Model switching update method and apparatus

Abstract

The present disclosure provides a model switching update method and apparatus. The method comprises: a first device performing performance monitoring on a first model currently used by the first device and / or a second model currently used by a second device; if a performance monitoring result satisfies a first preset condition, determining that model switching update needs to be performed; the first device determining at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter; the first device performing model switching update and sending indication information to the second device, the indication information indicating the target second model and / or the second model update parameter. According to the method of the present disclosure, models used in real time by the first device and / or the second device can match a current channel scenario or data distribution characteristics, and CSI feedback accuracy is improved.

Description

Model switching and updating method and device Technical Field

[0001] The present disclosure relates to the field of communication technology, and in particular to a model switching update method and device. Background Art

[0002] In communication systems, terminals typically need to report downlink channel state information (CSI) to network devices so that the network devices can determine the downlink channel quality based on the CSI. Alternatively, the 3rd Generation Partnership Project (3GPP) system introduces compressed CSI reporting and reconstruction based on a bilateral model.

[0003] Summary of the Invention

[0004] The present disclosure provides a model switching update method and device.

[0005] According to a first aspect of an embodiment of the present disclosure, a model switching update method is proposed, comprising:

[0006] The first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device; the first model and the second model correspond to each other, and the first model and the second model are used to implement compression and reconstruction of channel state information (CSI);

[0007] The performance monitoring result meets the first preset condition, and it is determined that a model switching update is required;

[0008] The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter; wherein the target first model is a target model to be switched by the first device, and the target second model is a target model to be switched by the second device;

[0009] The first device performs a model switching update and sends indication information to the second device, where the indication information indicates the target second model and / or the second model update parameter.

[0010] According to a second aspect of an embodiment of the present disclosure, a model switching update method is proposed, comprising:

[0011] The second device receives indication information sent by the first device, where the indication information indicates a target second model and / or a second model update parameter; wherein the target second model is a target model to be switched by the second device;

[0012] The second device performs model switching update based on the indication information.

[0013] According to a third aspect of an embodiment of the present disclosure, a model switching update method is provided for a communication system, the communication system including a first device and a second device, the method including at least one of the following:

[0014] The first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device; the first model and the second model correspond to each other, and the first model and the second model are used to implement compression and reconstruction of channel state information (CSI);

[0015] The performance monitoring result meets the first preset condition, and the first device determines that a model switching update is required;

[0016] The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter; wherein the target first model is a target model to be switched by the first device, and the target second model is a target model to be switched by the second device;

[0017] The first device performs a model switching update and sends indication information to the second device, where the indication information indicates the target second model and / or an update parameter of the second model;

[0018] The second device receives the indication information sent by the first device;

[0019] The second device performs model switching update based on the indication information.

[0020] According to a fourth aspect of an embodiment of the present disclosure, a first device is provided, including:

[0021] a processing module, configured to perform performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device; the first model and the second model correspond to each other, and the first model and the second model are used to implement compression and reconstruction of channel state information (CSI);

[0022] The processing module is further configured to determine that a model switching update is required when a performance monitoring result meets a first preset condition;

[0023] The processing module is further configured to determine at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter;

[0024] The sending module is configured to perform a model switching update and send indication information to the second device, where the indication information indicates the target second model and / or the second model update parameter.

[0025] According to a fifth aspect of the embodiments of the present disclosure, a second device is provided, including:

[0026] A receiving module, configured for the second device to receive indication information sent by the first device, wherein the indication information indicates a target second model and / or a second model update parameter; wherein the target second model is a target model to be switched by the second device;

[0027] A processing module is used for the second device to perform model switching update based on the indication information.

[0028] According to a sixth aspect of an embodiment of the present disclosure, a communication device is provided, including:

[0029] one or more processors;

[0030] The processor is used to call instructions to enable the communication device to execute the model switching update method described in the first aspect or the second aspect.

[0031] According to the seventh aspect of an embodiment of the present disclosure, a communication system is proposed, characterized in that it includes a first device and a second device, wherein the first device is configured to implement the model switching update method described in the first aspect, and the second device is configured to implement the model switching update method described in the first aspect or the second aspect.

[0032] According to an eighth aspect of an embodiment of the present disclosure, a storage medium is proposed, which stores instructions, and is characterized in that when the instructions are executed on a communication device, the communication device executes the model switching update method described in the first aspect or the second aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The above and / or additional aspects and advantages of the present disclosure will become apparent and readily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0034] FIG1A is a schematic diagram of the architecture of some communication systems provided by embodiments of the present disclosure;

[0035] FIG1B is a structural block diagram of a method for “CSI compression reporting and reconstruction based on a bilateral model” provided by one embodiment of the present disclosure;

[0036] 2A-2B are interactive diagrams of a model switching and updating method provided by an embodiment of the present disclosure;

[0037] 3A-3C are flowcharts of a model switching and updating method provided in yet another embodiment of the present disclosure;

[0038] 4A-4C are flowcharts of a model switching and updating method provided in yet another embodiment of the present disclosure;

[0039] 5A-5C are flowcharts of a model switching and updating method provided in yet another embodiment of the present disclosure;

[0040] FIG6A is a schematic diagram of the structure of a network device provided by an embodiment of the present disclosure;

[0041] FIG6B is a schematic structural diagram of a terminal provided by an embodiment of the present disclosure;

[0042] FIG7A is a schematic structural diagram of a communication device provided by an embodiment of the present disclosure;

[0043] FIG7B is a schematic structural diagram of a chip provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0044] The present disclosure provides a model switching update method, device, equipment, system, and computer storage medium.

[0045] In a first aspect, an embodiment of the present disclosure provides a model switching update method, the method comprising:

[0046] The first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device; the first model and the second model correspond to each other, and the first model and the second model are used to implement compression and reconstruction of channel state information (CSI);

[0047] The performance monitoring result meets the first preset condition, and it is determined that a model switching update is required;

[0048] The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter; wherein the target first model is a target model to be switched by the first device, and the target second model is a target model to be switched by the second device;

[0049] The first device performs a model switching update and sends indication information to the second device, where the indication information indicates the target second model and / or the second model update parameter.

[0050] In the above embodiment, the first device determines whether a model switch update is required by determining whether the performance monitoring results of the first model currently used by the first device and / or the second model currently used by the second device meet the first preset condition, and when it is determined that a model switch update is required, at least one of the target first model, the first model update parameter, the target second model, and the second model update parameter is determined. After that, the first device performs a model switch update and sends an instruction message to the second device to instruct the second device to also perform a model switch update. It can be seen that in the embodiment of the present disclosure, the first device can adaptively switch the model based on the real-time performance monitoring results of the first model currently used by the first device, and can also adaptively update the model parameters of the model. At the same time, the first device can also instruct the second device to adaptively switch the model based on the real-time performance monitoring results of the second model currently used by the second device, and / or adaptively update the model parameters of the model, so as to ensure that the models used in real time by the first device and / or the second device are both matched with the current channel scenario or data distribution characteristics, thereby improving the accuracy of CSI feedback.

[0051] In conjunction with some embodiments of the first aspect, in some embodiments, before the first device performs performance monitoring on the first model currently used by the first device and / or the second model currently used by the second device, the method further includes:

[0052] Determine model-related information corresponding to the second device; the model-related information includes at least one of the following:

[0053] a second model currently used by the second device;

[0054] at least one second model deployed by the second device;

[0055] The model structure corresponding to the at least one second model deployed by the second device.

[0056] In conjunction with some embodiments of the first aspect, in some embodiments, determining model-related information corresponding to the second device includes:

[0057] Receive the model-related information sent by the second device.

[0058] In the above embodiment, before the first device performs performance monitoring on the currently used first model and / or the currently used second model, it will first obtain model-related information corresponding to the second device, so that the first device can know the second model currently used by the second device, which second models are deployed on the second device, and the model structure of the second model deployed on the second device based on the model-related information. After the first device understands the model deployment and usage on the second device, the first device can subsequently successfully monitor the performance of the second model currently used by the second device, and can successfully determine the second model update parameters and / or the target second model, thereby successfully instructing the second device to switch and update the model, so that the second device can adaptively switch the model and / or update the model parameters of the model based on the instruction of the first device, so that the model used by the second device in real time can match the current channel scenario or data distribution characteristics, thereby improving the accuracy of CSI feedback.

[0059] In conjunction with some embodiments of the first aspect, in some embodiments, before determining that a model switching update is required, the method further includes:

[0060] The first device determines a first parameter value, where the first parameter value is used to determine whether a model switching update is required;

[0061] Among them, the first preset condition includes: the performance monitoring result is less than the first parameter value.

[0062] In conjunction with some embodiments of the first aspect, in some embodiments, the first device determines the first parameter value, including at least one of the following:

[0063] The first device receives the first parameter value sent by the second device;

[0064] The first device determines the first parameter value configured in the first device;

[0065] The first device determines the first parameter value based on a protocol agreement.

[0066] In the above embodiment, before the first device performs performance monitoring on the currently used first model and / or the currently used second model, it will first determine the first parameter value, so that the first device can determine whether a model switch update is currently required based on the first parameter value, and when it is determined that a model switch update is currently required, the subsequent model switch update process is executed, thereby ensuring the integrity and smooth execution of the model switch update process.

[0067] In conjunction with some embodiments of the first aspect, in some embodiments, the first device determines at least one of the target first model, the first model update parameter, the target second model, and the second model update parameter, including:

[0068] The first model currently used by the first device is the only first model deployed by the first device, and / or the second model currently used by the second device is the only second model deployed by the second device, and sample data is determined;

[0069] Training the first model currently used by the first device and / or the second model currently used by the second device based on the sample data until a performance monitoring result of the trained first model and / or the trained second model meets a second preset condition, and then stopping the training;

[0070] The model parameters of the trained first model are determined as the first model update parameters, and / or the model parameters of the trained second model are determined as the second model update parameters.

[0071] In conjunction with some embodiments of the first aspect, in some embodiments, the first device performs a model switching update and sends indication information to the second device, including at least one of the following:

[0072] The first device updates the model parameters of the currently used first model to the first model update parameters;

[0073] The first device sends indication information to the second device, where the indication information includes second model update parameters and / or a first indication; the first indication is used to instruct the second device to update the model parameters of the currently used second model to the second model update parameters.

[0074] In conjunction with some embodiments of the first aspect, in some embodiments, the first device determines at least one of the target first model, the first model update parameter, the target second model, and the second model update parameter, including:

[0075] The first device deploys at least two first models, and / or the second device deploys at least two second models, and determines at least one model group, wherein the model group includes first models and second models corresponding to each other;

[0076] Performing performance monitoring on the first model and / or the second model in each model group;

[0077] If the performance monitoring result of at least one of the model groups meets the second preset condition, the first model in the model group with the highest performance monitoring result is determined as the target first model, and the second model in the model group with the highest performance monitoring result is determined as the target second model.

[0078] In conjunction with some embodiments of the first aspect, in some embodiments, the first device performs a model switching update and sends indication information to the second device, including at least one of the following:

[0079] The first device deactivates the currently used first model and switches from the currently used first model to the target first model;

[0080] The first device sends indication information to the second device, where the indication information includes a second indication and / or relevant information of a target second model; the second indication is used to indicate at least one of the following:

[0081] The second device deactivates the second model currently in use;

[0082] The second device switches from the currently used second model to the target second model.

[0083] In conjunction with some embodiments of the first aspect, in some embodiments, the first device determines at least one of the target first model, the first model update parameter, the target second model, and the second model update parameter, including:

[0084] The first device deploys at least two first models, and / or the second device deploys at least two second models, and determines at least one model group, wherein the model group includes first models and second models corresponding to each other;

[0085] Monitor the performance of each model group;

[0086] If the performance monitoring results of each model group do not meet the second preset condition, the first model in the model group with the highest performance monitoring result is determined as the target first model, and the second model in the model group with the highest performance monitoring result is determined as the target second model;

[0087] Determine sample data;

[0088] Training the target first model and the target second model based on the sample data until a performance monitoring result of the trained first model and / or the trained second model meets a second preset condition, and then stopping the training;

[0089] The model parameters of the trained target first model are determined as the first model update parameters, and / or the model parameters of the trained target second model are determined as the second model update parameters.

[0090] In conjunction with some embodiments of the first aspect, in some embodiments, the first device performs a model switching update and sends indication information to the second device, including at least one of the following:

[0091] The first device deactivates the currently used first model, switches from the currently used first model to the target first model, and updates the model parameters of the target first model to the first model update parameters;

[0092] The first device sends indication information to the second device, where the indication information includes at least one of relevant information of the target second model, a second model update parameter, and a third indication, where the third indication is used to indicate at least one of the following:

[0093] The second device deactivates the second model currently in use;

[0094] The second device switches from the currently used second model to the target second model;

[0095] The second device updates the model parameters of the target second model to the second model update parameters.

[0096] In combination with some embodiments of the first aspect, in some embodiments, the second preset condition includes: the performance monitoring result is greater than or equal to a first parameter value.

