Method for analyzing compatibility between giant constellations based on Spark architecture

Abstract

The invention discloses a giant constellation compatibility analysis method based on a Spark architecture, and belongs to the field of compatibility analysis between satellite systems. According to the giant constellation compatibility analysis method based on the Spark architecture, the compatibility analysis process can be optimized, function decoupling can be completed, the algorithm parallelism degree can be improved, parallel calculation can be completed through the Spark architecture, the calculation efficiency can be improved, and the overall calculation time consumption can be effectively reduced. According to the technical key points, a parallel interference analysis architecture between giant constellations is established, and an interference analysis process is divided into a plurality of time slices according to set simulation starting and ending time and simulation step length; decomposing an interference analysis process into orbit deduction, satellite and earth station visibility calculation, earth station satellite selection, interference link and interfered link visibility calculation and interference index calculation in a time slice; based on parallel interference analysis of a Spark architecture, decoupling an interference analysis process between giant constellations according to a time dimension and a function dimension to form an independent module; and adding an action operator for the interference analysis process, wherein the adding positions of the action operator are a calculation end position of each time slice, an end position of orbit deduction, an end position of visibility calculation, an end position of earth station satellite selection, an end position of visibility calculation of an interference link and an interfered link, and an end position of interference index calculation.

Description

Technical Field

[0001] The present invention belongs to the field of compatibility analysis between satellite systems, and particularly relates to a method for analyzing the compatibility between large constellations based on the Spark architecture in this field. Background Art

[0002] The number and scale of large low-earth orbit (LEO) constellations planned internationally are increasing continuously. Among them, Starlink plans to have a scale of 42,000 satellites, and nearly 7,000 have been launched so far; Oneweb has deployed more than 400 satellites and plans to expand to 6,372. Other LEO large constellations such as TeleSat, Boing V-band, and Kuiper are also continuously expanding, resulting in an increasingly prominent interference problem between large LEO constellations. In the process of interference analysis, there are problems such as the large scale of LEO large constellations, complex interference scenarios, and large-scale deduction calculations. Existing interference analysis systems are all in a single-machine mode, with limited computing resources, low computing efficiency, and long time consumption, and cannot meet the timeliness requirements of interference analysis. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method for analyzing the compatibility between large constellations based on the Spark architecture, which can optimize the compatibility analysis process, complete function decoupling, improve the parallelism of the algorithm, and use the Spark architecture to complete parallel computing, improve computing efficiency, and effectively reduce the overall computing time consumption.

[0004] The present invention adopts the following technical solutions:

[0005] A method for analyzing the compatibility between large constellations based on the Spark architecture, which is improved in that it includes the following steps:

[0006] Step 1, build a parallel interference analysis architecture between large constellations:

[0007] Divide the interference analysis process into multiple time slices according to the set simulation start and end times and simulation step lengths, and decompose the interference analysis process into five sub-processes: orbit deduction, satellite and earth station visibility calculation, earth station satellite selection, interference link and interfered link visibility calculation, and interference index calculation within the time slice;

[0008] Step 2, parallel interference analysis based on the Spark architecture:

[0009] Step 21, decouple the interference analysis process between large constellations in terms of time dimension and function dimension to form separate modules; the function dimension includes five functions: orbit deduction, satellite and earth station visibility calculation, earth station satellite selection, interference link and interfered link visibility calculation, and interference index calculation.

[0010] Step 22: Add an action operator to the interference analysis process and divide the interference analysis process into five jobs;

[0011] The positions where the action operator is added are: the end position of the calculation for each time slice, the end position of the orbit deduction, the end position of the visibility calculation, the end position of the earth station satellite selection, the end position of the visibility calculation for the interference link and the interfered link, and the end position of the interference index calculation;

[0012] The implementation method for each job is as follows:

[0013] Job-1, orbit deduction: Use the orbit model to deduce the satellite orbit and calculate the position of the satellite at the current time slice;

[0014] Job-2, visibility calculation: Based on the position of the earth station and the satellite position deduced in Job-1, analyze whether the satellite and the earth station are visible at the current time slice:

[0015] Calculate the distance D between the satellite and the earth station:

[0016]

[0017] In the above formula, x 1 、y 1 、z 1 are the position vectors of the satellite, and x 2 、y 2 、z 2 are the position vectors of the earth station;

[0018] The distance R of the satellite from the earth's center 1 is: The distance R of the earth station from the earth's center 2 is: The distances D of the satellite and the earth station to the horizon i :

[0019]

[0020] In the above formula, R e is the earth's radius;

[0021] When the distance D between the satellite and the earth station is less than D 1 + D 2 , the satellite and the earth station are visible;

