Cross-layer collaborative toughness evaluation method and system for low-orbit giant constellation

Abstract

The invention relates to the technical field of space-based information networks, and discloses a cross-layer collaborative toughness evaluation method and system for a low-orbit giant constellation, and the method comprises the steps: collecting the physical layer resource state and link layer communication characteristics of a satellite node in the constellation in real time; constructing a high-order state tensor based on the physical layer resource state and the link layer communication features, performing dimension reduction processing, and extracting low-dimension key manifold data; based on the low-dimensional key manifold data, a cross-layer coupling seepage model is constructed through physical constraint mapping, a toughness critical point of global connectivity phase change of a constellation system is analyzed and determined, and a real-time toughness margin is calculated; the real-time toughness margin is compared with a preset safety threshold value, and when the system state enters a phase change early warning area, a resource reconstruction and topology adjustment strategy is generated and issued. The invention provides a complete giant constellation toughness evaluation method from state sensing, evaluation and early warning to control.

Description

Technical Field

[0001] This invention relates to the field of space-based information network technology, and in particular to a cross-layer collaborative resilience assessment method and system for low-Earth orbit mega-constellations. Background Technology

[0002] With the rapid development of commercial space technology, low-Earth orbit mega-constellations have become a key infrastructure for building a globally covered, low-latency, high-bandwidth space-based information network. These constellations are evolving to the tens of thousands of satellites, exhibiting highly dynamic topologies, strong periodicity in spatiotemporal evolution, and deep coupling of cross-layer resources. As the core of space-based information network security and security range verification, accurately and in real-time assessing the resilience of mega-constellations after being subjected to natural environmental fluctuations, electronic interference, or malicious attacks is a major scientific problem for achieving batch autonomous operation of satellites and continuous improvement of system effectiveness.

[0003] For reliability and security assessment of space-based networks, existing technologies mainly rely on maximum flow or connectivity algorithms from classical graph theory to calculate the network's survivability under random node failures or targeted attacks. While this method can quantify the resilience of static network topologies, it has significant limitations when dealing with large-scale constellation scenarios: First, the model only focuses on the geometric connectivity of the link layer, failing to consider the constraints of physical resources such as the energy charging and discharging status and thermal control node temperature of the satellite platform on the upper-layer service flow, making it difficult to characterize the risk of cross-layer cascading failures due to physical layer "resource depletion"; Second, the time complexity of maximum flow and path search algorithms is typically... The above (of which) For the number of nodes, When dealing with giant constellations with tens of thousands of nodes (where the number of edges is 1), the computational overhead of matrix operations increases exponentially with the number of nodes, making it impossible to meet the needs of real-time evaluation.

[0004] Furthermore, the paper "Discussion on Capability Adaptation Methods of Space-Based Information Networks" (Aerospace Engineering, 2020, No. 2) proposes a multi-layered capability adaptation method oriented towards network protocol architecture, addressing the highly dynamic and strongly constrained characteristics of space-based network topology. This method enhances the network's ability to resist disturbances through flexible design. This research provides an important reference for the systematic construction of my country's space-based network, but a technological blind spot remains in the crucial "assessment-early warning" stage. Existing solutions mostly employ graph analysis methods based on eigenvalue decomposition, but calculating the eigenvalue decomposition of the entire network's Laplacian matrix at the scale of tens of thousands of satellites incurs extremely high computational overhead, with a complexity reaching [insert value here]. This greatly limits its application on spaceborne constrained computing platforms.

[0005] More importantly, when giant constellations suffer large-scale disturbances, their system failure process is not a linear performance drop, but rather exhibits a "non-equilibrium phase transition" characteristic, progressing from local damage to global paralysis. Research has found that when network connectivity is damaged to near a critical point, the spectral gap of the system's Laplace matrix, i.e., the first non-zero eigenvalue, rapidly approaches zero, signifying a complete loss of the system's synchronization capability and robustness. However, existing graph theory methods and adaptive models often focus on recovery strategies after system failure, lacking a real-time mechanism for capturing these phase transition characteristics, thus failing to provide accurate early warnings before system collapse.