[0097] In the above embodiment, a method is provided for how a first device determines a target first model, a first model update parameter, a target second model, and a second model update parameter, so that the first device successfully determines at least one of the target first model, the first model update parameter, the target second model, and the second model update parameter. Moreover, the prerequisites for the first device to determine the target first model, the first model update parameter, the target second model, and the second model update parameter are all: the performance monitoring result of the first model and / or the second model meets the second preset condition. The second preset condition is: the performance monitoring result is greater than or equal to the first parameter value (i.e., the performance monitoring result is higher). It can be seen that in the embodiment of the present disclosure, the first device determines the target first model, the first model update parameter, the target second model, and the second model update parameter with the purpose of "higher model performance monitoring result". As a result, the target first model, the first model update parameter, the target second model, and the second model update parameter determined by the first device can all make the performance monitoring results of the first model and the second model higher, thereby improving the feedback accuracy of the CSI.

[0098] Moreover, in the above embodiment, a method is provided for how the first device performs model switching update and how to instruct the second device to perform model switching update, so that the first device can successfully implement model switching update, and the first device can successfully instruct the second device to perform model switching update, so that both the first device and the second device can adaptively switch the model and / or update the model parameters of the model, so that the models used in real time by the first device and the second device can both match the current channel scenario or data distribution characteristics, thereby improving the accuracy of CSI feedback.

[0099] In conjunction with some embodiments of the first aspect, in some embodiments, after the first device sends the indication information to the second device, the method further includes:

[0100] Receive confirmation information sent by the second device, where the confirmation information is used to indicate that the second device has completed the model switching update.

[0101] In conjunction with some embodiments of the first aspect, in some embodiments, the first device is a terminal, and the second device is a network device;

[0102] The first model is used to compress CSI, and the second model is used to reconstruct CSI.

[0103] In conjunction with some embodiments of the first aspect, in some embodiments, the first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device, including:

[0104] The first device uses the CSI data to monitor the performance of a first model currently used by the first device and / or a second model currently used by the second device.

[0105] In conjunction with some embodiments of the first aspect, in some embodiments, after the first device sends the indication information to the second device, the method further includes:

[0106] The first model after switching and updating is used for CSI compression.

[0107] In the above embodiment, the first device can be a terminal and the second device can be a network device, that is, the terminal determines at least one of the target first model, the first model update parameters, the target second model, and the second model update parameters. Since the CSI data on the terminal side is usually required when determining the target first model, the first model update parameters, the target second model, and the second model update parameters, the present disclosure uses the terminal to determine at least one of the target first model, the first model update parameters, the target second model, and the second model update parameters without using other devices (such as network devices) to determine them. This can avoid the situation where "when other devices are used to determine at least one of the target first model, the first model update parameters, the target second model, and the second model update parameters, the terminal needs to send CSI data to other devices", thereby saving communication overhead and simplifying the process.

[0108] In conjunction with some embodiments of the first aspect, in some embodiments, the first device is a network device, and the second device is a terminal;

[0109] The second model is used to compress CSI, and the first model is used to reconstruct CSI.

[0110] In conjunction with some embodiments of the first aspect, in some embodiments, the first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device, including:

[0111] The first device sends a first request to the second device, where the first request is used to request CSI data;

[0112] The first device receives the CSI data sent by the second device;

[0113] The first device uses the CSI data to monitor the performance of a first model currently used by the first device and / or a second model currently used by the second device.

[0114] In conjunction with some embodiments of the first aspect, in some embodiments, determining the sample data includes:

[0115] The first device sends a second request to the second device, where the second request is used to request the sample data;

[0116] The first device receives the sample data sent by the second device.

[0117] In conjunction with some embodiments of the first aspect, in some embodiments, after the first device sends the indication information to the second device, the method further includes:

[0118] The CSI reconstruction is performed using the first model after switching updates.

[0119] In the above embodiment, the first device can be a network device and the second device can be a terminal, that is, the network device determines at least one of the target first model, the first model update parameters, the target second model, and the second model update parameters. Since the computing resource configuration and hardware configuration of the network device are relatively high, when the network device is used to determine the target first model, the first model update parameters, the target second model, and the second model update parameters, it can better provide hardware support for the training and updating of the model.

[0120] In combination with some embodiments of the first aspect, in some embodiments, the CSI data is obtained by the terminal through measuring a pilot signal sent by a network device and through channel estimation.

[0121] In combination with some embodiments of the first aspect, in some embodiments, the sample data is: CSI data measured and estimated by the terminal within a first preset time period.

[0122] In combination with some embodiments of the first aspect, in some embodiments, the performance monitoring result is used to reflect the CSI feedback accuracy of the first model and the second model.

[0123] In a second aspect, an embodiment of the present disclosure proposes a model switching update method, the method comprising:

[0124] The second device receives indication information sent by the first device, where the indication information indicates a target second model and / or a second model update parameter; wherein the target second model is a target model to be switched by the second device;

[0125] The second device performs model switching update based on the indication information.

[0126] In the above embodiment, the first device sends instruction information to the second device to instruct the second device to also switch and update the model. Therefore, in the disclosed embodiment, the first device can instruct the second device to adaptively switch models and / or adaptively update model parameters, thereby ensuring that the model used in real time by the second device matches the current channel scenario or data distribution characteristics, thereby improving the accuracy of CSI feedback.

[0127] In combination with some embodiments of the second aspect, in some embodiments, a first model is deployed on the first device, and a second model is deployed on the second device; the first model and the second model correspond to each other, and the first model and the second model are used to realize compression and reconstruction of CSI.

[0128] In conjunction with some embodiments of the second aspect, in some embodiments, before the second device receives the indication information sent by the first device, the method further includes:

[0129] Sending model-related information corresponding to the second device to the first device; the model-related information includes at least one of the following:

[0130] a second model currently used by the second device;

[0131] at least one second model deployed by the second device;

[0132] The model structure corresponding to the at least one second model deployed by the second device.

[0133] In conjunction with some embodiments of the second aspect, in some embodiments, before the second device receives the indication information sent by the first device, the method further includes:

[0134] A first parameter value is sent to the first device, where the first parameter value is used to determine whether a model switching update is required.

[0135] With reference to some embodiments of the second aspect, in some embodiments, the indication information includes a second model update parameter and / or a first indication; the first indication is used to instruct the second device to update the model parameters of the currently used second model to the second model update parameter;

[0136] The performing model switching and updating based on the indication information includes:

[0137] The second device updates the model parameters of the currently used second model to the second model update parameters.

[0138] In conjunction with some embodiments of the second aspect, in some embodiments, the indication information includes a second indication and / or relevant information of the target second model; the second indication is used to indicate at least one of the following: the second device deactivates the currently used second model, or the second device switches from the currently used second model to the target second model;

[0139] The performing model switching and updating based on the indication information includes:

[0140] The second device deactivates the currently used second model and switches from the currently used second model to the target second model.

[0141] In conjunction with some embodiments of the second aspect, in some embodiments, the indication information includes at least one of relevant information of the target second model, second model update parameters, and a third indication, where the third indication is used to indicate at least one of the following: the second device deactivates the currently used second model, the second device switches from the currently used second model to the target second model, or the second device updates the model parameters of the target second model to the second model update parameters;

[0142] The performing model switching and updating based on the indication information includes:

[0143] The second device deactivates the currently used second model, switches from the currently used second model to the target second model, and updates the model parameters of the target second model to the second model update parameters.

[0144] In conjunction with some embodiments of the second aspect, in some embodiments, after the second device performs a model switching update based on the indication information, the method further includes:

[0145] Sending a confirmation message to the first device, where the confirmation message is used to indicate that the second device has completed the model switching update.

[0146] In conjunction with some embodiments of the second aspect, in some embodiments, the first device is a terminal, and the second device is a network device;

[0147] The first model is used to compress CSI, and the second model is used to reconstruct CSI.

[0148] In conjunction with some embodiments of the second aspect, in some embodiments, after the second device performs a model switching update based on the indication information, the method further includes:

[0149] The CSI is reconstructed using the second model after switching updates.

[0150] In conjunction with some embodiments of the second aspect, in some embodiments, the first device is a network device, and the second device is a terminal;

[0151] The second model is used to compress CSI, and the first model is used to reconstruct CSI.

[0152] In conjunction with some embodiments of the second aspect, in some embodiments, the method further includes:

[0153] receiving a first request sent by a first device, where the first request is used to request CSI data;

[0154] Send CSI data to the first device.

[0155] In conjunction with some embodiments of the second aspect, in some embodiments, the method further includes:

[0156] receiving a second request sent by the first device, where the first request is for requesting sample data;

[0157] Send sample data to the first device.

[0158] In conjunction with some embodiments of the second aspect, in some embodiments, after the second device performs a model switching update based on the indication information, the method further includes:

[0159] The second model after switching and updating is used for CSI compression.

[0160] In a third aspect, an embodiment of the present disclosure provides a model switching update method for a communication system, wherein the communication system includes a first device and a second device, and the method includes at least one of the following:

[0161] The first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device; the first model and the second model correspond to each other, and the first model and the second model are used to implement compression and reconstruction of channel state information (CSI);

[0162] The performance monitoring result meets the first preset condition, and the first device determines that a model switching update is required;

[0163] The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter;

[0164] The first device performs a model switching update and sends indication information to the second device, where the indication information indicates the target second model and / or an update parameter of the second model;

[0165] The second device receives the indication information sent by the first device;

[0166] The second device performs model switching update based on the indication information.

[0167] In a fourth aspect, an embodiment of the present disclosure provides a first device, including:

[0168] a processing module, configured to perform performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device; the first model and the second model correspond to each other, and the first model and the second model are used to implement compression and reconstruction of channel state information (CSI);

[0169] The processing module is further configured to determine that a model switching update is required when a performance monitoring result meets a first preset condition;

[0170] The processing module is further configured to determine at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter; wherein the target first model is a target model to be switched by the first device, and the target second model is a target model to be switched by the second device;

[0171] The sending module is configured to perform a model switching update and send indication information to the second device, where the indication information indicates the target second model and / or the second model update parameter.

[0172] In conjunction with some embodiments of the fourth aspect, in some embodiments, before the first device performs performance monitoring on the first model currently used by the first device and / or the second model currently used by the second device, the apparatus is further configured to:

[0173] Determine model-related information corresponding to the second device; the model-related information includes at least one of the following:

[0174] a second model currently used by the second device;

[0175] at least one second model deployed by the second device;

[0176] The model structure corresponding to the at least one second model deployed by the second device.

[0177] In conjunction with some embodiments of the fourth aspect, in some embodiments, the device is further used to:

[0178] Receive the model-related information sent by the second device.

[0179] In conjunction with some embodiments of the fourth aspect, in some embodiments, before determining that a model switching update is required, the apparatus is further configured to:

[0180] The first device determines a first parameter value, where the first parameter value is used to determine whether a model switching update is required;

[0181] Among them, the first preset condition includes: the performance monitoring result is less than the first parameter value.

[0182] In conjunction with some embodiments of the fourth aspect, in some embodiments, the device is further used for at least one of the following:

[0183] The first device receives the first parameter value sent by the second device;

[0184] The first device determines the first parameter value configured in the first device;

[0185] The first device determines the first parameter value based on a protocol agreement.

[0186] In conjunction with some embodiments of the fourth aspect, in some embodiments, the processing module is further configured to:

[0187] The first model currently used by the first device is the only first model deployed by the first device, and / or the second model currently used by the second device is the only second model deployed by the second device, and sample data is determined;

[0188] Training the first model currently used by the first device and / or the second model currently used by the second device based on the sample data until a performance monitoring result of the trained first model and / or the trained second model meets a second preset condition, and then stopping the training;

[0189] The model parameters of the trained first model are determined as the first model update parameters, and / or the model parameters of the trained second model are determined as the second model update parameters.

[0190] In conjunction with some embodiments of the fourth aspect, in some embodiments, the sending module is further used for at least one of the following:

[0191] The first device updates the model parameters of the currently used first model to the first model update parameters;

[0192] The first device sends indication information to the second device, where the indication information includes second model update parameters and / or a first indication; the first indication is used to instruct the second device to update the model parameters of the currently used second model to the second model update parameters.

[0193] In conjunction with some embodiments of the fourth aspect, in some embodiments, the processing module is further configured to:

[0194] The first device deploys at least two first models, and / or the second device deploys at least two second models, and determines at least one model group, wherein the model group includes first models and second models corresponding to each other;

[0195] Performing performance monitoring on the first model and / or the second model in each model group;

[0196] If the performance monitoring result of at least one of the model groups meets the second preset condition, the first model in the model group with the highest performance monitoring result is determined as the target first model, and the second model in the model group with the highest performance monitoring result is determined as the target second model.

[0197] In conjunction with some embodiments of the fourth aspect, in some embodiments, the sending module is further used for at least one of the following:

[0198] The first device deactivates the currently used first model and switches from the currently used first model to the target first model;

[0199] The first device sends indication information to the second device, where the indication information includes a second indication and / or relevant information of a target second model; the second indication is used to indicate at least one of the following:

[0200] The second device deactivates the second model currently in use;

[0201] The second device switches from the currently used second model to the target second model.

