Satellite formation system real-time fusion positioning method

Abstract

The application discloses a real-time fusion positioning method of a satellite formation system, and belongs to the technical field of space-based navigation and positioning and satellite formation. The application solves the problem of low positioning precision of the existing satellite formation positioning method based on pseudo-range observation values. The main technical scheme of the method is as follows: step one, the baseline observation values are preprocessed by using Kalman filtering to determine the integer ambiguity, and then the integer ambiguity is substituted into the baseline observation equation to obtain the preprocessed baseline observation values; step two, the pseudo-range observation values are preprocessed by using nonlinear Kalman filtering to obtain the initial estimation result of the user satellite position; and step three, the final estimation result of the user satellite position is obtained according to the preprocessed baseline observation values in step one and the initial estimation result in step two. The method can be applied to the real-time positioning task of the satellite formation system.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of space-based navigation and positioning and satellite formation, and particularly relates to a real-time fusion positioning method for a satellite formation system. BACKGROUND

[0002] The satellite formation system is a new satellite space operation mode emerging in the mid and late 1990s with the continuous development of science and technology. A plurality of satellites fly in close proximity to a reference satellite, form a stable relative configuration among each other, and cooperatively constitute a "big satellite" in a distributed manner, but are superior to the traditional single-satellite system in performance. The satellite formation system is increasingly widely applied in the field of spaceflight with its advantages, and gradually becomes a hot direction and development trend in the field of spaceflight.

[0003] Obtaining the position and velocity information of a satellite in space is one of the important guarantees for reliable on-orbit operation of the satellite formation system. The relatively traditional satellite positioning method is to obtain a set of distance observation values (pseudo-range observations) by simultaneously observing a single satellite by a plurality of (at least four) ground observation stations or space-based satellites, and then estimating the position and velocity by means of least squares, Kalman filtering and other estimation methods. However, this positioning method is usually based on pseudo-range observations for positioning, and due to the observation method and hardware limitations, the pseudo-range observation values are low in accuracy, usually in the order of m or dm, so the positioning result is also low in accuracy, and in some occasions with high accuracy requirements, the requirements cannot be met, so it has strong engineering significance to study a high-precision real-time positioning method suitable for the satellite formation system. SUMMARY

[0004] The purpose of the present application is to solve the problem of low positioning accuracy of the existing satellite formation positioning method based on pseudo-range observations, and a real-time fusion positioning method for a satellite formation system is proposed.

[0005] The technical scheme adopted by the present application to solve the above technical problem is:

[0006] A real-time fusion positioning method for a satellite formation system, which specifically comprises the following steps:

[0007] Step one, using Kalman filtering to pre-process the baseline observation values, determining the integer ambiguity, and then substituting the integer ambiguity into the baseline observation equation to obtain the pre-processed baseline observation values;

[0008] Step two, using nonlinear Kalman filtering to pre-process the pseudo-range observation values to obtain the initial estimation result of the user satellite position;

[0009] Step three, obtaining the final estimation result of the user satellite position according to the pre-processed baseline observation values in step one and the initial estimation result in step two.

[0010] The beneficial effects of the present application are:

[0011] (1) The present application adopts nonlinear Kalman filtering and nonlinear least squares algorithm to perform fusion processing of multiple observation data, which is simple and easy to implement, and has smaller calculation amount, and compared with intelligent algorithms such as genetic algorithm and neural network, the present application has smaller calculation amount and is more suitable for the computing power of an actual satellite computer to perform real-time position estimation.

[0012] (2) The fusion positioning algorithm of the present application has good real-time performance, retains the advantages of traditional positioning algorithms that recursively estimate one by one, and can better meet the real-time positioning needs of satellite formation configuration control.

[0013] (3) The fusion positioning algorithm of the present application greatly improves the positioning accuracy of space-based systems, and compared with traditional positioning methods, the baseline observation value is introduced for fusion positioning according to the running characteristics of the satellite formation system, so that the positioning accuracy is improved by up to 50%.

