CN106197477A - A kind of indoor positioning course error bearing calibration - Google Patents

Abstract

The invention belongs to indoor navigation field of locating technology, be specifically related to a kind of indoor positioning course error bearing calibration, this bearing calibration comprises the steps: step one, chooses orientation, record course actual value；Step 2, measures orientation, obtains corresponding heading measure value；Step 3, calculates course error and also sets up course error calibration model, poor to heading measure value and course actual value, obtains course error, and course error is regarded as the nonlinear function of heading measure value, sets up the course error calibration model of unknown parameters；Step 4, carries out curve fitting according to predetermined threshold value, determines the parameter of course error calibration model；Step 5, corrects the course of reality according to course error calibration model.The present invention can adaptively determine course error calibration model, accurately completes error correction and accuracy compensation, solves the problem of the correction of course in indoor positioning and navigation procedure.

Description

A kind of indoor positioning course error bearing calibration

Technical field

The invention belongs to indoor navigation field of locating technology, be specifically related to a kind of indoor positioning course error bearing calibration.

Background technology

Traditional outdoor positioning technology by means of global positioning system (Global Positioning System, GPS), Utilize the direct communication between satellite and mobile object, can complete to position accurately.Along with growth in the living standard and shifting The development of dynamic Internet technology, the demand that people position for indoor navigation is the strongest.At large-scale shopping plaza, building Internal structure is complicated, and miscellaneous retail shop is distributed in different corners, cannot find purpose commodity, now in client's short time Provided that indoor navigation positioning service, provide position and the merchandise news of retail shop for client, client can be helped to look for rapidly To target retail shop.In tier building during breaking out of fire, withdraw rapidly not a duck soup, at this moment need indoor navigation positioning service to refer to Lead be in danger crowd quickly and accurately withdraw to safety area, prevent great casualties and property loss.Pass The gps signal of system is due to blocking by building etc., it is impossible to accurately find range, and has been difficult to location under indoor environment.Currently Occur in that a lot of indoor navigation location technology, achieve good effect.Most important function of navigating is to provide carrier accurately Course angle information, due to zero offset error and the drift etc. of gyroscope of acceleration, exports according to Inertial Measurement Unit (IMU) merely Course angle can not meet requirement, the combination direction finding equipment being made up of acceleration and magnetometer can improve the accuracy in course And reliability.But magnetometer in use can be disturbed by external magnetic field, need corresponding correction and compensate.Accurately Real-time magnetometer correction and compensation technique are the key points of course output.

Utilize the combination direction finding device of acceleration and magnetometer by measuring the magnetic field of the earth component size at carrier three axle Determine course, have anti-interference, radiationless, volume is little and low cost and other advantages.But magnetometer can introduce alignment error, zero-bit Error and scale factor error and extraneous magnetic substance disturb equal error.First three error does not relies on actual application environment, once mends Repay and can forever use, but the magnetic substance in practical service environment disturbs according to intensity and the difference in direction, needs real time correction And compensation.

At present, occurring in that correction and the compensation technique much for magnetometer, certain methods has considered magnetometer Error, is modeled error, determines model parameter and then completes correction, but can not meet the correction of real-time high-precision Requirement；The magnetic substance interference modeling of environment is a general error model by other methods, does not take into full account varying environment Between diversity.

The present invention the most in this context, devises a kind of indoor positioning course error school based on nonlinear model Correction method, the method can adaptively determine corresponding error model parameters according to actual magnetic substance interference environment, accurate in real time Really complete correction.This method is simple to operate, it is not necessary to extra auxiliary device, has that self adaptation degree is high, real-time is good, pervasive Property the advantage such as strong.

Summary of the invention

The invention provides a kind of indoor positioning course error bearing calibration, this error calibration method can be the most true Vectoring error correction model, accurately completes error correction and accuracy compensation, solves course in indoor positioning and navigation procedure The problem of correction.

