Communication processing system and method for enhancing computing power of low-orbit communication satellite based on constellation cooperative computing

Abstract

The invention discloses a task scheduling and energy consumption optimization system and method based on distributed satellite calculation, and the method comprises the following steps: obtaining the floating point calculation amount and data size of each subtask, obtaining the calculation power and link rate of each satellite, and calculating the estimated execution time of the subtasks on the satellites according to an execution time formula; further obtaining the calculation power, the communication power and the corresponding duration of each satellite, and estimating the total energy consumption through an energy consumption formula; and optimizing a task allocation scheme based on the energy consumption evaluation, thereby reducing the overall energy consumption of the satellite cluster.

Description

Technical Field

[0001] This invention relates to the fields of satellite communication and edge computing, and in particular to a communication and processing method that achieves computing power expansion, low-latency processing and bandwidth optimization through inter-satellite collaborative scheduling and joint operations within a low Earth orbit (LEO) satellite constellation. Background Technology

[0002] With the widespread deployment of large-scale constellations of low-Earth orbit (LEO) satellites in fields such as remote sensing, communications, and the Internet of Things (IoT), the amount of data generated by sensors (such as synthetic aperture radar (SAR) and multispectral cameras) carried by a single satellite is growing exponentially. Taking SAR as an example, the raw echo data generated from a single imaging session can reach several gigabytes to tens of gigabytes. If all the raw data were transmitted back to the ground for processing, it would not only consume a large amount of downlink bandwidth and increase the burden on ground processing centers, but also bring significant latency and data leakage risks. Therefore, most existing systems adopt a "transparent forwarding / relay" approach: the satellite is responsible for data acquisition and forwarding, minimizing complex floating-point calculations on-board or performing only simple preprocessing. The main reason for this is that LEO communication satellites face strict constraints in terms of energy (solar cells and battery capacity), thermal management (passive / limited active cooling), and computing power (it is difficult to configure high-power, high-performance processors or accelerators on a single satellite). The limited bandwidth of inter-satellite links (ISL) and downlinks, which are affected by weather and the visibility time of ground stations, leads to bottlenecks in data transmission. When conducting computational collaboration within a constellation, data confidentiality, integrity, and access control must be considered to prevent the leakage or tampering of sensitive information. Determining how to partition complex tasks and distribute them across multiple satellites under constraints such as energy, heat, computing power, and bandwidth to minimize overall latency or energy consumption is a combinatorial optimization problem that must also meet real-time requirements.

[0003] Therefore, a collaborative computing method is needed that is secure, controllable, low-latency, and energy / thermal-constrained within a satellite constellation. This method would enable several resource-constrained satellites to "synthesize" processing power equivalent to strong ground-based computing capabilities, thereby reducing ground backhaul, improving real-time performance, and ensuring data security. Summary of the Invention This invention addresses the limitation of existing low-Earth orbit (LEO) communication satellites in efficiently processing large-scale data under computational constraints. It proposes a communication and processing system for LEO communication satellites based on constellation collaborative computing, characterized by comprising: The data acquisition module is used to acquire raw data generated by low-orbit satellite payloads and to cache and preprocess it. The task generation module is used to generate a task request package based on the original data and task requirements. The request package includes task type, computation requirements, latency threshold and priority information. The task decomposition module is used to divide the task request package into multiple subtasks with dependencies, the dependencies including data flow dependencies and computation order constraints; The distributed computing power scheduling module is used to collect computing power, energy and link status information of each satellite based on the high-speed inter-satellite link, and calculate candidate scheduling schemes according to the sub-task requirements to support distributed execution across satellites; The task allocation and optimization module is used to establish a multi-objective optimization model based on the scheduling results, with total delay and total energy consumption as optimization objectives. The inter-satellite communication module is used to transmit task input data and intermediate results through inter-satellite links under the control of the scheme generated by the allocation and optimization module. The mission execution module is used to execute sub-tasks locally on the target satellite using a general-purpose processor or hardware accelerator. The result fusion and encryption module is used to fuse, correct errors, and encrypt the execution results of multiple subtasks to obtain the final processing result and transmit it to the ground.