[0014] (4) The fusion positioning algorithm of the present application has strong practicability and adaptability, and can be designed according to the running and observation characteristics of the satellite formation system. In particular, the designed unscented Kalman filter + nonlinear least squares algorithm can be applied to different actual working conditions such as full sampling and under-sampling observation. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 is a simplified diagram of an observation model of a satellite formation system;

[0016] In the diagram, the thin line represents the pseudo-range observation value, and the thick line represents the baseline observation value.

[0017] Figure 2 is a principle diagram of a satellite formation system real-time fusion positioning method of the present application;

[0018] Figure 3 is a flowchart of a satellite formation system real-time fusion positioning method of the present application;

[0019] Figure 4 is a fusion positioning algorithm simulation result diagram under full sampling;

[0020] In the diagram, the point line is the traditional Kalman filter positioning result error, and the asterisk line is the fusion positioning result error.

[0021] Figure 5 is a fusion positioning algorithm simulation result diagram under under-sampling;

[0022] In the diagram, the point line is the traditional Kalman filter positioning result error, and the asterisk line is the fusion positioning result error. DETAILED DESCRIPTION

[0023] DETAILED DESCRIPTION Figure 2 and Figure 3 The embodiment is described. The real-time fusion positioning method of the satellite formation system described in the embodiment specifically comprises the following steps:

[0024] Step one, using Kalman filter to pre-process the baseline observation value, determine the integer ambiguity, and then substitute the integer ambiguity into the baseline observation equation to obtain the pre-processed baseline observation value;

[0025] Step two, using nonlinear Kalman filter to pre-process the pseudorange observation value to obtain the initial estimation result of the user satellite position of the satellite formation system;

[0026] Step three, according to the pre-processed baseline observation value in step one and the initial estimation result in step two, obtain the final estimation result of the user satellite position.

[0027] The embodiment adopts the method of Kalman filter + nonlinear least squares to combine low-precision pseudorange observation values and high-precision baseline observation values, which ensures the real-time positioning while improving the positioning accuracy, lays a foundation for the reliable operation of the satellite formation system, and meets the needs of actual engineering.

[0028] Specific implementation method two: the difference between the embodiment and the specific implementation method one is that in step one, Kalman filter is used to pre-process the baseline observation value to determine the integer ambiguity; the specific process is as follows:

[0029]

[0030] In the formula: represents the i-th baseline observation value, unit: rad; ρ i represents the i-th pseudorange observation value, unit: m; N i represents the i-th integer ambiguity, i=1, 2, …, L, L represents the number of integer ambiguities; λ1 represents the wavelength under the carrier phase observation frequency f1 (f1=1575.42MHz), unit: m;

[0031] Define the state vector X of the baseline observation model:

[0032] X=[ρ1 … ρ L N1 … N L ] T

[0033] Define the observation variable Y of the baseline observation model:

[0034]

[0035] The integer ambiguity N1, N2, …, N is obtained by solving the state vector X of the baseline observation model through the Kalman filtering formula. L :

[0036]

[0037] P(k+1|k) = FP(k)F T +ΓQΓ T

[0038] K(k+1) = P(k+1|k)H T [HP(k+1|k)H T +R] -1

[0039]

[0040] P(k+1) = [I-K(k+1)H]P(k+1|k)

[0041] wherein, X(k) is the state vector estimation of the baseline observation model at time k, F is the state transition matrix, the upper index T represents the transpose of the matrix, X(k|k-1) is the one-step prediction of the state vector of the baseline observation model at time k, Γ is the input noise driving matrix of the baseline observation model, P(k) is the state quantity covariance estimation of the baseline observation model at time k, Q is the system model (constant speed model) noise matrix, H is the observation matrix of the baseline observation model, P(k+1|k) is the one-step prediction of the state quantity covariance of the baseline observation model at time k, R is the baseline observation model noise matrix, the upper index -1 represents the inverse of the matrix, X(k+1) is the state vector estimation of the baseline observation model at time k+1, K(k+1) is the gain of the baseline observation model at time k+1, Y(k+1) is the observation variable of the baseline observation model at time k+1, P(k+1) is the state quantity covariance estimation of the baseline observation model at time k+1, I is the unit matrix.

[0042] State transition matrix F:

[0043]

[0044] wherein, represents the combination number;

[0045] Observation matrix:

[0046]

[0047] System model noise matrix Q:

[0048]

[0049] Observation model noise matrix R:

[0050]

[0051] Initial covariance matrix P:

[0052]

[0053] Other steps and parameters are the same as in the first embodiment.

