Massive constellation hierarchical routing planning method for concurrent tasks
By dividing the constellation into management domains and performing virtual address encoding and comprehensive cost calculation, the complexity of routing planning and load balancing issues in the Giant Star network are resolved, achieving efficient global concurrent transmission.
Patent Information
- Application Number
- CN202611246602.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2026-06-03
- Filing Date
- 2026-08-17
- Publication Date
- 2026-11-06
AI Technical Summary
Traditional routing planning architectures struggle to simultaneously meet the requirements of efficient response and load balancing in the Giants constellation network, especially in scenarios with large-scale nodes, high traffic, and high dynamism, where computational efficiency and real-time performance decline, and the on-board interaction capabilities of fully distributed planning are limited.
By employing domain-level shortest route planning, intra-domain boundary load balancing, and intra-domain multi-task collaborative route planning methods, the constellation is divided into regularly distributed management domains. Through virtual address encoding and comprehensive cost calculation, the route planning process is optimized, reducing complexity and achieving load balancing.
It effectively reduces the complexity of constellation routing planning under large-scale concurrent tasks, achieves network load balancing performance optimization, and provides an efficient global concurrent transmission solution for large-scale constellations.
Smart Images
Figure CN123055646A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a hierarchical routing planning method for large-scale constellations oriented towards concurrent tasks, belonging to the field of giant constellation management and routing planning technology. Background Technology
[0002] The Giants constellation network has become an important supplement to global communications services. To meet the needs of global users, the Giants constellation needs to have efficient data transmission capabilities under high concurrency. However, considering the large number of constellation nodes, the high dynamism of satellites, and the large scale of network traffic, the Giants constellation needs to solve problems such as low planning efficiency, local congestion caused by uneven network load, and excessive signaling overhead in order to achieve efficient response to user needs.
[0003] For concurrent routing planning, the computational mechanism and interaction mechanism jointly determine the routing planning architecture. In the massive concurrent task scenario of Giant Pixnet, the characteristics of large-scale nodes and multi-source long-distance transmission are significant, which amplifies the impact of the routing planning architecture on network response speed and performance. Traditional routing planning architectures are difficult to simultaneously meet the comprehensive requirements of network response efficiency and load balancing. The characteristics of Giant Pixnet—a large number of nodes, high traffic, and high dynamism—become key bottlenecks restricting routing efficiency.
[0004] In current constellation routing planning, the computational efficiency and real-time performance of classic centralized optimization will drop sharply as the constellation size increases. The fully distributed planning currently being widely studied relies on frequent information exchanges between a large number of nodes, which is difficult to achieve theoretical optimality under the constraints of on-board interaction capabilities and computing power. Summary of the Invention
[0005] To address the complexity of constellation routing planning under large-scale concurrent tasks and large-scale nodes, this invention provides a large-scale constellation hierarchical routing planning method for concurrent tasks.
[0006] This invention provides a large-scale constellation hierarchical routing planning method for concurrent tasks, comprising:
[0007] Domain-level shortest route planning: Divide the large-scale constellation into constellation management domains with regular distribution, consistent structure and balanced link capacity in all directions. Determine the corresponding starting constellation management domain and ending constellation management domain based on the starting point and ending point of the task, and determine the shortest domain-level path from the starting constellation management domain to the ending constellation management domain for each task.
[0008] Intra-domain boundary load balancing: Using the direction of the line connecting the start and end points of the task as the global routing reference direction for the current task, for each constellation management domain traversed by the shortest domain-level path, the satellite with the smallest angle between its position and the global routing reference direction is selected from the set of satellites at the domain boundary as the initial inbound endpoint and initial outbound endpoint of the constellation management domain; initial outbound endpoints that have been selected more than a threshold number of times within the current task cycle are designated as congested nodes. With the goal of minimizing the expected total delay in information transmission, the selection of outbound endpoints for tasks that exceed the threshold number of times at congested nodes is adjusted to determine the final outbound endpoint; at the same time, the final inbound endpoint of the next adjacent constellation management domain is adjusted accordingly.
