Methods, devices, equipment and storage media for predicting traffic patterns in spatial information networks

Abstract

This application discloses a method, apparatus, device, and storage medium for predicting traffic patterns in spatial information networks, relating to the field of big data analytics. The method involves acquiring traffic data from a spatial information network over multiple historical time periods. First, a probability distribution function is used to simulate the traffic characteristic maps and application performance indicators of each sub-network layer in each time period. Then, these are combined to obtain the traffic characteristic maps and end-to-end application performance indicators of the spatial information network in each time period. Next, the time series of traffic characteristic data and label data obtained from data processing are imported into a GCN+BiGRU model for training, resulting in a traffic characteristic data prediction model. Finally, the current time series of traffic characteristic data is input into the traffic characteristic data prediction model, outputting the traffic characteristic data and traffic pattern for the next time period. This improves the accuracy of spatial information network traffic prediction and enhances the ability to perceive traffic patterns.

Description

Technical Field

[0001] This invention belongs to the field of big data analysis technology, specifically relating to a method, device, equipment, and storage medium for predicting the traffic situation of spatial information networks. Background Technology

[0002] Space information networks are network systems that use space platforms (such as geostationary satellites or medium and low Earth orbit satellites, stratospheric balloons, and manned or unmanned aircraft) as carriers to acquire, transmit, and process space information in real time. With the rapid development of space information networks, their services are becoming more diverse, and their states are becoming more complex. Space information networks are characterized by heterogeneity, dynamism, and complexity; the form and state of their traffic development, i.e., traffic patterns, will vary with changes in network conditions. Existing technologies lack the ability to fuse and process multi-service data from space information networks, and machine learning models such as SVM and random forests suffer from inaccurate prediction accuracy, poor predictive ability of multi-source traffic patterns, insufficient generalization ability of prediction models, and an inability to accurately predict the changing trends of traffic patterns for different services based on varying service requirements. Summary of the Invention

[0003] The purpose of this invention is to provide a method, apparatus, computer equipment, and computer-readable storage medium for predicting the traffic situation of spatial information networks, so as to improve the accuracy of traffic prediction in spatial information networks and enhance the ability to perceive traffic situation.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] Firstly, a method for predicting the traffic situation of spatial information networks is provided, including:

[0006] Acquire traffic data of the spatial information network over M historical time periods, where M represents a positive integer and the M historical time periods are consecutive in time.

[0007] For each historical period in the M historical periods, based on the corresponding traffic data, the traffic characteristic map and application performance index of each sub-network in the spatial information network in the corresponding period are first obtained by using the probability distribution function. Then, based on the traffic characteristic map and application performance index of each sub-network in the corresponding period, the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period are combined to obtain the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period.

[0008] The spatial information network is preprocessed with traffic characteristic maps and end-to-end application performance indicators for the M historical time periods to obtain M traffic characteristic data corresponding one-to-one with the M historical time periods.

[0009] N sets of traffic characteristic data are extracted from the M sets of traffic characteristic data, each corresponding to one of the last N historical periods in the M historical periods. For each set of traffic characteristic data in the N sets of traffic characteristic data, K sets of traffic characteristic data are extracted from the M sets of traffic characteristic data, each corresponding to one of the K most recent historical periods before the corresponding period in the M historical periods, to obtain the corresponding traffic characteristic data time series. Here, N and K represent positive integers and N+K is less than or equal to M.

[0010] N time series of traffic feature data are used as N sample data, and N sets of traffic feature data are used as N label data corresponding one-to-one with the N sample data. These are then imported into a GCN+BiGRU model for model training to obtain a traffic feature data prediction model. The GCN+BiGRU model includes a graph convolutional neural network (GCN), a bidirectional gated recurrent unit (BiGRU), and a fully connected layer. The GCN is used to model the spatial characteristics of the spatial information network traffic after inputting the N sample data, and extract the spatial features of the spatial information network traffic. The Bidirectional gated recurrent unit (BiGRU) is used to model the temporal characteristics of the spatial information network traffic after inputting the N sample data, and extract the temporal features of the spatial information network traffic. The fully connected layer is used to process the spatial and temporal features of the spatial information network traffic to obtain the final prediction result.

[0011] Obtain the traffic data of the spatial information network for the K most recent historical time periods;

[0012] For each historical period in the current K most recent historical periods, based on the corresponding traffic data, the probability distribution function is first used to simulate and obtain the traffic characteristic map and application performance index of each sub-network in the spatial information network in the corresponding period. Then, based on the traffic characteristic map and application performance index of each sub-network in the corresponding period, the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period are combined to obtain the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period.

[0013] The spatial information network is preprocessed with the traffic characteristic maps and end-to-end application performance indicators of the current K most recent historical periods to obtain the current traffic characteristic data time series.

[0014] The current traffic characteristic data time series is input into the traffic characteristic data prediction model, and the traffic characteristic data for the next time period is output.

[0015] Based on the traffic characteristic data of the current next time period, the traffic situation of the spatial information network in the current next time period is obtained.

