A robust adaptive filtering method for autonomous orbit determination of satellites

Abstract

A method for autonomous orbit determination and adaptive filtering for satellites, comprising: obtaining satellite initial state information; predicting and calculating satellite predicted state values ​​and satellite predicted state variance values ​​at epoch t1; forming orbit determination observation values ​​from two-way observation values ​​between satellites, and establishing observation equations; calculating filtering innovation values ​​of all observation values ​​at epoch t1 using satellite predicted state information and observation information; determining whether there are gross errors in corresponding observation values ​​based on the size of the filtering innovation values, or establishing whether there are orbit maneuvers for a transmitting satellite or a receiving satellite of the corresponding observation value; if there are gross errors, calculating equivalent observation variances, and performing filtering observation updates using the equivalent observation variances to obtain updated observation values; if there are orbit maneuvers, calculating equivalent predicted state variances of maneuvering satellites, and performing filtering observation updates on maneuvering satellite state parameters using the equivalent predicted state variances to obtain updated observation values.

Description

Technical Field

[0001] The invention relates to the field of satellite navigation systems, and in particular to an anti-error adaptive filtering method for autonomous orbit determination of a satellite. Background Art

[0002] The BeiDou system is a global satellite navigation system independently developed and built by my country. In accordance with the "three-step" development strategy, on July 31, 2020, the BeiDou-3 global satellite navigation system was officially completed. Compared with BeiDou-2 satellites, BeiDou-3 satellites are equipped with inter-satellite link payloads. Through ranging, communication and data exchange between satellites, the BeiDou-3 system can achieve autonomous orbit determination without the support of ground monitoring stations. Autonomous orbit determination can effectively reduce the dependence of the navigation satellite system on the ground control system, reduce system construction costs, shorten the satellite calendar update cycle and improve system service performance. In the event of loss of contact with the ground control system, the BeiDou-3 system can maintain its normal service performance for at least 180 days through autonomous orbit determination. Therefore, autonomous orbit determination can not only improve system performance, but also effectively ensure the wartime vitality of the system, which is crucial to the stable operation of the BeiDou-3 system and the realization of global services.

[0003] At present, the autonomous orbit determination of the BeiDou-3 system is implemented using the Kalman filter algorithm. The process is as follows: when the initial orbit value is known, the satellite state is predicted in one step using the orbit integral; the orbit determination observation equation is composed of the two-way observation values ​​between satellites in this epoch, and the predicted satellite state is updated by observation; when the initial satellite clock error value is known, the satellite clock error is predicted in one step using the clock error prediction model; the time synchronization observation equation is composed of the two-way observation values ​​between satellites in this epoch, and the predicted satellite clock error is updated by observation. The current autonomous orbit determination filter algorithm for the BeiDou-3 system is only applicable to conventional situations.

[0004] However, due to the failure of inter-satellite link equipment and the orbit control requirements of the BeiDou-3 system's high-orbit (GEO and IGSO) satellites, observation anomalies (caused by satellite ranging equipment anomalies) and state anomalies (caused by the periodic orbital precession of GEO and IGSO satellites) are unavoidable during the autonomous orbit determination process. If these are not taken into account, the accuracy of autonomous orbit determination will be severely reduced, thereby affecting the global navigation and positioning service performance of the BeiDou-3 satellite navigation system. Summary of the invention

[0005] In view of the technical defects and technical drawbacks existing in the prior art, the embodiment of the present invention provides a method for autonomous orbit determination of satellites to overcome the above problems or at least partially solve the above problems. The basic idea is to use the state domain information and observation domain information to solve the estimated parameters according to the principle of Kalman filtering, and use the filter innovation value reflecting the difference between the two to calibrate the observation value with observation anomaly or state anomaly. When there is an observation anomaly, the expansion factor is calculated using the standardized observation residual as the control amount, and then the equivalent observation variance is calculated for filtering update; when there is a state anomaly, the normal filtering observation update is first performed on the satellite without abnormality. For the satellite with state anomaly, the adaptive factor is calculated using the standardized state residual as the control amount after the normal observation update, and then the equivalent predicted state variance is calculated for filtering update. The scheme proposed by the present invention can effectively distinguish between observation anomalies and state anomalies, and can greatly improve the autonomous orbit determination accuracy of the BeiDou-3 system under abnormal conditions without interrupting the real-time solution process of autonomous orbit determination, and ensure the reliability of its global service performance. The specific scheme is as follows:

