltage values by using the neural network weight matrix which is stored in the memory, and computes an azimuth angle by using the compensated voltage values. An accurate azimuth angle can be obtained even at slopes.

Claims

1. A fluxgate sensor comprising:

a two-axis fluxgate having an X-axis fluxgate and a Y-axis fluxgate in substantially perpendicular relation with the X-axis fluxgate, the two-axis fluxgate outputting voltage values of the X-axis and Y-axis fluxgates corresponding to a terrestrial magnetism;

a memory storing therein a neural network weight matrix which corresponds to the X-axis and Y-axis fluxgate voltage values, respectively; and

a control unit calibrating the X-axis and Y-axis fluxgate voltage values based on the neural network weight matrix stored in the memory, and computing an azimuth angle by using the calibrated voltage values.

2. The fluxgate sensor according to claim 1, wherein the X-axis and Y-axis fluxgate voltage values comprise mapping values normalized by the control unit based on:

${\mathrm{Xf}}_{\mathrm{norm}}=\frac{2<br><br>\mathrm{Xf}-{\mathrm{Xf}}_{\max }-{\mathrm{Xf}}_{\min }}{\left({\mathrm{Xf}}_{\max }-{\mathrm{Xf}}_{\min }\right)}*\cos <br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\lambda <br><br>}$ ${\mathrm{Yf}}_{\mathrm{norm}}=\frac{2<br><br>\mathrm{Yf}-{\mathrm{Yf}}_{\max }-{\mathrm{Yf}}_{\min }}{\left({\mathrm{Yf}}_{\max }-{\mathrm{Yf}}_{\min }\right)}*\cos <br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\lambda <br><br>}$

where ‘

Xf’

is an actual voltage value of the X-axis fluxgate, ‘

Yf’

is an actual voltage value of the Y-axis fluxgate, ‘

Xf_norm’ and

‘

Yf_norm’

are normalized values of ‘

Xf’ and

‘

Yf’

, ‘

Xf_max’ and

‘

Xf_min’

are maximum and minimum values of ‘

Xf’

, ‘

Yf_max’ and

‘

Yf_min’

are maximum and minimum values of ‘

Yf’

, and λ

is a dip angle.

3. The fluxgate sensor according to claim 1, wherein the neural network weight matrix is computed by a training of the neural network, in which a predetermined data sheet is applied to a neural network software,

the predetermined data sheet recording therein the X-axis and Y-axis fluxgate voltage values corresponding to changes of pitch, roll and yaw angles of the fluxgate sensor.

4. The fluxgate sensor according to claim 3, wherein the predetermined data sheet is made by:

with the pitch and roll angles of the fluxgate sensor at a certain fixed angle, rotating once the fluxgate sensor to vary the yaw angle by a predetermined degree and measuring and storing the X-axis and Y-axis fluxgate voltage values; and

repeating the rotating, measuring and storing while varying one of the pitch and the roll angles.

5. An azimuth angle measuring method of a fluxgate sensor which comprises an X-axis fluxgate and a Y-axis fluxgate, and a memory, the azimuth angle measuring method comprising:

(a) measuring voltage values of the X-axis and the Y-axis fluxgates corresponding to changes of pitch, roll and yaw angles of the fluxgate sensor, normalizing the measured X-axis and Y-axis fluxgate voltage values and accordingly making a predetermined data sheet;

(b) making a neural network weight matrix by training a neural network based on the predetermined data sheet;

(c) storing the neural network weight matrix in the memory; and

(d) measuring the normalized voltage values of current x-axis and Y-axis fluxgates of the fluxgate sensor, and computing an azimuth angle by using the neural network weight matrix.

6. The azimuth angle measuring method according to claim 5, wherein (a) comprises:

(a1) with the pitch and roll angles of the fluxgate sensor fixed at ‘

0’

angle, rotating the fluxgate sensor once to vary the yaw angle by a predetermined degree, and measuring and normalizing the voltage values of the X-axis and Y-axis fluxgates;

(a2) measuring and storing a maximum, a minimum and a dip angle of the voltage values of the X-axis and Y-axis fluxgates;

(a3) with the pitch and roll angles of the fluxgate sensor fixed at a predetermined degree, rotating the fluxgate sensor once to vary the yaw angle by a predetermined degree and measuring the voltage values of the X-axis and Y-axis fluxgates;

(a4) computing a cos-function value and a sin-function value of the yaw angle and recognizing the computed values as reference values;

(a5) varying one of the pitch angle and the roll angle of the fluxgate sensor by a predetermined degree and repeating (a3) and (a4); and

(a6) measuring, normalizing and storing the X-axis and Y-axis fluxgate voltage values in every change of one of the pitch angle, the roll angle and the yaw angle, thereby making a data sheet.

7. The azimuth angle measuring method according to claim 6, wherein the normalization is performed based on:

${\mathrm{Xf}}_{\mathrm{norm}}=\frac{2<br><br>\mathrm{Xf}-{\mathrm{Xf}}_{\max }-{\mathrm{Xf}}_{\min }}{\left({\mathrm{Xf}}_{\max }-{\mathrm{Xf}}_{\min }\right)}*\cos <br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\lambda <br><br>}$ ${\mathrm{Yf}}_{\mathrm{norm}}=\frac{2<br><br>\mathrm{Yf}-{\mathrm{Yf}}_{\max }-{\mathrm{Yf}}_{\min }}{\left({\mathrm{Yf}}_{\max }-{\mathrm{Yf}}_{\min }\right)}*\cos <br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\lambda <br><br>}$

where ‘

Xf’

is an actual voltage value of the X-axis fluxgate, ‘

Yf’

is an actual voltage value of the Y-axis fluxgate, ‘

Xf_norm’ and

‘

Yf_norm’

are normalized values of ‘

Xf’ and

‘

Yf’

, ‘

Xf_max’ and

‘

Xf_min’

are maximum and minimum values of ‘

Xf’

, ‘

Yf_max’ and

‘

Yf_min’

are maximum and minimum values of ‘

Yf’

, and λ

is a dip angle.

8. The azimuth angle measuring method according to claim 5, wherein (b) comprises:

(b1) executing a neural network software;

(b2) forward processing of inputting output values of the X-axis and Y-axis fluxgates of the data sheet to the neural network, applying a predetermined weight and outputting as a cos-function value and a sin-function value;

(b3) backward processing of obtaining an error value by comparing the cos-function value and the sin-function value with a corresponding reference value recorded in the data sheet, and if the error value is determined to exceed a predetermined allowable range, correcting the weight so that the cos-function value and the sin-function value can be brought close to the reference value;

(b4) repeating (b2) and (b3) until the error value is within the predetermined allowable range; and

(b5) when the error value is within the predetermined allowable range, storing the corresponding weight and thereby making a neural network weight matrix.

9. The azimuth angle measuring method according to claim 5, wherein (d) comprises:

(d1) measuring a normalized value of current X-axis and Y-axis fluxgate voltage values;

(d2) computing a cos-function value by applying a corresponding weight of the neural network weight matrix, which is stored in the memory, to an input of the X-axis fluxgate voltage values;

(d3) computing a sin-function value by applying a corresponding weight of the neural network weight matrix, which is stored in the memory, to an input of the Y-axis fluxgate voltage values; and

(d4) computing an azimuth angle by using the cos-function value and the sin-function value.

10. The azimuth angle measuring method according to claim 9, wherein (d4) computes the azimuth angle by:

ψ

=tan<sup>−

1</sup>(sin<sub>—

out/cos—

out)</sub>

where, ψ

is the azimuth angle, sin_out is the sin-function value, and cos_out is the cos-function value.