Service type prediction method and device for space-based information network

Abstract

The invention provides a service type prediction method and a service type prediction device for a space-based information network. Performing time dimension service type prediction on the service traffic data of each link currently corresponding to the space-based information network, and performing spatial dimension service type prediction on a traffic matrix formed by the service traffic data; and generating target service type prediction result data corresponding to the service flow data of each link according to the time dimension service type prediction result and the space dimension service type prediction result corresponding to the service flow data of each link. According to the method and the device, the service type prediction for the space-based information network can be realized, the joint prediction of the comprehensive service of the space information network can be more adapted, and the prediction precision and efficiency of the service type prediction of the space-based information network can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of network communication technologies, and particularly to a method and device for predicting service types for a space-based information network. Background Art

[0002] A space-based information network is a multi-layer heterogeneous dynamic complex network. Conventional communication services have the same self-similarity as those of terrestrial networks, but the generation and flow rules of services are more complex, and they carry diverse service flows with comprehensive characteristics. This self-similarity of traffic in the space-based network will have a great negative impact on the performance of the entire network, such as increased network congestion and packet loss rate. For terrestrial networks with abundant bandwidth, the impact of these problems may not be obvious, but for space-based information networks with limited bandwidth and tight resources, it will lead to a large loss of network performance. To address the problems brought about by resource constraints and service diversification in space-based information networks, an effective solution is to add a service prediction mechanism to the system, so as to achieve early warning of possible burst data and then take effective countermeasures. This is particularly important for satellite communication systems with large transmission delays.

[0003] Currently, traditional service type prediction methods usually adopt time series algorithms or Boosting algorithm clusters in ensemble learning. These service prediction methods are proposed based on terrestrial networks with abundant bandwidth, and only use a single scenario or a single service, which will result in poor multi-service comprehensive prediction ability of traditional service type prediction methods; moreover, the requirements for reliability and transmission rate of services in space-based information networks are significantly different from those of terrestrial networks. In the context of service diversification in space-based information networks, single-service prediction methods are obviously difficult to meet the service and traffic prediction requirements.

[0004] Based on this, there is an urgent need to design a method that can meet the prediction accuracy and real-time requirements of service type prediction for space-based information networks. Summary of the Invention

[0005] In view of this, embodiments of the present application provide a method and device for predicting service types for a space-based information network to eliminate or improve one or more defects existing in the prior art.

[0006] One aspect of the present application provides a method for predicting service types for a space-based information network, including:

[0007] Based on a preset multi-mode combination service prediction method, respectively perform service type prediction in the time dimension on the service traffic data of each corresponding link of the space-based information network to obtain the time dimension service type prediction results corresponding to the service traffic data of each of the links;

[0008] And, based on the multi-mode combined service prediction method, perform service type prediction in the spatial dimension on a traffic matrix jointly composed of the service traffic data of each of the links, so as to obtain the spatial dimension service type prediction results corresponding to the service traffic data of each of the links;

[0009] Generate the target service type prediction result data corresponding to the service traffic data of each of the links according to the time dimension service type prediction results and the spatial dimension service type prediction results corresponding to the service traffic data of each of the links.

[0010] In some embodiments of the present application, the multi-mode combined service prediction method includes: the ensemble empirical mode decomposition method; the time dimension service type prediction results include: the first service type prediction result;

[0011] Correspondingly, the step of performing service type prediction in the time dimension on the service traffic data of each of the current links of the space-based information network based on the preset multi-mode combined service prediction method to obtain the time dimension service type prediction results corresponding to the service traffic data of each of the links includes:

[0012] Using the ensemble empirical mode decomposition method, perform service type prediction in the time dimension on the service traffic data of each of the current links of the space-based information network, so as to obtain the first service type prediction results corresponding to the service traffic data of each of the links.

[0013] In some embodiments of the present application, the multi-mode combined service prediction method further includes: a service prediction method based on control charts; the time dimension service type prediction results include: the second service type prediction result;

[0014] Correspondingly, the step of performing service type prediction in the time dimension on the service traffic data of each of the current links of the space-based information network based on the preset multi-mode combined service prediction method to obtain the time dimension service type prediction results corresponding to the service traffic data of each of the links includes:

[0015] Using the service prediction method based on control charts, perform service type prediction in the time dimension on the service traffic data of each of the current links of the space-based information network, so as to obtain the second service type prediction results corresponding to the service traffic data of each of the links.

[0016] In some embodiments of the present application, the multi-mode combined service prediction method further includes: a service prediction method based on the spatial characteristics of satellite network traffic;

[0017] Correspondingly, based on the multi-mode combined service prediction method, a service type prediction in the spatial dimension is performed on a traffic matrix jointly composed of the traffic data of each of the links, so as to obtain a spatial dimension service type prediction result corresponding to the traffic data of each of the links, including:

[0018] Adopt the service prediction method based on the spatial characteristics of satellite network traffic to perform a service type prediction in the spatial dimension on a traffic matrix jointly composed of the traffic data of each of the links, so as to obtain a spatial dimension service type prediction result corresponding to the traffic data of each of the links.

[0019] In some embodiments of the present application, the method of using the ensemble empirical mode decomposition method to perform a service type prediction in the time dimension on the traffic data of each link corresponding to the current space-based information network, so as to obtain a first service type prediction result corresponding to the traffic data of each of the links, includes:

[0020] Based on the ensemble empirical mode decomposition method, signal decomposition is respectively performed on the traffic data of each link corresponding to the current space-based information network to obtain an intrinsic mode function component group corresponding to each of the traffic data, wherein each intrinsic mode function component group contains a plurality of intrinsic mode function components;

[0021] Each of the intrinsic mode function components in each of the intrinsic mode function component groups is respectively input into a preset service type prediction model, so that the service type prediction model respectively outputs service type prediction data corresponding to each of the intrinsic mode function components;

[0022] According to the mean value between the service type prediction data of each of the intrinsic mode function components in each of the intrinsic mode function component groups, a first service type prediction result corresponding to each of the traffic data is respectively determined.