[0022] Job-3, earth station satellite selection: The earth station selects a satellite from the visible satellites according to the satellite selection strategy for link establishment;

[0023] job-4, Visibility calculation for the interference link and the interfered link: When the earth stations of the interference system and the interfered system have both completed satellite selection and link establishment through job-3, the visibility calculation method of job-2 is used to analyze whether there is visibility between the transmitting end of the interference system and the receiving end of the interfered system. If there is visibility, there is interference within this time slice, and the calculation of job-5 needs to be performed;

[0024] job-5, Interference index calculation;

[0025] Step 23, The Spark architecture automatically identifies whether the RDD inside the job is a wide dependency or a narrow dependency. When a wide dependency is identified, it divides the stage, thus automatically dividing the job into stages with the wide dependency as the boundary;

[0026] Step 24, According to the number of times each stage in the Spark architecture needs to be repeatedly calculated, each stage is divided into corresponding tasks;

[0027] Step 25, Distribute each Task to the specified Executor for execution.

[0028] Furthermore, for job-1, the orbit models for orbit deduction include J2, SGP4, and SDP4.

[0029] Furthermore, for job-3, the satellite selection strategies for earth station satellite selection include the longest tracking time, the maximum communication elevation angle, the shortest communication distance, and the maximum separation angle.

[0030] Furthermore, for job-5, the interference indexes for interference index calculation include C / N, C / I, I / N, C / (I + N), △T / T, PFD, and EPFD.

[0031] Furthermore, in Step 24, the number of tasks for orbit deduction is equal to the number of satellites that need to undergo orbit deduction, which is equal to the number of satellites in the interference system + the number of satellites in the interfered system; the number of tasks for satellite-earth station visibility calculation is equal to the number of earth stations in the interference system × the number of satellites in the interference system + the number of earth stations in the interfered system × the number of satellites in the interfered system; the number of tasks for earth station satellite selection is equal to the number of earth stations in the interference system + the number of earth stations in the interfered system; the number of tasks for interference link and interfered link visibility calculation is equal to the number of interference links × the number of interfered links; the number of tasks for interference index calculation is equal to the number of visible pairs between the receiving end of the interference link and the transmitting end of the interfered link.

[0032] The beneficial effects of the present invention are:

[0033] The method disclosed by the present invention designs parallel acceleration in two dimensions for the compatibility analysis algorithm, optimizes the compatibility analysis process, completes function decoupling, improves the parallelism of the compatibility analysis algorithm, and maps it to the Spark architecture capable of parallel computing. The Spark architecture is used to complete parallel computing, greatly improving the interference analysis calculation efficiency and the utilization rate of computing resources, and providing strong support for frequency and orbit resource declaration, coordination, and maintenance. Description of the Drawings

[0034] Figure 1 It is a schematic diagram of the architecture for parallel interference analysis between megaconstellations;

[0035] Figure 2 It is a schematic diagram of the process for Spark to achieve parallel computing;

[0036] Figure 3 It is a schematic diagram of the visibility between a satellite and an earth station;

[0037] Figure 4 It is a schematic diagram of the visibility between an interfering link and an interfered link. Detailed Embodiments

[0038] In order to make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0039] Embodiment 1. Aiming at the problems of complex interference scenarios, large interference analysis calculation volume, high calculation complexity, and long time consumption between megaconstellations in low earth orbit, this embodiment discloses a method for compatibility analysis between megaconstellations based on the Spark architecture, including the following steps:

[0040] Step 1, build an architecture for parallel interference analysis between megaconstellations:

[0041] As Figure 1 shown, in order to improve the interference analysis efficiency between megaconstellations and reduce the calculation time consumption, parallel computing of interference analysis tasks between megaconstellations is realized from two dimensions:

[0042] (1) Parallel computing between different time slices:

[0043] According to the set start and end times and simulation step length of the simulation, the interference analysis process is divided into multiple time slices. The calculations between time slices do not affect each other, and the calculation tasks of different time slices can be assigned to different computing nodes through Spark for parallel computing.

[0044] (2) Parallel computing within a time slice:

[0045] Within a time slice, the interference analysis process is decomposed into five sub-processes: orbit deduction, satellite and earth station visibility calculation, earth station satellite selection, interference link and interfered link visibility calculation, and interference index calculation;

[0046] Each sub-process requires multiple calculations, so parallel calculations can be performed within each sub-process. Taking orbit deduction as an example, the number of orbit deductions required within a single time slice is equal to the total number of satellites. Therefore, the number of orbit deductions that can be parallelized is min(total number of satellites, number of parallel threads allocated to this time slice by the Spark architecture).

