Prediction method and system for relay satellite mission conflict probability distribution

Abstract

The present invention relates to the field of satellite measurement, operation and control technology, and in particular discloses a method and system for predicting the probability distribution of relay satellite mission conflicts, including constructing a portrait label system for the relay field based on historical data; constructing a neural network prediction model using the historical data classified by the portrait label system; determining the predicted application probability of relay resources for each user within a target time period based on the portrait label system and the neural network prediction model; and predicting the probability of conflicts within the target time period from the perspective of the user and the perspective of the operation management center based on the predicted application probability. The user's application probability is predicted by constructing a neural network prediction model, and then the conflict probability within the target time period is predicted from the perspective of the user and the perspective of the operation management center based on the predicted application probability. The conflict prediction results can be used to guide the user center to reasonably apply for relay resources to avoid possible conflicts in advance.

Description

Technical Field

[0001] The present invention relates to the field of satellite measurement, operation and control technology, and in particular to a method and system for predicting the probability distribution of relay satellite mission conflicts. Background Art

[0002] Currently, the Relay Satellite Operations Management Center (OMC) sends information about the next week's available windows to each user center at a fixed time each week. Users can then submit requests for tasks based on their needs. This process is often unaware of each other's submission status, leading to information barriers and potential conflicts. Furthermore, once a user's request is submitted, the OMC can only adjust resources within the sliding window, alternative window, and other conditions specified in the user's request, resulting in limited redundancy. Summary of the Invention

[0003] Based on this, it is necessary to provide a method and system for predicting the probability distribution of relay satellite mission conflicts in order to address the mission conflicts that may occur when the user center submits an application.

[0004] A method for predicting relay satellite mission conflict probability distribution, comprising:

[0005] Build a portrait label system in the relay field based on historical data;

[0006] Using the historical data classified by the portrait label system, a neural network prediction model is constructed;

[0007] Determine the predicted probability of each user applying for relay resources within a target time period based on the portrait label system and the neural network prediction model;

[0008] The conflict probability within the target time period is predicted from the perspective of the user and the perspective of the operation management center according to the predicted application probability.

[0009] In one embodiment, the portrait label system includes user portrait, resource portrait, and business portrait. The user portrait includes user center portrait and user target portrait. The resource portrait includes relay satellite portrait and relay ground station portrait.

[0010] In one embodiment, constructing a neural network prediction model using the historical data classified by the portrait label system includes:

[0011] Preprocessing the historical data, and dividing the preprocessed historical data into a training set and a test set according to a preset ratio;

[0012] Create neural network prediction models based on deep learning frameworks;

[0013] The training set is used to train and optimize the neural network prediction model.

[0014] In one embodiment, when the operation management center publishes an idle window for a target time period and no user has submitted an application for the target time period, performing a conflict probability prediction within the target time period from the perspective of a single user target based on the predicted application probability includes:

[0015] Get the predicted application probability p of each other user target i applying to use a single relay satellite in a single time block during the target time period i ;

[0016] According to the mathematical model of the conflict probability of a single user target i applying for a single time block in the target time period to use a single relay satellite, the conflict probability f of the user target applying for a single time block is determined. t ;

[0017] The mathematical model of the conflict probability of a single user target i applying to use a single relay satellite in a single time block of a target time period is:

[0018]

[0019] Where p is the actual application probability of a single user target in a single time block, s is the total number of historical applications of a single user target in a single time block; s max The total number of weeks for historical data request; f t The conflict probability of applying for a single relay satellite to use a single time block in the target time period for a single user target, p i The probability of applying for prediction for another user target in the same time block, n is the number of other user targets except the currently calculated user target, and total is the number of all user targets.

[0020] In one embodiment, when the operation management center publishes an idle window for a target time period and a user submits an application for the target time period, performing a conflict probability prediction within the target time period from the perspective of a single user target based on the predicted application probability includes:

[0021] Looping through the user application queue to obtain the time blocks that a single user target has applied for, and adding 1 to the historical application count s of each of the time blocks that have been applied for;

[0022] Calculate the actual application probability p of the single user target for the applied time block according to formula (1);

[0023] For other user targets i that have not submitted applications, the predicted application probability p of the applied time block is calculated one by one according to the neural network prediction model. i ;

[0024] Calculate n according to formula (3);

[0025] According to formula (2), the conflict probability f of the user's target application for the applied time block is calculated t .

[0026] In one embodiment, after the operation management center completes the planning for the target time period, it performs emergency insertion for temporary applications and emergency applications of users based on the priority and first-come-first-served strategy according to preset constraints. The conflict probability prediction within the target time period is performed from the perspective of a single user target based on the predicted application probability, including:

[0027] Looping through the user application queue to obtain the time blocks that a single user target has applied for, and adding 1 to the historical application count s of each of the time blocks that have been applied for;

[0028] Calculate the actual application probability p of the single user target for the applied time block according to formula (1);

[0029] Circularly take out the emergency time blocks applied for by a single user from the emergency application queue, and add 1 to the historical application times s of each emergency time block;

[0030] Recalculate the actual application probability p of the single user target for the emergency time block according to formula (1);

[0031] Calculate n according to formula (3);

[0032] According to formula (2), the conflict probability f of the user's target application for the emergency time block is calculated t .

[0033] In one embodiment, when the operation management center publishes an idle window for a target time period and no user has submitted an application for the target time period, performing a conflict probability prediction within the target time period from the perspective of a single user center based on the predicted application probability includes:

[0034] Get the predicted application probability p of other user targets j who are not the current user center and apply to use a single relay satellite in a single time block during the target time period j ;

[0035] According to the mathematical model of the conflict probability of a single user center applying to use a single relay satellite in a single time block during the target time period, the conflict probability f of a single user center applying to use a single relay satellite in a single time block during the target time period is determined. c ;

[0036] The mathematical model of the conflict probability of a single user center applying to use a single relay satellite in a single time block of a target time period is:

[0037]

[0038] Where, f c The conflict probability of a single user center applying to use a single relay satellite in a single time block during the target time period; p j The probability of applying for a single user target j belonging to other user centers to use the same relay satellite in the same time block; m is the number of all user targets belonging to other user centers; Total is the number of all user targets, and Sum0 is the number of user targets in the current user center;

[0039] Circularly calculate the conflict probability of the current user center applying to use a single relay satellite in all time blocks of the target time period;

[0040] Circularly calculate the conflict probability of the current user center applying to use all relay satellites in all time blocks of the target time period.

