Satellite collision avoidance system and collision avoidance method based on satellite edge computing network

Abstract

The invention discloses a satellite collision avoidance system and method based on a satellite edge computing network. The system comprises a space monitoring network, a space traffic management subsystem and the satellite edge computing network. The space monitoring network is used for observing the running state of a space target to obtain multi-source observation data; the space traffic management subsystem is used for integrating the multi-source observation data to generate space target orbit data; the satellite edge computing network further comprises the steps that a GEO satellite broadcasts space target orbit data to an LEO satellite constellation and forwards collision risk screening information to a ground cloud computing center; the LEO satellite constellation calculates collision risk screening information according to the space target orbit data and formulates a collision avoidance control strategy; the LEO / GEO earth station is used for data transceiving and protocol conversion among the GEO satellite, the LEO satellite constellation and the cloud computing center; and the ground cloud computing center stores constellation orbit data through protocol conversion and satellite communication. According to the invention, the real-time performance of satellite collision avoidance is improved.

Description

Technical Field

[0001] The present application relates to the field of satellite collision avoidance technology, and more specifically, to a satellite collision avoidance system and collision avoidance method based on a satellite edge computing network. Background Art

[0002] With the large-scale deployment of global low-Earth orbit (LEO) satellite constellations (such as Starlink and OneWeb), the number of satellites in low-Earth orbit (LEO) has grown exponentially, and the density of space traffic has increased dramatically. According to the European Space Agency (ESA), as of 2023, the number of satellites in orbit has exceeded 10,000, and an additional 100,000 are expected over the next decade. At the same time, the number of space debris (including defunct satellites, rocket debris, etc.) exceeds 35,000 pieces, posing a potential collision risk. In this context, satellite collision avoidance technology has become a core requirement for ensuring the safe operation of spacecraft, and is particularly crucial for the sustainable development of low-Earth orbit satellite constellations.

[0003] Current satellite obstacle avoidance technology primarily relies on ground-based centralized computing. This involves using ground-based space surveillance networks (such as radar and optical telescopes) to track and monitor space targets in a timely manner. This centralized computing uses measurement data to screen and publicly release collision risk information. Satellite operators then use this data to implement collision avoidance control based on the collision risk. For example, the U.S. Space Surveillance Network (SSN) and the European Space Situational Awareness Center (SSA) both employ this model.

[0004] However, existing ground-based centralized computing relies on two-way satellite-ground-satellite communication, with single-transmission data transmission latency ranging from seconds to minutes. This system consumes significant satellite communication bandwidth (e.g., single-orbit data transmission throughput reaches megabytes). The highly dynamic motion of satellites prevents continuous communication between ground stations and satellites, making existing systems unable to meet the collision avoidance requirements of high-density satellite constellations. When extreme events disrupt communication, satellites completely lose their obstacle avoidance capabilities, posing a collision risk. Summary of the Invention

[0005] In response to at least one defect or improvement need in the existing technology, the present invention provides a satellite collision avoidance system and collision avoidance method based on a satellite edge computing network, which is used to solve the problem that the collision avoidance method in the existing technology relies on continuous communication between the ground station and the satellite, and there is a delay in data transmission, which cannot meet the real-time obstacle avoidance requirements under the satellite's high dynamic orbit. When an extreme event occurs and causes communication interruption, the satellite will completely lose its obstacle avoidance capability, causing the risk of chain collisions.

[0006] To achieve the above objectives, according to a first aspect of the present invention, a satellite collision avoidance system based on a satellite edge computing network is provided, comprising: a space monitoring network, a space traffic management subsystem, and a satellite edge computing network;

[0007] The space monitoring network is used to observe the operational status of space targets and obtain multi-source observation data;

[0008] The space traffic management subsystem is used to integrate multi-source observation data to generate space target orbit data;

[0009] The satellite edge computing network also includes:

[0010] GEO satellites deployed in geosynchronous orbit are used to respond to computing tasks through edge computing servers, broadcast space target orbit data to the LEO satellite constellation through inter-satellite link signal transmitters, and forward collision risk screening information to ground cloud computing centers;

[0011] A LEO satellite constellation consisting of multiple low-orbit satellites is used to screen and calculate collision risk screening information based on space target orbit data and formulate collision avoidance control strategies;

[0012] LEO / GEO earth stations deployed on the ground are used for data transmission and reception and protocol conversion between GEO satellites, LEO satellite constellations and cloud computing centers;

[0013] The ground cloud computing center is connected to the GEO satellite and LEO satellite constellation through the LEO / GEO earth station, and is used to communicate with the satellite through protocol conversion to store the constellation orbit data.

[0014] In one possible implementation, the space surveillance network includes ground-based radars and optical telescopes;

[0015] Land-based radar is used to scan space targets using multi-band radar beams to obtain target range and velocity vector data of the targets;

[0016] Optical telescopes are used to capture the angular and shape characteristics of space targets through high-resolution imaging technology.

[0017] In one possible implementation, the LEO satellite constellation collision warning screening process includes geometric screening, temporal screening, and probability calculation;

[0018] Among them, geometric screening is used to perform perigee / apogee screening and orbit intersection distance analysis based on space target orbit data to determine suspected collision targets;

[0019] Time screening analyzes the time difference between suspicious collision targets passing through the track intersection line and screens out suspicious collision targets with non-overlapping time windows;

[0020] The probability calculation uses the preset orbit prediction model to calculate the collision probability of the remaining suspected collision targets.

[0021] In one possible implementation, the LEO satellite constellation further includes an early warning module and a confidence module;

[0022] Among them, the warning module is used to trigger a warning when the collision probability exceeds a preset probability threshold;

[0023] The confidence module is used to initiate an encrypted observation request to reacquire the space target orbit data when the confidence of the space target orbit data is lower than a preset confidence threshold.

[0024] In one possible implementation, the LEO satellite constellation further includes a collision avoidance decision module;

[0025] The collision avoidance decision module is used to set constraints, establish a constrained multi-objective optimization model, formulate a collision avoidance control strategy, calculate the change in satellite velocity vector, and execute the collision avoidance control strategy with the smallest change in satellite velocity vector.