[0202] In conjunction with some embodiments of the fourth aspect, in some embodiments, the processing module is further configured to:

[0203] The first device deploys at least two first models, and / or the second device deploys at least two second models, and determines at least one model group, wherein the model group includes first models and second models corresponding to each other;

[0204] Monitor the performance of each model group;

[0205] If the performance monitoring results of each model group do not meet the second preset condition, the first model in the model group with the highest performance monitoring result is determined as the target first model, and the second model in the model group with the highest performance monitoring result is determined as the target second model;

[0206] Determine sample data;

[0207] Training the target first model and the target second model based on the sample data until a performance monitoring result of the trained first model and / or the trained second model meets a second preset condition, and then stopping the training;

[0208] The model parameters of the trained target first model are determined as the first model update parameters, and / or the model parameters of the trained target second model are determined as the second model update parameters.

[0209] In conjunction with some embodiments of the fourth aspect, in some embodiments, the sending module is further used for at least one of the following:

[0210] The first device deactivates the currently used first model, switches from the currently used first model to the target first model, and updates the model parameters of the target first model to the first model update parameters;

[0211] The first device sends indication information to the second device, where the indication information includes at least one of relevant information of the target second model, a second model update parameter, and a third indication, where the third indication is used to indicate at least one of the following:

[0212] The second device deactivates the second model currently in use;

[0213] The second device switches from the currently used second model to the target second model;

[0214] The second device updates the model parameters of the target second model to the second model update parameters.

[0215] In combination with some embodiments of the fourth aspect, in some embodiments, the second preset condition includes: the performance monitoring result is greater than or equal to the first parameter value.

[0216] In conjunction with some embodiments of the fourth aspect, in some embodiments, after the first device sends the indication information to the second device, the apparatus is further configured to:

[0217] Receive confirmation information sent by the second device, where the confirmation information is used to indicate that the second device has completed the model switching update.

[0218] In conjunction with some embodiments of the fourth aspect, in some embodiments, the first device is a terminal, and the second device is a network device;

[0219] The first model is used to compress CSI, and the second model is used to reconstruct CSI.

[0220] In conjunction with some embodiments of the fourth aspect, in some embodiments, the processing module is further configured to:

[0221] The first device uses the CSI data to monitor the performance of a first model currently used by the first device and / or a second model currently used by the second device.

[0222] In conjunction with some embodiments of the fourth aspect, in some embodiments, after the first device sends the indication information to the second device, the apparatus is further configured to:

[0223] The first model after switching and updating is used for CSI compression.

[0224] In conjunction with some embodiments of the fourth aspect, in some embodiments, the first device is a network device, and the second device is a terminal;

[0225] The second model is used to compress CSI, and the first model is used to reconstruct CSI.

[0226] In conjunction with some embodiments of the fourth aspect, in some embodiments, the processing module is further configured to:

[0227] The first device sends a first request to the second device, where the first request is used to request CSI data;

[0228] The first device receives the CSI data sent by the second device;

[0229] The first device uses the CSI data to monitor the performance of a first model currently used by the first device and / or a second model currently used by the second device.

[0230] In conjunction with some embodiments of the fourth aspect, in some embodiments, the device is further used to:

[0231] The first device sends a second request to the second device, where the second request is used to request the sample data;

[0232] The first device receives the sample data sent by the second device.

[0233] In conjunction with some embodiments of the fourth aspect, in some embodiments, after the first device sends the indication information to the second device, the apparatus is further configured to:

[0234] The CSI reconstruction is performed using the first model after switching updates.

[0235] In combination with some embodiments of the fourth aspect, in some embodiments, the CSI data is obtained by the terminal through measuring a pilot signal sent by a network device and through channel estimation.

[0236] In combination with some embodiments of the fourth aspect, in some embodiments, the sample data is: CSI data measured and estimated by the terminal within a first preset time period.

[0237] In combination with some embodiments of the fourth aspect, in some embodiments, the performance monitoring result is used to reflect the CSI feedback accuracy of the first model and the second model.

[0238] In a fifth aspect, an embodiment of the present disclosure provides a second device, including:

[0239] A receiving module, configured for the second device to receive indication information sent by the first device, wherein the indication information indicates a target second model and / or a second model update parameter; wherein the target second model is a target model to be switched by the second device;

[0240] A processing module is used for the second device to perform model switching update based on the indication information.

[0241] In combination with some embodiments of the fifth aspect, in some embodiments, a first model is deployed on the first device, and a second model is deployed on the second device; the first model and the second model correspond to each other, and the first model and the second model are used to realize compression and reconstruction of CSI.

[0242] In conjunction with some embodiments of the fifth aspect, in some embodiments, before the second device receives the indication information sent by the first device, the apparatus is further configured to:

[0243] Sending model-related information corresponding to the second device to the first device; the model-related information includes at least one of the following:

[0244] a second model currently used by the second device;

[0245] at least one second model deployed by the second device;

[0246] The model structure corresponding to the at least one second model deployed by the second device.

[0247] In conjunction with some embodiments of the fifth aspect, in some embodiments, before the second device receives the indication information sent by the first device, the apparatus is further configured to:

[0248] A first parameter value is sent to the first device, where the first parameter value is used to determine whether a model switching update is required.

[0249] With reference to some embodiments of the fifth aspect, in some embodiments, the indication information includes a second model update parameter and / or a first indication; the first indication is used to instruct the second device to update the model parameters of the currently used second model to the second model update parameter;

[0250] The processing module is further configured to:

[0251] The second device updates the model parameters of the currently used second model to the second model update parameters.

[0252] In conjunction with some embodiments of the fifth aspect, in some embodiments, the indication information includes a second indication and / or relevant information of the target second model; the second indication is used to indicate at least one of the following: the second device deactivates the currently used second model, or the second device switches from the currently used second model to the target second model;

[0253] The processing module is further configured to:

[0254] The second device deactivates the currently used second model and switches from the currently used second model to the target second model.

[0255] In conjunction with some embodiments of the fifth aspect, in some embodiments, the indication information includes at least one of relevant information of the target second model, second model update parameters, and a third indication, where the third indication is used to indicate at least one of the following: the second device deactivates the currently used second model, the second device switches from the currently used second model to the target second model, or the second device updates the model parameters of the target second model to the second model update parameters;

[0256] The processing module is further configured to:

[0257] The second device deactivates the currently used second model, switches from the currently used second model to the target second model, and updates the model parameters of the target second model to the second model update parameters.

[0258] In conjunction with some embodiments of the fifth aspect, in some embodiments, after the second device performs a model switching update based on the indication information, the apparatus is further configured to:

[0259] Sending a confirmation message to the first device, where the confirmation message is used to indicate that the second device has completed the model switching update.

[0260] In conjunction with some embodiments of the fifth aspect, in some embodiments, the first device is a terminal, and the second device is a network device;

[0261] The first model is used to compress CSI, and the second model is used to reconstruct CSI.

[0262] In conjunction with some embodiments of the fifth aspect, in some embodiments, after the second device performs a model switching update based on the indication information, the method further includes:

[0263] The CSI is reconstructed using the second model after switching updates.

[0264] With reference to some embodiments of the fifth aspect, in some embodiments, the first device is a network device, and the second device is a terminal;

[0265] The second model is used to compress CSI, and the first model is used to reconstruct CSI.

[0266] In conjunction with some embodiments of the fifth aspect, in some embodiments, the device is further used to:

[0267] receiving a first request sent by a first device, where the first request is used to request CSI data;

[0268] Send CSI data to the first device.

[0269] In conjunction with some embodiments of the fifth aspect, in some embodiments, the device is further used to:

[0270] receiving a second request sent by the first device, where the first request is for requesting sample data;

[0271] Send sample data to the first device.

[0272] In conjunction with some embodiments of the fifth aspect, in some embodiments, after the second device performs a model switching update based on the indication information, the apparatus is further configured to:

[0273] The second model after switching and updating is used for CSI compression.

[0274] In the sixth aspect, an embodiment of the present disclosure proposes a communication device, which includes: one or more processors; one or more memories for storing instructions; wherein the processor is used to call the instructions so that the communication device executes the model switching update method described in the first aspect, the optional implementation of the first aspect, the second aspect, and the optional implementation of the second aspect.

[0275] In the seventh aspect, an embodiment of the present disclosure proposes a communication system, which includes: a terminal and a network device; wherein the terminal is configured to execute the method described in the first aspect and the optional implementation of the first aspect, and the network device is configured to execute the method described in the second aspect and the optional implementation of the second aspect.

[0276] In an eighth aspect, an embodiment of the present disclosure proposes a storage medium, which stores instructions. When the instructions are executed on a communication device, the communication device executes the method described in the first aspect, the optional implementation of the first aspect, the second aspect, and the optional implementation of the second aspect.

[0277] In a ninth aspect, an embodiment of the present disclosure proposes a program product. When the program product is executed by a communication device, the communication device executes the method described in the first aspect, the optional implementation of the first aspect, the second aspect, and the optional implementation of the second aspect.

[0278] In a tenth aspect, an embodiment of the present disclosure proposes a computer program, which, when executed on a computer, enables the computer to execute the method described in the first aspect, the optional implementation of the first aspect, the second aspect, and the optional implementation of the second aspect.

[0279] It is understandable that the above-mentioned terminals, network devices, communication devices, communication systems, storage media, program products, and computer programs are all used to execute the methods proposed in the embodiments of the present disclosure. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects of the corresponding methods and will not be repeated here.

[0280] The present disclosure provides invention titles. In some embodiments, the terms "sending method" and "receiving method," "information sending method," and "information receiving method" are interchangeable; the terms "communication device" and "information processing device," "information sending device," and "information receiving device" are interchangeable; and the terms "information processing system," "communication system," "information sending system," and "information receiving system" are interchangeable.

[0281] The embodiments of the present disclosure are not exhaustive and are merely illustrative of some embodiments, and are not intended to be a specific limitation on the scope of protection of the present disclosure. In the absence of contradiction, each step in a certain embodiment can be implemented as an independent embodiment, and the steps can be arbitrarily combined. For example, a solution after removing some steps in a certain embodiment can also be implemented as an independent embodiment, and the order of the steps in a certain embodiment can be arbitrarily exchanged. In addition, the optional implementation methods in a certain embodiment can be arbitrarily combined; in addition, the embodiments can be arbitrarily combined. For example, some or all steps of different embodiments can be arbitrarily combined, and a certain embodiment can be arbitrarily combined with the optional implementation methods of other embodiments.

[0282] In each embodiment of the present disclosure, unless otherwise specified or provided for by logic, the terms and / or descriptions between the embodiments are consistent and can be referenced by each other. The technical features in different embodiments can be combined to form a new embodiment based on their inherent logical relationships.

[0283] The terms used in the embodiments of the present disclosure are only for the purpose of describing specific embodiments and are not intended to limit the present disclosure.

[0284] In the embodiments of the present disclosure, unless otherwise specified, elements expressed in the singular, such as "a", "an", "the", "above", "said", "the", "the", etc., may mean "one and only one", or "one or more", "at least one", etc. For example, when using articles such as "a", "an", "the" in English in translation, the noun following the article may be understood as a singular expression or a plural expression.

[0285] In the embodiments of the present disclosure, “plurality” refers to two or more.

[0286] In some embodiments, the terms "at least one of", "at least one of", "at least one of", "one or more", "a plurality of", "multiple", etc. can be used interchangeably.

[0287] In the embodiments of the present disclosure, descriptions such as “at least one of A, B, C…”, “A and / or B and / or C…”, etc. include the situation where any one of A, B, C… exists alone, and also include any combination of any multiple of A, B, C…, and each situation can exist alone; for example, “at least one of A, B, C” includes the situation where A exists alone, B exists alone, C exists alone, the combination of A and B, the combination of A and C, the combination of B and C, and the combination of A, B, and C; for example, A and / or B includes the situation where A exists alone, B exists alone, and the combination of A and B.

[0288] In some embodiments, descriptions such as "in one case A, in another case B," or "in response to one case A, in response to another case B," may include the following technical solutions depending on the situation: executing A independently of B (in some embodiments, A); executing B independently of A (in some embodiments, B); selectively executing A and B (in some embodiments, selecting between A and B); and executing both A and B (in some embodiments, A and B). The same applies when there are more branches, such as A, B, and C.

[0289] The prefixes such as "first" and "second" in the embodiments of the present disclosure are only used to distinguish different description objects and do not constitute any restriction on the position, order, priority, quantity or content of the description objects. For the statement of the description object, please refer to the description in the context of the claims or embodiments, and no unnecessary restriction should be constituted due to the use of prefixes. For example, if the description object is a "field", the ordinal number before the "field" in the "first field" and the "second field" does not limit the position or order between the "fields". "First" and "second" do not limit whether the "fields" they modify are in the same message, nor do they limit the order of the "first field" and the "second field". For another example, if the description object is a "level", the ordinal number before the "level" in the "first level" and the "second level" does not limit the priority between the "levels". For another example, the number of description objects is not limited by the ordinal number and can be one or more. Taking "first device" as an example, the number of "devices" can be one or more. In addition, the objects modified by different prefixes can be the same or different. For example, if the description object is "device", then the "first device" and the "second device" can be the same device or different devices, and their types can be the same or different; for another example, if the description object is "information", then the "first information" and the "second information" can be the same information or different information, and their contents can be the same or different.

