Space target fusion positioning precision improvement method and system based on initialization optimization

Abstract

The invention provides a space target fusion positioning precision improvement method and system based on initialization optimization, and the method comprises the steps: carrying out the initialization of a target prediction value at a current moment when a target is converted from single-star observation to multi-star observation or the precision of the target is in a divergence trend; the initialization is to initialize the prediction position, the process noise covariance matrix and the expected covariance matrix of the target at the current moment; in the step of initializing the predicted position of the target at the current moment, the three-dimensional position of the target is calculated by adopting a line-of-sight cross positioning method according to the visual vector of multiple satellites to the target at the current moment, and the predicted position of the target at the current moment is replaced. According to the method, the target prediction position is initialized by using the multi-star sight line result, the target position is revised, and the process noise covariance matrix and the expected covariance matrix are initialized, so that the technical problem of accumulative increase of trajectory prediction errors caused by high maneuverability of an active section and incomplete matching of a target dynamic model is solved.

Description

Technical Field

[0001] This invention belongs to the field of on-board information fusion technology, specifically relating to a method and system for improving the positioning accuracy of space targets based on initialization optimization. Background Technology

[0002] The high maneuverability of the active phase and the imperfect matching of the target dynamics model lead to a cumulative increase in trajectory prediction error.

[0003] Reference 1, "Liu Zhiyong, A Target Tracking Method Based on Self-Constructed Fuzzy EKF, Computational Technology and Automation, Vol. 41, No. 3, 2022," discloses a target tracking method based on self-constructed fuzzy EKF. First, a control model of the UAV system is established based on the dynamic equations of the UAV system, and the high-frequency random interference signals existing in the UAV control system are treated as observation interference noise. Second, a self-constructed fuzzy extended Kalman filter is designed, using the error between the EKF estimated variance and the actual observation variance as an input to the self-constructed fuzzy system, which then identifies the error.

[0004] The method in Reference 1 is for the problem of UAV flight trajectory tracking and positioning, but it does not provide a method to improve the target fusion positioning accuracy when the target changes from single-satellite observation to multi-satellite observation or when the target accuracy shows a divergent trend.

[0005] Reference 2, "Zhao Jubo, Space-Based Two-Star Stereo Astronomical Positioning of Space Targets, Optics and Precision Engineering, Vol. 29, No. 12, 2021," discloses a space-based two-star stereo astronomical positioning method. First, based on the inherent parameters of the optical payload, the imaging characteristics of the space target on the optical sensor are analyzed, and the threshold centroid method is selected to accurately extract the target's position on the two-dimensional image plane. Next, based on the full-link coordinate projection transformation relationship from the target to the observation sensor, an observation vector model of the target in the Earth's inertial coordinate system is established. Then, combined with the least squares criterion, a two-star stereo astronomical geometric positioning model is established, completing the projection transformation of the space target from two-dimensional image information to three-dimensional spatial information. Finally, a ground experiment is built to generate a starry sky image containing the space target, and the positioning algorithm is verified and error simulation analysis is completed.

[0006] The method in Reference 2 is designed for positioning space targets using two satellites and is not suitable for single-satellite detection. It does not provide a method to improve the accuracy of target fusion positioning when the target is transitioned from single-satellite observation to multi-satellite observation or when the target accuracy shows a divergent trend.

[0007] Reference 3, "Gao Chunyan, Application of EKF Interactive Multi-Model Algorithm in Target Tracking, Mechanical Design and Manufacturing, No. 2, 2020," discloses an extended Kalman filter interactive multi-model algorithm, EKF-IMM. This algorithm is based on the interactive multi-model algorithm and incorporates the EKF algorithm for filtering.

[0008] The method in Reference 3 addresses the problem of noise in sensor perception information and sudden changes in motion trajectory during moving target tracking, which can lead to target observation distortion or even loss. However, it does not provide a method to improve the target fusion positioning accuracy when the target shifts from single-satellite observation to multi-satellite observation or when the target accuracy shows a divergent trend.