[0016] Based on the above-mentioned invention, a training and application scheme for a spatial information network traffic situation prediction model based on hierarchical modeling and GCN+BiGRU model is provided. Specifically, after acquiring traffic data of the spatial information network over multiple historical time periods, the traffic characteristic maps and application performance indicators of each sub-network layer in each time period are first simulated using a probability distribution function. Then, these are combined to obtain the traffic characteristic maps and end-to-end application performance indicators of the spatial information network in each time period. Next, the time series of traffic characteristic data and label data obtained from data processing are imported into the GCN+BiGRU model for model training, resulting in a traffic characteristic data prediction model. Finally, the current time series of traffic characteristic data is input into the traffic characteristic data prediction model, and the output is the traffic characteristic data and traffic situation for the next time period. This improves the accuracy of spatial information network traffic prediction, enhances the ability to perceive traffic situation, and facilitates practical application and promotion.

[0017] In one possible design, the traffic feature map and application performance indicators include service type, service data size, service priority, link bandwidth, service characteristics under gateway load rate, transmission delay and / or packet loss rate.

[0018] In one possible design, when the traffic characteristic map and application performance indicators include transmission delay, for a certain historical time period among the M historical time periods, based on the corresponding traffic data, the traffic characteristic map and application performance indicators of each sub-network layer in the spatial information network in the corresponding time period are first simulated using a probability distribution function. Then, based on the traffic characteristic map and application performance indicators of each sub-network layer in the corresponding time period, the traffic characteristic map and end-to-end application performance indicators of the spatial information network in the corresponding time period are composited, including:

[0019] Based on the traffic data of a certain historical period, the transmission delay of each sub-network in the spatial information network is obtained by integral of the offset gamma probability density function during the certain historical period. The traffic data packet of the certain historical period contains the server response time collected during the certain historical period and the service transmission time after the service connection is established between the two ends of the communication.

[0020] Based on the transmission delays of each sub-network layer during a certain historical period, the end-to-end transmission delay of the spatial information network during that historical period is obtained by combining the following formulas. :

[0021]

[0022] In the formula, This indicates the server response time collected during a specific historical period. This indicates the number of subnetworks in the spatial information network. Indicates less than or equal to positive integers, In the spatial information network, the first The transmission delay of each sub-network during a certain historical period.

[0023] In one possible design, based on the traffic data for a certain historical period, the transmission delay of each sub-network layer in the spatial information network during that historical period is obtained by integrating the offset gamma probability density function, including:

[0024] Based on the traffic data of a certain historical period, the transmission delay of each sub-network in the spatial information network in the certain historical period is obtained by integral of the offset gamma probability density function based on a preset confidence level. The traffic data packet of the certain historical period contains the server response time and the service transmission time after the service connection is established at both ends of the communication during the certain historical period.

[0025] In one possible design, when the traffic characteristic map and application performance indicators include packet loss rate, for a certain historical period among the M historical periods, based on the corresponding traffic data, the traffic characteristic map and application performance indicators of each sub-network layer in the spatial information network at the corresponding time period are first simulated using a probability distribution function. Then, based on the traffic characteristic map and application performance indicators of each sub-network layer at the corresponding time period, the traffic characteristic map and end-to-end application performance indicators of the spatial information network at the corresponding time period are composited, including:

[0026] Based on the traffic data for a certain historical period, the packet loss rate of each sub-network in the spatial information network during that historical period is obtained using the normal distribution probability density function.

[0027] Based on the packet loss rates of each sub-network layer during a certain historical period, the end-to-end packet loss rate of the spatial information network during that historical period is obtained by combining the results using the following formula. :

[0028]

[0029] In the formula, This indicates the number of subnetworks in the spatial information network. Indicates less than or equal to positive integers, In the spatial information network, the first The packet loss rate of each subnetwork during a certain historical period.

[0030] In one possible design, the data preprocessing includes data cleaning and standardization.

[0031] In one possible design, the sub-networks of the space information network include a satellite network, a terrestrial fiber optic network, and a wireless ad hoc network.

[0032] Secondly, a spatial information network traffic situation prediction device is provided, including a data acquisition module, a hierarchical composite module, a data processing module, a data extraction module, a model training module, a traffic feature prediction module, and a traffic situation determination module.

[0033] The data acquisition module is used to acquire traffic data of the spatial information network in M ​​historical time periods, where M represents a positive integer and the M historical time periods are consecutive in time.

[0034] The layered composite module is connected to the data acquisition module and is used to, for each historical period in the M historical periods, first use the probability distribution function to simulate the traffic characteristic map and application performance index of each sub-network in the spatial information network in the corresponding period according to the corresponding traffic data, and then, based on the traffic characteristic map and application performance index of each sub-network in the corresponding period, composite the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period.

[0035] The data processing module is communicatively connected to the hierarchical composite module and is used to preprocess the traffic characteristic maps and end-to-end application performance indicators of the spatial information network in the M historical time periods to obtain M traffic characteristic data corresponding one-to-one with the M historical time periods.