[0290] In some embodiments, “including A,” “comprising A,” “used to indicate A,” and “carrying A” can be interpreted as directly carrying A or indirectly indicating A.

[0291] In some embodiments, terms such as "in response to...", "in response to determining...", "in the case of...", "at the time of...", "when...", "if...", "if...", etc. can be used interchangeably.

[0292] In some embodiments, terms such as "greater than", "greater than or equal to", "not less than", "more than", "more than or equal to", "not less than", "higher than", "higher than or equal to", "not less than", and "above" can be replaced with each other, and terms such as "less than", "less than or equal to", "not greater than", "less than", "less than or equal to", "not more than", "lower than", "lower than or equal to", "not higher than", and "below" can be replaced with each other.

[0293] In some embodiments, devices, etc. can be interpreted as physical or virtual, and their names are not limited to the names recorded in the embodiments. Terms such as "device", "equipment", "device", "circuit", "network element", "node", "function", "unit", "section", "system", "network", "chip", "chip system", "entity", and "subject" can be used interchangeably.

[0294] In some embodiments, "network" can be interpreted as devices included in the network (eg, access network equipment, core network equipment, etc.).

[0295] In some embodiments, the terms "access network device (AN device)", "radio access network device (RAN device)", "base station (BS)", "radio base station" "fixed station", "node", "access point", "transmission point (TP)", "reception point (RP)", "transmission / reception point (TRP)", "panel", "antenna panel", "antenna array", "cell", "macro cell", "small cell", "femto cell", "pico cell", "sector", "cell group", "carrier", "component carrier", "bandwidth part (BWP)" and the like may be used interchangeably.

[0296] In some embodiments, the terms "terminal", "terminal device", "user equipment (UE)", "user terminal", "mobile station (MS)", "mobile terminal (MT)", subscriber station, mobile unit, subscriber unit, wireless unit, remote unit, mobile device, wireless device, wireless communication device, remote device, mobile subscriber station, access terminal, mobile terminal, wireless terminal, remote terminal, handset, user agent, mobile client, client, etc. can be used interchangeably.

[0297] In some embodiments, the terminal may be replaced by an access network device, a core network device, or a network device. In this case, the access network device, the core network device, or the network device may have a structure that has all or part of the functions of the terminal.

[0298] In some embodiments, obtaining data, information, etc. may comply with the laws and regulations of the country where the data is obtained.

[0299] In some embodiments, data, information, etc. may be obtained with the user's consent.

[0300] In addition, each element, each row, or each column in the table of the embodiment of the present disclosure can be implemented as an independent embodiment, and the combination of any elements, any rows, and any columns can also be implemented as an independent embodiment.

[0301] The correspondences shown in the tables of the present disclosure can be configured or predefined. The values ​​of the information in each table are merely examples and can be configured to other values, which are not limited by the present disclosure. When configuring the correspondences between information and parameters, it is not necessarily required to configure all the correspondences shown in each table. For example, in the tables of the present disclosure, the correspondences shown in certain rows may not be configured. For another example, appropriate deformation adjustments can be made based on the above tables, such as splitting, merging, etc. The names of the parameters shown in the titles of the above tables may also adopt other names that can be understood by the communication device, and the values ​​or representations of the parameters may also adopt other values ​​or representations that can be understood by the communication device. When implementing the above tables, other data structures may also be used, such as arrays, queues, containers, stacks, linear lists, pointers, linked lists, trees, graphs, structures, classes, heaps, hash tables or hash tables, etc.

[0302] The predefined in the present disclosure may be understood as defined, predefined, stored, pre-stored, pre-negotiated, pre-configured, solidified, or pre-burned.

[0303] Figure 1A is a schematic diagram illustrating the architecture of a communication system according to an embodiment of the present disclosure. As shown in Figure 1A, communication system 100 may include a first device 101 and a second device 102. Optionally, the first device may be, for example, a terminal, and the second device may be, for example, a network device; alternatively, the second device may be, for example, a terminal, and the first device may be, for example, a network device; and the network device may include an access network device and / or a core network device.

[0304] In some embodiments, the terminal includes, for example, a mobile phone, a wearable device, an Internet of Things device, a car with communication function, a smart car, a tablet computer, a computer with wireless transceiver function, a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal device in industrial control, a wireless terminal device in self-driving, a wireless terminal device in remote medical surgery, a wireless terminal device in a smart grid, a wireless terminal device in transportation safety, a wireless terminal device in a smart city, and at least one of a wireless terminal device in a smart home, but is not limited thereto.

[0305] In some embodiments, the access network device is, for example, a node or device that accesses a terminal to a wireless network. The access network device may include an evolved NodeB (eNB), a next generation evolved NodeB (ng-eNB), a next generation NodeB (gNB), a node B (NB), a home node B (HNB), a home evolved nodeB (HeNB), a wireless backhaul device, a radio network controller (RNC), a base station controller (BSC), a base transceiver station (BTS), a base band unit (BBU), a mobile switching center, a base station in a 6G communication system, an open base station (Open RAN), a cloud base station (Cloud RAN), a base station in other communication systems, and at least one of an access node in a wireless fidelity (WiFi) system, but is not limited thereto.

[0306] In some embodiments, the technical solution of the present disclosure can be applied to the Open RAN architecture. In this case, the interfaces between or within the access network devices involved in the embodiments of the present disclosure can be transformed into internal interfaces of the Open RAN, and the processes and information interactions between these internal interfaces can be implemented through software or programs.

[0307] In some embodiments, the access network device can be composed of a centralized unit (CU) and a distributed unit (DU), where the CU can also be called a control unit. The CU-DU structure can be used to split the protocol layer of the access network device, with the functions of some protocol layers centrally controlled by the CU, and the functions of the remaining part or all of the protocol layers distributed in the DU, which is centrally controlled by the CU, but is not limited to this.

[0308] In some embodiments, the core network device may be a device including one or more network elements, or may be multiple devices or a group of devices, each including all or part of one or more network elements. The network element may be virtual or physical. The core network, for example, includes at least one of an Evolved Packet Core (EPC), a 5G Core Network (5GCN), and a Next Generation Core (NGC). Alternatively, the core network device may also be a location management function network element. Exemplarily, the location management function network element includes a location server (location server), which may be implemented as any one of the following: Location Management Function (LMF), Enhanced Serving Mobile Location Centre (E-SMLC), Secure User Plane Location (SUPL), and Secure User Plane Location Platform (SUPLLP).

[0309] It can be understood that the communication system described in the embodiment of the present disclosure is for the purpose of more clearly illustrating the technical solution of the embodiment of the present disclosure, and does not constitute a limitation on the technical solution proposed in the embodiment of the present disclosure. Ordinary technicians in this field can know that with the evolution of the system architecture and the emergence of new business scenarios, the technical solution proposed in the embodiment of the present disclosure is also applicable to similar technical problems.

[0310] The following embodiments of the present disclosure may be applied to the communication system 100 shown in FIG1A , or a portion thereof, but are not limited thereto. The entities shown in FIG1A are illustrative only. The communication system may include all or part of the entities shown in FIG1A , or may include other entities other than those shown in FIG1A . The number and form of the entities may be arbitrary. The connection relationship between the entities is illustrative only. The entities may be connected or disconnected, and the connection may be in any manner, including direct or indirect, wired or wireless.

[0311] The embodiments of the present disclosure can be applied to Long Term Evolution (LTE), LTE-Advanced (LTE-A), LTE-Beyond (LTE-B), SUPER 3G, IMT-Advanced, 4th generation mobile communication system (4G), 5th generation mobile communication system (5G), 5G new radio (NR), future radio access (FRA), new radio access technology (RAT), new radio (NR), new radio access (NX), future generation radio access (FX), Global System for Mobile communications (GSM (registered trademark)), CDMA2000, Ultra Mobile Broadband (UMB), IEEE 802.11 (Wi-Fi (registered trademark)), IEEE 802.16 (WiMAX (registered trademark)), IEEE 802.20, Ultra-WideBand (UWB), Bluetooth (registered trademark), Public Land Mobile Network (PLMN) networks, Device-to-Device (D2D) systems, Machine-to-Machine (M2M) systems, Internet of Things (IoT) systems, Vehicle-to-Everything (V2X), systems using other model switching and update methods, and next-generation systems based on and extending these systems. Furthermore, multiple systems may be combined (for example, a combination of LTE or LTE-A with 5G).

[0312] Optionally, FIG1B is a structural block diagram of a method for "compressed reporting and reconstruction of CSI based on a bilateral model" in the above-mentioned background technology provided by an embodiment of the present disclosure. In the method for "compressed reporting and reconstruction of CSI based on a bilateral model", an encoder model is deployed on the terminal side, and a decoder model is deployed on the network device side. The encoder model and the decoder model together constitute a feedback network model to implement CSI feedback. Optionally, as shown in FIG1B , the CSI feedback process may include: the encoder model deployed on the terminal side performs feature extraction and dimensionality compression on the original CSI data to compress the original CSI data into codewords with a smaller data volume, and then transmits the codeword information to the network device via an uplink feedback link. Since the codeword information transmitted by the terminal is the codeword compressed from the original CSI data, the feedback overhead can be reduced. In addition, after receiving the codeword information, the network device can perform CSI reconstruction through the decoder model to output CSI data of the original dimension. Optionally, to improve the accuracy of CSI feedback, a data set of a certain size is usually used to train the above-mentioned feedback network model, so that the feedback network model can fully learn the characteristic distribution of the CSI data in the current training set. Then, when faced with CSI data with the same or similar characteristic distribution, it can minimize the difference between the CSI data reconstructed by the decoder model and the original CSI data input by the encoder model, thereby achieving CSI compression, feedback, and reconstruction with lower error.

[0313] Alternatively, a current training method for the feedback network model is an offline model training method, that is, using a channel scenario i A large-scale data set collected in advance i Complete offline training of the feedback network model and deploy the feedback network model in the terminals and network devices in the actual system. During the model application process in the actual system, the feedback network model parameters remain fixed.

[0314] However, the actual channel environment is complex and time-varying, and terminals are highly mobile. Therefore, a variety of channel scenarios may exist in actual communication systems, and the characteristic distribution of the channel data is unknown. Moreover, real-time channel characteristic changes may occur within a channel scenario. If a deep learning-based offline model training method is used to train the feedback network model, since the trained feedback network model parameters are fixed, the trained model may have difficulty processing CSI data with an unknown distribution that differs from the characteristic distribution of the training dataset. This means that the model is prone to mismatch between the model and the data, resulting in a decrease in CSI feedback accuracy (i.e., the CSI data reconstructed by the decoder model differs significantly from the original CSI data input by the encoder model), which in turn affects the overall system performance.

[0315] Moreover, since the actual channel environment is relatively complex, in order to ensure the overall performance of the feedback network under various circumstances, it is usually necessary to independently train multiple sets of feedback network models for a variety of different channel scenarios to complete the CSI feedback tasks in different channel scenarios. For example, due to the large differences between indoor and outdoor channel scenarios, it is difficult to achieve high-precision CSI feedback in both scenarios using only one model. Therefore, the indoor and outdoor feedback network models are trained independently using their respective data sets in the two channel scenarios, and are used to complete the CSI compression and reconstruction work in indoor and outdoor channel scenarios, respectively. However, the offline training method only trains a feedback network model for one channel scenario. The trained feedback network model cannot be applied to CSI feedback in a variety of different channel scenarios. There will also be a mismatch between the model and the data distribution characteristics, which will lead to a significant decrease in feedback accuracy, thereby affecting the operation of the system.

[0316] FIG2A is an interactive diagram illustrating a model switching and updating method according to an embodiment of the present disclosure. As shown in FIG2A , the embodiment of the present disclosure relates to a model switching and updating method, which is used in a communication system 100. In the embodiment of FIG2A , the first device is a terminal and the second device is a network device. The method includes:

[0317] Step 2101: The first device determines model-related information corresponding to the second device.

[0318] Optionally, in some embodiments, one or more first models may be deployed on the first device, and one or more second models may be deployed on the second device, the first model and the second model corresponding to each other, and optionally, the corresponding first model and second model can be used to implement CSI compression and reconstruction. Optionally, when the first device is a terminal and the second device is a network device, the first model can be used to compress CSI, and the second model can be used to reconstruct CSI; for example, the first model can be an encoder model, and the second model can be a decoder model, but is not limited thereto; for example, if the first model #1 corresponds to the second model 1, then after determining the original CSI data, the first device can use the first model #1 to compress the original CSI data into codeword information and send it to the second device, and after the second device receives the codeword information, it can use the second model #1 to reconstruct the CSI data of the original dimension (i.e., data close to the original CSI data) based on the codeword information. Optionally, when the first device is a network device and the second device is a terminal, the first model can be used to reconstruct CSI, and the second model can be used to compress CSI. For example, the first model can be a decoder model, and the second model can be an encoder model, but is not limited thereto. In some embodiments, the first model and the second model can be artificial intelligence (AI) or machine learning (ML) models.

[0319] It should be noted that, in the embodiment of FIG. 2A , the first device is a terminal and the second device is a network device.

