CN103148813B - For the treatment of the method for GPS deformation measurement data - Google Patents

Abstract

The present invention proposes a kind of method for the treatment of GPS deformation measurement data, comprising: the first step: obtain base station and monitoring station current epoch observation data and broadcast ephemeris data, utilize carrier phase observable to form two difference observation equation; Second step: the maximum deformation quantity between two epoch is set; 3rd step: adopt Cholesky to decompose and build Ambiguity Search Space; 4th step: search blur level, and make variance ratio be greater than 3; 5th step: single epoch integer ambiguity static solution coordinate X, Y, the Z and its covariance matrix that obtain monitoring point; 6th step: utilize mean gap method to carry out deformation examination to the deformation values between current epoch and last epoch; 7th step: after the inspection of the 6th step, using the robust least square steady-state solution of two observation equation superpositions epoch as the result of current epoch or the result obtaining current epoch according to robust sequential adjustment.This method is very accurate.

Description

For the treatment of the method for GPS deformation measurement data

Technical field

The present invention relates to a kind of method for the treatment of data, particularly a kind of method for the treatment of GPS deformation measurement data.

Background technology

Along with the development of GPS technology, GPS is more and more for deformation monitoring.But in deformation monitoring environment, usual multipath or diffracted signal stronger, cause cycle slip occurrence frequency higher, and the observation data information that conventional least square search procedure (LS), fast ambiguity search's method (FARA), the irrelevant method of adjustment of least square (LAMBDA) reducing correlativity between blur level, Cholesky decomposition search procedure etc. all need the multiple epoch utilized in a period of time, and the observation data of this multiple epoch can not occur cycle slip, otherwise determine integer ambiguity by being difficult to.

Therefore in deformation monitoring application, the process of gps data generally adopts simple epoch solution method to avoid Detection of Cycle-slip, namely little according to monitoring point change in location feature, utilizes deflection to retrain and carries out single epoch ambiguity resolution.Equally due to the restriction of monitoring of environmental, visual gps satellite is usually less, causes satellite distribution graphic structure intensity more weak, add the impact of multipath effect, make obtained result precision not high, in Deformation Series, there will be the situation of saltus step, thus be difficult to carry out early warning according to monitoring result.

For the deficiency of the method for the treatment of GPS deformation measurement data of the prior art above, research precision and the new gps data disposal route that all gets a promotion of reliability, have positive meaning to the development of technology for deformation monitoring and its practical application in engineering.

Summary of the invention

Cannot saltus step be avoided for the method for the treatment of GPS deformation measurement data of the prior art, cause being difficult to the deficiency according to the accurate early warning of monitoring result, the present invention proposes a kind of method for the treatment of GPS deformation measurement data of novelty.

The present invention proposes a kind of method for the treatment of GPS deformation measurement data, comprising: the first step: obtain base station and monitoring station current epoch observation data and broadcast ephemeris data, utilize carrier phase observable to form two difference observation equation; Second step: the maximum deformation quantity between two epoch is set; 3rd step: adopt Cholesky to decompose and build Ambiguity Search Space; 4th step: search blur level, and make variance ratio be greater than 3; 5th step: single epoch integer ambiguity static solution coordinate X, Y, the Z and its covariance matrix that obtain monitoring point; 6th step: utilize mean gap method to carry out deformation examination to the deformation values between current epoch and last epoch; 7th step: after the inspection of the 6th step, if be not subjected to displacement between two epoch, then using the result of the robust least square steady-state solution of two observation equation superpositions epoch as current epoch; If the assay of the 6th step is have displacement to occur between two epoch, then obtain the result of current epoch according to robust sequential adjustment.

