Method and system for distributed routing in multi-layer satellite networks

Abstract

The application provides a multi-layer satellite network distributed routing method and system, the method comprising: dividing a low-orbit satellite set into multiple low-orbit satellite regions, each low-orbit satellite region corresponding to at least one medium-high-orbit satellite as a regional control node; multiple medium-high-orbit satellites cooperatively generating an exit selection strategy for each low-orbit satellite region and issuing the exit selection strategy to the corresponding low-orbit satellite region; low-orbit satellites in a low-orbit satellite region taking the corresponding exit selection strategy as an inter-regional forwarding constraint and performing hop-by-hop autonomous forwarding in the low-orbit satellite region; and multiple medium-high-orbit satellites updating the exit selection strategy for each low-orbit satellite region based on updated regional network state. The multi-layer satellite network distributed routing method and system of the application can reduce the frequency of cross-region path reconfiguration while retaining the real-time response capability of hop-by-hop routing, thereby enhancing the overall stability and scalability of the system.

Description

Technical Field

[0001] This invention belongs to the technical field of wireless communication and relates to a distributed routing method and system for multi-layer satellite networks. Background Technology

[0002] Integrated space-air-ground networks represent a crucial development direction for mobile communication systems towards 2030 and beyond. As a vital component of these networks, satellite communication boasts broad development and application prospects due to its wide-area coverage, robustness, and flexible deployment capabilities. From the early Iridium system to recent giant low-Earth orbit (LEO) constellations like Starlink, satellite network communication technology has continuously evolved, gradually transforming from small-scale experimental systems into large-scale communication infrastructures with global coverage. LEO satellites, with their low latency and flexible networking capabilities, effectively support high-speed data services and global broadband access. Existing satellite network routing technologies largely revolve around LEO satellite networks, deploying numerous LEO satellites to reduce communication latency and increase network capacity. However, as constellation size continues to expand, the number of routing nodes and network state dimensions increase dramatically. Traditional centralized or node-by-node distributed routing mechanisms are gradually facing bottlenecks in terms of computational complexity, control overhead, and system stability. Relying solely on LEO satellites is no longer sufficient to achieve a balance between capacity, stability, and controllability. To further enhance network coverage and overall carrying capacity, existing satellite communication systems are beginning to introduce medium-Earth orbit and geostationary orbit satellites to form a multi-layered satellite communication system.

[0003] In a multi-layered satellite communication system, satellites at different orbital altitudes exhibit significant differences in coverage, motion characteristics, and resource capabilities, thus undertaking different functions. Low-Earth orbit (LEO) satellites operate at low altitudes and have low link latency, making them suitable for real-time forwarding of high-speed service data. However, their high-speed motion leads to frequent topology changes, limiting the observable range of a single satellite and making it difficult to maintain a stable global network view over long periods. Medium-Earth orbit (MEO) and geostationary orbit (GEO) satellites, on the other hand, have large coverage areas and relatively stable orbits, possessing more continuous link availability and stronger onboard computing capabilities, making them naturally suitable for undertaking network coordination and control functions within a regional area. Based on these characteristics, multi-layered satellite networks have gradually formed a hierarchical collaborative architecture where LEO satellites handle data forwarding, and MEO and GO satellites participate in regional coordination and control.

[0004] To address the aforementioned multi-layered structure, some existing technologies introduce clustering, partitioning, or domain-based mechanisms, employing a combination of centralized and distributed routing control methods. They also incorporate reinforcement learning algorithms for adaptive routing to improve network optimization. For example, control nodes are set up at the regional layer to make unified decisions on cross-regional routing: cross-domain paths are generated directly between regions, or the cross-regional path structure is determined first, and then the regional controller centrally calculates and executes the regional-level path after reaching the target region. This type of hierarchical routing method reduces the scale of routing solutions through regional management and adaptively enhances different routing levels through reinforcement learning algorithms. However, cross-regional routing still primarily relies on path-level reconfiguration for adjustment. Changes in local link states or service loads can easily and directly affect the cross-regional routing structure, gradually spreading to multiple regions under the distributed feedback mechanism. This triggers frequent cross-regional path reconstruction, leading to a significant increase in forwarding table updates, a rapid increase in control signaling overhead with network size, and a decrease in routing convergence speed. Meanwhile, despite the introduction of regional control nodes or intelligent decision-making modules, existing solutions have not effectively constrained cross-regional traffic behavior at the system feedback structure level, making it difficult to suppress the propagation of local disturbances within the network, thus affecting overall stability and scalability. Summary of the Invention