[0006] In summary, given the complexity of the deep coupling between physical layer resource constraints and network layer topology evolution in the context of mega-constellation environments, there is an urgent need for a resilience assessment method that can overcome computational bottlenecks, finely characterize cross-layer physical constraints, and capture critical characteristics of non-equilibrium phase transitions in real time, in order to build a resilience assessment foundation that supports end-to-end security and trustworthiness. Summary of the Invention

[0007] The purpose of this invention is to address the shortcomings of the existing technology and provide a cross-layer collaborative resilience assessment method for low-Earth orbit mega-constellations. The method aims to construct an assessment system through cross-layer mapping of physical and information, high-order tensor dimensionality reduction prediction, and non-equilibrium seepage analysis to improve the intrinsic safety level of mega-constellations under extreme disturbances.

[0008] On the one hand, this invention provides a method for assessing cross-layer collaborative resilience for low-Earth orbit mega-constellations, comprising the following steps: S1: Real-time acquisition of physical layer resource status and link layer communication characteristics of satellite nodes in the constellation; S2: Construct a high-order state tensor based on the physical layer resource state and link layer communication characteristics, and perform dimensionality reduction processing to extract low-dimensional key manifold data; S3: Based on the low-dimensional key manifold data, a cross-layer coupled seepage model is constructed through physical constraint mapping to analytically determine the toughness critical point of the global connectivity phase transition of the constellation system and calculate the real-time toughness margin. S4: Compare the real-time resilience margin with the preset safety threshold. When the system state enters the phase transition warning zone, generate and issue resource reconstruction and topology adjustment strategies.

[0009] Further, in step S2, the construction of the higher-order state tensor and the dimensionality reduction process include: The collected data is mapped to a fourth-order tensor. ,in, The total number of satellite nodes. For the observation time step, The number of physical state characteristics. The number of link communication features; The fourth-order tensor is decomposed using the Tucker decomposition algorithm. Compressed into key tensor With each dimension factor matrix The product form, i.e. ; The key tensor Input the Transformer self-attention prediction engine, utilize its multi-head self-attention mechanism to capture the spatiotemporal dependency features of interstellar topology, and output the low-dimensional key manifold data.

[0010] Preferably, in the dimensionality reduction process, a randomized Tucker decomposition algorithm and a linear self-attention mechanism based on kernel function approximation are employed, and the computational complexity of self-attention is reduced from [previous level] to [current level] by utilizing the matrix multiplication associative law. Reduce to This reduces the computational complexity of network resilience assessment to the level of critical tensor operations. .

[0011] Further, in step S3, the construction of the cross-layer coupled seepage model through physical constraint mapping includes: Based on the aforementioned low-dimensional key manifold data, a resource mapping function is constructed, expressed as follows: , in, For real-time satellite power supply, The battery is fully charged. For real-time thermal control of satellite temperature, The preset thermal control temperature safety threshold, For the real-time available bandwidth of the satellite link, To design for maximum bandwidth, As a resource sensitivity factor, , and These are the normalized weighted coefficients corresponding to the factors mentioned above; Using the resource mapping function The physical layer resource status of each satellite node is mapped to the node occupancy probability in the seepage model. Mapping link-layer communication characteristics to edge penetration probabilities .

[0012] Preferably, in step S3, the analytical determination of the toughness critical point includes: Degree distribution function based on constellation network The occupancy probability and edge penetration probability A self-consistent equation describing the evolution of the global branch size of the system is constructed as follows: , in, This represents the probability that any node cannot connect to a branch of the giant network. Let the degree of the node be . The average degree of the network; Linear stability analysis was performed on the self-consistent equations to obtain the occupancy probability of the critical node where the system undergoes a phase transition. ,in, This represents the current average edge penetration probability across the entire network. , is the network topology heterogeneity parameter, representing the heterogeneity of the degree distribution in the constellation network, which is determined by the first and second moments of the network topology.

[0013] More preferably, in step S3, the calculation of the real-time toughness margin includes: Calculate the average survival probability of all nodes in the current network. , This indexes satellite nodes, thereby obtaining real-time resilience margins. ,when When the value approaches zero, it indicates that the constellation system is on the verge of a resilience collapse.