[0036] The data extraction module is communicatively connected to the data processing module. It is used to extract N sets of traffic feature data from the M sets of traffic feature data, which correspond one-to-one with the last N historical periods in the M historical periods. For each set of traffic feature data in the N sets of traffic feature data, it extracts K sets of traffic feature data from the M sets of traffic feature data, which correspond one-to-one with the K most recent historical periods before the corresponding period in the M historical periods, to obtain the corresponding traffic feature data time series. Here, N and K represent positive integers and N+K is less than or equal to M.

[0037] The model training module, communicatively connected to the data extraction module, is used to input N time series of traffic feature data as N sample data and N sets of traffic feature data as N label data corresponding one-to-one with the N sample data into the GCN+BiGRU model for model training to obtain a traffic feature data prediction model. The GCN+BiGRU model includes a Graph Convolutional Neural Network (GCN), a Bidirectional Gated Recurrent Unit (BiGRU), and a fully connected layer. The GCN is used to model the spatial characteristics of the spatial information network traffic after inputting the N sample data, extracting the spatial features of the spatial information network traffic. The Bidirectional Gated Recurrent Unit (BiGRU) is used to model the temporal characteristics of the spatial information network traffic after inputting the N sample data, extracting the temporal features of the spatial information network traffic. The fully connected layer processes the spatial and temporal features of the spatial information network traffic to obtain the final prediction result.

[0038] The data acquisition module is also used to acquire the traffic data of the spatial information network in the most recent K historical time periods;

[0039] The layered composite module is further configured to, for each historical period in the current K most recent historical periods, first use the probability distribution function to simulate and obtain the traffic characteristic map and application performance index of each sub-network in the spatial information network in the corresponding period based on the corresponding traffic data, and then, based on the traffic characteristic map and application performance index of each sub-network in the corresponding period, composite the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period.

[0040] The data processing module is also used to perform the data preprocessing on the traffic characteristic map and end-to-end application performance index of the spatial information network in the current K most recent historical periods to obtain the current traffic characteristic data time series;

[0041] The traffic feature prediction module is communicatively connected to the data processing module and the model training module, respectively, and is used to input the current traffic feature data time series into the traffic feature data prediction model and output the traffic feature data for the next time period.

[0042] The traffic situation determination module is communicatively connected to the traffic feature prediction module and is used to obtain the traffic situation of the spatial information network in the current next time period based on the traffic feature data of the current next time period.

[0043] Thirdly, the present invention provides a computer device comprising a memory, a processor, and a transceiver connected in sequence for communication, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the spatial information network traffic situation prediction method as described in the first aspect or any possible design in the first aspect.

[0044] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, perform the spatial information network traffic situation prediction method as described in the first aspect or any possible design within the first aspect.

[0045] Fifthly, the present invention provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform the spatial information network traffic situation prediction method as described in the first aspect or any possible design in the first aspect.

[0046] The beneficial effects of the above scheme are:

[0047] (1) This invention creatively provides a training and application scheme for a spatial information network traffic situation prediction model based on hierarchical modeling and GCN+BiGRU model. That is, after obtaining traffic data of spatial information network in multiple historical periods, the traffic feature map and application performance index of each sub-network in each period are first simulated using probability distribution function. Then, the traffic feature map and end-to-end application performance index of spatial information network in each period are obtained by combining them. Then, the traffic feature data time series and label data obtained by data processing are imported into GCN+BiGRU model for model training to obtain traffic feature data prediction model. Finally, the current traffic feature data time series is input into the traffic feature data prediction model to output the traffic feature data and traffic situation of the next period. In this way, the accuracy of spatial information network traffic prediction can be improved and the ability to perceive traffic situation can be enhanced.

[0048] (2) It can also solve the problem of multi-service data fusion processing of spatial information networks, improve the prediction ability of multi-source traffic situation, and enhance the generalization ability of the model, which is convenient for practical application and promotion. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart illustrating the spatial information network traffic situation prediction method provided in an embodiment of this application.

[0051] Figure 2 This is a schematic diagram of the spatial information network traffic situation prediction device provided in the embodiments of this application.

[0052] Figure 3 A schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0053] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0054] It should be understood that although the terms "first" and "second", etc., may be used herein to describe various objects, these objects should not be limited by these terms. These terms are only used to distinguish one object from another. For example, the first object may be referred to as the second object, and similarly, the second object may be referred to as the first object, without departing from the scope of the exemplary embodiments of the invention.

[0055] It should be understood that the term "and / or" that may appear in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, or A and B exist simultaneously. Another example is A, B and / or C, which can mean that any one of A, B, and C or any combination thereof exists. The term " / and" that may appear in this document describes another relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone or A and B exist simultaneously. In addition, the character " / " that may appear in this document generally indicates that the related objects before and after it are in an "or" relationship.

[0056] Example:

[0057] like Figure 1As shown, the spatial information network traffic situation prediction method provided in the first aspect of this embodiment can be executed, but is not limited to, by computer devices with certain computing resources, such as platform servers, personal computers (PCs, referring to a type of multi-purpose computer suitable for personal use in terms of size, price, and performance; desktop computers, laptops, mini-laptops, tablets, and ultrabooks are all considered personal computers), smartphones, personal digital assistants (PDAs), or wearable devices. Figure 1 As shown, the spatial information network traffic situation prediction method may include, but is not limited to, the following steps S1 to S10.