[0005] The purpose of this invention is to provide a distributed routing method and system for multi-layer satellite networks, which can reduce the frequency of cross-regional path reconfiguration while retaining the real-time response capability of hop-by-hop routing, thereby reducing control signaling overhead and enhancing the overall stability and scalability of the system.

[0006] In a first aspect, the present invention provides a distributed routing method for a multi-layer satellite network, the method comprising the following steps: dividing a set of low-Earth orbit (LEO) satellites into multiple LEO satellite regions, each LEO satellite region corresponding to at least one medium-high orbit (MEO) satellite as a region control node; multiple MEO satellites collaboratively generating an exit selection strategy for each LEO satellite region based on the region-level network status of each LEO satellite region, and distributing the exit selection strategy to the corresponding LEO satellite region; the exit selection strategy is used to provide cross-regional forwarding direction constraints for service flows entering the LEO satellite region, indicating the set of exit nodes and the set of exit directions for the service flow within the current LEO satellite region; the LEO satellites within the LEO satellite region perform hop-by-hop autonomous forwarding within their respective LEO satellite regions using the corresponding exit selection strategy as an inter-regional forwarding constraint; and the multiple MEO satellites updating the exit selection strategy of each LEO satellite region based on the updated region-level network status.

[0007] In one implementation of the first aspect, the collaborative generation of exit selection strategies for each low-Earth orbit (LEO) satellite region by multiple medium- and high-Earth orbit (MEO) satellites based on the regional-level network status of each LEO satellite region includes the following steps:

[0008] The medium and high orbit satellites periodically collect the regional network status of each low orbit satellite region;

[0009] The medium- and high-orbit satellites construct regional-level status observations based on the regional-level network status.

[0010] Based on the aforementioned regional state observations, multiple medium- and high-orbit satellites employ a distributed multi-agent reinforcement learning strategy for collaborative training to acquire joint regional actions.

[0011] A reward function is generated based on the joint regional actions and the updated regional network state.

[0012] The medium- and high-orbit satellites exchange local policy parameters or local state summaries of the agents through inter-satellite control links, and perform collaborative parameter updates based on the reward function, so as to converge the exit selection strategy of each low-orbit satellite region under the framework of centralized training, distributed execution and parameter collaborative update.

[0013] In one implementation of the first aspect, each medium-high orbit satellite acquires the regional network state based on a partially observable Markov decision process, the regional network state including one or more combinations of regional average link congestion, regional service load level, regional egress reachability, and connectivity status of adjacent regions.

[0014] In one implementation of the first aspect, the regional-level state observation ,in This indicates the regional network status of the low-Earth orbit satellite area. This represents a partially observable mapping function.

[0015] In one implementation of the first aspect, the low-Earth orbit satellites within the low-Earth orbit satellite region perform hop-by-hop autonomous forwarding within their respective low-Earth orbit satellite regions using a corresponding exit selection strategy as an inter-regional forwarding constraint, comprising the following steps:

[0016] The low-orbit satellite will use the exit node and exit direction of the corresponding exit selection strategy as the routing target within the region.

[0017] The low-Earth orbit satellite selects the next-hop satellite within the low-Earth orbit satellite region based on local link status information using a distributed adaptive routing algorithm, so that the service flow advances hop by hop along the exit direction and converges to the exit node.

[0018] In one implementation of the first aspect, the local link status information includes one or more combinations of link delay with adjacent low-Earth orbit satellites, queue length, available bandwidth, link availability, and link error / packet loss indication.