[0320] Optionally, in some embodiments, the model-related information may be used to reflect the model deployment and usage of the second device. Optionally, the model-related information may include at least one of the following:

[0321] The second model currently used by the second device may include, for example, a model number of the second model currently used by the second device;

[0322] The at least one second model deployed by the second device may include, for example, a model number of the at least one second model deployed by the second device;

[0323] The model structure corresponding to at least one second model deployed by the second device.

[0324] Optionally, in some embodiments, the model structure may be a model structure to which the second model belongs, such as: the second model belongs to a neural network model or an encoder model, etc.; and / or, in some embodiments, the model structure may be an internal model structure of the second model, such as: the number of convolutional layers of the second model, etc.

[0325] Optionally, in some embodiments, the method for the first device to determine the model-related information corresponding to the second device may include: the first device receiving the model-related information corresponding to the second device sent by the second device.

[0326] Optionally, in some other embodiments, the first device may also send model-related information corresponding to the first device to the second device.

[0327] Step 2102: The first device determines a first parameter value.

[0328] Optionally, in some embodiments, the first parameter value may be used to determine whether a model switching update is required.

[0329] Optionally, in some embodiments, the method for the first device to determine the first parameter value may include at least one of the following:

[0330] The first device receives a first parameter value sent by the second device;

[0331] The first device determines a first parameter value configured in itself;

[0332] The first device determines a first parameter value based on a protocol agreement.

[0333] Optionally, in some embodiments, when the model-related information of the second device is sent by the second device to the first device, and when the first parameter value is sent by the second device to the first device, the second device can use the same message to send the first parameter value and the model-related information of the second device to the first device at the same time; or, the first parameter value and the model-related information of the second device can also be included in different messages, such as the second device can first send the first parameter value and then send the model-related information of the second device; or, the second device can first send the model-related information of the second device and then send the first parameter value.

[0334] Step 2103: The first device performs performance monitoring on the first model currently used by the first device and / or the second model currently used by the second device.

[0335] Optionally, in some embodiments, the first device may determine real-time CSI data, which may be obtained by the first device (i.e., the terminal) by measuring the pilot signal sent by the second device (i.e., the network device) and estimating the channel. In addition, the first device may use the determined CSI data to perform performance monitoring on the first model currently used by the first device and / or the second model currently used by the second device to obtain a performance monitoring result. Optionally, in some embodiments, the performance monitoring result may be used to reflect the CSI feedback accuracy of the first model and the second model. Optionally, the feedback accuracy may be understood, for example, as: the error value between the CSI data reconstructed by the second model and the original CSI data input to the first model.

[0336] Step 2104: The performance monitoring result meets the first preset condition, and the first device determines that a model switch update is required.

[0337] Optionally, in some embodiments, the first preset condition may include: the performance monitoring result is less than the first parameter value. Optionally, the first parameter value may be an accuracy threshold value; when the performance monitoring result corresponding to the first model currently used by the first device and / or the second model currently used by the second device meets the first preset condition, it indicates that the performance monitoring result corresponding to the first model currently used by the first device and / or the second model currently used by the second device is low, that is, the CSI feedback accuracy of the first model currently used by the first device and / or the second model currently used by the second device is low, and at this time, it is considered necessary to perform a model switching update.

[0338] Step 2105: The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter.

[0339] Optionally, the above-mentioned first target model may be a target model to be switched by the first device, and the second target model may be a target model to be switched by the second device.

[0340] Optionally, the above-mentioned first model update parameters can be used to update the model parameters of the first model deployed on the first device. For example, the first model update parameters can be used to update the model parameters of the first model currently used by the first device, or can be used to update the model parameters of the target first model; the above-mentioned second model update parameters can be used to update the model parameters of the second model deployed on the second device. For example, the second model update parameters can be used to update the model parameters of the second model currently used by the second device, or can be used to update the model parameters of the target second model.

[0341] Optionally, in some embodiments, the method for the first device to determine at least one of the target first model, the first model update parameter, the target second model, and the second model update parameter may include:

[0342] Step a: The first model currently used by the first device is the first model uniquely deployed by the first device, and / or the second model currently used by the second device is the second model uniquely deployed by the second device, and sample data is determined.

[0343] Optionally, the sample data may be: CSI data obtained by the first device measuring a pilot signal and estimating a channel within a first preset time period. Optionally, the first preset time period may be pre-set, for example, a period before the current time period.

[0344] Step b: training the first model currently used by the first device and / or the second model currently used by the second device based on the sample data until the performance monitoring results of the trained first model and / or the trained second model meet a second preset condition, and then stopping the training;

[0345] Optionally, in some embodiments, the performance monitoring results of the trained first model and / or the trained second model may be the feedback accuracy of the model on the training data set after completing one cycle of model training, and the feedback accuracy may be directly obtained during the model training process.

[0346] Optionally, the second preset condition may be, for example, that the performance monitoring result is greater than or equal to the first parameter value. Optionally, when the performance monitoring result of the trained first model and / or the trained second model satisfies the second preset condition, indicating that the CSI feedback accuracy of the trained first model and / or the trained second model is high, the CSI feedback accuracy can be ensured when the trained first model and / or the trained second model are used for CSI feedback, and therefore training may be stopped.

[0347] Step c: determining the model parameters of the trained first model as first model update parameters, and / or determining the model parameters of the trained second model as second model update parameters.

[0348] Optionally, in some other embodiments, the method for the first device to determine at least one of the target first model, the first model update parameter, the target second model, and the second model update parameter may include:

[0349] Step 1: The first device deploys at least two first models, and / or the second device deploys at least two second models, and determines at least one model group, where the model group includes the first model and the second model corresponding to each other;

[0350] Step 2: monitoring the performance of the first model and / or the second model in each model group;

[0351] Step 3: If the performance monitoring result of at least one model group meets the second preset condition, the first model in the model group with the highest performance monitoring result is determined as the target first model, and the second model in the model group with the highest performance monitoring result is determined as the target second model.

[0352] Optionally, in some further embodiments, the method for the first device to determine at least one of the target first model, the first model update parameter, the target second model, and the second model update parameter may include:

[0353] Step 1: The first device deploys at least two first models, and / or the second device deploys at least two second models, and determines at least one model group, where the model group includes the first model and the second model corresponding to each other;

[0354] Step 2: Monitor the performance of each model group;

[0355] Step 3: If the performance monitoring results of each model group do not meet the second preset condition, the first model in the model group with the highest performance monitoring result is determined as the target first model, and the second model in the model group with the highest performance monitoring result is determined as the target second model;

[0356] Step 4: Determine sample data;

[0357] Step 5: Training the target first model and the target second model based on the sample data until the performance monitoring results of the trained first model and / or the trained second model meet the second preset condition, and then stopping the training;

[0358] Step 6: Determine the model parameters of the trained target first model as first model update parameters, and / or determine the model parameters of the trained target second model as second model update parameters.

[0359] Step 2106: The first device performs a model switching update and sends an indication message to the second device.

[0360] Optionally, the indication information may be used to indicate a target second model and / or second model update parameters.

[0361] Optionally, in some embodiments, when the first model currently used by the first device is the only first model deployed by the first device, and / or the second model currently used by the second device is the only second model deployed by the second device, the method for the first device to perform a model switching update and send indication information to the second device may include at least one of the following:

[0362] The first device updates the model parameters of the currently used first model to first model update parameters;

[0363] The first device sends instruction information to the second device, where the instruction information may include the second model update parameters and / or the first instruction; the first instruction is used to instruct the second device to update the model parameters of the second model currently in use to the second model update parameters. Optionally, in some embodiments, the first device may quantize the second model update parameters, and the second model update parameters included in the instruction information may be the quantized second model update parameters.

[0364] Optionally, in some other embodiments, when the first device deploys at least two first models, and / or the second device deploys at least two second models, and the performance monitoring result of at least one model group (i.e., the model group includes the first model and the second model corresponding to each other) meets the second preset condition, the first device performs a model switching update and sends an indication message to the second device, including at least one of the following:

[0365] The first device deactivates the currently used first model and switches from the currently used first model to the target first model;

[0366] The first device sends indication information to the second device, where the indication information may include the second indication and / or relevant information of the target second model; optionally, the relevant information of the target second model may be: a model number of the target second model.

[0367] The second indication may be used to indicate at least one of the following:

[0368] The second device deactivates the second model currently in use;

[0369] The second device switches from the currently used second model to the target second model.

[0370] Optionally, in some further embodiments, when the first device deploys at least two first models, and / or the second device deploys at least two second models, and the performance monitoring results of each group (i.e., the model group includes the first model and the second model corresponding to each other) do not meet the second preset condition, the first device performs a model switching update and sends an indication message to the second device, including at least one of the following:

[0371] The first device deactivates the currently used first model, switches from the currently used first model to the target first model, and updates the model parameters of the target first model to the first model update parameters;

[0372] The first device sends indication information to the second device, where the indication information may include at least one of relevant information of the target second model, second model update parameters, and a third indication, where the third indication is used to indicate at least one of the following:

[0373] The second device deactivates the second model currently in use;

[0374] The second device switches from the currently used second model to the target second model;

[0375] The second device updates the model parameters of the target second model to second model update parameters.

[0376] Step 2107: The second device performs model switching update based on the indication information.

[0377] Optionally, in some embodiments, when the indication information includes the second model update parameter and / or the first indication; the above-mentioned method for model switching update based on the indication information may include: the second device updates the model parameters of the currently used second model to the second model update parameter.

[0378] Optionally, in other embodiments, when the indication information includes relevant information of the second indication and / or the target second model; the above-mentioned method of model switching update based on the indication information may include: the second device deactivates the currently used second model and switches from the currently used second model to the target second model.

[0379] Optionally, in some embodiments, when the indication information includes relevant information of the target second model, second model update parameters, and at least one of the third indications; the above-mentioned method for model switching update based on the indication information may include: the second device deactivates the currently used second model, switches from the currently used second model to the target second model, and updates the model parameters of the target second model to the second model update parameters.

[0380] Step 2108: The second device sends a confirmation message.

[0381] Optionally, in some embodiments, the confirmation information may be used to indicate that the second device has completed the model switching update.

[0382] Step 2109: The first device uses the first model after switching and updating to perform CSI compression.

[0383] Optionally, the first device may perform CSI compression using the target first model determined above.

[0384] Step 2110: The second device uses the second model after switching and updating to perform CSI reconstruction.

[0385] Optionally, the second device may use the target second model determined above to perform CSI reconstruction.

[0386] The model switching update method involved in the embodiment of the present disclosure may include at least one of steps S2101 to S2110. For example, step S2101 may be implemented as an independent embodiment, step S2102 may be implemented as an independent embodiment, and step S2103 may be implemented as an independent embodiment, but is not limited thereto.

[0387] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0388] FIG2B is an interactive diagram illustrating a model switching and updating method according to an embodiment of the present disclosure. As shown in FIG2B , the embodiment of the present disclosure relates to a model switching and updating method, which is used in a communication system 100. In the embodiment of FIG2A , the first device is a network device and the second device is a terminal.

[0389] Step 2201: The first device determines model-related information corresponding to the second device.

[0390] Step 2202: The first device determines a first parameter value.

[0391] Step 2203: The first device sends a first request to the second device, where the first request is used to request CSI data.

[0392] Optionally, in some embodiments, the first request may be used to request real-time CSI data from the second device. The CSI data may be obtained by the second device (i.e., terminal) by measuring a pilot signal sent by the first device (i.e., network device) and performing channel estimation.

[0393] Step 2204: The second device sends CSI data to the first device.

[0394] Optionally, in some embodiments, the second device sends the real-time CSI data of the second device to the first device. Optionally, in some embodiments, the second device may quantize the real-time CSI data and send the quantized CSI data to the first device.

[0395] Step 2205: The first device performs performance monitoring on the first model currently used by the first device and / or the second model currently used by the second device.

[0396] Step 2206: The performance monitoring result meets the first preset condition, and the first device determines that a model switch update is required.

[0397] Step 2207: The first device determines at least one of the target first model, the first model update parameters, the target second model, and the second model update parameters.

[0398] Optionally, when the first device is a network device, the method by which the first device determines at least one of the target first model, the first model update parameters, the target second model, and the second model update parameters is consistent with the method of step 2105 above, except that, when the first device is a network device, when determining the sample data for training the first model and / or the second model, the first device does not directly determine the CSI data measured and estimated within the first preset time period as in step 2105 above, but the first device (i.e., the network device) needs to send a second request to the second device (i.e., the terminal), and the second request is used to request the sample data; thereafter, the first device receives the sample data sent by the second device. Optionally, in some embodiments, after the second device receives the second request and determines the sample data, it can quantize the sample data and send the quantized sample data to the first device. Other steps are consistent with step 2105 above.

[0399] Step 2208: The first device performs a model switching update and sends an indication message to the second device.

[0400] Step 2209: The second device performs model switching update based on the indication information.

[0401] Step 2210: The second device sends a confirmation message.

[0402] Step 2211: The second device uses the second model after switching and updating to perform CSI compression.

[0403] Step 2212: The first device uses the first model after switching and updating to perform CSI reconstruction.

[0404] For a detailed description of steps 2201 - 2212 , please refer to the above embodiment description.

[0405] The model switching and updating method according to the embodiment of the present disclosure may include at least one of steps S2201 to S2212. For example, step S2201 may be implemented as an independent embodiment, and step S2202 may be implemented as an independent embodiment, but the present invention is not limited thereto.

