Time delay prediction method, communication optimization method and system
By convolution and cyclic layer processing of the historical delay information of user equipment, combined with attention mechanism and autoregression model, the delay is accurately predicted, and the problem of inaccurate delay budgets between IMS and GEO satellites is solved, and the call is improved Quality and user satisfaction.
Patent Information
- Application Number
- CN202411155338.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-22
- Publication Date
- 2024-12-03
AI Technical Summary
Within the 3GPP framework, when IMS and GEO satellites are connected, the one-way transmission delay significantly affects the R score in the E-Model, resulting in a decline in call quality. The existing technology cannot accurately calculate the delay budget, especially due to the movement of users and low-orbit satellites. sexual interference.
By obtaining the ID information and historical delay information of the user equipment, performing convolution processing to extract short-term features, combining the recurrent layer processing to determine long-term features, using attention mechanism and autoregressive model to fusion information, accurately predict the delay of the user equipment, and The core network side sets up the service link delay budget to optimize communication.
Improves the accuracy of delay prediction, improves call quality, ensures user satisfaction and R scores of network design.
Smart Images

Figure CN119071798A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of 5G communication technology, and in particular to a delay prediction method, a communication optimization method and a system. Background Art
[0002] In the 3GPP framework, voice calls using the IMS platform have become a standard feature; for IMS and GEO satellite access, one-way transmission delay significantly affects the R score in the E-Model, which is used to evaluate the quality of voice calls. Managing this delay is key to achieving a higher R score in network design. Due to the mobility of users and low-orbit satellites (which may cause interference), related technologies cannot accurately calculate the delay budget, which in turn affects call quality. Summary of the invention
[0003] The purpose of the present invention is to solve one of the technical problems existing in the prior art to at least a certain extent.
[0004] To this end, an object of the present invention is to provide a highly accurate delay prediction method, communication optimization method and system.
[0005] In order to achieve the above technical objectives, one aspect of an embodiment of the present invention provides a delay prediction method, comprising the following steps: obtaining user device information; the user device information includes the ID information of the user device and the historical delay information related to the user device; performing convolution processing on the user device information to determine short-term information; the short-term information includes the short-term characteristics of the delay; processing the short-term information through a loop layer to determine long-term information; the long-term information includes the long-term characteristics of the delay; and determining the delay corresponding to the user device based on the long-term information. The embodiment of the present application processes the historical delay information of the user device through convolution and loop layer processing to accurately predict the delay of the user device. The embodiment of the present application is conducive to improving the accuracy of delay prediction and improving call quality.
[0006] In some embodiments, in the delay prediction method of the embodiment of the present invention, the short-term information is processed through a loop layer to determine the long-term information:
[0007] Processing the short-term information through a cyclic unit to determine first information;
[0008] Processing the short-term information through a cyclic skip unit to determine second information;
[0009] The first information and the second information are fused to determine long-term information.
[0010] In some embodiments, in an embodiment of the present invention, determining the delay corresponding to the user equipment according to the long-term information includes:
[0011] The long-term information is processed through an attention mechanism to determine hidden information;
[0012] Performing autoregressive model processing on the user equipment information to determine regression information;
[0013] The hidden information and the regression information are fused to determine a time delay.
[0014] In some embodiments, in one embodiment of the present invention, the method further comprises:
[0015] Determining input dimension information of the model input information based on the number of user devices and the duration corresponding to the historical delay information;
[0016] According to the input dimension, the output dimension information is matched.
[0017] In some embodiments, in one embodiment of the present invention, the method further comprises:
[0018] Acquire a user equipment sample; the user equipment sample includes a real result of a delay corresponding to the user equipment;
[0019] Processing the user equipment sample through convolution and recurrent layers to determine a prediction result;
[0020] The predicted result and the true result are subjected to CORR to calculate the loss value, and the parameters of the convolution and circulation layers are updated according to the loss value.
[0021] On the other hand, an embodiment of the present application provides a communication optimization method, including:
[0022] Obtaining an access request from a user device;
[0023] If the satellite type corresponding to the access request is a geosynchronous orbit, a first delay of a preset time is determined by the delay prediction method as described above on the core network side;
[0024] The service link delay budget between the user equipment and the satellite is set to the first delay.