[0006] A method for autonomous orbit determination of a satellite with robust adaptive filtering, the method comprising:

[0007] Step 1, obtaining satellite initial status information;

[0008] Step 2: Based on the initial state information of the satellite, the predicted state value of the satellite and the predicted state variance value of the satellite at epoch t1 are predicted and calculated;

[0009] Step 3, composing orbit determination observations from bidirectional observations between satellites, and establishing observation equations based on orbit determination observations and satellite state parameters;

[0010] Step 4, using the satellite prediction state information and observation information to calculate the filtered innovation values ​​of all observation values ​​at epoch t1, wherein the satellite prediction state information is obtained by the satellite prediction state value, and the satellite observation information is obtained by the observation equation;

[0011] Step 5: determine whether there is a gross error in the corresponding observation value based on the size of the filtered innovation value, or whether there is an orbital maneuver of the transmitting satellite or the receiving satellite of the corresponding observation value;

[0012] Step 6: If there is a gross error, the equivalent observation variance is calculated, and the equivalent observation variance is used to perform filtered observation update to obtain an updated observation value; if there is an orbital maneuver, the equivalent predicted state variance of the maneuvering satellite is calculated, and the equivalent predicted state variance is used to perform filtered observation update on the maneuvering satellite state parameters to obtain an updated observation value;

[0013] Step 7: Use the updated observations to perform autonomous orbit determination calculations for the next epoch.

[0014] Furthermore, in step 1, the satellite initial state information includes the satellite initial state value Initial parameter prior expectation The prior variance of the observations Σ0 and the prior process noise of the state equation Σ W .

[0015] Furthermore, step 2 is specifically as follows:

[0016] According to the in-orbit force of the satellite, the perturbation motion equations of the GEO, IGSO and MEO satellites of the BeiDou-3 system are established using an accurate perturbation force model. Based on the premise, the perturbation motion equations of GEO, IGSO and MEO satellites are integrated by numerical integration method to obtain the state transfer matrix Φ from epoch t1 to epoch t0. 1,0 , using the state transfer matrix Φ 1,0 Calculate the satellite prediction status value at epoch t1 And the satellite prediction state variance

[0017]

[0018]

[0019] Furthermore, in step 3, the orbit determination observation values ​​formed by the two-way observation values ​​between satellites are specifically:

[0020] The two-way observations between satellites at epoch t1 are combined into orbit determination observations, and the process is as follows:

[0021] Choose any two satellite SVs A With SV B , let SV A Transmit ranging signal and SV B The observation value composed of the received signal is the forward observation value D AB ; Let SV B Transmit ranging signal and SV A The observation value composed of the received signal is the reverse observation value D BA , then the satellite's SV A With SV B The orbital observation value between AB +D BA ;

[0022] All two-way observations in the corresponding epoch are added up in turn to obtain the orbit determination observations of all satellite pairs.

[0023] Furthermore, in step 3, the observation equation based on orbit-determined observation values ​​and satellite state parameters is established as:

[0024]

[0025] Where L1 represents all orbit determination observations within epoch t1, represents the satellite state parameter value at epoch t1, and A1 represents the design matrix of the observation equation.

[0026] Furthermore, in step 4, the satellite prediction state information and observation information are used to calculate the filtered innovation value of all observation values ​​at epoch t1, which is expressed as the following equation:

[0027]

[0028] Furthermore, step 5 is specifically as follows:

[0029] Set the threshold l_max of the filter innovation value. Since the filter innovation value is used as a parameter to characterize the difference between state information and observation information, if the filter innovation value of the observation value l>l_max, the observation quality flag Obs_Quality is calibrated to 1. If l≤l_max, the observation quality flag Obs_Quality is calibrated to 0. i =1 means that there is a gross error in the corresponding observation value or that the transmitting satellite or the receiving satellite has orbital maneuvers.