[0406] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0407] FIG3A is an interactive diagram of a model switching update method according to an embodiment of the present disclosure. As shown in FIG3A , the embodiment of the present disclosure relates to a model switching update method for a first device, which is a terminal. The method includes:

[0408] Step 3101: The first device determines model-related information corresponding to the second device.

[0409] Step 3102: The first device determines a first parameter value.

[0410] Step 3103: The first device performs performance monitoring on the first model currently used by the first device and / or the second model currently used by the second device;

[0411] Step 3104: The performance monitoring result meets the first preset condition, and the first device determines that a model switching update is required.

[0412] Step 3105: The first device determines at least one of the target first model, the first model update parameters, the target second model, and the second model update parameters.

[0413] Step 3106: The first device performs a model switching update and sends an indication message to the second device.

[0414] Step 3107: The first device receives confirmation information sent by the second device.

[0415] Step 3108: The first device uses the first model after switching and updating to perform CSI compression.

[0416] For a detailed description of steps 3101 - 3108 , please refer to the above embodiment description.

[0417] The model switching and updating method according to the embodiment of the present disclosure may include at least one of steps S3101 to S3108. For example, step S3101 may be implemented as an independent embodiment, and step S3102 may be implemented as an independent embodiment, but the present invention is not limited thereto.

[0418] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0419] FIG3B is an interactive diagram of a model switching update method according to an embodiment of the present disclosure. As shown in FIG3B , the embodiment of the present disclosure relates to a model switching update method for a first device, which is a network device, and the method includes:

[0420] Step 3201: The first device determines model-related information corresponding to the second device.

[0421] Step 3202: The first device determines a first parameter value.

[0422] Step 3203: The first device sends a first request to the second device.

[0423] Step 3204: The first device receives CSI data sent by the second device.

[0424] Step 3205: The first device performs performance monitoring on the first model currently used by the first device and / or the second model currently used by the second device.

[0425] Step 3206: The performance monitoring result meets the first preset condition, and the first device determines that a model switch update is required.

[0426] Step 3207: The first device determines at least one of the target first model, the first model update parameters, the target second model, and the second model update parameters.

[0427] Step 3208: The first device performs a model switching update and sends an indication message to the second device.

[0428] Step 3209: The first device receives confirmation information sent by the second device.

[0429] Step 3210: The first device uses the first model after switching and updating to perform CSI reconstruction.

[0430] For a detailed description of steps 3201 - 3210 , please refer to the above embodiment description.

[0431] The model switching and updating method according to the embodiment of the present disclosure may include at least one of steps S3201 to S3210. For example, step S3201 may be implemented as an independent embodiment, and step S3202 may be implemented as an independent embodiment, but the present invention is not limited thereto.

[0432] Optionally, the above step 3206 may be an optional step and may be performed optionally, for example, may be performed or not performed.

[0433] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0434] FIG3C is an interactive diagram of a model switching update method according to an embodiment of the present disclosure. As shown in FIG3C , the embodiment of the present disclosure relates to a model switching update method for a first device, the method comprising:

[0435] Step 3301: The first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device.

[0436] Step 3302: The performance monitoring result meets the first preset condition, and it is determined that a model switching update is required;

[0437] Step 3303: The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter.

[0438] Step 3304: The first device performs a model switching update and sends indication information to the second device, where the indication information indicates the target second model and / or second model update parameters.

[0439] Optionally, the first model and the second model correspond to each other, and the first model and the second model are used to implement compression and reconstruction of channel state information CSI;

[0440] Before the first device performs performance monitoring on the first model currently used by the first device and / or the second model currently used by the second device, the method further includes:

[0441] Determine model-related information corresponding to the second device; the model-related information includes at least one of the following:

[0442] a second model currently used by the second device;

[0443] at least one second model deployed by the second device;

[0444] The model structure corresponding to the at least one second model deployed by the second device.

[0445] Optionally, determining model-related information corresponding to the second device includes:

[0446] Receive the model-related information sent by the second device.

[0447] Optionally, before determining that a model switch update is required, the method further includes:

[0448] The first device determines a first parameter value, where the first parameter value is used to determine whether a model switching update is required;

[0449] Among them, the first preset condition includes: the performance monitoring result is less than the first parameter value.

[0450] Optionally, the first device determines the first parameter value, including at least one of the following:

[0451] The first device receives the first parameter value sent by the second device;

[0452] The first device determines the first parameter value configured in the first device;

[0453] The first device determines the first parameter value based on a protocol agreement.

[0454] Optionally, the first device determines at least one of the target first model, the first model update parameter, the target second model, and the second model update parameter, including:

[0455] The first model currently used by the first device is the only first model deployed by the first device, and / or the second model currently used by the second device is the only second model deployed by the second device, and sample data is determined;

[0456] Training the first model currently used by the first device and / or the second model currently used by the second device based on the sample data until a performance monitoring result of the trained first model and / or the trained second model meets a second preset condition, and then stopping the training;

[0457] The model parameters of the trained first model are determined as the first model update parameters, and / or the model parameters of the trained second model are determined as the second model update parameters.

[0458] Optionally, the first device performs model switching update and sends indication information to the second device, including at least one of the following:

[0459] The first device updates the model parameters of the currently used first model to the first model update parameters;

[0460] The first device sends indication information to the second device, where the indication information includes second model update parameters and / or a first indication; the first indication is used to instruct the second device to update the model parameters of the currently used second model to the second model update parameters.

[0461] Optionally, the first device determines at least one of the target first model, the first model update parameter, the target second model, and the second model update parameter, including:

[0462] The first device deploys at least two first models, and / or the second device deploys at least two second models, and determines at least one model group, wherein the model group includes first models and second models corresponding to each other;

[0463] Performing performance monitoring on the first model and / or the second model in each model group;

[0464] If the performance monitoring result of at least one of the model groups meets the second preset condition, the first model in the model group with the highest performance monitoring result is determined as the target first model, and the second model in the model group with the highest performance monitoring result is determined as the target second model.

[0465] Optionally, the first device performs model switching update and sends indication information to the second device, including at least one of the following:

[0466] The first device deactivates the currently used first model and switches from the currently used first model to the target first model;

[0467] The first device sends indication information to the second device, where the indication information includes a second indication and / or relevant information of a target second model; the second indication is used to indicate at least one of the following:

[0468] The second device deactivates the second model currently in use;

[0469] The second device switches from the currently used second model to the target second model.

[0470] Optionally, the first device determines at least one of the target first model, the first model update parameter, the target second model, and the second model update parameter, including:

[0471] The first device deploys at least two first models, and / or the second device deploys at least two second models, and determines at least one model group, wherein the model group includes first models and second models corresponding to each other;

[0472] Monitor the performance of each model group;

[0473] If the performance monitoring results of each model group do not meet the second preset condition, the first model in the model group with the highest performance monitoring result is determined as the target first model, and the second model in the model group with the highest performance monitoring result is determined as the target second model;

[0474] Determine sample data;

[0475] Training the target first model and the target second model based on the sample data until a performance monitoring result of the trained first model and / or the trained second model meets a second preset condition, and then stopping the training;

[0476] The model parameters of the trained target first model are determined as the first model update parameters, and / or the model parameters of the trained target second model are determined as the second model update parameters.

[0477] Optionally, the first device performs model switching update and sends indication information to the second device, including at least one of the following:

[0478] The first device deactivates the currently used first model, switches from the currently used first model to the target first model, and updates the model parameters of the target first model to the first model update parameters;

[0479] The first device sends indication information to the second device, where the indication information includes at least one of relevant information of the target second model, a second model update parameter, and a third indication, where the third indication is used to indicate at least one of the following:

[0480] The second device deactivates the second model currently in use;

[0481] The second device switches from the currently used second model to the target second model;

[0482] The second device updates the model parameters of the target second model to the second model update parameters.

[0483] Optionally, the second preset condition includes: the performance monitoring result is greater than or equal to a first parameter value.

[0484] Optionally, after the first device sends the indication information to the second device, the method further includes:

[0485] Receive confirmation information sent by the second device, where the confirmation information is used to indicate that the second device has completed the model switching update.

[0486] Optionally, the first device is a terminal, and the second device is a network device;

[0487] The first model is used to compress CSI, and the second model is used to reconstruct CSI.

[0488] Optionally, the first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device, including:

[0489] The first device uses the CSI data to monitor the performance of a first model currently used by the first device and / or a second model currently used by the second device.

[0490] Optionally, after the first device sends the indication information to the second device, the method further includes:

[0491] The first model after switching and updating is used for CSI compression.

[0492] Optionally, the first device is a network device, and the second device is a terminal;

[0493] The second model is used to compress CSI, and the first model is used to reconstruct CSI.

[0494] Optionally, the first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device, including:

[0495] The first device sends a first request to the second device, where the first request is used to request CSI data;

[0496] The first device receives the CSI data sent by the second device;

[0497] The first device uses the CSI data to monitor the performance of a first model currently used by the first device and / or a second model currently used by the second device.

[0498] Optionally, determining the sample data includes:

[0499] The first device sends a second request to the second device, where the second request is used to request the sample data;

[0500] The first device receives the sample data sent by the second device.

[0501] Optionally, after the first device sends the indication information to the second device, the method further includes:

[0502] The CSI reconstruction is performed using the first model after switching updates.

[0503] Optionally, the CSI data is obtained by the terminal through measuring a pilot signal sent by a network device and through channel estimation.

[0504] Optionally, the sample data is: CSI data measured and estimated by the terminal within a first preset time period.

[0505] Optionally, the performance monitoring result is used to reflect the CSI feedback accuracy of the first model and the second model.

[0506] For a detailed description of steps 3301 - 3304 , please refer to the above embodiment description.

[0507] The model switching and updating method according to the embodiment of the present disclosure may include at least one of steps S3301 to S3310. For example, step S3301 may be implemented as an independent embodiment, and step S3302 may be implemented as an independent embodiment, but the present invention is not limited thereto.

[0508] Optionally, the above step 3306 may be an optional step and may be performed optionally, for example, it may be performed or not performed.

[0509] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0510] FIG4A is an interactive diagram of a model switching update method according to an embodiment of the present disclosure. As shown in FIG4A , the embodiment of the present disclosure relates to a model switching update method for a second device 102, which is a network device. The method includes:

[0511] Step 4101: The second device sends model-related information corresponding to the second device to the first device.

[0512] Step 4102: The second device sends a first parameter value to the first device.

[0513] Step 4103: The second device receives the indication information sent by the first device.

[0514] Step 4104: The second device performs model switching update based on the indication information.

[0515] Step 4105: The second device sends a confirmation message to the first device.

[0516] Step 4106: The second device uses the second model after switching and updating to perform CSI reconstruction.

[0517] For a detailed description of steps 4101 - 4106 , please refer to the above embodiment description.

[0518] The model switching and updating method according to the embodiment of the present disclosure may include at least one of steps S4101 to S4106. For example, step S4101 may be implemented as an independent embodiment, and step S4102 may be implemented as an independent embodiment, but the present invention is not limited thereto.

[0519] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0520] FIG4B is an interactive diagram of a model switching update method according to an embodiment of the present disclosure. As shown in FIG4B , the embodiment of the present disclosure relates to a model switching update method for a second device 102, where the second device is a terminal. The method includes:

[0521] Step 4201: The second device sends model-related information corresponding to the second device to the first device.

[0522] Step 4202: The second device sends a first parameter value to the first device.

[0523] Step 4203: The second device receives the first request sent by the first device.

[0524] Step 4204: The second device sends CSI data to the first device.

[0525] Step 4205: The second device receives the indication information sent by the first device.

[0526] Step 4206: The second device performs model switching update based on the indication information.

[0527] Step 4207: The second device sends a confirmation message to the first device.

[0528] Step 4208: The second device uses the second model after switching and updating to perform CSI compression.

[0529] For a detailed description of steps 4201-4208, please refer to the above embodiment description.

[0530] The model switching and updating method according to the embodiment of the present disclosure may include at least one of steps S4201 to S4206. For example, step S4201 may be implemented as an independent embodiment, and step S4202 may be implemented as an independent embodiment, but the present invention is not limited thereto.

[0531] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0532] FIG4C is an interactive diagram of a model switching update method according to an embodiment of the present disclosure. As shown in FIG4C , the embodiment of the present disclosure relates to a model switching update method for the second device 102, the method comprising:

[0533] Step 4301: The second device receives indication information sent by the first device.

[0534] Step 4302: The second device performs model switching update based on the indication information.

[0535] Optionally, the indication information indicates the target second model and / or second model update parameters, wherein the target second model is the target model to be switched by the second device.

[0536] Optionally, a first model is deployed on the first device, and a second model is deployed on the second device; the first model and the second model correspond to each other, and the first model and the second model are used to realize compression and reconstruction of CSI.

[0537] Optionally, before the second device receives the indication information sent by the first device, the method further includes:

[0538] Sending model-related information corresponding to the second device to the first device; the model-related information includes at least one of the following:

[0539] a second model currently used by the second device;

[0540] at least one second model deployed by the second device;

[0541] The model structure corresponding to the at least one second model deployed by the second device.