[0025] In some embodiments, in an embodiment of the present invention, the core network side determines the first delay by the following steps:
[0026] If the satellite type corresponding to the access request is a geosynchronous orbit, sending a request to a unified data management network element through an access and mobility management function network element to obtain historical delay information related to the user equipment;
[0027] A request is sent to a network data analysis function network element through an access and mobility management function network element, so that the network data analysis function network element calculates the first delay.
[0028] On the other hand, an embodiment of the present invention provides a delay prediction system, including:
[0029] The first module is used to obtain user equipment information; the user equipment information includes ID information of the user equipment and historical delay information related to the user equipment;
[0030] The second module is used to perform convolution processing on the user equipment information to determine short-term information; the short-term information includes short-term characteristics of delay;
[0031] The third module is used to process the short-term information through a cyclic layer to determine long-term information; the long-term information includes long-term characteristics of the delay;
[0032] The fourth module is used to determine the delay corresponding to the user equipment according to the long-term information.
[0033] On the other hand, an embodiment of the present invention provides a delay prediction device, including:
[0034] at least one processor;
[0035] at least one memory for storing at least one program;
[0036] When the at least one program is executed by the at least one processor, the at least one processor implements the above-mentioned delay prediction method.
[0037] On the other hand, an embodiment of the present invention provides a storage medium, in which a program executable by a processor is stored. When the program executable by the processor is executed by the processor, it is used to implement the above-mentioned delay prediction method.
[0038] The method provided in the embodiment of the present application includes: obtaining user device information; the user device information includes the ID information of the user device and the historical delay information related to the user device; performing convolution processing on the user device information to determine short-term information; the short-term information includes the short-term characteristics of the delay; the short-term information is processed through a loop layer to determine long-term information; the long-term information includes the long-term characteristics of the delay; according to the long-term information, the delay corresponding to the user device is determined. The embodiment of the present application processes the historical delay information of the user device through convolution and loop layer processing to accurately predict the delay of the user device. The embodiment of the present application is conducive to improving the accuracy of delay prediction and improving call quality. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the embodiments of the present invention or the drawings of related technical solutions in the prior art are introduced below. It should be understood that the drawings introduced below are only for the convenience of clearly describing some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.
[0040] Figure 1 A schematic diagram of a flow chart of an embodiment of a delay prediction method provided by the present invention;
[0041] Figure 2 A schematic diagram of a flow chart of another embodiment of the delay prediction method provided by the present invention;
[0042] Figure 3 A schematic diagram of the time delay of a user equipment of an embodiment of the multi-step prediction provided by the present invention;
[0043] Figure 4 A schematic diagram of the structure of an embodiment of the delay prediction system provided by the present invention;
[0044] Figure 5 A schematic diagram of the structure of an embodiment of the delay prediction device provided by the present invention. DETAILED DESCRIPTION
[0045] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and are not to be construed as limitations of the present invention. For the step numbers in the following embodiments, they are only provided for the convenience of explanation, and the order between the steps is not limited in any way, and the execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.
[0046] First, the technical terms involved in the embodiments of the present application are explained:
[0047] Access and Mobility Management Function (AMF);
[0048] 3rd Generation Partnership Project (3GPP);
[0049] User Equipment (UE);
[0050] Radio Access Network (RAN);
[0051] Unified Data Management (UDM);
[0052] Network Data Analytics Function (NWDAF);
[0053] High orbit / geosynchronous orbit (Geostationary Earth Orbit, GEO);
[0054] IP Multimedia Subsystem (IMS);
[0055] Session Management Function (SMF);
[0056] Policy Control Function (PCF);
[0057] The Proxy-Call Session Control Function (P-CSCF) is an important network element in the IP Multimedia Subsystem (IMS). It is mainly responsible for processing the signaling interaction between the terminal and the network, and plays the role of proxy and forwarding;
[0058] Auto Regressive (AR).
[0059] Using general-purpose mobile phones to make voice calls via satellite access is becoming increasingly common, especially due to satellite's advantages in providing coverage in rural and remote areas. The satellite voice call service in 3GPP provides a consistent network connection that can be accessed at any time and place, facilitating smooth communication between terrestrial and satellite networks.