[0058] S1. Obtain traffic data of the spatial information network in M ​​historical time periods, where M represents a positive integer and the M historical time periods are consecutive in time sequence.

[0059] S2. For each historical period in the M historical periods, based on the corresponding traffic data, first use the probability distribution function to simulate and obtain the traffic characteristic map and application performance index of each sub-network layer in the spatial information network in the corresponding period. Then, based on the traffic characteristic map and application performance index of each sub-network layer in the corresponding period, combine them to obtain the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period.

[0060] In step S2, considering the significant differences and fluctuations in traffic distribution and application performance of each sub-network layer due to varying network states, channel environments, and end-to-end application delivery performance, a probability distribution function is needed to simulate the traffic pattern and application performance fluctuations of each sub-network layer. Specifically, the traffic characteristic map and application performance indicators include, but are not limited to, service type, service data size, service priority, link bandwidth, service characteristics under gateway load rate, transmission delay, and / or packet loss rate. Furthermore, the sub-networks in the spatial information network include, but are not limited to, satellite networks, terrestrial fiber optic networks, and wireless ad hoc networks.

[0061] In step S2, specifically, when the traffic feature map and application performance indicators include transmission delay, for a certain historical period among the M historical periods, based on the corresponding traffic data, the traffic feature map and application performance indicators of each sub-network layer in the spatial information network in the corresponding period are first simulated using a probability distribution function. Then, based on the traffic feature map and application performance indicators of each sub-network layer in the corresponding period, the traffic feature map and end-to-end application performance indicators of the spatial information network in the corresponding period are composited, including but not limited to the following steps S211 to S212.

[0062] S211. Based on the traffic data of a certain historical period, the transmission delay of each sub-network in the spatial information network in the certain historical period is obtained by integrating the offset gamma probability density function. The traffic data of the certain historical period includes, but is not limited to, the server response time and the service transmission time after the service connection is established at both ends of the communication during the certain historical period.

[0063] In step S211, the traffic situation of the sub-network and the application transmission latency are considered to be represented by the server response time and the sub-network latency, as shown in the following equation:

[0064]

[0065] In the formula, Indicates transmission delay. Indicates the server's response time. This represents the transmission time of services after the connection is established between the two ends of the communication. Furthermore, since the offset gamma probability distribution function has typical heavy-tailed distribution characteristics, it can accurately describe the delay jitter and other issues in spatial information networks. Therefore, this embodiment uses the following offset gamma probability density function to integrate and obtain the sub-network delay fluctuation:

[0066]

[0067] In the formula, Indicates the minimum delay. This represents the time delay variance, i.e., the shape factor. This represents the average time delay, i.e., the scale parameter. Represents function variables, This indicates the solution for the gamma function. This represents the base of the natural logarithm. Based on the detected data, these parameters can be updated to obtain the traffic characteristics and service transmission delays of each sub-network under different network conditions.

[0068] In step S211, a confidence level is introduced to obtain the specific values ​​of traffic pattern distribution and service transmission delay through integration. The confidence level represents the probability of the range of values ​​for the traffic pattern distribution and service transmission delay. By selecting a confidence level within the range of 10-95%, the range of values ​​for the spatial information network traffic pattern distribution and service transmission delay can be obtained (generally, the higher the selected confidence level, the higher the reliability of the obtained traffic pattern distribution and service transmission delay values, but the lower the numerical accuracy). Therefore, more specifically, based on the traffic data of a certain historical period, the transmission delay of each sub-network layer in the spatial information network in that historical period is obtained by integration using the offset gamma probability density function, including: based on the traffic data of a certain historical period, using the offset gamma probability density function with a preset confidence level to obtain the transmission delay of each sub-network layer in the spatial information network in that historical period, wherein the traffic data packets of the certain historical period contain the server response time collected in that historical period and the service transmission time after the service connection is established between the two ends of the communication.

[0069] S212. Based on the transmission delay of each sub-network in a certain historical period, the end-to-end transmission delay of the spatial information network in that certain historical period is obtained by combining the following formulas. :

[0070]

[0071] In the formula, This indicates the server response time collected during a specific historical period. This indicates the number of subnetworks in the spatial information network. Indicates less than or equal to positive integers, In the spatial information network, the first The transmission delay of each sub-network during a certain historical period.

[0072] In step S2, specifically, when the traffic feature map and application performance indicators include packet loss rate, for a certain historical period among the M historical periods, based on the corresponding traffic data, the traffic feature map and application performance indicators of each sub-network layer in the spatial information network in the corresponding period are first simulated using a probability distribution function. Then, based on the traffic feature map and application performance indicators of each sub-network layer in the corresponding period, the traffic feature map and end-to-end application performance indicators of the spatial information network in the corresponding period are composited, including but not limited to the following steps S221 to S223.

[0073] S221. Based on the traffic data of a certain historical period, use the normal distribution probability density function to obtain the packet loss rate of each sub-network in the spatial information network during that historical period.

