Mobile switching method and device based on federated learning, electronic equipment and medium

Abstract

The application discloses a mobile switching method and device based on federated learning, electronic equipment and medium. Through the application of the technical solution, the future moving track of the terminal can be predicted based on the real-time position information of the terminal by the base station currently accessed by the terminal, and the appropriate base station for the terminal is selected by using the moving track and the mobile prediction model obtained by collaborative training under the federated learning framework in the subsequent, and the base station switching for the terminal is automatically performed. Further, the model training of the mobile prediction is achieved by the local data, the model parameters are uploaded to the upper base station for aggregation, and the user track data does not need to be continuously shared between the base stations. At the same time, the problem that the terminal service progress is easily affected due to the insufficient switching management method between the base stations in the related art is avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to data processing technology, and in particular to a mobile switching method and device based on federated learning, electronic equipment and medium. BACKGROUND

[0002] With the rapid growth of the number of user equipment and traffic demand, wireless cellular networks are developing towards denser deployment, with an increase in the number of cells, a decrease in cell radius, and enhanced user mobility, which leads to more frequent and costly mobile switching and more complex mobility management problems.

[0003] Current traditional mobility management solutions are mainly designed for macro cells and cells that do not frequently switch, and are difficult to provide reliable switching performance for medium and high-speed mobile users in ultra-dense networks (UDN) to ensure service continuity and user quality of service.

[0004] Current researchers introduce user mobility prediction based on artificial intelligence for intelligent mobility management, but since each base station can only obtain local user mobility trajectory information, good mobility prediction performance depends on the cooperation of base stations to continuously share user mobility trajectory data. However, a large number of micro base stations (including pico base stations, fly base stations, etc.) are deployed wirelessly, and the data transmission capability between base stations is limited, so the switching performance of the intelligent mobility management method based on mobility prediction, such as switching delay, is difficult to guarantee. SUMMARY

[0005] The embodiments of the present application provide a mobile switching method and device based on federated learning, electronic equipment and medium. To solve the problem that the switching management method between base stations in the related art is insufficient and easily affects the progress of terminal services.

[0006] According to an aspect of the embodiments of the present application, a mobile switching method based on federated learning is provided, applied to a source base station, comprising:

[0007] receiving a configuration report reported by a terminal connected to itself, the configuration report comprising current position information of the terminal and channel state information of other base stations;

[0008] determining a future mobility trajectory of the terminal using the current position information and a preset target mobility prediction model, the target mobility prediction model being a model obtained by the source base station through collaborative training under a federated learning framework;

[0009] selecting a target base station for the terminal based on the future mobility trajectory of the terminal and the channel state information of the other base stations;

[0010] switching the access base station of the terminal to the target base station.

[0011] Optionally, in another embodiment based on the above method of the application, before the obtaining of the configuration report reported by the terminal accessing the self, further comprising:

[0012] receiving an initial configuration report reported by the terminal accessing the self, the initial configuration report comprising current position information of the terminal and channel state information of other base stations;

[0013] training an initial mobile prediction model by using the initial configuration report to obtain initial model parameters;

[0014] obtaining the target mobile prediction model by using the initial model parameters.

[0015] Optionally, in another embodiment based on the above method of the application, the obtaining of the target mobile prediction model by using the initial model parameters comprises:

[0016] sending the initial model parameters to a macro base station associated with the source base station, so that the macro base station and the edge device federatedly aggregate each initial model parameter received to obtain global model parameters;

[0017] receiving the global model parameters sent by the macro base station, and obtaining the target mobile prediction model based on the global model parameters.

[0018] Optionally, in another embodiment based on the above method of the application, the obtaining of the target mobile prediction model based on the global model parameters comprises:

[0019] performing a preset model accuracy test on the initial mobile prediction model by using the global model parameters;

[0020] if it is determined that the initial mobile prediction model meets the preset model accuracy, taking the initial mobile prediction model as the target mobile prediction model; or,

[0021] if it is determined that the initial mobile prediction model does not meet the preset model accuracy, training the initial mobile prediction model by using the global model parameters to obtain the target mobile prediction model.

