CN114895336A

CN114895336A - Forecasting method for observed value of reference station in GNSS real-time dynamic positioning system

Info

Publication number: CN114895336A
Application number: CN202210813342.XA
Authority: CN
Inventors: 陈国�; 瞿子扬; 姜益昊
Original assignee: Kepler Satellite Technology Wuhan Co ltd
Current assignee: Kepler Satellite Technology Wuhan Co ltd
Priority date: 2022-07-12
Filing date: 2022-07-12
Publication date: 2022-08-12
Anticipated expiration: 2042-07-12
Also published as: CN114895336B

Abstract

The invention provides a forecasting method of observed values of a reference station in a GNSS real-time dynamic positioning system, which belongs to the technical field of global satellite navigation systems, and is characterized in that historical observation data of a plurality of GNSS satellites provided by the reference station in the latest 2T time duration are collected, the observation data are preprocessed, abnormal values in the observation data are identified and removed, and phase cycle slip is detected and marked in a segmented manner; forecasting an observed value in the T duration according to the forecasted troposphere delay rate and ionosphere delay rate and positions and rates of a satellite and a reference station; and establishing a compensation model of the observation value prediction error by using the difference between the predicted data and the actually measured observation data, obtaining the compensation error through the compensation model, and applying the compensation error to the observation value prediction of the future T duration. The invention provides a method for forecasting with high precision by using historical observation data of a single reference station, and provides guarantee for RTK positioning of a rover user by using non-delay reference station data.

Description

Forecasting method for observed value of reference station in GNSS real-time dynamic positioning system

Technical Field

The invention relates to the technical field of global satellite navigation systems, in particular to a forecasting method of an observed value of a reference station in a GNSS real-time dynamic positioning system.

Background

The position of any point location can be determined by utilizing a GNSS absolute point location technology only depending on single receiving equipment, but the positioning accuracy is limited by ephemeris error and pseudo range noise, so that the positioning accuracy is not high easily, and the accuracy of the positioning method is in a meter level. In order to improve the positioning accuracy, a Real-time Kinematic (RTK) positioning method is generally adopted, and since a GNSS observation error and an ephemeris error have correlation in a certain spatial range, the RTK positioning method obtains a centimeter-level positioning technical means by using the correlation and a double-difference method.

However, a user using the real-time dynamic positioning method needs to receive not only the ranging signal of the satellite but also the reference station observation information broadcasted from the service system by using the communication device, and the reference station observation data is inevitably affected by communication delay, receiver failure, and the like.

Aiming at the problems, generally, an asynchronous RTK means is adopted, namely outdated reference station observation data and current latest data of a rover station are directly utilized to form a baseline, and when the data delay is 30s, the positioning error of the method in the elevation direction can reach 30cm, the requirement of high-precision positioning still cannot be met, and the method is mainly used for post-processing.

The forecasting of the observed value of the reference station in the GNSS real-time dynamic positioning system mainly has the following two technical difficulties or problems:

firstly, how to comprehensively utilize the change rate of the carrier phase and the Doppler observation value among epochs to separate out the influence of orbit error, satellite clock error, ionosphere error rate and troposphere error rate on observation value prediction; and aiming at various time variation characteristics of error rate influencing the forecasting precision of the observed value, a corresponding error model is designed.

And secondly, evaluating the error of the historical forecast observation value in real time, and how to further optimize the forecast precision of the current observation value by utilizing the forecast error.

The two technical difficulties directly influence the forecasting precision of the RTK reference station observation value, further influence the use experience of the rover user on the RTK service, and are problems which need to be solved in the current RTK continuous stable high-precision service system.

Therefore, for a user with a very high real-time requirement, how to generate a high-precision observation data prediction value in a short period of several minutes by using the historical data of the reference station is a problem to be solved urgently by using the real-time dynamic positioning method to achieve the user positioning continuity and high-precision position information.

