A method for hierarchical early warning of space target collisions based on rapid screening

Abstract

This invention discloses a method, apparatus, and device for graded early warning of space target collisions based on rapid screening. The method includes: constructing a graded early warning template containing target size parameters and collision probability thresholds; standardizing a single set of TLE orbital data and constructing a list of potential rendezvous events; employing a multi-layered filtering strategy from coarse to fine, sequentially using orbital envelope, orbital plane difference, and time window coarse propagation distance upper bound criteria to eliminate risk-free targets, generating a candidate rendezvous pair list; performing precise evaluation on candidate targets to calculate collision probabilities; and finally, combining the graded early warning template to output graded early warning results. This invention significantly reduces the computational scale through a multi-layered rapid screening strategy, effectively solving the combinatorial explosion problem in large-scale space target early warning and significantly improving the timeliness of early warnings; simultaneously, by combining a graded judgment mechanism based on target size, it avoids alarm imbalance caused by a single threshold, improving the engineering practicality of the early warning results.

Description

Technical Field

[0001] This application relates to the field of collision warning technology, and in particular to a method, apparatus and equipment for graded early warning of space target collisions based on rapid screening. Background Technology

[0002] With the increasing frequency of space activities, the number of spacecraft and space debris in orbit continues to grow. According to publicly available data, the number of categorizable space targets has reached hundreds of thousands, and the number of debris larger than centimeters with damaging capabilities is even greater. Faced with such a dense group of targets, when conducting collision warnings for specific spacecraft, if high-precision orbital predictions are performed for all targets individually, and the closest approach time and minimum distance are calculated for each pair, the computational scale will increase explosively with the number of targets. Even predicting only the next few days may result in unbearable computational overhead, thus affecting the timeliness and update frequency of the warning results. Therefore, how to quickly narrow down the candidate rendezvous pair set using low-cost methods while ensuring the reliability of risk identification, and how to classify and output the risk level, has become a key issue in engineering applications. Existing technology proposes a space debris collision risk level prediction method, which extracts features from rendezvous event data and performs supervised learning to classify and distinguish risk levels. However, this type of method requires several days of historical orbital data and is not suitable for short-term, rapid calculation of collision probabilities. On the other hand, patent CN115578889A, prior art 2, proposes a collision warning process for a large-scale target library. It improves computational efficiency by setting up preliminary screening and fine screening steps, but it focuses more on the computational framework and parameterized screening implementation. Summary of the Invention

[0003] The main objective of this application is to provide a method, apparatus, and device for hierarchical early warning of space target collisions based on rapid screening, which aims to combine target scale with collision risk probability to form an interpretable hierarchical early warning output.

[0004] To achieve the above objectives, this application provides a space target collision classification early warning method based on rapid screening, comprising: A hierarchical early warning template is constructed, which includes the protected object, the intersecting target, the target size parameters, the equivalent collision radius, the collision probability threshold, and the early warning level output; wherein, the early warning level output is determined based on the target size classification and the collision probability threshold. The input space target orbit data is standardized and quality controlled. The target size and position uncertainty parameters are supplemented by the target catalog information to form a list of key target states. A list of potential rendezvous events is also constructed within the warning time window. A multi-layered filtering strategy, from coarse to fine, is used to screen target pairs in the potential rendezvous event list. Target pairs that are unlikely to approach are eliminated by using orbital envelope, orbital plane difference, and upper bound of coarse propagation distance within the time window as criteria, thus generating a candidate rendezvous pair list. For target pairs in the candidate rendezvous pair list, perform actuarial propagation and search for the nearest approach point, calculate the nearest approach time, minimum distance and relative velocity, establish a characterization of relative position uncertainty, and calculate the collision probability of the rendezvous event under the constraint of equivalent collision radius; Based on the constructed hierarchical early warning template, the collision probability and the acquired target size information are used to determine the level of the intersection event and output a hierarchical early warning result including the risk level.

[0005] Optionally, the input space target orbit data is standardized and quality controlled, including: The TLE data is parsed to extract the NORAD number, name, epoch time, and orbital root number, and the units and fields of the orbital root number are standardized. Query the target directory to obtain the target's equivalent size and location standard deviation. If the information is missing, fill it in using the default parameters. The key states of the targets are combined in pairs and encapsulated into key states of rendezvous events that include target pairs, time windows, orbital parameters and uncertainty parameters.