[0054] For a satellite formation system, the satellites in the system are called member satellites, and the satellites outside the system that observe the member satellites are called navigation satellites. The L pseudorange observations participating in the operation of the embodiment are mutual pseudorange observations between the member satellites.

[0055] The third embodiment is different from the first or second embodiment in that the baseline observation is obtained by carrier phase measurement.

[0056] Other steps and parameters are the same as in the first or second embodiment.

[0057] The fourth embodiment is different from any one of the first to third embodiments in that the integer ambiguity is substituted into the baseline observation equation to obtain the preprocessed baseline observation, and the specific process is as follows:

[0058]

[0059] In the formula, d i represents the preprocessed i-th baseline observation, and the unit is m.

[0060] Other steps and parameters are the same as in any one of the first to third embodiments.

[0061] The fifth embodiment is different from any one of the first to fourth embodiments in that the specific process of step two is as follows:

[0062] The state vector X' of the pseudorange observation model is defined as follows:

[0063]

[0064] In the formula, x, y, and z represent the three-dimensional coordinates of the user satellite, and the unit is m; represents the three-dimensional velocity of the user satellite, and the unit is m / s; t represents the clock difference between the user satellite (referring to a satellite flying close to a reference satellite) and the satellite (referring to the reference satellite), and the unit is s; represents the clock difference change rate between the user satellite and the satellite;

[0065] The three-dimensional coordinates of the user star are calculated by extending the Kalman filter (EKF) to obtain an initial estimation of the user star position;

[0066]

[0067] P'(k+1|k)=Φ(k)P'(k)Φ T (k)+Γ'Q'Γ ′T

[0068] K'(k+1)=P'(k+1|k)H ′T (k+1)[H'(k+1|k)P'(k+1|k)H'(k+1|k)+R'] T (k+1|k)+R'] -1

[0069]

[0070]

[0071] P'(k+1)=[I-K'(k+1)H'(k+1|k)]P'(k+1|k)

[0072] wherein, is the final position estimation result at k (which can be calculated by step three and then brought back, and for the first time when positioning is required, the 0 vector can be taken), f[·] is a linear system model, is a one-step prediction of the state quantity of the pseudo-range observation model at k, Φ(k) is a state transition matrix at k, the upper index T represents the transpose of the matrix, P'(k) is a state quantity covariance estimation of the pseudo-range observation model at k, Q' is an input noise covariance matrix of the pseudo-range observation model, Γ' is an input noise driving matrix of the pseudo-range observation model, P'(k+1|k) is a one-step prediction of the state quantity covariance of the pseudo-range observation model at k, H'(k+1) is a Jacobi matrix of the pseudo-range observation equation at k, R' is an observation noise covariance matrix of the pseudo-range observation model, the upper index -1 represents the inverse of the matrix, h[·] is a pseudo-range observation equation, is a state vector estimation of the pseudo-range observation model at k+1, K'(k+1) is a gain of the pseudo-range observation model at k+1, P'(k+1) is a state quantity covariance estimation of the pseudo-range observation model at k+1, I is an identity matrix, and Y'(k+1) is an observation quantity of the pseudo-range observation model at k+1.

[0073] Assuming that there are M member satellites in the system, full sampling refers to that all baseline data between the member satellites can be obtained at the observation time, i.e. M(M-1) / 2 baseline observation data, in which case the observation data is complete and positioning is easy, and the present application solves the problem of low positioning accuracy of the existing method by using extended Kalman filtering + nonlinear least squares.

[0074] The other steps and parameters are the same as one of the first to fourth embodiments.

[0075] State transition matrix:

[0076]

[0077] In the formula:

[0078] dt - time interval between adjacent two observation epochs;

[0079] q - number of observation equations at the same time;

[0080] Input noise matrix:

[0081]

[0082] In the formula:

[0083] - three-dimensional acceleration of the user satellite, unit: m / s 2 , is a Gaussian white noise.