[0041] In one embodiment, when the operation management center publishes an idle window for a target time period and a user submits an application for the target time period, performing a conflict probability prediction within the target time period from the perspective of a single user center based on the predicted application probability includes:

[0042] Circularly take out the time blocks applied for by a single user target from the application queue, and add 1 to the historical application count s of the applied time blocks;

[0043] Calculate the actual application probability p of the single user target for the applied time block according to formula (1);

[0044] For other user targets j that have not submitted applications, the predicted application probability p of the applied time block is calculated one by one according to the neural network prediction model. j ;

[0045] Calculate m according to formula (5);

[0046] According to formula (4), the conflict probability f of the current user center applying to use a single relay satellite in a single time block is calculated as follows: c ;

[0047] Circularly calculate the conflict probability of the current user center applying to use a single relay satellite in all time blocks of the target time period;

[0048] The conflict probability of the current user center applying to use all relay satellites in all time blocks of the target time period is cyclically calculated.

[0049] In one embodiment, predicting the conflict probability within the target time period from the perspective of the operation management center based on the predicted application probability includes:

[0050] According to the neural network prediction model, the predicted application probability p of a single user target k applying for a single time block in the target time period is calculated one by one k ;

[0051] According to the probability mathematical model of all users applying to use a single relay satellite a in the same time block of the target time period, the probability r of a single relay satellite a being applied in the same time block of the target time period is calculated. a ;

[0052] The probability mathematical model of all user target applications using a single relay satellite a in the same time block of the target time period is:

[0053]

[0054] Where p k The predicted application probability for user target k to use a single relay satellite a in the same time block, z is the total number of user targets;

[0055] Circularly calculate the probability of a single relay satellite a being applied for in all time blocks of the target time period;

[0056] The probability of all relay satellites being applied for in all time blocks of the target time period is cyclically calculated.

[0057] In one embodiment, a system for predicting relay satellite mission conflict probability distribution includes:

[0058] Data processing module, used to build a portrait label system in the relay field based on historical data;

[0059] A model building module, for building a neural network prediction model using the historical data classified by the portrait label system;

[0060] The prediction analysis module is used to determine the predicted application probability of relay resources for each user within the target time period based on the portrait label system and the neural network prediction model, and is also used to predict the conflict probability within the target time period from the perspective of the user and the perspective of the operation management center based on the predicted application probability.

[0061] The aforementioned method for predicting the probability distribution of relay satellite mission conflicts analyzes user resource usage habits based on historical data and constructs a relay domain profile tagging system. Using the historical data categorized by the profile tagging system, a neural network prediction model is constructed. Based on the profile tagging system and the neural network prediction model, the predicted application probability for each user within a target time period is determined. Furthermore, based on the predicted application probability, the conflict probability within the target time period is predicted from both the user's perspective and the operations management center's perspective. The conflict prediction results can be used to guide user centers in making appropriate relay resource applications to avoid potential conflicts in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the implementation methods of this specification or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the implementation methods or the description of the prior art. Obviously, the drawings described below are only some implementation methods recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0063] Figure 1 This is a flow chart of a method for predicting relay satellite mission conflict probability distribution according to one embodiment of the present application;

[0064] Figure 2 A flowchart of a method for constructing a neural network prediction model according to one embodiment of the present application;

[0065] Figure 3 This is a flowchart of a method for predicting the probability of conflict within a target time period from the perspective of a single user target in one embodiment of the present application;

[0066] Figure 4 This is a flowchart of a method for predicting the probability of conflict within a target time period from the perspective of a single user target in another embodiment of the present application;

[0067] Figure 5 This is a flowchart of a method for predicting the probability of conflict within a target time period from the perspective of a single user target according to another embodiment of the present application;

[0068] Figure 6 This is a flowchart of a method for predicting the probability of conflict within a target time period from a single user-centric perspective in one embodiment of the present application;

[0069] Figure 7 This is a flowchart of a method for predicting the probability of conflict within a target time period from a single user-centric perspective according to another embodiment of the present application;

[0070] Figure 8This is a flowchart of a method for predicting the probability of conflict within a target time period from the perspective of an operation management center in one embodiment of the present application;

[0071] Figure 9 This is a schematic diagram of the structure of a system for predicting relay satellite mission conflict probability distribution in one embodiment of the present application. DETAILED DESCRIPTION

[0072] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. Preferred embodiments of the present invention are shown in the accompanying drawings. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the present disclosure.

[0073] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0074] In response to the task conflict problem that may occur when the user center submits an application, this application provides a prediction method for the probability distribution of relay satellite task conflicts to guide users to submit reasonable applications. Figure 1 This is a flow chart of a method for predicting the probability distribution of relay satellite mission conflicts according to one embodiment of the present application. In one embodiment, the method for predicting the probability distribution of relay satellite mission conflicts may include the following steps S100 to S400.

[0075] Step S100: Construct a portrait label system for the relay field based on historical data.

[0076] Collect data on users, resources, orbits, and other aspects to obtain historical data. In this embodiment, historical data may include, but is not limited to, historical data on relay satellite resource planning, historical user data, and orbital information related to relay satellites and user targets. By analyzing user resource usage habits (including device resources and time resources), a profile labeling system for the relay sector is established, providing a data foundation for the prediction model.

[0077] Step S200: Build a neural network prediction model using historical data classified by the portrait label system.

[0078] Using historical data classified by the portrait labeling system, a neural network prediction model is constructed. The neural network prediction model is based on deep learning theory, which enables the model's prediction results to achieve high accuracy and practicality.

[0079] Step S300: Based on the portrait label system and the neural network prediction model, determine the predicted application probability of each user within the target time period.

[0080] Based on the relay domain's portrait analysis and neural network prediction model, the predicted application probability of each user within the target time period can be calculated. Considering that the current operation management center usually sends the idle window information for the next week to each user center within a fixed time period each week, and each user center also submits a weekly application plan, in this embodiment of the application, the target time period can refer to the next week. In other words, the predicted application probability of each user within the next week is predicted.

[0081] Step S400: Conflict probability prediction within a target time period is performed from the perspective of the user and the perspective of the operation management center according to the predicted application probability.

[0082] A neural network prediction model is used to predict the application probability within the target time period, thereby predicting resource usage within the target time period. The task conflict probability distribution is predicted from multiple dimensions, including time distribution and device matching. This helps guide subsequent users to avoid periods with high conflict probabilities when applying for relay resources. Within the service time range (consisting of one or more consecutive time blocks) that meets the task requirements, the period with the lowest conflict probability is selected to rationally apply for relay resources.

