Spacecraft landing disturbance estimation method and device based on SVDCKF algorithm

Abstract

The invention provides a spacecraft landing disturbance estimation method and device based on an SVDCKF algorithm, relates to the field of aerospace, and aims at generating disturbed state output by establishing a dynamic model of a spacecraft state, simulating external disturbance based on a real or simulated landing environment and inputting the external disturbance into the dynamic model. The disturbance injection module is used for simulating external disturbance in a simulation or real landing environment, and output of the spacecraft in a disturbed state is obtained. And then, performing time updating and measurement updating on the disturbed state through an SVDCKF algorithm, executing singular value decomposition to optimize a covariance matrix, generating a volume point, realizing real-time estimation of the spacecraft state and the external disturbance, and outputting the corrected estimation result of the spacecraft state and the external disturbance so as to provide high-precision disturbance estimation. And the stability and the safety of the spacecraft in a complex landing environment are ensured.

Description

Technical Field

[0001] This invention relates to the aerospace field, and more particularly to a method and apparatus for estimating spacecraft landing disturbances based on the SVDCKF algorithm. Background Technology

[0002] During spacecraft landing, the environment is complex and variable, including various types of atmospheric disturbances, terrain undulations, sensor noise, and many other uncertainties. Traditional disturbance estimation techniques mainly rely on the Kalman filter (KF) method. While it performs well in handling linear systems and Gaussian noise, it often suffers from insufficient estimation accuracy and robustness when facing strongly nonlinear, non-Gaussian noise, and multi-source coupled disturbances, making it difficult to meet the requirements for high-precision and high-reliability spacecraft landings. To address some nonlinear problems, improved methods such as the Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF) have been proposed. Although these methods can improve the estimation performance in nonlinear systems, they still have the following limitations under strong disturbances or multiple noise coupling conditions:

[0003] (1) Computational complexity and real-time performance: The linearization step of the extended Kalman filter for complex nonlinear equations may generate a large number of approximation errors in the calculation; while the unscented Kalman filter requires more volume points in high-dimensional states or multi-source disturbance scenarios, which leads to a rapid increase in computational complexity and makes it difficult to achieve real-time performance.

[0004] (2) Adaptability and robustness: When the disturbance distribution deviates from the Gaussian assumption or the system model has large uncertainties, traditional linear filtering is prone to accuracy decline or even divergence; when faced with terrain impact, turbulent change or large-scale sensor noise, the algorithm lacks the ability to adapt to complex scenarios.

[0005] (3) Insufficient adaptability to new or unseen disturbances: Although data-driven methods such as neural networks can show good estimation results under certain conditions, they often rely on a large amount of prior training data and have limited generalization ability to unknown or rare disturbance patterns. Once they encounter disturbances that are significantly different from the training data, the model performance is prone to a sharp decline, making it difficult to ensure the safety of spacecraft landing.

[0006] Therefore, current technologies struggle to provide a high-precision and robust disturbance estimation solution for spacecraft landing while simultaneously addressing strong nonlinearity, non-Gaussian noise, multiple disturbance types, and real-time requirements. Failure to accurately detect and compensate for these disturbances can easily lead to excessive landing deviations or attitude instability, thereby threatening the safety and reliability of the spacecraft and the mission. Summary of the Invention

[0007] To overcome the shortcomings of existing technologies, the technical problem to be solved by this invention is to propose a spacecraft landing disturbance estimation method and device based on the SVDCKF algorithm, and the following technical solution is adopted:

[0008] This invention provides a spacecraft landing perturbation estimation method based on the SVDCKF algorithm, comprising the following steps:

[0009] S1: Establish a spacecraft dynamics model to characterize several state parameters of the spacecraft. The aforementioned state parameters include at least position parameters, velocity parameters, orientation parameters, force parameters, and physical parameters, and obtain the initial state output.

[0010] S2: Based on the real or simulated landing environment, inject several external disturbances into the above spacecraft dynamics model to obtain the disturbed state output that represents the actual response of the spacecraft under the action of external disturbances.