[0026] According to a second aspect of the present invention, a satellite collision avoidance method based on a satellite edge computing network is also provided, and a satellite collision avoidance system based on a satellite edge computing network according to any one of the possible implementations described above includes:

[0027] The ground-based radar uses a multi-band radar beam to scan space targets to obtain observation data on target range and velocity vector, and the optical telescope uses high-resolution imaging to capture observation data on target angular characteristics and shape characteristics;

[0028] Integrate observation data from ground-based radars and optical telescopes to generate orbital data of space targets, and upload the data synchronously to GEO satellites in geosynchronous orbit;

[0029] The GEO satellite broadcasts the orbital data of the space target to the LEO satellite constellation, and performs geometric screening, time window matching, and collision probability calculation in sequence;

[0030] When the collision probability exceeds the preset threshold, an early warning is triggered, and a multi-objective optimization model is established to generate a collision avoidance control strategy.

[0031] In a possible implementation, performing geometric screening further includes:

[0032] Extract the semi-major axis, eccentricity, and orbital inclination parameters from the space target orbit data and calculate the perigee and apogee heights;

[0033] Based on the orbital plane equation of the target satellite, calculate the orbital intersection distance and intersection point coordinates;

[0034] If the orbital altitude difference of the target satellite's perigee or apogee altitude is less than the preset altitude, and the orbital intersection distance is less than the preset safety distance, it will be marked as a suspected collision target.

[0035] In a possible implementation, performing time window matching further includes:

[0036] Calculate the time difference between the suspected collision target passing through the orbital intersection line based on the space target orbit data and the satellite's ephemeris;

[0037] If the time difference between the suspected collision target and the track intersection is greater than the preset safety time window, the target is judged to have no collision risk.

[0038] In a possible implementation, sequentially performing geometric screening, time window matching, and collision probability calculation further includes:

[0039] Calculate the confidence level of the space target orbit data based on the space target orbit covariance of the observation satellite, the target imaging clarity and the orbit prediction model error parameters;

[0040] If the confidence level is lower than the preset confidence level, an encrypted observation request is initiated, and land-based radar and optical telescopes are designated to track and observe the target satellite.

[0041] The encrypted observation data is fused with historical data through the Kalman filter algorithm to update the space target orbit data.

[0042] In one possible implementation, when the collision probability exceeds a preset threshold, an early warning is triggered, and a multi-objective optimization model is established to generate a collision avoidance control strategy, which also includes:

[0043] Set constraints and establish a multi-objective optimization model based on the constraints;

[0044] The multi-objective optimization model is solved by particle swarm optimization algorithm to generate a set of candidate avoidance strategies;

[0045] The strategy that minimizes the change in the satellite velocity vector is selected from the candidate avoidance strategy set as the target solution, and the target solution is sent to the target satellite.

[0046] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0047] The present invention provides a satellite collision avoidance system based on a satellite edge computing network. After integrating multi-source observation data (including radar observation data, telescope observation data, satellite GNSS measurement data, etc.), the space traffic management center directly broadcasts space target orbit data to the LEO satellite constellation via the GEO satellite. The LEO satellite completes collision risk screening and avoidance strategy generation based on local computing power, and performs collision avoidance control. During or after the event, the calculation results, control results and other information are transmitted to the ground cloud computing center. This forms a "data-decision-making" closed loop, avoiding the time delay of traditional "satellite-ground-satellite" two-way transmission and adapting to the instantaneous obstacle avoidance requirements of the high dynamic orbit of low-orbit satellites. The GEO satellite, the LEO constellation, and the ground cloud computing center realize data exchange with each other through the LEO / GEO earth station, forming an inter-satellite-satellite-ground dual-channel communication architecture. When ground communication is interrupted, the GEO satellite can still maintain data transmission with the LEO constellation, ensuring the continuous issuance of risk information and avoidance instructions, and avoiding the loss of obstacle avoidance capability due to communication failure. The present invention achieves obstacle avoidance in the high-dynamic orbit scenario of low-orbit satellites, gets rid of dependence on ground communications, and maintains system resilience through redundant links in extreme scenarios, providing more reliable technical support for space traffic safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0049] Figure 1 An interactive schematic diagram of an embodiment of a satellite collision avoidance system based on a satellite edge computing network provided by the present invention;

[0050] Figure 2 A schematic flow chart of an embodiment of a satellite collision avoidance method based on a satellite edge computing network provided by the present invention;

[0051] Figure 3 A schematic diagram of a flow chart of an embodiment of performing geometric screening provided by the present invention;

[0052] Figure 4 A schematic diagram of a flow chart of an embodiment of updating space target orbit data provided by the present invention;

[0053] Figure 5 A schematic diagram of a flow chart of an embodiment of generating a collision avoidance control strategy provided by the present invention;

[0054] Figure 6 A schematic diagram of the architecture of an embodiment of a satellite collision avoidance system based on a satellite edge computing network provided by the present invention. DETAILED DESCRIPTION

[0055] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0056] The terms "first," "second," "third," and the like in the specification and claims of this application and the accompanying drawings are used to distinguish between different objects, not to describe a particular order. Furthermore, the terms "including," "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or elements is not limited to the listed steps or elements, but may optionally include steps or elements not listed, or may optionally include other steps or elements inherent to the process, method, product, or apparatus.

[0057] The present invention provides a satellite collision avoidance system and collision avoidance method based on a satellite edge computing network, which are described below.

[0058] See also Figure 1 , Figure 1 This is an interactive diagram of an embodiment of a satellite collision avoidance system based on a satellite edge computing network provided by the present invention. In a specific embodiment of the present invention, a satellite collision avoidance system based on a satellite edge computing network is disclosed, including: a space monitoring network, a space traffic management subsystem, and a satellite edge computing network;

[0059] The space monitoring network is used to observe the operating status of targets in space and obtain multi-source observation data;

[0060] The space traffic management subsystem is used to integrate multi-source observation data to generate space target orbit data;

[0061] The satellite edge computing network also includes:

[0062] GEO satellites deployed in geosynchronous orbit are used to respond to computing tasks through edge computing servers, broadcast space target orbit data to the LEO satellite constellation through inter-satellite link signal transmitters, and forward collision risk screening information to ground cloud computing centers;

[0063] A LEO satellite constellation consisting of multiple low-orbit satellites is used to screen and calculate collision risk screening information based on space target orbit data and formulate collision avoidance control strategies;

[0064] LEO / GEO earth stations deployed on the ground are used for data transmission and reception and protocol conversion between GEO satellites, LEO satellite constellations and cloud computing centers;

[0065] The ground cloud computing center is connected to the GEO satellite and LEO satellite constellation through the LEO / GEO earth station, and is used to communicate with the satellite through protocol conversion to store the constellation orbit data.