[0083] The aforementioned method for predicting the conflict probability distribution of relay satellite missions analyzes user resource usage habits based on historical data and constructs a relay domain profile tagging system. Using the historical data categorized by the profile tagging system, a neural network prediction model is constructed. Based on this profile tagging system and the neural network prediction model, the predicted application probability for each user within a target time period is determined. Furthermore, based on the predicted application probability, the conflict probability within the target time period is predicted from both the user's perspective and the operations management center's perspective. The conflict prediction results can be used to guide user centers in making appropriate relay resource applications to proactively avoid potential conflicts.

[0084] In one embodiment, the portrait labeling system may include user portraits, resource portraits, and business portraits. Among them, the user portrait in the relay field mainly refers to the relay user center and the user targets of the relay user center. The relay user center can have one or more user targets. The user portrait may include the user center portrait and the user target portrait, and a user feature labeling system that conforms to the rules of user characteristics is constructed. The resource portrait can extract resource features by analyzing the historical operation records of the relay resources and performing trend calculations. The resource portrait may include the relay satellite portrait and the relay ground station portrait. A labeling system for user, resource, and business portraits in the relay field is constructed. Based on data such as the relay satellite resource planning history data, user history data, and relay satellite / user target related orbital information, a multi-dimensional feature description system for users, resources, and businesses is established, which can provide relatively accurate user center / user target portraits, relay satellite / ground station portraits, and resource application business portraits, providing a data foundation for fully utilizing the system's effectiveness.

[0085] In a preferred embodiment, as the user scale and business scope expand, a collaborative filtering recommendation algorithm can also be used to improve business profiles to maintain the efficiency and accuracy of resource application probability prediction.

[0086] In one embodiment, the label classification of the user center portrait may include the following types:

[0087] (1) User center type: relay operation management center, relay user center, etc.; (2) User center business type: measurement and control, operation control, measurement and operation control, data transmission, etc.; (3) Number of satellites managed by the user center: the number of managed satellites; (4) Normal user application submission time: Friday morning, Thursday afternoon, random, etc.; (5) Normal user application submission number: indicates the user's popularity; (6) Normal user application submission frequency: submitted in several times, once or multiple times: (7) Temporary user application submission time: Friday morning, Thursday afternoon, random, etc.; (8) Temporary user application submission number: indicates the user's popularity; (9) Temporary user application submission frequency: submitted in several times, once or multiple times: (10) Emergency user application submission time: Friday morning, Thursday afternoon, random, etc.; (11) Emergency user application submission number: indicates the user's popularity; (12) Emergency user application submission frequency: submitted in several times, once or multiple times; (13) User task priority and urgency: extremely urgent, urgent, ordinary; (14) User application adjustment status: no adjustment, occasional adjustment, frequent adjustment.

[0088] In one embodiment, the user targets of the relay satellite may include, but are not limited to, medium and low orbit satellites, space probes, spacecraft, space stations and space laboratories, space shuttles, space planes, and other spacecraft. The label classification of the user target profile may include the following types:

[0089] (1) User target type: manned spacecraft, unmanned spacecraft, etc.; (2) Service type: data relay, space-based measurement and control; (3) User target orbit: orbit altitude, inclination, etc.; (4) Number and type of terminals: number of S-band terminals, number of Ka-band terminals, number of laser terminals, etc.; (5) Data rate: type and rate of forward return data; (6) Data transmission quality: packet loss rate, bit error rate, disorder rate; (7) Data transmission frequency and transmission time: data transmission frequency and transmission time in a circle; (8) Security protection requirements: user encryption, remote control and telemetry both have clear and secret modes.

[0090] In one embodiment, the label classification of the relay satellite image may include the following types:

[0091] (1) Satellite type label: loads the satellite field and the Chinese name / code of the satellite and NoradID (NATO number); (2) Relay satellite transmission link label: divided into several identifiers according to the transmission rate; (3) Relay satellite frequency band label: transponder configuration information; (4) Mission execution label: the number of times and proportion of successful missions performed by the relay satellite; (5) Mission type label: calculates the number of times and proportion of mission types performed by the relay satellite; (6) Mission type label: calculates the number of times and proportion of mission types performed by the relay satellite; (7) Working mode label: calculates the number of times and proportion of each working mode executed by the relay satellite.

[0092] In one embodiment, the label classification of the relay ground station portrait may include the following types:

[0093] (1) Ground station latitude and longitude tag: relay ground station longitude, latitude, elevation and other geographic location information; (2) Ground station antenna tag: ground station antenna configuration information; (3) Ground station control event tag: statistics on the number and proportion of each control event type based on the daily, weekly and monthly dimensions; (4) Ground station management event tag: statistics on the number and proportion of each management event type based on the daily, weekly and monthly dimensions; (5) Ground station alarm tag: statistics on the number and proportion of each alarm type based on the daily, weekly and monthly dimensions.

[0094] In one embodiment, by combining user portraits and resource portraits, the user center conducts a characteristic analysis of the target application behavior of different types of users in long-term management tasks / major tasks / comprehensive long-term scenarios. At the same time, the user demand initiation behavior patterns, the spatiotemporal characteristics of user target demands, the characteristics of user target resource demands, the time characteristics of satellite-to-ground / inter-satellite link establishment, the degree of satisfaction of user target demands and other characteristics are analyzed to build a business portrait.

[0095] The label classification of business portraits can include the following types:

[0096] (1) User application tags: initiation time, initiation frequency, initiation demand, resource preference, time redundancy, user target category, priority distribution, demand initiation suddenness, demand initiation regularity, user importance; (2) User target demand spatiotemporal tags: satellite operation status, relative spatiotemporal relationship between user target and relay satellite resources, time distribution characteristics, mission duration, mission area distribution, etc.; (3) User target resource demand tags: mission type, priority, terminal type, real-time requirements, data volume, transmission rate requirements, service quality requirements, resource adaptability, resource exclusivity, etc.; (4) Satellite-to-ground / inter-satellite link establishment time tags: average satellite-to-ground link establishment time, minimum satellite-to-ground link establishment time, maximum satellite-to-ground link establishment time, average inter-satellite link establishment time, minimum inter-satellite link establishment time, maximum inter-satellite link establishment time, etc.; (5) User target demand satisfaction degree tags: scheduling success rate, actual execution completion rate, etc.