[0014] Further, in step S4, the generated resource reconstruction and topology adjustment strategy includes: Based on the inter-satellite long-range correlation weight matrix in the low-dimensional key manifold data Identify critical links and non-core links; The system generates instructions to disconnect non-core links that are in a high-energy-consuming state and whose associated weight is lower than a preset security threshold, and reroutes business traffic to backup paths with high resilience margins.

[0015] Preferably, in step S4, the strategy further includes power scheduling instructions: Reduce the power of low-priority payloads to prioritize ensuring the critical survivability of the satellite platform.

[0016] On the other hand, the present invention provides a cross-layer collaborative resilience assessment system for low-Earth orbit mega-constellations, comprising: The multi-dimensional state awareness module is used to collect the physical layer resource status and link layer communication characteristics of satellite nodes in the constellation in real time. The state data compression module is used to construct a high-order state tensor based on the physical layer resource state and link layer communication characteristics and perform dimensionality reduction processing to extract low-dimensional key manifold data. The cross-layer toughness assessment module is used to construct a cross-layer coupled seepage model based on the low-dimensional key manifold data through physical constraint mapping, analyze and determine the toughness critical point of the global connectivity phase transition of the constellation system, and calculate the real-time toughness margin. The resilience strategy generation module is used to compare the real-time resilience margin with the preset safety threshold. When the system state enters the phase change warning zone, it generates and issues resource reconstruction and topology adjustment strategies.

[0017] In addition, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the cross-layer collaborative resilience assessment method for low-Earth orbit mega-constellations as described above.

[0018] Meanwhile, an electronic device is provided, comprising: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement the cross-layer collaborative resilience assessment method for low-Earth orbit mega-constellations as described above.

[0019] Compared with the prior art, the beneficial effects of the present invention are: This invention combines multidimensional state perception with cross-layer resilience assessment to uniformly model and map the physical layer states of a satellite, such as energy and thermal control, with the characteristics of the link communication information layer. This breaks through the limitations of traditional methods that only analyze network topology, enabling the assessment results to truly reflect the risk of cascading failures caused by the depletion of physical resources, and significantly improving the intrinsic accuracy and authenticity of the assessment model. This invention utilizes tensor decomposition and attention mechanisms to reduce the dimensionality and extract features from massive state data, reducing the computational complexity of the whole network assessment from the exponential level of traditional graph theory methods to the near-linear level. This makes it possible to conduct real-time resilience assessment of constellations with tens of thousands of nodes under limited onboard computing power, providing engineering feasibility for online monitoring and situational awareness. This invention analyzes the phase transition critical point of system collapse using non-equilibrium seepage theory, enabling the system to quantitatively perceive its distance from the collapse boundary before global paralysis occurs, thereby achieving early warning of early risks and gaining a critical time window for proactive intervention. This invention generates resource reconstruction and topology adjustment strategies. Based on the assessment and early warning results, it automatically generates and distributes optimization strategies such as topology adjustment and resource reconstruction, transforming the traditional passive assessment into active defense, and directly improving the adaptive survival and recovery capabilities of the constellation system when attacked or disturbed. Attached Figure Description

[0020] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the process for evaluating the cross-layer collaborative resilience of a low-Earth orbit mega-constellation according to the present invention. Figure 2This is a schematic diagram of the logical architecture of a physical and information cross-layer mapping mechanism of the present invention; Figure 3 This is a flowchart of a high-dimensional state reduction process based on Tucker decomposition and Transformer according to the present invention; Figure 4 This invention relates to the occupancy probability of a constellation system evolving to a critical node. Phase transition characteristic curve at time; Figure 5 This is a comparison chart of the computational complexity of the evaluation method of the present invention on a spaceborne embedded platform. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] This invention is applied to low-Earth orbit satellite networks with tens of thousands of satellites. The method, run on an onboard central processor or a ground-based simulation verification platform, enables dynamic resilience assessment of the constellation system under conditions of debris impact, malicious interference, or natural failure.

[0023] The specific embodiments of the present invention will be described below with reference to the accompanying drawings and examples.