In sum, for reaching above-mentioned purpose, the technical solution adopted in the present invention is as follows:

A kind of indoor positioning course error bearing calibration, this bearing calibration use Course Measuring Unit include accelerometer and Magnetometer,

It is characterized in that, realize in accordance with the following steps:

Step one: choose orientation, uniformly chooses n orientation in correction environment, and course actual value is designated as

Step 2: measure orientation, utilize in step one described in accelerometer and magnetometer combination direction finding measurement device The orientation chosen, obtains corresponding heading measure value

Step 3: calculate course error and also set up course error calibration model, to the heading measure value in described step 2 and Course actual value in described step one is poor, obtains course error And described course errorRegard course measured value in described step 2 asNonlinear function, set up the boat of unknown parameters To error correction model；

Step 4: carry out curve fitting according to predetermined threshold value, determines the ginseng of course error correction model in described step 3 Number；

Step 5: correct the course of reality according to course error correction model in described step 4.

Further, the nonlinear function in described step 3 is:

WhereinFor described course error,For described heading measure value, A, B ... K, M are the correction of described course error The parameter of model, N is the high math power of representative function, and N ≠ 0.

Further, the method determining course error calibration model parameter in described step 4 is:

1) described formula (9) is regarded as described heading measure valueN times function, being write as equation group form is:

WhereinRepresent the error of i-th orientation,Represent in i-th orientation Heading measure value, A, B ... K, M are equation parameter；

2) value of N is calculated according to auto-adapted fitting method；

3) method of least square is utilized to solve function parameter A, the B of described formula (9) ... K, M.

Further, described auto-adapted fitting method, it is characterised in that particularly as follows:

For the N value of certain matching, the initial value of N is 1, if

Then N is and is taken, otherwise N=N+1, continues matching next time；

WhereinFor the fitting function value utilizing described course actual value to try to achieve,For described course error.

Further, described method of least square solved function parameter A that utilizes, B ... K, M, it is characterised in that particularly as follows:

According to the match value N obtained, being write the equation group described in formula (10) as matrix form is:

Parameter A is calculated according to method of least square, B ... K, M be:

Further, in described step 2, the measuring method of course measured value is:

Taking navigational coordinate system is sky, northeast coordinate system (ENU), and carrier coordinate system is Ox_by_bz_b, wherein Oy_bAlong carrier forward direction Direction, Ox_bAxle is perpendicular to forward direction towards the right side, Oz_bAxle and Oy_bAxle and Ox_bAxle constitutes right-handed coordinate system, utilizes under carrier coordinate system three Axle acceleration value (a_x,a_y,a_z) calculate carrier roll angle roll and angle of pitch pitch, as follows:

$\left\{\begin{array}{c}roll=\arctan (a_y/a_z)\\ pitch=\arctan \left(\frac{-a_x}{\sqrt{a_y^2+a_z^2}}\right)\end{array}\right.---\left(1\right)$

Spin matrix according to described formula (1) calculating carrier coordinate system navigation coordinate system conversion:

$\begin{array}{c}{R}_b^t=\left(\begin{array}{ccc}1& 0& 0\\ 0& \cos r& \sin r\\ 0& -\sin r& \cos r\end{array}\right)\left(\begin{array}{ccc}\cos \mathrm{\θ}& 0& -\sin \mathrm{\θ}\\ 0& 1& 0\\ \sin \mathrm{\θ}& 0& \cos \mathrm{\θ}\end{array}\right)\\ =\left(\begin{array}{ccc}\cos \mathrm{\θ}& 0& -\sin \mathrm{\θ}\\ \sin \mathrm{\θ}\·\sin r& \cos r& \cos \mathrm{\θ}\·\sin r\\ \sin \mathrm{\θ}\·\cos r& -\sin r& \cos \mathrm{\θ}\·\cos r\end{array}\right)\end{array}---\left(2\right)$

Wherein, wherein θ=pitch, r=roll, calculated by formula (1) and obtain；

By three axle magnetic strength of carrier any attitudeIt is transformed into the magnetic field intensity under navigational coordinate system

$\left(\begin{array}{c}{M}_x^t\\ M_y^t\\ M_z^t\end{array}\right)=R_b^t\·\left(\begin{array}{c}{M}_x^b\\ M_y^b\\ M_z^b\end{array}\right)---\left(3\right)$

Field strength values calculating course angle according under navigational coordinate system:

Further, the method for building up of the course error calibration model described in described step 3 is:

The soft magnetism of magnetometer place environment and Hard Magnetic interference are different according to intensity and the direction in magnetic field, become course defeated The main source of error gone out.Assume that the x-axis ground magnetic component of magnetometer output and y-axis magnetic strength component are by structure around without magnetic interference Become a positive round；

Hard Magnetic disturbs the metal after resulting from permanent magnetic material or being magnetized, and it can make magnetic field intensity curve of output The center of circle offset, be periodically and nonlinear on the impact of final heading, error modeling is:

Soft magnetism interference results from the interaction near magnetometer between magnetic material, and it can make the song that magnetic field intensity exports Line is become oval by positive round, is also periodically and nonlinear on the impact of final heading, and error modeling is:

Therefore magnetometer is subject to Hard Magnetic and soft magnetism interference can be modeled as:

According to f_hmAnd f_smPeriodicity and the non-linear f that understands also have periodically and non-linear.Therefore, further can be by Course error is modeled as:

WhereinRepresent course error,Expression heading measure value, A, B ... K, M are the parameter of function, and N is representative function High math power.

From the scheme of foregoing invention, the beneficial effects of the present invention is:

1) present invention proposes method low cost, precision are high, and universality is strong, it is possible to significantly improves navigational course precision, enters And improve the reliability of navigation positioning system.

2) method that the present invention proposes has stronger adaptivity, can determine course error in real time according to correction environment Calibration model, accurately completes error correction and accuracy compensation.

3) assembly of the invention is constituted and conceptual design is simple, only utilizes a little several courses actual value to complete The determination of error model, thus the complexity of method is relatively low.

4) method that the present invention proposes is compared with traditional method, is not merely to be corrected magnetic field intensity, but right Final output course error modeling, has taken into full account course error influence factor, and therefore correction result is relatively reliable.

Accompanying drawing explanation

Fig. 1 is Hard Magnetic interference output curve diagram in magnetometer environment；

Fig. 2 is the Hard Magnetic interference PERIODIC INTERFERENCE figure to course in magnetometer environment；

Fig. 3 is soft magnetism interference output curve diagram in magnetometer environment；

Fig. 4 is the soft magnetism interference PERIODIC INTERFERENCE figure to course in magnetometer environment；

Fig. 5 is correction of course flow chart；

Fig. 6 is error model curve-fitting results figure；

Fig. 7 is error comparison diagram before and after correction.

Detailed description of the invention

The present invention will be further described in detail with specific embodiment below in conjunction with the accompanying drawings, in order to is more fully understood that this Invention, but protection scope of the present invention is not limited to this.

The measurement apparatus that embodiment is selected is MPU6050 accelerometer and HMC5883 magnetometer inertia measuring module, both There is identical or parallel coordinate axes definition.MPU6050 has ± the measurement range of 16g, and stability is 0.01g, differentiates Rate is up to 6.1 × 10^-5The measurement scope of g, HMC5883 magnetometer is [-8,8] Gauss, can be accurate to 1mgauss, utilize Du Pont Module and development board are connected through Mini USB line and store data with host computer communication by line.

As it is shown in figure 5, the implementation environment of the present embodiment is a typical indoor environment, owing to there is substantial amounts of electricity in environment Magnetic material and there is the interference of obvious magnetic substance, a kind of indoor positioning course error bearing calibration, the enforcement that this bearing calibration is concrete Step is:

Step one: every 30 ° of selected orientations in 0 °～360 °, course actual value is designated as

Step 2: utilize described measurement apparatus to measure acceleration and the magnetic of the orientation chosen in described step one respectively Field intensity data, obtain corresponding heading measure valueThe horizontal stroke of carrier is calculated according to acceleration information Roll angle and the angle of pitch, and then calculate the spin matrix of carrier coordinate system navigation coordinate system conversion.

By described spin matrix, the magnetic field intensity under carrier coordinate system is transformed under navigational coordinate system, according to formula (4) course angle is calculated.