[0097] Figure 2 This is a flow chart of a method for constructing a neural network prediction model in one embodiment of the present application. In one embodiment, constructing a neural network prediction model using historical data classified by a portrait label system may include the following steps S210 to S230.

[0098] Step S210: pre-process the historical data, and divide the pre-processed historical data into a training set and a test set according to a preset ratio.

[0099] After establishing a portrait labeling system for the relay domain, historical data can be categorized based on this labeling system. Furthermore, missing values ​​and outliers are processed for historical data from user-submitted label classifications, the number of applications is counted, the application probability is calculated, and normalization is performed. Sequences can also be segmented into samples (input-output pairs). In this embodiment, the preset ratio is preferably 8:2, meaning that the sequence samples are divided into a training set and a test set at an 8:2 ratio.

[0100] Step S220: Create a neural network prediction model based on a deep learning framework.

[0101] Relay resource application probability prediction predicts the probability of a relay satellite being applied for in a specific time block, and is therefore essentially a time series prediction. A time series is a chronological sequence of statistical values ​​representing relay service profiles. This time series data exhibits complex nonlinear relationships or long-term dependencies. Time series prediction methods compile and analyze time series data, extrapolating or extending the development process, direction, and trends revealed by the time series to predict the likely performance over the next period or several subsequent time periods.

[0102] LSTM (Long Short-Term Memory) networks excel in time series prediction, particularly for processing complex nonlinear relationships or long-term dependencies within long time series data. LSTM is a special type of RNN (Recurrent Neural Network) that addresses the vanishing and exploding gradient problems found in traditional RNNs by introducing gating mechanisms (forget gate, input gate, and output gate). The forget gate determines which past information is retained or discarded; the input gate determines which information from the current step is retained to update the network's memory or discarded; and the output gate uses historical information stored in the network's memory to process the current element of the sequence. Therefore, in this embodiment, an LSTM model is used to predict the probability of relay resource requests.

[0103] Using a deep learning framework (such as Keras or PyTorch), create a neural network prediction model containing an LSTM layer. The neural network prediction model can include an input layer, hidden layers, and fully connected layers. In one specific embodiment, the input layer can construct training samples using a sliding time window with a step size of 10 minutes and a time window size of 1008 (7 days). After the sequence data is sliced ​​along the time axis, it can enter the hidden layer. The hidden layer can contain multiple LSTM units, each with a forget gate, input gate, and output gate to control the flow and storage of information. The size of the hidden layer is preferably 128. Dropout regularization is applied to each LSTM prediction layer, with a parameter size of 0.2 to prevent overfitting to the training sample set. The fully connected layer can be used for regression prediction and output the final predicted value. All neural network layers use the ReLU function as the activation function, the MAE loss function is set, and the Adam algorithm is used to optimize the network parameters based on the error.

[0104] Step S230: using the training set to train and optimize the neural network prediction model.

[0105] The neural network prediction model is trained using training data, and weights are adjusted through backpropagation and gradient descent to minimize prediction error. At the same time, a sliding window and rolling forecasting technique can be combined. For each new prediction, the sliding window moves forward, updating the input sequence, and using the previous prediction as one of the inputs for the next prediction, continuously "rolling" to perform continuous multi-step predictions.

[0106] Taking into account the autocorrelation of the sequences in the training set and the test set, the true value curve will lag behind the predicted value curve. Therefore, in an embodiment of the present application, a dynamic difference operation can be used to eliminate the autocorrelation of the sequences in the training set and the test set. By taking the difference between the current moment and the previous moment as the regression target, the model is trained using the differenced data, and the model parameters are adjusted by the back propagation algorithm. The optimized model adds an update gate to the gating mechanism. After the current time window unit receives the differential information of the hidden state of the adjacent time window, it passes through the forget gate, input gate and update gate, and is fused with the long-term memory unit of the previous time window to form a differential feature. The output gate is then used to generate the long-term memory unit of the current time window and participate in the information update of the next time window.

[0107] In a preferred embodiment, when tuning the parameters of the LSTM model, based on actual application scenarios, a CNN (Convolutional Neural Network) combined with LSTM can be used to optimize the model's hyperparameters and improve the model's prediction or classification accuracy for specific tasks. This combination automatically finds the optimal model configuration, avoiding the tedious process of manual trial and error.

[0108] In one embodiment, when designing the prediction problem for the probability distribution of relay satellite mission conflicts, it is necessary to consider complex constraints, uncertainties, and priority relationships. The prediction method provided in this application combines the actual application characteristics of relay satellite missions to establish a mathematical model for mission conflicts.

[0109] Considering that some model targets have two user centers responsible for their respective TT&C and data transmission applications, the application probability for these model targets should be the combined application probability of both TT&C and data transmission applications. Furthermore, the weekly task conflict probability distribution prediction model should be adjusted based on the real-time application status of user centers within the weekly timeframe. This involves replacing the probability statistics in the prediction model based on historical data with actual application data and recalculating the weekly task conflict probability to better guide user resource applications.

[0110] A mathematical model for predicting the probability distribution of relay satellite mission conflicts is constructed based on the following constraints. For a single user target, its application to use any relay satellite in a single time block must meet the following conditions:

[0111] (1) Capacity constraints

[0112] Match relay ground station, relay satellite and user targets based on relay field resource capabilities.

[0113] (2) Visibility constraints

[0114] The user target and the requested relay satellite must be geometrically visible during the mission validity period. Strictly speaking, antenna visibility is required, but calculating antenna visibility is complex due to multiple factors, including the user target's attitude, the user antenna's mounting position, its rotation range, and any obstruction of the user antenna by other components. Furthermore, the accurate attitude of the user target for the requested period cannot be obtained during the weekly application phase. The antenna visibility calculated using the current attitude is likely to deviate from the actual situation during the requested period. Therefore, geometric visibility is often used as a substitute in practical applications.

[0115] (3) State constraints

[0116] The application period is within the idle window of the relay resource. The prediction of a single user target only needs to be performed within the intersection of its geometric visibility window and the idle window, and does not need to be calculated for the entire week.

[0117] (4) Duration constraints

[0118] The arc length should not be less than the minimum duration requirement of this task.

[0119] (5) Association constraint check

[0120] The total duration of multiple arcs shall not be less than the shortest task time; the time interval between two arcs shall meet the maximum / minimum task interval; the measurement and control task cannot be after the data transmission task.