[0019] In one implementation of the first aspect, the medium-high orbit satellite updates the exit selection strategy only when preset update conditions are met. The update conditions include one or more combinations of the following: regional exit congestion in the low orbit satellite region exceeds a preset threshold, inter-regional reachability changes, and regional service load fluctuations exceed a preset value.

[0020] In one implementation of the first aspect, the exit selection strategy is updated according to a first time scale, and the hop-by-hop autonomous forwarding of the low-orbit satellite is performed according to a second time scale, wherein the second time scale is smaller than the first time scale.

[0021] In one implementation of the first aspect, the exit selection strategy has a preset freeze window and a preset validity period; within the freeze window, cross-regional route reconfiguration is prohibited, and low-Earth orbit (LEO) satellites are only allowed to adaptively adjust to link congestion, node failure, or service load changes within their respective LEO satellite regions through hop-by-hop autonomous forwarding, so as to limit the propagation of local disturbances generated within the region to their respective LEO satellite regions and avoid the cross-regional spread of local disturbances and the triggering of cross-regional path reconstruction; within the preset validity period, the LEO satellites perform hop-by-hop autonomous forwarding within the LEO satellite region according to the exit selection strategy, without triggering immediate updates to the exit selection strategy.

[0022] Secondly, the present invention provides a multi-layer satellite network distributed routing system, including multiple medium-high orbit satellites and multiple low orbit satellites; the multiple low orbit satellites are divided into multiple low orbit satellite regions, and each low orbit satellite region corresponds to at least one medium-high orbit satellite as a region control node.

[0023] The multiple medium- and high-orbit satellites are used to collaboratively generate exit selection strategies for each low-orbit satellite region based on the regional network status of each low-orbit satellite region, and to distribute the exit selection strategies to the corresponding low-orbit satellite regions. The exit selection strategies are used to provide cross-regional forwarding direction constraints for service flows entering the low-orbit satellite region, indicating the set of exit nodes and the set of exit directions for the service flow in the current low-orbit satellite region.

[0024] The low-orbit satellite is used to perform hop-by-hop autonomous forwarding within its respective low-orbit satellite region, with the corresponding exit selection strategy as the inter-regional forwarding constraint.

[0025] The medium and high orbit satellites are also used to update the exit selection strategy for each low orbit satellite region based on the updated regional network status.

[0026] As described above, the multi-layer satellite network distributed routing method and system of the present invention have the following beneficial effects:

[0027] (1) It can reduce the frequency of cross-regional path reconfiguration while retaining the real-time response capability of hop-by-hop routing, thereby reducing control signaling overhead and enhancing the overall stability and scalability of the system;

[0028] (2) The cross-regional route adjustment was changed from path-level reconfiguration to regional exit-level control, which reduced the probability of frequent cross-regional path reconstruction;

[0029] (3) By retaining the node-by-node autonomous forwarding mechanism of low-orbit satellites, the real-time response capability to local link changes has been improved;

[0030] (4) Utilize the wide-area observability advantage of medium and high orbit satellites to achieve regional coordination and effectively suppress the spread of local disturbances within the network. Attached Figure Description

[0031] Figure 1 The flowchart shown is an embodiment of the multi-layer satellite network distributed routing method of the present invention;

[0032] Figure 2 The diagram shown is a schematic representation of the hierarchical control architecture in one embodiment of the multi-layer satellite network distributed routing method of the present invention;

[0033] Figure 3 This diagram illustrates hop-by-hop forwarding within an area in one embodiment of the multi-layer satellite network distributed routing method of the present invention.

[0034] Figure 4 The diagram shown is a structural schematic of a multi-layer satellite network distributed routing system according to an embodiment of the present invention. Detailed Implementation

[0035] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0036] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0037] Furthermore, in this invention, descriptions involving "first," "second," etc., are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature. Additionally, the technical solutions of the various embodiments can be combined with each other, but only on the basis of being achievable by those skilled in the art. When the combination of technical solutions is contradictory or impossible to implement, such a combination of technical solutions should be considered non-existent and not within the scope of protection claimed by this invention.