[0009] Reference 4, "Ye Zehao, Application of Improved Square Root UKF in Reentry Gliding Target Tracking, Journal of Astronautics, Vol. 40, No. 2, 2019," discloses an improved square root UKF filtering algorithm, ISR-UKF, based on a new aerodynamic model. First, the aerodynamic model is transformed. Second, based on the traditional square root UKF, a spherical unscented transformation is used to calculate the weight coefficients and sigma points; the decomposition method of the square root matrix is ​​improved; and to address the problem of singular values ​​easily arising during matrix inversion causing filter failure, a multi-order stabilization factor is introduced in the covariance matrix update. Finally, the algorithm is compared with ISR-UKF based on the original aerodynamic model, the square root UKF based on the new aerodynamic model, and the square root UKF based on the original aerodynamic model through simulation.

[0010] Reference 4 addresses the reentry gliding target tracking problem during dual-satellite observations, but does not provide a method to improve target fusion positioning accuracy when the target shifts from single-satellite observation to multi-satellite observation or when the target accuracy shows a divergent trend.

[0011] Reference 5, "Luo Zhaoxian, Research on Improved Algorithm for Multi-Target Tracking in Dual-Satellite Detection, Electro-Optics and Control, Vol. 25, No. 4, 2018," discloses an improved algorithm for multi-target tracking in dual-satellite detection. This algorithm studies the tracking problem of multiple targets in dual-satellite cooperative detection. First, based on the observability analysis of dual-satellite cooperative detection, the active segment state equation and observation equation based on the gravity turning model are established. Then, the BPGM-SME algorithm based on the idea of ​​bivariate polynomials is proposed to solve the track intersection tracking anomaly problem in the case of multi-target tracking. On this basis, in order to improve the single-target tracking accuracy and convergence speed, an improved unscented Kalman filter algorithm based on the iterative idea is proposed.

[0012] Reference 5 describes a fusion positioning method for targets observed by two satellites, but it does not provide a method to improve the accuracy of target fusion positioning when the target is switched from single-satellite observation to multi-satellite observation or when the target accuracy shows a diverging trend.

[0013] Patent document CN109597841A utilizes the positioning accuracy of a single satellite observation and the nominal overall positioning accuracy to establish an error model. It standardizes the residuals to correct the weight function and experimentally determines the optimal harmonic coefficient through satellite simulation data and actual observation data. It eliminates or reduces the weight of observations with abnormal error distributions and optimizes the positioning results of repeated target observations to improve its positioning accuracy.

[0014] None of the above methods can improve the accuracy of space target fusion positioning when the target shifts from single-satellite observation to multi-satellite observation or when the target accuracy shows a divergent trend. Therefore, a new technical solution is needed to improve the above technical problems. Summary of the Invention

[0015] To address the shortcomings of existing technologies, the purpose of this invention is to provide a method and system for improving spatial target fusion positioning accuracy based on initialization optimization.

[0016] A method for improving spatial target fusion positioning accuracy based on initialization optimization according to the present invention includes:

[0017] First initialization step: When the target is switched from single-star observation to multi-star observation, initialize the target prediction value at the current moment;

[0018] or

[0019] The second initialization step: When the target accuracy shows a divergent trend, initialize the target prediction value at the current moment.

[0020] Preferably, the initialization involves initializing the predicted position of the target at the current moment, the process noise covariance matrix, and the expected covariance matrix.

[0021] The initialization of the target's predicted position at the current moment is achieved by calculating the target's three-dimensional position using the line-of-sight cross-location method based on the target's view vectors from multiple stars at the current moment, and then replacing the target's predicted position at the current moment.

[0022] The initialization of the process noise covariance matrix and the expected covariance matrix is ​​to replace the current values ​​with the initial values ​​of the process noise covariance matrix and the expected covariance matrix.

[0023] Preferably, the multi-satellite observation refers to receiving data on the same target from multiple satellites at the same or similar times; wherein, the similar times refer to less than half of the shortest satellite imaging time interval.

[0024] The term "single-satellite observation to multi-satellite observation" refers to receiving target data from one satellite in the previous moment and receiving the same target data from multiple satellites in the current moment.

[0025] Preferably, the target accuracy refers to the difference between the target fused position and velocity and the target's true value;

[0026] The divergent trend in target accuracy means that the errors in target position and velocity gradually increase over time.