[0060] Starting from Release 17 (Rel-17), GEO satellite access has been included in the 3GPP standard as a 5G access technology, and all types of media such as voice, data, and video are supported by default. However, due to the unique challenges of GEO satellites, such as a signal delay of about 285ms at a distance of 35,786 kilometers from the earth and atmospheric attenuation, the data rate cannot be increased simply by expanding the bandwidth, which means that the services supported by non-geostationary satellite orbits (NGSO) may not be supported on GEO satellites.
[0061] Voice calls using the IMS platform have been a standard feature since 3GPP Release 5 (Rel-5). Since IMS is able to support a variety of multimedia services and ensure interoperability, several services have been added subsequently, including IMS emergency calls, messaging, group management, push-to-talk, and real-time communications. These enhancements make IMS a key tool for connecting different operators and service providers for voice calls across various types of access networks. However, when integrating IMS with GEO satellite access within the 3GPP framework, the reasons that affect the quality of experience include: one-way transmission delay.
[0062] One-way propagation delay significantly affects the R score in the E-Model, which is used to evaluate the quality of voice calls. Managing this delay is key to achieving higher R scores in network design. ITU-T recommends a maximum delay of 400 milliseconds for network planning. However, when using GEO satellite access, the propagation delay is much longer than other technologies (285ms), and a careful calculation of the delay budget is required. Due to the mobility of users and low-orbit satellites (which may cause interference), existing technologies cannot accurately calculate the delay budget, which is critical to manage, and an accurate budget can ensure higher user satisfaction when adjusting IMS voice calls to GEO satellite access.
[0063] In this regard, an embodiment of the present application proposes a solution for optimizing IMS communications based on high-orbit satellite access, which provides management with accurate predictions of delay budgets by analyzing the historical real delays collected between UE and GEO satellites to ensure user satisfaction.
[0064] The delay prediction method and system proposed in the embodiments of the present invention are described in detail below with reference to the accompanying drawings. First, the delay prediction method proposed in the embodiments of the present invention will be described with reference to the accompanying drawings.
[0065] Reference Figure 1 , a delay prediction method is provided in an embodiment of the present invention. The delay prediction method in the embodiment of the present invention can be applied to a terminal, can be applied to a server, can also be software running in a terminal or a server, etc. The terminal can be a tablet computer, a laptop computer, a desktop computer, etc., but is not limited to this. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The delay prediction method in the embodiment of the present invention mainly includes the following steps:
[0066] S100: Acquire user equipment information; the user equipment information includes ID information of the user equipment and historical delay information related to the user equipment;
[0067] S200: performing convolution processing on the user equipment information to determine short-term information; the short-term information includes short-term characteristics of the delay;
[0068] S300: Process the short-term information through the loop layer to determine the long-term information; the long-term information includes the long-term characteristics of the delay;
[0069] S400: Determine a delay corresponding to the user equipment according to the long-term information.
[0070] In some possible implementations, reference Figure 2 As shown, the embodiment of the present application extracts short-term delay information from historical delay information through convolution layer processing; and determines long-term delay information from historical delay information through the loop layer. It is understandable that, for call scenarios, the call density within 24 hours may be affected by people's work and life; and the call density within one month or half a year may be affected by relevant holidays. Of course, the above are illustrative examples. For different influencing parameters / regions of calls, the call delay can be predicted by the delay prediction method provided in the embodiment of the present application. It is understandable that the embodiment of the present application can provide a delay prediction model, and the delay prediction model includes a convolution layer and a loop layer, which provides user equipment information for delay prediction.
[0071] Optionally, in one embodiment of the present invention, the short-term information is processed through a recurrent layer to determine the long-term information:
[0072] Processing the short-term information through the recurrent unit to determine the first information;
[0073] Processing the short-term information through a cyclic skip unit to determine second information;
[0074] The first information and the second information are fused to determine long-term information.
[0075] In some possible implementations, the cyclic skip unit is used to add a skip link between the current hidden unit and the hidden unit of the same phase to extract long-term information. In some embodiments, the cyclic unit in the embodiment of the present application can be GRU and LSTM, which is not specifically limited in the present application.