[0004] Preferably, the task allocation and optimization module uses the following objective function to solve the task allocation scheme: Where T is the total delay, E is the total system energy consumption, and α and β are weighting factors that are dynamically adjusted according to the task type; The solution process is based on a set of feasibility constraints, including a task integrity constraint to ensure that each subtask is assigned to a single satellite, a computing resource constraint to limit the computational requirements of the subtasks to not exceed the available computing power of the target satellite, an energy constraint to ensure that the energy consumption of the subtasks does not exceed the remaining available energy of the satellite, and a link reachability constraint to ensure that the subtask data is transmitted within the link availability window.

[0005] Preferably, the distributed computing power scheduling module calculates the execution time of the subtask on the satellite using the following formula: in, The floating-point computation cost of the subtask. In order to allocate satellite computing power, For data size, For link speed, Preferably, the task allocation and optimization module estimates the energy consumption E using the following energy consumption formula: ; in, For satellite Calculated power, Calculate the duration for it. For its communication power, The duration of their communication.

[0006] Preferably, the task allocation and optimization module performs scheduling optimization using the following objective function: ; Where T is the total delay, E is the total system energy consumption, and α and β are weighting factors that are dynamically adjusted according to the task type; The following constraints are introduced: Mission integrity constraint: Each sub-mission is assigned to only one satellite; Computing resource constraints: The computational workload of the allocated tasks shall not exceed the computing capacity of the satellite; Energy constraint: The energy consumption of the sub-task execution shall not exceed the remaining energy of the satellite; Link constraint: Subtask data transmission must be completed within the available window of the link.

[0007] Another aspect of the present invention is to provide a communication and processing method for enhancing the computing power of low-Earth orbit communication satellites based on constellation collaborative computing, characterized by comprising the following steps: S1. Raw data is collected by low-orbit satellites and a mission request package is generated. The request package includes mission type, computational requirements, latency threshold and priority information. S2. Decompose the task request package into multiple subtasks with dependencies. S3. Collect computing power, energy and link status information of each satellite in the constellation through inter-satellite links, and generate candidate scheduling schemes based on this information to realize distributed computing power scheduling; S4. Based on the candidate scheduling schemes, construct a multi-objective optimization model with total delay and total energy consumption as optimization objectives, and introduce task integrity constraints, computing power constraints, energy constraints and link constraints to solve for the optimal task allocation scheme. S5. According to the optimal task allocation scheme, the inter-satellite communication module is used to complete the cross-satellite transmission of sub-task data and intermediate results; the corresponding satellite executes the allocated sub-task and transmits the execution results to the result fusion and encryption module; S6. The results of the sub-tasks are fused, error corrected, and encrypted to generate complete task results and then transmitted to the ground.

[0008] Preferably, for each subtask, its floating-point computational cost is obtained. and data size Obtain the available computing power of each satellite. and link speed Calculate the estimated execution time of the sub-task on the satellite using the following formula. : ; The subtasks are scheduled and allocated based on the execution time, enabling the tasks to be executed efficiently on the satellite constellation.

[0009] Preferably, the following steps are also included: Obtain the computational power of each satellite With communication power and its corresponding computation time and communication duration ; Estimate the total energy consumption E of subtask allocation using the following formula: ; Based on the energy consumption assessment, the task allocation scheme is optimized to reduce the overall energy consumption of the satellite constellation.

[0010] Preferably, in step S4, the task allocation and optimization process aims to minimize latency and energy consumption, and adopts the following objective function: ; Where T is the total delay, β is the total system energy consumption, and α and β are weighting factors that are dynamically adjusted according to the task type.

[0011] Another aspect of the present invention is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method.