[0066] In the above-mentioned embodiment, the space monitoring network relies on globally distributed optical and radar observatories to coordinate observations, acquiring full-dimensional data on target satellites, including their position, velocity, attitude, and electromagnetic spectrum. This network supports millimeter-level positioning accuracy and sub-second sampling cycles, addressing the blind spots and latency issues inherent in traditional single-source monitoring methods. Synchronizing multi-source data to a unified spatiotemporal reference eliminates coordinate system deviations between different observation sources, providing a reliable data foundation for subsequent orbit predictions.

[0067] The space traffic management subsystem performs spatiotemporal correlation matching and outlier removal on multiple sources of information, including radar point clouds, optical imagery, and ephemeris data, to generate centimeter-level precision orbital data for space targets, supporting the subsequent precise calculation of collision probabilities. Using a rapid collision cone algorithm and a dynamic collision probability threshold model, the fused orbital data is screened in real time, sending only high-risk targets with a collision probability greater than 1e-6 and their associated orbital data to edge computing nodes, reducing the burden of ineffective data transmission.

[0068] The radiation-hardened FPGA array deployed on the GEO satellite has TOPS-level AI computing power, supports real-time orbit propagation algorithms and collision probability rapid assessment models, and completes secondary verification and compression encoding of space target orbit data on-orbit. Through the Ka / laser hybrid intersatellite link, directional beamforming technology is used to broadcast orbit data to the LEO constellation at a rate of MB / s, while the TDMA time division multiple access protocol is used to ensure the reliability of multi-satellite concurrent transmission. High-confidence collision events (including time windows and avoidance strategies) reported by the LEO constellation are forwarded to the ground cloud computing center via the S-band satellite-to-ground link, supporting breakpoint resumption and QoS priority scheduling.

[0069] The LEO satellite constellation is based on a lightweight reinforcement learning model, combining local orbit data with GEO broadcast data to complete dynamic assessment of collision probability and generation of avoidance strategies within seconds. The avoidance plan is executed in real time through the onboard thruster / electric propulsion system.

[0070] Based on a software-defined radio (SDR) architecture, the LEO / GEO earth station dynamically adapts to the CCSDS satellite-to-ground protocol, TCP / IP terrestrial protocol, and custom lightweight protocols. It supports gigabit-level satellite-to-ground data throughput and features built-in traffic shaping and congestion control algorithms to address sudden communication peaks. Using a dual-satellite + multi-band hybrid networking mode, it automatically switches to a backup link if the primary link is interrupted by space interference, ensuring 99.999% satellite-to-ground communication availability.

[0071] Based on a high-precision orbital dynamics model, the ground-based cloud computing center replays millions of scenarios to verify the avoidance strategies reported by the LEO constellation. The center then outputs the globally optimal strategy, which is then transmitted back to the constellation via the Earth station. Combining solar activity forecast data with a space debris evolution model, a collision risk heat map is generated to guide the dynamic allocation of constellation resources and the reservation of avoidance windows.

[0072] Compared to existing technologies, this embodiment provides a satellite collision avoidance system based on a satellite edge computing network. After integrating multi-source observation data (including radar observation data, telescope observation data, and satellite GNSS measurement data), the space traffic management center directly broadcasts space target orbit data to the LEO satellite constellation via GEO satellites. LEO satellites, based on their local computing power, perform collision risk screening and generate avoidance strategies. During or after the event, they transmit the calculation results, control results, and other information to the ground-based cloud computing center. This creates a closed "data-decision-making" loop, avoiding the latency of traditional "satellite-ground-satellite" bidirectional transmission and adapting to the instantaneous obstacle avoidance requirements of low-orbit satellites' highly dynamic orbits. GEO satellites exchange data with the LEO constellation and the ground-based cloud computing center via LEO / GEO earth stations, forming an inter-satellite-satellite-ground dual-channel communication architecture. When ground communications are interrupted, GEO satellites can maintain data transmission with the LEO constellation, ensuring the continuous delivery of risk information and avoidance instructions, and preventing the loss of obstacle avoidance capabilities due to communication failures. The present invention achieves obstacle avoidance in the high-dynamic orbit scenario of low-orbit satellites, gets rid of dependence on ground communications, and maintains system resilience through redundant links in extreme scenarios, providing more reliable technical support for space traffic safety.

[0073] In some embodiments of the present invention, the space surveillance network includes ground-based radars and optical telescopes;

[0074] Land-based radar is used to scan space targets using multi-band radar beams, and to obtain target range and velocity vector data;

[0075] Optical telescopes are used to capture the angular and shape characteristics of space targets through high-resolution imaging technology.

[0076] In the above embodiment, the land-based radar subsystem uses a phased array radar array deployed at globally distributed radar sites. It uses the L-band (1-2 GHz) to penetrate the ionosphere to achieve long-range target acquisition, the S-band (2-4 GHz) to balance resolution and effective range, and the X-band (8-12 GHz) to provide centimeter-level high-precision ranging, forming full altitude layer coverage of 100 km to 40,000 km.

[0077] Using broadband linear frequency modulation signals and time-frequency analysis algorithms, the target's micro-vibration frequency and attitude change rate are extracted from the echo signal, assisting in identifying the physical state of non-cooperative targets. For reentry targets or high-density plasma environments, dual-band cross-polarization modulation and pulse compression coding are used to suppress ionospheric scintillation interference and ensure continuous tracking of targets below 100 km altitude.

[0078] The optical telescope achieves angular distance measurement accuracy of ≤0.1 arc second through star sensor calibration and image processing algorithms. Multi-frame image differential calculations support the inversion of target relative motion parameters. Edge detection algorithms extract the target's 2D outline, assisting in target type identification. Combining angular distance and distance data, the target's 3D dimensions are estimated to determine its shape characteristics.