[0011] S3: Based on capacitive Kalman filtering, singular value decomposition is performed on the state covariance matrix describing the system in the above-mentioned disturbed state to obtain the decomposition result;

[0012] S4: Using the above decomposition results, generate several volume points for predicted states, and combine them with the above spacecraft landing dynamics model to update the above disturbed states over time, and calculate the prior state estimate and prior covariance matrix.

[0013] S5: Based on the measurement equation, the measurement propagation volume point is mapped to the measurement space, and the measurement estimate, measurement error covariance matrix, and state-measurement interaction covariance matrix are calculated.

[0014] S6: Based on the calculation results of step S5, calculate the Kalman gain and correct the prior state estimate and prior covariance matrix. Output the corrected spacecraft state and the real-time estimate of several external disturbances mentioned above. Repeat steps S3 to S6 until landing is completed.

[0015] As a further improvement, in step S3, singular value decomposition is performed on the system state covariance matrix, specifically in the form of...

[0016]

[0017] Among them, U k It is an orthogonal matrix. It is a diagonal matrix that retains only non-zero singular values.

[0018] As a further improvement, in step S4, volume points are generated based on the covariance decomposition results in step S3, and the calculation formula is:

[0019]

[0020] Among them, the volume point direction set ξ i and weight ω i Determined by the following formula:

[0021]

[0022] The propagation volume point at the next time step is obtained through the state transition function f(·) based on the above volume point:

[0023]

[0024] As a further improvement, in step S4, the propagation volume point is... Calculate the prior state estimate and the prior covariance matrix P k+1 / k The formula for its calculation is:

[0025]

[0026] Among them, Q k Let be the process noise covariance matrix.

[0027] As a further improvement, in step S5, the prior error covariance matrix P obtained in step S4 is... k+1 / k Perform singular value decomposition, specifically in the form of:

[0028]

[0029] Among them, U k+1 / k It is an orthogonal matrix. A diagonal matrix that retains only non-zero singular values;

[0030] Based on the aforementioned prior error covariance matrix P k+1 / k The decomposition results are used to calculate the prior volume point X. k+1 / k The formula for its calculation is:

[0031]

[0032] The a priori volume point X is obtained by measuring the equation H(·). k+1 / k Mapping to the measurement space yields the measurement propagation volume point.

[0033]

[0034] As a further improvement, in step S5, the formula for calculating the above measurement estimate is:

[0035]

[0036] The formula for calculating the above measurement error covariance matrix is:

[0037]

[0038] Among them, R k To measure the noise covariance matrix;

[0039] The formula for calculating the interaction covariance matrix between the above states and measurements is:

[0040]

[0041] As a further improvement, in step S6, the formula for calculating the Kalman gain is:

[0042]

[0043] From the Kalman gain-corrected prior state estimate and the prior covariance matrix, we obtain the state estimate and the state estimate error covariance matrix at time k+1. The formula for calculating the state estimate at time k+1 is:

[0044]

[0045] The formula for calculating the state estimation error covariance matrix at time k+1 is:

[0046]

[0047] As a further improvement, in step S1 above,

[0048] The aforementioned location parameters include at least the longitude, latitude, and mirror distance of the spacecraft's location from the target ground.

[0049] The aforementioned velocity parameters include at least the spacecraft's velocity relative to the target ground.

[0050] The aforementioned directional parameters include at least the flight track angle and the flight heading angle;

[0051] The aforementioned force parameters include at least lift acceleration and drag acceleration;

[0052] The aforementioned physical parameters include at least the total mass and aerodynamic area of ​​the spacecraft.

[0053] As a further improvement, in step S2, the aforementioned external disturbances include at least atmospheric turbulence disturbances, impact disturbances caused by terrain undulations, and sensor noise disturbances.