In one embodiment, in the third step:

$-\frac{{\mathrm{\δd}}_{\max }}{{\mathrm{\λl}}_{11}}\≤{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}\≤\frac{{\mathrm{\δd}}_{\max }}{{\mathrm{\λl}}_{11}}$

$\frac{-\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2}-l_{21}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}}{{\mathrm{\λl}}_{22}}\≤{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_2}\≤\frac{\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2}-l_{21}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}}{{\mathrm{\λl}}_{22}}$

$\frac{-\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2-C^2}-l_{31}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}-l_{32}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_2}}{\mathrm{\λ}{l}_{33}}\≤{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_3}\≤\frac{\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2-C^2}-l_{31}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}-l_{32}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_2}}{{\mathrm{\λl}}_{33}}$

Wherein:

$B=l_{11}\·\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1,}$ $C=l_{21}\·\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}+l_{22}\·\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_2},$

δ d _maxfor the maximum deformation quantity of monitoring point, for two poor integer ambiguity medial error, λ is

The wavelength of carrier phase,

$L^{-1}=\left[\begin{array}{ccc}{l}_{11}& & \\ l_{21}& l_{22}& \\ l_{31}& l_{32}& l_{33}\end{array}\right]$ For the system of equations matrix A A of the symmetric positive definite that solving condition system of equations and error equation group form ^tcholesky decompose lower triangular matrix.

In one embodiment, in the 6th step, following sub-step is also comprised:

First sub-step: by observed reading correction acquisition two epoch empiric variance:

$S_0^2=\frac{{\left(V^T\mathrm{PV}\right)}_I+{\left(V^T\mathrm{PV}\right)}_{\mathrm{II}}}{f}$

Wherein f be two epoch degree of freedom sum, i.e. f=n ₁-3+n ₂-3, wherein n ₁, n ₂be the number of two difference observation equation two epoch, V is observed reading residual vector, and P is observed reading power battle array;

Second sub-step: by the coordinate difference d of two epoch _i(i=1,2,3) obtain weight unit empiric variance

${\stackrel{\‾}{S}}_0^2=\frac{d^T{P}_{d}d}{3}$

Wherein, d=X _iI-X _i, represent the difference vector of monitoring point at the coordinate of front and back two epoch:

$d=\left(\begin{array}{cc}-1& 1\end{array}\right)\left(\begin{array}{c}{X}_I\\ X_{\mathrm{II}}\end{array}\right)$

Wherein P _dpower battle array for d:

$P_d=Q_d^{-1}$

Q herein _dby $d=\left(\begin{array}{cc}-1& 1\end{array}\right)\left(\begin{array}{c}{X}_I\\ X_{\mathrm{II}}\end{array}\right)$ Propagate law according to covariance to obtain:

$Q_d=\left(\begin{array}{cc}-1& 1\end{array}\right)\left[\begin{array}{cc}{Q}_{X_I{X}_I}& 0\\ 0& Q_{X_{\mathrm{II}}{X}_{\mathrm{II}}}\end{array}\right]\left(\begin{array}{c}-1\\ 1\end{array}\right)=Q_{X_I{X}_I}+Q_{X_{\mathrm{II}}{X}_{\mathrm{II}}}$

3rd sub-step: build statistic:

$F_{h,f}=\frac{{\stackrel{\‾}{S}}_0^2}{S_0^2}=\frac{d^T{P}_{d}d}{h{S}_0^2}$

4th sub-step: select level of signifiance α=0.01, by the F obtained _h,fwith the F found from F distribution table _α(h, f) tantile compares, if

F _h,f＞F _α(h,f)

Then show be greater than namely have:

P{[F _h,f＞F _α(h,f)]}＝α

Thus judge have displacement to occur between two epoch;

If cannot F be drawn _h,f> F _α(h, f), then judge not to be subjected to displacement between two epoch.

In one embodiment, in the 4th step, utilize least square method to search for blur level.

In one embodiment, in the 5th step, utilize least square method to obtain single epoch integer ambiguity static solution coordinate X, Y, Z and its covariance matrix of monitoring point.

Method hinge structure for the treatment of GPS deformation measurement data according to the present invention brings following progress: avoid Detection of Cycle-slip; Adopt Cholesky to decompose and build Ambiguity Search Space, search efficiency is higher; Can judge that the coordinate difference between epoch is caused by displacement or caused by error by deformation examination, improve the reliability of result; Forecast precision is higher, improves the treatment effeciency of GPS deformation measurement data.