[0024] Example 1 Please see Figure 1 The technical solution for a cross-layer collaborative resilience assessment method for low-Earth orbit mega-constellations provided in this embodiment includes: Step S1, Multi-dimensional State Awareness: Real-time acquisition of the physical layer resource status and link layer communication characteristics of satellite nodes in the constellation.

[0025] Specifically, during the operation of the low-Earth orbit mega-constellation, onboard sensors are used to monitor the health status of each node in real time. For the satellite numbered **, its physical layer resource status is collected. This includes energy and electricity. Current remaining battery percentage; thermal control temperature Real-time temperature of the payload compartment.

[0026] Simultaneously collect link layer communication characteristics Including bandwidth : Available bandwidth of the inter-satellite link (ISL); Signal-to-noise ratio (SNR): Link quality metric at the receiver.

[0027] The above data were sampled at a period of 100ms and aggregated to form an initial observation matrix.

[0028] Step S2, State data compression and low-complexity feature extraction: Based on the physical layer resource state and link layer communication features, construct a high-order state tensor and perform dimensionality reduction processing to extract low-dimensional key manifold data.

[0029] In step S2, the construction of the higher-order state tensor and the dimensionality reduction process are performed, such as... Figure 3 As shown, it includes: S21: Map the collected data into a fourth-order tensor ,in, This embodiment represents the total number of satellite nodes. , For the observation time step, The number of physical state characteristics, in this embodiment Each corresponds to a different amount of energy. With thermal control temperature , This embodiment represents the number of link communication features. Each corresponds to available bandwidth With signal-to-noise ratio (SNR); S22: The fourth-order tensor is decomposed using the Tucker decomposition algorithm. Compressed into key tensor With each dimension factor matrix The product form, i.e. ; Among them, the key tensor The core manifold characterizing the state of the constellation system is used to eliminate high-entropy noise reflecting local random disturbances by setting a compression ratio.

[0030] S23: Transfer the key tensor Input the Transformer self-attention prediction engine, utilize its multi-head self-attention mechanism to capture the spatiotemporal dependency features of interstellar topology, and output the low-dimensional key manifold data.

[0031] This step captures the potential cascading failure trend of the system by calculating the association weight of each node with the states of other nodes.

[0032] Furthermore, in the dimensionality reduction process, by employing a randomized Tucker decomposition algorithm and a linear self-attention mechanism based on kernel function approximation, the computational complexity of self-attention is reduced from [previous level] to [current level] by utilizing the matrix multiplication associative law. Reduce to This reduces the computational complexity of network resilience assessment to the level of critical tensor operations. .

[0033] Specifically, such as Figure 5 As shown, this invention employs incremental randomized Tucker decomposition to remove noise, and combines it with a Performer self-attention operator with linear time complexity. This eliminates the need for eigenvalue decomposition of the entire Laplacian matrix in the computational state evolution operation, effectively reducing the computational complexity to [missing information]. Level. In simulations involving tens of thousands of stars, the evaluation time was reduced by more than 85% compared to traditional methods, ensuring the feasibility of real-time operation in spaceborne constrained computing modules.

[0034] Step S3, Physical constraint mapping and phase transition critical point analysis: Based on the low-dimensional key manifold data, a cross-layer coupled seepage model is constructed through physical constraint mapping, the toughness critical point of global connectivity phase transition of the constellation system is determined analytically, and the real-time toughness margin is calculated.

[0035] Specifically, the low-dimensional key manifold data output from step 2 is received, and the manifold data includes key feature tensors reflecting the evolution trend of resources across the entire network. And the correlation weight matrix that captures the correlation of node failures To eliminate the modeling isolation between information flow and physical field, this embodiment constructs a nonlinear resource mapping function based on the decoupled real-time satellite state values ​​from the manifold data.

[0036] Furthermore, the correlation weight matrix The multi-head attention score, derived from the Transformer self-attention prediction engine, is used to characterize the long-range spatiotemporal dependencies of interstellar topology. This embodiment uses the aforementioned correlation weight matrix... It is mapped to the edge permeation probability correction factor in the permeation model to reflect the heterogeneity of the link's contribution to the overall network connectivity.