[0027] Preferably, it includes:

[0028] Estimation steps: The extended Kalman filter algorithm is used to estimate the target position and velocity at the current moment.

[0029] A spatial target fusion positioning accuracy improvement system based on initialization optimization, provided by the present invention, includes:

[0030] First initialization module: When the target is switched from single-star observation to multi-star observation, the target prediction value at the current moment is initialized;

[0031] or

[0032] The second initialization module initializes the target prediction value at the current moment when the target accuracy shows a divergent trend.

[0033] Preferably, the initialization involves initializing the predicted position of the target at the current moment, the process noise covariance matrix, and the expected covariance matrix.

[0034] The initialization of the target's predicted position at the current moment is achieved by calculating the target's three-dimensional position using the line-of-sight cross-location method based on the target's view vectors from multiple stars at the current moment, and then replacing the target's predicted position at the current moment.

[0035] The initialization of the process noise covariance matrix and the expected covariance matrix is ​​to replace the current values ​​with the initial values ​​of the process noise covariance matrix and the expected covariance matrix.

[0036] Preferably, the multi-satellite observation refers to receiving data on the same target from multiple satellites at the same or similar times; wherein, the similar times refer to less than half of the shortest satellite imaging time interval.

[0037] The term "single-satellite observation to multi-satellite observation" refers to receiving target data from one satellite in the previous moment and receiving the same target data from multiple satellites in the current moment.

[0038] Preferably, the target accuracy refers to the difference between the target fused position and velocity and the target's true value;

[0039] The divergent trend in target accuracy means that the errors in target position and velocity gradually increase over time.

[0040] Preferably, it includes:

[0041] Estimation module: The extended Kalman filter algorithm is used to estimate the target position and velocity at the current moment.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. This invention uses multi-star line-of-sight results to initialize the target prediction position and revise the target position.

[0044] 2. The noise covariance matrix and expected covariance matrix of the initialization process of this invention solve the technical problem that the trajectory prediction error accumulates due to the strong maneuverability of the active segment and the incomplete matching of the target dynamic model.

[0045] 3. The method of the present invention is applicable when the target is shifted from single-satellite observation to multi-satellite observation or when the target accuracy shows a divergent trend, thereby improving the fusion positioning accuracy of space targets. Attached Figure Description

[0046] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0047] Figure 1 This is a flowchart illustrating a method for improving spatial target fusion positioning accuracy based on initialization optimization.

[0048] Figure 2 A comparison curve of the target position error before and after initialization;

[0049] Figure 3 A graph showing the comparison of target velocity error before and after initialization. Detailed Implementation

[0050] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0051] This invention uses multi-satellite line-of-sight results to initialize the predicted target position, revise the target position, and initialize the process noise covariance matrix and the expected covariance matrix. When the target is switched from single-satellite observation to multi-satellite observation or the target accuracy shows a divergent trend, the accuracy of target fusion positioning is improved.

[0052] Specifically, embodiments of the present invention provide a method for improving spatial target fusion positioning accuracy based on initialization optimization, referring to... Figure 1 The flowchart in the document includes the following steps:

[0053] Step 1: Based on the input satellite observation data, determine whether the target is being observed by multiple satellites at the current moment, and obtain the judgment result;

[0054] Step 2: Based on the judgment result, access the target trajectory database to determine whether the target at the current moment is transitioning from single-satellite observation to multi-satellite observation;

[0055] Step 3: If the target at the current moment is transitioning from single-satellite observation to multi-satellite observation, then initialize the target prediction value at the current moment;

[0056] Step 4: If the target is currently transitioning from single-star observation to multi-star observation, determine whether the target accuracy shows a divergent trend.

[0057] Step 5: If the target accuracy shows a divergent trend, initialize the target prediction value at the current moment;

[0058] Step 6: Calculate the target's three-dimensional position and velocity at the current moment.

[0059] Furthermore, in step 1: multi-satellite observation refers to receiving data on the same target from multiple satellites at the same or similar time (less than half the shortest satellite imaging time interval).