[0023] In some embodiments of the present application, the method of using the service prediction method based on control charts to perform a service type prediction in the time dimension on the traffic data of each link corresponding to the current space-based information network, so as to obtain a second service type prediction result corresponding to the traffic data of each of the links, includes:

[0024] Based on the control chart, data separation is respectively performed on the traffic data of each link corresponding to the current space-based information network to obtain non-burst data and burst data corresponding to each of the traffic data;

[0025] Input each of the non-burst data into a preset service type prediction model respectively, so that the service type prediction model outputs the service type prediction data corresponding to each of the non-burst data respectively;

[0026] And, perform self-adaptive template matching prediction on each of the burst data respectively to obtain the service type prediction data corresponding to each of the burst data respectively;

[0027] Integrate the service type prediction data of the non-burst data and the burst data corresponding to each of the service traffic data respectively to obtain the second service type prediction result corresponding to each of the service traffic data respectively.

[0028] In some embodiments of the present application, the service type prediction method based on the traffic space characteristics of the satellite network is adopted to perform service type prediction in the spatial dimension on a traffic matrix jointly composed of the service traffic data of each of the links, so as to obtain the spatial dimension service type prediction results corresponding to the service traffic data of each of the links, including:

[0029] Perform traffic classification on a traffic matrix jointly composed of the service traffic data of each of the links to obtain corresponding multiple sub-traffic matrices;

[0030] Based on the preset locally linear embedding non-negative matrix factorization method, perform dimensionality reduction processing on each of the sub-traffic matrices respectively to form corresponding low-dimensional matrices, and perform service type prediction on each of the low-dimensional matrices, and then determine the service type prediction data corresponding to each of the sub-traffic matrices respectively according to the service type prediction data corresponding to each of the low-dimensional matrices;

[0031] Obtain the service type prediction data of the traffic matrix according to the service type prediction data corresponding to each of the sub-traffic matrices, and use the service type prediction data of the traffic matrix as the spatial dimension service type prediction results corresponding to the service traffic data of each of the links.

[0032] Another aspect of the present application provides a service type prediction device for a space-based information network, including:

[0033] A time dimension prediction module, configured to perform service type prediction in the time dimension on the service traffic data of each of the links corresponding to the current space-based information network respectively based on a preset multi-mode combination service prediction method, so as to obtain the time dimension service type prediction results corresponding to the service traffic data of each of the links;

[0034] And a spatial dimension prediction module, configured to perform a service type prediction in the spatial dimension on a traffic matrix jointly composed of the traffic data of each of the links based on the multi-mode combined service prediction method, so as to obtain a spatial dimension service type prediction result corresponding to the traffic data of each of the links;

[0035] A prediction result generation module, configured to generate a target service type prediction result data corresponding to the traffic data of each of the links according to the time dimension service type prediction result and the spatial dimension service type prediction result corresponding to the traffic data of each of the links.

[0036] A third aspect of the present application provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where when the processor executes the computer program, the service type prediction method for a space-based information network described above is implemented.

[0037] A fourth aspect of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the service type prediction method for a space-based information network described above is implemented.

[0038] The service type prediction method for a space-based information network provided by the present application, based on a preset multi-mode combined service prediction method, respectively performs a service type prediction in the time dimension on the traffic data of each link corresponding to the current space-based information network, so as to obtain a time dimension service type prediction result corresponding to the traffic data of each of the links; and based on the multi-mode combined service prediction method, performs a service type prediction in the spatial dimension on a traffic matrix jointly composed of the traffic data of each of the links, so as to obtain a spatial dimension service type prediction result corresponding to the traffic data of each of the links; according to the time dimension service type prediction result and the spatial dimension service type prediction result corresponding to the traffic data of each of the links, generates a target service type prediction result data corresponding to the traffic data of each of the links, can realize the service type prediction for the space-based information network, can be more suitable for the joint prediction of the integrated services of the space information network, and can effectively improve the prediction accuracy and efficiency of the service type prediction of the space-based information network, thereby laying a foundation for the space-based information network to provide a more powerful integrated service and guarantee ability.

[0039] The additional advantages, objectives, and features of the present application will be partially described below, and will become partially apparent to those of ordinary skill in the art after studying the following part, or can be learned through the practice of the present application. The objectives and other advantages of the present application can be achieved and obtained by the structures specifically pointed out in the specification and the drawings.

[0040] Those skilled in the art will understand that the objectives and advantages achievable with the present application are not limited to those specifically described above, and the above and other objectives achievable with the present application will be more clearly understood from the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The drawings described herein are used to provide a further understanding of the present application, form a part of the present application, and do not limit the present application. The components in the drawings are not drawn to scale, but are only for showing the principles of the present application. For the convenience of showing and describing some parts of the present application, the corresponding parts in the drawings may be enlarged, that is, may become larger relative to other components in the exemplary device actually manufactured according to the present application. In the drawings:

[0042] Figure 1 FIG. 11 is a first schematic flowchart of a service type prediction method for a space-based information network according to an embodiment of the present application.

[0043] Figure 2 FIG. 15 is a second schematic flowchart of a service type prediction method for a space-based information network according to an embodiment of the present application.

[0044] Figure 3 FIG. 19 is a third schematic flowchart of a service type prediction method for a space-based information network according to an embodiment of the present application.

[0045] Figure 4 FIG. 23 is a schematic diagram for judging burst data in traffic volume through a control chart in an example of the present application.

[0046] Figure 5 FIG. 27 is a schematic diagram of the effect after removing bursts through a control chart method in an example of the present application.

[0047] Figure 6 FIG. 31 is a schematic flowchart for predicting separated non-burst data and burst data respectively using an artificial neural network (ANN) and an adaptive template matching method in an example of the present application.

[0048] Figure 7 FIG. 35 is a schematic diagram of a compression prediction process based on LPNMF in an example of the present application.