[0006] Optionally, a multi-layered filtering strategy, from coarse to fine, is employed to filter target pairs in the potential intersection event list, including: The first layer is orbital envelope screening, the second layer is orbital plane screening, and the third layer is time window coarse propagation screening; The first-level orbital envelope screening includes: Calculate the perigee and apogee altitude ranges of the two targets in the target pair; Determine whether the perigee and apogee height ranges of two targets overlap. If there is no overlap, determine that the target pair is unlikely to approach and eliminate it. The second-level track plane filtering includes: Calculate the difference in orbital inclination and the difference in right ascension of the ascending node between the two targets in the target alignment; Determine whether the difference in orbital inclination and the difference in right ascension of the ascending node are both greater than a preset threshold. If so, determine that the target is unapproachable due to the excessive difference in orbital plane space and eliminate it. The third-level time window coarse propagation screening includes: A low-precision orbit model and a coarse sampling time step are used to propagate and sample the target position within the warning time window; The minimum distance upper bound of the target pair formed by the two targets within the warning time window is dynamically estimated. If the minimum distance upper bound is greater than the preset screening threshold, the target pair is determined to be unlikely to approach and is removed. If the minimum distance during propagation is less than the preset screening threshold, the propagation is terminated early and the pair is identified as a candidate intersection pair.

[0007] Optionally, perform actuarial propagation and nearest proximity search on target pairs in the candidate intersection pair list, including: The orbital status of candidate rendezvous pairs in the candidate rendezvous pair list is synchronously propagated within the early warning time window using a precisely calculated step size. The system iterates through and calculates the distance between two targets, searches for and records the closest approach time, the minimum distance, and the relative velocity at that time to complete the search for the closest approach point.

[0008] Optionally, a characterization of relative position uncertainty is established, including: Construct the three-dimensional position error covariance matrix of a single target based on the isotropic assumption; By superimposing the position error covariance matrices of the two targets, a unified position error model based on the relative position error covariance is obtained. A uniform position error model is used for each target to establish a characterization of the relative position uncertainty of each target; The position error model sets the position standard deviation to increase linearly with the time difference between the epoch of the orbital data and the closest approach time.

[0009] Optionally, the calculation of the equivalent collision radius includes: Obtain the equivalent diameter of the protected object and the equivalent diameter of the intersecting target; If the equivalent diameter of the intersecting target is unknown, the default size will be used. Calculate the equivalent collision radius, which is half the sum of the equivalent diameter of the protected object and the equivalent diameter of the intersecting target.

[0010] Optionally, a level determination is made for the intersection event, including: Based on the size parameters of the rendezvous targets, the targets are classified into corresponding size categories, which include at least micro target category, small target category, medium target category, large target category, and unknown size category; Obtain the set of collision probability thresholds corresponding to this size category; The calculated collision probability is compared with the corresponding threshold. If the collision probability is greater than or equal to the threshold, the corresponding level of warning is triggered.

[0011] Optionally, for targets with unknown dimensions, the default dimensions are used to classify them into the default category, and a collision probability threshold is set according to a conservative principle for early warning determination; the output graded early warning results also include the closest approach time, minimum distance, relative speed, and traceable data identifier.

[0012] Furthermore, to achieve the above objectives, this application also provides a space target collision classification early warning device based on rapid screening, comprising: The template construction module is configured to build a hierarchical early warning template that includes the protected object, the intersection target, the target size parameters, the equivalent collision radius, the collision probability threshold, and the early warning level output; wherein, the early warning level output is determined based on the target size classification and the collision probability threshold. The rendezvous event list construction module is configured to include standardizing and quality-controlling the input space target orbit data, supplementing the target size and position uncertainty parameters with the target catalog information, forming a target key status list, and constructing a potential rendezvous event list within the warning time window; The rendezvous pair list construction module is configured to use a multi-layer filtering strategy from coarse to fine to filter target pairs in the potential rendezvous event list. It uses orbital envelope, orbital plane difference and time window coarse propagation distance upper bound criteria to eliminate target pairs that are unlikely to approach each other and generate a candidate rendezvous pair list. The collision probability calculation module is used to perform precise propagation and nearest approach point search on target pairs in the candidate rendezvous pair list, calculate the nearest approach time, minimum distance and relative velocity, establish a relative position uncertainty characterization, and calculate the collision probability of the rendezvous event under the constraint of equivalent collision radius. The graded early warning module is used to determine the grade of a meeting event based on the constructed graded early warning template, the collision probability, and the acquired target size information, and output a graded early warning result that includes the risk level.

[0013] To achieve the above objectives, this application also provides an electronic device, which includes: at least one processor, a memory, and an input / output unit; wherein the memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the space target collision classification early warning method based on rapid screening provided in any of the foregoing embodiments.