[0054] Another aspect of the present invention provides a spacecraft landing perturbation estimation device based on the SVDCKF algorithm, which applies the estimation method described in any one of the above claims, including...

[0055] The spacecraft dynamics model module outputs the initial state of the spacecraft and updates the dynamic characteristics of the spacecraft during the landing process based on the spacecraft's position, velocity, orientation, attitude and other state parameters and physical parameters.

[0056] The disturbance injection module is used to inject external disturbances into the aforementioned spacecraft dynamics model module in a real or simulated landing environment to obtain the spacecraft's output under disturbed conditions.

[0057] The disturbance estimation module is used to perform time and measurement updates on the above-mentioned disturbed state using the SVDCKF algorithm. It performs singular value decomposition on the state covariance matrix, generates volume points, and completes real-time estimation of the spacecraft state and disturbance parameters, thereby outputting the corrected spacecraft state and external disturbance estimation results.

[0058] Compared with the prior art, the beneficial effects of the present invention are:

[0059] Firstly, addressing the issues of insufficient estimation accuracy and instability in existing technologies when facing nonlinear and complex disturbances, this invention introduces Singular Value Decomposition (SVD) to replace the traditional Cholesky Decomposition within the Capacitive Kalman Filter (CKF) framework. By performing SVD on the covariance matrix, the invention avoids nondefiniteness or numerical instability, significantly enhancing the robustness and accuracy of the filter in various nonlinear and non-Gaussian noise disturbance environments. This effectively solves the problem of decreased estimation accuracy in existing Kalman filtering methods under strongly nonlinear scenarios.

[0060] Secondly, addressing the issues of insufficient adaptability under multiple disturbance types and limited estimation results due to reliance on a single model or noise source, this invention constructs a disturbance injection module capable of simulating various disturbance forms, including atmospheric turbulence, terrain impact, and sensor noise, in real or simulated landing environments. By injecting rich and diverse disturbance data into the spacecraft dynamics model, it provides a more realistic training and testing scenario for the subsequent disturbance estimation module based on the SVDCKF algorithm, enabling the algorithm to maintain high accuracy and stability even when facing complex and variable disturbances.

[0061] Third, addressing the issue that data-driven methods are prone to drastic performance degradation in the face of unseen types of disturbances, this invention does not simply rely on pure data training. Instead, it utilizes a spacecraft dynamics model constructed from real physical mechanisms and multi-source disturbance injection, combined with the robustness of the SVDCKF algorithm in nonlinear systems. Thus, even when various new or extreme disturbances occur, it can still accurately estimate the spacecraft state and unknown disturbances, significantly reducing the dependence on massive training data and the difficulty of adapting to new scenarios.

[0062] Fourth, addressing the challenges of real-time updates, parallel computation, and adaptive parameter tuning during spacecraft landing, this invention jointly models the state space and noise statistical characteristics within the disturbance estimation module. The covariance matrix obtained from SVD decomposition allows for flexible adjustment of noise estimation and filtering gain. Through multiple simulations and comparisons with on-orbit data, filtering parameters can be adaptively adjusted based on real-time errors and changes in the external environment. This ensures that the filtering process can still converge quickly and maintain state estimation accuracy even during sudden changes in the landing environment or increased disturbance intensity. Attached Figure Description

[0063] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0064] Figure 1 This is a schematic diagram of the spacecraft landing disturbance estimation method based on the SVDCKF algorithm of the present invention;

[0065] Figure 2 This is a block diagram of the spacecraft landing disturbance estimation device based on the SVDCKF algorithm of the present invention.

[0066] Figure 3 This is a schematic diagram of the disturbance injection of the present invention;

[0067] Figure 4 This is a schematic diagram of the disturbance estimation of the present invention. Detailed Implementation

[0068] To facilitate understanding by those skilled in the art, the structure of the present invention will now be described in further detail with reference to the accompanying drawings:

[0069] In the description of this invention, the terms “part,” “side,” “end,” etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0070] like Figure 1 As shown, this application provides a spacecraft landing disturbance estimation method and apparatus based on the SVDCKF algorithm, comprising the following steps:

[0071] S1: Establish a spacecraft dynamics model to characterize several state parameters of the spacecraft. The state parameters include at least position parameters, velocity parameters, orientation parameters, force parameters, and physical parameters, and obtain the initial state output.