Taking navigational coordinate system is sky, northeast coordinate system (ENU), and carrier coordinate system is Ox_by_bz_b, wherein Oy_bAlong carrier forward direction Direction, Ox_bAxle is perpendicular to forward direction towards the right side, Oz_bAxle and Oy_bAxle and Ox_bAxle constitutes right-handed coordinate system, utilizes under carrier coordinate system three Axle acceleration value (a_x,a_y,a_z) calculate carrier roll angle roll and angle of pitch pitch, as follows:

$\left\{\begin{array}{c}roll=arctan(a_y/a_z)\\ pitch=arctan\left(\frac{-a_x}{\sqrt{a_y^2+a_z^2}}\right)\end{array}\right.---\left(1\right)$

Spin matrix according to formula (1) calculating carrier coordinate system navigation coordinate system conversion:

$\begin{array}{c}{R}_b^t=\left(\begin{array}{ccc}1& 0& 0\\ 0& \cos r& \sin r\\ 0& -\sin r& \cos r\end{array}\right)\left(\begin{array}{ccc}\cos \mathrm{\θ}& 0& -\sin \mathrm{\θ}\\ 0& 1& 0\\ \sin \mathrm{\θ}& 0& \cos \mathrm{\θ}\end{array}\right)\\ =\left(\begin{array}{ccc}\cos \mathrm{\θ}& 0& -\sin \mathrm{\θ}\\ \sin \mathrm{\θ}\·\sin r& \cos r& \cos \mathrm{\θ}\·\sin r\\ \sin \mathrm{\θ}\·\cos r& -\sin r& \cos \mathrm{\θ}\·\cos r\end{array}\right)\end{array}---\left(2\right)$

Wherein, wherein θ=pitch, r=roll, calculated by formula (1) and obtain.

By three axle magnetic strength of carrier any attitudeIt is transformed into the magnetic field intensity under navigational coordinate system

$\left(\begin{array}{c}{M}_x^t\\ M_y^t\\ M_z^t\end{array}\right)=R_b^t\·\left(\begin{array}{c}{M}_x^b\\ M_y^b\\ M_z^b\end{array}\right)---\left(3\right)$

Field strength values calculating course angle according under navigational coordinate system:

Step 3: calculate course error and set up course error calibration model, to the heading measure value obtained in described step 2 Poor with course actual value in described step one, obtain course error And described course errorRegard course measured value in described step 2 asNonlinear function, set up the boat of unknown parameters To error correction model；

The method for building up of course error calibration model is:

The soft magnetism of magnetometer place environment and Hard Magnetic interference are different according to intensity and the direction in magnetic field, become course defeated The main source of error gone out.Assume that the x-axis ground magnetic component of magnetometer output and y-axis magnetic strength component are by structure around without magnetic interference Become a positive round.

Hard Magnetic interference result from permanent magnetic material or be magnetized after metal, this interference can make magnetic field intensity with Actual value produces deviation, as it is shown in figure 1, the circle with X-direction magnetic-field component as transverse axis, with Y-direction magnetic-field component as the longitudinal axis, in figure The center of circle there occurs skew.And Hard Magnetic interference presents non-linear relation to the impact in course, as in figure 2 it is shown, with course as transverse axis, With course error as the longitudinal axis, in magnetometer environment, Hard Magnetic interference is periodically and nonlinear, by it on the impact of final heading It is modeled as:

Soft magnetism interference results from the interaction near magnetometer between magnetic material, and it can make the song that magnetic field intensity exports Line is become oval by positive round, as it is shown on figure 3, with X-direction magnetic-field component as transverse axis, with Y-direction magnetic-field component as the longitudinal axis, originally should Should become oval for circular magnetic strength distribution；Fig. 4 shows the soft magnetism interference influence curve to course error, with course is Transverse axis, with course error as the longitudinal axis, soft magnetism interference is also periodically and nonlinear on the impact of final heading, error modeling For:

Therefore magnetometer is subject to Hard Magnetic and soft magnetism interference can be modeled as:

According to f_hmAnd f_smPeriodicity and the non-linear f that understands also have periodically and non-linear.Therefore, further can be by Course error is modeled as:

WhereinRepresent course error,Expression heading measure value, A, B ... K, M are the parameter of function, and N is representative function High math power.

In sum, described nonlinear function is:

WhereinFor described course error,For described heading measure value, A, B ... K, M are the correction of described course error The parameter of model, N is the high math power of representative function, and N ≠ 0.