Disclosure of Invention

In view of the above, the present invention provides a method for predicting an observed value of a reference station in a GNSS real-time dynamic positioning system, and therefore, the present invention performs classification analysis on predictability and delay sensitivity of error sources, such as a satellite ephemeris error, an ionosphere difference rate, a troposphere difference rate, and the like, which affect the prediction of the GNSS observed value, and provides a method for performing high-precision prediction by using historical observation data of a single reference station, so as to provide a guarantee for a rover user to perform RTK positioning by using data of a non-delay reference station.

The technical scheme of the invention is realized as follows: a forecasting method of an observation value of a reference station in a GNSS real-time dynamic positioning system comprises the steps of collecting historical observation data of a plurality of GNSS satellites provided by the reference station within the latest 2T time, preprocessing the observation data, identifying abnormal values in the observation data, removing the abnormal values, detecting phase cycle slip and marking in a segmented mode;

the method is characterized in that: the observation data comprises pseudo range, phase, Doppler and signal to noise ratio and accurate reference station coordinates;

determining a receiver clock error rate, a troposphere delay rate and an ionosphere delay rate;

modeling the residual errors of the troposphere delay rate and the ionosphere delay rate respectively, calculating the delay rate of the T duration of a forecast part, calculating the prior correction quantity of the troposphere delay rate and the ionosphere delay rate of the satellite-earth sight direction of the forecast part, and calculating the troposphere delay rate and the ionosphere delay rate of the forecast part by utilizing the prior correction quantity and the delay rate;

forecasting an observed value in the T duration according to the forecasted troposphere delay rate and ionosphere delay rate and positions and rates of a satellite and a reference station;

and establishing a compensation model of the observation value prediction error by using the difference between the predicted data and the actually measured observation data, obtaining the compensation error through the compensation model, and applying the compensation error to the observation value prediction of the future T duration.

In addition to the above technical means, preferably, T is 10 seconds or more and 10 minutes or less.

On the basis of the technical scheme, preferably, a speed measurement observation equation is established by utilizing historical observation data of a plurality of satellites of the GNSS and the limit epoch difference observation value.

On the basis of the above technical solution, preferably, in the observation data within the T duration, the value of i is from-299 to 0, and the troposphere delay rate and the ionosphere delay rate time series

And

eliminating the constant offset and linear components in the troposphere delay rate and the ionosphere delay rate by utilizing a linear model, obtaining zero-mean-value-change troposphere delay rate residual errors and ionosphere delay rate residual errors, and respectively recording the zero-mean-value-change troposphere delay rate residual errors and the ionosphere delay rate residual errors as

And

(ii) a And then, respectively adopting a first-order autoregressive model and a second-order autoregressive model to establish a fitting model for the convective stratum delay and the ionospheric delay rate residual.

On the basis of the technical scheme, preferably, the receiver clock difference rate is ignored, and the reference station phase observation value is forecasted epoch by epoch according to the calculated satellite rate, the satellite clock rate, the ionosphere delay rate and the troposphere delay rate, wherein the observation forecasting time is T.

On the basis of the above technical solution, preferably, on the premise of not interrupting the actual observation data, the prediction error is calculated by using the actual observation value and the prediction observation value.

On the basis of the above technical solution, preferably, when the actual observation data and the predicted observation data are two different reference stations, the two reference stations are subtracted at the same time to obtain an inter-station difference, at this time, the influence of the receiver clock error rate is eliminated, but the residual errors are still influenced by the ephemeris error and the atmospheric error, and the like, the residual errors within the time length T are fitted by using a quadratic polynomial model, the observation data prediction error is fitted by using a formula, and the prediction error of the next prediction arc segment is predicted by using the fitting model.

On the basis of the above technical solution, preferably, the above process is repeated until the ranging data of all satellites observable at the reference station are completely forecasted, the forecast data is output, and the forecast data of the next T duration is started when the next forecast time comes.