[0022] Optionally, in another embodiment based on the above method of the application, the switching of the access base station of the terminal to the target base station comprises:

[0023] sending a switching request to the target base station;

[0024] If the target base station sends a handover agreement message, a reconfiguration message is sent to the terminal, and the reconfiguration message includes the handover execution condition and configuration parameters of the target base station.

[0025] When it is determined that the terminal performs a handover process with the target base station, the access state with the terminal is canceled.

[0026] Optionally, in another embodiment based on the above-mentioned method of the application, after the reconfiguration message is sent to the terminal, the method further includes:

[0027] A service message is sent to the target base station, and the service message includes attribute information and SN state information of the terminal.

[0028] Optionally, in another embodiment based on the above-mentioned method of the application, the source base station and the target base station are at least one base station in a micro cell base station cluster.

[0029] According to another aspect of the embodiments of the application, a mobile handover device based on federated learning is provided, which is applied to a source base station and includes:

[0030] The receiving module is configured to receive a configuration report reported by a terminal accessing the source base station, and the configuration report includes current location information of the terminal and channel state information of other base stations;

[0031] The determining module is configured to determine a future moving track of the terminal by using the current location information and a preset target moving prediction model, and the target moving prediction model is a model obtained by the source base station through collaborative training under a federated learning framework;

[0032] The selecting module is configured to select a target base station for the terminal based on the future moving track of the terminal and the channel state information of the other base stations;

[0033] The handover module is configured to switch an access base station of the terminal to the target base station.

[0034] According to another aspect of the embodiments of the application, an electronic device is provided, which includes:

[0035] A memory is configured to store executable instructions; and

[0036] A processor is configured to execute the executable instructions with the memory to complete the operations of any of the above-mentioned mobile handover methods based on federated learning.

[0037] According to still another aspect of the embodiments of the present application, a computer readable storage medium is provided for storing computer readable instructions, which, when executed, perform the operations of any of the above-described mobile handover method based on federated learning.

[0038] In the present application, a configuration report reported by a terminal accessing a source base station can be received by the source base station, and the configuration report includes current location information of the terminal and channel state information of other base stations; a future moving track of the terminal is determined by using the current location information and a preset target moving prediction model, the target moving prediction model being a model obtained by the source base station through collaborative training under a federated learning framework; a target base station is selected for the terminal based on the future moving track of the terminal and the channel state information of the other base stations; and the access base station of the terminal is switched to the target base station. By applying the technical solution of the present application, the future moving track of the terminal can be predicted by the base station currently accessed by the terminal based on real-time location information of the terminal, and in the subsequent process, the moving track and a moving prediction model obtained by collaborative training under a federated learning framework are used to select a suitable base station for the terminal, and the base station switching for the terminal is automatically performed. Thus, the model training for moving prediction by using local data is achieved, the model parameters are uploaded to an upper base station for aggregation, and the user track data does not need to be continuously shared between base stations. Meanwhile, the problem of easily affecting the progress of terminal services caused by the insufficient base station switching management method in the related art is avoided.

[0039] The technical solution of the present application is described in further detail below by means of the accompanying drawings and embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0040] The accompanying drawings, which form a part of the specification, illustrate embodiments of the present application and, together with the description, serve to explain the principles of the present application.

[0041] The present application can be understood more clearly with reference to the following detailed description when considered in conjunction with the accompanying drawings, in which:

[0042] Figure 1 A mobile handover method based on federated learning provided by an embodiment of the present application is shown in a schematic diagram;

[0043] Figure 2 A mobile handover system architecture based on federated learning provided by an embodiment of the present application is shown in a schematic diagram;

[0044] Figure 3 A mobile handover method based on federated learning provided by an embodiment of the present application is shown in a flowchart;

[0045] Figure 4 A structure of an electronic device provided by an embodiment of the present application is shown in a schematic diagram;

[0046] Figure 5 Fig. 1 shows a structural schematic diagram of an electronic device according to an embodiment of the present application;

[0047] Figure 6 Fig. 2 shows a schematic diagram of a storage medium according to an embodiment of the present application. DETAILED DESCRIPTION

[0048] Various exemplary embodiments of the present application will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement, numerical expressions, and numerical values of the components and steps set forth in these embodiments are not intended to limit the scope of the present application unless otherwise specifically stated.