[0012] This invention establishes a multi-objective optimization model with "delay minimization" and "energy consumption minimization" as optimization objectives. The constraints include: energy supply, thermal design capability, and communication bandwidth utilization of each satellite. Through integer programming or distributed optimization algorithms, the optimal allocation scheme of tasks among multiple satellites is determined. Under dynamic task load and constellation operation conditions, the overall computing efficiency and energy consumption balance of the constellation are guaranteed.

[0013] By leveraging inter-satellite links and distributed computing power scheduling, task processing has shifted from the traditional "satellite-to-ground" model to "satellite-to-satellite" collaboration, significantly reducing the need for raw data to be transmitted back to the ground and achieving task processing response times ranging from seconds to minutes. Furthermore, through collaborative optimization allocation under energy and computing power constraints, task allocation is optimized at the constellation level, avoiding overload or energy waste on some satellites, thereby improving the overall constellation's task processing efficiency and lifespan. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments or related technologies of this specification, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 : A framework diagram of a communication processing system for enhancing the computing power of low-Earth orbit communication satellites using constellation collaborative computing, as described in this invention; Figure 2 Flowchart of a communication processing method for enhancing the computing power of low-Earth orbit communication satellites through constellation collaborative computing. Detailed Implementation

[0016] The technical solutions in the embodiments of this specification 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.

[0017] It should be noted that the terms "comprising" and "having," and any variations thereof, in the embodiments and drawings of this specification are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the steps or units listed, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0018] See Figure 1 The constellation collaborative computing system proposed in this invention mainly consists of the following parts: Data acquisition module 10 The data acquisition module 10 is deployed on each low-Earth orbit communication satellite. Its input is the satellite's local payload data, including synthetic aperture radar (SAR) echoes, optical remote sensing images, communication relay data, or signals uploaded by IoT terminals. The module output is a formatted and preliminarily corrected data block.

[0019] This module standardizes the raw data and packages it according to a preset time slice or data block size for use by the task request generation module. The module employs a cache management mechanism to ensure that data is stored and forwarded in an orderly manner even with inter-satellite task allocation delays.

[0020] Task request generation module 11 The task request generation module 11 takes as input the data block output by the data acquisition module and the task instructions issued by the user or the ground task center. The module outputs a task request package containing data characteristics, computational requirements, and deadlines.

[0021] This module creates a task request descriptor, including parameters such as: data type (SAR / optical / communication), computational requirements (floating-point operations (FLOPs), time constraints (deadline), and priority. This task request descriptor is then passed to the task decomposition module.

[0022] Task Decomposition Module 12 The task decomposition module 12 receives the task request package output by the task request generation module and divides the task into multiple subtasks according to a predefined decomposable model. The input is the task request package; the output is several subtask units, each containing data size, computational complexity, and parallelism metrics.

[0023] The module employs a graph partitioning approach, modeling the overall task as a directed acyclic graph (DAG), where nodes represent subtasks and edges represent dependencies. For example, a SAR imaging task can be decomposed into four stages: pulse compression, range processing, azimuth processing, and imaging reconstruction.

[0024] Distributed computing power scheduling module 13 The distributed computing power scheduling module 13 takes as input the sub-task units generated by the task decomposition module and the computing power, energy, and bandwidth status of each satellite in the constellation; and outputs the task scheduling scheme, i.e., the mapping relationship between sub-tasks and target satellites.

[0025] This module utilizes inter-satellite link transmission capabilities to model task scheduling as a resource scheduling problem: Where T is the overall task completion time, and C... i D represents the computation time of the subtask on satellite i. i This refers to the data transmission latency of subtasks. This scheduling mechanism enables rapid parallel processing of complex tasks across the constellation, significantly reducing processing latency and improving task real-time performance.

[0026] Task Allocation and Optimization Module 14 The task allocation and optimization module 14, based on the output of the distributed computing power scheduling module, further incorporates energy and thermal design constraints to complete the optimization decision. The inputs are the scheduling scheme, satellite energy margin, battery status, and inter-satellite bandwidth parameters; the output is the final optimized task execution path and allocation table.