[0009] Intra-domain multi-task collaborative routing planning: Based on the determination of the final inbound endpoint, the tasks in the constellation management domain are divided into rounds according to the estimated time order of the tasks arriving in the constellation management domain. Considering multi-task path conflicts, a low-load next-hop satellite node is selected for each round of tasks until the final outbound endpoint is reached.
[0010] According to the large-scale constellation hierarchical routing planning method for concurrent tasks of the present invention, in the domain-level shortest route planning, the divided constellation management domains are virtual address encoded, and the corresponding starting constellation management domain and ending constellation management domain are determined based on the virtual address encoding.
[0011] According to the large-scale constellation hierarchical routing planning method for concurrent tasks of the present invention, in the process of intra-domain boundary load balancing, the method for adjusting the final inbound endpoint of the next adjacent constellation management domain is as follows: detect the link reachability between the initial inbound endpoint of the next adjacent constellation management domain and the final outbound endpoint of the previous adjacent constellation management domain, discard the initial inbound endpoint that cannot establish a direct link with the current final outbound endpoint, and select the new inbound endpoint that has a direct link with the current final outbound endpoint as the final inbound endpoint.
[0012] According to the large-scale constellation hierarchical routing planning method for concurrent tasks of the present invention, in the multi-task collaborative routing planning within the domain, after dividing the tasks within the constellation management domain into rounds, the next-hop satellite node is selected from the candidate adjacent satellites in the order of rounds based on the calculation of comprehensive cost for each round of tasks; the calculation of comprehensive cost includes the calculation of link delay, remaining path cost, link packet loss risk and task progress penalty.
[0013] The beneficial effects of this invention are as follows: Based on the partitioning of large-scale constellation management domains, this invention determines the shortest path at the domain level, providing a global reference for intra-domain routing. Then, it determines the start and end points of tasks within the domain and balances the traffic distribution at the domain boundaries. Furthermore, it divides the planning process into rounds, considering multi-task path conflicts, and guides tasks to select the next-hop node with low load in each round. This invention effectively reduces the complexity of constellation routing planning for large-scale concurrent tasks and nodes, optimizes network load balancing performance, and provides a feasible routing planning scheme for large-scale constellations to efficiently handle high-volume concurrent global transmission.
[0014] The method of this invention constructs a routing planning architecture that fits the characteristics of the Giants constellation network. By reasonably decomposing the planning hierarchy and interaction dimensions, it reduces the planning complexity while achieving a balance between global fast response and performance optimization. Attached Figure Description
[0015] Figure 1 This is a flowchart of the large-scale constellation hierarchical routing planning method for concurrent tasks described in this invention;
[0016] Figure 2 This is a schematic diagram illustrating the division of constellation management domains for large-scale constellations;
[0017] Figure 3 This is a diagram illustrating the shortest domain-level path in the domain-level shortest route planning;
[0018] Figure 4 This is a schematic diagram of the global routing reference direction for the task;
[0019] Figure 5 This is a schematic diagram of load redistribution at the domain boundary;
[0020] Figure 6 This is a diagram illustrating cross-domain endpoint matching for tasks. Detailed Implementation
[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0022] Specific Implementation Method 1: Combination Figures 1 to 6 As shown, this invention provides a large-scale constellation hierarchical routing planning method for concurrent tasks, including:
[0023] Domain-level shortest route planning: Divide the large-scale constellation into constellation management domains with regular distribution, consistent structure and balanced link capacity in all directions. Determine the corresponding starting constellation management domain and ending constellation management domain based on the starting point and ending point of the task, and determine the shortest domain-level path from the starting constellation management domain to the ending constellation management domain for each task.