[0049] It should be understood, however, that the sizes of the various portions shown in the drawings are not intended to bear a strict proportion to reality.

[0050] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the scope of the application or its application or uses.

[0051] Techniques, methods, and devices known to those of ordinary skill in the relevant art can not be discussed in detail herein. However, where appropriate, such techniques, methods, and devices can be considered part of the specification.

[0052] It should be noted that like reference numerals and letters in the various figures indicate similar items, and, thus, once an item is defined in one figure, it should not require further discussion in subsequent figures.

[0053] In addition, the technical solutions in the various embodiments of the present application can be combined with each other, but it must be based on the fact that a person of ordinary skill in the art can realize it, and when the combination of technical solutions contradicts each other or cannot be realized, it should be considered that the combination of technical solutions does not exist, nor is it within the scope of protection required by the present application.

[0054] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present application are only used to explain the relative positional relationship, movement condition, etc. between the components, and if the specific posture changes, the directional indications will also change accordingly.

[0055] The method for performing mobile switching based on federated learning according to the exemplary embodiments of the present application will be described below with reference to the accompanying drawings. Figures 1-3 It should be noted that the following application scenarios are only shown to facilitate understanding of the spirit and principles of the present application, and the embodiments of the present application are not limited in this respect. On the contrary, the embodiments of the present application can be applied to any applicable scenario.

[0056] The application also provides a mobile switching method and device based on federated learning, an electronic device and a medium.

[0057] Figure 1 A flowchart of a mobile switching method based on federated learning according to an embodiment of the application is schematically shown. As shown in the figure, the method is applied to a source base station and includes the following steps. Figure 1

[0058] S101, receiving a configuration report reported by a terminal accessing itself, the configuration report including current position information of the terminal and channel state information of other base stations.

[0059] S102, determining a future moving track of the terminal by using the current position information and a preset target moving prediction model, the target moving prediction model being a model obtained by the source base station through collaborative training under a federated learning framework.

[0060] S103, selecting a target base station for the terminal based on the future moving track of the terminal and the channel state information of the other base stations.

[0061] S104, switching the access base station of the terminal to the target base station.

[0062] With the rapid growth of the number of user equipment and traffic demand, wireless cellular networks are developing towards denser deployment, with an increase in the number of cells, a decrease in cell radius and an increase in user mobility, which leads to more frequent and costly mobile switching and more complex mobility management.

[0063] Current traditional mobility management schemes are mainly designed for macro cells and cells with infrequent handover, and are difficult to provide reliable handover performance for medium and high-speed mobile users in ultra-dense networks (UDN) to ensure service continuity and user service quality.

[0064] Current researchers introduce user moving prediction based on artificial intelligence to conduct intelligent mobility management, but since each base station can only obtain local user moving track information, good moving prediction performance depends on the cooperation and continuous sharing of user moving track data among base stations. However, a large number of micro base stations (including pico base stations, fly base stations, etc.) are deployed through wireless means, and the data transmission capacity among base stations is limited, so the switching performance such as switching delay of the intelligent mobility management method based on moving prediction is difficult to guarantee.

[0065] ​To solve the above problems, the application provides a mobile switching method based on federated learning. The idea is to predict user movement through the base station side under the federated learning architecture. Each base station can only train the model through the local user movement trajectory data it owns, and does not need to continuously share user movement trajectory data between base stations to achieve the effect of distributed collaborative training of mobile prediction model, effectively reducing the data transmission overhead of intelligent mobile switching. At the same time, according to the prediction accuracy, time delay and other requirements, the model aggregation of federated learning is optimized to be deployed at different network levels to realize distributed intelligent mobile switching under the dynamic hierarchical federated learning architecture, reduce the mobile switching delay, and improve the mobile switching performance.