[0027] In a preferred embodiment, the module may be used to establish a multi-objective optimization model: ; Where T is the total latency, E is the total system energy consumption, and α and β are weighting factors that are dynamically adjusted according to the task type (real-time task / non-real-time task).

[0028] The formula for calculating energy consumption E is: ; in, For satellite Calculated power, Calculate the duration for it. For its communication power, The duration of their communication.

[0029] This optimization process ensures optimal performance in terms of both latency and energy consumption while meeting energy constraints, thereby improving the overall processing efficiency of the constellation.

[0030] Inter-satellite communication module 15 The input to the inter-satellite communication module 15 is the task allocation table and sub-task data generated by the task allocation and optimization module; the output is the sub-task data block forwarded via the inter-satellite link.

[0031] This module uses laser or microwave inter-satellite links to support high-speed, low-error-rate data transmission. It incorporates congestion control and encryption mechanisms to ensure link stability and security even during large-scale parallel task scheduling.

[0032] Task Execution Module 16 The task execution module 16 is located in the local processing unit of each satellite. Its input is the assigned subtask data, and its output is the subtask calculation results.

[0033] This module relies on embedded GPUs, FPGAs, or ASICs on the satellite to perform tasks. For example, pulse compression in the SAR subtask can be accelerated using FFT, with a complexity of O(nlogn), which can reduce processing latency by more than 70% compared to traditional CPU computation.

[0034] Result fusion and encryption module 17 The results fusion and encryption module 17 receives sub-task results from different satellites, performs result stitching, error correction, and consistency checks. The output is complete mission result data, which is encrypted before being transmitted to the ground or other satellites.

[0035] The module supports distributed erasure coding verification to ensure the robustness of data during inter-satellite transmission; and uses lightweight encryption algorithms (such as elliptic curve cryptography ECC) to ensure the security of the results during transmission.

[0036] The implementation process of the method of the present invention is illustrated below with reference to a specific embodiment. Assume that a low-Earth orbit communication satellite constellation consists of four satellites, numbered Sat-A, Sat-B, Sat-C, and Sat-D, all in orbits at an altitude of 500 km, and interconnected via inter-satellite links.

[0037] Step S1: Task Generation Assuming the Sat-A satellite carries a SAR payload and acquires approximately 50GB of raw echo data in a single imaging mission, the mission request generation module generates a mission request packet on the Sat-A satellite, as described below: Mission type: SAR imaging; Data volume: 50GB; Computational requirement: Approximately 1 TFLOPs (number of floating-point operations); Delay constraint: within 60 seconds; Priority: High.

[0038] Step S2: Task Breakdown The task decomposition module divides the SAR imaging task into multiple sub-tasks, such as four sub-tasks: Table 1. Schematic diagram of SAR imaging task subtask decomposition These subtasks are interdependent: the pipeline-style distributed task flow of raw data undergoing preprocessing → reconstruction / processing → analysis → integrated output is PC→RP→AP→IR.

[0039] Step S3: Distributed computing power scheduling Assume the computing resources of the four satellites are as follows: Table 2 Computing power performance and power consumption parameters of each satellite in the constellation Meanwhile, the inter-satellite link rate is 2Gbps, and the average one-way transmission delay is 0.05 s.

[0040] In a preferred embodiment, the scheduling module calculates the completion time of each subtask as follows: = , = ; in, The floating-point computation cost of the subtask. In order to allocate satellite computing power, R represents the data size and R represents the link speed.

[0041] For example, if the pulse compression PC task is assigned to Sat-B: Calculation time: ; Transfer time (12 GB → 96 Gb): ; Total Time 48.3 seconds.

[0042] In a preferred embodiment of the present invention, under the scheduling of the task allocation and optimization module, each subtask is rationally allocated to different satellites for execution through enumeration and dynamic optimization: the pulse compression (PC) subtask is allocated to satellite B (PC→Sat-B), the range processing (RP) subtask is allocated to satellite C (RP→Sat-C), the azimuth processing (AP) subtask is allocated to satellite D (AP→Sat-D), and the imaging reconstruction (IR) subtask is allocated to satellite A (IR→Sat-A). This cross-satellite distributed execution method fully utilizes the heterogeneous computing power and remaining energy of each satellite within the constellation, avoiding the problems of insufficient processing power or excessive energy consumption of a single satellite, thereby significantly improving the overall task processing efficiency and energy efficiency ratio.