Step 4: carry out curve fitting according to pre-determined threshold threshold value, determines error correction model in course in described step 3 Parameter；

1) method determining course error calibration model parameter is:

Regard described formula (9) as described heading measure valueN times function, being write as equation group form is:

Wherein Δ φ_i, i=[1:n] represents the error of i-th orientation,Represent i-th orientation On heading measure value, A, B ... K, M are equation parameter.

2) value of N is calculated according to auto-adapted fitting method；

For the N value of certain matching, the initial value of N is 1, if

Then N is and is taken, otherwise N=N+1, continues matching next time.

WhereinFor the fitting function value utilizing described course actual value to try to achieve,For described course error.

3) method of least square is utilized to solve function parameter A, the B of described formula (9) ... K, M are particularly as follows: according to the plan obtained Close

Value N, being write the equation group described in described formula (10) as matrix form is:

Parameter A is calculated according to method of least square, B ... K, M be:

As shown in Figure 6, with course value of sampling as transverse axis, the actual measurement profile made for the longitudinal axis with course error and matching Curve, matching obtains N=4, and corresponding Error model coefficients is: Coef=[-0.0068,0.0944 ,-0.3545,0.2051, 0.2844]。

Step 5: correct the course of reality according to course error correction model in described step 4.

Step 6: described measurement apparatus rotated a circle, gathers secondary data, correct for the method for inspection every 15 ° Property；

As shown in table 1, the error contrast situation before and after table 1 shows the method for the invention correction.

Error contrast before and after table 1 correction of course

Fig. 7 is the error comparison diagram before and after utilizing described method to be corrected course, with course sampled value as transverse axis, With course error as the longitudinal axis, before the correction made course error curve and correction after course error curve.

Table 2 is the data adding up the mean error before and after course error correction and error to standard deviation.

Main precision index contrast before and after table 2 correction of course

From table 2 it can be seen that before utilizing the method described in invention to the correction of course, mean error is 12.9 °, and standard deviation is 15.4 °, utilizing the mean error after the correction of described method is 3.2 °, and standard deviation is 3.4 °.Improve course precision significantly.

Being noted that above-described embodiment is illustrative not limiting to technical solution of the present invention, art is general The equivalent leading to technical staff or other amendment made according to prior art, as long as beyond technical solution of the present invention Thinking and scope, within should be included in interest field of the presently claimed invention.

Claims (7)

1. an indoor positioning course error bearing calibration, this bearing calibration uses Course Measuring Unit to include accelerometer and magnetic Power meter, it is characterised in that realize in accordance with the following steps:

Step one: choose orientation, uniformly chooses n orientation in correction environment, and course actual value is designated as

Step 2: measure orientation, utilize in step one described in accelerometer and magnetometer combination direction finding measurement device and choose Orientation, obtain corresponding heading measure value

Step 3: calculate course error and also set up course error calibration model, to the heading measure value in described step 2 and described Course actual value in step one is poor, obtains course error And described course errorRegard course measured value in described step 2 asNonlinear function, set up the boat of unknown parameters To error correction model；

Step 4: carry out curve fitting according to predetermined threshold value, determines the parameter of course error correction model in described step 3；

Step 5: correct the course of reality according to course error correction model in described step 4.

A kind of indoor positioning course error bearing calibration the most according to claim 1, it is characterised in that in described step 3 Nonlinear function be:

WhereinFor described course error,For described heading measure value, A, B ... K, M are described course error calibration model Parameter, N is the high math power of representative function, and N ≠ 0.

A kind of indoor positioning course error bearing calibration the most according to claim 1, it is characterised in that in described step 4 The method determining course error calibration model parameter is:

1) described formula (9) is regarded as described heading measure valueN times function, being write as equation group form is:

Wherein Δ φ_i, i=[1:n] represents the error of i-th orientation,Represent in i-th orientation Heading measure value, A, B ... K, M are equation parameter；

2) value of N is calculated according to auto-adapted fitting method；

3) method of least square is utilized to solve function parameter A, the B of described formula (9) ... K, M.