[0024] Intra-domain boundary load balancing: Using the direction of the line connecting the start and end points of a task as the global routing reference direction for the current task, for each constellation management domain traversed by the shortest domain-level path, the satellite with the smallest angle between its position and the global routing reference direction is selected from the set of satellites at the domain boundary as the initial inbound and initial outbound endpoints of the constellation management domain. Initial outbound endpoints that have been selected more than a threshold number of times within the current task cycle are designated as congested nodes. With the goal of minimizing the expected total delay in information transmission, the selection of outbound endpoints for tasks exceeding the threshold number at congested nodes is adjusted to determine the final outbound endpoint. Simultaneously, the final inbound endpoint of the next adjacent constellation management domain is adjusted accordingly. Intra-domain boundary load balancing achieves balanced traffic distribution at the domain boundary.
[0025] Intra-domain multi-task collaborative routing planning: Based on the determination of the final inbound endpoint, the tasks in the constellation management domain are divided into rounds according to the estimated time order of the tasks arriving in the constellation management domain. Considering multi-task path conflicts, a low-load next-hop satellite node is selected for each round of tasks until the final outbound endpoint is reached.
[0026] Furthermore, combined with Figure 3 As shown, in domain-level shortest route planning, the divided constellation management domains are virtual address encoded, and the corresponding starting constellation management domain and ending constellation management domain are determined based on the virtual address encoding.
[0027] This implementation uses the constellation management domain as the basic unit for on-board routing planning; virtual address encoding can provide a quick reference for deducing adjacency relationships for domain-level routing planning; then, the shortest domain-level path is calculated based on the distribution of management domains; thus, the shortest domain-level path is provided for global concurrent tasks, and a global reference direction is provided for intra-domain routing planning.
[0028] Combination Figure 6 As shown, during the intra-domain boundary load balancing process, the method for adjusting the final inbound endpoint of the next adjacent constellation management domain is as follows: detect the link reachability between the initial inbound endpoint of the next adjacent constellation management domain and the final outbound endpoint of the previous adjacent constellation management domain, discard the initial inbound endpoint that cannot establish a direct link with the current final outbound endpoint, and select the new inbound endpoint that has a direct link with the current final outbound endpoint as the final inbound endpoint.
[0029] Intra-domain boundary load balancing consists of three methods: initial selection of task entry and exit endpoints, redistribution of satellite load at the domain boundary, and task cross-domain endpoint matching. Initial selection of task entry and exit endpoints involves choosing the satellite with the smallest angle to the global routing reference direction from the set of satellites at the domain boundary as the initial entry and exit endpoints for tasks passing through the current management domain. Redistribution of satellite load at the domain boundary adjusts the selection of the task's exit endpoint based on the number of times a domain boundary node is assigned as an exit endpoint, achieving a balanced distribution of traffic at the domain boundary. Cross-domain endpoint matching involves detecting the reachability of the link between the selected entry endpoint of the next management domain and the exit endpoint of the current domain for cross-domain tasks, and adjusting the entry endpoint of the next domain to establish a direct link with the exit endpoint of the current domain if no direct link exists. This method can determine the intra-domain start and end points for multiple tasks in intra-domain routing planning and ensure balanced traffic across the management domain boundary.
[0030] Furthermore, in the multi-task collaborative routing planning within the domain, after dividing the tasks within the constellation management domain into rounds, the next-hop satellite node is selected from the candidate adjacent satellites according to the round order based on the calculation of the comprehensive cost. The calculation of the comprehensive cost includes the calculation of link delay, remaining path cost, link packet loss risk, and task progress penalty.
[0031] The multi-task collaborative routing planning method is executed in parallel within each management domain and consists of a task round-by-round activation method and a next-hop decision method based on comprehensive cost. This method can achieve low-complexity concurrent task routing planning, optimize network node load distribution, and thus improve task transmission latency and reduce packet loss rate.
[0032] As an example, combined Figures 2 to 4 As shown, the constellation management domain is a quadrilateral management domain.
[0033] Divide large constellations into Each constellation management domain: In the formula Let be the number of fields along the orbital direction. The number of domains along the transorbital direction; each constellation management domain is composed of It consists of several satellites. The number of satellites along the orbital direction within each constellation management domain. The number of satellites along the transorbital direction is considered; the constellation management domain division takes into account uniform coverage of the entire constellation and balanced link capacity in all directions between domains, while balancing the planning complexity at the domain level and within the domain, in order to minimize the computational complexity of the overall routing planning.