[0043] Step S4: Task Allocation and Optimization Based on the scheduling results, the task allocation and optimization module introduces energy consumption constraints.

[0044] After completing the initial scheduling and candidate mapping in step S3, the task allocation and optimization module, based on the scheduling results, introduces energy consumption and thermal constraints to jointly optimize the candidate allocation schemes and determine the final task allocation and execution plan. The mathematical model and principles constructed by this module are as follows.

[0045] Let the set of subtasks be T = {1, ..., n}, and the set of satellites participating in the collaboration be... For any subtask With satellite Define decision variables: The computation time of subtask j on satellite i is denoted as ; in The computational cost (FLOP) for subtask j. Let be the computing power (FLOP / s) available to satellite i within the scheduling window. The transmission time of subtask j from the source node (acquisition satellite or previous processing node) to satellite i is denoted as: ; in The amount of data to be transmitted (bits), For the corresponding inter-satellite or uplink / downlink rates (bits / s), The propagation and scheduling overhead is given by (s). The estimated energy consumption of subtask j on satellite i is: , in Let be the average computational power consumption (W) of satellite i. Let be the transmit power (W) of the corresponding link. The objective function is defined as the weighted sum of the total mission delay T and the total mission energy consumption E, with the following optimization objective: ; in , in For necessary result fusion / synchronization overhead; and The coefficients α and β ≥ 0 are adjustable weights, which can be adjusted by the implementer according to the real-time nature of the task or the energy priority (for example, a larger α is used for real-time tasks, and a larger β is used when energy is limited).

[0046] The model should meet the following constraints to ensure feasibility and platform security: Table 3 Examples of Model Constraints The core function of these constraint formulas is to limit the theoretical optimal allocation (minimum latency, minimum energy consumption) to a practically feasible range: for example, to prevent tasks from being missed or repeated; to prevent satellite computing power or energy from being overloaded; to prevent overheating or link blockage; and to guide the solver to select only feasible allocation schemes during the optimization process.

[0047] To ensure the real-time solvability of this optimization problem in a spaceborne environment, the following solution strategy is preferably adopted in the embodiments: First, a near-optimal global solution is obtained at the constellation control center or cooperative domain controller using an exact or approximate integer linear programming (ILP) / mixed integer programming (MIP) solver; when sudden changes occur in the execution window or link, the satellite uses a lightweight heuristic algorithm (e.g., greedy, Min-Min, local search, or distributed iterative method based on Lagrangian relaxation) for rapid correction and reallocation. If the solver cannot return a feasible solution within the specified time, the system triggers a degradation strategy: increasing the compression ratio r and reducing the data volume D; or adopting a task priority strategy, completing high-priority subtasks first and delaying or transferring low-priority subtasks to ground processing.

[0048] In this embodiment, the output of step S4 is the final task allocation table {x}. j,i The allocation table and the corresponding transmission timing table will be sent to the inter-satellite communication module for execution in step S5.

[0049] This implementation method closely integrates the weighted optimization formula described above with actual energy, computing power and link constraints, so that task allocation has both theoretical optimality and engineering feasibility, thereby achieving the technical effect of "reduced latency and controlled energy consumption" expected by this invention.

[0050] In a preferred embodiment, in the above optimization model, the total energy consumption consists of the computational energy consumption and communication energy consumption of each sub-task on the target satellite, and the calculation formula is as follows: ; in, For the computational power consumption of satellite i, t c,i P represents the computation time of the subtask on satellite i. t,i For the power consumption of the communication link, t t,i This refers to the corresponding transmission duration.