[0049] Figure 8 FIG. 39 is a flowchart of a multi-mode combined service prediction in an application example of the present application.

[0050] Figure 9 FIG. 43 is a schematic structural diagram of a service type prediction device for a space-based information network according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0051] To make the objectives, technical solutions, and advantages of this application clearer and more understandable, the following further elaborates on this application in conjunction with the embodiments and the accompanying drawings. Herein, the illustrative embodiments of this application and their descriptions are used to explain this application, but do not limit this application.

[0052] Here, it should also be noted that to avoid obscuring this application with unnecessary details, only the structures and / or processing steps closely related to the solution of this application are shown in the drawings, while other details less relevant to this application are omitted.

[0053] It should be emphasized that the term "comprising / including" when used herein refers to the presence of features, elements, steps, or components, but does not exclude the presence or addition of one or more other features, elements, steps, or components.

[0054] Here, it should also be noted that if not otherwise specified, the term "connection" in this text can not only refer to a direct connection, but also represent an indirect connection with an intermediate.

[0055] In the following, embodiments of this application will be described with reference to the accompanying drawings. In the drawings, the same reference numerals represent the same or similar components, or the same or similar steps.

[0056] Traditional ground-oriented business prediction methods cannot meet the requirements for the prediction accuracy and efficiency of the service types in the space-based information network. For example, through time series algorithms, a large amount of research and exploration has been carried out on the prediction of ground-oriented users and service traffic. Most of these methods abstract the service traffic prediction problem in the scenario into a time series problem to solve. Among them, relatively typical ones are the autoregressive integrated moving average (ARIMA) algorithm, the Holt-Winter algorithm, the FBProphet algorithm, etc. In addition to time series algorithms, machine learning is also one of the important methods for ground-oriented business prediction. The most commonly used is the Boosting algorithm cluster in ensemble learning. Typical representatives include the gradient boosting decision tree (GBDT) algorithm, the extreme gradient boosting (XGBoost) algorithm, the light gradient boosting (LightGBM) algorithm, etc. In addition, there are also predictions based on chaotic sequences and combined predictions. These types of business prediction methods are all proposed based on the ground network with abundant bandwidth. Using only a single scenario or a single service will result in poor multi-service comprehensive prediction capabilities of traditional service type prediction methods; moreover, the requirements for reliability and transmission rate of the services in the space-based information network are significantly different from those of the ground network. In the context of the diversification of services in the space-based information network, a single service prediction method is obviously difficult to meet the service and traffic prediction requirements.

[0057] Based on this, in order to meet the requirements of prediction accuracy and real-time performance for the service type prediction of the space-based information network, the embodiments of the present application respectively provide a service type prediction method for the space-based information network, a service type prediction device for the space-based information network for executing the service type prediction method for the space-based information network, an entity device, and a computer-readable storage medium. By absorbing the advantages of multiple methods and adopting the strategy of combining multiple methods for service prediction, it is the only way to improve prediction accuracy and real-time performance.

[0058] Specific details are described in the following embodiments.

[0059] Based on this, the embodiments of the present application provide a service type prediction method for the space-based information network that can be implemented by a service type prediction device for the space-based information network. Refer to Figure 1 , the service type prediction method for the space-based information network specifically includes the following content:

[0060] Step 100: Based on a preset multi-mode combined service prediction method, perform service type prediction in the time dimension on the service traffic data of each corresponding link of the space-based information network respectively, so as to obtain the time dimension service type prediction results corresponding to the service traffic data of each of the links.

[0061] And, Step 200: Based on the multi-mode combined service prediction method, perform service type prediction in the space dimension on a traffic matrix composed of the service traffic data of each of the links, so as to obtain the space dimension service type prediction results corresponding to the service traffic data of each of the links.

[0062] In one or more embodiments of the present application, the multi-mode combined service prediction method includes at least one service type prediction method in the time dimension and at least one service type prediction method in the space dimension, so as to use a variety of and comprehensive combined service prediction methods to perform type prediction on the services of the space-based information network.

[0063] Step 300: Generate the target service type prediction result data corresponding to the service traffic data of each of the links according to the time dimension service type prediction results and the space dimension service type prediction results corresponding to the service traffic data of each of the links.

[0064] In one implementation of step 300, the time - dimension service type prediction results and the space - dimension service type prediction results corresponding to the service traffic data of each of the links can be aggregated, and then the aggregated result can be used as the target service type prediction result data corresponding to the service traffic data of each of the links, so as to achieve early warning of possible burst data based on the target service type prediction result data corresponding to the service traffic data of each of the links, and then take effective countermeasures. This is particularly important for satellite communication systems with large transmission delays.

[0065] In order to further improve the convenience and effectiveness of checking the target service type prediction result data, in another implementation of step 300, the time - dimension service type prediction results and the space - dimension service type prediction results corresponding to the service traffic data of each of the links can be processed based on preset screening or selection rules, algorithms, etc., so as to obtain a unique service type prediction result corresponding to the service traffic data of each link from the time - dimension service type prediction results and the space - dimension service type prediction results corresponding to the service traffic data of each of the links, and use it as the corresponding target service type prediction result data. It can be understood that the preset screening or selection rules can be pre - specified by humans and can adopt more intelligent analysis methods.

[0066] As can be seen from the above description, the service type prediction method for a space - based information network provided by the embodiments of the present application can achieve service type prediction for a space - based information network, be more adaptable to the joint prediction of comprehensive services of a space information network, and effectively improve the prediction accuracy and efficiency of service type prediction for a space - based information network. Furthermore, it can lay a foundation for the space - based information network to provide stronger comprehensive service and guarantee capabilities.

[0067] In order to further improve the accuracy and reliability of the service type prediction in the time dimension for the service traffic data of each link currently corresponding to the space - based information network, in a service type prediction method for a space - based information network provided by the embodiments of the present application, the multi - mode combined service prediction method includes: the ensemble empirical mode decomposition method; the time - dimension service type prediction result includes: the first service type prediction result.