[0079] In some embodiments of the present invention, the LEO satellite constellation collision warning screening process includes geometric screening, temporal screening, and probability calculation;

[0080] The geometric screening module is used to perform perigee / apogee screening and track intersection distance analysis based on the space target orbit data to determine suspected collision targets.

[0081] The time screening module analyzes the time difference between suspicious collision targets passing through the track intersection line and screens out suspicious collision targets with non-overlapping time windows;

[0082] The probability calculation module uses a preset orbit prediction model to calculate the collision probability of the remaining suspected collision targets.

[0083] In the above embodiment, perigee / apogee screening involves calculating the target satellite's perigee and apogee altitudes based on the compressed orbit data packets broadcast by the GEO satellites, along with parameters such as the semi-major axis, eccentricity, and orbital inclination. This is then combined with a locally stored historical orbit database to calculate the target's perigee and apogee altitudes. Based on the current orbital altitude distribution of the LEO constellation and a heat map of space debris density, the screening threshold is dynamically adjusted to avoid false positives in low-density areas and missed negatives in high-density areas.

[0084] Orbital intersection distance analysis: Using the Cartesian coordinate system and orbital root conversion algorithm, the orbits of the target satellite and the LEO constellation are projected into the J2000 geocentric inertial coordinate system, and the intersection equation and minimum intersection distance of the two orbital planes are calculated.

[0085] Based on DBSCAN density clustering and trajectory similarity metrics, suspicious targets are identified as clusters, prioritizing multi-target threats in the same domain. Convolutional neural networks are used to extract target angular and shape features, which are then compared with a library of known spatial targets to flag unknown high-risk targets.

[0086] Orbital intersection time difference calculation: Using the SGP4 / HPOP hybrid model, the orbits of the target satellite and the LEO constellation are propagated in a short period of time with high precision. The theoretical and actual passing time windows of the orbital intersection are calculated to screen out suspected collision targets whose time windows do not overlap.

[0087] Dynamic Collision Probability Calculation: Based on spatial geometric probability theory, the collision probability between the target satellite and the LEO constellation is calculated within millisecond latency. Based on the collision probability, the target is classified as high risk, medium risk, or low risk, triggering corresponding avoidance strategies.

[0088] Through a logic AND gate mechanism, only targets that have passed three levels of verification, namely, geometric screening, temporal screening, and probability calculation, are retained, preventing ineffective circumventions caused by misjudgments in a single module. The weights of the three levels of screening are dynamically adjusted based on the complexity of the spatial environment to optimize system response efficiency.

[0089] In some embodiments of the present invention, the LEO satellite constellation further includes an early warning module and a confidence module;

[0090] Among them, the warning module is used to trigger a warning when the collision probability exceeds a preset probability threshold;

[0091] The confidence module is used to initiate an encrypted observation request to reacquire the space target orbit data when the confidence of the space target orbit data is lower than a preset confidence threshold.

[0092] In the above embodiment, the risk classification threshold is set as follows:

[0093] Red warning threshold: triggers an emergency evasive maneuver, calls on thrusters to make a rapid orbit change, and transmits the information to the ground control center.

[0094] Orange warning threshold: Initiate delayed avoidance strategy, call on electric thrusters to make small orbit adjustments, and continue tracking the target.

[0095] Yellow warning threshold: Marked as a potential risk target, included in the priority observation queue, and no evasive action is triggered.

[0096] It is understandable that any thruster that meets the avoidance control conditions may include chemical thrusters, ion thrusters, cold gas thrusters, etc. The specific warning threshold may be set according to actual needs, and the present invention does not further limit this.

[0097] The confidence level of radar observation data is calculated based on signal-to-noise ratio, Doppler shift consistency, and cross-validation matching. The confidence level of optical observation data is calculated based on measurement accuracy, image signal-to-noise ratio, and multispectral consistency. Based on historical data source reliability statistics and current observation conditions, the weight of each data source is dynamically assigned to calculate the overall confidence level.

[0098] When the comprehensive confidence of the target satellite is lower than the threshold, an encrypted observation mission request is initiated to other satellites or ground tracking and control stations in the constellation, giving priority to the observation needs of high-risk, low-confidence targets.

[0099] In some embodiments of the present invention, the LEO satellite constellation further includes a collision avoidance decision module;

[0100] The collision avoidance decision module is used to set constraints, establish a constrained multi-objective optimization model, formulate a collision avoidance control strategy, calculate the change in satellite velocity vector, and execute the collision avoidance control strategy with the smallest change in satellite velocity vector.

[0101] In the above embodiment, constraints such as fuel balance, mission priority, constellation configuration, and orbital perturbation effects are comprehensively considered to establish a constrained multi-objective optimization model to generate a set of candidate avoidance strategies that take into account avoidance effectiveness, fuel consumption, and mission constraints.

[0102] The satellite velocity vector change is the difference between the final and initial values ​​of the velocity vector caused by the avoidance maneuver. The strategy with the smallest satellite velocity vector change is selected from the candidate avoidance strategy set as the target solution. For each selected strategy, the orbital dynamics constraints and mission constraints are revalidated to ensure there are no hidden conflicts.

[0103] See also Figure 2 , Figure 2 The present invention provides a flow chart of an embodiment of a satellite collision avoidance method based on a satellite edge computing network. The present invention also provides a satellite collision avoidance method based on a satellite edge computing network, which is based on a satellite collision avoidance system based on a satellite edge computing network as described in any one of the above embodiments, including:

[0104] S201. Scanning space targets with a multi-band radar beam using a land-based radar to obtain observation data of target range and velocity vector data, and capturing observation data of target angular characteristics and shape characteristics using an optical telescope with high-resolution imaging;

[0105] S202, integrating observation data from ground-based radar and optical telescope to generate space target orbit data, and synchronously uploading the data to a GEO satellite in geosynchronous orbit;

[0106] S203, the GEO satellite distributes the space target orbit data to the LEO satellite constellation by broadcasting, and performs geometric screening, time window matching, and collision probability calculation in sequence;

[0107] S204: When the collision probability exceeds a preset threshold, a warning is triggered, and a multi-objective optimization model is established to generate a collision avoidance control strategy.