[0542] Optionally, before the second device receives the indication information sent by the first device, the method further includes:

[0543] A first parameter value is sent to the first device, where the first parameter value is used to determine whether a model switching update is required.

[0544] Optionally, the indication information includes a second model update parameter and / or a first indication; the first indication is used to instruct the second device to update the model parameters of the second model currently in use to the second model update parameter;

[0545] The performing model switching and updating based on the indication information includes:

[0546] The second device updates the model parameters of the currently used second model to the second model update parameters.

[0547] Optionally, the indication information includes a second indication and / or relevant information of the target second model; the second indication is used to indicate at least one of the following: the second device deactivates the currently used second model, or the second device switches from the currently used second model to the target second model;

[0548] The performing model switching and updating based on the indication information includes:

[0549] The second device deactivates the currently used second model and switches from the currently used second model to the target second model.

[0550] Optionally, the indication information includes at least one of relevant information of the target second model, second model update parameters, and a third indication, wherein the third indication is used to indicate at least one of the following: the second device deactivates the currently used second model, the second device switches from the currently used second model to the target second model, and the second device updates the model parameters of the target second model to the second model update parameters;

[0551] The performing model switching and updating based on the indication information includes:

[0552] The second device deactivates the currently used second model, switches from the currently used second model to the target second model, and updates the model parameters of the target second model to the second model update parameters.

[0553] Optionally, after the second device performs model switching update based on the indication information, the method further includes:

[0554] Sending a confirmation message to the first device, where the confirmation message is used to indicate that the second device has completed the model switching update.

[0555] Optionally, the first device is a terminal, and the second device is a network device;

[0556] The first model is used to compress CSI, and the second model is used to reconstruct CSI.

[0557] Optionally, after the second device performs model switching update based on the indication information, the method further includes:

[0558] The CSI is reconstructed using the second model after switching updates.

[0559] Optionally, the first device is a network device, and the second device is a terminal;

[0560] The second model is used to compress CSI, and the first model is used to reconstruct CSI.

[0561] Optionally, after the second device performs model switching update based on the indication information, the method further includes:

[0562] The second model after switching and updating is used for CSI compression.

[0563] For a detailed description of steps 4301-4302, please refer to the above embodiment description.

[0564] The model switching and updating method involved in the embodiment of the present disclosure may include at least one of steps S4301 and S4302. For example, step S4301 may be implemented as an independent embodiment, and step S4302 may be implemented as an independent embodiment, but the present invention is not limited thereto.

[0565] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0566] FIG5A is a flow chart of a model switching update method according to an embodiment of the present disclosure. As shown in FIG5A , the embodiment of the present disclosure relates to a model switching update method for a communication system, wherein the method includes at least one of the following:

[0567] Step 5101: A first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by a second device;

[0568] Step 5102: The performance monitoring result meets the first preset condition, and the first device determines that a model switching update is required;

[0569] Step 5103: The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter.

[0570] Step 5104: The first device performs a model switching update and sends indication information to the second device, where the indication information indicates the target second model and / or the second model update parameters.

[0571] Step 5105: The second device receives the indication information sent by the first device.

[0572] Step 5106: The second device performs model switching update based on the indication information.

[0573] Optional implementations of steps 5101 to 5106 may refer to the description of the above embodiment.

[0574] In some embodiments, the above method may include the methods of the above embodiments of the communication system side, terminal side, network terminal side, etc., which will not be repeated here.

[0575] The model switching and updating method involved in the embodiment of the present disclosure may include at least one of steps S5101 to S5107. For example, step S5101 may be implemented as an independent embodiment, and step S5102 may be implemented as an independent embodiment, but the present invention is not limited thereto.

[0576] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0577] The following is an exemplary introduction to the above method.

[0578] FIG5B is a flow chart of a model switching update method according to an embodiment of the present disclosure. As shown in FIG5B , the method includes:

[0579] 1. Network equipment (base station)

[0580] 1. Receive information such as the AI / ML model type and parameters for CSI feedback reported by the terminal, including: Model number currently used by the terminal current , model structure information, the number of the AI / ML model deployed by the terminal for the CSI feedback task Model1, Model2,…, Model n .

[0581] 2. Send relevant parameters for AI / ML model switching and updating to the terminal. These parameters include: the primary basis for the terminal to determine whether a model switch or update is necessary, such as the AI / ML model accuracy threshold λ. If the current feedback accuracy is higher than λ, the current AI / ML model will continue to be used for CSI compression and reconstruction. If the current feedback accuracy is lower than λ, a model switch or update is required.

[0582] 3. Receive the model switching and update request information and related parameters sent by the terminal, including: the instruction Indicatordeactivate to deactivate the current CSI feedback network decoder model, the model switching instruction Indicator switch Or model update instruction Indicator update ; Upon receiving the model update instruction Indicator update In the case of , the decoder model parameters θ′ after quantization are received at the same time. de .

[0583] Disable the currently working CSI feedback network decoder model and switch and update the network device model:

[0584] a) If the received instruction is Indicator switch , according to the model number Model carried in the instruction opt , switch the decoder model to the model corresponding to this number.

[0585] b) If the received instruction is an Indicator update , dequantize the received model parameters, and then replace the currently used decoder model parameters with θ′ de .

[0586] 4. Send model switching, update confirmation information to the terminal, the confirmation information includes: the network device has completed the model switching or update confirmation indicator complete .

[0587] 5. Use the switched or updated AL / ML decoder model for subsequent CSI reception.

[0588] 2. Terminal

[0589] 1. Report AI / ML model type, parameters, and other information for CSI feedback to the network device, including the model number currently used by the terminal. current , model structure information, model number Model1, Model2,…, Model deployed by the terminal n .

[0590] 2. Receive relevant parameters for AI / ML model switching and updating sent by the network device. The parameters include: the main basis for the terminal to determine whether to switch or update the model, such as the accuracy threshold λ of the AI / ML model.

[0591] 3. Use real-time CSI data to monitor model performance and obtain real-time feedback accuracy acc. If the real-time feedback accuracy is lower than λ, the terminal switches and updates the model in two ways:

[0592] a) The terminal switches and updates the AI / ML model based on the relevant parameters for switching and updating the AI / ML model issued by the network device.

[0593] b) The terminal autonomously switches and updates the model based on the parameters of its own AI / ML model switching and update. The specific operations of model switching and updating are as follows:

[0594] a) According to the model number Model1, Model2, ..., Model deployed by the terminal n Make a judgment, if only Model is deployed current For the corresponding AI / ML model currently in use, a certain amount of real-time sample data is collected first, and the model parameters are updated based on the current model parameters using the real-time sample data set data until the feedback accuracy is higher than the accuracy threshold λ, and the updated encoder model θ′ is obtained. en and decoder model θ′ de, replace the current encoder model parameters with θ′ en , and generate model update instructions Indicator update , and the decoder model parameters θ′ de Perform quantitative processing;

[0595] b) According to the model number Model1, Model2, ..., Model deployed by the terminal n To make a judgment, if multiple groups of AI / ML models are deployed, use real-time test samples to calculate the feedback accuracy of each group of models.

[0596] i. If there is one or more groups of models whose feedback accuracy results are higher than or equal to the accuracy threshold λ: From the models whose feedback accuracy results are higher than or equal to the accuracy threshold λ, select the model with the highest feedback accuracy The terminal switches the current encoder model to At the same time, generate the model switching instruction Indicator switch This command carries the model number of the AI / ML model with the highest accuracy. opt .

[0597] ii. If the feedback accuracy of the models deployed by the terminal is lower than the accuracy threshold λ, the model with the highest feedback accuracy is selected Use real-time sample dataset data based on encoder model and decoder model Update the model parameters until the feedback accuracy is higher than the accuracy threshold λ, and train the encoder and decoder model parameters θ′ en and θ′ de , replace the current encoder model parameters with θ′ en , and generate the model update instruction Indicator update , and the decoder model parameters θ′ de Perform quantitative processing;

[0598] 4. Deactivate the currently working CSI feedback network encoder model and send model switching and update request information and related parameters to the network device. The information and parameters include: Indicatordeactivate command to deactivate the current CSI feedback network decoder model, Indicator switch Or model update instruction Indicator update ; Send model update instruction Indicator update In the case of , the updated decoder model parameters θ′ after quantization are sent at the same time de .

[0599] 5. Receive the model switching and updating confirmation information sent by the network device, the confirmation information includes: the confirmation indication that the network device has completed the model switching or updating. complete .

[0600] 6. Use the switched or updated AL / ML encoder model for subsequent CSI transmission.

[0601] In actual communication systems, deep neural network models are used to compress downlink CSI on the UE side and reconstruct it on the base station side. When the distribution characteristics of the real-time data that the model needs to process differ significantly from those of the model training data, a model-data mismatch occurs, significantly reducing the accuracy of CSI feedback and, in turn, impacting the overall performance of the massive MIMO system. Furthermore, considering the model training data acquisition methods used by the UE and base station, the UE can directly obtain real-time downlink CSI data through downlink channel estimation. Therefore, real-time feedback network model performance testing, switching, and updates on the UE side do not require the transmission of CSI training datasets from the UE to the base station, thus helping to reduce data transmission overhead.

[0602] The present disclosure is dedicated to overcoming the significant performance degradation problem caused by the mismatch between the model and the data in the offline training mode, so as to improve the practical value of the channel state information feedback network based on deep learning in the actual system. The present disclosure designs a terminal-triggered CSI feedback network switching and update scheme, and completes the process design of the specific terminal and network equipment for CSI feedback network switching and updating. The scheme proposed in the present disclosure can ensure the feedback accuracy of the CSI feedback network in the actual time-varying channel environment with low communication overhead and model training overhead, fully improve the adaptive ability of the feedback network model, effectively guarantee the performance of the CSI feedback network model, and improve the practicality of the feedback network model.

[0603] Optionally, FIG5C is a flow chart of a model switching update method according to an embodiment of the present disclosure. As shown in FIG5C , the method includes:

[0604] 1. Network equipment (base station)

[0605] 1. Receive information such as the AI / ML model type and parameters for CSI feedback reported by the terminal, including: Model number currently used by the terminal current , model structure information, the number of the AI / ML model deployed by the terminal for the CSI feedback task Model1, Model2,…, Model n .

[0606] 2. Send a CSI test data request command to the terminal.

[0607] 3. Receive the quantized CSI test data sent by the terminal and dequantize it. Use the test data to monitor model performance and obtain the CSI feedback accuracy acc of the current model. Compare it with the accuracy threshold λ. If the current feedback accuracy acc is lower than λ, the network device performs the following model switching and update operations:

[0608] a) According to the model number of the network equipment deployed, Model1, Model2, ..., Model n Make a judgment, if only Model is deployed current The corresponding AI / ML model currently in use,

[0609] i. Generate a CSI training set request instruction Indicatordata_request and send it to the terminal;

[0610] ii. Receive the quantized CSI training data set sent by the terminal and perform dequantization processing. Use the real-time sample data set to update the model parameters based on the current model to obtain the updated encoder and decoder models θ′ en and θ′ de , replace the current decoder model parameters with θ′ de , and generate the model update instruction Indicator update , and the encoder model parameters θ′ en Perform quantitative processing;

[0611] b) Model numbering Model1, Model2, ..., Model n To make a judgment, if multiple groups of AI / ML models are deployed, use real-time test samples to calculate the feedback accuracy of each group of models.

[0612] i. If there is one or more groups of models whose feedback accuracy results are higher than or equal to the accuracy threshold λ: From the models whose feedback accuracy results are higher than or equal to the accuracy threshold λ, select the model with the highest feedback accuracy The network device switches the current decoder model to At the same time, generate the model switching instruction Indicator switch This command carries the model number of the AI / ML model with the highest accuracy. opt .

[0613] ii. If the feedback accuracy of the models deployed by the network devices is lower than the accuracy threshold λ, the model with the highest feedback accuracy is selected.

[0614] i. Generate a CSI training set request instruction Indicatordata_request and send it to the terminal;

[0615] ii. Receive the quantized CSI training data set sent by the terminal and dequantize it, using the real-time sample data set data based on the encoder model and decoder model Update the model parameters until the feedback accuracy is higher than the accuracy threshold λ, and train the encoder and decoder model parameters θ′ en and θ′ de , replace the current decoder model parameters with θ′ de , and generate the model update instruction Indicator update , and the encoder model parameters θ′ en Perform quantitative processing;

[0616] 4. Deactivate the currently working CSI feedback network decoder model and send model switching and update request information and related parameters to the terminal. The information and parameters include: Indicatordeactivate command to deactivate the current CSI feedback network encoder model, Indicator switch Or model update instruction Indicator update ; Send model update instruction Indicator update In the case of , the updated quantized encoder model parameters θ′ are sent at the same time en .

[0617] 5. Receive the model switching and update confirmation information sent by the terminal, which includes: the confirmation indication that the terminal has completed the model switching or update. complete

[0618] 6. Use the switched or updated AL / ML decoder model for subsequent CSI reception.

[0619] 2. Terminal

[0620] 1. Report the AI / ML model type, parameters, and other information used for CSI feedback to the network device, including the model number currently used by the terminal. current , model structure information, model number Model1, Model2,…, Model deployed by the terminal n .