[0068] Correspondingly, referring to Figure 2 , step 100 in the service type prediction method for a space - based information network specifically includes the following content:

[0069] Step 110: Use the ensemble empirical mode decomposition method to perform business type prediction for the service traffic data of each current link in the space-based information network in the time dimension, so as to obtain the first service type prediction results corresponding to the service traffic data of each of the links.

[0070] To further improve the accuracy and reliability of the business type prediction for the service traffic data of each current link in the space-based information network in the time dimension, in a service type prediction method for a space-based information network provided in an embodiment of the present application, the multi-mode combined service prediction method further includes: a service prediction method based on a control chart; the time dimension service type prediction results include: second service type prediction results;

[0071] Correspondingly, see Figure 2 , the step 100 in the service type prediction method for a space-based information network specifically further includes the following content:

[0072] Step 120: Use the service prediction method based on the control chart to perform business type prediction for the service traffic data of each current link in the space-based information network in the time dimension, so as to obtain the second service type prediction results corresponding to the service traffic data of each of the links.

[0073] To further improve the accuracy and reliability of the business type prediction for the service traffic data of each current link in the space-based information network in the space dimension, in a service type prediction method for a space-based information network provided in an embodiment of the present application, the multi-mode combined service prediction method further includes: a service prediction method based on the spatial characteristics of satellite network traffic;

[0074] Correspondingly, see Figure 2 , the step 200 in the service type prediction method for a space-based information network specifically further includes the following content:

[0075] Step 210: Use the service prediction method based on the spatial characteristics of satellite network traffic to perform business type prediction for a traffic matrix jointly composed of the service traffic data of each of the links in the space dimension, so as to obtain the spatial dimension service type prediction results corresponding to the service traffic data of each of the links.

[0076] To further improve the accuracy and efficiency of using the ensemble empirical mode decomposition method to perform business type prediction for the service traffic data of each current link in the space-based information network in the time dimension, in a service type prediction method for a space-based information network provided in an embodiment of the present application, see Figure 3, step 110 in the service type prediction method for the space-based information network specifically includes the following content:

[0077] Step 111: Based on the ensemble empirical mode decomposition method, perform signal decomposition on the service traffic data of each current link corresponding to the space-based information network respectively, so as to obtain an intrinsic mode function component group corresponding to each piece of the service traffic data. Among them, each intrinsic mode function component group contains multiple intrinsic mode function components.

[0078] Step 112: Input each of the intrinsic mode function components in each intrinsic mode function component group into a preset service type prediction model respectively, so that the service type prediction model outputs service type prediction data corresponding to each of the intrinsic mode function components respectively.

[0079] Step 113: Determine the first service type prediction result corresponding to each piece of the service traffic data respectively according to the mean value among the service type prediction data of each of the intrinsic mode function components in each intrinsic mode function component group.

[0080] Specifically, the prototype of the ensemble empirical mode decomposition method EEMD (Ensemble Empirical Mode Decomposition) is the empirical mode decomposition (Empirical Mode Decomposition, EMD). EMD is an adaptive signal decomposition method, which can be used not only for the analysis of stationary signals but also for the analysis of non-stationary signals. After the original signal is decomposed by EMD, multiple intrinsic mode functions (Intrinsic Mode Function, IMF) can be obtained, and each IMF can ensure that after its Hilbert transform, the calculated instantaneous frequency has practical physical significance.

[0081] Due to the existence of environmental noise and observation errors, the actually obtained signal observation data are actually all "polluted" data. If the observation error is also regarded as noise, the observed signal is as shown in Equation (1):

[0082]

[0083] Among them, x(t) represents the value of the actual signal, n(t) represents the noise term, and represents the observed signal that can actually be obtained after being "polluted" by noise.

[0084] When dealing with such signals with noise, in order to make the observed results closer to the true values, the commonly used method is to observe the noisy signal multiple times and then take its average value as the observation result. The same method can be used in EMD decomposition to avoid the occurrence of mode mixing, so that each IMF obtained by decomposition can truly reflect each mode in the signal.

[0085] The basic idea of ensemble empirical mode decomposition is as follows: By artificially adding amplitude-limited Gaussian white noise to the original signal and taking the average value through multiple decompositions to achieve the decomposition of the original signal. Multiple EMDs with added white noise can make the decomposition results traverse all possible situations, while taking the average value processing can not only make the white noise cancel each other out during the summation process, but also enable the true IMF components of the original signal to be presented in the form of the average value.

[0086] Specifically, in EEMD, since the added white noise is uniformly distributed in the entire time-frequency space, and the time-frequency space is composed of different-scale components segmented by a filter bank. When the signal is added with a uniformly distributed white noise background, the signal regions and white noise of different scales will automatically map to the relevant and appropriate scales established with the background white noise. Of course, each independent test may produce a very "noisy" result because each added noise component includes the signal and the added white noise. Since the noise is different in each independent test, when using the overall average of a sufficient number of tests, according to the characteristics of white noise, the noise will be eliminated. The overall average will finally be regarded as the true result, and the only persistent and stable part is the signal itself. Many tests are added to eliminate the added noise.

[0087] The specific calculation process of EEMD is as follows:

[0088] (a) Add the random white noise n(t) to the original signal x(t). Among them, the signal x(t) is fixed, while each realization n i (t) of n(t) is random. Let the signal after adding noise for the i-th time be:

[0089]

[0090] (b) Use the EMD method to analyze x i (t) and decompose it into several IMF components;

[0091] (c) Repeat steps (a) and (b) N times;

[0092] (d) Calculate the (ensemble) average value of the decomposed IMF components as the final result, that is:

[0093]

[0094] Among them, is the j-th IMF component obtained after the original signal is decomposed by EEMD, and N is the number of repeated analyses. The following statistical law is obeyed between the standard deviation of the added noise in EEMD and the variance of the noise in the signal after taking the mean value through multiple repetitions:

[0095]

[0096] Among them, σ 2 is the variance of the added noise, is the variance of the reconstruction error of the EEMD decomposition. When the noise amplitude is certain, the more the number of repeated analyses, the smaller the variance of the added noise, and the smaller the distortion finally obtained by decomposition. For the added noise, if its variance is too small, the addition of the noise will not affect the selection of poles during the EMD decomposition, and thus lose its added role. Therefore, the variance of the added noise each time cannot be too small, and the accuracy of the final decomposition can be guaranteed through multiple repeated analyses.