[0072] In this step, a corresponding spacecraft dynamics model is constructed based on the spacecraft's position, velocity, orientation, forces, and other physical parameters involved in the landing phase. This spacecraft dynamics model can output the spacecraft's dynamic response under ideal conditions, providing a basis for subsequent injection and estimation of external disturbances. Specifically, the position parameters include at least the longitude and latitude of the spacecraft's location and its radial distance from the target ground; the velocity parameters include at least the spacecraft's velocity relative to the target ground; the orientation parameters include at least the flight path angle and flight heading angle; the force parameters include at least lift acceleration and drag acceleration; and the physical parameters include at least the spacecraft's total mass and aerodynamic area.

[0073] S2: Based on the real or simulated landing environment, inject several external disturbances into the spacecraft dynamics model to obtain the disturbed state output, which represents the actual response of the spacecraft under the action of external disturbances.

[0074] In this step, external disturbances are injected into the spacecraft dynamics model based on a real or simulated spacecraft landing environment. Specifically, disturbance sources may include, but are not limited to, atmospheric turbulence, terrain impact, and sensor noise. These external disturbances, input into the spacecraft's state equations, affect the spacecraft's trajectory and dynamic characteristics, thereby causing changes in the spacecraft's state.

[0075] The purpose of injecting atmospheric turbulence disturbance is to simulate the impact of atmospheric instability on spacecraft and to simulate the possible fluctuations in spacecraft state parameters.

[0076] The purpose of injecting terrain impact disturbances is to simulate the impact effects of terrain factors such as ground undulations and obstacles on the spacecraft landing process, as well as their impact on the spacecraft's stability.

[0077] Injecting sensor noise disturbance takes into account the noise that may occur during spacecraft sensor measurements, such as sensor systematic errors and random noise, and simulates its impact on the accuracy of observation data.

[0078] S3: Based on capacitive Kalman filtering, singular value decomposition is performed on the state covariance matrix describing the system in the disturbed state to obtain the decomposition result.

[0079] In this step, the capacitive Kalman filter (CKF) algorithm is used to perform singular value decomposition (SVD) on the state covariance matrix of the spacecraft under disturbed conditions.

[0080] Specifically, the singular value decomposition method is used to decompose the system state covariance matrix P. k Obtain the orthogonal matrix U k and diagonal matrix Right now

[0081]

[0082] Where S k The singular values ​​of the system state covariance matrix reflect uncertainties in different directions. Singular value decomposition avoids the numerical instability issues that arise in traditional Kalman filtering's Cholesky decomposition under high perturbations or non-positive definite covariance matrices, ensuring the accuracy and stability of subsequent calculations. The key to this step is using the decomposition results to generate volume points and provide a stable numerical basis for further estimations.

[0083] S4: Use the decomposition results to generate several volume points for the predicted states, and combine them with the spacecraft landing dynamics model to update the disturbed state over time, and calculate the prior state estimate and the prior covariance matrix.

[0084] In this step, based on the SVD decomposition results in step S3, several predicted state volume points are generated. Specifically, volume points are first generated based on the singular values ​​and orthogonal matrices obtained from the covariance matrix decomposition. These volume points represent the possible distribution of the spacecraft's state. The generated volume points will be propagated in the state space to characterize the spacecraft's state changes under uncertainty. The formula for calculating the volume points is as follows:

[0085]

[0086] Among them, X k Let ξ represent the volume point at time k, and the direction set of the volume point. i and weight ω i Determined by the following formula:

[0087]

[0088] Where m is the number of volume points and n is the state dimension, the set of volume points and their corresponding weights are calculated based on the third-order spherical radial volume criterion. The volume points are then input into the spacecraft landing dynamics model. Combined with the given control inputs and disturbance data in the model, time updates are performed. The time update process calculates the prior state estimate and covariance matrix for the next time step by calculating the state of each volume point. Specifically, this includes obtaining the propagation volume point for the next time step through the state transition function f(·).