A kind of indoor positioning course error bearing calibration the most according to claim 3, it is characterised in that described self adaptation is intended Legal, particularly as follows:

For the N value of certain matching, the initial value of N is 1, if

Then N is and is taken, otherwise N=N+1, continues matching next time；

WhereinFor the fitting function value utilizing described course actual value to try to achieve,For described course error.

A kind of indoor positioning course error bearing calibration the most according to claim 3, it is characterised in that described utilization is minimum Square law solved function parameter A, B ... K, M, it is characterised in that particularly as follows:

According to the match value N obtained, being write the equation group described in formula (10) as matrix form is:

Parameter A is calculated according to method of least square, B ... K, M be:

A kind of indoor positioning course error bearing calibration the most according to claim 1, it is characterised in that in described step 2 The measuring method of heading measure value is:

Taking navigational coordinate system is sky, northeast coordinate system (ENU), and carrier coordinate system is Ox_by_bz_b, wherein Oy_bAlong carrier forward direction, Ox_bAxle is perpendicular to forward direction towards the right side, Oz_bAxle and Oy_bAxle and Ox_bAxle constitutes right-handed coordinate system, utilizes three axles under carrier coordinate system to accelerate Angle value (a_x,a_y,a_z) calculate carrier roll angle roll and angle of pitch pitch, as follows:

$\left\{\begin{array}{c}roll=\arctan (a_y/a_z)\\ pitch=\arctan \left(\frac{-a_x}{\sqrt{a_y^2+a_z^2}}\right)\end{array}\right.---\left(1\right)$

Spin matrix according to described formula (1) calculating carrier coordinate system navigation coordinate system conversion:

$\begin{array}{c}{R}_b^t=\left(\begin{array}{ccc}1& 0& 0\\ 0& \cos r& \sin r\\ 0& -\sin r& \cos r\end{array}\right)\left(\begin{array}{ccc}\cos \mathrm{\θ}& 0& -\sin \mathrm{\θ}\\ 0& 1& 0\\ \sin \mathrm{\θ}& 0& \cos \mathrm{\θ}\end{array}\right)\\ =\left(\begin{array}{ccc}\cos \mathrm{\θ}& 0& -\sin \mathrm{\θ}\\ \sin \mathrm{\θ}\·\sin r& \cos r& \cos \mathrm{\θ}\·\sin r\\ \sin \mathrm{\θ}\·\cos r& -\sin r& \cos \mathrm{\θ}\·\cos r\end{array}\right)\end{array}---\left(2\right)$

Wherein, wherein θ=pitch, r=roll, calculated by formula (1) and obtain.

By three axle magnetic strength of carrier any attitudeIt is transformed into the magnetic field intensity under navigational coordinate system

$\left(\begin{array}{c}{M}_x^t\\ M_y^t\\ M_z^t\end{array}\right)=R_b^t\·\left(\begin{array}{c}{M}_x^b\\ M_y^b\\ M_z^b\end{array}\right)---\left(3\right)$

Field strength values calculating course angle according under navigational coordinate system:

A kind of indoor positioning course error bearing calibration the most according to claim 1, it is characterised in that in described step 3 The method for building up of described course error calibration model is:

The soft magnetism of magnetometer place environment and Hard Magnetic interference are different according to intensity and the direction in magnetic field, become course output Main source of error.Assume that, around without magnetic interference, the x-axis ground magnetic component of magnetometer output and y-axis magnetic strength component will constitute one Individual positive round.Hard Magnetic disturbs the metal after resulting from permanent magnetic material or being magnetized, and it can make magnetic field intensity curve of output The center of circle offset, be periodically and nonlinear on the impact of final heading, error modeling is:

Soft magnetism interference results near magnetometer the interaction between magnetic material, it can make curve that magnetic field intensity exports by Positive round becomes oval, is also periodically and nonlinear on the impact of final heading, and error modeling is:

Therefore magnetometer is subject to Hard Magnetic and soft magnetism interference can be modeled as:

According to f_hmAnd f_smPeriodicity and the non-linear f that understands also have periodically and non-linear.Therefore, further can by course by mistake Difference is modeled as:

WhereinRepresent course error,Represent heading measure value, A, B ... K, M are the parameter of function, N be representative function High math power.