[0014] This application proposes a method, apparatus, and device for graded early warning of space target collisions based on rapid screening. Addressing the combinatorial explosion problem caused by a large-scale space target catalog, this invention rapidly eliminates the vast majority of target pairs that are unlikely to approach each other before proceeding to the costly actuarial assessment. This approach transforms the computational problem of combining all pairs of targets into an actuarial problem targeting only a small number of candidate intersection pairs, significantly reducing the computational scale. This allows for rapid risk assessment within a short-term early warning window based on a single set of Time-Like Point (TLE) data, meeting the high-frequency update requirements of the early warning results. This invention innovatively designs a graded early warning template that combines target size parameters with collision probabilities. By introducing target size tiers and configuring differentiated collision probability thresholds, it avoids the imbalance in alarm strategies caused by using the same threshold for small and large targets (e.g., false alarms for small targets or missed alarms for large targets). The output graded early warning results not only include the risk level but also key evidence fields and traceable data identifiers, making the early warning results more intuitive and interpretable, facilitating operators to formulate subsequent response strategies (e.g., evasive maneuvers) based on different risk levels. This invention establishes a standardized data quality control and uncertainty characterization mechanism. This method ensures that the system can still output quantitative and reasonable risk assessment results even under imperfect input data, guaranteeing the system's robustness. This invention performs actuarial propagation and nearest-approach point search on target pairs in the candidate intersection pair list, calculates the nearest approach time, minimum distance, and relative velocity, and establishes a characterization of relative position uncertainty. Under the constraint of equivalent collision radius, it calculates the collision probability of intersection events. Employing a serial processing architecture combining fast screening and actuarial calculation, it leverages the efficiency of fast screening while retaining the accuracy of actuarial evaluation, effectively resolving the contradiction between efficiency and accuracy inherent in traditional methods. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating an embodiment of the space target collision classification and early warning method based on rapid screening in this application; Figure 2 This is a schematic diagram of the orbital model provided in an embodiment of the space target collision classification and early warning method based on rapid screening in this application; Figure 3 This is a schematic diagram of the orbital propagation results provided in an embodiment of the space target collision classification and early warning method based on rapid screening in this application.

[0016] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0017] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0018] To achieve rapid collision risk assessment and graded early warning for a large-scale space target catalog, and given the rapidly increasing number of targets and the increasing requirements for early warning update frequency, this invention provides a space target collision graded early warning method based on rapid screening. This method uses only a single set of TLE orbit data at the current moment as input. For a short-term warning window of 3 days, it first standardizes and quality-controls the input spatial target orbit data, completing target identification association, time-scale consistency, and outlier removal to form a target orbit state set suitable for batch calculation. Second, orbit propagation is performed within the warning window, and a rapid screening mechanism from coarse to fine is constructed. Low-cost criteria such as orbit envelope, orbital plane difference, and distance upper bound are used to quickly eliminate target pairs that are unlikely to approach, thus transforming the problem of pairwise combinations into a precise evaluation of the candidate rendezvous pair set. Based on this, precise propagation and nearest-approach point search are performed on the candidate target pairs, calculating rendezvous geometry such as the nearest approach time, minimum distance, and relative velocity. Furthermore, a characterization of relative position uncertainty is established, and the collision probability of rendezvous events is calculated under hardware radius constraints, achieving a probabilistic assessment of rendezvous risk. Finally, target size information and collision probability thresholds are combined to construct a hierarchical warning criterion, outputting hierarchical warning results including risk level, key evidence fields, and traceable data identifiers. It also supports incremental reassessment and rapid updating of warning results under data update conditions. This method balances computational efficiency under large-scale catalog conditions with the interpretability of hierarchical results, and is suitable for rapid hierarchical early warning application scenarios based on single-moment data updates.

[0019] Example 1 Reference Figure 1 The first embodiment of this application provides a space target collision classification early warning method based on rapid screening. This method uses only a single set of TLE orbital data at the current moment as input. The method can be executed by a processor, which can be located on a server or a terminal device. The method includes the following steps: S101. Design of a hierarchical early warning template based on space target collision: Construct a hierarchical early warning template that includes the protected object, the intersecting target, the target size parameters, the equivalent collision radius, the collision probability threshold, and the early warning level output; wherein, the early warning level output is determined based on the target size classification and the collision probability threshold. The graded early warning template is used to uniformly express the key parameters of space target rendezvous events and output the early warning level based on the target size and collision probability. The graded early warning template consists of six parts: protected object, rendezvous target, target size parameters, equivalent collision radius, collision probability threshold, and early warning level output.

[0020] Specifically, the protected object refers to the on-orbit spacecraft target that requires safety protection and early warning, denoted by S, where Si represents the i-th protected object.

[0021] A rendezvous target refers to a spatial target that has a potential close approach to the protected object, denoted by F, where Fij represents the j-th rendezvous target of the protected object Si.

[0022] The target size parameter refers to the equivalent diameter or equivalent size used to characterize the physical scale of the target, denoted by D. Dij represents the size parameter of the intersection target Fij. When the target size is known, it is directly assigned a value. When the target size is unknown, the default size D0 is used and its source is marked as the default value.

[0023] The equivalent collision radius refers to the radius of the hard object used for collision probability calculation. It is denoted by H. Hij represents the equivalent collision radius of the intersection pair Si and Fij. It can be calculated from the equivalent diameter Ds of the protected object and the target size Dij. For example, Hij is equal to 0.5 multiplied by Ds plus Dij. It also applies when Dij is the default value D0.

[0024] Collision probability threshold refers to the set of probability thresholds corresponding to different warning levels, denoted by P, where Pk represents the probability threshold corresponding to the k-th level. The threshold can be varied according to the target size to avoid the imbalance of alarm strategy caused by using the same threshold for small and large targets.