[0074] In step S221, considering that the packet loss rate changes in the dynamic distribution of traffic patterns in the sub-networks conform to a normal distribution, the following normal distribution probability density function can be used to obtain the packet loss fluctuations of each sub-network layer:

[0075]

[0076] In the formula, This represents the average packet loss rate when the traffic distribution is in the subnetwork. This represents the standard deviation of the packet loss rate. This represents the variance of the packet loss rate.

[0077] S222. Based on the packet loss rate of each sub-network in a certain historical period, the end-to-end packet loss rate of the spatial information network in that historical period is obtained by combining the following formulas. :

[0078]

[0079] In the formula, This indicates the number of subnetworks in the spatial information network. Indicates less than or equal to positive integers, In the spatial information network, the first The packet loss rate of each subnetwork during a certain historical period.

[0080] S3. Perform data preprocessing on the traffic characteristic maps and end-to-end application performance indicators of the spatial information network in the M historical time periods to obtain M sets of traffic characteristic data corresponding one-to-one with the M historical time periods.

[0081] In step S3, specifically, the data preprocessing includes, but is not limited to, data cleaning and standardization, so that the traffic characteristic data is suitable for input into the model.

[0082] S4. Extract N sets of traffic characteristic data from the M sets of traffic characteristic data that correspond one-to-one with the last N historical periods in the M historical periods. For each set of traffic characteristic data in the N sets of traffic characteristic data, extract K sets of traffic characteristic data from the M sets of traffic characteristic data that correspond one-to-one with the K most recent historical periods before the corresponding period in the M historical periods to obtain the corresponding traffic characteristic data time series. Here, N and K represent positive integers and N+K is less than or equal to M.

[0083] S5. Take the N time series of traffic feature data as N sample data and the N traffic feature data as N label data corresponding one-to-one with the N sample data, and import them into the GCN+BiGRU model for model training to obtain a traffic feature data prediction model. The GCN+BiGRU model includes, but is not limited to, a graph convolutional neural network (GCN), a bidirectional gated recurrent unit (BiGRU), and a fully connected layer. The graph convolutional neural network (GCN) is used to model the spatial characteristics of the spatial information network traffic after inputting the N sample data and extract the spatial features of the spatial information network traffic. The bidirectional gated recurrent unit (BiGRU) is used to model the temporal characteristics of the spatial information network traffic after inputting the N sample data and extract the temporal features of the spatial information network traffic. The fully connected layer is used to process the spatial and temporal features of the spatial information network traffic to obtain the final prediction result.

[0084] In step S5, the Graph Convolutional Network (GCN), as an existing feature extractor for graph data, is mainly used to extract features from graph data. Therefore, in this embodiment, the spatial features of the spatial information network traffic can be extracted using the GCN. The Bidirectional Gated Recurrent Unit (BiGRU) is an existing bidirectional recurrent neural network based on the GRU (Gated Recurrent Unit, a type of recurrent neural network with optimized structure based on Long Short-Term Memory networks) and composed of forward and backward neural networks. It is mainly used to acquire and utilize feature information from both directions at a given moment. Therefore, in this embodiment, the temporal features of the spatial information network traffic can be extracted using the BiGRU. The fully connected layer is an essential unit in existing machine learning models, mainly used to process the extracted features to obtain the final prediction result. Thus, the GCN+BiGRU model can be conventionally built based on the existing GCN, BiGRU, and fully connected layer to train the traffic feature data prediction model.

[0085] S6. Obtain the traffic data of the spatial information network in the most recent K historical time periods.

[0086] S7. For each historical period in the current K most recent historical periods, based on the corresponding traffic data, first use the probability distribution function to simulate and obtain the traffic characteristic map and application performance index of each sub-network layer in the spatial information network in the corresponding period. Then, based on the traffic characteristic map and application performance index of each sub-network layer in the corresponding period, combine them to obtain the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period.

[0087] S8. Perform the data preprocessing on the traffic characteristic maps and end-to-end application performance indicators of the spatial information network in the current K most recent historical periods to obtain the current traffic characteristic data time series.

[0088] The technical details of the aforementioned steps S7 and S8 can be derived by referring to the aforementioned steps S2 and S3, and will not be repeated here.

[0089] S9. Input the current traffic characteristic data time series into the traffic characteristic data prediction model, and output the traffic characteristic data for the next time period.

[0090] S10. Based on the traffic characteristic data of the current next time period, obtain the traffic status of the spatial information network in the current next time period.

[0091] In step S10, since the traffic characteristic data of the current next time period has been obtained, the traffic situation of the spatial information network in the current next time period can be directly obtained based on conventional traffic situation analysis methods.

[0092] Therefore, based on the spatial information network traffic situation prediction method described in steps S1 to S10 above, a training and application scheme for a spatial information network traffic situation prediction model based on hierarchical modeling and the GCN+BiGRU model is provided. Specifically, after acquiring traffic data of the spatial information network over multiple historical time periods, the traffic characteristic maps and application performance indicators of each sub-network layer in each time period are first simulated using a probability distribution function. Then, these are combined to obtain the traffic characteristic maps and end-to-end application performance indicators of the spatial information network in each time period. Next, the time series of traffic characteristic data and label data obtained from data processing are imported into the GCN+BiGRU model for model training, resulting in a traffic characteristic data prediction model. Finally, the current traffic characteristic data time series is input into the traffic characteristic data prediction model, outputting the traffic characteristic data and traffic situation for the next time period. This improves the accuracy of spatial information network traffic prediction and enhances the ability to perceive traffic situation. Furthermore, it can solve the problem of multi-service data fusion processing of spatial information networks, improve the prediction capability of multi-source traffic situation, and enhance the model's generalization ability, facilitating practical application and promotion.