[0051] To ensure the feasibility of task allocation, energy constraints need to be added: ; This means that the total energy consumption of the mission cannot exceed the satellite's remaining available energy. This prevents the satellite from running out of energy or affecting other functions.

[0052] The total energy consumption E of all tasks must satisfy: ; This total energy consumption can be used as an optimization objective (aiming to minimize total energy consumption), or it can be combined with latency to form a weighted optimization objective: where α and β are adjustable weights used to balance task latency and energy consumption priority. By solving this optimization problem, the system generates the final task allocation table and execution timing table, and sends them to the satellite for execution.

[0053] In a preferred embodiment, α=0.8 and β=0.2 can be used for real-time SAR imaging tasks; in another preferred embodiment, α=0.3 and β=0.7 can be used for energy-saving priority tasks.

[0054] Assuming the PC executes on Sat-B: Calculate energy consumption: ; In the above formula The computational energy consumption for executing subtask i on satellite j; The computing power consumption of the satellite processor (here, we take 90 W). The computation time is equal to the computation amount of the subtask. Divide by satellite computing power Communication power consumption (calculated based on 15W laser link power consumption): 15 × 48 = 720 J; In the above formula Communication energy consumption for subtask transmission; The power consumption of the communication link (here, we assume a 15W laser link power consumption). Transmission time is equal to the data volume divided by the link rate. The sum of energy consumption can be calculated from the above formula as follows: In task allocation and optimization, after obtaining the scheduling results of each subtask, the optimization module estimates the energy consumption of each subtask based on the aforementioned calculation and communication energy consumption formulas. Specifically, for each subtask, based on the computing power and link rate of the satellite it is allocated to, the corresponding calculation time and transmission time are first calculated, and then substituted into the calculation and communication energy consumption formulas to obtain the energy consumption value of that subtask. Subsequently, the energy consumption results of all subtasks are summed to obtain the total energy consumption of the entire task process. In this embodiment, the total energy consumption obtained by the optimization module after the above calculation is approximately 2100J.

[0055] As a comparative embodiment of the present invention, if the traditional "single-satellite processing" method is used, Sat-A alone completes 1 TFLOPs of computation in approximately 1 / 0.5 = 2 seconds, but needs to transmit 50GB of data back to the ground station (at a rate of 1Gbps), with a transmission time of approximately 400 seconds and energy consumption exceeding 5000J. Therefore, the optimized intra-constellation coordination scheme of the present invention reduces latency by approximately 85% and energy consumption by approximately 58%.

[0056] Step S5: Inter-satellite transmission and execution Based on the optimized allocation results, mission data is transmitted sequentially among the four satellites and executed by the corresponding satellites. After each sub-task is completed, its results are transmitted to the next execution node via inter-satellite links.

[0057] Step S6: Result Fusion and Encryption The final imaging results are fused and encrypted by Sat-A to generate a complete SAR image. The results are completed within 60 seconds and transmitted to the ground control center via an encrypted link.

[0058] As can be seen from the above embodiments, the method of the present invention utilizes distributed computing power within the constellation to collaboratively process tasks, and combines optimized scheduling and task allocation schemes. Under the condition of completing the same SAR imaging task, the total processing latency is approximately 58 seconds, and the total energy consumption is approximately 2100 joules. In contrast, the traditional method of processing and transmitting all data independently by a single satellite results in a total processing latency of approximately 402 seconds and a total energy consumption of approximately 5000 joules. Therefore, the present invention can significantly reduce the latency of task completion while effectively saving the overall energy consumption of the constellation, thereby significantly improving the efficiency and reliability of low-Earth orbit satellite constellations in large-scale data processing.