[0076] Optionally, in an embodiment of the present invention, determining the delay corresponding to the user equipment according to the long-term information includes:
[0077] Long-term information is processed through the attention mechanism to determine hidden information;
[0078] Performing autoregressive model processing on user equipment information to determine regression information;
[0079] The hidden information and the regression information are fused to determine the delay.
[0080] Optionally, in one embodiment of the present invention, the method further comprises:
[0081] Determine input dimension information of model input information based on the number of user devices and the duration corresponding to the historical delay information;
[0082] According to the input dimension, match the output dimension information.
[0083] In some possible implementations, the embodiments of the present application provide a multiple prediction method to adapt the output dimension.
[0084] Optionally, in one embodiment of the present invention, the method further comprises:
[0085] Obtaining a user equipment sample; the user equipment sample includes a real result of a delay corresponding to the user equipment;
[0086] Process the user device samples through convolution and recurrent layers to determine the prediction results;
[0087] The predicted results and the actual results are used to calculate the loss value using CORR, and the parameters of the convolution and circulation layers are updated according to the loss value.
[0088] In some possible implementations, the embodiments of the present application may update the delay prediction model through CORR or RSE to improve the accuracy of the delay prediction model.
[0089] On the other hand, an embodiment of the present application provides a communication optimization method, including:
[0090] Obtaining an access request from a user device;
[0091] If the satellite type corresponding to the access request is a geosynchronous orbit, a first delay of a preset time is determined by the delay prediction method as described above on the core network side;
[0092] The service link delay budget between the user equipment and the satellite is set to the first delay.
[0093] In some possible implementations, the embodiments of the present application may deploy the delay prediction model in the network data analysis function network element of the core network, and deploy the user equipment information in the unified data management network element of the core network.
[0094] Optionally, in one embodiment of the present invention, the core network side determines the first delay by the following steps:
[0095] If the satellite type corresponding to the access request is a geosynchronous orbit, a request is sent to a unified data management network element through an access and mobility management function network element to obtain historical delay information related to the user equipment;
[0096] A request is sent to the network data analysis function network element through the access and mobility management function network element, so that the network data analysis function network element calculates the first delay.
[0097] In some possible implementations, a registration request is sent to a session management function network element through an access and mobility management function network element; the session management function network element forwards the registration request to a policy control function network element, and the policy control function network element sends a first delay to a proxy call session control function network element of the IP multimedia subsystem, so that the proxy call session control function network element of the IP multimedia subsystem sets the service link delay budget between the user equipment and the satellite to the first delay.
[0098] The following is a detailed introduction to the delay prediction method and communication optimization method involved in the embodiment of the present application using a specific embodiment:
[0099] First, the delay prediction model proposed in the embodiment of the present application is introduced. Figure 2 As shown in Figure 3, the building blocks of the model include: convolutional layer, recurrent layer, recurrent skip layer, temporal attention layer, and autoregressive layer.
[0100] Convolutional layer: The first layer is a convolutional layer without pooling. Its main purpose is to extract short-term patterns in the time dimension and local dependencies of delays between different UE IDs. The convolutional layer consists of multiple filters with width ω and height n (the height is the same as the number of variables). The kth filter sweeps over the input matrix X and produces h k , where * represents the convolution operation, the output hk will be a vector, and RELU is the activation function. The size of the output matrix of the convolution layer is d c ×T, where d c Indicates the number of filters.
[0101] h k =RELU(W k *X+b k )
[0102] Recurrent layer: The output of the convolutional layer is used as the input of both the recurrent unit and the recurrent skip unit. The recurrent unit is a recurrent layer with a gated recurrent unit (GRU) and uses the RELU function as the hidden update activation function. Where ⊙ is the product of the elements, σ is the sigmoid function, and x i is the input of the layer at time t, and the output of the layer is the hidden state at each time. Based on experience, we have found that RELU as an activation function can provide more reliable performance, making the gradient easier to back propagate.
[0103] r t =σ(x t W xr +h t-1 W hr +b r )
[0104] u t =σ(x t W xu +h t-1 W hu +b u )
[0105] c t =RELU(x t W xc +r t ⊙(h t-1 W hc )+b c )
[0106] h t =(1-u)⊙h t-1 +u t ⊙c t
[0107] Recurrent skip layer: By adding skip links between the current hidden unit and the hidden unit of the same phase, it alleviates the problem of GRU and LSTM being unable to capture very long-term correlations caused by gradient disappearance.