[0066] In one way, as shown in Figure 2 , it is a system architecture diagram applied in the mobile switching method based on federated learning. Taking the base station as the micro cell base station as an example, it can be seen from the figure that the user terminal is served by the source micro cell base station cluster and constantly moves out of the coverage range of the source micro cell base station cluster, and needs to perform mobile switching. The micro cell base station cluster is connected to the upper layer macro base station, and the upper layer macro base station is connected to the core network cloud server or edge server.

[0067] Further, as shown in Figure 3 , it is a flowchart of the mobile switching method based on federated learning proposed by the application, which includes:

[0068] Step 1, after the user terminal is connected to each source base station in the source micro cell base station cluster, the source base station sends measurement configuration information to the user terminal.

[0069] Step 2, the user terminal measures to obtain a configuration report, wherein the configuration report includes the current position information of the terminal and the channel state information of other base stations.

[0070] Step 3, the user terminal reports the configuration report to each source base station in the source micro cell base station cluster.

[0071] Step 4: each source base station in the source micro cell base station cluster cooperatively trains an initial mobile prediction model under the federated learning framework according to the data input of the measurement report and the initial federated learning model parameters.

[0072] Step 5, each source base station in the source micro cell base station cluster uploads the local trained initial model parameters to the macro cell base station.

[0073] Step 6, the macro cell base station performs low-level federated aggregation of the initial model parameters to obtain low-level global model parameters. Then go to step 7 or step 10.

[0074] It should be noted that after the macro cell base station performs federal aggregation on the initial model parameters sent by each source base station to obtain low-level global model parameters, the macro cell base station can execute one of the following two steps according to the user terminal's requirements for the prediction result standard:

[0075] The steps in the first case are:

[0076] If the macro cell base station detects that the user terminal has a higher requirement for the accuracy of the handover prediction result for selecting a base station for it, or has a lower requirement for the time delay. In this case, in order to improve the handover accuracy of the source base station, step 7 can be continued in the embodiment of the present application. That is, the model aggregation of federal learning is deployed in a higher level cloud core network or edge device. So that the low-level global model parameters sent by multiple macro cell base stations are aggregated by the cloud core network or edge device to obtain the final global model parameters. Thus, the purpose of using more user trajectory data and collecting more extensive model parameters to support the training of the initial mobile prediction model is achieved.

[0077] The steps in the second case are:

[0078] If the macro cell base station detects that the user terminal has a lower requirement for the accuracy of the handover prediction result for selecting a base station for it, or has a higher requirement for the time delay. In this case, in order to ensure the rapid aggregation of model parameters, step 10 (that is, the low-level global model parameters are used as the final global model parameters) can be continued in the embodiment of the present application.

[0079] Step 7, multiple macro cell base stations upload the model parameters aggregated by each low-level federal to the cloud core network or edge device.

[0080] Step 8, the cloud core network or edge device performs high-level federal aggregation of model parameters to obtain high-level global model parameters.

[0081] Step 9, the cloud core network or edge device distributes the global model parameters aggregated by the high-level federal to each macro cell base station.

[0082] It can be understood that the high-level federal aggregation of model parameters by the cloud core network or edge device can obtain more model parameters possessed by the source base station, so as to achieve the purpose of jointly training and aggregating the target global model, and thus the prediction accuracy of the final target mobile prediction is higher.

[0083] Step 10, the macro cell base station distributes the global model parameters to each source base station in the source micro cell base station cluster.

[0084] Step 11: Each source base station performs a mobile prediction model precision test on the initial mobile prediction model of itself, judges whether the model meets the fitting requirements, if it meets the requirements, terminates the federated learning model training process, and continues to step 12, if it does not meet the requirements, continues to train the initial mobile prediction model based on the newly obtained global model parameters, until the target mobile prediction model that meets the conditions is obtained.