[0089]

[0090] Where the subscript k+1 / k represents the prediction at time k+1, obtained through propagation volume points. Calculate the prior state estimate and the prior error covariance matrix P k+1 / k The formula for its calculation is:

[0091]

[0092] Among them, Q k Let be the process noise covariance matrix.

[0093] This step allows for a prediction of the spacecraft's disturbed state at the next moment, providing necessary information for subsequent measurement updates.

[0094] S5: Based on the measurement equation, the measurement propagation volume point is mapped to the measurement space, and the measurement estimate, measurement error covariance matrix, and state-measurement interaction covariance matrix are calculated.

[0095] Specifically, firstly, the prior error covariance matrix P obtained in step S4 is... k+1 / k Perform singular value decomposition, in the form of:

[0096]

[0097] Among them, U k+1 / k It is an orthogonal matrix. A diagonal matrix that retains only non-zero singular values;

[0098] Based on the prior error covariance matrix P k+1 / k The decomposition results are used to calculate the prior volume point X. k+1 / k The formula for its calculation is:

[0099]

[0100] The prior volume point X is obtained by measuring the equation H(·). k+1 / k Mapping to the measurement space yields the measurement propagation volume point.

[0101]

[0102] The formula for calculating the measurement estimate is:

[0103]

[0104] The formula for calculating the measurement error covariance matrix is:

[0105]

[0106] Among them, R k To measure the noise covariance matrix;

[0107] The formula for calculating the interaction covariance matrix between state and measurement is:

[0108]

[0109] These statistics will serve as input data for subsequent filtering and updating processes, ensuring real-time estimation of spacecraft status corrections and disturbances.

[0110] S6: Calculate the Kalman gain based on the calculation results of step S5.

[0111]

[0112] By correcting the prior state estimate with Kalman gain and the prior covariance matrix, we obtain the state estimate and the state estimate error covariance matrix at time k+1. The formula for calculating the state estimate at time k+1 is:

[0113]

[0114] The formula for calculating the state estimation error covariance matrix at time k+1 is:

[0115]

[0116] Output the corrected spacecraft status and real-time estimates of several external disturbances, and repeat steps S3 to S6 until landing is completed.

[0117] like Figure 2 As shown, another aspect of the present invention provides a spacecraft landing disturbance estimation device based on the SVDCKF algorithm. Applying the above-described estimation method, the device includes...

[0118] The spacecraft dynamics model module outputs the initial state of the spacecraft based on its position, velocity, orientation, attitude and other state parameters and physical parameters, and updates the dynamic characteristics of the spacecraft during the landing process at any time, providing a basis for subsequent disturbance estimation.

[0119] The disturbance injection module is used to inject external disturbances, including but not limited to atmospheric turbulence, terrain impact, and sensor noise, into the aforementioned spacecraft dynamics model module in a real or simulated landing environment. This module is used to apply disturbances to the spacecraft, obtaining the spacecraft's output data under disturbed conditions, which reflects the spacecraft's true dynamic behavior in complex environments.

[0120] The disturbance estimation module is used to perform time and measurement updates on the above-mentioned disturbed state using the SVDCKF algorithm. It performs singular value decomposition on the state covariance matrix, generates volume points, and completes real-time estimation of the spacecraft state and disturbance parameters, thereby outputting the corrected spacecraft state and external disturbance estimation results.

[0121] Figure 3This is a schematic diagram of the disturbance injection of the present invention. The disturbance injection module combines external disturbances with the spacecraft dynamics model, so that the state of the spacecraft can reflect the actual dynamic behavior in complex environments and provide the disturbed state that needs to be estimated and corrected.