[0059] In a preferred embodiment, step S4, combined with the available computing power, remaining energy, and communication bandwidth of each satellite in the constellation, generates a candidate allocation scheme. This allocation scheme determines which satellite should perform each task, as well as the execution order and transmission path of the tasks. Based on the allocation results, the computation time and communication time of each task on the corresponding satellite can be determined: the computation time is determined by the task's computational load and the satellite's computing power, and the communication time is determined by the task's data volume and link bandwidth. Subsequently, the task allocation and optimization module uses the following energy consumption calculation method to evaluate the energy consumption of each sub-task: The satellite's computing power describes its energy consumption rate when processing subtasks, while the link's communication power describes its energy consumption rate when transmitting data. To ensure the feasibility of task allocation, the calculated energy consumption needs to be incorporated into an energy constraint: the energy consumption of each subtask must not exceed the satellite's remaining available energy. This constraint ensures that tasks will not exceed the satellite's available energy after execution, thus preventing energy depletion or interference with other functions. The total energy consumption is calculated as follows: The total energy consumption can be used as part of the optimization objective (to minimize the possibility of total energy depletion), or it can be combined with the satellite's remaining energy constraints to ensure the feasibility of the optimization results. The optimization model weighs the total energy consumption and the total mission delay to form a weighted optimization objective: Here, α and β are adjustable parameters, which implementers can choose according to different priorities of real-time performance or energy saving for the task. By solving this optimization problem, the system generates the final task allocation table and execution timing table, and sends them to the inter-satellite communication module for execution.

[0060] In this invention, task allocation and constraints are not executed independently, but rather form an iterative, interconnected process: each candidate allocation scheme first calculates the corresponding computation time and communication time to obtain energy consumption and latency, and then verifies whether constraints such as computing power, energy, and links are met; if not, the allocation scheme is eliminated or the allocation path is adjusted; if met, it is included in the calculation scope of the optimization objective function. Through this cyclical linkage mechanism of "allocation-constraint-optimization," this invention can dynamically find the task allocation scheme with optimal latency and energy consumption while ensuring that all constraints are met. For example, in the example, the preprocessing task (PC) is allocated to Sat-B for execution: the computation time is 0.286 seconds, corresponding to a computation energy consumption of 25.7 joules; the communication time is 48 seconds, corresponding to a communication energy consumption of 720 joules, and the total energy consumption is approximately 746 joules. In this way, task allocation, time calculation, energy consumption calculation, and constraint judgment are naturally connected in the embodiment, ensuring that the distributed task execution meets both latency requirements and energy constraints.

[0061] Therefore, this invention significantly improves processing real-time performance and reduces system energy consumption, thus meeting the needs of real-time SAR imaging and large-scale data processing.

[0062] This embodiment also provides a computer-readable storage medium, which may be a read-only memory (ROM), random access memory (RAM), flash memory, optical disk, hard disk, or other storage device. A computer program is stored on this storage medium, and when the program is executed by a computer or processor, the methods described above in this application can be implemented.

[0063] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of one embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing the present invention.

[0064] Those skilled in the art will understand that the modules in the apparatus of the embodiments can be distributed in the apparatus of the embodiments as described in the embodiments, or they can be located in one or more devices different from this embodiment with corresponding changes. The modules of the above embodiments can be combined into one module, or they can be further divided into multiple sub-modules.

[0065] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A communication processing system for enhancing the computing power of low-Earth orbit communication satellites based on constellation collaborative computing, characterized in that, include: The data acquisition module is used to acquire raw data generated by low-orbit satellite payloads and to cache and preprocess it. The task generation module is used to generate a task request package based on the original data and task requirements. The request package includes task type, computation requirements, latency threshold and priority information. The task decomposition module is used to divide the task request package into multiple subtasks with dependencies, the dependencies including data flow dependencies and computation order constraints; The distributed computing power scheduling module is used to collect computing power, energy and link status information of each satellite based on the high-speed inter-satellite link, and calculate candidate scheduling schemes according to the sub-task requirements to support distributed execution across satellites; The task allocation and optimization module is used to establish a multi-objective optimization model based on the scheduling results, with total delay and total energy consumption as optimization objectives; The inter-satellite communication module is used to transmit task input data and intermediate results via inter-satellite links under the control of the scheme generated by the allocation and optimization module. The mission execution module is used to execute sub-tasks locally on the target satellite using a general-purpose processor or hardware accelerator. The result fusion and encryption module is used to fuse, correct errors, and encrypt the execution results of multiple subtasks to obtain the final processing result and transmit it to the ground.