Compared with the prior art, the forecasting method of the observed value of the reference station in the GNSS real-time dynamic positioning system has the following beneficial effects:

(1) the method comprises the steps of obtaining high-precision inter-epoch observed value speed by utilizing a phase and Doppler observed value combination model, further obtaining estimated values of various error variations, establishing a targeted autoregressive fitting model by utilizing time characteristics of troposphere and ionosphere error variations, modeling prediction errors of historical observed values, further improving the prediction precision of a current prediction arc section, and finally solving short-term high-precision prediction of reference station observed data in an RTK service system;

(2) the RTK reference station continuously operates and stores historical observation data within the time length of not less than 2T, an observation value forecasting model can be effectively established, and high-precision short-term forecasting of observation data of the reference station is realized;

(3) by means of the GNSS speed measurement principle, various error rates are respectively measured by comprehensively utilizing phase and Doppler observed values, and a basis is provided for forecasting the observed value of the reference station;

(4) respectively establishing fitting models of ionospheric difference rate and tropospheric error rate by adopting an autoregressive model, and realizing the prediction of irregular change parts in the tropospheric difference delay difference rate and the ionospheric difference delay rate;

and the prediction error modeling is carried out by utilizing the measured data and the prediction value, so that the prediction precision of the observed value is further improved.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a flowchart illustrating a method for forecasting observed values of a reference station in a GNSS real-time kinematic positioning system according to the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.

As shown in fig. 1, a method for forecasting an observed value of a reference station in a GNSS real-time dynamic positioning system includes collecting historical observation data of multiple satellites of the GNSS provided by the reference station within a latest 2T duration, preprocessing the observation data, identifying and eliminating abnormal values in the observation data, and detecting and segmentally marking a phase cycle slip.

In a preferred embodiment, T is 10 seconds or more and 5 minutes or less, since the transmission of the observation data by the satellite requires time, and T is a minimum of 10 seconds for more accurate prediction of the observation value, and T is not more than 10 minutes due to an accumulated error in the transmission process, and it is found by experiment that T is preferably 5 minutes, and therefore T is defaulted to 5 minutes in the present embodiment.

The observation data comprises pseudo range, phase, Doppler and signal to noise ratio and accurate reference station coordinates; determining a receiver clock error rate, a troposphere delay rate and an ionosphere delay rate; measuring speed by using phase and Doppler data, determining receiver clock error rate, troposphere and ionosphere delay rate, calculating fitting residual error of troposphere and ionosphere delay rate by using a second-order linear fitting model, respectively using first-order and second-order autoregressive models,

modeling a troposphere delay rate and an ionosphere delay rate residual error respectively, calculating the delay rate of the T duration of a forecast part, calculating the troposphere delay rate and the ionosphere delay rate prior correction quantity of the satellite-earth sight line direction of the forecast part by utilizing a publicly-obtained troposphere experience correction model and a quick ionosphere grid product, and calculating the troposphere delay rate and the ionosphere delay rate of the forecast part by utilizing the prior correction quantity and the delay rate; forecasting an observation value within T duration according to the forecasted troposphere delay rate and ionosphere delay rate as well as positions and rates of a satellite and a reference station; and repeating the steps in the next T duration, establishing a compensation model of the observation value forecasting error by using the difference between the forecasted data and the actually measured observation data, obtaining the compensation error through the compensation model, and applying the compensation error to the observation value forecasting of the future T duration to realize high-precision forecasting of the observation value of the reference station.

The present embodiment assumes that the reference station can observe multiple satellites for a certain GNSS system and provide satellite, phase, doppler and snr observations, which assumption is very easy to satisfy in practice. Because the basic requirement is that the RTK reference station is constructed, the RTK reference station is generally constructed in a field with a wide view and less signal interference, as a preferred embodiment, the data preprocessing is performed on the collected latest historical observation data within a 2T time duration, the position and the speed of a fixed satellite are used as an ultra-fast orbit and a broadcast ephemeris, the clock error of the satellite is a real-time clock error or a broadcast clock error, the observed value of the satellite-ground distance and the theoretical value difference OMC are calculated, the pseudorange gross error is detected and eliminated based on the OMC, the phase observed value is used to form the inter-epoch difference, and the phase cycle slip is detected and marked.