[0122] The target command is the desired state that the spacecraft needs to achieve during landing. This command is received and processed by the controller, which adjusts various control variables of the spacecraft according to the target command to ensure that the spacecraft lands along the predetermined trajectory or landing point. The spacecraft's dynamic state variables reflect its state parameters at a specific moment. The controller calculates the required control variables based on the target command and the spacecraft's current state variables. During this process, the spacecraft dynamic model describes the spacecraft's motion to ensure that it lands along the predetermined trajectory. The aforementioned signal sources are injected into the spacecraft dynamic model in normal and disturbed states via a switching circuit, while the spacecraft dynamic model simultaneously provides real-time feedback of the spacecraft's dynamic state variables to the controller.

[0123] Figure 4 This is a schematic diagram of the disturbance estimation of the present invention. The diagram illustrates the process from the input control quantity to the final output corrected spacecraft state and disturbance estimation. The diagram includes the following main parts:

[0124] The spacecraft dynamics model is used to receive the input control quantity u, to simulate the state of the spacecraft when no external disturbance is injected, and to generate the output state y.

[0125] The spacecraft discrete-time state-space model is based on the spacecraft dynamics model and is transformed into a discrete-time state-space form. Specifically, it includes a state transition function h′ and discrete-time state variables x′, which are used to predict the spacecraft's state at the next moment.

[0126] The system receives the output information from the discrete-time state-space model, combines it with the actual observations y′ of the system, and processes the covariance matrix of the spacecraft state using the SVDCKF algorithm to generate accurate volumetric sampling points. It then updates the spacecraft's state and perturbations in real time. Through iterative optimization of the Kalman filter gain, the deviation between the predicted and observed states is corrected.

[0127] After completing the state and measurement updates, the system outputs an estimate containing the unmeasurable unknown disturbance d′. This result integrates the system input u, the dynamic model h′, and the measurement data y′ to accurately estimate environmental disturbances and other unknown factors during the spacecraft landing process.

[0128] In the practical application architecture of the estimation method and apparatus provided by this invention, all modules run in the Simulink environment, making full use of Simulink's powerful modeling and simulation capabilities. This allows the algorithm to be seamlessly embedded into the entire spacecraft landing disturbance estimation system, ensuring the efficiency and synergy of system integration. It can significantly reduce the complexity of data interaction between different platforms and make debugging and optimization between modules more convenient, thereby greatly improving the operating efficiency of the entire system and providing stable, reliable and efficient technical support for disturbance estimation during spacecraft landing.

[0129] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the invention by those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the invention should be included within the scope of protection of the invention.

Claims

1. A spacecraft landing perturbation estimation method based on the SVDCKF algorithm, characterized in that, Includes the following steps: S1: Establish a spacecraft dynamics model to characterize several state parameters of the spacecraft, including at least position parameters, velocity parameters, orientation parameters, force parameters, and physical parameters, and obtain the initial state output; S2: Based on the real or simulated landing environment, inject several external disturbances into the spacecraft dynamics model to obtain the disturbed state output, which represents the actual response of the spacecraft dynamics model under the action of external disturbances. S3: Based on capacitive Kalman filtering, perform singular value decomposition on the state covariance matrix of the disturbed state to obtain the decomposition result; S4: Using the decomposition results, generate several volume points for predicted states, and combine them with the spacecraft landing dynamics model to update the disturbed state over time, and calculate the prior state estimate and prior covariance matrix. S5: Based on the measurement equation, the measurement propagation volume point is mapped to the measurement space, and the measurement estimate, measurement error covariance matrix, and state-measurement interaction covariance matrix are calculated. S6: Based on the calculation results of step S5, calculate the Kalman gain and correct the prior state estimate and prior covariance matrix. Output the corrected spacecraft state and the real-time estimate of several external disturbances. Repeat steps S3 to S6 until landing is completed.