[0084] Input noise driving matrix Γ' of the pseudo-range observation model:

[0085]

[0086] The covariance matrix Q' of the input noise W is:

[0087]

[0088] The observation noise covariance matrix R' of the pseudo-range observation model is a q×q diagonal matrix:

[0089]

[0090] Initial covariance matrix P0:

[0091]

[0092] The sixth embodiment is different from the first to fifth embodiments in that the Jacobi matrix of the pseudo-range observation equation is:

[0093]

[0094] wherein, x i , y i , z i represents the ith pseudo-range observation value of the user star position, r i represents the distance between (x i , y i , z i ) and (x, y, z), and L+q represents the total number of pseudo-range observation values.

[0095] The other steps and parameters are the same as one of the first to fifth embodiments.

[0096] The number of pseudo-range observation values participating in the operation of the embodiment is q, and the q pseudo-range observation values participating in the operation of the embodiment are all the pseudo-range observation values between the navigation satellite and the member satellite.

[0097] The seventh embodiment is different from one of the first to sixth embodiments in that the specific process of the second step is as follows:

[0098] The state vector X' of the pseudo-range observation model is defined as follows:

[0099]

[0100] In the formula, x, y, and z represent the three-dimensional coordinates of the user star, with the unit of m; represents the three-dimensional velocity of the user star, with the unit of m / s; t represents the clock difference between the user star and the satellite, with the unit of s; represents the clock difference change rate between the user star and the satellite;

[0101] The augmented state vector X is defined as follows: wherein, n represents a given parameter, 2n+1 is the number of sampling points, P' represents the state quantity covariance matrix, λ=α 2 (n+k)-n, k is a constant, and k satisfies that (n+λ)P' is positive definite, and α is a constant, which can be adjusted according to the filtering effect;

[0102] The unscented Kalman filter (UKF) is used to solve X Sigma to obtain the three-dimensional coordinates of the user star, that is, the initial estimation result of the user star position;

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111] P'(k+1) = P'(k+1|k) - K'(k+1)P Y (k+1|k)K' T (k+1)

[0112] wherein, is the state vector estimate of the pseudo-range observation model at time k, is the estimate of X Sigma at time k,

[0113] is the one-step prediction of the augmented state vector at time k, f[·] is the linear system model, is the one-step prediction of the state vector of the pseudo-range observation model at time k, is the one-step prediction of the augmented state vector at time k for the i-th sampling point, ω i′ represents the calculation weight of the i'-th sampling point, i' = 0, 1,..., 2n, errX is an intermediate variable, the upper index T represents the transpose, Q' is the covariance matrix of the input noise of the pseudo-range observation model, Γ' is the input noise driving matrix of the pseudo-range observation model, h[·] is the pseudo-range observation equation, Y Sigma (k+1) represents the observation value prediction point set obtained by K'(k+1) is the gain of the pseudo-range observation model at time k+1, Y'(k+1) represents the observation value at time k+1, Y p re (k+1) is the mean of the prediction point set Y Sigma (k+1), P Y (k+1) is the covariance matrix of Y'(k+1), P XY (k+1) is the cross-covariance matrix of X'(k) and Y'(k+1), P Y (k+1|k) is the one-step prediction of P Y (k+1).

[0114] The other steps and parameters are the same as one of the first to sixth embodiments.

[0115] ​In some cases, such as obstruction shielding, observation device measurement range limitation, etc., the observation value will be lost, resulting in only part of the star and star baseline measurement data, i.e. the baseline observation data is less than M(M-1) / 2, this working condition is called undersampling. Under-sampling often occurs during the operation of the satellite formation system, and data loss increases the difficulty of positioning. The embodiment based on the method of unscented Kalman filter + nonlinear least squares can avoid the problem of filter failure caused by the number of equations being less than the number of unknowns, effectively deal with the problem of missing baseline observation data under undersampling working condition, and ensure the implementation of real-time fusion positioning of satellite formation system. At the same time, since the embodiment chooses to solve the problem from the perspective of probability and statistics, it effectively avoids the linearization process and overcomes the shortcomings of low precision and poor stability of extended Kalman filter.