As a preferred embodiment, since the phase observation value is susceptible to cycle slip, which causes phase difference to generate jump, the doppler observation value and the phase epoch difference observation value are comprehensively utilized to establish a velocity measurement observation equation as follows:

（1）

in formula (1):

j is the satellite number;

deriving a Doppler measurement based on the phase epoch difference;

receiving raw Doppler observations directly received by a device for a reference station;

and

position vectors of a reference station and a satellite under a geostationary system respectively;

and

the velocity vectors of the reference station and the satellite, respectively, in the Earth's fixed frame, which are stationary relative to the Earth for the RTK stationary reference station, i.e.

；

And

respectively representing the carrier wavelength of the observed value frequency point and the light speed in vacuum;

and

respectively, reference station receiver clock speed and satellite clock speed;

and

ionospheric delay rate and tropospheric delay rate on the satellite-ground line of sight, respectively, expressed as the product of a projection function and zenith delay rate;

and

respectively, the noise of the derived doppler and the original doppler.

A speed measurement observation equation as shown in a formula (1) is established according to n observation satellites, wherein the positions and the speeds of the satellites are from broadcast ephemeris or real-time orbit clock error products received by a reference station, and estimated values of parameters to be estimated, including ionospheric delay difference rate, tropospheric delay difference rate and receiver clock error rate, are obtained based on the least square principle.

Modeling tropospheric delay difference rate and ionospheric delay difference rate, as a preferred embodiment, determining a time series of tropospheric delay difference rate and ionospheric delay difference rate

And

for the observation data in the time length of T, the value of i ranges from-299 to 0. Firstly, eliminating the constant offset and linear components in the troposphere delay difference rate and the ionosphere delay difference rate by utilizing a linear model, obtaining the troposphere delay difference rate and the ionosphere delay difference rate residual error with zero mean change, and respectively recording the residual errors as

And

(ii) a Then, a first-order autoregressive model and a second-order autoregressive model are respectively adopted to establish a fitting model for the convective stratum delay and the ionospheric delay rate residual error as follows:

and

noise of tropospheric delay rate and ionospheric delay rate residuals respectively;

first order autoregressive coefficients for tropospheric delay rate residuals;

and

second order auto-regressive system for ionospheric delay rate residualAnd (4) counting.

Reference station observation prediction, as a preferred embodiment, the satellite rate, the satellite clock rate, the ionosphere delay difference rate and the troposphere delay difference rate calculated in the above steps are used for performing reference station phase observation prediction epoch by epoch, the prediction duration is T,

(3)

in formula (3):

is composed of

A phase prediction value of the time;

is composed of

A phase prediction value of the time, which is an actual observation value at an initial prediction time (i = 0);

is composed of

Predicting an ionospheric delay rate at a time;

is composed of

A tropospheric delay rate prediction value at a time;

this value is 1 second for the reference station for a time interval of two epochs before and after.

Because the influence of the clock difference rate of the receiver on all satellites is the same, the technical scheme ignores the influence of the clock difference rate of the receiver, and the corresponding forecast error of the observation value can be absorbed by the clock difference rate of the user receiver, so that the formula (3)

The direct value is 0.

Observation prediction error compensation as a preferred embodiment, the prediction error is calculated using the measured value and the predicted value without interruption of the actual observation data of the reference station.

（4）

In formula (4):

to represent

The difference between the forecast errors of the time reference satellite ref and the satellite j;

and represents

A single difference between the actual observed values of the time reference satellite ref and the satellite j;

to represent

Single difference between time reference satellite ref and satellite j predicted value。

If the actual observation value and the forecast observation value are regarded as two different observation stations, the two observation stations are subtracted at the same time to obtain the inter-station difference, at the moment, the influence of the clock error rate of the receiver is eliminated, but the difference is still influenced by residual errors such as ephemeris error and atmospheric error, and the residual errors can be well fitted by a quadratic polynomial model in the time length of T, so that the forecast error of the observation value is fitted by using the following formula, and the forecast error of the next forecast arc segment is forecasted by using the fitting model.