[0108] r t =σ(x t W xr +h t-p W hr +b r )
[0109] u t =σ(x t W xu +h t-p W hu +b u )
[0110] c t =RELU(x t W xc +r t ⊙(h t-p W hc )+b c )
[0111] h t =(1-u)⊙h t-p +u t ⊙ct
[0112] The input of this layer is the output of the convolutional layer, and p is the number of hidden units skipped, where p is 24. Then, the output of the recurrent unit and the recurrent skip unit are integrated through a fully connected layer. The input of the fully connected layer includes the hidden state of the recurrent unit at time t and the hidden states of p (from t-p+1 to t) recurrent skip units. The output calculation process of the full connection is:
[0113]
[0114] where h D t is the predicted delay value of the neural network at time t.
[0115] Temporal Attention Layer: An attention mechanism that learns a weighted combination of the hidden representations at each window position of the input matrix. The attention weight α at the current time t is t ∈R q Calculated as:
[0116]
[0117] Among them, H R t =[h R t-q ,...,h R t-1 ] is a matrix that stacks the hidden representations of the RNN in columns, and AttnScore is some similarity function, such as dot product, cosine, or parameterized by a simple multilayer perceptron. The final output of the temporal attention layer is the weighted context vector c t =H t α t Hide the last window R t-1 The concatenation of , and the linear projection operator.
[0118]
[0119] Autoregressive layer: The prediction result of the AR component is represented as hLt∈Rn, and the coefficient of the AR model is represented as W ar ∈R q_ar and b ar ∈R, where qar is the size of the input window on the input matrix. The autoregressive model is as follows:
[0120]
[0121] The LSTNet model integrates the output of the neural network and the autoregressive model as the final output:
[0122]
[0123] where Y ^ t is the final prediction result of the model at time t.
[0124] There are two scoring criteria, namely RSE and CORR, which are calculated as follows:
[0125]
[0126] Next, the multi-step prediction provided by the embodiment of the present application is introduced:
[0127] In order to predict the latency of multiple time steps at a time, we need to adapt the output dimension. The input X is 3-dimensional (X[n,P,m], n is the batch size, P is the time series window, that is, how long the historical data is needed to predict the future data, and m is the width of the data, that is, how many columns are in a row). Then we need to shape the output Y into 3 dimensions. Suppose there are 5 UEs, and the latency values of the 5 UEs are counted every 10 minutes, for a total of 30 days (each row is the latency of a time step, and each column is the latency of a UE). Refer to Figure 3 As shown, the X input size is [128, 144, 5], there are 4320 data in total, and it is necessary to predict the next five time steps, so the output Y is [128, 5, 5].
[0128] Then, the IMS communication optimization process for high-orbit satellite access provided by this application is introduced:
[0129] The trained prediction model is deployed in the NWDAF network element on the ground core network side. The historical delay data of the service link corresponding to the UE ID is stored in the UDM network element on the core network side. When the UE registers for the IMS service and specifies access by the GEO satellite, the following process is triggered:
[0130] S21. The UE sends an access request to the base station (GEO satellite or ground base station), carrying UE ID information and access satellite type requirements;
[0131] S22. The base station receives the request message from the UE and forwards the request to the AMF on the core network side;
[0132] S23. After receiving the request, the AMF on the core network side checks whether the access satellite type is GEO. If so, it sends a request to the UDM to obtain the historical delay data of the service link corresponding to the UE ID;
[0133] S24. After receiving the acquisition request, UDM returns the result R to AMF;
[0134] S25. AMF sends a delay budget prediction request to NWDAF, carrying the result R obtained from UDM.
[0135] S26. When the prediction model in NWDAF receives the prediction request, it uses the historical delay data of result R as input to predict the delay D at time t in the future, and returns the delay D to AMF.
[0136] S27. AMF sends the UE registration service request to SMF, including UE ID and delay D information;
[0137] S28. After receiving the request, SMF forwards it to PCF;
[0138] S29. After receiving the request, the PCF sends the delay D at the future time t to the P-CSCF of the IMS Core.