[0034] Combination Figure 3 As shown, a two-bit virtual address can be assigned to each constellation management domain. Where m is the index across orbital directions and n is the index within orbital directions, the communication relationships between management domains can be quickly derived using virtual addresses; within each management domain, a local virtual address is configured for each satellite. ,in Indicates the index across the orbital direction. Indicates the index in the same orbital direction.
[0035] Combination Figure 3 As shown, the method for determining the shortest domain-level path from the starting constellation management domain to the ending constellation management domain for each task is as follows:
[0036] Each constellation management domain is abstracted as a domain node. The corresponding starting domain node and ending domain node are determined based on the task's starting point and ending point. Based on adjacency relationships and the shortest path algorithm, the shortest domain-level path from the starting domain node to the ending domain node is calculated.
[0037] Furthermore, combining Figure 4 and Figure 5 As shown, the method for determining the initial inbound endpoint and the initial outbound endpoint is as follows:
[0038] Define global routing reference direction vector for:
[0039] ,
[0040] In the formula As the starting satellite within the starting constellation management domain, For the endpoint satellite within the endpoint constellation management domain;
[0041] Represent the constellation management domains traversed by the shortest domain-level path as follows: If the starting satellite Located in the constellation management domain Then constellation management domain Initial inbound endpoint ;
[0042] If constellation management domain Non-originating satellite The constellation it is in is the next constellation management domain. Set up constellation management domain With constellation management domain The satellite set for inter-domain inter-satellite links is For satellite collection satellites in Construct direction vector Select satellite set make the direction vector closest reference direction vector satellite As the initial entry endpoint :
[0043] ,
[0044] If the destination satellite Located in the constellation management domain Then constellation management domain initial outbound endpoint ;
[0045] If constellation management domain Non-terminal satellite The current constellation's governing domain, and the next adjacent constellation's governing domain is... Set up constellation management domain With constellation management domain The satellite set for inter-domain inter-satellite links is For satellite collection satellites in Construct direction vector Select satellite set make the direction vector closest reference direction vector satellite As the initial outbound endpoint :
[0046] .
[0047] Combination Figure 5 As shown, the method for determining the final outbound endpoint and the final inbound endpoint is as follows:
[0048] Count the number of times each initial outbound endpoint is selected by a task. For tasks at congested nodes that exceed the selection threshold, adjust the selection of outbound endpoints. Calculate the evaluation index with the goal of minimizing the expected total delay in information transmission. :
[0049] ,
[0050] In the formula The estimated time for the task to arrive. For the amount of task data, For satellite link bandwidth, For satellite The number of tasks; the evaluation index for selecting tasks that exceed the threshold at congested nodes. The satellite corresponding to the minimum As the final outbound endpoint;
[0051] Perform task alignment at nodes across domain boundaries, and check the constellation management domain through inter-domain interactions. Initial inbound endpoint and constellation management domain If the final outbound endpoint has a direct link, the initial inbound endpoint is retained as the final inbound endpoint; otherwise, a new inbound endpoint with a direct link to the current final outbound endpoint is selected as the final inbound endpoint.
[0052] In this embodiment, the method for dividing tasks within the constellation management domain into rounds in the intra-domain multi-task collaborative routing planning is as follows:
[0053] Initialize constellation management domain The planning rounds will determine the task cycle. Discretized into multiple time slices , For the first A time slice; constellation management domain The discrete representation of the task within is Rounds ;
[0054] If the task reaches the constellation management domain Estimated time The corresponding time slice is , Then the current task will be planned as a round. .