[0085] Step 12: Each source base station in the source micro cell base station cluster makes a handover decision based on the terminal current position information included in the configuration report and the channel state information of other base stations, the predicted next time user terminal movement trajectory information, and the channel state of surrounding micro cells in the current time user measurement report, and selects a target micro cell base station for the user terminal.

[0086] Step 13: Each source base station in the source micro cell base station cluster sends a handover request to the target micro cell base station cluster that meets the handover condition;

[0087] Steps 14-15: Each target base station in the target micro cell base station cluster performs admission control, and if it agrees to handover, it feeds back a handover request response to each source base station;

[0088] Step 16: Each source base station sends a handover configuration to the user terminal through an RRC reconfiguration message, including the handover execution condition, configuration parameters, etc. of the target micro cell base station cluster;

[0089] Step 17: The user terminal sends an RRC reconfiguration complete message to each source base station;

[0090] Step 18: The user terminal uses conditional handover (CHO), when the target micro cell base station meets the handover condition, directly starts the handover execution process, separates from each source base station, initiates random access to the target base station, and successfully accesses the target base station, while each source base station sends the data and SN state information of the user terminal to the target base station through the Early / Late data forwarding method.

[0091] At this point, the data transmission between the user terminal and the target micro cell base station cluster is established, and the handover is completed.

[0092] In the present application, a configuration report reported by a terminal accessed by a source base station can be received, and the configuration report includes current position information of the terminal and channel state information of other base stations; a future moving track of the terminal is determined by using the current position information and a preset target moving prediction model, and the target moving prediction model is a model obtained by the source base station through collaborative training under a federated learning framework; a target base station is selected for the terminal based on the future moving track of the terminal and the channel state information of other base stations; and the access base station of the terminal is switched to the target base station. By applying the technical solution of the present application, the future moving track of the terminal can be predicted based on the real-time position information of the terminal by the base station currently accessed by the terminal, and a suitable base station can be selected for the terminal by using the moving track and a moving prediction model obtained by collaborative training under a federated learning framework in the subsequent stage, and the base station switching for the terminal is automatically performed. Thus, the model training of moving prediction by local data is achieved, the model parameters are uploaded to an upper base station for aggregation, and the user track data does not need to be continuously shared between base stations. Meanwhile, the problem that the switching management method between base stations in the related art is insufficient and easily affects the progress of terminal services is avoided.

[0093] Optionally, in another embodiment based on the above method of the present application, before the configuration report reported by the terminal accessed by the source base station is obtained, the method further includes:

[0094] receiving an initial configuration report reported by the terminal accessed by the source base station, and the initial configuration report includes current position information of the terminal and channel state information of other base stations;

[0095] performing model training on an initial moving prediction model by using the initial configuration report to obtain initial model parameters;

[0096] obtaining the target moving prediction model by using the initial model parameters.

[0097] Optionally, in another embodiment based on the above method of the present application, the obtaining of the target moving prediction model by using the initial model parameters includes:

[0098] sending the initial model parameters to a macro base station associated with the source base station, so that the macro base station and an edge device perform federated aggregation on the received initial model parameters to obtain global model parameters;

[0099] receiving the global model parameters sent by the macro base station, and obtaining the target moving prediction model based on the global model parameters.

[0100] Optionally, in another embodiment based on the above method of the present application, the obtaining of the target moving prediction model based on the global model parameters includes:

[0101] performing a preset model accuracy test on the initial mobile prediction model using the global model parameters;

[0102] If it is determined that the initial mobile prediction model meets the preset model accuracy, the initial mobile prediction model is taken as the target mobile prediction model; or,

[0103] If it is determined that the initial mobile prediction model does not meet the preset model accuracy, the initial mobile prediction model is trained using the global model parameters to obtain the target mobile prediction model.

[0104] Optionally, in another embodiment based on the above method of the present application, the switching of the access base station of the terminal to the target base station comprises:

[0105] sending a switching request to the target base station;

[0106] If the switching permission message sent by the target base station is received, a reconfiguration message is sent to the terminal, wherein the reconfiguration message contains the switching execution condition and configuration parameters of the target base station;

[0107] When it is determined that the terminal and the target base station perform the switching process, the access state with the terminal is cancelled.