[0060] Furthermore, in step 2: switching from single-satellite observation to multi-satellite observation means that the target data observed by one satellite was received at the previous moment, and the same target data observed by multiple satellites was received at the current moment.

[0061] Furthermore, in step 3: initialization is to initialize the predicted position of the target at the current time, the process noise covariance matrix, and the expected covariance matrix.

[0062] Furthermore, the initialization of the predicted position of the target at the current moment is achieved by calculating the three-dimensional position of the target using the line-of-sight cross-location method based on the view vectors of the target from multiple satellites at the current moment, and replacing the predicted position of the target at the current moment.

[0063] Furthermore, the initialization process noise covariance matrix and expected covariance matrix are obtained by replacing the current values ​​with the initial values ​​of the process noise covariance matrix and expected covariance matrix.

[0064] Furthermore, in step 4: target accuracy refers to the difference between the target fusion position and velocity and the target's true value.

[0065] Furthermore, in step 4: the divergence trend of target accuracy means that the target position and velocity errors gradually increase over time.

[0066] Furthermore, in step 5: initialization is to initialize the predicted position of the target at the current time, the process noise covariance matrix, and the expected covariance matrix.

[0067] Furthermore, in step 6: the extended Kalman filter (EKF) algorithm is used to estimate the target position and velocity at the current moment.

[0068] The following description, in conjunction with the accompanying drawings, further illustrates this embodiment, which satisfies the requirement to improve the fusion positioning accuracy of space targets when the target shifts from single-satellite observation to multi-satellite observation or when the target accuracy shows a divergent trend.

[0069] See Figure 1 , Figure 2 and Figure 3 See the description below for details:

[0070] Assuming two satellites simultaneously observe a space target, at the 40th second of observation, the target is switched from single-satellite observation to dual-satellite observation, and the target is initialized. At the 60th second of observation, when the target accuracy shows a diverging trend, the target is initialized again. A comparison of the fused positioning accuracy between the uninitialized and initialized targets is shown below. Figure 2 and Figure 3 As shown.

[0071] Depend on Figure 2 and Figure 3 It can be seen that initialization improves the target fusion positioning accuracy when the target is switched from single-star observation to dual-star observation or when the target accuracy shows a divergent trend.

[0072] This embodiment achieves the following beneficial effects:

[0073] (1) This invention uses multi-star line-of-sight results to initialize the target prediction position, revise the target position, and initialize the process noise covariance matrix and the expected covariance matrix, which solves the technical problem that the trajectory prediction error accumulates due to the strong maneuverability of the active segment and the incomplete matching of the target dynamics model.

[0074] (2) The method of the present invention is applicable to improving the fusion positioning accuracy of space targets when the target is changed from single-star observation to multi-star observation or the target accuracy shows a divergent trend.

[0075] The present invention also provides a spatial target fusion positioning accuracy improvement system based on initialization optimization. The spatial target fusion positioning accuracy improvement system based on initialization optimization can be implemented by executing the process steps of the spatial target fusion positioning accuracy improvement method based on initialization optimization. That is, those skilled in the art can understand the spatial target fusion positioning accuracy improvement method based on initialization optimization as a preferred embodiment of the spatial target fusion positioning accuracy improvement system based on initialization optimization.

[0076] A spatial target fusion positioning accuracy improvement system based on initialization optimization, provided by the present invention, includes:

[0077] The first initialization module initializes the target prediction value at the current moment when the target observation transitions from single-satellite observation to multi-satellite observation. Multi-satellite observation refers to receiving the same target data from multiple satellites at the same or similar time. The similar time refers to less than half of the shortest satellite imaging time interval. The transition from single-satellite observation to multi-satellite observation means that the target data was received from one satellite at the previous time and the same target data was received from multiple satellites at the current time.

[0078] or

[0079] The second initialization module initializes the target prediction value at the current moment when the target accuracy shows a diverging trend. The target accuracy refers to the difference between the fused target position and velocity and the true target value; the diverging trend in target accuracy means that the target position and velocity errors gradually increase over time.