[0139] S210. After receiving the delay, the P-CSCF sets the service link delay budget between the UE and the satellite to D.
[0140] In this way, the accuracy of the service link delay budget in the IMS can be improved, so that when the user makes a voice call, there will be no impact (such as call interruption, reduced call quality, etc.) due to the delay not being consistent with the delay budget in the IMS.
[0141] The embodiment of the present application proposes an IMS optimization method based on high-orbit satellite access, constructs and applies a delay prediction model based on the LSTNet network, and outputs adaptation of multiple time-step predictions of delay. The embodiment of the present application can optimize the call quality degradation caused by the delay not being consistent with the delay budget in the IMS; can obtain higher user satisfaction; and can predict the delay of multiple time steps at a time.
[0142] In summary, the method provided by the embodiment of the present application includes: obtaining user device information; the user device information includes the ID information of the user device and the historical delay information related to the user device; performing convolution processing on the user device information to determine short-term information; the short-term information includes the short-term characteristics of the delay; processing the short-term information through a loop layer to determine the long-term information; the long-term information includes the long-term characteristics of the delay; and determining the delay corresponding to the user device based on the long-term information. The embodiment of the present application processes the historical delay information of the user device through convolution and loop layer processing to accurately predict the delay of the user device. The embodiment of the present application is conducive to improving the accuracy of delay prediction and improving call quality.
[0143] Secondly, refer to the attached Figure 4 A delay prediction system proposed according to an embodiment of the present invention is described.
[0144] Figure 4: is a schematic diagram of the structure of a delay prediction system according to an embodiment of the present invention, and the system specifically includes:
[0145] The first module 410 is used to obtain user equipment information; the user equipment information includes ID information of the user equipment and historical delay information related to the user equipment;
[0146] The second module 420 is used to perform convolution processing on the user equipment information to determine short-term information; the short-term information includes short-term characteristics of the delay;
[0147] The third module 430 is used to process the short-term information through the recurrent layer to determine the long-term information; the long-term information includes the long-term characteristics of the delay;
[0148] The fourth module 440 is used to determine the delay corresponding to the user equipment according to the long-term information.
[0149] Optionally, in one embodiment of the present invention, the system further includes a fifth module, configured to determine input dimension information of the model input information based on the number of user devices and the duration corresponding to the historical delay information;
[0150] According to the input dimension, match the output dimension information.
[0151] Optionally, in one embodiment of the present invention, the system further includes a sixth module, configured to obtain a user equipment sample; the user equipment sample includes a real result of a delay corresponding to the user equipment;
[0152] Process the user device samples through convolution and recurrent layers to determine the prediction results;
[0153] The predicted results and the actual results are used to calculate the loss value using CORR, and the parameters of the convolution and circulation layers are updated according to the loss value.
[0154] It can be seen that the contents of the above method embodiments are all applicable to the present system embodiments, the functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.
[0155] Reference Figure 5 , an embodiment of the present invention provides a delay prediction device, comprising:
[0156] at least one processor 510;
[0157] At least one memory 520, used to store at least one program;
[0158] When the at least one program is executed by the at least one processor 510, the at least one processor 510 implements the delay prediction method.
[0159] Similarly, the contents of the above method embodiments are all applicable to the present device embodiments. The functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.
[0160] An embodiment of the present invention further provides a computer-readable storage medium, in which a program executable by a processor is stored. The program executable by the processor is used to execute the above-mentioned delay prediction method when executed by the processor.
[0161] Similarly, the contents of the above method embodiments are all applicable to the present storage medium embodiments. The functions specifically implemented by the present storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.
[0162] In some selectable embodiments, the function / operation mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the function / operation involved, the two boxes shown in succession can actually be executed substantially simultaneously or the boxes can sometimes be executed in reverse order. In addition, the embodiment presented and described in the flow chart of the present invention is provided by way of example, for the purpose of providing a more comprehensive understanding of technology. The disclosed method is not limited to the operation and logic flow presented herein. Selectable embodiments are expected, wherein the order of various operations is changed and the sub-operation of a part for which is described as a larger operation is performed independently.