[0108] Optionally, in another embodiment based on the above method of the present application, after the reconfiguration message is sent to the terminal, it further comprises:

[0109] sending a service message to the target base station, wherein the service message contains the attribute information and SN state information of the terminal.

[0110] Optionally, in another embodiment based on the above method of the present application, the source base station and the target base station are respectively at least one base station in a micro cell base station cluster.

[0111] In one way, the source base station and the target base station in the present application include a memory and a processor, wherein the memory is used to store an initial mobile prediction model and corresponding model parameters, and store the mobile position information reported by the user terminal. In another way, the processor is used to implement the model training of mobile prediction (such as the mobile prediction model based on LSTM). The parameter uploading module is used to upload the trained model parameters to the upper layer macro cell base station or the edge server or the core network cloud server.

[0112] Further, the upper layer macro cell base station or the edge server or the core network cloud server in the present application can be deployed with a federal aggregation module and a parameter issuing module, and the model training module can include a memory and a processor.

[0113] Optionally, in another implementation of the present application, as shown in Figure 4 The present application also provides a mobile switching device based on federated learning. It includes:

[0114] The receiving module 201 is configured to receive a configuration report reported by a terminal accessing itself, wherein the configuration report includes current position information of the terminal and channel state information of other base stations;

[0115] The determining module 202 is configured to determine a future moving track of the terminal by using the current position information and a preset target moving prediction model, wherein the target moving prediction model is a model obtained by collaborative training of the source base station under a federated learning framework;

[0116] The selecting module 203 is configured to select a target base station for the terminal based on the future moving track of the terminal and the channel state information of the other base stations;

[0117] The switching module 204 is configured to switch an access base station of the terminal to the target base station.

[0118] By applying the technical solution of the present application, the future moving track of a terminal can be predicted based on real-time position information of the terminal by a base station currently accessed by the terminal, and a suitable base station can be selected for the terminal by using the moving track and a moving prediction model obtained by collaborative training under a federated learning framework, and the base station of the terminal can be automatically switched. Thus, the model training of moving prediction by local data is achieved, and model parameters are uploaded to an upper base station for aggregation, and user track data does not need to be continuously shared between base stations. Meanwhile, the problem of affecting terminal service progress caused by insufficient switching management method between base stations in related technologies is avoided.

[0119] In another implementation of the present application, the receiving module 201 is configured to perform the following steps:

[0120] receive an initial configuration report reported by a terminal accessing itself, wherein the initial configuration report includes current position information of the terminal and channel state information of other base stations;

[0121] perform model training on an initial moving prediction model by using the initial configuration report to obtain initial model parameters;

[0122] obtain the target moving prediction model by using the initial model parameters.

[0123] In another implementation of the present application, the receiving module 201 is configured to perform the following steps:

[0124] send the initial model parameters to a macro base station associated with the source base station, so that the macro base station and the edge device federatedly aggregate the received respective initial model parameters to obtain global model parameters;

[0125] receive the global model parameters sent by the macro base station, and obtain the target mobile prediction model based on the global model parameters.

[0126] In another embodiment of the present application, the receiving module 201 is configured to perform the following steps:

[0127] performing preset model accuracy verification on the initial mobile prediction model using the global model parameters;

[0128] if it is determined that the initial mobile prediction model meets the preset model accuracy, taking the initial mobile prediction model as the target mobile prediction model; or,

[0129] if it is determined that the initial mobile prediction model does not meet the preset model accuracy, training the initial mobile prediction model using the global model parameters to obtain the target mobile prediction model.

[0130] In another embodiment of the present application, the receiving module 201 is configured to perform the following steps:

[0131] sending a handover request to the target base station;

[0132] if the handover agreement message sent by the target base station is received, sending a reconfiguration message to the terminal, wherein the reconfiguration message contains the handover execution condition and configuration parameters of the target base station;

[0133] when it is determined that the terminal and the target base station perform the handover process, canceling the access state with the terminal.

