CN-121432484-B - Satellite clock difference resolving method for rapid convergence under regional station distribution

CN121432484BCN 121432484 BCN121432484 BCN 121432484BCN-121432484-B

Abstract

The invention relates to a satellite clock difference resolving method capable of fast converging under regional station distribution, which belongs to the field of satellite navigation, and comprises the steps of constructing a normal equation by using known precise orbit and station measurement coordinates and combining pseudo-range and carrier phase without an ionosphere in a regional monitoring network, carrying out covariance forecasting on satellite clock difference, receiver clock difference and troposphere delay in a Kalman filtering stage, carrying out cycle slip detection and covariance forecasting on ambiguity, resetting the forecasting covariance to a minimum value to apply strong constraint when the ambiguity has no cycle slip and converges, obviously compressing reinitialization time, carrying out parameter estimation to obtain an optimal value and a pseudo-range post-test residual, carrying out multi-station multi-epoch residual joint resolving on satellite clock difference constant deviation, merging the satellite clock difference constant deviation into original code deviation to form spread code deviation parameter broadcasting, and enabling a user to realize instantaneous centimeter level positioning by only adding the spread correction in a pseudo-range observation value. The invention realizes high-precision and high-availability real-time precise single-point positioning service under the condition of regional network.

Inventors

Tang Chengpan
ZHOU SHANSHI
ZHAO LIQIAN

Assignees

中国科学院上海天文台

Dates

Publication Date: 20260512
Application Date: 20260104

Claims (8)