[0116] Specific implementation method eight: different from one of the specific implementation methods one to seven, the specific process of step three is:

[0117] Step three one, the baseline observation equation set is:

[0118] Y″=F(X″)

[0119] Wherein, F(·) is the baseline observation equation;

[0120] Definition:

[0121] Y″=[d1 d2... d L ] T

[0122] X″=[P1 P2... P M ] T

[0123] In the formula: M represents the number of user stars, L≤M(M-1) / 2, d i represents the pre-processed i-th baseline observation value, unit: m; P j represents the to-be-solved variable of the j-th user star, P j =[x′ j y′ j z′ j t j ], 1≤j≤M, wherein x′ j , y′ j , z′ j represent the three-dimensional coordinates of the j-th user star, t j represents the clock difference between the j-th user star and the satellite;

[0124] Step three two, expand the baseline observation equation set and omit the high-order terms to obtain:

[0125]

[0126] wherein:

[0127] F0= F(X"0)

[0128] wherein, X"0 represents the initial value of the variable to be solved, i.e. the initial estimation result of the user star position obtained in step two, X"0= [P 10 P 20 ... P M0 ] T , F0 is the initial value of the baseline observation vector Y", P j0 represents the initial estimation result of the position of the jth user star;

[0129] Step three, constructing the l+1th iteration error equation of the baseline observation equation group:

[0130] X" l+1 = X" l + δX" l+1

[0131] wherein: X" l+1 represents the variable to be solved after the l+1th iteration correction, X" l represents the variable to be solved after the lth iteration correction, δX" l+1 = [δP 1(l+1) δP 2(l+1) ... δP M(l+1) ] T , δP j(l+1) represents the variable correction value of the jth user star in the l+1th iteration, with the unit of m, which is obtained according to the least square principle:

[0132] δX" l+1 = (B l+1 T B l+1 ) -1 B l+1 T L l+1

[0133] wherein:

[0134] L l+1 = [d1-d 1l d2 -d 2l ... d L -d Ll ] T

[0135]

[0136] d ilrepresents the calculated value of the i-th baseline after the l-th iteration of correction, in units of m; P jl represents the to-be-solved variable of the j-th user satellite after the l-th iteration of correction; for d i0 , P 10 may be calculated by substituting the baseline observation equation set. 20 , P M0 may be calculated by substituting the baseline observation equation set. il , l > 0, P 1l may be calculated by substituting the baseline observation equation set. 2l , P Ml may be calculated by substituting the baseline observation equation set.

[0137] until the set iteration stopping condition is met, and X" l+1 obtained in the last iteration is taken as the final position estimation result at the current time.

[0138] For any time, based on the processing results of steps one and two corresponding to the time, the method of the embodiment can be used for processing to obtain the final position estimation result at the time, so as to improve the positioning accuracy.

[0139] The other steps and parameters are the same as one of embodiments one to seven.

[0140] Embodiment nine: The embodiment is different from one of embodiments one to eight in that the iteration stopping condition is: ε0 is a constant, and a small value is given in advance.

[0141] wherein |·| represents taking the absolute value, σ l+1 and V l+1 are intermediate variables.

[0142] The other steps and parameters are the same as one of embodiments one to eight.

[0143] Embodiment ten: The embodiment is different from one of embodiments one to nine in that the intermediate variable V l+1 is:

[0144] V l+1 = B l+1 δX" l+1 + L l+1

[0145] The other steps and parameters are the same as one of embodiments one to nine.

[0146] Experimental part

[0147] Taking space-based observation as an example, the simplified observation model is as follows: Figure 1, the navigation stars in the figure are known position observation satellites, and the member stars are satellites to be positioned. Figure 3 First, the initial observation data is preprocessed, then the processed pseudo-range observation values are subjected to non-linear Kalman filtering to obtain an initial position estimate, and finally, the processed baseline observation values are taken as reference values, and the position estimate values are taken as initial values to perform non-linear least squares iteration, and when the error is less than a given value, the high-precision position estimate value is output, and the high-precision position estimate value is output to the pseudo-range observation value part for the next epoch fusion positioning.