[0025] The warning level output refers to the result of classifying the rendezvous event according to the collision probability and target size, denoted by L, where Lij represents the warning level of the rendezvous pair Si and Fij.

[0026] In summary, under the hierarchical early warning template proposed in this invention, a single rendezvous event can be characterized as Eij. Eij is equal to the combination of S, F, D, H, Pc, and L, where Pc is the collision probability assessment result of the rendezvous pair.

[0027] Table 1 Basic Elements of Collision Classification Early Warning Template

[0028] For ease of engineering application, the target size classification and probability threshold can be configured using the template shown in Table 2. When the target size is unknown, the default size D0 is used and assigned to the default classification, and the threshold is set according to a conservative principle.

[0029] Table 2 Collision Probability Grading Threshold Templates by Target Size

[0030] D1, D2, D0, P0, P1, P2, and P3 are configurable parameters. In actual implementation, it can be further expanded into a five-level rating table, and the probability threshold corresponding to each level can be set as a set of multiple thresholds to achieve more granular hierarchical early warning output.

[0031] Understandably, by pre-designing hierarchical templates that include target size and probability thresholds, the processor can expand the early warning results from a single distance criterion to a multi-dimensional risk assessment that combines the target's physical characteristics, thus enabling refined configuration of early warning strategies.

[0032] Clearly, this step, by constructing a standardized hierarchical early warning template, provides a benchmark for subsequently converting collision probability calculation results into intuitive early warning levels. It solves the problem of imbalance in alarm strategies caused by the failure to combine target scale with collision risk probability in existing technologies, and improves the engineering practicality of early warning results.

[0033] S102. Standardize and quality control the input space target orbit data, supplement the target size and position uncertainty parameters with the target catalog information, form a list of key target states, and build a list of potential rendezvous events within the warning time window; In one embodiment of this application, the process of standardizing and quality-controlling the input space target orbit data may include the following: The TLE data is parsed to extract the NORAD number, name, epoch time, and orbital root number, and the units and fields of the orbital root number are standardized. Query the target directory to obtain the target's equivalent size and location standard deviation. If the information is missing, fill it in using the default parameters. The key states of the targets are combined in pairs and encapsulated into key states of rendezvous events that include target pairs, time windows, orbital parameters and uncertainty parameters.

[0034] This step only requires inputting the latest TLE data snapshot obtained at the start time of the current warning calculation, without relying on multi-day historical TLE sequences for orbit accuracy improvement or error statistical analysis.

[0035] When performing collision classification early warning for space targets, the processor first needs to extract key information for rapid screening and probability assessment from the Two Line Element (TLE) data and the target directory. This invention defines the "critical state of rendezvous events" for each pair of space targets within the early warning time window as a set of fields consisting of orbital parameters, propagation time windows, relative geometric quantities, and uncertainty parameters. This set of fields is stored as a rendezvous event list in the form of event nodes for subsequent rapid screening and precise assessment. This step only requires inputting the latest TLE data snapshot obtained at the start time of the current early warning calculation, without relying on multi-day historical TLE sequences for orbital accuracy improvement or error statistical analysis. This processing achieves a unified representation of the risk state of a specific target pair within a specific time window, and the algorithm involved is defined as Algorithm 1.

[0036] Algorithm 1: Intersection Event Information Extraction Algorithm

[0037] In other words, the processor first parses and standardizes the two lines of input root data line by line, unifying the scattered orbital description information into a structured target basic information table. Based on this, it further combines target catalog information to supplement equivalent size and positional uncertainty parameters for each spatial target, and fills in missing information using default parameters, thus forming a set of key target states that meets the needs of early warning calculations. Subsequently, by pairwise combining the key target states, a candidate set of potential rendezvous events is constructed within a given early warning time window, and the orbital parameters and uncertainty information of each pair of targets within that time window are encapsulated as independent key rendezvous event states. In fact, by standardizing and quality-controlling the input data, and using default values ​​to fill in missing items, a unified representation of the risk state of a specific target pair within a specific time window is achieved, ensuring the robustness and continuity of subsequent calculations even with incomplete data.

[0038] S103. A multi-layered filtering strategy from coarse to fine is adopted to screen the target pairs in the potential rendezvous event list. The target pairs that are unlikely to approach are eliminated by using the orbital envelope, orbital plane difference and time window coarse propagation distance upper bound criteria, and a candidate rendezvous pair list is generated. Specifically, the process of filtering target pairs in the potential intersection event list using a multi-layered filtering strategy from coarse to fine can include the following: The first layer is orbital envelope screening, the second layer is orbital plane screening, and the third layer is time window coarse propagation screening; The first-level orbital envelope screening includes: Calculate the perigee and apogee altitude ranges of the two targets in the target pair; Determine whether the perigee and apogee height ranges of two targets overlap. If there is no overlap, determine that the target pair is unlikely to approach and eliminate it. The second-level track plane filtering includes: Calculate the difference in orbital inclination and the difference in right ascension of the ascending node between the two targets in the target alignment; Determine whether the difference in orbital inclination and the difference in right ascension of the ascending node are both greater than a preset threshold. If so, determine that the target is unapproachable due to the excessive difference in orbital plane space and eliminate it. The third-level time window coarse propagation screening includes: A low-precision orbit model and a coarse sampling time step are used to propagate and sample the target position within the warning time window; The minimum distance upper bound of the target pair formed by the two targets within the warning time window is dynamically estimated. If the minimum distance upper bound is greater than the preset screening threshold, the target pair is determined to be unlikely to approach and is removed. If the minimum distance during propagation is less than the preset screening threshold, the propagation is terminated early and the pair is identified as a candidate intersection pair.