[0055] The method for selecting a low-load next-hop satellite node for each mission round is as follows:
[0056] Construct a subset of tasks for the current round ; by task subset Current node satellite for each mission Corresponding next-hop candidate node satellite Construct a set of candidate decision nodes ;
[0057] Constructing a comprehensive cost function For the set of candidate decision nodes Candidate node satellites Performance evaluation:
[0058] ,
[0059] In the formula For link latency, For the cost of the remaining path, To mitigate the risk of packet loss in the link, Penalty for task progress;
[0060] In the set of candidate decision nodes Choose to make the comprehensive cost function Minimum candidate node satellite As the next-hop satellite node Then update the intra-domain paths and satellite payloads; check Have all tasks in this round of decision-making been completed? If not, return to the above process to continue processing the remaining tasks; if all tasks are completed, some tasks still need to be planned in subsequent rounds. Then return to the next round of planning until the final outbound endpoint is reached.
[0061] The method of this invention divides all satellites in the constellation into multiple management domains, decomposing the large-scale concurrent routing planning problem into two levels: domain-level fast search and intra-domain load balancing optimization. This effectively reduces planning complexity and optimizes network load balancing performance, providing an effective solution for rapid on-board response in subsequent large-scale missions.
[0062] While the invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and features described herein can be combined in ways different from those described in the original claims. It is also understood that features described in conjunction with individual embodiments can be used in other described embodiments.
Claims
1. A large-scale constellation hierarchical routing planning method for concurrent tasks, characterized in that, include, Domain-level shortest route planning: Divide the large-scale constellation into constellation management domains with regular distribution, consistent structure and balanced link capacity in all directions. Determine the corresponding starting constellation management domain and ending constellation management domain based on the starting point and ending point of the task, and determine the shortest domain-level path from the starting constellation management domain to the ending constellation management domain for each task. Intra-domain boundary load balancing: Using the direction of the line connecting the start and end points of the task as the global routing reference direction for the current task, for each constellation management domain traversed by the shortest domain-level path, the satellite with the smallest angle between its position and the global routing reference direction is selected from the set of satellites at the domain boundary as the initial inbound endpoint and initial outbound endpoint of the constellation management domain; initial outbound endpoints that have been selected more than a threshold number of times within the current task cycle are designated as congested nodes. With the goal of minimizing the expected total delay in information transmission, the selection of outbound endpoints for tasks that exceed the threshold number of times at congested nodes is adjusted to determine the final outbound endpoint; at the same time, the final inbound endpoint of the next adjacent constellation management domain is adjusted accordingly. Intra-domain multi-task collaborative routing planning: Based on the determination of the final inbound endpoint, the tasks in the constellation management domain are divided into rounds according to the estimated time order of the tasks arriving in the constellation management domain. Considering multi-task path conflicts, a low-load next-hop satellite node is selected for each round of tasks until the final outbound endpoint is reached.
2. The large-scale constellation hierarchical routing planning method for concurrent tasks according to claim 1, characterized in that, In domain-level shortest route planning, the divided constellation management domains are virtual address encoded, and the corresponding starting constellation management domain and ending constellation management domain are determined based on the virtual address encoding.
3. The large-scale constellation hierarchical routing planning method for concurrent tasks according to claim 2, characterized in that, During the intra-domain boundary load balancing process, the method for adjusting the final inbound endpoint of the next adjacent constellation management domain is as follows: detect the link reachability between the initial inbound endpoint of the next adjacent constellation management domain and the final outbound endpoint of the previous adjacent constellation management domain, discard the initial inbound endpoint that cannot establish a direct link with the current final outbound endpoint, and select the new inbound endpoint that has a direct link with the current final outbound endpoint as the final inbound endpoint.
4. The large-scale constellation hierarchical routing planning method for concurrent tasks according to claim 3, characterized in that, In intra-domain multi-task collaborative routing planning, after dividing the tasks within the constellation management domain into rounds, the next-hop satellite node is selected from the candidate adjacent satellites according to the round order based on the calculation of comprehensive cost. The calculation of comprehensive cost includes link delay, remaining path cost, link packet loss risk, and task progress penalty.