[0038] In the multi-layer satellite network distributed routing method and system of this invention, the satellite network adopts a hierarchical control architecture that combines upper-layer inter-regional routing with lower-layer intra-regional hop-by-hop forwarding. Medium- and high-orbit satellites, based on the regional network state, collaboratively generate regional exit selection strategies through distributed multi-agent reinforcement learning, providing directional constraints for cross-regional forwarding. Low-orbit satellites, under the constraints of the exit selection strategy, perform hop-by-hop autonomous forwarding based on their local link state, achieving rapid adaptive adjustment within the region and limiting local link disturbances to propagation within their respective regions. Therefore, this invention effectively improves the overall system stability and transmission performance through the collaborative work of cross-regional slow-timescale decision-making and intra-regional fast-timescale adaptation.

[0039] like Figure 1 As shown, in one embodiment, the multi-layer satellite network distributed routing method of the present invention includes steps S1-S4.

[0040] Step S1: Divide the low-Earth orbit satellite set into multiple low-Earth orbit satellite regions, with each low-Earth orbit satellite region corresponding to at least one medium-high orbit satellite as a region control node.

[0041] Specifically, this invention breaks down the end-to-end service routing process into upper-layer inter-regional routing for region-level forwarding direction decision-making and lower-layer intra-regional routing for hop-by-hop forwarding within the intra-region. The upper-layer inter-regional routing determines the target LEO satellite region or forwarding direction that the service flow should point to within the LEO satellite region, while the lower-layer intra-regional routing completes hop-by-hop forwarding within the current LEO satellite region under given forwarding direction constraints.

[0042] In one embodiment, the upper-layer inter-regional routing is generated by medium- and high-Earth orbit satellites based on regional network status to provide cross-regional forwarding direction constraints for service flows entering low-Earth orbit satellite regions from one low-Earth orbit satellite region to another. The lower-layer intra-regional routing is performed by low-Earth orbit satellites using hop-by-hop autonomous forwarding based on local link state information, allowing service flows to gradually advance along the forwarding direction without re-performing cross-regional decisions. This approach decomposes the originally end-to-end centralized routing process into two decoupled sub-processes: cross-regional direction decision-making and intra-regional hop-by-hop forwarding, thereby achieving a synergy between stable cross-regional routing structure and rapid intra-regional adaptation. Figure 2 As shown, the multi-layer satellite network distributed routing system of the present invention consists of a low-Earth orbit (LEO) satellite network and a medium-high Earth orbit (MEO) satellite network. In the LEO satellite layer, each LEO satellite establishes communication connections with its adjacent satellites in the same orbit and across orbits through inter-satellite links to perform hop-by-hop autonomous forwarding within the region. In the MEO satellite layer, each MEO satellite acts as a regional control node to perform regional-level status acquisition and collaborative generation of exit strategies.

[0043] In this invention, the low-Earth orbit (LEO) satellites are divided into multiple LEO regions based on orbital parameters, inter-satellite link connectivity, and a preset coverage area. Each LEO region is associated with at least one medium-to-high orbit (MEO) satellite as a region control node. In one embodiment, a multi-layer satellite network is modeled. Let the LEO satellite set be: Low-Earth orbit satellites are divided into multiple regions: Inter-region connectivity constitutes regional topology: ,in, This indicates the reachability relationship between regions. For any region... Its regional network state is defined as follows: ,in, Indicates the average link congestion level within the region. Indicates the regional business load level. Indicates regional export accessibility. This indicates the connectivity status of adjacent regions.

[0044] Step S2: Multiple medium- and high-orbit satellites collaboratively generate exit selection strategies for each low-orbit satellite region based on the regional network status of each low-orbit satellite region, and distribute the exit selection strategies to the corresponding low-orbit satellite regions; the exit selection strategies are used to provide cross-regional forwarding direction constraints for service flows entering the low-orbit satellite region, indicating the set of exit nodes and the set of exit directions for the service flow in the current low-orbit satellite region.