[0121] (6) Resource Exclusivity Constraints

[0122] A beam, a link, or a channel of a relay satellite can only be used by one user target at a time.

[0123] (7) Task interval constraints

[0124] The interval between adjacent tasks of the same set of equipment should be greater than the shortest task switching time.

[0125] (8) The time blocks for submitting multiple applications to a user center will not overlap

[0126] If a user center applies for a time block for a certain model target, then its other model targets will not apply for this time block.

[0127] (9) The application period follows the “greedy principle”

[0128] 1) Arc selection rules: Prefer non-reserved arcs (reserved relay resources and outbound arcs). Based on the forward-looking strategy, prioritize arcs with minimal conflicts with other arcs. Based on user preferences, prioritize arcs with user-specified devices and good tracking quality.

[0129] 2) Conflict resolution rules: Eliminate arcs occupied by tasks with low priority; exclude arcs occupied by tasks with more available arcs; and re-enter the queue for scheduling tasks whose arcs have been excluded.

[0130] 3) Emergency Response Rules. For temporary and emergency applications, the Transportation Management Center adopts a priority and first-come, first-served strategy. It prioritizes idle and reserved segments; prioritizes rescheduling adjusted tasks to their original stations and cycles; and limits the depth of adjustments.

[0131] In one embodiment, based on relay domain profile analysis and LSTM application probability prediction results, the predicted application probability p of each target for relay resources in any time block (minimum granularity: minute) of a single relay satellite within the next week is calculated. Then, the weekly application conflict probability is predicted from the user's perspective and the operation management center's perspective. The user perspective can be further subdivided into predictions for a specific user target and an overall prediction for all model targets of a user center.

[0132] In one embodiment, when predicting a single user target, the mathematical model of the conflict probability of a single user target i applying to use a single relay satellite in a single time block of a target time period is:

[0133]

[0134] Where p is the actual application probability of a single user target in a single time block, s is the total number of historical applications of a single user target in a single time block; s max The total number of weeks for historical data request; f t The conflict probability of applying for a single relay satellite to use a single time block in the target time period for a single user target, p i The probability of applying for prediction for another user target in the same time block, n is the number of other user targets except the currently calculated user target, and total is the number of all user targets. i It will have different values ​​at different stages.

[0135] Figure 3 This is a flow chart of a method for predicting the probability of conflict within a target time period from the perspective of a single user target in one of the embodiments of the present application. In one embodiment, when the operation management center releases an idle window for the target time period and all users have not submitted an application for the target time period, predicting the probability of conflict within the target time period from the perspective of a single user target based on the predicted application probability may include the following steps S401 to S403.

[0136] Step S401: Obtain the predicted application probability p of each other user target i applying to use a single relay satellite in a single time block in the target time periodi .

[0137] Step S403: Determine the conflict probability f of a user target i applying for a single time block in a target time period based on the conflict probability mathematical model of a single user target i applying for a single relay satellite in a single time block. t .

[0138] Specifically, if the operation management center has released the idle window for the next week and all users have not submitted weekly applications, the predicted application probability p of each other user target i applying for a certain time block in the next week can be calculated one by one based on the LSTM model. i , and calculate n according to formula (3). Further, calculate the conflict probability f of the current computing target applying for this time block according to formula (2) t .

[0139] Figure 4 The present invention is a flowchart of a method for predicting the probability of conflict within a target time period from the perspective of a single user target in another embodiment of the present application. In one embodiment, when the operation management center releases an idle window for the target time period and all users have not submitted an application for the target time period, predicting the probability of conflict within the target time period from the perspective of a single user target based on the predicted application probability may include the following steps S405 to S413.

[0140] Step S405: cyclically obtain the time blocks applied for by a single user target from the user application queue, and add 1 to the historical application times s of each applied time block.

[0141] Step S407: Calculate the actual application probability p of a single user target for the applied time block according to formula (1).

[0142] Step S409: For other user targets i that have not submitted applications, the predicted application probability p of the applied time blocks is calculated one by one according to the neural network prediction model. i .

[0143] Step S411: Calculate n according to formula (3).

[0144] Step S413: Calculate the conflict probability f of the time block that the user's target application has applied for according to formula (2) t .

[0145] Specifically, for users who have submitted weekly applications, the algorithm flow is as follows: a time block that a target has applied for is taken out from the application queue, the number of historical applications s for this time block is increased by 1, and the actual application probability p of this target for this time block is calculated according to formula (1). For other user targets i that have not submitted applications, the predicted application probability p of this time block is calculated one by one according to the LSTM model.i Calculate n according to formula (3), and further calculate the conflict probability f of the current computing target applying for this time block according to formula (2) t .

[0146] Figure 5 This is a flow chart of a method for predicting the probability of conflict within a target time period from the perspective of a single user target in another embodiment of the present application. In one embodiment, after the operation management center completes the plan for the target time period, the operation management center adopts a priority and first-come-first-served strategy for emergency insertion for temporary applications and emergency applications submitted by users based on preset constraints. That is, after the weekly plan of the operation center is completed, the operation management center can adopt a priority and first-come-first-served strategy for temporary applications and emergency applications submitted by users based on the constraint (9) as described in the above embodiment, and preferably select idle arcs and reserved arcs for emergency insertion. Predicting the probability of conflict within a target time period from the perspective of a single user target based on the predicted application probability can include the following steps S415 to S425.

[0147] Step S415: cyclically obtain the time blocks applied for by a single user target from the user application queue, and add 1 to the historical application times s of each applied time block.

[0148] Step S417: Calculate the actual application probability p of a single user target for the applied time block according to formula (1).

[0149] Step S419: cyclically extract the emergency time blocks applied for by a single user from the emergency application queue, and add 1 to the historical application times s of each emergency time block.

[0150] Step S421: recalculate the actual application probability p of a single user target for the emergency time block according to formula (1).

[0151] Step S423: Calculate n according to formula (3).

[0152] Step S425: Calculate the conflict probability f of the user's target emergency time block according to formula (2) t .

[0153] Specifically, a target is looped out of the application queue for the next week's time blocks. The historical application count s for each time block is incremented by 1, and the target's actual application probability for this time block is calculated according to formula (1). A target is looped out of the emergency application queue for the next week's time blocks, the historical application count s for each time block is incremented by 1, and the target's actual application probability p for this time block is recalculated according to formula (1). Calculate n according to formula (3), and calculate the conflict probability f of the current target applying for this time block according to formula (2). t .