[0097] Essentially, EEMD is still a signal decomposition method of layer-by-layer peeling. This method makes the signal basically not bring distortion after decomposition and reconstruction. Although the addition of noise will bring some distortion, due to the adoption of multiple superpositions and mean value processing, the influence of noise on the final separation result can be basically ignored.

[0098] After being decomposed by EEMD, each IMF component presents short-term correlation. This application will study the method of combining ANN and ARMA to predict the IMF components, and use the set mean value of each IMF component as the final prediction result.

[0099] In order to further improve the business prediction method based on the control chart, and improve the accuracy and efficiency of the business type prediction for the service traffic data of each current link of the space-based information network in the time dimension, in a business type prediction method for a space-based information network provided in an embodiment of this application, refer to Figure 3 The step 120 in the business type prediction method for the space-based information network specifically includes the following contents:

[0100] Step 121: Based on the control chart, separately perform data separation on the service traffic data of each current link of the space-based information network to obtain the non-burst data and burst data corresponding to each of the service traffic data.

[0101] Step 122: Input each of the non-burst data into a preset business type prediction model, so that the business type prediction model outputs the business type prediction data corresponding to each of the non-burst data respectively.

[0102] And, step 123: Perform self-adaptive template matching prediction on each of the burst data to obtain the service type prediction data corresponding to each of the burst data.

[0103] Step 124: Integrate the service type prediction data of the non-burst data and the burst data corresponding to each of the service traffic data to respectively obtain the second service type prediction result corresponding to each of the service traffic data.

[0104] Specifically, the biggest problem causing difficult accurate prediction of services lies in the existence of many bursts in the data. The existence of multiple bursts not only causes difficulties in predicting burst data by classical prediction models, but also causes a decrease in the prediction accuracy of non-burst data. This application will quantitatively divide the burstiness of services with the help of control chart theory, and respectively adopt the adaptive template matching method and the classical ARMA model to model burst services and non-burst services based on this division, so as to realize the prediction of services.

[0105] A control chart is a graphical tool used to indicate whether the current random process is in a controllable state. It is an important means and tool for statistical quality management. Here, controllable means that this random process is stable, and its fluctuations are only normal fluctuations caused by the inherent random causes of the process, rather than abnormal fluctuations. In the control field, when a process is in a controllable state, generally, the control parameters of this process do not need to be modified. In addition, when a process is in an uncontrollable state, if the current data is used to predict future data, the accuracy of the prediction cannot be guaranteed. Many new methods have been proposed in the field of control charts and are gradually applied to various research applications.

[0106] For actual traffic data, in some cases, burst data is significantly different from non-burst data. In addition, there are also some cases where burst data is not very easy to identify from other data. Therefore, for burst data separation, an important issue is how to define burst data. For this problem, control chart theory gives a good solution, that is, to realize the division of burst data from a statistical perspective.

[0107] Figure 4 The results of using the Individual X and Moving Range Chart (XmR) method to judge burst data are given. UCL and LCL are the upper control line and the lower control line obtained by the XmR method respectively. It can be seen that all traffic data falls above the LCL, while some data showing burstiness breaks through the UCL. This part of the data is Figure 4 marked by circles in. To further show the judgment effect of the control chart method on burst data, Figure 4The circled burst traffic values are set to the mean CL, as Figure 5 shown. It can be found from Figure 5 that extremely abrupt observed values no longer appear, and the fluctuation range of the traffic values slows down.

[0108] After completing the burst data separation process, the separated burst data is then subjected to a non-linear transformation while keeping the non-burst data unchanged, and a non-burst data set can be obtained. The separated non-burst traffic data is smoother than the original traffic data. The burst traffic data is 0 at most time nodes and has relatively similar sharp convex characteristics. Based on these characteristics of the separation results, the separated non-burst data and burst data will be predicted by an Artificial Neural Network (ANN) and an adaptive template matching method respectively. The process is as Figure 6 shown.

[0109] In order to further improve the service prediction method based on the spatial characteristics of satellite network traffic and improve the accuracy and efficiency of service type prediction in the spatial dimension for a traffic matrix jointly composed of the service traffic data of each of the said links, in a service type prediction method for a space-based information network provided in an embodiment of the present application, refer to Figure 3 , step 210 in the service type prediction method for a space-based information network specifically includes the following contents:

[0110] Step 211: Classify the traffic matrix jointly composed of the service traffic data of each of the said links to obtain corresponding multiple sub-traffic matrices.

[0111] Step 212: Based on a preset locally linear embedding non-negative matrix factorization method, perform dimensionality reduction processing on each of the said sub-traffic matrices to form corresponding low-dimensional matrices, and perform service type prediction on each of the said low-dimensional matrices. Then, determine the service type prediction data corresponding to each of the said sub-traffic matrices respectively according to the service type prediction data corresponding to each of the said low-dimensional matrices.

[0112] Step 213: Obtain the service type prediction data of the traffic matrix according to the service type prediction data corresponding to each of the said sub-traffic matrices, and use the service type prediction data of the traffic matrix as the spatial dimension service type prediction results corresponding to the service traffic data of each of the said links.

[0113] Specifically, the prediction processes of steps 110 and 120 mainly utilize the temporal correlation of traffic volume, without considering the spatial correlation among various types of traffic volume. Existing research has shown that in satellite networks, different link services are not completely independent, but there is a strong correlation, that is, spatial correlation. Step 210, on the other hand, studies the spatial correlation characteristics in space-based information network services and uses this characteristic to study the dimensionality reduction prediction method for traffic.