[0108] In the above-described embodiment, a land-based radar system deployed at a ground station uses a multi-band collaborative scanning mode, leveraging the ability of low-frequency radar waves to penetrate cloud and fog interference and the high-precision positioning capabilities of high-frequency radar waves to perform three-dimensional scanning of space targets, continuously acquiring dynamic parameters such as the target's radial range, relative velocity, and acceleration. Simultaneously, a high-resolution optical telescope is used to conduct staring observations of key target areas. Adaptive optical compensation technology eliminates the effects of atmospheric turbulence, and multispectral imaging and laser ranging technologies are combined to extract static parameters such as the target's angular range, geometric profile, and surface material reflectance characteristics. The two observation methods achieve data alignment through a time synchronization mechanism, forming a complete observation set covering the target's kinematic and morphological characteristics.

[0109] Based on a heterogeneous data fusion algorithm, the dynamic parameters provided by the radar are spatiotemporally aligned with the static features acquired by the optical telescope. Measurement noise is eliminated through extended Kalman filtering and particle filtering algorithms, resulting in a spatial target state model that includes the six elements of the target's orbit, attitude characteristics, and motion trends. The fused data is broadcast to a geosynchronous orbit (GEO) satellite via a high-speed communication link, employing forward error correction coding and adaptive modulation to ensure the reliability of deep-space communication. The GEO satellite acts as a relay node, performing integrity checks and redundancy removal on the received data, generating standardized data packets for subsequent processing.

[0110] GEO satellites distribute pre-processed target data to a constellation of low-Earth orbit (LEO) satellites, forming a distributed edge computing network. Each LEO satellite acts as a computing node, collaboratively calculating the target orbit based on a parallel processing architecture. First, a recursive orbit prediction algorithm performs geometric screening to identify potential threatening targets that may enter the danger zone. A time window matching algorithm, combined with the satellite's attitude adjustment capabilities and propellant remaining, calculates the time interval for rendezvous between the target and the satellite. The collision probability is calculated based on orbital position and covariance information, quantitatively assessing the collision probability within each time window to form a multi-dimensional risk assessment matrix.

[0111] When the collision probability exceeds a preset safety threshold, the system automatically triggers a three-level warning mechanism: Level 1 sends a risk alert to ground control, Level 2 activates the satellite's autonomous decision-making module, and Level 3 activates the avoidance strategy generation engine. This engine builds a multi-objective optimization model that comprehensively considers constraints such as minimum fuel consumption, mission interruption risk, and space debris contamination. It employs an improved genetic algorithm and deep reinforcement learning techniques to generate an avoidance control strategy that includes velocity vector adjustments, attitude control commands, and an execution time window. The strategy generation process incorporates digital twin technology, and the feasibility of the strategy is verified using a virtual satellite model.

[0112] The avoidance control strategy is autonomously calculated by the satellite and transmitted to the ground, providing an overall understanding of the constellation's operational safety. Upon receiving the command, the target satellite activates its autonomous execution module, employing a closed-loop control algorithm to fine-tune the propulsion system. Coordinated verification of strategy execution is achieved via an intersatellite laser link. During execution, the satellite collects real-time data from attitude sensors, propulsion system status, and target tracking, generating an execution log containing execution timestamps, control command sequences, and environmental parameter changes. This log data is synchronously transmitted to a ground-based cloud computing center via a bidirectional transmission protocol and immutably stored using a distributed file system and blockchain technology, providing data support for subsequent mission review and algorithm optimization.

[0113] See also Figure 3 , Figure 3 This is a flow chart of an embodiment of performing geometric screening provided by the present invention. In some embodiments of the present invention, performing geometric screening further includes:

[0114] S301. Extract the semi-major axis, eccentricity, and orbital inclination parameters from the space target orbit data and calculate the perigee and apogee heights;

[0115] S302, calculating the orbital intersection distance and the intersection coordinates based on the orbital plane equation of the target satellite;

[0116] S303: If the orbital height difference between the perigee and apogee of the target satellite is less than a preset height, and the orbital intersection distance is less than a preset safety distance, the target satellite is marked as a suspected collision target.

[0117] In the above embodiment, key parameters characterizing orbital geometry are extracted from multi-source fused space target orbital data, including but not limited to the semi-major axis length (reflecting the overall scale of the orbit), the size of the eccentricity (characterizing the eccentricity of the orbital shape), and the orbital inclination angle (describing the angular relationship between the orbital plane and the reference plane). Based on the Kepler orbital dynamics principle, combined with the real-time position and velocity vector of the satellite, the extreme height characteristics of the target orbit - the perigee height (characterizing the distance of the lowest point of the orbit relative to the earth's surface) and the apogee height (characterizing the distance of the highest point of the orbit relative to the earth's surface) are calculated through an iterative optimization algorithm. This process needs to take into account the long-term influence of complex factors such as the non-spherical gravitational perturbation of the earth and the attenuation of atmospheric drag on the orbital parameters to ensure the dynamic accuracy of the extreme point calculation.

[0118] The orbital plane equations for the target satellite and the space target are constructed. The normal vectors of the two orbital planes are determined through vector cross product operations, and the parameters of the plane angle are then solved. Based on spatial geometric analytical methods, the equation of the intersection line of the two orbital planes is calculated. This intersection line represents the potential spatial region where the two satellites may intersect. The coordinates of the intersection point of the intersection line with the Earth's reference ellipsoid are further solved using parametric equations to determine the three-dimensional position information of the intersection point in the Earth's inertial coordinate system. At the same time, the spatiotemporal distribution of the target satellite and the space target near the intersection line is predicted by combining satellite ephemeris data, and the probability of spatial overlap between the two orbits is quantitatively assessed.