[0154] In one embodiment, for the overall prediction of a single user center, a mathematical model of conflict probability constructed for a single user center applying to use a single relay satellite in a single time block of a target time period is:

[0155]

[0156] Where, f c The conflict probability of a single user center applying to use a single relay satellite in a single time block during the target time period; p j The probability of applying for a single user target j belonging to other user centers to use the same relay satellite in the same time block; m is the number of all user targets belonging to other user centers; Total is the number of all user targets, and Sum0 is the number of user targets in the current user center. j It can have different values ​​at different stages.

[0157] Figure 6 This is a flow chart of a method for predicting the probability of conflict within a target time period from the perspective of a single user center in one of the embodiments of the present application. In one of the embodiments, when the operation management center releases an idle window for the target time period and all users have not submitted an application for the target time period, predicting the probability of conflict within the target time period from the perspective of a single user center based on the predicted application probability may include the following steps S427 to S433.

[0158] Step S427: Obtain the predicted application probability p of other user targets j who are not the current user center and apply to use a single relay satellite in a single time block in the target time period j .

[0159] Step S429: Determine the conflict probability f of a single user center applying to use a single relay satellite in a single time block of the target time period based on the conflict probability mathematical model of a single user center applying to use a single relay satellite in a single time block of the target time period. c .

[0160] Step S431: cyclically calculating the conflict probability of the current user center applying to use a single relay satellite in all time blocks of the target time period.

[0161] Step S433: cyclically calculating the conflict probability of the current user center applying to use all relay satellites in all time blocks of the target time period.

[0162] Specifically, when the operation control center releases the idle window for the next week and all users have not submitted weekly applications, the predicted application probability p of other users who are not members of the current user center applying for a certain time block can be calculated one by one according to the LSTM model. jCalculate m according to formula (5), and calculate the conflict probability f of the current user center applying for a certain relay satellite in a certain time block according to formula (4) c Then, the conflict probability of the current user center applying for a certain relay satellite in all time blocks in the next week is cyclically calculated, and the conflict probability of the current user center applying for all relay satellites in all time blocks in the next week is cyclically calculated.

[0163] Figure 7 A flow chart of a method for predicting the probability of conflict within a target time period from the perspective of a single user center is provided for another embodiment of the present application. In one embodiment, when the operation management center releases an idle window for the target time period and a user submits an application for the target time period, predicting the probability of conflict within the target time period from the perspective of a single user center based on the predicted application probability may include the following steps S435 to S447.

[0164] Step S435: cyclically taking out the time blocks applied for by a single user target from the application queue, and adding 1 to the historical application times s of the applied time blocks.

[0165] Step S437: Calculate the actual application probability p of a single user target for the applied time block according to formula (1).

[0166] Step S439: For other user targets j that have not submitted applications, the predicted application probability p of the applied time blocks is calculated one by one according to the neural network prediction model. j .

[0167] Step S441: Calculate m according to formula (5).

[0168] Step S443: Calculate the conflict probability f of the current user center applying to use a single relay satellite in a single time block according to formula (4): c .

[0169] Step S445: cyclically calculating the conflict probability of the current user center applying to use a single relay satellite in all time blocks of the target time period.

[0170] Step S447: cyclically calculating the conflict probability of the current user center applying to use all relay satellites in all time blocks of the target time period.

[0171] Specifically, if a user center has submitted a weekly application, the algorithm process can be to cyclically take out a time block that a target has applied for from the application queue, add 1 to the historical application number s of this time block, and calculate the actual application probability p of this target for this time block according to formula (1). For other user targets that have not submitted applications, the predicted application probability p of this time block being applied for is calculated one by one according to the LSTM model. jAccording to formula (5), calculate the conflict probability f when the current user center applies for a certain relay satellite in a certain time block according to formula (4): c Then, the conflict probability of the current user center applying for a certain relay satellite in all time blocks in the next week is cyclically calculated, and the conflict probability of the current user center applying for all relay satellites in all time blocks in the next week is cyclically calculated.

[0172] Based on the above prediction results, the user center can make a probabilistic prediction of the user center's weekly resource application. It should be noted that the user center's prediction can also fully consider mission scenarios, such as manned flight missions, long-term control periods, major exercises, and major mission support periods.

[0173] In one embodiment, for the overall prediction of the operation control center, the probability mathematical model constructed for all user target applications to use a single relay satellite a in the same time block of the target time period is:

[0174]

[0175] Where p k The predicted application probability of user target k applying to use a single relay satellite a in the same time block, z is the total number of user targets.

[0176] Figure 8 This is a flow chart of a method for predicting the probability of conflict within a target time period from the perspective of an operation management center in one embodiment of the present application. In one embodiment, predicting the probability of conflict within a target time period from the perspective of an operation management center based on the predicted application probability may include the following steps S449 to S455.

[0177] Step S449: Calculate the predicted application probability p of a single user target k applying for a single time block in the target time period one by one according to the neural network prediction model k .

[0178] Step S451: Calculate the probability r of a single relay satellite a being applied for in the same time block of the target time period based on the probability mathematical model of all users applying to use a single relay satellite a in the same time block of the target time period. a .

[0179] Step S453: cyclically calculating the probability of a single relay satellite a being applied for in all time blocks of the target time period.

[0180] Step S455: cyclically calculating the probability of all relay satellites being applied for in all time blocks in the target time period.

[0181] Specifically, the predicted application probability p for any time block in the next week is calculated one by one based on the LSTM model.k The probability of a relay satellite a being applied for the same time block in the next week is calculated according to formula (6). Furthermore, the probability of a relay satellite being applied for all time blocks in the next week is cyclically calculated, and the probability of all relay satellites being applied for all time blocks in the next week is cyclically calculated.

[0182] The operation and management center can use the above prediction results to predict the application probability distribution of all relay satellites in all time blocks in the next week, so that the resources that need to be reserved can be arranged as much as possible in the period with low user application probability, leaving more available resources for users.

[0183] In one embodiment, different prediction angles may have different calculation timings.

[0184] The calculation timing rules for single user target prediction and single user center prediction can be:

[0185] 1) Automatic calculation after the operation control center releases the idle window for the next week for the first time;

[0186] 2) Automatically calculate after detecting a new user weekly application and determining that the current round of user applications has ended;

[0187] 3) Before the operations control center releases the next week's plan to users, each change in the idle window automatically triggers calculation.