[0114] Most previous traffic predictions have taken single-channel traffic volume as the research object and attempted to achieve prediction by utilizing the temporal correlation of traffic data, thus ignoring the spatial correlation information among the data. Since these prediction methods need to independently model each data, the computational complexity of the prediction process is relatively high. Especially when there is a large amount of user terminal data in the satellite communication system, it will impose a great burden on the network control center. To reduce the computational overhead of the prediction process, the embodiments of this application will study a method based on locally linear preserving non-negative matrix factorization to achieve effective dimensionality reduction processing of data and perform prediction on the basis of dimensionality reduction.

[0115] If only analyzed from the perspective of dimensionality reduction and compression, the principal component analysis (PCA) method used for analyzing the spatial correlation of traffic can achieve a good compression effect on the origin-destination (OD) flow, achieving the purpose of obtaining a low-dimensional representation of traffic from high-dimensional data. However, PCA achieves dimensionality reduction of the original data by successively extracting the principal components in the direction with the largest variance change in the data. Therefore, when performing PCA processing on the traffic matrix, the inherent temporal correlation in the traffic data is not considered, resulting in the destruction of the temporal correlation structure crucial for traffic prediction during PCA processing and affecting the subsequent prediction processing.

[0116] Considering the above problems, this application proposes to use the locally linear preserving non-negative matrix factorization (LPNMF) method to achieve dimensionality reduction and compression processing of traffic data. The LPNMF method can not only effectively achieve dimensionality reduction of the matrix, but also preserve the original temporal correlation structure of the data in the low-dimensional space. In addition, after being processed by LPNMF, the original information will not exhibit the phenomenon of excessive aggregation during PCA processing in the low-dimensional space.

[0117] By using the spatial correlation between OD flows to perform dimensionality reduction and compression processing on the matrix, low-dimensional data with fewer numbers than the original OD flows can be obtained. Subsequently, through the prediction processing of the low-dimensional data and by reversing the dimensionality reduction process for the prediction results, the prediction of the original OD flows of each path can be achieved. In addition, considering that different types of service traffic may be carried in each OD flow, different service characteristics will be presented. Therefore, before performing dimensionality reduction processing, it is necessary to first perform clustering processing on each OD according to its spatial correlation. The entire compression and prediction process based on LPNMF is as Figure 7 shown.

[0118] In an actual satellite system, especially in a VPN network, the category information of OD flows can be determined at the user side according to the actual statistical results and transmitted to the NCC, thus simplifying the classification process. Specifically, the classification information can be obtained through prior information such as actual service types, such as FTP, WWW, video conferencing, etc., and empirical statistics of the empirical distribution of OD flows.

[0119] To further illustrate the implementation process of the above-mentioned service type prediction method for the space-based information network, this application also provides a specific application example of the multi-mode combined service prediction method for the space-based information network. The multi-mode combined service prediction method specifically includes the following contents:

[0120] Based on the EEMD method to well predict stationary and non-stationary services;

[0121] Based on the control chart method to better predict burst services;

[0122] At the same time, based on the method of traffic space to make full use of the regional spatial distribution characteristics of the space-based information network to predict the traffic characteristics of services.

[0123] This application combines the prediction methods of the above three modes. By comprehensively analyzing the prediction results of the three methods, accurate prediction of diversified services is achieved.

[0124] Figure 8 The multi-mode combined service prediction flow chart is given. The traffic data is respectively used to predict stationary and non-stationary signals through EEMD signal decomposition, predict burst data through burst data separation based on the control chart, and predict the spatial correlation of traffic through traffic classification and LPNMF decomposition. Then, the prediction results obtained by the three methods respectively are analyzed and judged by a service traffic and characteristic comprehensive analysis and prediction module, and finally the service prediction result is obtained.

[0125] Table 1 shows the comparison of the prediction accuracies of four methods, namely EEMD prediction, prediction based on control charts, prediction based on spatial characteristics, and multi-mode combination, for five service types: WWW, P2P, Email, Internet of Things burst services, and multimedia services. It can be seen from the table that the prediction method based on multi-mode combination has better prediction accuracy than other prediction methods under different service types.

[0126] Table 1 Prediction Accuracies (%) of Different Prediction Methods under Five Service Conditions

[0127]

[0128] That is to say, by proposing a service awareness method based on comprehensive features and machine learning, a prediction method based on ensemble empirical mode decomposition, a service prediction method based on control charts, a service prediction method based on the spatial characteristics of satellite network traffic, and a multi-mode combination service prediction method, the present application has developed a more comprehensive and general service prediction model and method. By using multiple methods for combined prediction of system services and realizing the optimal configuration and scheduling of resources based on the prediction, compared with the ground-based service prediction method based on time series and machine learning, the multi-mode service prediction modeling and prediction method for multi-service integration of space-based information networks proposed in the present application converts single-species single-mode service prediction into multiple multi-mode service predictions, enabling it to be more adaptable to the joint prediction of integrated services in space information networks, thus laying a foundation for the space information network to provide stronger comprehensive service and guarantee capabilities.

[0129] At the software level, the present application also provides a service type prediction device for space-based information networks for executing all or part of the service type prediction method for space-based information networks. See Figure 9 The service type prediction device for space-based information networks specifically includes the following:

[0130] A time dimension prediction module 10, configured to perform service type prediction in the time dimension on the service traffic data of each current link corresponding to the space-based information network respectively based on a preset multi-mode combination service prediction method, so as to obtain the time dimension service type prediction results corresponding to the service traffic data of each of the links;

[0131] And a space dimension prediction module 20, configured to perform service type prediction in the space dimension on a traffic matrix composed of the service traffic data of each of the links together based on the multi-mode combination service prediction method, so as to obtain the space dimension service type prediction results corresponding to the service traffic data of each of the links;

[0132] The prediction result generation module 30 is configured to generate target service type prediction result data corresponding to the service traffic data of each of the links according to the time - dimension service type prediction results and the space - dimension service type prediction results corresponding to the service traffic data of each of the links.