[0045] Specifically, the collaborative generation of exit selection strategies for each low-Earth orbit (LEO) satellite region by multiple medium- and high-Earth orbit (MEO) satellites based on the regional network status of each LEO satellite region includes the following steps:

[0046] 21) The medium- and high-orbit satellites periodically collect the regional network status of each low-orbit satellite region.

[0047] Each medium-high orbit satellite obtains the regional network status based on a partially observable Markov decision process (Dec-POMDP). The regional network status includes one or more combinations of the following: average link congestion level within the region, regional service load level, regional egress reachability, and connectivity status of adjacent regions.

[0048] 22) The medium- and high-orbit satellites construct regional-level status observations based on the regional-level network status.

[0049] Let the set of intelligent agents corresponding to medium and high orbit satellites be: Each intelligent agent Corresponding to a low-Earth orbit satellite region For any given moment , No. These high-orbit satellites obtain regional-level status observations of their corresponding areas. ,in This indicates the regional network status of the low-Earth orbit satellite area. This represents a partially observable mapping function. Wherein, Used for At least some of the state variables are filtered, normalized, encoded, or vectorized to obtain the regional state observation of the k-th medium-high orbit satellite at time t. In this invention... Instead of limiting to a specific algorithm, multiple algorithmic strategies can be used to address the issue. Process it.

[0050] 23) Based on the aforementioned regional state observations, multiple medium- and high-orbit satellites employ a distributed multi-agent reinforcement learning strategy for collaborative training to acquire regional joint actions.

[0051] Each medium- and high-orbit satellite is based on its regional-level state observations. Selecting an exit area: ,in Indicates low-Earth orbit satellite region At any moment The selected export node and export direction This represents the exit selection strategy function for the corresponding low-Earth orbit satellite region. Multiple medium- and high-Earth orbit satellites jointly form a regional coordinated action: .

[0052] 24) Generate a reward function based on the joint regional actions and the updated regional network state.

[0053] Specifically, the regional network state is updated based on joint regional actions, and a reward function is returned: The reward function mentioned above It is associated with at least one or more combinations of cross-regional service latency, regional load balancing, and regional egress congestion levels.

[0054] 25) The medium and high orbit satellites exchange local policy parameters or local state summaries of the agents through inter-satellite control links, and perform collaborative parameter updates based on the reward function, so as to converge the exit selection strategy of each low orbit satellite region under the framework of centralized training, distributed execution and parameter collaborative update.

[0055] In this process, each medium-to-high orbit satellite exchanges local state summaries or policy parameters of its respective agent through inter-satellite control links, and performs collaborative parameter updates based on the reward function, so as to gradually converge to obtain the regional exit selection strategy of each low orbit satellite under the framework of centralized training and distributed execution. This enables joint optimization of cross-regional forwarding directions.

[0056] Step S3: Low-Earth orbit satellites within the low-Earth orbit satellite region use the corresponding exit selection strategy as the inter-regional forwarding constraint and perform hop-by-hop autonomous forwarding within their respective low-Earth orbit satellite regions.

[0057] Specifically, when a service flow enters a certain LEO satellite region, the LEO satellites within that region determine the target exit node and exit direction for the service flow based on the corresponding region's exit selection strategy. This maps the routing results between upper-layer regions to hop-by-hop forwarding constraints within the LEO satellite region. Subsequently, each LEO satellite within the region performs hop-by-hop autonomous forwarding based solely on its local link state information. This local link state information includes one or more combinations of link delay with neighboring LEO satellites, queue length, available bandwidth, link availability, and link error / packet loss indications.

[0058] like Figure 3 As shown, after receiving the exit selection policy for the corresponding region, the low-Earth orbit satellite takes the exit node and exit direction as the forwarding target within the region. Based on local link status information such as link delay, queue length, available bandwidth, link availability, or link packet loss indication with neighboring low-Earth orbit satellites, it uses a distributed adaptive routing algorithm to perform hop-by-hop autonomous forwarding, so that the service flow gradually advances within the low-Earth orbit satellite region and converges to the exit node.