Accompanying drawing explanation

Also with reference to accompanying drawing, the present invention is described in more detail based on the embodiment being only indefiniteness hereinafter.Wherein:

Fig. 1 is the FB(flow block) according to method of the present invention.

Embodiment

Come below with reference to accompanying drawings to introduce the present invention in detail.

The object of the invention is to the deficiency overcoming the method for the treatment of GPS deformation measurement data of the prior art, a kind of GPS single epoch Method of Deformation Monitoring Data Processing based on deformation examination is provided.

Fig. 1 shows the process flow diagram according to method of the present invention.

With reference to Fig. 1, method according to the present invention mainly comprises the steps:

The first step: obtain base station and monitoring station current epoch observation data and broadcast ephemeris data, utilizes carrier phase observable to form two difference observation equation.

Second step: according to deformation monitoring Properties of Objects, arranges the generable maximum deformation quantity be out of shape between two epoch.

3rd step: adopt Cholesky to decompose and build Ambiguity Search Space.

Wherein:

$-\frac{{\mathrm{\δd}}_{\max }}{{\mathrm{\λl}}_{11}}\≤{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}\≤\frac{{\mathrm{\δd}}_{\max }}{{\mathrm{\λl}}_{11}}$

$\frac{-\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2}-l_{21}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}}{{\mathrm{\λl}}_{22}}\≤{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_2}\≤\frac{\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2}-l_{21}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}}{{\mathrm{\λl}}_{22}}$

$\frac{-\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2-C^2}-l_{31}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}-l_{32}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_2}}{\mathrm{\λ}{l}_{33}}\≤{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_3}\≤\frac{\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2-C^2}-l_{31}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}-l_{32}\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_2}}{{\mathrm{\λl}}_{33}}$

Wherein:

$B=l_{11}\·\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1,}$ $C=l_{21}\·\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_1}+l_{22}\·\mathrm{\λ}{\mathrm{\δ}}_{\▿\mathrm{\Δ}{N}_2},$

δ d _maxfor the maximum deformation quantity of monitoring point, for two poor integer ambiguity medial error, λ is the wavelength of carrier phase,

$L^{-1}=\left[\begin{array}{ccc}{l}_{11}& & \\ l_{21}& l_{22}& \\ l_{31}& l_{32}& l_{33}\end{array}\right]$ For system of equations (normal equation) the matrix A A of the symmetric positive definite that solving condition system of equations and error equation group form ^tcholesky decompose lower triangular matrix.

4th step: utilize least square method to search for blur level, wherein variance ratio (Ratio value) needs to be greater than 3.

5th step: utilize least square method to obtain single epoch integer ambiguity static solution coordinate X, Y, Z and its covariance matrix of monitoring point.

6th step: utilize mean gap method to carry out deformation examination to the deformation values between current epoch and last epoch.

The detailed process of inspection is as follows:

1. the empiric variance obtained by two observed reading correction epoch (residual error):

$S_0^2=\frac{{\left(V^T\mathrm{PV}\right)}_I+{\left(V^T\mathrm{PV}\right)}_{\mathrm{II}}}{f}$

Wherein V is observed reading residual vector, P be observed reading power battle array, f be two epoch degree of freedom sum, i.e. f=n ₁-3+n ₂-3, wherein n ₁, n ₂it is the number of two difference observation equation two epoch.

2. by coordinate difference (the i.e. so-called gap) d of two epoch _i(i=1,2,3) component unit power empiric variance

${\stackrel{\‾}{S}}_0^2=\frac{d^T{P}_{d}d}{3}$

D=X _iI-X _i, represent the difference vector of the coordinate of two epoch before and after monitoring point, also can be expressed as:

$d=\left(\begin{array}{cc}-1& 1\end{array}\right)\left(\begin{array}{c}{X}_I\\ X_{\mathrm{II}}\end{array}\right)$

P _dpower battle array for d:

$P_d=Q_d^{-1}$

Q herein _dby $d=\left(\begin{array}{cc}-1& 1\end{array}\right)\left(\begin{array}{c}{X}_I\\ X_{\mathrm{II}}\end{array}\right)$ Propagate law according to covariance to obtain:

$Q_d=\left(\begin{array}{cc}-1& 1\end{array}\right)\left[\begin{array}{cc}{Q}_{X_I{X}_I}& 0\\ 0& Q_{X_{\mathrm{II}}{X}_{\mathrm{II}}}\end{array}\right]\left(\begin{array}{c}-1\\ 1\end{array}\right)=Q_{X_I{X}_I}+Q_{X_{\mathrm{II}}{X}_{\mathrm{II}}}$

3. statistic is built:

$F_{h,f}=\frac{{\stackrel{\‾}{S}}_0^2}{S_0^2}=\frac{d^T{P}_{d}d}{h{S}_0^2}$

This statistic obeys F distribution, and its degree of freedom is respectively degree of freedom 3 He degree of freedom f.

4. displacement judges.Select level of signifiance α=0.01, carry out right-tailed test judgement whether be greater than be about to the F calculated _h,fwith the F found from F distribution table _α(h, f) tantile compares, if

F _h,f＞F _α(h,f)

Then show be greater than namely have:

P{[F _h,f＞F _α(h,f)]}＝α

Illustrate and have displacement to occur between two epoch, otherwise, show not to be subjected to displacement between two epoch.

7th step: be not subjected to displacement between two epoch after the inspection of the 6th step, then the robust least square steady-state solution superposed by the observation equation of two epoch is as the result of current epoch;

If the assay of the 6th step is have displacement to occur between two epoch, then obtain current epoch result by robust sequential adjustment.

Following progress is brought: avoid Detection of Cycle-slip according to the method hinge structure for the treatment of GPS deformation measurement data of the present invention; Adopt Cholesky to decompose and build Ambiguity Search Space, search efficiency is higher; Can judge that the coordinate difference between epoch causes by displacement or by error by deformation examination, improve the reliability of result; Forecast precision is higher.

Although invention has been described with reference to preferred embodiment, without departing from the scope of the invention, various improvement can be carried out to it and parts wherein can be replaced with equivalent.The present invention is not limited to specific embodiment disclosed in literary composition, but comprises all technical schemes fallen in the scope of claim.

Claims (5)

1., for the treatment of a method for GPS deformation measurement data, comprising:

The first step: obtain base station and monitoring station current epoch observation data and broadcast ephemeris data, utilizes carrier phase observable to form two difference observation equation;

Second step: the maximum deformation quantity between two epoch is set;

3rd step: adopt Cholesky to decompose and build Ambiguity Search Space;

4th step: search blur level, and make variance ratio be greater than 3;

5th step: single epoch integer ambiguity static solution coordinate X, Y, the Z and its covariance matrix that obtain monitoring point;

6th step: utilize mean gap method to carry out deformation examination to the deformation values between current epoch and last epoch;

7th step: after the inspection of the 6th step, if be not subjected to displacement between two epoch, then carried out robust least square static state by two difference observation equation superposition of two epoch and resolves the result obtaining current epoch; If the assay of the 6th step is have displacement to occur between two epoch, then obtain the result of current epoch according to robust sequential adjustment.

2. method according to claim 1, is characterized in that, in the third step:

$-\frac{{\mathrm{\δd}}_{max}}{{\mathrm{\λl}}_{11}}\≤{\mathrm{\δ}}_{\▿{\mathrm{\ΔN}}_1}\≤\frac{{\mathrm{\δd}}_{max}}{{\mathrm{\λl}}_{11}}$

$\frac{-\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2}-l_{21}{\mathrm{\λ\δ}}_{\▿{\mathrm{\ΔN}}_1}}{{\mathrm{\λl}}_{22}}\≤{\mathrm{\δ}}_{\▿{\mathrm{\ΔN}}_2}\≤\frac{\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2}-l_{21}{\mathrm{\λ\δ}}_{\▿{\mathrm{\ΔN}}_1}}{{\mathrm{\λl}}_{22}}$