[0148] For the high-precision real-time fusion positioning algorithm designed by the application, simulation is completed through an STK / MATLAB multi-platform simulation environment. In STK, a simulation scene is established and satellite objects are inserted, first, relevant data in the simulation environment is read into the MATLAB environment through relevant interfaces for algorithm processing, and then the calculation results are input into the STK scene for verification. In the simulation, four satellites are selected from the scene as space-based navigation stars with known position information, and nine satellites are selected as user satellites to be positioned with unknown position information, that is, the number of baseline observation values is 36 in full sampling, and the number of baseline observation values is less than 36 in under-sampling. For data setting, the pseudo-range observation value noise is set to 2, and the baseline observation value noise is set to 0.005.

[0149] From the nine user stars, No. 3 satellite is selected for simulation result analysis, and the simulation results in full sampling working condition are as follows Figure 4 , and the simulation results in under-sampling working condition are as follows Figure 5 From the simulation results of Figure 4 and Figure 5 , it can be seen that the high-precision real-time fusion positioning algorithm designed by the application has obvious improvement in three-dimensional positioning accuracy compared with the traditional positioning algorithm, and the error covariance is obviously reduced, which shows that the fusion positioning algorithm of Kalman filtering and non-linear least squares has achieved the expected effect.

[0150] The above examples of the application only serve to illustrate the calculation model and calculation process of the application, and are not intended to limit the embodiments of the application. For those skilled in the art, other different forms of changes or modifications can be made on the basis of the above description, and it is impossible to enumerate all the embodiments here, and any obvious changes or modifications derived from the technical solutions of the application still fall within the protection scope of the application.

Claims

1. A real-time fusion positioning method for a satellite formation system, characterized in that, The method specifically comprises the following steps: Step one, the baseline observation value is pretreated by using Kalman filter to determine the integer ambiguity, and then the integer ambiguity is substituted into the baseline observation equation to obtain the pretreated baseline observation value; Step two, the pseudorange observation value is pretreated by using nonlinear Kalman filter to obtain the initial estimation result of the user star position; Step three, the final estimation result of the user star position is obtained according to the pretreated baseline observation value in step one and the initial estimation result in step two; The specific process of step three is as follows: Step three one, the baseline observation equation set is as follows: wherein is the baseline observation equation; Definition: In the formula: represents the number of user stars, , represents the pre-processed baseline observation value of the th user star, and the unit is ; represents the to-be-solved variable of the th user star, , , wherein, represents the three-dimensional coordinates of the th user star, represents the clock difference between the th user star and the reference satellite; Step three two, the baseline observation equation set is expanded and the high-order term is omitted to obtain: In the formula, an initial value representing a variable to be solved, , an initial value of the baseline observation vector , an initial estimate of the position of the th user satellite; Step three, construct the baseline observation equation set Second iteration error equation: In the formula: represent the first iterative correction of the to-be-solved variable, represent the first iterative correction of the to-be-solved variable, , represent the first user star in the first iterative correction of the variable, the unit is , and the formula is obtained according to the least square principle: In the formula, represent the first baseline after the first iterative correction, in units of ; represent the first user star after the first iterative correction of the unknown variable; until a set iteration stop condition is satisfied, the final iteration obtained as the final position estimation result at the current time point.

2. The real-time fusion positioning method for a satellite formation system according to claim 1, characterized in that, In step one, the baseline observation value is pretreated by using Kalman filter to determine the integer ambiguity, and the specific process is as follows: wherein: represents the first baseline observation value, in rad; represents the first pseudo-range observation value, in m; represents the first baseline observation value, in rad; represents the first pseudo-range observation value, in m; represents the first integer ambiguity, , , represents the number of integer ambiguities; represents the wavelength at the carrier phase observation frequency , in m; defining a state vector of a baseline observation model : Defining observation variables for a baseline observation model : Solving the state vector of the baseline observation model by Kalman filtering formula Obtaining integer ambiguity , ,… : wherein is the state vector estimate of the baseline observation model at time k, is a state transition matrix, the superscript T denotes the transpose of a matrix, is the one-step prediction of the state vector estimate of the baseline observation model at time k, is the input noise driving matrix of the baseline observation model, is the state vector covariance estimate of the baseline observation model at time k, is a system model noise matrix, is the observation matrix of the baseline observation model, is the one-step prediction of the state vector covariance of the baseline observation model at time k, is the noise matrix of the baseline observation model, the superscript -1 denotes the inverse of a matrix, is the state vector estimate of the baseline observation model at time k, is the gain of the baseline observation model at time k, is the observation variable of the baseline observation model at time k, is the state vector covariance estimate of the baseline observation model at time k, is an identity matrix.