[0059] In one embodiment, the hop-by-hop autonomous forwarding process within the low-Earth orbit satellite region is formally described using a local forwarding selection function. Let the service flow... At any moment Located in the low Earth orbit satellite region Current low-Earth orbit satellite nodes within And set the exit selection strategy issued by the upper level as follows: The corresponding export constraints within the region can be represented as a set of exports: ,in, This represents a mapping function that maps exit actions to a set of exit nodes and a set of exit directions. Let the current node be... The set of adjacent nodes is: ,in, This represents the set of inter-satellite links in the lower Earth orbit. For any candidate next hop... The local link state information is defined as follows: ,in, For link latency, The length of the queue. For available bandwidth, As an indicator of link availability, This indicates packet loss or bit error in the link. Based on the local link status information, the current node... Calculate the local forwarding cost function for candidate next hops In one implementation, the cost function can be expressed as: ,in And at least one of them is 1, used to characterize the selected subset of indicators; These are the corresponding weighting coefficients. Then, the next-hop node is selected while satisfying the exit constraints: ,in, Represents the set of exports The derived set of feasible next hops is used to ensure that the business flow converges hop-by-hop toward the set of exit nodes within the region. Preferably, to ensure that the business flow gradually converges to the set of exit nodes, a regional distance metric to the set of exit nodes can be defined. And add monotonically convergent constraints: This means that only the next-hop node that strictly reduces the distance to the exit set can be selected, thereby avoiding loops or invalid oscillations in the business flow within the region.

[0060] Step S4: The multiple medium- and high-orbit satellites update the exit selection strategy for each low-orbit satellite region based on the updated regional-level network status.

[0061] Specifically, once the system reaches the preset stable conditions, the distributed routing method of the present invention enters the online execution phase. Under the constraints of real-time satellite network environment information and real-time inter-satellite link topology, the low-orbit satellites cyclically execute hop-by-hop autonomous forwarding within the region according to the current regional exit selection strategy, so that the service flow continues to advance along the exit direction. This achieves rapid adaptation to changes in link status within the region while maintaining the stability of the cross-regional routing structure, thus completing the continuous optimization and stable operation of the cross-regional forwarding process.

[0062] Specifically, the medium-to-high orbit (MEO) satellites update the egress selection strategy only when preset update conditions are met. These update conditions include one or more combinations of the following: regional egress congestion in the low-Earth orbit (LEO) satellite region exceeds a preset threshold, inter-regional reachability changes, and regional service load fluctuations exceed a preset value. When the preset update conditions are met, the MEO satellites re-collect regional-level network status and generate a new egress selection strategy, which is then distributed to the corresponding LEO region. If the preset update conditions are not met, the current egress selection strategy remains unchanged.

[0063] In one embodiment, the exit selection strategy is updated according to a first time scale, and the hop-by-hop autonomous forwarding of the low-orbit satellite is performed according to a second time scale, wherein the second time scale is smaller than the first time scale, thereby keeping the cross-regional routing decision relatively stable, while retaining the ability to respond quickly to changes in link status within the region, and achieving synergy between cross-regional structural stability and adaptive adjustment within the region.

[0064] In one embodiment, the egress selection strategy has a preset freeze window and a preset validity period. Within the freeze window, cross-regional route reconfiguration is prohibited. Low-Earth orbit (LEO) satellites are only allowed to adaptively adjust to link congestion, node failures, or changes in service load within their respective LEO satellite regions through hop-by-hop autonomous forwarding. This restricts local disturbances generated within the region to propagate within the LEO satellite region, preventing the spread of local disturbances across regions and the triggering of cross-regional path reconstruction. Within the preset validity period, the LEO satellites perform hop-by-hop autonomous forwarding within their LEO satellite regions according to the egress selection strategy, without triggering immediate updates to the egress selection strategy.