$\frac{-\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2-C^2}-l_{31}{\mathrm{\λ\δ}}_{\▿{\mathrm{\ΔN}}_1}-l_{32}{\mathrm{\λ\δ}}_{\▿{\mathrm{\ΔN}}_2}}{{\mathrm{\λl}}_{33}}\≤{\mathrm{\δ}}_{\▿{\mathrm{\ΔN}}_3}\≤\frac{\sqrt{{{\mathrm{\δd}}_{\max }}^2-B^2-C^2}-l_{31}{\mathrm{\λ\δ}}_{\▿{\mathrm{\ΔN}}_1}-l_{32}{\mathrm{\λ\δ}}_{\▿{\mathrm{\ΔN}}_2}}{{\mathrm{\λl}}_{33}}$

Wherein:

$B=l_{11}\·{\mathrm{\λ\δ}}_{\▿{\mathrm{\ΔN}}_1},C=l_{21}\·{\mathrm{\λ\δ}}_{\▿{\mathrm{\ΔN}}_1}+l_{22}\·{\mathrm{\λ\δ}}_{\▿{\mathrm{\ΔN}}_2},$

δ d _maxfor the maximum deformation quantity of monitoring point, δ _{▽ Δ N}for two poor integer ambiguity medial error, λ is the wavelength of carrier phase,

$L^{-1}=\left[\begin{array}{ccc}{l}_{11}& & \\ 1_{21}& l_{22}& \\ l_{31}& 1_{32}& l_{33}\end{array}\right]$ For the system of equations matrix A A of the symmetric positive definite that solving condition system of equations and error equation group form ^tcholesky decompose lower triangular matrix.

3. method according to claim 1 and 2, is characterized in that, in the 6th step, also comprise following sub-step:

First sub-step: by observed reading correction acquisition two epoch empiric variance:

$S_0^2=\frac{{\left(V^{T}PV\right)}_I+{\left(V^{T}PV\right)}_{II}}{f}$

Wherein f be two epoch degree of freedom sum, i.e. f=n ₁-3+n ₂-3, wherein n ₁, n ₂be the number of two difference observation equations of two epoch, V is observed reading residual vector, and P is observed reading power battle array;

Second sub-step: by the coordinate difference d of two epoch _i(i=1,2,3) obtain weight unit empiric variance ${\stackrel{\‾}{S}}_0^2:$

${\stackrel{\‾}{S}}_0^2\frac{d^T{P}_{d}d}{3}$

Wherein, d=X _iI-X _i, represent the difference vector of monitoring point at the coordinate of front and back two epoch:

$d=\left(\begin{array}{cc}-1& 1\end{array}\right)\left(\begin{array}{c}{X}_I\\ X_{II}\end{array}\right)$

Wherein P _dpower battle array for d:

$P_d=Q_d^{-1}$

Q herein _dby $d=\left(\begin{array}{cc}-1& 1\end{array}\right)\left(\begin{array}{c}{X}_I\\ X_{II}\end{array}\right)$ Propagate law according to covariance to obtain:

$Q_d=\left(\begin{array}{cc}-1& 1\end{array}\right)\left[\begin{array}{cc}{Q}_{X_I{X}_I}& 0\\ 0& Q_{X_{II}{X}_{II}}\end{array}\right]\left(\begin{array}{c}-1\\ 1\end{array}\right)=Q_{X_I{X}_I}+Q_{X_{II}{X}_{II}}$

3rd sub-step: build statistic:

$F_{h,f}=\frac{{\stackrel{\‾}{S}}_0^2}{S_0^2}=\frac{d^T{P}_{d}d}{{\mathrm{hS}}_0^2}$

4th sub-step: select level of signifiance α=0.01, by the F obtained _h,fwith the F found from F distribution table _α(h, f) tantile compares, if

F _h,f＞F _α(h,f)

Then show be greater than namely have:

P{[F _h,f＞F _α(h,f)]}＝α

Thus judge have displacement to occur between two epoch;

If cannot F be drawn _h,f> F _α(h, f), then judge not to be subjected to displacement between two epoch.

4. method according to claim 1 and 2, is characterized in that, in the 4th step, utilizes least square method to search for blur level.

5. method according to claim 1 and 2, is characterized in that, in the 5th step, utilizes least square method to obtain single epoch integer ambiguity static solution coordinate X, Y, Z and its covariance matrix of monitoring point.