[0047] Step 2, parallel interference analysis based on the Spark architecture:

[0048] As Figure 2 shown, first, the interference analysis process is divided into multiple jobs with the action operator as the boundary; within each job, by analyzing the dependencies between RDDs, the job is divided into multiple stages; finally, the stage is divided into tasks, and tasks can be executed in parallel with each other.

[0049] Job: A parallel calculation consisting of multiple tasks, often generated by the Spark Action operator. Multiple jobs are often generated in one Application;

[0050] RDD: Resilient Distributed Datasets;

[0051] Stage: Each job is split into multiple groups of tasks, which are regarded as a TaskSet, and its name is stage;

[0052] Task: A unit of work sent to a certain Executor;

[0053] Step 21, decouple the interference analysis process between giant constellations in terms of time dimension and functional dimension to form separate modules, and perform parallel calculations through Spark, which greatly speeds up the simulation speed; the time dimension ensures that the interference analysis processes between different time slices do not interfere with each other; the functional dimension includes five functions: orbit deduction, satellite and earth station visibility calculation, earth station satellite selection, interference link and interfered link visibility calculation, and interference index calculation;

[0054] Step 22, add action operators to the interference analysis process and divide the interference analysis process into five jobs;

[0055] The added positions of the action operator are: the end position of the calculation for each time slice, the end position of the orbit deduction, the end position of the visibility calculation, the end position of the satellite selection for the earth station, the end position of the visibility calculation for the interfering link and the interfered link, and the end position of the interference index calculation;

[0056] The implementation method for each job is as follows:

[0057] job-1, Orbit Deduction: Use orbit models such as J2, SGP4 / SDP4, etc. to deduce the satellite orbit and calculate the position of the satellite at the current time slice;

[0058] job-2, Visibility Calculation: Based on the position of the earth station and the satellite position deduced in job-1, analyze whether the satellite and the earth station are visible at the current time slice:

[0059] Calculate the distance D between the satellite and the earth station:

[0060]

[0061] In the above formula, x 1 , y 1 , z 1 are the position vectors of the satellite, and x 2 , y 2 , z 2 are the position vectors of the earth station;

[0062] The distance R of the satellite from the earth's center 1 is: The distance R of the earth station from the earth's center 2 is: The distances D of the satellite and the earth station to the horizon i are:

[0063]

[0064] In the above formula, R e is the radius of the earth;

[0065] As Figure 3 shown, when the distance D between the satellite and the earth station is less than D 1 + D 2 , that is: D < D 1 + D 2 , the satellite and the earth station are visible;

[0066] job-3, Satellite Selection for the Earth Station: The earth station selects a satellite for link establishment from the visible satellites according to the satellite selection / link establishment strategy. The satellite selection strategies include the longest tracking time, the maximum communication elevation angle, the shortest communication distance, the maximum separation angle, etc.;

[0067] job-4, Interference Link and Interfered Link Visibility Calculation: As Figure 4 shown, when the earth stations of the interference system and the interfered system have both completed satellite selection and link establishment through job-3, the visibility calculation method of job-2 is used to analyze whether there is visibility between the transmitting end of the interference system and the receiving end of the interfered system. If there is visibility, interference exists within this time slice, and the calculation of job-5 needs to be performed;

[0068] Suppose in a scenario, system a interferes with system b. Then system a is the interference system and system b is the interfered system.

[0069] job-5, Interference Index Calculation: Calculate interference indexes specified by ITU such as C / N, C / I, I / N, C / (I + N), △T / T, PFD, and EPFD according to requirements;

[0070] Step 23, The Spark architecture automatically identifies whether the RDD inside the job is a wide dependency or a narrow dependency. When a wide dependency is identified, the stage is divided, and thus the job is automatically divided into stages with the wide dependency as the boundary;

[0071] Step 24, According to the number of times each stage needs to be repeatedly calculated in the Spark architecture, each stage is divided into corresponding tasks;

[0072] The number of tasks for orbit deduction is equal to the number of satellites that need to perform orbit deduction, which is equal to the number of satellites in the interference system + the number of satellites in the interfered system; the number of tasks for satellite and earth station visibility calculation is equal to the number of earth stations in the interference system × the number of satellites in the interference system + the number of earth stations in the interfered system × the number of satellites in the interfered system; the number of tasks for earth station satellite selection is equal to the number of earth stations in the interference system + the number of earth stations in the interfered system; the number of tasks for interference link and interfered link visibility calculation is equal to the number of interference links × the number of interfered links; the number of tasks for interference index calculation is equal to the number of visible links between the receiving end of the interference link and the transmitting end of the interfered link.