[0119] A dual-threshold collaborative judgment mechanism is established. First, a dynamic threshold comparison is performed on the perigee / apogee altitude of the target satellite. When the difference between the orbital extreme altitude of the space target and the corresponding extreme point altitude of the target satellite is less than the preset vertical safety margin (this margin needs to comprehensively consider factors such as the satellite's physical size, attitude maneuverability, and propulsion system response delay), a horizontal safety check is triggered. Then, the closest distance between the intersection of the space target's orbit and the predicted orbit of the target satellite is calculated. If this horizontal distance is less than the preset lateral safety distance (this distance needs to be dynamically adjusted based on risk assessment results such as the space debris environment density and collision kinetic energy threshold), the space target is judged to be a potential collision threat object. The judgment result needs to be accompanied by a confidence score, which comprehensively considers multiple factors such as the quality of observation data, orbit prediction error, and the satellite's autonomous avoidance capability to provide a decision-making basis for subsequent risk classification and disposal.

[0120] In some embodiments of the present invention, performing time window matching further includes:

[0121] Calculate the time difference between the suspected collision target passing through the orbital intersection line based on the space target orbit data and the satellite's ephemeris;

[0122] If the time difference between the suspected collision target and the track intersection is greater than the preset safety time window, the target is judged to have no collision risk.

[0123] In the above embodiment, based on the multi-source fusion of space target orbital data (six orbital elements, covariance, update time and other information) and the high-precision ephemeris of the target satellite (covering the orbital element changes and attitude adjustment plan within the forecast period), a multi-body dynamics simulation model is used to parallelly deduce the future trajectories of the target and satellite. The orbital intersection position of the two is calculated through an orbital recursion algorithm (which needs to integrate complex perturbation models such as the Earth's non-spherical gravity, atmospheric drag, and solar pressure), and the time and space coordinates of both parties arriving at the intersection are predicted. This process requires the introduction of uncertainty quantification technology, and Monte Carlo sampling is used to generate multiple groups of possible trajectory clusters to cover the orbit prediction errors and the influence of environmental perturbations.

[0124] For each suspected collision target, the time difference between its predicted orbital intersection moment and the target satellite's corresponding moment is calculated to form a time difference sequence. A two-tiered judgment mechanism is constructed: first, a basic safety time window is set based on mission priority and the satellite's autonomous avoidance capability (this window must comprehensively consider factors such as propulsion system response delay, attitude adjustment time, and collision kinetic energy threshold). If the target time difference exceeds this threshold, the target is directly judged as a risk-free object. If the time difference is within the critical range, a secondary dynamic assessment is initiated, combining the density of the space debris environment, the stability of the target orbit, and the continuity requirements of the satellite mission to perform a weighted score on the time difference (for example, assigning a stricter time tolerance to high-value mission satellites).

[0125] Based on the time difference analysis, auxiliary parameters such as orbit prediction accuracy, observation data quality, and satellite autonomous decision-making capabilities are further integrated to construct a risk confidence model. Specifically, this includes: attenuation weighting of the timestamp freshness of the target orbit data, covariance propagation analysis of the ephemeris prediction error, and historical data fitting of the success rate of satellite avoidance strategy execution. When the comprehensive confidence is lower than the preset threshold, even if the time difference is within the critical range, the target can still be judged as a low-risk object, avoiding mission interruption caused by excessive avoidance. The judgment result needs to generate a time-stamped risk assessment report, including a time difference distribution histogram, a confidence heat map, and recommended disposal suggestions, providing structured input for subsequent collision probability calculations.

[0126] This time window matching process achieves a fine-grained spatiotemporal decoupling of potential collision threats by constructing a three-level evaluation architecture of "spatiotemporal trajectory deduction-dynamic threshold comparison-confidence fusion", minimizing unnecessary avoidance operations while ensuring satellite safety, and providing a reliable time benchmark for subsequent collision probability quantification and strategy generation.

[0127] See also Figure 4 , Figure 4 This is a flow chart of an embodiment of updating space target orbit data provided by the present invention. In some embodiments of the present invention, sequentially performing geometric screening, time window matching, and collision probability calculation further includes:

[0128] S401, calculating the confidence level of the space target orbit data based on the observed space target orbit covariance, target imaging clarity, and orbit prediction model error parameters;

[0129] S402: If the confidence level is lower than the preset confidence level, an encrypted observation request is initiated to designate a land-based radar and an optical telescope to track and observe the target satellite.

[0130] S403: Fusing the encrypted observation data with the historical data through the Kalman filter algorithm to update the space target orbit data.

[0131] In the above embodiment, based on the multimodal sensor characteristics of the observation satellite, a confidence assessment framework is constructed that integrates physical parameters and empirical models. The echo signal-to-noise ratio of the ground-based radar (reflecting the contrast between the target reflection intensity and the background noise) and the imaging clarity of the optical telescope (affected by atmospheric turbulence, target relative motion, and detector resolution) are analyzed to quantify the effective information content of the data. The historical deviation statistics of the orbit prediction model (such as the short-term prediction error distribution and the long-term perturbation model residual) are combined with the current space environment disturbance (such as the solar activity index and geomagnetic storm warning) to evaluate the degree of reliability degradation of the orbit prediction. Confidence weight coefficients are set for different mission stages (such as initial screening, warning confirmation, and avoidance execution). For example, the signal quality weight is increased in the avoidance decision stage, and the model error weight is increased in the long-term monitoring stage. Finally, a timestamped confidence score report is generated, which includes the sensitivity analysis of each parameter and the confidence degradation trend prediction.

[0132] When the confidence score falls below the preset "critical safety threshold" (which requires comprehensive consideration of the mission risk level and avoidance operation costs), an encrypted observation request is automatically triggered; if the confidence score is in the "low quality but acceptable" range, only an observation suggestion is generated and reported to the ground control center; for low-confidence targets, coordinated observation instructions are issued to ground-based radars and optical telescopes, specifying that the radar adopt high-repetition-rate pulse compression mode (to enhance weak target detection capabilities) and polarization diversity technology (to extract target material characteristics), and the optical telescope switch to adaptive optical compensation mode (to eliminate atmospheric turbulence interference) and multispectral imaging mode (to identify target morphological details); the observation window is dynamically adjusted according to the target's motion characteristics (such as maneuverability and orbital inclination change rate), and encrypted observations are prioritized when the target passes through the perigee or near the orbit intersection line, while avoiding the Earth's shadow area and strong background radiation areas.