[0188] The calculation timing rules for the overall forecast of the operation control center can be: any time this week before reserving next week's resources, automatic calculation can be initiated regularly, or calculation can be manually initiated by the operator.

[0189] The prediction method for the relay satellite mission conflict probability distribution provided in this application can be based on the relay satellite resource planning historical data, user historical data, and relay satellite / user target-related orbital information, and can analyze the user's habitual characteristics of resource use by constructing a relay domain portrait. Based on the domain portrait, an LSTM model is established to predict the application probability of relay resources in the next week, and further calculates the task conflict probability in the next week based on the application probability, thereby realizing the prediction of resource usage within the weekly period. The LSTM model is based on deep learning theory and effectively handles the complex nonlinearity and long-term dependencies of long-term series data through a gating mechanism. Combined with sliding window and rolling prediction technology, the resource application probability predicted by the LSTM model can achieve high accuracy and practicality.

[0190] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts of the accompanying drawings may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times. The order of execution of these steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a portion of the steps or stages in other steps.

[0191] Based on the description of the embodiment of the method for predicting the probability distribution of relay satellite mission conflicts, the present disclosure also provides a system for predicting the probability distribution of relay satellite mission conflicts. The system may include a system (including a distributed system), software (application), modules, components, servers, clients, etc. using the method described in the embodiments of this specification and combined with necessary implementation hardware. Based on the same innovative concept, the system in one or more embodiments provided by the embodiments of the present disclosure is as described in the following embodiments. Since the implementation scheme and method for solving the problem of the system are similar, the implementation of the specific system of the embodiments of this specification can refer to the implementation of the aforementioned method, and the repeated parts will not be repeated. As used below, the term "unit" or "module" can be a combination of software and / or hardware that implements the predetermined function. Although the device described in the following embodiments is preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and conceivable.

[0192] This application also provides a prediction system for the probability distribution of relay satellite mission conflicts. Figure 9 This is a structural diagram of a system for predicting the probability distribution of relay satellite mission conflicts in one embodiment of the present application. The system for predicting the probability distribution of relay satellite mission conflicts may include a data processing module 100, a model building module 200 and a prediction analysis module 300.

[0193] The data processing module 100 can be used to construct a portrait label system in the relay field based on historical data.

[0194] The model building module 200 can be used to build a neural network prediction model using historical data classified by the portrait label system.

[0195] The prediction analysis module 300 can be used to determine the predicted application probability of relay resources within the target time period based on the portrait label system and the neural network prediction model. It can also be used to predict the conflict probability within the target time period from the user's perspective and the operation management center's perspective based on the predicted application probability.

[0196] Based on relay domain profiling and predictive analysis using a neural network prediction model, we can determine the probability of each user target applying for any time block on a single relay satellite within the next week. Using a probability distribution prediction model, we can predict the conflict probability of weekly applications from a single user target, a single user center, or an operations management center for all time blocks across multiple relay satellites. This assists users in submitting weekly applications, guiding them to use time blocks with minimal conflict probability, thus avoiding potential conflicts in advance.

[0197] It should be understood that the various embodiments of the above-mentioned methods, devices, etc. in this specification are described in a progressive manner. The same / similar parts between the various embodiments can be referred to in detail. Each embodiment focuses on the differences from other embodiments. For related parts, refer to the descriptions of other method embodiments.

[0198] In an exemplary embodiment, a computer-readable storage medium including instructions is further provided, such as a memory S22 including instructions. The instructions may be executed by a processor of the relay satellite mission conflict probability distribution prediction system to perform the above method. The storage medium may be a computer-readable storage medium, such as a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, or optical data storage device.

[0199] In an exemplary embodiment, a computer program product is further provided. The computer program product includes instructions, and the instructions can be executed by a processor of a system for predicting relay satellite mission conflict probability distribution to implement the above method.

[0200] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), data processing logic devices based on quantum computing, and the like.

[0201] The various embodiments in this specification are described in a progressive manner. Similar portions between the various embodiments can be referenced to each other. Each embodiment focuses on the differences between the other embodiments. In particular, the hardware + program embodiments are described briefly because they are generally similar to the method embodiments. For relevant portions, refer to the description of the method embodiments.

[0202] It should be noted that the aforementioned systems, electronic devices, servers, etc., according to the description of the method embodiments, may also include other implementation methods. For specific implementation methods, reference can be made to the description of the relevant method embodiments. Furthermore, new embodiments formed by combining features of various method, system, device, and server embodiments remain within the scope of implementation of this disclosure and are not detailed here.

[0203] Throughout this specification, references to terms such as "some embodiments," "other embodiments," and "desired embodiments" indicate that a particular feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. Although these terms are used interchangeably throughout this specification, they do not necessarily refer to the same embodiment or example.

[0204] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0205] The above-described embodiments merely represent several implementations of the present invention. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent. It should be noted that a person skilled in the art would be able to make various modifications and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the patent for this invention shall be determined by the appended claims.

Claims

1. A method for predicting the probability distribution of relay satellite mission conflicts, characterized in that: include: Build a portrait label system in the relay field based on historical data; Using the historical data classified by the portrait label system, a neural network prediction model is constructed; Determine the predicted probability of each user applying for relay resources within a target time period based on the portrait label system and the neural network prediction model; Predicting the conflict probability within the target time period from the user's perspective and the operation management center's perspective based on the predicted application probability; The portrait label system includes user portrait, resource portrait, and business portrait. The user portrait includes user center portrait and user target portrait. The resource portrait includes relay satellite portrait and relay ground station portrait. When the operation management center publishes an idle window for the target time period and no user has submitted an application for the target time period, the conflict probability prediction within the target time period is performed from the perspective of a single user target based on the predicted application probability, including: Get the predicted application probability p of each other user target i applying to use a single relay satellite in a single time block during the target time period i ; According to the mathematical model of the conflict probability of a single user target i applying for a single time block in the target time period to use a single relay satellite, the conflict probability f of the user target applying for a single time block is determined. t ; The mathematical model of the conflict probability of a single user target i applying to use a single relay satellite in a single time block of a target time period is: Where p is the actual application probability of a single user target in a single time block, s is the total number of historical applications of a single user target in a single time block; s max The total number of weeks for historical data request; f t The conflict probability of applying for a single relay satellite to use a single time block in the target time period for a single user target, p i The probability of applying for prediction for another user target in the same time block, n is the number of other user targets except the currently calculated user target, and total is the number of all user targets.