[0073] Step 25, Distribute each Task to the specified Executor for execution.

[0074] Executor: A process in which an Application runs on a worker node.

[0075] Next, select the Starlink constellation and the Oneweb constellation as the interference simulation objects to analyze the downlink interference of the Starlink constellation on the Oneweb constellation. The simulation period is 1 day and the simulation step is 1 s. The orbital parameters of the Oneweb constellation are shown in Table 1 below.

[0076] Table 1: Orbital Parameters of Oneweb Constellation

[0077]

[0078] The system parameters of the Oneweb constellation are shown in Table 2 below.

[0079] Table 2: Oneweb constellation system parameters

[0080] Parameter Value Peak gain of satellite transmitting antenna / dBi 28 Half-power beamwidth of satellite transmitting antenna / (°) 4.9 Peak gain of ground station receiving antenna / dBi 38 Half-power beamwidth of ground station receiving antenna / (°) 1.6 Satellite transmitting power / dBW 3 Communication frequency / GHz 17.8 Communication bandwidth / MHz 250 System noise temperature / K 300 Ground station location / (°E, °N) (120,40)

[0081] The orbital parameters of the Starlink constellation are shown in Table 3 below.

[0082] Table 3: Starlink constellation orbital parameters

[0083]

[0084] The system parameters of the Starlink constellation are shown in Table 4 below.

[0085] Table 4: Starlink constellation system parameters

[0086]

[0087] (1) Based on the designed parallel dimensions, decouple the parallel interference analysis process between giant constellations, including the time dimension and the functional dimension:

[0088] 1) Parallel computing in the time dimension:

[0089] In this embodiment, the simulation period is 1 day and the simulation step is 1 s, which can be divided into 24 * 60 * 60 = 86,400 time slices. For each time slice, the corresponding computing resources can be allocated by the Spark cluster for parallel computing.

[0090] 2) Parallel computing in the functional dimension:

[0091] Decouple the functional dimension into functions such as orbit deduction, visibility calculation, earth station satellite selection, interference link and interfered link visibility calculation, and interference index calculation. Each function corresponds to a job, and parallel computing can be performed within each job.

[0092] job-1, orbit deduction: Use the J2, SGP4 / SDP4 orbit models to deduce the satellite trajectories and calculate the satellite positions at each moment;

[0093] In this scenario, the number of satellites is 1969 + 11908 = 13877. Therefore, the orbit deduction needs to be repeated 13877 times for each time slice. The orbit deductions of different satellites are independent of each other, and the corresponding computing resources can be allocated by the Spark cluster for parallel computing.

[0094] job-2, Visibility calculation: Based on the location of the earth station and the satellite location obtained in job-1, calculate whether the satellite and the earth station are visible to each other.

[0095] In this scenario, the number of visibility calculations required for each time slice is 1969×1 + 11908×5 = 61509 times. The visibility between different satellites and the earth station is not related to each other, and the corresponding computing resources can be allocated by the Spark cluster for parallel computing.

[0096] job-3, Earth station satellite selection: The earth station selects a satellite for link establishment from the visible satellites obtained in job-2 according to the satellite selection strategy. The specific satellite selection strategies include: longest tracking time, maximum communication elevation angle, shortest communication distance, maximum separation angle, etc.

[0097] Assume that after the calculation in (2), the number of satellites visible to the earth station accounts for 10% of the total number of satellites. Then, at a single moment, 1969×1%×1 + 11908×1%×5 = 615 times are required. The satellite selection calculations between different earth stations and satellites are not related to each other, and the corresponding computing resources can be allocated by the Spark cluster for parallel computing.

[0098] job-4, Visibility calculation of the interfering link and the interfered link: When the earth stations of the interfering system and the interfered system have both completed satellite selection and link establishment through job-3, it is necessary to analyze whether the transmitting end of the interfering system and the receiving end of the interfered system are visible through the visibility calculation method introduced in job-2.

[0099] Since only 5 interfering links and 1 interfered link are constructed in this scenario, only 5 visibility analyses are required. The visibility between different satellites and the earth station is not related to each other, and the corresponding computing resources can be allocated by the Spark cluster for parallel computing.

[0100] job-5, Interference index calculation: Calculate the interference indexes specified by ITU such as C / N, C / I, I / N, C / (I + N), △T / T, PFD, EPFD, etc. according to the requirements.

[0101] Since 5 interfering links and 1 interfered link are constructed in this scenario, M single-entry interference calculations + 1 lumped interference calculation are required (0 ≤ M ≤ 5). The M single-entry interferences are not related to each other, and the corresponding computing resources can be allocated by the Spark cluster for parallel computing.