[0065] The scope of protection of the multi-layer satellite network distributed routing method described in this embodiment is not limited to the execution order of the steps listed in this embodiment. Any solution implemented by adding, deleting, or replacing steps in the prior art based on the principles of this invention is included within the scope of protection of this invention.

[0066] This invention also provides a multi-layer satellite network distributed routing system, which can implement the multi-layer satellite network distributed routing method described in this invention. However, the implementation device of the multi-layer satellite network distributed routing system described in this invention includes, but is not limited to, the structure of the multi-layer satellite network distributed routing system listed in this embodiment. All structural modifications and substitutions of the prior art made in accordance with the principles of this invention are included within the protection scope of this invention.

[0067] like Figure 4 As shown, in one embodiment, the multi-layer satellite network distributed routing system of the present invention includes multiple medium-high orbit satellites 41 and multiple low-Earth orbit satellites 42. The multiple low-Earth orbit satellites 42 are divided into multiple low-Earth orbit satellite regions, and each low-Earth orbit satellite region corresponds to at least one medium-high orbit satellite 41 as a region control node.

[0068] The multiple medium- and high-orbit satellites 41 are used to collaboratively generate exit selection strategies for each low-orbit satellite region based on the regional network status of each low-orbit satellite region, and to distribute the exit selection strategies to the corresponding low-orbit satellite regions. The exit selection strategies are used to provide cross-regional forwarding direction constraints for service flows entering the low-orbit satellite region, indicating the set of exit nodes and the set of exit directions for the service flows in the current low-orbit satellite region.

[0069] The low-orbit satellite 42 is used to perform hop-by-hop autonomous forwarding within its respective low-orbit satellite region, with the corresponding exit selection strategy as the inter-regional forwarding constraint.

[0070] The medium-high orbit satellite 41 is also used to update the exit selection strategy for each low orbit satellite region based on the updated regional network status.

[0071] In the embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, or methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or modules or units may be electrical, mechanical, or other forms.

[0072] The modules / units described as separate components may or may not be physically separate. The components shown as modules / units may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules / units can be selected to achieve the objectives of the embodiments of the present invention, depending on actual needs. For example, the functional modules / units in the various embodiments of the present invention may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0073] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0074] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A distributed routing method for multi-layer satellite networks, characterized in that, The method includes the following steps: The low-Earth orbit satellite set is divided into multiple low-Earth orbit satellite regions, and each low-Earth orbit satellite region corresponds to at least one medium-high orbit satellite as a regional control node. Multiple medium- and high-orbit satellites collaboratively generate exit selection strategies for each low-orbit satellite region based on the regional network status of each low-orbit satellite region, and distribute the exit selection strategies to the corresponding low-orbit satellite regions. The exit selection strategies are used to provide cross-regional forwarding direction constraints for service flows entering the low-orbit satellite region, indicating the set of exit nodes and the set of exit directions for the service flow in the current low-orbit satellite region. Low-Earth orbit satellites within the low-Earth orbit satellite region use the corresponding exit selection strategy as the inter-regional forwarding constraint and perform hop-by-hop autonomous forwarding within their respective low-Earth orbit satellite regions. The multiple medium- and high-orbit satellites update the exit selection strategy for each low-orbit satellite region based on the updated regional-level network status.

2. The distributed routing method for multi-layer satellite networks according to claim 1, characterized in that, Multiple medium- and high-orbit satellites collaboratively generate exit selection strategies for each low-orbit satellite region based on the regional-level network status of each low-orbit satellite region, including the following steps: The medium and high orbit satellites periodically collect the regional network status of each low orbit satellite region; The medium- and high-orbit satellites construct regional-level status observations based on the regional-level network status. Based on the aforementioned regional state observations, multiple medium- and high-orbit satellites employ a distributed multi-agent reinforcement learning strategy for collaborative training to acquire joint regional actions. A reward function is generated based on the joint regional actions and the updated regional network state. The medium- and high-orbit satellites exchange local policy parameters or local state summaries of the agents through inter-satellite control links, and perform collaborative parameter updates based on the reward function, so as to converge the exit selection strategy of each low-orbit satellite region under the framework of centralized training, distributed execution and parameter collaborative update.