[0030] Furthermore, step 5 also includes:

[0031] Since the gross error occurs in the observation value domain, and the maneuver occurs in the satellite state value domain, in order to distinguish the cause of the filter innovation exceeding the limit, a quality flag SV_Quality is set for each satellite and assigned an initial value of 0, and all observations of the corresponding epoch are traversed: if an observation value corresponds to the transmitting signal satellite PRN number j and the corresponding receiving signal satellite PRN number p, and its quality flag Obs_Quality = 0, then the quality flags of satellite j and satellite p are increased by 1 (SV_Quality j +1, SV_Quality p +1); if the observation value quality flag Obs_Quality = 1, skip and do not perform any processing;

[0032] If no satellite maneuvers occur in the corresponding epoch, the filter innovation limit will be exceeded only by the observation gross error. According to the randomness and small probability of the observation gross error, the quality mark SV_Quality of all satellites is a value greater than 0 at this time; if satellite j maneuvers occur in this epoch, the quality mark Obs_Quality of all observations related to satellite j is 1, and the quality mark SV_Quality of satellite j is 1. j will be equal to 0;

[0033] In this way, all satellites are traversed and their quality flags are judged: if SV_Quality j > 0, it means that satellite j has not maneuvered; if SV_Quality j =0, it means that satellite j has maneuvered. At this time, the quality flag Obs_Quality of the observation value related to satellite j (satellite j serves as a transmitting satellite or a receiving satellite) is assigned a value of 2.

[0034] Furthermore, in step 6, if there is a gross error, the equivalent observation variance is calculated, and the equivalent observation variance is used to perform filter observation update, and the updated observation value is specifically:

[0035] Step 611: Let Σ1=Σ0, and calculate the filter gain matrix K:

[0036]

[0037] Step 612: Calculate the observation correction:

[0038]

[0039]

[0040] Step 613: Calculate the variance of the observed value correction:

[0041]

[0042] make Calculate the error in the corrections to the observed values:

[0043]

[0044] Where m represents the number of orbit determination observations in epoch t1; represents the mean error of the correction of the ith observation in epoch t1;

[0045] Step 614: Calculate the observed standard residual:

[0046]

[0047] Step 615: Calculate the observed value variance inflation factor:

[0048]

[0049] In the formula, a is an artificially set threshold, usually in the range of [1,1.5];

[0050] Step 616: Calculate the equivalent observation variance:

[0051]

[0052] Step 617: Use the equivalent observation variance to perform filtered observation update:

[0053]

[0054]

[0055]

[0056] Furthermore, in step 6, if there is an orbital maneuver, the equivalent predicted state variance of the maneuvering satellite is calculated, and the equivalent predicted state variance is used to perform filtering observation update on the maneuvering satellite state parameters, and the updated observation value is specifically:

[0057] Step 622: Calculate the parameter standardized deviation:

[0058]

[0059] In the formula, is the trace of the variance matrix of the satellite predicted state parameters.

[0060] Step 623: Calculate the adaptive factor:

[0061]

[0062] Where c is an artificially set threshold, which is usually set to 2.5.

[0063] Step 623: Calculate the equivalent predicted state variance:

[0064]

[0065] Step 624: Filter observation and update the maneuvering satellite state parameters using the equivalent predicted state variance:

[0066]

[0067]

[0068]

[0069] The present invention has the following beneficial effects:

[0070] 1) In the autonomous orbit determination calculation of BDS-3 satellite, both observation anomalies and state anomalies are taken into account, expanding the applicable scenarios of the autonomous orbit determination algorithm;

[0071] 2) The algorithm proposed in this invention can effectively reduce the impact of observation anomalies and state anomalies on the accuracy of BDS-3 autonomous orbit determination;

[0072] 3) The algorithm proposed in this invention can effectively improve the availability and reliability of the autonomous orbit determination results of the BDS-3 satellite. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 A flow chart of a robust adaptive filtering method for autonomous orbit determination of a satellite provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0074] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0075] like Figure 1 As shown, a method for autonomous orbit determination and robust adaptive filtering for a satellite provided by an embodiment of the present invention specifically includes the following steps:

[0076] Step 1: Given the initial state information of the satellite: including the satellite state parameter values ​​at the initial epoch t0 Satellite status parameter values Contains satellite 3D position, 3D velocity, transmit and receive hardware delay, and solar pressure, which can be obtained by fitting the precise satellite ephemeris; Given a priori random information: state parameter prior variance The prior variance of the observations is Σ0, and the prior process noise of the state equation is Σ W , whose prior variance values ​​are given artificially according to experience, where Σ W It remains unchanged throughout the calculation.