5. The large-scale constellation hierarchical routing planning method for concurrent tasks according to claim 4, characterized in that, The constellation management domain is a quadrilateral management domain.
6. The large-scale constellation hierarchical routing planning method for concurrent tasks according to claim 5, characterized in that, The method for determining the shortest domain-level path for each task from the originating constellation management domain to the ending constellation management domain is as follows: Each constellation management domain is abstracted as a domain node. The corresponding starting domain node and ending domain node are determined based on the task's starting point and ending point. Based on adjacency relationships and the shortest path algorithm, the shortest domain-level path from the starting domain node to the ending domain node is calculated.
7. The large-scale constellation hierarchical routing planning method for concurrent tasks according to claim 6, characterized in that, The method for determining the initial inbound endpoint and the initial outbound endpoint is as follows: Define global routing reference direction vector for: ， In the formula As the starting satellite within the starting constellation management domain, For the endpoint satellite within the endpoint constellation management domain; Represent the constellation management domains traversed by the shortest domain-level path as follows: If the starting satellite Located in the constellation management domain Then constellation management domain Initial inbound endpoint ; If constellation management domain Non-originating satellite The constellation it is in is the next constellation management domain. Set up constellation management domain With constellation management domain The satellite set for inter-domain inter-satellite links is For satellite collection satellites in Construct direction vector Select satellite set make the direction vector closest reference direction vector satellite As the initial entry endpoint : ， If the destination satellite Located in the constellation management domain Then constellation management domain initial outbound endpoint ; If constellation management domain Non-terminal satellite The current constellation's governing domain, and the next adjacent constellation's governing domain is... Set up constellation management domain With constellation management domain The satellite set for inter-domain inter-satellite links is For satellite collection satellites in Construct direction vector Select satellite set make the direction vector closest reference direction vector satellite As the initial outbound endpoint : 。 8. The large-scale constellation hierarchical routing planning method for concurrent tasks according to claim 7, characterized in that, The method for determining the final outbound endpoint and the final inbound endpoint is as follows: Count the number of times each initial outbound endpoint is selected by a task. For tasks at congested nodes that exceed the selection threshold, adjust the selection of outbound endpoints. Calculate the evaluation index with the goal of minimizing the expected total delay in information transmission. : ， In the formula The estimated time for the task to arrive. For the amount of task data, For satellite link bandwidth, For satellite The number of tasks; For tasks exceeding the frequency threshold at congested nodes, select an evaluation index. The satellite corresponding to the minimum As the final outbound endpoint; Check constellation management domain Initial inbound endpoint and constellation management domain If the final outbound endpoint has a direct link, the initial inbound endpoint is retained as the final inbound endpoint; otherwise, a new inbound endpoint with a direct link to the current final outbound endpoint is selected as the final inbound endpoint.
9. The large-scale constellation hierarchical routing planning method for concurrent tasks according to claim 8, characterized in that, In intra-domain multi-task collaborative routing planning, the method for dividing tasks within the constellation management domain into rounds is as follows: Initialize constellation management domain The planning rounds will determine the task cycle. Discretized into multiple time slices , For the first A time slice; constellation management domain The discrete representation of the task within is Rounds ; If the task reaches the constellation management domain Estimated time The corresponding time slice is , Then the current task will be planned as a round. .
10. The large-scale constellation hierarchical routing planning method for concurrent tasks according to claim 9, characterized in that, The method for selecting a low-load next-hop satellite node for each mission round is as follows: Construct a subset of tasks for the current round ; by task subset Current node satellite for each mission Corresponding next-hop candidate node satellite Construct a set of candidate decision nodes ; Constructing a comprehensive cost function For the set of candidate decision nodes Candidate node satellites Performance evaluation: ， In the formula For link latency, For the cost of the remaining path, To mitigate the risk of packet loss in the link, Penalty for task progress; In the set of candidate decision nodes Choose to make the comprehensive cost function Minimum candidate node satellite As the next-hop satellite node Then update the intradomain path and satellite payload; and repeat the above process until the final outbound endpoint is reached.