1. The satellite clock difference resolving method for fast convergence under regional station distribution is characterized by comprising the following steps: S1, under the condition of regional monitoring stations, constructing a normal equation based on a GNSS satellite precise orbit and accurate coordinates of the monitoring stations by using pseudo-range and carrier phase observation values combined without an ionosphere; S2, carrying out Kalman filtering state prediction, wherein covariance prediction is carried out on satellite clock error, receiver clock error and troposphere delay parameters, cycle slip detection and covariance prediction are carried out on ambiguity parameters, wherein when the ambiguity parameters do not cycle slip and are converged to a suboptimal solution, the prediction covariance is reset to apply strong constraint, and when the ambiguity parameters do not cycle slip and are not converged to the suboptimal solution, the original covariance is maintained for prediction; S3, carrying out parameter estimation based on the normal equation and the state forecast result to obtain an optimal estimated value of each parameter and a pseudo-range post-test residual error; s4, calculating satellite clock error constant deviation of the same satellite in pseudo-range post-test residual errors of a plurality of epochs by utilizing a plurality of regional monitoring stations, and fusing the clock error constant deviation into satellite code deviation parameters to form spread code deviation parameters and broadcasting the spread code deviation parameters; s5, the user side applies correction to the pseudo-range and carrier phase observation values, wherein the correction of the pseudo-range observation values comprises correction of the spread code deviation parameters; The step S2 comprises the following steps: s201, establishing a random walk prediction model, adopting a unit matrix as a state transition matrix, directly inheriting the optimal estimated value of the previous epoch into the predicted value of the current epoch, synchronously transmitting the covariance matrix of the previous epoch to the current epoch, and superposing a process noise matrix to quantify the error accumulation caused by the uncertainty of the model; the diagonal elements of the forecast covariance matrix are respectively the forecast variances of receiver clock error, satellite clock error, zenith troposphere delay and phase ambiguity, and the non-diagonal elements are the cross covariance among the diagonal elements; S202, implementing fixed incremental covariance expansion on three parameters, namely satellite clock error, receiver clock error and zenith troposphere delay; s203, cycle slip detection and convergence judgment are carried out on the phase ambiguity parameters, and the phase ambiguity parameters are obtained by epoch To the direction of Performing variance resetting if cycle slip occurs; If no cycle slip occurs, the covariance forecast of the ambiguity parameter is: ; Is the phase ambiguity variance; the covariance is predicted for the phase ambiguity, For the first threshold, selecting an upper variance limit that has been close to the full-week search region, Selecting a variance lower limit slightly larger than measurement noise for a second empirical threshold; S204, integrating the sub-block covariance elements processed in the S202 and the S203 into a unified matrix according to parameter sequence bits, generating a complete forecast covariance matrix, and using the complete forecast covariance matrix as input of parameter estimation of the next link to realize rapid convergence and high-precision maintenance of clock error calculation in an area monitoring network environment.
2. The method for rapidly converging satellite clock correction under regional distribution according to claim 1, In S1, the normal equation construction process comprises the following steps: S101, taking accurate station coordinates of a GNSS satellite accurate orbit and a ground monitoring station receiver as input, and calculating pseudo-range and phase data residual errors of the regional monitoring receiver to the satellite; S102, under a least square frame, constructing an observation equation by designing a coefficient matrix and a constant vector, establishing a parameter system to be estimated comprising satellite clock error, receiver clock error, ambiguity parameters and troposphere parameters, constructing a normal equation matrix and a normal equation right matrix according to the observation equation, and constructing a fast convergence and clock error deviation estimation normal equation serving the regional network.
3. The method for rapidly converging satellite clock correction under regional distribution according to claim 2, S101 includes: 1) Establishing a current epoch Ionosphere-free combination equations of the regional ground monitoring receiver of (1) for GNSS satellite pseudorange and carrier phase observations: ; In the formula, 、 The GNSS satellites and the ground monitoring receiver respectively, And Observing the GNSS satellite pseudo range and the carrier phase of the regional ground monitoring receiver; For the geometrical distance between the GNSS satellite and the ground monitoring receiver, the precise station coordinates of the GNSS satellite precise orbit and the ground monitoring station receiver are used as input to calculate, And The ground monitoring receiver clock difference and the GNSS satellite clock difference, For the speed of light in vacuum, As a projection function of the tropospheric delay, To calculate zenith directional tropospheric delay using empirical models, Delaying the residual zenith direction troposphere; For the satellite wavelengths, In order to provide a phase ambiguity, And Pseudo-range measurement errors and phase measurement errors are respectively; 2) Calculating pseudo-range and phase data residual errors of the regional monitoring receiver to satellites according to the ionosphere-free combination equation; ; Wherein, the And The pseudo-range ionosphere free combined residual and the phase ionosphere free combined residual are represented respectively.
4. The method for rapidly converging satellite clock correction under regional distribution according to claim 2, In S102, the method includes: 1) Constructing an observation equation; ; Wherein, the Is that Time residual vector , In order to design the coefficient matrix, As a result of the state vector to be estimated, Is a constant vector; 2) Constructing a normal equation matrix and a normal equation right matrix according to the observation equation; Wherein the normal equation matrix The method comprises the following steps: ; french right matrix The method comprises the following steps: ; Wherein, the Is an observation value weight matrix; 3) The simultaneous normal equation matrix and the normal equation right matrix form a normal equation, and the optimal parameter estimation is obtained by solving ; The normal equation is: 。
5. The method for rapidly converging satellite clock correction under regional distribution according to claim 1, In S202, the fixed incremental covariance expansion is implemented on three parameters of satellite clock error, receiver clock error and zenith troposphere delay, comprising: ; Wherein, the The variance is forecasted for the current epoch receiver clock differential, Receiver clock difference for the last epoch; the variance is forecasted for the current epoch satellite clock difference, Satellite clock difference for the last epoch; The variance is predicted for zenith tropospheric delay for the current epoch, For zenith tropospheric delay of the last epoch, In order to provide the time difference between the epochs, For the noise intensity coefficient of the clock-difference random walk process, Is the noise intensity coefficient of the troposphere random walk process.
6. The method for rapidly converging satellite clock difference solution under regional distribution according to claim 5, wherein in S3, parameter estimation is performed to obtain an optimal solution of parameter estimation: Wherein, the In the form of a matrix of normal equations, For the right matrix of the normal equation, In the form of a covariance matrix, Obtaining the optimal estimation of the parameters; pseudo-range post-test residual The calculation method is as follows: Wherein, the For the raw pseudorange ionosphere-free combined observations, To calculate the geometric distance using the a priori precision orbit and the station precision coordinates, A projection function that is a tropospheric delay; For residual zenith directional tropospheric delay, For the satellite clock-difference, Is the receiver clock difference.
7. The method for rapidly converging satellite clock correction under regional distribution according to claim 6, S4, spreading code deviation parameter The method comprises the following steps: Wherein, the The satellite true code deviation of the hardware delay is; Clock difference constant deviation calculated for the server; ; Wherein, the To participate in satellites The total number of ground monitoring stations observed, For each station pair satellite In a continuous manner The number of samples of valid pseudorange residuals within an epoch.
8. The method for rapidly converging satellite clock correction under regional distribution according to claim 7, S5, correcting the precise satellite orbit, the group delay parameter and the precise clock error parameter of the pseudo-range observation value; ; Wherein, the To correct the pseudo-range; The original pseudo-range observed value; Is the track correction; Is a precise satellite clock error; is a group delay parameter; correcting parameters of a precise satellite orbit and a precise clock error for the phase observation value; ; Wherein, the Is the corrected phase; is the original phase observation.