[0077] Step 2: According to the satellite in-orbit force conditions, the perturbation motion equations of the BeiDou-3 system GEO, IGSO and MEO satellites are established using the accurate perturbation force model. Under the premise of , the perturbation motion equations of the three types of satellites are integrated by numerical integration method to obtain the state transfer matrix Φ from epoch t1 to epoch t0 1,0 , using the state transfer matrix Φ1,0 Calculate the satellite prediction status value at epoch t1 And the satellite prediction state variance

[0078]

[0079]

[0080] Step 3: Combine the two-way observations between satellites at epoch t1 into orbit determination observations. The process is as follows: arbitrarily select two satellite SVs A With SV B , if SV A Transmit ranging signal, SV B The observation value composed of the received signal is the forward observation value D AB ; Let SV B Transmit ranging signal, SV A The observation value composed of the received signal is the reverse observation value D BA , then the satellite pair (SV A With SV B ) is the orbital observation value between AB +D BA In this way, all two-way observations in the epoch are added together to obtain the orbit determination observations for all satellite pairs.

[0081] Step 4: Establish the observation equation and give the functional relationship between the orbit determination observation value and the satellite state parameters:

[0082]

[0083] Where L1 represents all orbit determination observations in epoch t1, represents the satellite state parameter value at epoch t1, and A1 represents the design matrix of the observation equation.

[0084] Step 5: Use the satellite prediction status information and observation information to calculate the filtered innovation value of all observations at epoch t1:

[0085]

[0086] Step 6: Since the filtered innovation value can be used as a parameter to characterize the difference between state information and observation information, if the filtered innovation value of the observation value is too large (over the limit), it can be considered that the observation value has a gross error or the observation value corresponds to an orbital maneuver of the transmitting signal satellite / receiving signal satellite. The present invention sets a threshold l_max to determine the filtered innovation value of each observation value: if l>l_max, the observation value quality flag Obs_Quality is calibrated to 1; if l≤l_max, the observation value quality label Obs_Quality is calibrated to 0, and Obs_Quality is i =1 indicates that there is a gross error in the i-th observation value or the transmitting satellite SV that established the i-th observation value A Or receive satellite SV B There is mobility.

[0087] Step 7: Since the gross error occurs in the observation value domain, and the maneuver occurs in the satellite state value domain. In order to further distinguish the cause of the filter innovation exceeding the limit, the present invention sets a quality flag SV_Quality for each satellite and assigns an initial value of 0. Traverse all observations of this epoch: if an observation corresponds to a transmitting signal satellite PRN number j and a receiving signal satellite PRN number p, and its quality flag Obs_Quality = 0, then add 1 to the quality flags of satellite j and satellite p respectively (SV_Quality j +1, SV_Quality p +1); if the observation value quality flag Obs_Quality=1, then skip and do not perform any processing.

[0088] If no satellite maneuvers occur in this epoch, the filter innovation limit will be exceeded only by the observation gross error. According to the randomness and small probability of the observation gross error, the quality mark SV_Quality of all satellites will be a value greater than 0 at this time; if satellite j maneuvers occur in this epoch, the quality mark Obs_Quality of all observations related to satellite j will be 1. According to the processing strategy of the present invention, the quality mark SV_Quality of satellite j will be 1 at this time. j Will be equal to 0, so next traverse all satellites and judge their quality flags: If SV_Quality j > 0, it means that satellite j has not maneuvered; if SV_Quality j =0, it means that satellite j has maneuvered. At this time, the quality flag Obs_Quality of the observation value related to satellite j (satellite j serves as a transmitting satellite or a receiving satellite) is assigned a value of 2.

[0089] Step 8: Determine the quality mark Obs_Quality of all observations: If Obs_Quality i=2, it means that the observation value includes a maneuvering satellite (a maneuvering satellite). In order to prevent the abnormal state of the maneuvering satellite from affecting the orbit determination accuracy of other satellites, the observation value is skipped and not processed first; if Obs_Quality i ≠2, then calculate the equivalent observation variance of the observation value, the steps are as follows:

[0090] Step 8.1: Let Σ1 = Σ0 and calculate the filter gain matrix K:

[0091]

[0092] Step 8.2: Calculate the corrections to the observations:

[0093]

[0094]

[0095] Step 8.3: Calculate the variance of the observed corrections:

[0096]

[0097] make Calculate the error in the corrections to the observed values:

[0098]

[0099] Where m represents the number of orbit determination observations in epoch t1; It represents the mean error of the correction of the ith observation in epoch t1.