2. The spacecraft landing perturbation estimation method based on the SVDCKF algorithm as described in claim 1, characterized in that: In step S3, singular value decomposition is performed on the system state covariance matrix, specifically in the form of: Among them, U k It is an orthogonal matrix. It is a diagonal matrix that retains only non-zero singular values.

3. The spacecraft landing perturbation estimation method based on the SVDCKF algorithm as described in claim 2, characterized in that: In step S4, volume points are generated based on the covariance decomposition results in step S3, and the calculation formula is as follows: Among them, the volume point direction set ξ i and weight ω i Determined by the following formula: The propagation volume point at the next time step is obtained through the state transition function f(·):

4. The spacecraft landing perturbation estimation method based on the SVDCKF algorithm as described in claim 3, characterized in that: In step S4, through the propagation volume point Calculate the prior state estimate and the prior covariance matrix P k+1 / k The formula for its calculation is: Among them, Q k Let be the process noise covariance matrix.

5. The spacecraft landing perturbation estimation method based on the SVDCKF algorithm as described in claim 4, characterized in that: In step S5, the prior error covariance matrix P obtained in step S4 is... k+1 / k Perform singular value decomposition, specifically in the form of: Among them, U k+1 / k It is an orthogonal matrix. A diagonal matrix that retains only non-zero singular values; Based on the prior error covariance matrix P k+1 / k The decomposition results are used to calculate the prior volume point X. k+1 / k The formula for its calculation is: The prior volume point X is measured using the equation H(·). k+1 / k Mapping to the measurement space yields the measurement propagation volume point.

6. The spacecraft landing perturbation estimation method based on the SVDCKF algorithm as described in claim 5, characterized in that: In step S5, the formula for calculating the measurement estimate is: The formula for calculating the measurement error covariance matrix is: Among them, R k To measure the noise covariance matrix; The formula for calculating the state-measurement interaction covariance matrix is ​​as follows:

7. The spacecraft landing perturbation estimation method based on the SVDCKF algorithm as described in claim 6, characterized in that: In step S6, the Kalman gain is calculated as follows: The state estimate and state estimate error covariance matrix at time k+1 are obtained from the Kalman gain-corrected prior state estimate and the prior covariance matrix. The formula for calculating the state estimate at time k+1 is: The formula for calculating the state estimation error covariance matrix at time k+1 is:

8. The spacecraft landing perturbation estimation method based on the SVDCKF algorithm as described in claim 1, characterized in that: In step S1, The location parameters include at least the longitude, latitude, and mirror distance of the spacecraft's location from the target ground. The velocity parameters include at least the spacecraft's velocity relative to the target ground; The directional parameters include at least the flight track angle and the flight heading angle; The force parameters include at least lift acceleration and drag acceleration; The physical parameters include at least the total mass and aerodynamic area of ​​the spacecraft.

9. The spacecraft landing perturbation estimation method based on the SVDCKF algorithm as described in claim 1, characterized in that: In step S2, the external disturbances include at least atmospheric turbulence disturbances, impact disturbances caused by terrain undulations, and sensor noise disturbances.

10. A spacecraft landing disturbance estimation device based on the SVDCKF algorithm, employing the spacecraft landing disturbance estimation method based on the SVDCKF algorithm as described in any one of claims 1-9, characterized in that: include The spacecraft dynamics model module outputs the initial state of the spacecraft and updates the dynamic characteristics of the spacecraft during the landing process based on the spacecraft's position, velocity, orientation, attitude and other state parameters and physical parameters. The disturbance injection module is used to inject external disturbances into the spacecraft dynamics model module in a real or simulated landing environment to obtain the output of the spacecraft under the disturbed state. The disturbance estimation module is used to perform time and measurement updates on the disturbed state using the SVDCKF algorithm. It performs singular value decomposition on the state covariance matrix, generates volume points, and completes real-time estimation of the spacecraft state and disturbance parameters, thereby outputting the corrected spacecraft state and external disturbance estimation results.

Images