[0133] The embodiments of the service type prediction device for the space - based information network provided in this application can specifically be used to execute the processing procedures of the embodiments of the service type prediction method for the space - based information network in the above - mentioned embodiments. Its functions will not be elaborated here and can refer to the detailed description of the embodiments of the service type prediction method for the space - based information network above.

[0134] The part of the service type prediction device for the space - based information network for performing service type prediction for the space - based information network can be executed in the server or completed in the client device. Specifically, it can be selected according to the processing capabilities of the client device and the limitations of the user usage scenario, etc. This application does not make any limitations in this regard. If all operations are completed in the client device, the client device may further include a processor for specific processing of service type prediction for the space - based information network.

[0135] The above - mentioned client device may have a communication module (i.e., communication unit) and can communicate with a remote server to achieve data transmission with the server. The server may include a server on the task scheduling center side. In other implementation scenarios, it may also include a server on the intermediate platform, such as a server on a third - party server platform with a communication link to the task scheduling center server. The server may include a single computer device, or a server cluster composed of multiple servers, or a server structure of a distributed device.

[0136] Any suitable network protocol can be used for communication between the above - mentioned server and the client device end, including network protocols not yet developed on the filing date of this application. The network protocol may, for example, include TCP / IP protocol, UDP / IP protocol, HTTP protocol, HTTPS protocol, etc. Of course, the network protocol may, for example, also include RPC protocol (Remote Procedure Call Protocol) and REST protocol (Representational State Transfer) used on top of the above - mentioned protocols.

[0137] As can be seen from the above description, the service type prediction device for the space-based information network provided by the embodiments of the present application can implement the service type prediction for the space-based information network, be more adaptable to the joint prediction of the integrated services of the space information network, and effectively improve the prediction accuracy and efficiency of the service type prediction of the space-based information network. Furthermore, it can lay a foundation for the space-based information network to provide a more powerful integrated service and guarantee ability.

[0138] The embodiments of the present application also provide an electronic device, which may include a processor, a memory, a receiver, and a transmitter. The processor is configured to execute the service type prediction method for the space-based information network mentioned in the above embodiments. The processor and the memory may be connected through a bus or other means. Taking the connection through the bus as an example, the receiver may be connected to the processor and the memory in a wired or wireless manner.

[0139] The processor may be a central processing unit (CPU). The processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. chips, or a combination of the above various types of chips.

[0140] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the service type prediction method for the space-based information network in the embodiments of the present application. The processor executes various functional applications and data processing of the processor by running the non-transitory software programs, instructions, and modules stored in the memory, that is, implements the service type prediction method for the space-based information network in the above method embodiments.

[0141] The memory may include a program storage area and a data storage area. Among them, the program storage area can store an operating system and application programs required for at least one function; the data storage area can store data created by the processor and the like. In addition, the memory may include high-speed random access memory and may also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0142] The one or more modules are stored in the memory and, when executed by the processor, perform the service type prediction method for the space-based information network in the embodiments.

[0143] In some embodiments of the present application, the user equipment may include a processor, a memory, and a transceiver unit. The transceiver unit may include a receiver and a transmitter. The processor, the memory, the receiver, and the transmitter may be connected through a bus system. The memory is used to store computer instructions, and the processor is used to execute the computer instructions stored in the memory to control the transceiver unit to transmit and receive signals.

[0144] As an implementation manner, the functions of the receiver and the transmitter in the present application can be considered to be implemented through a transceiver circuit or a dedicated transceiver chip, and the processor can be considered to be implemented through a dedicated processing chip, a processing circuit, or a general-purpose chip.

[0145] As another implementation manner, it can be considered to use a general-purpose computer to implement the server provided in the embodiments of the present application. That is, the program codes for implementing the functions of the processor, the receiver, and the transmitter are stored in the memory, and the general-purpose processor implements the functions of the processor, the receiver, and the transmitter by executing the codes in the memory.

[0146] The embodiments of the present application further provide a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the foregoing service type prediction method for the space-based information network are implemented. The computer-readable storage medium may be a tangible storage medium, such as a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a floppy disk, a hard disk, a removable storage disk, a CD-ROM, or any other form of storage medium well-known in the technical field.

[0147] Those of ordinary skill in the art should understand that the various exemplary components, systems, and methods described in connection with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Specifically, whether to implement it in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application. When implemented in hardware, it can be, for example, an electronic circuit, an application-specific integrated circuit (ASIC), appropriate firmware, a plug-in, a functional card, and so on. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. The program or code segment can be stored in a machine-readable medium or transmitted through a data signal carried in a carrier wave over a transmission medium or a communication link.

[0148] It should be clear that this application is not limited to the specific configurations and processes described above and illustrated in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and illustrated as examples. However, the method process of this application is not limited to the specific steps described and illustrated. Those skilled in the art can make various changes, modifications, and additions, or change the order between steps after understanding the spirit of this application.

[0149] In this application, the features described and / or illustrated for one embodiment can be used in the same or a similar manner in one or more other embodiments, and / or combined with the features of other embodiments or replace the features of other embodiments.

[0150] The above are only the preferred embodiments of this application and are not used to limit this application. For those skilled in the art, various changes and variations can be made to the embodiments of this application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this application shall be included within the protection scope of this application.

Claims

1. A service type prediction method for space-based information networks, characterized in that, Including: Based on a preset multi-mode combined service prediction method, perform service type predictions for the service traffic data of each current link in the space-based information network in the time dimension respectively, so as to obtain the time-dimension service type prediction results corresponding to the service traffic data of each of the links; And, based on the multi-mode combined service prediction method, perform service type predictions for a traffic matrix jointly composed of the service traffic data of each of the links in the space dimension, so as to obtain the space-dimension service type prediction results corresponding to the service traffic data of each of the links; According to the time-dimension service type prediction results and the space-dimension service type prediction results corresponding to the service traffic data of each of the links, generate the target service type prediction result data corresponding to the service traffic data of each of the links.