[0093] like Figure 2 As shown, the second aspect of this embodiment provides a virtual device for implementing the spatial information network traffic situation prediction method described in the first aspect, including a data acquisition module, a hierarchical composite module, a data processing module, a data extraction module, a model training module, a traffic feature prediction module, and a traffic situation determination module.

[0094] The data acquisition module is used to acquire traffic data of the spatial information network in M ​​historical time periods, where M represents a positive integer and the M historical time periods are consecutive in time.

[0095] The layered composite module is connected to the data acquisition module and is used to, for each historical period in the M historical periods, first use the probability distribution function to simulate the traffic characteristic map and application performance index of each sub-network in the spatial information network in the corresponding period according to the corresponding traffic data, and then, based on the traffic characteristic map and application performance index of each sub-network in the corresponding period, composite the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period.

[0096] The data processing module is communicatively connected to the hierarchical composite module and is used to preprocess the traffic characteristic maps and end-to-end application performance indicators of the spatial information network in the M historical time periods to obtain M traffic characteristic data corresponding one-to-one with the M historical time periods.

[0097] The data extraction module is communicatively connected to the data processing module. It is used to extract N sets of traffic feature data from the M sets of traffic feature data, which correspond one-to-one with the last N historical periods in the M historical periods. For each set of traffic feature data in the N sets of traffic feature data, it extracts K sets of traffic feature data from the M sets of traffic feature data, which correspond one-to-one with the K most recent historical periods before the corresponding period in the M historical periods, to obtain the corresponding traffic feature data time series. Here, N and K represent positive integers and N+K is less than or equal to M.

[0098] The model training module, communicatively connected to the data extraction module, is used to input N time series of traffic feature data as N sample data and N sets of traffic feature data as N label data corresponding one-to-one with the N sample data into the GCN+BiGRU model for model training to obtain a traffic feature data prediction model. The GCN+BiGRU model includes a Graph Convolutional Neural Network (GCN), a Bidirectional Gated Recurrent Unit (BiGRU), and a fully connected layer. The GCN is used to model the spatial characteristics of the spatial information network traffic after inputting the N sample data, extracting the spatial features of the spatial information network traffic. The Bidirectional Gated Recurrent Unit (BiGRU) is used to model the temporal characteristics of the spatial information network traffic after inputting the N sample data, extracting the temporal features of the spatial information network traffic. The fully connected layer processes the spatial and temporal features of the spatial information network traffic to obtain the final prediction result.

[0099] The data acquisition module is also used to acquire the traffic data of the spatial information network in the most recent K historical time periods;

[0100] The layered composite module is further configured to, for each historical period in the current K most recent historical periods, first use the probability distribution function to simulate and obtain the traffic characteristic map and application performance index of each sub-network in the spatial information network in the corresponding period based on the corresponding traffic data, and then, based on the traffic characteristic map and application performance index of each sub-network in the corresponding period, composite the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period.

[0101] The data processing module is also used to perform the data preprocessing on the traffic characteristic map and end-to-end application performance index of the spatial information network in the current K most recent historical periods to obtain the current traffic characteristic data time series;

[0102] The traffic feature prediction module is communicatively connected to the data processing module and the model training module, respectively, and is used to input the current traffic feature data time series into the traffic feature data prediction model and output the traffic feature data for the next time period.

[0103] The traffic situation determination module is communicatively connected to the traffic feature prediction module and is used to obtain the traffic situation of the spatial information network in the current next time period based on the traffic feature data of the current next time period.

[0104] The working process, working details and technical effects of the aforementioned device provided in the second aspect of this embodiment can be found in the spatial information network traffic situation prediction method described in the first aspect, and will not be repeated here.

[0105] like Figure 3 As shown, the third aspect of this embodiment provides a computer device for executing the spatial information network traffic situation prediction method as described in the first aspect. The device includes a memory, a processor, and a transceiver connected in sequence. The memory stores a computer program, the transceiver sends and receives messages, and the processor reads the computer program and executes the spatial information network traffic situation prediction method as described in the first aspect. Specifically, the memory may include, but is not limited to, random-access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or first-in-last-out (FILO) memory, etc.; the processor may include, but is not limited to, a microprocessor of the STM32F105 series. Furthermore, the computer device may also include, but is not limited to, a power module, a display screen, and other necessary components.

[0106] The working process, working details and technical effects of the aforementioned computer equipment provided in the third aspect of this embodiment can be found in the spatial information network traffic situation prediction method described in the first aspect, and will not be repeated here.