（5）

In formula (5):

，

and

fitting coefficients to be estimated;

is composed of

The difference between the time and the initial forecast time, i.e.

。

Estimating model coefficients using the prediction error carry-over (5) calculated by equation (4)

，

And

and forecasting the forecasting error of the next forecasting arc segment by using the quadratic model, and using the forecasting error as a priori correction value of the next forecasting arc segment.

And repeating the steps until the distance measurement values of all the satellites observable by the reference station are completely forecasted, and outputting forecast data. And waiting for the next forecast time to come, and starting forecast data in the next T duration.

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, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims

1. A forecasting method of an observation value of a reference station in a GNSS real-time dynamic positioning system comprises the steps of collecting historical observation data of a plurality of GNSS satellites provided by the reference station within the latest 2T time, preprocessing the observation data, identifying abnormal values in the observation data, removing the abnormal values, detecting phase cycle slip and marking in a segmented mode;

2. The method as claimed in claim 1, wherein the method for forecasting the observed values of the reference station in the GNSS real-time dynamic positioning system comprises: t is not less than 10 seconds and not more than 10 minutes.

3. A method as claimed in claim 1 or 2, wherein the method for forecasting the observed value of the reference station in the GNSS real-time dynamic positioning system comprises: the method comprises the following steps of establishing a speed measurement observation equation by utilizing historical observation data of a plurality of satellites of the GNSS and a limit epoch difference observation value as follows:

j is the satellite number;

deriving a Doppler measurement based on the phase epoch difference;

and

and

；

And

respectively the carrier wavelength of the observed value frequency point and the medium speed in vacuum;

and

respectively, reference station receiver clock speed and satellite clock speed;

and

and

respectively, the noise of the derived doppler and the original doppler.

4. The method as claimed in claim 3, wherein the method for forecasting the observed value of the reference station in the GNSS real-time dynamic positioning system comprises: in the observation data in the T duration, i takes a value from-299 to 0, and the troposphere delay rate and the ionosphere delay rate time sequence

And

And

and

for delaying troposphereFirst order autoregressive coefficients of rate residuals;

and

the second order autoregressive coefficients are the ionospheric delay rate residuals.

5. The method as claimed in claim 4, wherein the method for forecasting the observed value of the reference station in the GNSS real-time dynamic positioning system comprises: and ignoring the clock difference rate of the receiver, and forecasting the phase observed value of the reference station by epochs according to the satellite rate, the satellite clock rate, the ionosphere delay rate and the troposphere delay rate which are obtained by calculation, wherein the observation and forecasting duration is T.

6. The method as claimed in claim 5, wherein the method for forecasting the observed value of the reference station in the GNSS real-time dynamic positioning system comprises: and on the premise of not interrupting the actual observation data, calculating a forecast error by using the actual observation value and the forecast observation value.

7. The method as claimed in claim 6, wherein the method for forecasting the observed value of the reference station in the GNSS real-time dynamic positioning system comprises: when the actual observation data and the forecast observation data are two different reference stations, the actual observation data and the forecast observation data are subtracted at the same time to obtain the inter-station difference, the influence of the clock difference rate of the receiver is eliminated at the moment, but the actual observation data and the forecast observation data are still influenced by residual errors such as ephemeris error and atmospheric error, the residual errors in the T duration are fitted by a quadratic polynomial model, the forecast errors of the observation data are fitted by a formula, and the forecast errors of the next forecast arc segment are forecasted by the fitting model.

8. The method as claimed in claim 7, wherein the method for forecasting the observed value of the reference station in the GNSS real-time dynamic positioning system comprises: and repeating the process until the distance measurement data of all satellites observable by the reference station are completely forecasted, outputting forecast data, and starting forecast data of the next T duration when the next forecast time comes.