[0037] In step S3, constructing the cross-layer coupled seepage model through physical constraint mapping includes: Based on the aforementioned low-dimensional key manifold data, a resource mapping function is constructed, expressed as follows: , in, For real-time satellite power supply, The battery is fully charged. For real-time thermal control of satellite temperature, The preset thermal control temperature safety threshold, For the real-time available bandwidth of the satellite link, To design for maximum bandwidth, As a resource sensitivity factor, , and These are the normalized weighted coefficients corresponding to the factors mentioned above; This function couples the star's energy, temperature, and communication resources across layers. In specific engineering numerical implementations, to avoid... or For computational singularities where the denominator approaches zero, regularization can be used to replace the denominator with zero. and ,in It is a preset machine precision floating-point number to ensure the stability of the calculation.

[0038] Assuming current satellite energy Even if the communication bandwidth drops to 50% of the critical value, Normal, the survival probability after mapping It will also decay exponentially, which truly simulates the cross-layer coupling process of "insufficient physical energy efficiency leading to communication interruption".

[0039] Using the resource mapping function The physical layer resource status of each satellite node is mapped to the node occupancy probability in the seepage model. Mapping link-layer communication characteristics to edge penetration probabilities .

[0040] like Figure 2 and Figure 4 As shown, the survival probability of each node is substituted into the non-equilibrium seepage self-consistent equation. In this embodiment, the node occupancy probability... Defined as the average survival probability after mapping all satellite nodes in the entire network, i.e. Edge penetration probability It is obtained by normalizing the average signal-to-noise ratio of the inter-satellite links.

[0041] Secondly, the analytical determination of the toughness critical point includes: Degree distribution function based on constellation network The occupancy probability and edge penetration probability A self-consistent equation describing the evolution of the global branch size of the system is constructed as follows: , in, This represents the probability that any node cannot connect to a branch of the giant network. Let the degree of the node be . The average degree of the network; The equation is then solved using Newton's iterative method, and combined with linear stability analysis, the critical threshold for the system to experience global collapse is calculated.

[0042] Specifically, a linear stability analysis is performed on the self-consistent equation to obtain the occupancy probability of the critical node where the system undergoes a phase transition. ,in, This represents the current average edge penetration probability across the entire network. , is the network topology heterogeneity parameter, representing the heterogeneity of the degree distribution in the constellation network, which is determined by the first and second moments of the network topology.

[0043] Secondly, the calculation of the real-time toughness margin in step S3 includes: Calculate the average survival probability of all nodes in the current network. , This indexes satellite nodes, thereby obtaining real-time resilience margins. ,when When the value approaches zero, it indicates that the constellation system is on the verge of a resilience collapse.

[0044] This evaluation method based on phase transition theory can detect signs of system instability 3-5 time steps earlier than traditional algorithms based on connectivity mean.

[0045] Step S4, Resilience Strategy Generation and Evaluation Optimization: Compare the real-time resilience margin with the preset safety threshold. When the system state enters the phase transition warning zone, generate and issue resource reconstruction and topology adjustment strategies.

[0046] In step S4, the generated resource reconstruction and topology adjustment strategy includes: Based on the inter-satellite long-range correlation weight matrix in the low-dimensional key manifold data Identify critical links and non-core links; The system generates instructions to disconnect non-core links that are in a high-energy-consuming state and whose associated weight is lower than a preset security threshold, and reroutes business traffic to backup paths with high resilience margins.

[0047] In addition, the strategy also includes power scheduling instructions: Reduce the power of low-priority payloads to prioritize ensuring the critical survivability of the satellite platform.

[0048] Specifically, the system monitors resilience margin in real time. .like Among them, the preset safety threshold Preferred setting is If the threshold of 15% is reached, the system is considered to have entered the instability warning zone of a non-equilibrium phase transition. At this point, although the system has not yet experienced a global connectivity collapse, it is near the critical point of the phase transition, and even minor disturbances can induce cascading failures.