[0107] This fourth aspect of the embodiment provides a computer-readable storage medium storing instructions comprising the spatial information network traffic situation prediction method as described in the first aspect. Specifically, the computer-readable storage medium stores instructions that, when executed on a computer, perform the spatial information network traffic situation prediction method as described in the first aspect. The computer-readable storage medium refers to a data storage medium, which may include, but is not limited to, floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or Memory Sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0108] The working process, working details and technical effects of the aforementioned computer-readable storage medium provided in the fourth aspect of this embodiment can be found in the spatial information network traffic situation prediction method described in the first aspect, and will not be repeated here.

[0109] This fifth aspect of the embodiment provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform the spatial information network traffic situation prediction method as described in the first aspect. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0110] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for predicting the traffic situation of a spatial information network, characterized in that, include: Acquire traffic data of the spatial information network over M historical time periods, where M represents a positive integer and the M historical time periods are consecutive in time. For each historical period in the M historical periods, based on the corresponding traffic data, the traffic characteristic map and application performance index of each sub-network in the spatial information network in the corresponding period are first obtained by using the probability distribution function. Then, based on the traffic characteristic map and application performance index of each sub-network in the corresponding period, the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period are combined to obtain the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period. The spatial information network is preprocessed with traffic characteristic maps and end-to-end application performance indicators for the M historical time periods to obtain M traffic characteristic data corresponding one-to-one with the M historical time periods. N sets of traffic characteristic data are extracted from the M sets of traffic characteristic data, each corresponding to one of the last N historical periods in the M historical periods. For each set of traffic characteristic data in the N sets of traffic characteristic data, K sets of traffic characteristic data are extracted from the M sets of traffic characteristic data, each corresponding to one of the K most recent historical periods before the corresponding period in the M historical periods, to obtain the corresponding traffic characteristic data time series. Here, N and K represent positive integers and N+K is less than or equal to M. N time series of traffic feature data are used as N sample data, and N sets of traffic feature data are used as N label data corresponding one-to-one with the N sample data. These are then imported into a GCN+BiGRU model for model training to obtain a traffic feature data prediction model. The GCN+BiGRU model includes a graph convolutional neural network (GCN), a bidirectional gated recurrent unit (BiGRU), and a fully connected layer. The GCN is used to model the spatial characteristics of the spatial information network traffic after inputting the N sample data, and extract the spatial features of the spatial information network traffic. The Bidirectional gated recurrent unit (BiGRU) is used to model the temporal characteristics of the spatial information network traffic after inputting the N sample data, and extract the temporal features of the spatial information network traffic. The fully connected layer is used to process the spatial and temporal features of the spatial information network traffic to obtain the final prediction result. Obtain the traffic data of the spatial information network for the K most recent historical time periods; For each historical period in the current K most recent historical periods, based on the corresponding traffic data, the probability distribution function is first used to simulate and obtain the traffic characteristic map and application performance index of each sub-network in the spatial information network in the corresponding period. Then, based on the traffic characteristic map and application performance index of each sub-network in the corresponding period, the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period are combined to obtain the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period. The spatial information network is preprocessed with the traffic characteristic maps and end-to-end application performance indicators of the current K most recent historical periods to obtain the current traffic characteristic data time series. The current traffic characteristic data time series is input into the traffic characteristic data prediction model, and the traffic characteristic data for the next time period is output. Based on the traffic characteristic data of the current next time period, the traffic situation of the spatial information network in the current next time period is obtained.

2. The spatial information network traffic situation prediction method according to claim 1, characterized in that, The traffic feature map and application performance indicators include service type, service data size, service priority, link bandwidth, service characteristics under gateway load rate, transmission delay and / or packet loss rate.

3. The spatial information network traffic situation prediction method according to claim 1, characterized in that, When the traffic characteristic map and application performance indicators include transmission delay, for a certain historical period among the M historical periods, based on the corresponding traffic data, the traffic characteristic map and application performance indicators of each sub-network layer in the spatial information network in the corresponding period are first simulated using a probability distribution function. Then, based on the traffic characteristic map and application performance indicators of each sub-network layer in the corresponding period, the traffic characteristic map and end-to-end application performance indicators of the spatial information network in the corresponding period are composited, including: Based on the traffic data of a certain historical period, the transmission delay of each sub-network in the spatial information network is obtained by integral of the offset gamma probability density function during the certain historical period. The traffic data packet of the certain historical period contains the server response time collected during the certain historical period and the service transmission time after the service connection is established between the two ends of the communication. Based on the transmission delays of each sub-network layer during a certain historical period, the end-to-end transmission delay of the spatial information network during that historical period is obtained by combining the following formulas. : In the formula, This indicates the server response time collected during a specific historical period. This indicates the number of sub-networks in the spatial information network. Indicates less than or equal to positive integers, In the spatial information network, the first The transmission delay of each sub-network during a certain historical period.

4. The spatial information network traffic situation prediction method according to claim 3, characterized in that, Based on the traffic data for a certain historical period, the transmission delay of each sub-network layer in the spatial information network during that historical period is obtained by integrating the offset gamma probability density function, including: Based on the traffic data of a certain historical period, the transmission delay of each sub-network in the spatial information network in the certain historical period is obtained by integral of the offset gamma probability density function based on a preset confidence level. The traffic data packet of the certain historical period contains the server response time and the service transmission time after the service connection is established at both ends of the communication during the certain historical period.