[0102] (2) Spark automatically divides the job into stages with wide dependencies as the boundary by analyzing the dependencies of the RDDs within the job.

[0103] (3) Divide each stage into corresponding tasks according to the number of times each stage needs to be recalculated in the Spark architecture. (4) Distribute each task to the specified Executor for execution.

Claims

1. A method for analyzing compatibility between giant constellations based on Spark architecture, characterized in that: The steps include: Step 1: Build a parallel interference analysis architecture between giant constellations: The interference analysis process is divided into multiple time slices according to the set simulation start and end time and simulation step size. Within the time slice, the interference analysis process is decomposed into five sub-processes: orbit deduction, satellite and earth station visibility calculation, earth station selection, interference link and interfered link visibility calculation, and interference index calculation. Step 2: Parallel interference analysis based on Spark architecture: Step 21, decoupling the giant constellation interference analysis process according to the time dimension and the functional dimension to form a separate module; the functional dimension includes five functions: orbit deduction, satellite and earth station visibility calculation, earth station selection, interference link and interfered link visibility calculation, and interference index calculation; Step 22, add an action operator to the interference analysis process and divide the interference analysis process into five jobs; The action operator is added at the following locations: the end of calculation of each time slice, the end of orbit deduction, the end of visibility calculation, the end of earth station selection, the end of visibility calculation of the interfering link and the interfered link, and the end of interference index calculation; The implementation method of each job is as follows: Job-1, orbit deduction: Use the orbit model to deduce the satellite orbit and calculate the satellite position in the current time slice; Job-2, visibility calculation: Based on the earth station position and the satellite position derived in job-1, analyze whether the satellite and the earth station are visible in the current time slice: Calculate the distance D between the satellite and the earth station: In the above formula, x1, y1, z1 are the position vectors of the satellite, and x2, y2, z2 are the position vectors of the earth station; The distance R1 from the satellite to the center of the earth is: The distance R2 from the earth station to the center of the earth is: The distance D from the satellite and the earth station to the horizon i : In the above formula, R e is the radius of the Earth; When the distance D between the satellite and the earth station is less than D1+D2, the satellite and the earth station are visible; Job-3, earth station satellite selection: The earth station selects a satellite from the visible stars to establish a link based on the satellite selection strategy; Job-4, visibility calculation of the interfering link and the interfered link: After both the earth station of the interfering system and the earth station of the interfered system have completed satellite selection and link establishment through job-3, the visibility calculation method of job-2 is used to analyze whether the transmitter of the interfering system and the receiver of the interfered system are visible. If they are visible, there is interference in this time slice, and job-5 calculation is required; job-5, interference index calculation; Step 23: The Spark architecture automatically identifies whether the RDD in the job is a wide dependency or a narrow dependency. If a wide dependency is identified, the job is divided into stages, thereby automatically dividing the job into stages based on the wide dependency as the boundary. Step 24: Divide each stage into corresponding tasks according to the number of times each stage of the Spark architecture needs to be repeatedly calculated; Step 25: Distribute each Task to the specified Executor for execution.

2. The method for analyzing compatibility between giant constellations based on the Spark architecture according to claim 1, characterized in that: Job-1, the orbit models for orbit deduction include J2, SGP4 and SDP4.

3. The method for analyzing compatibility between giant constellations based on the Spark architecture according to claim 1, characterized in that: Job-3, the satellite selection strategy of the earth station includes the longest tracking time, the maximum communication belief angle, the shortest communication distance and the maximum separation angle.

4. The method for analyzing compatibility between giant constellations based on the Spark architecture according to claim 1, characterized in that: Job-5, interference index calculation includes C / N, C / I, I / N, C / (I+N), △T / T, PFD, and EPFD.

5. The method for analyzing compatibility between giant constellations based on Spark architecture according to claim 1, characterized in that: In step 24, the number of tasks for orbital deduction is equal to the number of satellites that need to perform orbital deduction, which is equal to the number of satellites in the interfering system + the number of satellites in the interfered system; the number of tasks for calculating the visibility of satellites and earth stations is equal to the number of earth stations in the interfering system × the number of satellites in the interfering system + the number of earth stations in the interfered system × the number of satellites in the interfered system; the number of tasks for selecting earth stations for earth stations is equal to the number of earth stations in the interfering system + the number of earth stations in the interfered system; the number of tasks for calculating the visibility of interfering links and interfered links is equal to the number of interfering links × the number of interfered links; the number of tasks for calculating the interference index is equal to the number of visible interfering link receivers and interfered link transmitters.

Images