[0163] In addition, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise specified, one or more of the functions and / or features can be integrated into a single physical device and / or software module, or one or more functions and / or features can be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the present invention. More specifically, in view of the properties, functions and internal relationships of the various functional modules in the device disclosed herein, the actual implementation of the module will be understood within the conventional skills of the engineer. Therefore, those skilled in the art can implement the present invention set forth in the claims without excessive experimentation using ordinary techniques. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.
[0164] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several programs to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.
[0165] The logic and / or steps represented in the flowchart or otherwise described herein, for example, may be considered as an ordered list of executable programs for implementing the logical functions, and may be embodied in any computer-readable medium for use by a program execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch and execute a program from a program execution system, device or apparatus), or in conjunction with such program execution systems, devices or apparatuses. For purposes of this specification, a "computer-readable medium" may be any device that can contain, store, communicate, propagate or transmit a program for use by a program execution system, device or apparatus, or in conjunction with such program execution systems, devices or apparatuses.
[0166] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.
[0167] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable program execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0168] In the above description of this specification, the description with reference to the terms "one embodiment / example", "another embodiment / example" or "certain embodiments / examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.
[0169] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.
[0170] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of the present invention.
Claims
1. A delay prediction method, It is characterized in that The following steps are involved: Acquire user equipment information; the user equipment information includes ID information of the user equipment and historical delay information related to the user equipment; Performing convolution processing on the user equipment information to determine short-term information; the short-term information includes short-term characteristics of the delay; Processing the short-term information through a recurrent layer to determine long-term information; the long-term information includes long-term characteristics of the time delay; A delay corresponding to the user equipment is determined according to the long-term information.
2. The delay prediction method according to claim 1, It is characterized in that The short-term information is processed through a recurrent layer to determine the long-term information: Processing the short-term information through a cyclic unit to determine first information; Processing the short-term information through a cyclic skip unit to determine second information; The first information and the second information are fused to determine long-term information.
3. The delay prediction method according to claim 2, It is characterized in that The determining, according to the long-term information, a delay corresponding to the user equipment includes: The long-term information is processed through an attention mechanism to determine hidden information; Performing autoregressive model processing on the user equipment information to determine regression information; The hidden information and the regression information are fused to determine a time delay.
4. The delay prediction method according to claim 1, It is characterized in that The method further comprises: Determining input dimension information of the model input information based on the number of user devices and the duration corresponding to the historical delay information; According to the input dimension, the output dimension information is matched.
5. The delay prediction method according to claim 1, It is characterized in that The method further comprises: Acquire a user equipment sample; the user equipment sample includes a real result of a delay corresponding to the user equipment; Processing the user equipment sample through convolution and recurrent layers to determine a prediction result; The predicted result and the true result are subjected to CORR to calculate the loss value, and the parameters of the convolution and circulation layers are updated according to the loss value.
6. A communication optimization method, It is characterized in that The optimization method comprises: Obtaining an access request from a user device; If the satellite type corresponding to the access request is a geosynchronous orbit, determining a first delay of a preset time by a delay prediction method provided by any one of claims 1 to 5 on the core network side; The service link delay budget between the user equipment and the satellite is set to the first delay.
7. The communication optimization method according to claim 6, It is characterized in that The core network side determines the first delay by the following steps: If the satellite type corresponding to the access request is a geosynchronous orbit, sending a request to a unified data management network element through an access and mobility management function network element to obtain historical delay information related to the user equipment; A request is sent to a network data analysis function network element through an access and mobility management function network element, so that the network data analysis function network element calculates the first delay.
8. A delay prediction system, It is characterized in that include: The first module is used to obtain user equipment information; the user equipment information includes ID information of the user equipment and historical delay information related to the user equipment; The second module is used to perform convolution processing on the user equipment information to determine short-term information; the short-term information includes short-term characteristics of delay; The third module is used to process the short-term information through a cyclic layer to determine long-term information; the long-term information includes long-term characteristics of the delay; The fourth module is used to determine the delay corresponding to the user equipment according to the long-term information.
9. A delay prediction device, It is characterized in that include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the delay prediction method according to any one of claims 1 to 5.
10. A computer-readable storage medium having a program executable by a processor stored therein, It is characterized in that The program executable by the processor is used to implement the delay prediction method according to any one of claims 1 to 5 when executed by the processor.