2. The service type prediction method for the space-based information network according to claim 1, wherein The multi-mode combined service prediction method includes: an ensemble empirical mode decomposition method; the time-dimension service type prediction results include: first service type prediction results; Correspondingly, the step of performing service type predictions for the service traffic data of each current link in the space-based information network in the time dimension based on the preset multi-mode combined service prediction method to obtain the time-dimension service type prediction results corresponding to the service traffic data of each of the links includes: Adopt the ensemble empirical mode decomposition method to perform service type predictions for the service traffic data of each current link in the space-based information network in the time dimension respectively, so as to obtain the first service type prediction results corresponding to the service traffic data of each of the links.

3. The method for predicting service types for a space-based information network according to claim 2, wherein The multi-mode combined service prediction method further includes: a service prediction method based on control charts; the time-dimension service type prediction results include: second service type prediction results; Correspondingly, the step of performing service type predictions for the service traffic data of each current link in the space-based information network in the time dimension based on the preset multi-mode combined service prediction method to obtain the time-dimension service type prediction results corresponding to the service traffic data of each of the links includes: Adopt the service prediction method based on control charts to perform service type predictions for the service traffic data of each current link in the space-based information network in the time dimension respectively, so as to obtain the second service type prediction results corresponding to the service traffic data of each of the links.

4. The method for predicting service types for a space-based information network according to claim 3, wherein The multi-mode combined service prediction method further includes: a service prediction method based on the spatial characteristics of satellite network traffic; Correspondingly, the step of performing service type predictions for a traffic matrix jointly composed of the service traffic data of each of the links in the space dimension based on the multi-mode combined service prediction method to obtain the space-dimension service type prediction results corresponding to the service traffic data of each of the links includes: Adopt the service prediction method based on the spatial characteristics of satellite network traffic to perform service type predictions for a traffic matrix jointly composed of the service traffic data of each of the links in the space dimension, so as to obtain the space-dimension service type prediction results corresponding to the service traffic data of each of the links.

5. The service type prediction method for a space-based information network according to claim 2, characterized in that Using the above-mentioned ensemble empirical mode decomposition method, perform business type prediction for the service traffic data of each link corresponding to the current space-based information network in the time dimension, so as to obtain the first service type prediction results corresponding to the service traffic data of each of the links, including: Based on the ensemble empirical mode decomposition method, perform signal decomposition on the service traffic data of each link corresponding to the current space-based information network respectively, so as to obtain the intrinsic mode function component groups corresponding to the service traffic data of each of the links, where each of the intrinsic mode function component groups contains multiple intrinsic mode function components; Input each of the intrinsic mode function components in each of the intrinsic mode function component groups into a preset service type prediction model respectively, so that the service type prediction model outputs the service type prediction data corresponding to each of the intrinsic mode function components respectively; Determine the first service type prediction results corresponding to the service traffic data of each of the links respectively according to the mean values among the service type prediction data of the intrinsic mode function components in each of the intrinsic mode function component groups.

6. The method for predicting service types for a space-based information network according to claim 3, wherein Using the above-mentioned service prediction method based on control charts, perform business type prediction for the service traffic data of each link corresponding to the current space-based information network in the time dimension, so as to obtain the second service type prediction results corresponding to the service traffic data of each of the links, including: Based on the control chart, perform data separation on the service traffic data of each link corresponding to the current space-based information network respectively, so as to obtain the non-burst data and burst data corresponding to the service traffic data of each of the links; Input each of the non-burst data into a preset service type prediction model respectively, so that the service type prediction model outputs the service type prediction data corresponding to each of the non-burst data respectively; And perform self-adaptive template matching prediction on each of the burst data respectively, so as to obtain the service type prediction data corresponding to each of the burst data; Integrate the service type prediction data of the non-burst data and burst data corresponding to the service traffic data of each of the links respectively, so as to obtain the second service type prediction results corresponding to the service traffic data of each of the links respectively.

7. The method for predicting service types for a space-based information network according to claim 4, wherein Using the above-mentioned service prediction method based on the spatial characteristics of satellite network traffic, perform business type prediction in the spatial dimension on a traffic matrix composed of the service traffic data of each of the links, so as to obtain the spatial dimension service type prediction results corresponding to the service traffic data of each of the links, including: Perform traffic classification on a traffic matrix composed of the service traffic data of each of the links, so as to obtain corresponding multiple sub-traffic matrices; Based on the preset locally linear embedding non-negative matrix factorization method, perform dimensionality reduction processing on each of the sub-traffic matrices respectively to form corresponding low-dimensional matrices, perform business type prediction on each of the low-dimensional matrices, and then determine the service type prediction data corresponding to each of the sub-traffic matrices respectively according to the service type prediction data corresponding to each of the low-dimensional matrices; Obtain the service type prediction data of the traffic matrix according to the service type prediction data corresponding to each of the sub-traffic matrices, and use the service type prediction data of the traffic matrix as the spatial dimension service type prediction results corresponding to the service traffic data of each of the links.

8. A service type prediction device for a space-based information network, characterized in that, Including: A time dimension prediction module, configured to perform time dimension service type prediction on the service traffic data of each link currently corresponding to the space-based information network based on a preset multi-mode combined service prediction method, so as to obtain the time dimension service type prediction results corresponding to the service traffic data of each of the links; And, a spatial dimension prediction module, configured to perform spatial dimension service type prediction on a traffic matrix jointly composed of the service traffic data of each of the links based on the multi-mode combined service prediction method, so as to obtain the spatial dimension service type prediction results corresponding to the service traffic data of each of the links; A prediction result generation module, configured to generate target service type prediction result data corresponding to the service traffic data of each of the links according to the time dimension service type prediction results and the spatial dimension service type prediction results corresponding to the service traffic data of each of the links.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the service type prediction method for the space-based information network according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the service type prediction method for the space-based information network according to any one of claims 1 to 7.

Images