[0039] Specifically, when performing collision level warnings, directly performing pairwise calculations on all targets would result in combinatorial explosions and significantly increase the computational burden. This invention defines "rapid filtering" as a multi-layered filtering process from coarse to fine. First, it eliminates target pairs that are impossible to approach within the warning time window using low-cost criteria. Then, it performs precise calculations on the remaining candidate pairs, thereby achieving rapid calculation for collision warnings. Specifically, firstly, it uses orbital envelope filtering to quickly eliminate target pairs with excessively large orbital height differences based on whether the height ranges of the target's perigee and apogee overlap. Secondly, it performs orbital plane filtering by calculating the difference in inclination angles and the difference in right ascension of the ascending nodes of the two targets, eliminating target pairs that are impossible to approach due to excessive differences in orbital plane space. Finally, it implements coarse propagation filtering within a time window, using a low-precision orbital model and a large time step to rapidly sample target positions within the warning time window. If the upper bound of the minimum distance within the entire time window is still greater than a preset threshold, the target pair is determined to be impossible to approach. The algorithm involved in candidate pair generation is defined as Algorithm 2.

[0040] Algorithm 2: Fast Filtering and Candidate Crossover Pair Generation Algorithm

[0041] This step, based on the list of critical target states, employs a multi-layered screening strategy, from coarse to fine, to progressively compress potential intersecting target pairs. First, by comparing the perigee and apogee altitude ranges of the target orbits, target pairs whose orbital envelopes do not overlap spatially are quickly eliminated, thus excluding combinations with no geometric probability of approach. Building on this, orbital plane constraints are further introduced, using inclination difference and right ascension difference of the ascending node to determine the relative relationship of the target orbital planes, reducing invalid combinations. Subsequently, among the target pairs satisfying the above geometric conditions, the orbital states within the warning time window are propagated using a coarse time step, dynamically estimating the upper bound of the minimum distance between the two targets within the time window. The propagation calculation is terminated early when the coarse screening threshold is met to reduce computational overhead. Through this layered screening and early termination mechanism, the size of the target pairs required for subsequent precise evaluation is significantly reduced while ensuring the effectiveness of the screening.

[0042] This step, through a tiered screening and early termination mechanism, significantly reduces the size of target pairs that need to be processed in subsequent actuarial evaluations while ensuring the effectiveness of the screening. It solves the combinatorial explosion problem caused by pairwise calculations of all targets and greatly improves the efficiency of early warning calculations.

[0043] S104. Perform precise propagation and nearest approach point search on the target pairs in the candidate rendezvous pair list, calculate the nearest approach time, minimum distance and relative velocity, establish a characterization of relative position uncertainty, and calculate the collision probability of the rendezvous event under the constraint of equivalent collision radius. In one embodiment of this application, the process of performing actuarial propagation and nearest proximity point search on target pairs in the candidate intersection pair list may include the following: The orbital status of candidate rendezvous pairs in the candidate rendezvous pair list is synchronously propagated within the early warning time window using a precisely calculated step size. The system iterates through and calculates the distance between two targets, searches for and records the closest approach time, the minimum distance, and the relative velocity at that time to complete the search for the closest approach point.

[0044] In one embodiment of this application, the process of establishing a representation of relative position uncertainty may include the following: Construct the three-dimensional position error covariance matrix of a single target based on the isotropic assumption; By superimposing the position error covariance matrices of the two targets, a unified position error model based on the relative position error covariance is obtained. A uniform position error model is used for each target to establish a characterization of the relative position uncertainty of each target; The position error model sets the position standard deviation to increase linearly with the time difference between the epoch of the orbital data and the closest approach time.

[0045] In one embodiment of this application, the process of calculating the equivalent collision radius may include the following: Obtain the equivalent diameter of the protected object and the equivalent diameter of the intersecting target; If the equivalent diameter of the intersecting target is unknown, the default size will be used. Calculate the equivalent collision radius, which is half the sum of the equivalent diameters of the protected object and the intersecting target. The process of establishing a characterization of relative position uncertainty may include the following: Construct the three-dimensional position error covariance matrix of a single target based on the isotropic assumption; By superimposing the position error covariance matrices of the two targets, a unified position error model based on the relative position error covariance is obtained. A uniform position error model is used for each target to establish a characterization of the relative position uncertainty of each target; The position error model sets the position standard deviation to increase linearly with the time difference between the epoch of the orbital data and the closest approach time.