[0133] The encrypted observation data are spatiotemporally aligned (eliminating sampling time differences and coordinate system differences between different sensors) and quality screened (eliminating abnormal data points contaminated by noise) to generate a standardized observation data set; an improved Kalman filter framework is used to jointly estimate the dynamic parameters provided by the radar (such as radial velocity and acceleration) and the static characteristics obtained by the optical telescope (such as angular distance change rate and target geometric profile), and the interactive multi-model (IMM) technology is used to process the possible maneuvering behavior of the target; based on the fused observation data, the orbital elements (such as semi-major axis and eccentricity), perturbation parameters (such as atmospheric drag coefficient and solar pressure model) and collision risk characteristics (such as rendezvous time window and minimum distance probability distribution) of the space target are recursively updated to generate a corrected orbit model with uncertainty ellipsoid; the data source, fusion algorithm parameters and confidence improvement are recorded for each orbit update to form a traceable orbital state evolution chain, providing full life cycle data support for subsequent collision probability calculation and avoidance strategy generation.

[0134] See also Figure 5 , Figure 5 This is a flow chart of an embodiment of generating a collision avoidance control strategy provided by the present invention. In some embodiments of the present invention, when the collision probability exceeds a preset threshold, a warning is triggered, and a multi-objective optimization model is established to generate a collision avoidance control strategy, which also includes:

[0135] S501, setting constraints and establishing a multi-objective optimization model based on the constraints;

[0136] S502, solving a multi-objective optimization model by a particle swarm optimization algorithm to generate a set of candidate avoidance strategies;

[0137] S503: Select a strategy that minimizes the change in the satellite velocity vector from the candidate avoidance strategy set as a target solution, and send the target solution to the satellite.

[0138] In the above embodiment, based on the satellite's autonomous avoidance capability and mission assurance requirements, a multi-level constraint framework and composite optimization objectives are constructed, as follows:

[0139] Physical Constraint Layer:

[0140] Propulsion system constraints: Integrate the satellite's propellant reserve, thruster operating time limit, and attitude control mechanism motion range to generate the velocity vector change boundary allowed for avoidance maneuvers;

[0141] Thermal control constraints: Combine the satellite's solar panel pointing direction, the exposure angle of the thermal radiation surface, and the propellant temperature threshold to avoid maneuvering trajectories that may cause local overheating;

[0142] Communication constraints: Ensure continuous visibility with the ground station during avoidance and prevent the antenna pointing angle from exceeding the communication coverage range.

[0143] Task constraint layer:

[0144] Payload continuity: For remote sensing satellites, imaging area overlap must be maintained; for communication satellites, beam coverage must be uninterrupted, and a mission quality degradation tolerance curve must be generated;

[0145] Orbit maintenance requirements: Evaluate the impact of avoidance maneuvers on the semi-major axis and eccentricity of the satellite orbit to ensure that the orbital parameters after avoidance still meet the mission life requirements.

[0146] Optimize target layer:

[0147] Main goal: Minimize the velocity increment required for avoidance maneuvers to extend the satellite's on-orbit life and reduce propellant consumption;

[0148] Secondary goal: prioritize avoidance trajectories that have minimal impact on the satellite's attitude, reducing the energy consumption of the gyroscopes and flywheels required for attitude adjustments;

[0149] Additional goal: If there are multi-target avoidance requirements, the avoidance time window can be decentralized through weight distribution to reduce the risk of space debris environment.

[0150] An improved particle swarm optimization algorithm is used to perform distributed solution of the multi-objective optimization model, and the avoidance strategy parameters are encoded as particle position vectors. The control accuracy of the propulsion system is adapted through discretization processing. A composite evaluation function is generated based on the constraint conditions, which includes a velocity increment penalty term, an attitude stability score term, a mission quality attenuation term, and a collision probability reduction term. Multi-objective balance is achieved through dynamic weight adjustment. A particle population stratification strategy is introduced to decompose the avoidance problem into three subgroups: time window selection, velocity increment allocation, and attitude control. Information sharing is used to achieve collaborative optimization among subgroups. A simulated annealing algorithm is introduced in the later stage of PSO iteration to perform perturbation search on high-quality particles to avoid falling into local optimal solutions. A candidate avoidance strategy set is generated through mutation operators and crossover operators to ensure that the strategy set contains alternative plans with different avoidance directions, time windows, and operation durations.

[0151] Candidate strategies are quickly simulated and verified through the satellite digital twin model to eliminate high-risk strategies that may cause attitude loss of control or communication interruption; the impact of each strategy on the satellite's fuel balance, mission completion and on-orbit life is quantified to generate a strategy-cost curve; disturbance simulation is performed within the range of ±10% of the strategy parameters to evaluate the strategy's fault tolerance for orbit prediction errors and propulsion system response delays.

[0152] Candidate strategies are prioritized and ranked according to the mission urgency, satellite value level, and irreversibility of avoidance operations. Strategies with the smallest velocity vector change and the shortest avoidance time window are prioritized to reduce interference with the satellite's routine missions. For high-value satellites, 2-3 suboptimal strategies are retained as emergency backups to ensure rapid switching when the main strategy fails.

[0153] See also Figure 6 , Figure 6 The present invention provides an architectural diagram of an embodiment of a satellite collision avoidance system based on a satellite edge computing network, which includes a space monitoring network for observing the operating status of targets in space to obtain multi-source observation data; a space traffic management subsystem for integrating multi-source observation data to generate space target orbital data; wherein the satellite edge computing network also includes: a GEO satellite deployed in a geosynchronous orbit, which is used to respond to computing tasks through an edge computing server, broadcast space target orbital data to a LEO satellite constellation through an inter-satellite link signal transmitter, and forward collision risk screening information to a ground cloud computing center; a LEO satellite constellation composed of multiple low-orbit satellites, which is used to screen and calculate collision risk screening information based on space target orbital data and formulate collision avoidance control strategies; a LEO / GEO earth station deployed on the ground, which is used for data reception and transmission and protocol conversion between GEO satellites, LEO satellite constellations and cloud computing centers; a ground cloud computing center connected to GEO satellites and LEO satellite constellations through LEO / GEO earth stations, which is used to store constellation orbit data through protocol conversion and satellite communication.