[0100] Step 8.4: Calculate the observed standard residuals:

[0101]

[0102] Step 8.5: Calculate the observed variance inflation factor:

[0103]

[0104] Where a is an artificially set threshold, usually in the range of [1,1.5].

[0105] Step 8.6: Calculate the equivalent observation variance:

[0106]

[0107] Step 9: Update the filtered observation using the equivalent observation variance:

[0108]

[0109]

[0110]

[0111] Step 10: Determine the quality mark Obs_Quality of all observations: If Obs_Quality i =2, then the equivalent predicted state variance of the dynamic satellite is calculated as:

[0112] Step 10.1: Calculate the parameter standardized deviation:

[0113]

[0114] In the formula, is the trace of the variance matrix of the satellite predicted state parameters.

[0115] Step 10.2: Calculate the adaptive factor:

[0116]

[0117] Where c is an artificially set threshold, which is usually set to 2.5.

[0118] Step 10.3: Calculate the equivalent predicted state variance:

[0119]

[0120] Step 11: Use the equivalent predicted state variance to filter and update the state parameters of the maneuvering satellite:

[0121]

[0122]

[0123]

[0124] Step 12: Repeat steps 2 to 12 to perform autonomous orbit determination calculations for epoch t2, and so on to calculate the satellite orbits for all time periods.

[0125] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A robust adaptive filtering method for autonomous orbit determination of a satellite, characterized in that: The method comprises: Step 1, obtaining satellite initial status information; Step 2: Based on the initial state information of the satellite, the predicted state value of the satellite and the predicted state variance value of the satellite at epoch t1 are predicted and calculated; Step 3, composing orbit determination observations from bidirectional observations between satellites, and establishing observation equations based on orbit determination observations and satellite state parameters; Step 4, using the satellite prediction state information and observation information to calculate the filtered innovation values ​​of all observation values ​​at epoch t1, wherein the satellite prediction state information is obtained by the satellite prediction state value, and the satellite observation information is obtained by the observation equation; Step 5: determine whether there is a gross error in the corresponding observation value based on the size of the filtered innovation value, or whether there is an orbital maneuver of the transmitting satellite or the receiving satellite of the corresponding observation value; Step 6: If there is a gross error, the equivalent observation variance is calculated, and the equivalent observation variance is used to perform filtered observation update to obtain an updated observation value; if there is an orbital maneuver, the equivalent predicted state variance of the maneuvering satellite is calculated, and the equivalent predicted state variance is used to perform filtered observation update on the maneuvering satellite state parameters to obtain an updated observation value; Step 7: Use the updated observations to perform autonomous orbit determination calculations for the next epoch.

2. The method for autonomous orbit determination of a satellite according to claim 1, characterized in that: In step 1, the satellite initial state information includes the satellite initial state value Initial parameter prior expectation The prior variance of the observations Σ0 and the prior process noise of the state equation Σ W .

3. The robust adaptive filtering method for autonomous orbit determination of a satellite according to claim 2, characterized in that: Step 2 is as follows: According to the in-orbit force of the satellite, the perturbation motion equations of the GEO, IGSO and MEO satellites of the BeiDou-3 system are established using an accurate perturbation force model. Based on the premise, the perturbation motion equations of GEO, IGSO and MEO satellites are integrated by numerical integration method to obtain the state transfer matrix Φ from epoch t1 to epoch t0. 1,0 , using the state transfer matrix Φ 1,0 Calculate the satellite prediction status value at epoch t1 And the satellite prediction state variance 4. The method for autonomous orbit determination of satellite according to claim 3, characterized in that: In step 3, the orbit determination observations composed of the two-way observations between satellites are specifically: The two-way observations between satellites at epoch t1 are combined into orbit determination observations, and the process is as follows: Choose any two satellite SVs A With SV B , let SV A Transmit ranging signal and SV B The observation value composed of the received signal is the forward observation value D AB ; Let SV B Transmit ranging signal and SV A The observation value composed of the received signal is the reverse observation value D BA , then the satellite's SV A With SV B The orbital observation value between AB +D BA ; All two-way observations in the corresponding epoch are added up in turn to obtain the orbit determination observations of all satellite pairs.