[0046] In this embodiment, the present invention further performs precise propagation and nearest-neighbor point search to obtain key intersection geometry quantities such as the nearest approach time and minimum distance. Based on this, it combines the calculation of the equivalent collision radius and the uncertainty model to calculate the collision risk probability, providing a quantitative basis for graded early warning. The algorithm involved in the precise calculation and probability assessment of candidate pairs is defined as Algorithm 3.

[0047] Algorithm 3: Candidate intersection pair calculation and collision probability evaluation algorithm

[0048] In this step, the processor first propagates the target orbital state using a precise step size within the warning time window, searches for the closest approach time and minimum distance between the two targets, and records the corresponding relative velocity information. Then, it constructs an equivalent collision radius based on the target's equivalent size, and fills in missing size information using default parameters. Based on this, and combining the time difference between the closest approach time and the orbital epoch, it uniformly models the positional uncertainty of the two targets, constructs the relative position covariance, and calculates the collision probability. Finally, it outputs a rendezvous event evaluation result including the closest approach time, minimum distance, relative velocity, and collision probability, providing a quantitative basis for graded early warning.

[0049] This step transforms simple distance screening into collision probability calculation that includes positional uncertainty and target physical size by performing high-precision orbit propagation and probabilistic risk assessment on candidate target pairs. This provides accurate and quantitative data support for subsequent classification and judgment, ensuring the reliability of the early warning results.

[0050] S105. Based on the constructed graded early warning template, the collision probability and the acquired target size information, the level of the intersection event is determined, and the graded early warning result including the risk level is output.

[0051] In one embodiment of this application, the process of determining the level based on the graded early warning template may specifically include the following: Based on the size parameters of the target, the target is classified into the corresponding size class, which includes at least micro target class, small target class, medium target class, large target class and unknown size class; obtain the collision probability threshold set corresponding to the size class; compare the calculated collision probability with the corresponding threshold, and if the collision probability is greater than or equal to the threshold, trigger the corresponding level of warning.

[0052] Based on the obtained list of rendezvous event evaluation results (RL), this invention uses the established space target collision classification and early warning template to determine the level of each rendezvous event and output a structured early warning result. The algorithm involved in classification determination and output is defined as Algorithm 4.

[0053] Algorithm 4 Collision Level Warning Judgment and Output Algorithm

[0054] Understandably, the processor output includes not only the risk level, but also key evidence fields such as target pair ID, TCA, minimum distance, relative speed, Pc, HBR, and size classification, supporting incremental reassessment and rapid updating of warning results under data update conditions.

[0055] Therefore, this step achieves interpretable risk output by combining collision probability with target size into a graded judgment mechanism. This allows the warning results to quantify the degree of risk and correlate it with the physical attributes of the target, making it easier for operation and control personnel to intuitively judge the risk priority and formulate subsequent handling measures.

[0056] Example 2 This invention provides a method for hierarchical early warning of space target collisions based on rapid screening, the process of which is as follows: Figure 1 It includes the following steps: Based on step one, a multi-level space target collision classification early warning template is designed. The TLE data of the space target is used as the input data source, and the SGP4 / SDP4 propagation model, which matches the two-row root perturbation processing mechanism, is used for orbit prediction.

[0057] According to step two, the 3-day warning time window With fixed step size Generate an epoch sequence and for each target Propagation yields the state vector: , Quality control is performed on abnormal data during propagation: if propagation returns an invalid state or an error... If a target's invalid point is not identified, that moment is recorded as an invalid point. If the proportion of invalid points for a target within a time window exceeds a threshold, that target will not participate in subsequent screening in this round of warnings to ensure stable output results. For any two targets A and B, their orbital root numbers are respectively represented as... ; .in For the semi-major axis, For eccentricity, For the track inclination angle, Right ascension of the ascending node, The perigee argument, It is the angle closest to the point.

[0058] According to step three, the propagated state is used to calculate the relative geometric quantity. The candidate set is narrowed down step by step through a triple screening strategy to avoid performing high-cost actuarial calculations on all target pairs.

[0059] According to step four, the same location standard deviation is applied to all targets, and its linear growth with propagation time is considered: (1) in The initial position standard deviation, This represents the error growth rate. This represents the epoch of the orbital data. The corresponding three-dimensional position error covariance can be taken in isotropic form: (2) The covariance of the relative position errors of the two targets is taken as a superposition: (3) This error model serves as the input for subsequent collision probability Pc calculations. A collision event is defined as an event occurring within the encounter plane with a radius of... The circular regions whose relative positions in the meeting plane follow a two-dimensional normal distribution. Then the collision probability can be expressed by formula (4). (4) According to step five, the graded early warning judgment and result output are performed.