3. The real-time fusion positioning method for a satellite formation system according to claim 2, characterized in that, The baseline observation value is obtained by using carrier phase measurement method.

4. The real-time fusion positioning method for a satellite formation system according to claim 3, characterized in that, The integer ambiguity is substituted into the baseline observation equation to obtain the pretreated baseline observation value, and the specific process is as follows: In the formula: represents the pre-processed baseline observation value, and the unit is .​ 5. The real-time fusion positioning method for a satellite formation system according to claim 4, characterized in that, The specific process of step two is as follows: State vector defining the pseudorange observation model As follows: In the formula, represents the three-dimensional coordinates of the user star, with units of ; represents the three-dimensional velocity of the user star, with units of ; represents the clock difference between the user star and the reference satellite, with units of ; represents the clock difference variation rate between the user star and the satellite; The three-dimensional coordinates of the user star, i.e. the initial estimation result of the user star position, are obtained by solving through extended Kalman filter; in, for The final position estimation result at time [time] For a linear system model, for One-step prediction of state variables in a time pseudorange observation model yes The state transition matrix at time t, where the superscript T represents the transpose of the matrix. for State covariance estimation for time-major observation models This is the input noise covariance matrix of the pseudorange observation model. This is the input noise driving matrix for the pseudorange observation model. for One-step prediction of the state covariance of the time pseudorange observation model. for Jacobi matrix of the time pseudorange observation equation, This is the covariance matrix of the observation noise in the pseudorange observation model, where the superscript -1 represents the inverse of the matrix. For pseudorange observation equations, for State vector estimation of the time pseudorange observation model For pseudorange observation model Gain at any moment For pseudorange observation model State covariance estimation at time 1 It is the identity matrix. It is a pseudorange observation model Observations of time.

6. The real-time fusion positioning method for a satellite formation system according to claim 5, wherein, The Jacobi matrix of the pseudorange observation equation is as follows: ， wherein, , represents the first pseudo-range observation value to the user satellite position, represents and represents the total number of pseudo-range observation values.​​ 7. The real-time fusion positioning method for a satellite formation system according to claim 4, wherein, The specific process of step two is as follows: The three-dimensional coordinates of the user star, i.e. the initial estimation result of the user star position, are obtained by solving through extended Kalman filter; State vector defining the pseudorange observation model As follows: In the formula, represents the three-dimensional coordinates of the user star, with units of ; represents the three-dimensional velocity of the user star, with units of ; represents the clock difference between the user star and the satellite, with units of ; represents the clock difference rate between the user star and the satellite; Definition of augmented state vector wherein, is the number of sampling points, represents the state quantity covariance matrix, , is a constant, and satisfies is positive semi-definite, is a constant; Using unscented Kalman filtering The calculation is performed to obtain the three-dimensional coordinates of the user satellite, which is the initial estimate of the user satellite's position. in, For pseudorange observation model State vector estimation at time t. for time The estimate, for One-step prediction of the augmented state vector at time step. For a linear system model, For pseudorange observation model One-step prediction of the state vector at time step 1. It is the first Each sampling point is One-step prediction of the augmented state vector at time step 1. Representing the The weights of each sampling point are calculated. , For intermediate variables, the superscript T represents transpose. Let be the covariance matrix of the input noise for the pseudorange observation model. This is the input noise driving matrix for the pseudorange observation model. For pseudorange observation equations, Indicates passage The obtained observation prediction point set, For pseudorange observation model Gain at any moment express The observed value at time, For the prediction point set The mean, yes The covariance matrix, yes and The cross-covariance matrix, yes One step in the prediction.

8. The real-time fusion positioning method for a satellite formation system according to claim 6 or 7, characterized in that, The iteration stopping condition is: wherein, is a constant, represents taking the absolute value, and is an intermediate variable, .

9. The real-time fusion positioning method for a satellite formation system according to claim 8, wherein, The intermediate variable is: 。

Images