3. The distributed routing method for multi-layer satellite networks according to claim 2, characterized in that, Each medium-to-high orbit satellite acquires the regional network status based on a partially observable Markov decision process. The regional network status includes one or more combinations of the following: average link congestion level within the region, regional service load level, regional egress reachability, and connectivity status of adjacent regions.

4. The distributed routing method for multi-layer satellite networks according to claim 2, characterized in that, The regional-level state observation ,in This indicates the regional network status of the low-Earth orbit satellite area. This represents a partially observable mapping function.

5. The distributed routing method for multi-layer satellite networks according to claim 1, characterized in that, Low-Earth orbit (LEO) satellites within the LEO satellite region use the corresponding exit selection strategy as the inter-regional forwarding constraint, and perform hop-by-hop autonomous forwarding within their respective LEO satellite regions, including the following steps: The low-orbit satellite will use the exit node and exit direction of the corresponding exit selection strategy as the routing target within the region. The low-Earth orbit satellite selects the next-hop satellite within the low-Earth orbit satellite region based on local link status information using a distributed adaptive routing algorithm, so that the service flow advances hop by hop along the exit direction and converges to the exit node.

6. The distributed routing method for multi-layer satellite networks according to claim 5, characterized in that, The local link status information includes one or more combinations of link delay with adjacent low-Earth orbit satellites, queue length, available bandwidth, link availability, and link error / packet loss indication.

7. The distributed routing method for multi-layer satellite networks according to claim 1, characterized in that, The medium- and high-orbit satellites update the exit selection strategy only when preset update conditions are met. The update conditions include one or more combinations of the following: regional exit congestion in the low-orbit satellite region exceeds a preset threshold, inter-regional reachability changes, and regional service load fluctuations exceed a preset value.

8. The distributed routing method for multi-layer satellite networks according to claim 1, characterized in that, The exit selection strategy is updated according to a first time scale, and the hop-by-hop autonomous forwarding of the low-orbit satellite is performed according to a second time scale, wherein the second time scale is smaller than the first time scale.

9. The distributed routing method for multi-layer satellite networks according to claim 1, characterized in that, The export selection strategy has a preset freeze window and a preset validity period; Within the freeze window, cross-regional route reconfiguration is prohibited. Low-Earth orbit (LEO) satellites are only allowed to adaptively adjust to link congestion, node failures, or changes in service load within their respective LEO satellite regions through hop-by-hop autonomous forwarding. This aims to limit the propagation of local disturbances within the region to their respective LEO satellite regions, preventing the spread of local disturbances across regions and the triggering of cross-regional path reconstruction. Within the preset validity period, the LEO satellites perform hop-by-hop autonomous forwarding within their LEO satellite regions according to the egress selection policy, without triggering immediate updates to the egress selection policy.

10. A multi-layer satellite network distributed routing system, characterized in that, It includes multiple medium-high orbit satellites and multiple low orbit satellites; the multiple low orbit satellites are divided into multiple low orbit satellite regions, and each low orbit satellite region corresponds to at least one medium-high orbit satellite as a region control node. The multiple medium and high orbit satellites are used to collaboratively generate exit selection strategies for each low orbit satellite region based on the regional network status of each low orbit satellite region, and to distribute the exit selection strategies to the corresponding low orbit satellite regions. The exit selection strategy is used to provide cross-regional forwarding direction constraints for service flows entering the low-Earth orbit satellite region, indicating the set of exit nodes and the set of exit directions for the service flow within the current low-Earth orbit satellite region; The low-orbit satellite is used to perform hop-by-hop autonomous forwarding within its respective low-orbit satellite region, with the corresponding exit selection strategy as the inter-regional forwarding constraint. The medium and high orbit satellites are also used to update the exit selection strategy for each low orbit satellite region based on the updated regional network status.

Images