[0060] The beneficial effects of the present invention are verified by the following embodiments: At 16:55:59.82 UTC on February 10, 2009, Iridium 33 collided with the Cosmos 2251 satellite. TLE data from February 8, 2009, obtained from the Space-Track website, shows an orbital model with Iridium 33 as the protected object (S) and Cosmos 2251 as the rendezvous target (F). Figure 2 The TLE (epoch times of 2009-02-08 19:23:07.409Z and 2009-02-08 12:28:22.744Z) of both targets were input into the system of this invention for graded early warning assessment. In the rapid screening phase, the two targets were screened using their orbital envelopes and orbital planes. In the precise assessment, the system calculated the closest time to collision (TCA) to be 2009-02-10 16:55:59.794Z, with an absolute error Δt of less than 0.1 seconds compared to the historical collision time. At this time, the minimum distance was 841.432 meters, and the relative speed was 11.647 km / s.

[0061] Since the target size is unknown, it is processed according to the default size. A uniform position standard deviation σ_pos is set to 200 meters, and the equivalent collision radius (HBR) is set to 10 meters. The calculated collision probability Pc = 1.80091 × 10⁻⁶. -7 According to the tiered early warning template, the default probability threshold P0 for this tier is set to 1×10. -8Since Pc ≥ P0 = P1, the system determines that an early warning has been triggered, with the warning level being "notice level". Code execution time: 0.454906 s. A schematic diagram of its orbital propagation results is shown below. Figure 3 .

[0062] This embodiment demonstrates that the method of the present invention can accurately identify the meeting risk of this historical collision event within a 3-day warning time window by using only a single set of TLE data from about two days before the collision, and output a clear warning level based on probabilistic assessment and grading template, thus verifying the effectiveness and timeliness of the method in rapid screening and graded warning.

[0063] The third embodiment of this application provides a space target collision classification and early warning device based on rapid screening. The device may include: The template construction module is configured to build a hierarchical early warning template that includes the protected object, the intersection target, the target size parameters, the equivalent collision radius, the collision probability threshold, and the early warning level output; wherein, the early warning level output is determined based on the target size classification and the collision probability threshold. The rendezvous event list construction module is configured to include standardizing and quality-controlling the input space target orbit data, supplementing the target size and position uncertainty parameters with the target catalog information, forming a target key status list, and constructing a potential rendezvous event list within the warning time window; The rendezvous pair list construction module is configured to use a multi-layer filtering strategy from coarse to fine to filter target pairs in the potential rendezvous event list. It uses orbital envelope, orbital plane difference and time window coarse propagation distance upper bound criteria to eliminate target pairs that are unlikely to approach each other and generate a candidate rendezvous pair list. The collision probability calculation module is used to perform precise propagation and nearest approach point search on target pairs in the candidate rendezvous pair list, calculate the nearest approach time, minimum distance and relative velocity, establish a relative position uncertainty characterization, and calculate the collision probability of the rendezvous event under the constraint of equivalent collision radius. The graded early warning module is used to determine the grade of a meeting event based on the constructed graded early warning template, the collision probability, and the acquired target size information, and output a graded early warning result that includes the risk level.

[0064] A third embodiment of this application provides an electronic device, the electronic device comprising: At least one processor, memory, and input / output unit; The memory is used to store computer programs, and the processor is used to call the computer programs stored in the memory to execute the aforementioned space target collision classification and early warning method based on rapid screening.

[0065] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for hierarchical early warning of space target collisions based on rapid screening, characterized in that, include: A hierarchical early warning template is constructed, which includes the protected object, the intersecting target, the target size parameters, the equivalent collision radius, the collision probability threshold, and the early warning level output; wherein, the early warning level output is determined based on the target size classification and the collision probability threshold. The input space target orbit data is standardized and quality controlled. The target size and position uncertainty parameters are supplemented by the target catalog information to form a list of key target states. A list of potential rendezvous events is also constructed within the warning time window. A multi-layered filtering strategy, from coarse to fine, is used to screen target pairs in the potential rendezvous event list. Target pairs that are unlikely to approach are eliminated by using orbital envelope, orbital plane difference, and upper bound of coarse propagation distance within the time window as criteria, thus generating a candidate rendezvous pair list. For target pairs in the candidate rendezvous pair list, perform actuarial propagation and search for the nearest approach point, calculate the nearest approach time, minimum distance and relative velocity, establish a characterization of relative position uncertainty, and calculate the collision probability of the rendezvous event under the constraint of equivalent collision radius; Based on the constructed hierarchical early warning template, the collision probability and the acquired target size information are used to determine the level of the intersection event and output a hierarchical early warning result including the risk level.

2. The space target collision classification and early warning method based on rapid screening as described in claim 1, characterized in that, The standardization and quality control of the input space target orbit data includes: The TLE data is parsed to extract the NORAD number, name, epoch time, and orbital root number, and the units and fields of the orbital root number are standardized. Query the target directory to obtain the target's equivalent size and location standard deviation. If the information is missing, fill it in using the default parameters. The key states of the targets are combined in pairs and encapsulated into key states of rendezvous events that include target pairs, time windows, orbital parameters and uncertainty parameters.