[0080] The initialization involves initializing the predicted position, process noise covariance matrix, and expected covariance matrix of the target at the current moment. The initialization of the predicted position of the target at the current moment is achieved by calculating the three-dimensional position of the target using the line-of-sight cross-location method based on the view vectors of the target from multiple satellites at the current moment, and replacing the predicted position of the target at the current moment with the calculated position. The initialization of the process noise covariance matrix and the expected covariance matrix is ​​achieved by replacing the current values ​​of the process noise covariance matrix and the expected covariance matrix with their initial values.

[0081] The space target fusion positioning accuracy improvement system based on initialization optimization also includes:

[0082] Estimation module: The extended Kalman filter algorithm is used to estimate the target position and velocity at the current moment.

[0083] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0084] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for improving spatial target fusion positioning accuracy based on initialization optimization, characterized in that, include: First initialization step: When the target is switched from single-star observation to multi-star observation, initialize the target prediction value at the current moment; or The second initialization step: When the target accuracy shows a divergent trend, initialize the target prediction value at the current moment.

2. The method for improving spatial target fusion positioning accuracy based on initialization optimization according to claim 1, characterized in that, The initialization involves initializing the target's predicted position, process noise covariance matrix, and expected covariance matrix at the current moment. The initialization of the target's predicted position at the current moment is achieved by calculating the target's three-dimensional position using the line-of-sight cross-location method based on the target's view vectors from multiple stars at the current moment, and then replacing the target's predicted position at the current moment. The initialization of the process noise covariance matrix and the expected covariance matrix is ​​to replace the current values ​​with the initial values ​​of the process noise covariance matrix and the expected covariance matrix.

3. The method for improving spatial target fusion positioning accuracy based on initialization optimization according to claim 1, characterized in that, The multi-satellite observation refers to receiving data on the same target from multiple satellites at the same or similar times; where "similar times" means less than half of the shortest satellite imaging time interval. The term "single-satellite observation to multi-satellite observation" refers to receiving target data from one satellite in the previous moment and receiving the same target data from multiple satellites in the current moment.

4. The method for improving spatial target fusion positioning accuracy based on initialization optimization according to claim 1, characterized in that, The target accuracy refers to the difference between the fused target position and velocity and the target's true value; The divergent trend in target accuracy means that the errors in target position and velocity gradually increase over time.

5. The method for improving spatial target fusion positioning accuracy based on initialization optimization according to claim 1, characterized in that, include: Estimation steps: The extended Kalman filter algorithm is used to estimate the target position and velocity at the current moment.

6. A spatial target fusion positioning accuracy improvement system based on initialization optimization, characterized in that, include: First initialization module: When the target is switched from single-star observation to multi-star observation, the target prediction value at the current moment is initialized; or The second initialization module initializes the target prediction value at the current moment when the target accuracy shows a divergent trend.

7. The spatial target fusion positioning accuracy improvement system based on initialization optimization according to claim 6, characterized in that, The initialization involves initializing the target's predicted position, process noise covariance matrix, and expected covariance matrix at the current moment. The initialization of the target's predicted position at the current moment is achieved by calculating the target's three-dimensional position using the line-of-sight cross-location method based on the target's view vectors from multiple stars at the current moment, and then replacing the target's predicted position at the current moment. The initialization of the process noise covariance matrix and the expected covariance matrix is ​​to replace the current values ​​with the initial values ​​of the process noise covariance matrix and the expected covariance matrix.

8. The spatial target fusion positioning accuracy improvement system based on initialization optimization according to claim 6, characterized in that, The multi-satellite observation refers to receiving data on the same target from multiple satellites at the same or similar times; where "similar times" means less than half of the shortest satellite imaging time interval. The term "single-satellite observation to multi-satellite observation" refers to receiving target data from one satellite in the previous moment and receiving the same target data from multiple satellites in the current moment.

9. The spatial target fusion positioning accuracy improvement system based on initialization optimization according to claim 6, characterized in that, The target accuracy refers to the difference between the fused target position and velocity and the target's true value; The divergent trend in target accuracy means that the errors in target position and velocity gradually increase over time.

10. The spatial target fusion positioning accuracy improvement system based on initialization optimization according to claim 6, characterized in that, include: Estimation module: The extended Kalman filter algorithm is used to estimate the target position and velocity at the current moment.

Images