[0049] At this point, the control center issues an instruction based on the resource reconfiguration algorithm: Instruction 1: Topology Adjustment: Based on the inter-satellite long-range correlation weight matrix generated in step S2 Identify critical links; disconnect non-core links that are in a high-temperature, high-energy-consumption state and whose association weight is lower than the preset safety threshold, and reroute traffic to backup paths with higher resilience margins, thereby reducing thermal control pressure while ensuring system connectivity; Command 2 Power Scheduling: Reduce the power of low-priority payloads and prioritize the critical survivability energy consumption of the satellite platform.

[0050] Through the above assessment and adjustment, the system state is brought back to the safe zone, achieving intrinsic security.

[0051] In summary, this invention provides a complete method for assessing the resilience of giant constellations, from state perception to assessment and early warning, and then to control, by introducing a non-equilibrium physical model and computational dimensionality reduction techniques.

[0052] Based on the above method, the present invention also provides a cross-layer collaborative resilience assessment system for low-Earth orbit mega-constellations, comprising: The multi-dimensional state awareness module is used to collect the physical layer resource status and link layer communication characteristics of satellite nodes in the constellation in real time. The state data compression module is used to construct a high-order state tensor based on the physical layer resource state and link layer communication characteristics and perform dimensionality reduction processing to extract low-dimensional key manifold data. The cross-layer toughness assessment module is used to construct a cross-layer coupled seepage model based on the low-dimensional key manifold data through physical constraint mapping, analyze and determine the toughness critical point of the global connectivity phase transition of the constellation system, and calculate the real-time toughness margin. The resilience strategy generation module is used to compare the real-time resilience margin with the preset safety threshold. When the system state enters the phase change warning zone, it generates and issues resource reconstruction and topology adjustment strategies.

[0053] It should be noted that the steps in the cross-layer collaborative resilience assessment method for low-Earth orbit mega-constellations provided in this embodiment can be implemented based on the corresponding modules in the cross-layer collaborative resilience assessment system for low-Earth orbit mega-constellations. Those skilled in the art can refer to the technical solution of the system to implement the steps of the method. That is, the embodiments in the system can be understood as preferred examples of implementing the method, and will not be elaborated here.

[0054] Besides implementing the system and its various devices provided by this invention in purely computer-readable program code, the same functions can be achieved by logically programming the method steps, making the system and its various devices of this invention appear as logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices provided by this invention can be considered as a hardware component, and the devices included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0055] Finally, it should be noted that the above description is only a preferred embodiment of the present invention, and the scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be pointed out that for those skilled in the art, any improvements and modifications made without departing from the principle of the present invention should also be considered within the scope of protection of the present invention.

[0056] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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.

Claims

1. A method for assessing cross-layer collaborative resilience for low-Earth orbit mega-constellations, characterized in that, Includes the following steps: S1: Real-time acquisition of physical layer resource status and link layer communication characteristics of satellite nodes in the constellation; S2: Construct a high-order state tensor based on the physical layer resource state and link layer communication characteristics, and perform dimensionality reduction processing to extract low-dimensional key manifold data; S3: Based on the low-dimensional key manifold data, a cross-layer coupled seepage model is constructed through physical constraint mapping to analytically determine the toughness critical point of the global connectivity phase transition of the constellation system, and calculate the real-time toughness margin. S4: Compare the real-time resilience margin with the preset safety threshold. When the system state enters the phase transition warning zone, generate and issue resource reconstruction and topology adjustment strategies.

2. The method for assessing cross-layer collaborative resilience for low-Earth orbit mega-constellations according to claim 1, characterized in that, In step S2, the construction of the higher-order state tensor and the dimensionality reduction process include: The collected data is mapped to a fourth-order tensor. ,in, The total number of satellite nodes. For the observation time step, The number of physical state characteristics. The number of link communication features; The fourth-order tensor is decomposed using the Tucker decomposition algorithm. Compressed into key tensor With each dimension factor matrix The product form, i.e. ; The key tensor Input the Transformer self-attention prediction engine, utilize its multi-head self-attention mechanism to capture the spatiotemporal dependency features of interstellar topology, and output the low-dimensional key manifold data.