3. The space target collision classification and early warning method based on rapid screening as described in claim 1, characterized in that, The process of using a multi-layered filtering strategy, from coarse to fine, to filter target pairs in the potential intersection event list includes: The first layer is orbital envelope screening, the second layer is orbital plane screening, and the third layer is time window coarse propagation screening; The first-layer orbital envelope filtering includes: Calculate the perigee and apogee altitude ranges of the two targets in the target pair; Determine whether the perigee and apogee height ranges of two targets overlap. If there is no overlap, determine that the target pair is unlikely to approach and eliminate it. The second-level track plane filtering includes: Calculate the difference in orbital inclination and the difference in right ascension of the ascending node between the two targets in the target alignment; Determine whether the difference in orbital inclination and the difference in right ascension of the ascending node are both greater than a preset threshold. If so, determine that the target is unapproachable due to the excessive difference in orbital plane space and eliminate it. The third-level time window coarse propagation screening includes: A low-precision orbit model and a coarse sampling time step are used to propagate and sample the target position within the warning time window; The minimum distance upper bound of the target pair formed by the two targets within the warning time window is dynamically estimated. If the minimum distance upper bound is greater than the preset screening threshold, the target pair is determined to be unlikely to approach and is removed. If the minimum distance during propagation is less than the preset screening threshold, the propagation is terminated early and the pair is identified as a candidate intersection pair.

4. The space target collision classification and early warning method based on rapid screening as described in claim 1, characterized in that, The actuarial propagation and nearest proximity search of target pairs in the candidate intersection pair list includes: The orbital status of candidate rendezvous pairs in the candidate rendezvous pair list is synchronously propagated within the early warning time window using a precisely calculated step size. The system iterates through and calculates the distance between two targets, searches for and records the closest approach time, the minimum distance, and the relative velocity at that time to complete the search for the closest approach point.

5. The space target collision classification and early warning method based on rapid screening as described in claim 1, characterized in that, The establishment of the relative position uncertainty characterization includes: Construct the three-dimensional position error covariance matrix of a single target based on the isotropic assumption; By superimposing the position error covariance matrices of the two targets, a unified position error model based on the relative position error covariance is obtained. A uniform position error model is used for each target to establish a characterization of the relative position uncertainty of each target; The position error model sets the position standard deviation to increase linearly with the time difference between the epoch of the orbital data and the closest approach time.

6. The space target collision classification and early warning method based on rapid screening as described in claim 1, characterized in that, The equivalent collision radius includes: Obtain the equivalent diameter of the protected object and the equivalent diameter of the intersecting target; If the equivalent diameter of the intersecting target is unknown, the default size will be used. Calculate the equivalent collision radius, which is half the sum of the equivalent diameter of the protected object and the equivalent diameter of the intersecting target.

7. The space target collision classification and early warning method based on rapid screening as described in claim 1, characterized in that, The determination of the level of the intersection event includes: The targets are classified into corresponding size categories based on their size parameters. The size categories include at least micro target category, small target category, medium target category, large target category, and unknown size category. Obtain the set of collision probability thresholds corresponding to this size category; The calculated collision probability is compared with the corresponding threshold. If the collision probability is greater than or equal to the threshold, the corresponding level of warning is triggered.

8. The space target collision classification and early warning method based on rapid screening as described in claim 7, characterized in that, For targets with unknown dimensions, a default size is used to classify them into a default category, and a collision probability threshold is set according to a conservative principle for early warning determination; the output graded early warning results also include the closest approach time, minimum distance, relative speed, and traceable data identifier.

9. A space target collision classification and early warning device based on rapid screening, characterized in that, include: The template construction module is configured to build a hierarchical early warning template that includes the protected object, the intersection target, the target size parameters, the equivalent collision radius, the collision probability threshold, and the early warning level output; wherein, the early warning level output is determined based on the target size classification and the collision probability threshold. The rendezvous event list construction module is configured to include standardizing and quality-controlling the input space target orbit data, supplementing the target size and position uncertainty parameters with the target catalog information, forming a target key status list, and constructing a potential rendezvous event list within the warning time window; The rendezvous pair list construction module is configured to use a multi-layer filtering strategy from coarse to fine to filter target pairs in the potential rendezvous event list. It uses orbital envelope, orbital plane difference and time window coarse propagation distance upper bound criteria to eliminate target pairs that are unlikely to approach each other and generate a candidate rendezvous pair list. The collision probability calculation module is used to perform precise propagation and nearest approach point search on target pairs in the candidate rendezvous pair list, calculate the nearest approach time, minimum distance and relative velocity, establish a relative position uncertainty characterization, and calculate the collision probability of the rendezvous event under the constraint of equivalent collision radius. The graded early warning module is used to determine the grade of a meeting event based on the constructed graded early warning template, the collision probability, and the acquired target size information, and output a graded early warning result that includes the risk level.

10. An electronic device, characterized in that, include: At least one processor, memory, and input / output unit; The memory is used to store computer programs, and the processor is used to call the computer programs stored in the memory to execute the space target collision classification and early warning method based on rapid screening according to any one of claims 1 to 8.

Images