[0154] The present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above method. The computer-readable storage medium may include, but is not limited to, any type of disk, including a floppy disk, an optical disk, a DVD, a CD-ROM, a microdrive, a magneto-optical disk, a ROM, a RAM, an EPROM, an EEPROM, a DRAM, a VRAM, a flash memory device, a magnetic card or an optical card, a nanosystem (including a molecular memory IC), or any type of medium or device suitable for storing instructions and / or data.

[0155] It should be noted that for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

[0156] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0157] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interface, and the indirect coupling or communication connection of the device or unit can be electrical or other forms.

[0158] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0159] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0160] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: various media that can store program codes, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0161] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0162] The above is only an exemplary embodiment of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure herein, those skilled in the art will easily think of the implementation scheme of the present disclosure. This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art that are not recorded in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

[0163] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0164] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A satellite collision avoidance system based on a satellite edge computing network, characterized in that: include: Space monitoring network, space traffic management subsystem, and satellite edge computing network; The space monitoring network is used to observe the operating status of targets in space and obtain multi-source observation data; The space traffic management subsystem is used to integrate the multi-source observation data to generate space target orbit data; The satellite edge computing network also includes: GEO satellites deployed in geosynchronous orbit are used to respond to computing tasks through edge computing servers, broadcast space target orbit data to the LEO satellite constellation through inter-satellite link signal transmitters, and forward collision risk screening information to ground cloud computing centers; A LEO satellite constellation consisting of multiple low-orbit satellites is used to screen and calculate collision risk information based on space target orbit data and formulate collision avoidance control strategies; LEO / GEO earth stations deployed on the ground are used for data transmission and reception and protocol conversion between GEO satellites, LEO satellite constellations and cloud computing centers; A ground cloud computing center connected to the GEO satellite and LEO satellite constellation through the LEO / GEO earth station is used to communicate with the satellites through protocol conversion and store constellation orbit data.

2. The satellite collision avoidance system based on the satellite edge computing network according to claim 1, characterized in that: The space monitoring network includes ground-based radars and optical telescopes; The land-based radar is used to scan space targets using a multi-band radar beam to obtain target distance and velocity vector data; The optical telescope is used to capture the angular characteristics and shape characteristics of space targets through high-resolution imaging technology.

3. The satellite collision avoidance system based on satellite edge computing network according to claim 1, characterized in that: The LEO satellite constellation collision warning screening process includes geometric screening, time screening and probability calculation; The geometric screening is used to perform perigee / apogee screening and orbit intersection distance analysis based on the space target orbit data to determine a suspected collision target; The time screening analyzes the time difference of the suspected collision targets passing through the track intersection line, and screens out suspected collision targets whose time windows do not overlap; The probability calculation uses a preset orbit prediction model to calculate the collision probability of the remaining collision targets.

4. The satellite collision avoidance system based on satellite edge computing network according to claim 3, characterized in that: The LEO satellite constellation also includes an early warning module and a confidence module; Wherein, the warning module is used to trigger a warning when the collision probability exceeds a preset probability threshold; The confidence module is used to initiate an encrypted observation request to re-acquire the space target orbit data when the confidence of the space target orbit data is lower than a preset confidence threshold.

5. The satellite collision avoidance system based on satellite edge computing network according to claim 3, characterized in that: The LEO satellite constellation also includes a collision avoidance decision module; The collision avoidance decision module is used to set constraints, establish a constrained multi-objective optimization model, formulate a collision avoidance control strategy, calculate the satellite velocity vector change, and execute the collision avoidance control strategy with the smallest satellite velocity vector change.

6. A satellite collision avoidance method based on a satellite edge computing network, based on the satellite collision avoidance system based on a satellite edge computing network as described in any one of claims 1 to 5 above, characterized in that: include: The ground-based radar uses a multi-band radar beam to scan space targets to obtain observation data on target range and velocity vector, and the optical telescope uses high-resolution imaging to capture observation data on target angular characteristics and shape characteristics; Integrate observation data from ground-based radars and optical telescopes to generate orbital data of space targets, and upload the data synchronously to GEO satellites in geosynchronous orbit; The GEO satellite broadcasts the orbital data of the space target to the LEO satellite constellation, and performs geometric screening, time window matching, and collision probability calculation in sequence; When the collision probability exceeds the preset threshold, an early warning is triggered, and a multi-objective optimization model is established to generate a collision avoidance control strategy.

7. The satellite collision avoidance system based on satellite edge computing network according to claim 6, characterized in that: The performing geometric screening further comprises: Extract the semi-major axis, eccentricity, and orbital inclination parameters from the space target orbit data and calculate the perigee and apogee heights; Based on the orbital plane equation of the target satellite, calculate the orbital intersection distance and intersection point coordinates; If the orbital altitude difference of the target satellite's perigee or apogee altitude is less than the preset altitude, and the orbital intersection distance is less than the preset safety distance, it will be marked as a suspected collision target.

8. The satellite collision avoidance system based on satellite edge computing network according to claim 6, characterized in that: The execution time window matching further includes: Calculate the time difference between the suspected collision target passing through the orbital intersection line based on the space target orbit data and the satellite's ephemeris; If the time difference between the suspected collision target and the track intersection is greater than the preset safety time window, the target is judged to have no collision risk.

9. The satellite collision avoidance system based on satellite edge computing network according to claim 6, characterized in that: The sequential execution of geometric screening, time window matching, and collision probability calculation further includes: Calculate the confidence level of space target orbit data based on the observed space target orbit covariance, target imaging clarity, and orbit prediction model error parameters; If the confidence level is lower than the preset confidence level, an encrypted observation request is initiated, and land-based radar and optical telescopes are designated to track and observe the target satellite. The encrypted observation data is fused with historical data through the Kalman filter algorithm to update the space target orbit data.

10. The satellite collision avoidance system based on satellite edge computing network according to claim 6, characterized in that: The method of triggering an early warning when the collision probability exceeds a preset threshold and establishing a multi-objective optimization model to generate a collision avoidance control strategy also includes: Setting constraints and establishing a multi-objective optimization model based on the constraints; The multi-objective optimization model is solved by particle swarm optimization algorithm to generate a set of candidate avoidance strategies; A strategy that minimizes the change in the satellite velocity vector is selected from the candidate avoidance strategy set as a target solution, and the target solution is sent to the target satellite.

Images