Description

Satellite clock difference resolving method for rapid convergence under regional station distribution Technical Field The invention relates to the fields of satellite navigation, space measurement and control, in particular to a satellite clock difference resolving method for rapid convergence under regional station distribution. Background The satellite navigation system is a national key space-time infrastructure, can provide continuous and real-time accurate position and time information in the global scope, and is deeply integrated into a plurality of fields such as automatic driving, accurate agriculture, military operations, mass consumption and the like. With the rapid increase of the requirements of emerging applications such as intelligent transportation, unmanned systems and the like for Real-time high-precision positioning, real-time precise single point positioning (Real-TIME PRECISE Point Positioning, RT-PPP) services capable of providing dynamic decimeter-level and static centimeter-level positioning precision have become the core direction of the construction and application of global satellite navigation systems. Real-time precise satellite clock skew is one of the precondition parameters of RT-PPP. The satellite clock error presents remarkable random walk characteristics under the influence of factors such as high-frequency noise, temperature transient, orbit dynamics errors and the like of a satellite-borne atomic clock, generally adopts a recursive estimation algorithm such as Kalman filtering and the like, and depends on a monitoring network which is uniformly distributed worldwide to carry out continuous tracking and real-time estimation. However, global site distribution requires a large amount of overseas site resources, construction and maintenance costs are extremely high, and PPP service operators in most countries and regions cannot autonomously control global observation data and can only rely on posterior or delay products provided by a few international institutions, so that service autonomy, instantaneity and safety are insufficient. In order to get rid of the dependence on the global monitoring network, the regional distribution (Regional Monitoring Network) mode becomes a research hotspot in recent years. The mode is only provided with a limited number of monitoring stations in the home country or the periphery, and satellite clock error real-time estimation is realized through local observation arc segments. However, the coverage of the regional network is limited, and a single satellite can only be continuously observed for 4-6 hours each day, and the rest of the time is in an 'exit-entry' intermittent state, so that the following outstanding problems are caused: the filtering is frequently initialized, namely, the clock filter needs to be initialized again after the satellite enters the field of view of the regional network again each time, the traditional random walk model needs 3-5 hours to be converged to the centimeter-level precision, and the satellite usually goes out before convergence, so that the effective service period is almost zero. Convergence contradiction with usability, namely increasing process noise to shorten convergence time, obviously reducing clock stability although parameter adjustment can be accelerated, and slow convergence otherwise, failing to meet the rigid requirement of real-time PPP on instantaneous availability. The constant deviation is indistinguishable, namely, the observed geometric intensity of the regional network is weak, an absorption-coupling phenomenon occurs between clock error and hardware Delay (DCB), and systematic constant deviation exists in the estimated clock error. If the deviation can not be calibrated and broadcast in real time, the deviation can be directly introduced into a user positioning equation, so that systematic errors ranging from a few decimeters to a few meters are caused. In the existing improvement thought aiming at the difficult problem, the high-frequency epoch differential speed is increased, the global station network is used for redundancy, or real-time flow polynomial forecast is relied on for refinement, but filtering repeated initialization, convergence time consumption and constant deviation caused by short arc sections of regional monitoring networks and frequent satellite entry and exit are generally ignored, so that real-time calibration and broadcasting of the basic pain point are difficult, and the high-precision real-time clock difference service in the regional environment still cannot be considered in terms of precision, usability and autonomy. In summary, the prior art can better operate under the condition of the global network, but three bottlenecks of long convergence time, short available period, difficult real-time calibration of constant deviation exist in the regional station distribution mode generally, and the precision, availability and autonomy of the real-time PPP service ca