[0134] In another embodiment of the present application, the receiving module 201 is configured to perform the following steps:

[0135] sending a service message to the target base station, wherein the service message contains attribute information and SN state information of the terminal.

[0136] In another embodiment of the present application, the receiving module 201 is configured to perform the following steps:

[0137] The source base station and the target base station are at least one base station in a micro cell base station cluster.

[0138] The embodiments of the present application also provide an electronic device for executing the above mobile handover method based on federated learning. Please refer to Figure 5This illustrates a schematic diagram of an electronic device provided by some embodiments of this application. For example... Figure 5 As shown, the electronic device 3 includes: a processor 300, a memory 301, a bus 302, and a communication interface 303. The processor 300, the communication interface 303, and the memory 301 are connected through the bus 302. The memory 301 stores a computer program that can run on the processor 300. When the processor 300 runs the computer program, it executes the mobile handover method based on federated learning provided in any of the foregoing embodiments of this application.

[0139] The memory 301 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this device network element and at least one other network element is achieved through at least one communication interface 303 (which can be wired or wireless), such as the Internet, wide area network, local area network, or metropolitan area network.

[0140] Bus 302 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. The memory 301 is used to store programs. After receiving an execution instruction, the processor 300 executes the program. The data recognition method disclosed in any of the foregoing embodiments of this application can be applied to the processor 300, or implemented by the processor 300.

[0141] The processor 300 can be an integrated circuit chip with a processing capability of signals. In the implementation process, each step of the above method can be completed by the integrated logic circuit of hardware in the processor 300 or the instruction in the form of software. The processor 300 described above can be a general processor, including a central processing unit (CPU), a network processor (NP), etc.; can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a ready programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. Each method, step and logic block disclosed in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module can be located in a random access memory, a flash memory, a read only memory, a programmable read only memory or an electrically erasable programmable memory, a register or other mature storage medium in the art. The storage medium is located in the memory 301, and the processor 300 reads the information in the memory 301 and combines the hardware to complete the steps of the above method.

[0142] The electronic device provided by the embodiments of the present application and the mobile switching method based on federated learning provided by the embodiments of the present application have the same beneficial effects as the method they adopt, run or implement.

[0143] The embodiments of the present application also provide a computer readable storage medium corresponding to the data recognition method provided by the preceding embodiments. Please refer to Figure 6 The computer readable storage medium shown in the figure is an optical disc 40, and a computer program (i.e. program product) is stored on the optical disc 40. When the computer program is run by a processor, the mobile switching method based on federated learning provided by any of the preceding embodiments is executed.

[0144] It should be noted that examples of the computer readable storage medium can also include, but are not limited to, a phase change memory (PRAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), other types of random access memory (RAM), a read only memory (ROM), an electrically erasable programmable read only memory (EEPROM), a flash memory or other optical, magnetic storage medium, which will not be described one by one here.

[0145] The computer readable storage medium provided by the above embodiments of the present application has the same inventive concept as the method of data recognition provided by the embodiments of the present application, and has the same beneficial effects as the method adopted, run or implemented by the application stored therein.

[0146] It should be noted that:

[0147] In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the application can be practiced without these specific details. In some instances, well-known structures and techniques have not been shown in detail in order not to obscure the understanding of this description.

[0148] Similarly, it is to be understood that the above description is illustrative of the various aspects of the present application and certain specific embodiments thereof are shown by way of example in the accompanying drawings and will not limit the application since modifications and variations can be apparent to those skilled in the art from this disclosure.

[0149] Further, one skilled in the art will appreciate that a mechanism comprising a feature or features not specifically described herein is intended to fall within the scope of the present application and form a different embodiment of the present application. For example, in the claims below any of the claimed embodiments can be used in any combination.