2. The method for predicting relay satellite mission conflict probability distribution according to claim 1, wherein: The historical data classified by the portrait label system is used to construct a neural network prediction model, including: Preprocessing the historical data, and dividing the preprocessed historical data into a training set and a test set according to a preset ratio; Create neural network prediction models based on deep learning frameworks; The training set is used to train and optimize the neural network prediction model.

3. The method for predicting relay satellite mission conflict probability distribution according to claim 1, wherein: When the operation management center publishes an idle window for the target time period and a user submits an application for the target time period, conflict probability prediction within the target time period is performed from the perspective of a single user target based on the predicted application probability, including: Looping through the user application queue to obtain the time blocks that a single user target has applied for, and adding 1 to the historical application count s of each of the time blocks that have been applied for; Calculate the actual application probability p of the single user target for the applied time block according to formula (1); For other user targets i that have not submitted applications, the predicted application probability p of the applied time block is calculated one by one according to the neural network prediction model. i ; Calculate n according to formula (3); According to formula (2), the conflict probability f of the user's target application for the applied time block is calculated t .

4. The method for predicting relay satellite mission conflict probability distribution according to claim 1, wherein: After the operation management center completes the plan for the target time period, it uses a priority and first-come-first-served strategy to perform emergency insertion for users' temporary and emergency applications based on preset constraints. The conflict probability prediction within the target time period is performed from the perspective of a single user's goal based on the predicted application probability, including: Looping through the user application queue to obtain the time blocks that a single user target has applied for, and adding 1 to the historical application count s of each of the time blocks that have been applied for; Calculate the actual application probability p of the single user target for the applied time block according to formula (1); Circularly take out the emergency time blocks applied for by a single user from the emergency application queue, and add 1 to the historical application times s of each emergency time block; Recalculate the actual application probability p of the single user target for the emergency time block according to formula (1); Calculate n according to formula (3); According to formula (2), the conflict probability f of the user's target application for the emergency time block is calculated t .

5. The method for predicting relay satellite mission conflict probability distribution according to claim 1, wherein: When the operation management center publishes an idle window for the target time period and no user has submitted an application for the target time period, the conflict probability prediction for the target time period is performed from the perspective of a single user center based on the predicted application probability, including: Get the predicted application probability p of other user targets j who are not the current user center and apply to use a single relay satellite in a single time block during the target time period j ; According to the mathematical model of the conflict probability of a single user center applying to use a single relay satellite in a single time block during the target time period, the conflict probability f of a single user center applying to use a single relay satellite in a single time block during the target time period is determined. c ; The mathematical model of the conflict probability of a single user center applying to use a single relay satellite in a single time block of a target time period is: Where, f c The conflict probability of a single user center applying to use a single relay satellite in a single time block of the target time period; p j The probability of applying for a single user target j belonging to other user centers to use the same relay satellite in the same time block; m is the number of all user targets belonging to other user centers; Total is the number of all user targets, and Sum0 is the number of user targets in the current user center; Circularly calculate the conflict probability of the current user center applying to use a single relay satellite in all time blocks of the target time period; Circularly calculate the conflict probability of the current user center applying to use all relay satellites in all time blocks of the target time period.

6. The method for predicting relay satellite mission conflict probability distribution according to claim 5, characterized in that: When the operation management center publishes an idle window for the target time period and a user submits an application for the target time period, the conflict probability prediction within the target time period is performed from the perspective of a single user center based on the predicted application probability, including: Circularly take out the time blocks applied for by a single user target from the application queue, and add 1 to the historical application count s of the applied time blocks; Calculate the actual application probability p of the single user target for the applied time block according to formula (1); For other user targets j that have not submitted applications, the predicted application probability p of the applied time block is calculated one by one according to the neural network prediction model. j ; Calculate m according to formula (5); According to formula (4), the conflict probability f of the current user center applying to use a single relay satellite in a single time block is calculated as follows: c ; Circularly calculate the conflict probability of the current user center applying to use a single relay satellite in all time blocks of the target time period; Circularly calculate the conflict probability of the current user center applying to use all relay satellites in all time blocks of the target time period.

7. The method for predicting relay satellite mission conflict probability distribution according to claim 5, characterized in that: Predicting the conflict probability within the target time period from the perspective of the operation management center based on the predicted application probability includes: According to the neural network prediction model, the predicted application probability p of a single user target k applying for a single time block in the target time period is calculated one by one k ; According to the probability mathematical model of all users applying to use a single relay satellite a in the same time block of the target time period, the probability r of a single relay satellite a being applied in the same time block of the target time period is calculated. a ; The probability mathematical model of all user target applications using a single relay satellite a in the same time block of the target time period is: Where p k The predicted application probability for user target k to use a single relay satellite a in the same time block, z is the total number of user targets; Circularly calculate the probability of a single relay satellite a being applied for in all time blocks of the target time period; The probability of all relay satellites being applied for in all time blocks of the target time period is cyclically calculated.

8. A prediction system for relay satellite mission conflict probability distribution, characterized in that: include: Data processing module, used to build a portrait label system in the relay field based on historical data; A model building module, for building a neural network prediction model using the historical data classified by the portrait label system; A prediction analysis module, configured to determine the predicted application probability of relay resources for each user within a target time period based on the portrait label system and the neural network prediction model, and to predict the conflict probability within the target time period from the perspective of the user and the perspective of the operation management center based on the predicted application probability; The portrait label system includes user portrait, resource portrait, and business portrait. The user portrait includes user center portrait and user target portrait. The resource portrait includes relay satellite portrait and relay ground station portrait. When the operation management center releases the idle window of the target time period and all users have not submitted applications for the target time period, the prediction analysis module is further used to: Get the predicted application probability p of each other user target i applying to use a single relay satellite in a single time block during the target time period i ; According to the mathematical model of the conflict probability of a single user target i applying for a single time block in the target time period to use a single relay satellite, the conflict probability f of the user target applying for a single time block is determined. t ; The mathematical model of the conflict probability of a single user target i applying to use a single relay satellite in a single time block of a target time period is: Where p is the actual application probability of a single user target in a single time block, s is the total number of historical applications of a single user target in a single time block; s max The total number of weeks for historical data request; f t The conflict probability of applying for a single relay satellite to use a single time block in the target time period for a single user target, p i The probability of applying for prediction for another user target in the same time block, n is the number of other user targets except the currently calculated user target, and total is the number of all user targets.

Images