3. The method for assessing cross-layer collaborative resilience for low-Earth orbit mega-constellations according to claim 2, characterized in that, In the dimensionality reduction process, a randomized Tucker decomposition algorithm and a linear self-attention mechanism based on kernel function approximation are employed, utilizing the matrix multiplication associative law to reduce the computational complexity of self-attention from... Reduce to This reduces the computational complexity of network resilience assessment to the level of critical tensor operations. .

4. The method for assessing cross-layer collaborative resilience for low-Earth orbit mega-constellations according to claim 1, characterized in that, In step S3, constructing the cross-layer coupled seepage model through physical constraint mapping includes: Based on the aforementioned low-dimensional key manifold data, a resource mapping function is constructed, expressed as follows: ， in, For real-time energy and power of the satellite, The battery is fully charged. For real-time thermal control of satellite temperature, The preset thermal control temperature safety threshold, For the real-time available bandwidth of the satellite link, To design for maximum bandwidth, As a resource sensitivity factor, , and These are the normalized weighted coefficients corresponding to the factors mentioned above; Using the resource mapping function The physical layer resource status of each satellite node is mapped to the node occupancy probability in the seepage model. Mapping link-layer communication characteristics to edge penetration probabilities .

5. The method for assessing cross-layer collaborative resilience for low-Earth orbit mega-constellations according to claim 4, characterized in that, In step S3, the analytical determination of the toughness critical point includes: Degree distribution function based on constellation network The occupancy probability and edge penetration probability A self-consistent equation describing the evolution of the global branch size of the system is constructed as follows: ， in, This represents the probability that any node cannot connect to a branch of the giant network. Let be the degree of the node. The average degree of the network; Linear stability analysis was performed on the self-consistent equations to obtain the occupancy probability of the critical node where the system undergoes a phase transition. ,in, This represents the current average edge penetration probability across the entire network. , is the network topology heterogeneity parameter, representing the heterogeneity of the degree distribution in the constellation network, which is determined by the first and second moments of the network topology.

6. The method for assessing cross-layer collaborative resilience for low-Earth orbit mega-constellations according to claim 5, characterized in that, In step S3, the calculation of the real-time toughness margin includes: Calculate the average survival probability of all nodes in the current network. , This indexes satellite nodes, thereby obtaining real-time resilience margins. ,when When the value approaches zero, it indicates that the constellation system is on the verge of a resilience collapse.

7. The method for assessing cross-layer collaborative resilience for low-Earth orbit mega-constellations according to claim 1, characterized in that, In step S4, the generated resource reconstruction and topology adjustment strategy includes: Based on the inter-satellite long-range correlation weight matrix in the low-dimensional key manifold data Identify critical links and non-core links; The system generates instructions to disconnect non-core links that are in a high-energy-consuming state and whose associated weight is lower than a preset security threshold, and reroutes business traffic to backup paths with high resilience margins.

8. The method for assessing cross-layer collaborative resilience for low-Earth orbit mega-constellations according to claim 7, characterized in that, In step S4, the strategy further includes power scheduling instructions: Reduce the power of low-priority payloads to prioritize ensuring the critical survivability of the satellite platform.

9. A cross-layer collaborative resilience assessment system for low-Earth orbit mega-constellations, characterized in that, include: The multi-dimensional state awareness module is used to collect the physical layer resource status and link layer communication characteristics of satellite nodes in the constellation in real time. The state data compression module is used to construct a high-order state tensor based on the physical layer resource state and link layer communication characteristics and perform dimensionality reduction processing to extract low-dimensional key manifold data. The cross-layer toughness assessment module is used to construct a cross-layer coupled seepage model based on the low-dimensional key manifold data through physical constraint mapping, analyze and determine the toughness critical point of the global connectivity phase transition of the constellation system, and calculate the real-time toughness margin. The resilience strategy generation module is used to compare the real-time resilience margin with the preset safety threshold. When the system state enters the phase change warning zone, it generates and issues resource reconstruction and topology adjustment strategies.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the cross-layer collaborative resilience assessment method for low-Earth orbit mega-constellations as described in any one of claims 1-8.

11. An electronic device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the cross-layer collaborative resilience assessment method for low-Earth orbit mega-constellations as described in any one of claims 1-8.

Images