[0150] The above descriptions are only the preferred embodiments of the present application, but the protection scope of the present application is not limited to this, and any changes or replacements within the technical scope disclosed by the present application should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for mobile switching based on federated learning, characterized in that, Applied to a source base station, comprising: receiving a configuration report reported by a terminal accessing itself, the configuration report comprising current position information of the terminal and channel state information of other base stations; determining a future moving track of the terminal by using the current position information and a preset target moving prediction model, the target moving prediction model being a model obtained by the source base station under a federated learning framework for collaborative training; selecting a target base station for the terminal based on the future moving track of the terminal and the channel state information of the other base stations; switching an access base station of the terminal to the target base station; before obtaining the configuration report reported by the terminal accessing itself, further comprising: receiving an initial configuration report reported by the terminal accessing itself, the initial configuration report comprising current position information of the terminal and channel state information of other base stations; performing model training on an initial moving prediction model by using the initial configuration report to obtain initial model parameters; and obtaining the target moving prediction model by using the initial model parameters; the obtaining the target moving prediction model by using the initial model parameters comprises: sending the initial model parameters to a macro base station associated with the source base station, so that the macro base station performs federated aggregation on the received initial model parameters to obtain global model parameters; and receiving the global model parameters sent by the macro base station and obtaining the target moving prediction model based on the global model parameters; the obtaining the target moving prediction model based on the global model parameters comprises: performing preset model accuracy verification on the initial moving prediction model by using the global model parameters; if it is determined that the initial moving prediction model meets the preset model accuracy, taking the initial moving prediction model as the target moving prediction model; or, if it is determined that the initial moving prediction model does not meet the preset model accuracy, training the initial moving prediction model by using the global model parameters to obtain the target moving prediction model.

2. The method of claim 1, wherein, the switching the access base station of the terminal to the target base station comprises: sending a switching request to the target base station; if a switching permission message sent by the target base station is received, sending a reconfiguration message to the terminal, the reconfiguration message comprising switching execution conditions and configuration parameters of the target base station; when it is determined that the terminal and the target base station perform a switching process, canceling an access state with the terminal.

3. The method of claim 2, wherein, after the sending the reconfiguration message to the terminal, further comprising: sending a service message to the target base station, the service message comprising attribute information and SN state information of the terminal.

4. The method according to any one of claims 1 to 3, characterized in that, the source base station and the target base station are respectively at least one base station in a micro cell base station cluster. 5.A mobile switching device based on federated learning, characterized in that, applied to a source base station, comprising: a receiving module configured to receive a configuration report reported by a terminal accessing itself, the configuration report comprising current position information of the terminal and channel state information of other base stations; The determining module is configured to determine a future moving track of the terminal by using the current position information and a preset target moving prediction model, the target moving prediction model being a model obtained by the source base station through collaborative training under a federated learning framework; The selecting module is configured to select a target base station for the terminal based on the future moving track of the terminal and channel state information of the other base stations; The switching module is configured to switch an access base station of the terminal to the target base station; The apparatus is further configured to: receive an initial configuration report reported by a terminal accessing itself, the initial configuration report including current position information of the terminal and channel state information of other base stations; perform model training on an initial moving prediction model by using the initial configuration report to obtain initial model parameters; and obtain the target moving prediction model by using the initial model parameters; The obtaining of the target moving prediction model by using the initial model parameters includes: sending the initial model parameters to a macro base station associated with the source base station, so that the macro base station performs federated aggregation on the received initial model parameters to obtain global model parameters; receiving the global model parameters sent by the macro base station, and obtaining the target moving prediction model based on the global model parameters; The obtaining of the target moving prediction model based on the global model parameters includes: performing preset model accuracy verification on the initial moving prediction model by using the global model parameters; if it is determined that the initial moving prediction model meets the preset model accuracy, taking the initial moving prediction model as the target moving prediction model, or if it is determined that the initial moving prediction model does not meet the preset model accuracy, training the initial moving prediction model by using the global model parameters to obtain the target moving prediction model.

6. An electronic device, comprising: including: a memory for storing executable instructions; and a processor for executing the executable instructions with the memory to complete the operations of the federated learning-based moving switching method in any one of claims 1-4.

7. A computer readable storage medium storing computer readable instructions, wherein, The instructions are executed to perform the operations of the federated learning-based moving switching method in any one of claims 1-4.

Images