2. The system according to claim 1, characterized in that, The task allocation and optimization module uses the following objective function to solve for the task allocation scheme: ， Where T is the total delay, E is the total system energy consumption, and α and β are weighting factors that are dynamically adjusted according to the task type; The solution process is based on a set of feasibility constraints, including a task integrity constraint to ensure that each subtask is assigned to a single satellite, a computing resource constraint to limit the computational requirements of the subtasks to not exceed the available computing power of the target satellite, an energy constraint to ensure that the energy consumption of the subtasks does not exceed the remaining available energy of the satellite, and a link reachability constraint to ensure that the subtask data is transmitted within the link availability window.

3. The system according to claim 1, characterized in that, The distributed computing power scheduling module calculates the execution time of subtasks on the satellite using the following formula: ， in, The floating-point computation cost of the subtask. In order to allocate satellite computing power, For data size, This refers to the link speed.

4. The system according to claim 1, characterized in that, The task allocation and optimization module estimates energy consumption E using the following energy consumption formula: ； in, For satellite Calculated power, Calculate the duration for it. For its communication power, The duration of their communication.

5. The system according to any one of claims 1 to 3, characterized in that, The task allocation and optimization module performs scheduling optimization using the following objective function: ； Where T is the total delay, E is the total system energy consumption, and α and β are weighting factors that are dynamically adjusted according to the task type; The following constraints are introduced: Mission integrity constraint: Each sub-mission is assigned to only one satellite; Computing resource constraints: The computational workload of the allocated tasks shall not exceed the computing capacity of the satellite; Energy constraint: The energy consumption of the sub-task execution shall not exceed the remaining energy of the satellite; Link constraint: Subtask data transmission must be completed within the available window of the link.

6. A communication and processing method for enhancing the computing power of low-Earth orbit communication satellites based on constellation collaborative computing, characterized in that, Includes the following steps: S1. Raw data is collected by low-orbit satellites and a mission request package is generated. The request package includes mission type, computational requirements, latency threshold and priority information. S2. Decompose the task request package into multiple subtasks with dependencies. S3. Collect computing power, energy and link status information of each satellite in the constellation through inter-satellite links, and generate candidate scheduling schemes based on this information to realize distributed computing power scheduling; S4. Based on the candidate scheduling schemes, construct a multi-objective optimization model with total delay and total energy consumption as optimization objectives, and introduce task integrity constraints, computing power constraints, energy constraints and link constraints to solve for the optimal task allocation scheme. S5. According to the optimal task allocation scheme, the inter-satellite communication module is used to complete the cross-satellite transmission of sub-task data and intermediate results; the corresponding satellite executes the allocated sub-task and transmits the execution results to the result fusion and encryption module; S6. The results of the sub-tasks are fused, error corrected, and encrypted to generate complete task results and then transmitted to the ground.

7. The method according to claim 6, for each subtask in step S4, obtain its floating-point computational quantity. and data size ; Obtain the available computing power of each satellite and link speed ; The estimated execution time of the sub-task on the satellite is calculated using the following formula. : ； The subtasks are scheduled and allocated based on the execution time, enabling the tasks to be executed efficiently on the satellite constellation.

8. The method according to claim 7, characterized in that, Step S4 also includes the following steps: Obtain the computational power of each satellite With communication power and its corresponding computation time and communication duration ; Estimate the total energy consumption E of subtask allocation using the following formula: ； Based on the energy consumption assessment, the task allocation scheme is optimized to reduce the overall energy consumption of the satellite constellation.

9. The method according to claim 6, characterized in that, In step S4, the task allocation and optimization process aims to minimize latency and energy consumption, and adopts the following objective function: ； Where T is the total delay, E is the total system energy consumption, and α and β are weighting factors that are dynamically adjusted according to the task type.

10. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any one of claims 6-9.

Images