5. The spatial information network traffic situation prediction method according to claim 1, characterized in that, When the traffic feature map and application performance indicators include packet loss rate, for a certain historical period among the M historical periods, based on the corresponding traffic data, the traffic feature map and application performance indicators of each sub-network layer in the spatial information network at the corresponding time period are first simulated using a probability distribution function. Then, based on the traffic feature map and application performance indicators of each sub-network layer at the corresponding time period, the traffic feature map and end-to-end application performance indicators of the spatial information network at the corresponding time period are composited, including: Based on the traffic data for a certain historical period, the packet loss rate of each sub-network in the spatial information network during that historical period is obtained using the normal distribution probability density function. Based on the packet loss rates of each sub-network layer during a certain historical period, the end-to-end packet loss rate of the spatial information network during that historical period is obtained by combining the results using the following formula. : In the formula, This indicates the number of sub-networks in the spatial information network. Indicates less than or equal to positive integers, In the spatial information network, the first The packet loss rate of each subnetwork during a certain historical period.

6. The spatial information network traffic situation prediction method according to claim 1, characterized in that, The data preprocessing includes data cleaning and standardization.

7. The spatial information network traffic situation prediction method according to claim 1, characterized in that, The sub-networks in the space information network include satellite networks, terrestrial fiber optic networks, and wireless ad hoc networks.

8. A spatial information network traffic situation prediction device, characterized in that, It includes a data acquisition module, a hierarchical composite module, a data processing module, a data extraction module, a model training module, a traffic feature prediction module, and a traffic situation determination module; The data acquisition module is used to acquire traffic data of the spatial information network in M ​​historical time periods, where M represents a positive integer and the M historical time periods are consecutive in time. The layered composite module is connected to the data acquisition module and is used to, for each historical period in the M historical periods, first use the probability distribution function to simulate the traffic characteristic map and application performance index of each sub-network in the spatial information network in the corresponding period according to the corresponding traffic data, and then, based on the traffic characteristic map and application performance index of each sub-network in the corresponding period, composite the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period. The data processing module is communicatively connected to the hierarchical composite module and is used to preprocess the traffic characteristic maps and end-to-end application performance indicators of the spatial information network in the M historical time periods to obtain M traffic characteristic data corresponding one-to-one with the M historical time periods. The data extraction module is communicatively connected to the data processing module. It is used to extract N sets of traffic feature data from the M sets of traffic feature data, which correspond one-to-one with the last N historical periods in the M historical periods. For each set of traffic feature data in the N sets of traffic feature data, it extracts K sets of traffic feature data from the M sets of traffic feature data, which correspond one-to-one with the K most recent historical periods before the corresponding period in the M historical periods, to obtain the corresponding traffic feature data time series. Here, N and K represent positive integers and N+K is less than or equal to M. The model training module, communicatively connected to the data extraction module, is used to input N time series of traffic feature data as N sample data and N sets of traffic feature data as N label data corresponding one-to-one with the N sample data into the GCN+BiGRU model for model training to obtain a traffic feature data prediction model. The GCN+BiGRU model includes a Graph Convolutional Neural Network (GCN), a Bidirectional Gated Recurrent Unit (BiGRU), and a fully connected layer. The GCN is used to model the spatial characteristics of the spatial information network traffic after inputting the N sample data, extracting the spatial features of the spatial information network traffic. The Bidirectional Gated Recurrent Unit (BiGRU) is used to model the temporal characteristics of the spatial information network traffic after inputting the N sample data, extracting the temporal features of the spatial information network traffic. The fully connected layer processes the spatial and temporal features of the spatial information network traffic to obtain the final prediction result. The data acquisition module is also used to acquire the traffic data of the spatial information network in the most recent K historical time periods; The layered composite module is further configured to, for each historical period in the current K most recent historical periods, first use the probability distribution function to simulate and obtain the traffic characteristic map and application performance index of each sub-network in the spatial information network in the corresponding period based on the corresponding traffic data, and then, based on the traffic characteristic map and application performance index of each sub-network in the corresponding period, composite the traffic characteristic map and end-to-end application performance index of the spatial information network in the corresponding period. The data processing module is also used to perform the data preprocessing on the traffic characteristic map and end-to-end application performance index of the spatial information network in the current K most recent historical periods to obtain the current traffic characteristic data time series; The traffic feature prediction module is communicatively connected to the data processing module and the model training module, respectively, and is used to input the current traffic feature data time series into the traffic feature data prediction model and output the traffic feature data for the next time period. The traffic situation determination module is communicatively connected to the traffic feature prediction module and is used to obtain the traffic situation of the spatial information network in the current next time period based on the traffic feature data of the current next time period.

9. A computer device, characterized in that, The device includes a memory, a processor, and a transceiver connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the spatial information network traffic situation prediction method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that... The computer-readable storage medium stores instructions that, when executed on a computer, perform the spatial information network traffic situation prediction method as described in any one of claims 1 to 7.

Images