5. The robust adaptive filtering method for autonomous orbit determination of a satellite according to claim 4, characterized in that: In step 3, the observation equation based on orbit-determined observations and satellite state parameters is established as: Where L1 represents all orbit determination observations within epoch t1, represents the satellite state parameter value at epoch t1, and A1 represents the design matrix of the observation equation.

6. The robust adaptive filtering method for autonomous orbit determination of a satellite according to claim 5, characterized in that: In step 4, the satellite prediction state information and observation information are used to calculate the filtered innovation value of all observations at epoch t1, which is expressed as the following equation:

7. The robust adaptive filtering method for autonomous orbit determination of a satellite according to claim 1, characterized in that: Step 5 is as follows: Set the threshold l_max of the filter innovation value. Since the filter innovation value is used as a parameter to characterize the difference between state information and observation information, if the filter innovation value l of the observation value is l>l_max, the observation quality flag Obs_Quality is calibrated to 1. If l≤l_max, the observation quality flag Obs_Quality is calibrated to 0. i =1 means that there is a gross error in the corresponding observation value or that the transmitting satellite or the receiving satellite has orbital maneuvers.

8. The robust adaptive filtering method for autonomous orbit determination of a satellite according to claim 7, characterized in that: Step 5 also includes: Since the gross error occurs in the observation value domain, and the maneuver occurs in the satellite state value domain, in order to distinguish the cause of the filter innovation exceeding the limit, the quality flag SV_Quality is set for each satellite and the initial value is 0, and all the observation values ​​of the corresponding epoch are traversed: if the corresponding PRN number of the transmitting signal satellite of an observation value is j, the corresponding PRN number of the receiving signal satellite is p, and its quality flag Obs_Quality = 0, then the quality flags of satellite j and satellite p are increased by 1 respectively; if the observation value quality flag Obs_Quality = 1, it is skipped without any processing; If no satellite maneuvers occur in the corresponding epoch, the filter innovation limit will be exceeded only by the observation gross error. According to the randomness and small probability of the observation gross error, the quality mark SV_Quality of all satellites is a value greater than 0 at this time; if satellite j maneuvers occur in this epoch, the quality mark Obs_Quality of all observations related to satellite j is 1, and the quality mark SV_Quality of satellite j is 1. j will be equal to 0; In this way, all satellites are traversed and their quality flags are judged: if SV_Quality j > 0, it means that satellite j has not maneuvered; if SV_Quality j =0, it means that satellite j has maneuvered, and the quality flag Obs_Quality of the observation value related to satellite j is assigned a value of 2.

9. The robust adaptive filtering method for autonomous orbit determination of a satellite according to claim 6, characterized in that: In step 6, if there is a gross error, the equivalent observation variance is calculated, and the equivalent observation variance is used to update the filtered observation to obtain the updated observation value: Step 611: Let Σ1=Σ0, and calculate the filter gain matrix K: Step 612: Calculate the observation correction: Step 613: Calculate the variance of the observed value correction: make Calculate the error in the corrections to the observed values: Where m represents the number of orbit determination observations in epoch t1; represents the mean error of the correction of the ith observation in epoch t1; Step 614: Calculate the observed standard residual: Step 615: Calculate the observed value variance inflation factor: In the formula, a is an artificially set threshold, usually in the range of [1,1.5]; Step 616: Calculate the equivalent observation variance: Step 617: Use the equivalent observation variance to perform filtered observation update:

10. The robust adaptive filtering method for autonomous orbit determination of a satellite according to claim 9, characterized in that: In step 6, if there is an orbital maneuver, the equivalent predicted state variance of the maneuvering satellite is calculated, and the equivalent predicted state variance is used to perform filtered observation update on the maneuvering satellite state parameters to obtain the updated observation value: Step 622: Calculate the parameter standardized deviation: In the formula, is the trace of the variance matrix of the satellite predicted state parameters; Step 623: Calculate the adaptive factor: In the formula, c is an artificially set threshold, which is usually set to 2.5; Step 623: Calculate the equivalent predicted state variance: Step 624: Filter observation and update the maneuvering satellite state parameters using the equivalent predicted state variance:

Images