[0621] 2. Receive the CSI test data request instruction sent by the network device.

[0622] 3. Collect real-time CSI test sample data and send quantized CSI test data to network devices.

[0623] 4. If a training set request instruction Indicatordata_reques is received from a network device, a small-scale real-time CSI sample data is collected and quantized, and the quantized CSI data set is sent to the network device.

[0624] 5. Receive model switching and update request information and related parameters sent by the network device, including: Indicatordeactivate instruction to deactivate the current CSI feedback network encoder model, Indicator switch Or model update instruction Indicator update ; Upon receiving the model update instruction Indicator update In the case of , the updated encoder model parameters θ′ are received at the same time after quantization en .

[0625] Disable the currently working CSI feedback network encoder model and perform model switching and update operations on network devices:

[0626] a) If the received instruction is Indicator switch , according to the model number Model carried in the instruction opt , switches the encoder model to the model corresponding to this number.

[0627] b) If the received instruction is an Indicator update , dequantize the received model parameters and replace the encoder model parameters with θ′ en .

[0628] 6. Send model switching, update confirmation information to the network device, the confirmation information includes: the terminal has completed the model switching or update confirmation indicator complete .

[0629] 7. Use the switched or updated AL / ML encoder model for subsequent CSI transmission.

[0630] In actual communication systems, deep neural network models are used to compress downlink CSI on the UE side and reconstruct it on the base station side. When the distribution characteristics of the real-time data the model needs to process differ significantly from those of the model training data, a model-data mismatch occurs, significantly reducing the accuracy of CSI feedback and thus affecting the overall performance of the massive MIMO system. Furthermore, considering the computing resource allocation of the UE and base station, the UE's real-time computing resources are relatively limited, while the base station can better provide hardware support for training and updating the deep neural network model.

[0631] The present disclosure is dedicated to overcoming the significant performance degradation problem caused by the mismatch between the model and the data in the offline training mode, so as to improve the practical value of the channel state information feedback network based on deep learning in the actual system. The present disclosure designs a CSI feedback network switching and updating scheme triggered by a network device, and completes the process design of the CSI feedback network switching and updating between the specific terminal and the network device. The scheme proposed in the present disclosure can ensure the feedback accuracy of the CSI feedback network in the actual time-varying channel environment with low communication overhead and model training overhead, fully improve the adaptive ability of the feedback network model, effectively guarantee the performance of the CSI feedback network model, and improve the practicality of the feedback network model.

[0632] The embodiments of the present disclosure also provide apparatuses for implementing any of the above methods. For example, an apparatus is provided that includes units or modules for implementing each step performed by a terminal in any of the above methods. For another example, another apparatus is provided that includes units or modules for implementing each step performed by a network device in any of the above methods. The network device may be, for example, an access network device, a core network function node, or a core network device.

[0633] It should be understood that the division of the various units or modules in the above device is merely a division of logical functions. In actual implementation, they may be fully or partially integrated into a physical entity, or they may be physically separated. In addition, the units or modules in the device may be implemented in the form of a processor calling software: for example, the device includes a processor, the processor is connected to a memory, and the memory stores instructions. The processor calls the instructions stored in the memory to implement any of the above methods or implement the functions of the various units or modules of the above device, wherein the processor is, for example, a general-purpose processor, such as a central processing unit (CPU) or a microprocessor, and the memory is a memory within the device or a memory outside the device. Alternatively, the units or modules in the device can be implemented in the form of hardware circuits, and the functions of some or all of the units or modules can be realized by designing the hardware circuits. The above-mentioned hardware circuits can be understood as one or more processors; for example, in one implementation, the above-mentioned hardware circuit is an application-specific integrated circuit (ASIC), which realizes the functions of some or all of the above units or modules by designing the logical relationship of the components in the circuit; for example, in another implementation, the above-mentioned hardware circuit can be realized by a programmable logic device (PLD). Taking a field programmable gate array (FPGA) as an example, it can include a large number of logic gate circuits, and the connection relationship between the logic gate circuits is configured by configuring the configuration file, thereby realizing the functions of some or all of the above units or modules. All units or modules of the above devices can be realized in the form of software called by the processor, or in the form of hardware circuits, or in part by the form of software called by the processor, and the rest by hardware circuits.

[0634] In the embodiments of the present disclosure, the processor is a circuit with signal processing capabilities. In one implementation, the processor can be a circuit with instruction reading and execution capabilities, such as a central processing unit (CPU), a microprocessor, a graphics processing unit (GPU) (which can be understood as a microprocessor), or a digital signal processor (DSP). In another implementation, the processor can implement certain functions through the logical relationship of the hardware circuit. The logical relationship of the above-mentioned hardware circuit is fixed or reconfigurable. For example, the processor is a hardware circuit implemented by an application-specific integrated circuit (ASIC) or a programmable logic device (PLD), such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document and implementing the hardware circuit configuration can be understood as the process of the processor loading instructions to implement the functions of some or all of the above units or modules. In addition, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as a neural network processing unit (NPU), a tensor processing unit (TPU), a deep learning processing unit (DPU), etc.

[0635] Figure 6A is a schematic diagram of the structure of the first device proposed in an embodiment of the present disclosure. As shown in Figure 6A, it includes: a processing module for monitoring the performance of the first model currently used by the first device and / or the second model currently used by the second device; the first model and the second model correspond to each other, and the first model and the second model are used to implement the compression and reconstruction of the channel state information (CSI); the processing module is also used to determine that a model switching update is required when the performance monitoring result meets a first preset condition; the processing module is also used to determine at least one of the target first model, the first model update parameters, the target second model, and the second model update parameters; and a sending module is used to perform the model switching update and send indication information to the second device, the indication information indicating the target second model and / or the second model update parameters. Optionally, the processing module is used to perform the steps related to "processing" performed in any of the above methods, and the sending module is used to perform the steps related to "sending" performed in any of the above methods. Optionally, the first device may also include a receiving module, which is used to perform the steps related to "receiving" performed in any of the above methods, which will not be repeated here.

[0636] Figure 6B is a schematic diagram of the structure of the network device proposed in the embodiment of the present disclosure. As shown in Figure 6B, it includes: a receiving module, which is used for the second device to receive the indication information sent by the first device, and the indication information indicates the target second model and / or the second model update parameters; a processing module, which is used for the second device to perform model switching and updating based on the indication information. Optionally, the above-mentioned receiving module is used to execute the steps related to "receiving" executed in any of the above methods, and the above-mentioned processing module is used to execute the steps related to "processing" executed in any of the above methods. Optionally, the above-mentioned first device may also include a sending module, and the above-mentioned sending module is used to execute the steps related to "sending" executed in any of the above methods, which will not be repeated here.

[0637] Figure 7A is a schematic diagram of the structure of a communication device 7100 proposed in an embodiment of the present disclosure. Communication device 7100 can be a network device (e.g., an access network device, a core network device, etc.), a terminal (e.g., a user equipment, etc.), a chip, a chip system, or a processor that supports a network device to implement any of the above methods, or a chip, a chip system, or a processor that supports a terminal to implement any of the above methods. Communication device 7100 can be used to implement the methods described in the above method embodiments. For details, please refer to the description of the above method embodiments.

[0638] As shown in Figure 7A, the communication device 7100 includes one or more processors 7101. The processor 7101 can be a general-purpose processor or a dedicated processor, for example, a baseband processor or a central processing unit. The baseband processor can be used to process communication protocols and communication data, and the central processing unit can be used to control the communication device (such as a base station, a baseband chip, a terminal device, a terminal device chip, a DU or a CU, etc.), execute programs, and process program data. The processor 7101 is used to call instructions to enable the communication device 7100 to perform any of the above methods.

[0639] In some embodiments, the communication device 7100 further includes one or more memories 7102 for storing instructions. Optionally, all or part of the memories 7102 may be located outside the communication device 7100.

[0640] In some embodiments, the communication device 7100 further includes one or more transceivers 7103. When the communication device 7100 includes one or more transceivers 7103, the communication steps such as sending and receiving in the above method are performed by the transceiver 7103, and the other steps are performed by the processor 7101.

[0641] In some embodiments, a transceiver may include a receiver and a transmitter, which may be separate or integrated. Optionally, the terms transceiver, transceiver unit, transceiver, and transceiver circuit may be used interchangeably; the terms transmitter, transmitting unit, transmitter, and transmitting circuit may be used interchangeably; and the terms receiver, receiving unit, receiver, and receiving circuit may be used interchangeably.

[0642] Optionally, the communication device 7100 further includes one or more interface circuits 7104, which are connected to the memory 7102. The interface circuits 7104 may be configured to receive signals from the memory 7102 or other devices, and may be configured to send signals to the memory 7102 or other devices. For example, the interface circuits 7104 may read instructions stored in the memory 7102 and send the instructions to the processor 7101.

[0643] The communication device 7100 described in the above embodiment may be a network device or a terminal, but the scope of the communication device 7100 described in the present disclosure is not limited thereto, and the structure of the communication device 7100 may not be limited by FIG. 7a. The communication device may be an independent device or may be part of a larger device. For example, the communication device may be: 1) an independent integrated circuit IC, or a chip, or a chip system or subsystem; (2) a collection of one or more ICs, optionally, the above IC collection may also include a storage component for storing data or programs; (3) an ASIC, such as a modem; (4) a module that can be embedded in other devices; (5) a receiver, a terminal device, an intelligent terminal device, a cellular phone, a wireless device, a handheld device, a mobile unit, an in-vehicle device, a network device, a cloud device, an artificial intelligence device, etc.; (6) others, etc.

[0644] 7B is a schematic diagram of the structure of a chip 7200 proposed in an embodiment of the present disclosure. If the communication device 7100 can be a chip or a chip system, please refer to the schematic diagram of the structure of the chip 7200 shown in FIG7B , but the present disclosure is not limited thereto.

[0645] The chip 7200 includes one or more processors 7201 , and the processor 7201 is used to call instructions so that the chip 7200 executes any of the above methods.

[0646] In some embodiments, chip 7200 further includes one or more interface circuits 7202, which are connected to memory 7203. Interface circuit 7202 can be used to receive signals from memory 7203 or other devices, and can be used to send signals to memory 7203 or other devices. For example, interface circuit 7202 can read instructions stored in memory 7203 and send the instructions to processor 7201. Optionally, the terms interface circuit, interface, transceiver pin, and transceiver are interchangeable.

[0647] In some embodiments, the chip 7200 further includes one or more memories 7203 for storing instructions. Alternatively, all or part of the memories 7203 may be located outside the chip 7200.

[0648] The present disclosure also proposes a storage medium having instructions stored thereon. When the instructions are executed on the communication device 7100, the communication device 7100 executes any of the above methods. Optionally, the storage medium is an electronic storage medium. Optionally, the storage medium is a computer-readable storage medium, but is not limited thereto and may also be a storage medium readable by other devices. Optionally, the storage medium may be a non-transitory storage medium, but is not limited thereto and may also be a temporary storage medium.

[0649] The present disclosure also provides a program product, which, when executed by the communication device 7100, enables the communication device 7100 to perform any of the above methods. Optionally, the program product is a computer program product.

[0650] The present disclosure also proposes a computer program, which, when executed on a computer, causes the computer to perform any one of the above methods.

[0651] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented using software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer programs. When the computer program is loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present disclosure are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer program can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer program can be transmitted from one website, computer, server or data center to another website, computer, server or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated therein. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a high-density digital video disc (DVD)), or a semiconductor medium (eg, a solid state disk (SSD)).

[0652] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0653] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0654] The above description is merely a specific embodiment of the present disclosure, but the scope of protection of the present disclosure is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this disclosure should be included in the scope of protection of the present disclosure. Therefore, the scope of protection of the present disclosure should be based on the scope of protection of the claims.

Claims

1. A model switching update method, characterized in that: The method comprises: The first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device; the first model and the second model correspond to each other, and the first model and the second model are used to implement compression and reconstruction of channel state information CSI; The performance monitoring result meets the first preset condition, and it is determined that a model switching update is required; The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter; wherein the target first model is a target model to be switched by the first device, and the target second model is a target model to be switched by the second device; The first device performs a model switching update and sends indication information to the second device, where the indication information indicates the target second model and / or an update parameter of the second model.

2. The method according to claim 1, characterized in that Before the first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device, the method further includes: Determine the model-related information corresponding to the second device; the model-related information includes at least one of the following: a second model currently used by the second device; at least one second model deployed by the second device; The model structure corresponding to at least one second model deployed by the second device.

3. The method according to claim 2, characterized in that The determining the model-related information corresponding to the second device includes: Receive the model-related information sent by the second device.

4. The method according to claim 1, characterized in that Before determining that a model switching update is required, the method further includes: The first device determines a first parameter value, where the first parameter value is used to determine whether a model switching update is required; Among them, the first preset condition includes: the performance monitoring result is less than the first parameter value.

5. The method according to claim 4, characterized in that The first device determines the first parameter value, including at least one of the following: The first device receives the first parameter value sent by the second device; The first device determines the first parameter value configured by itself; The first device determines the first parameter value based on a protocol agreement.

6. The method according to any one of claims 1 to 5, characterized in that: The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter, including: The first model currently used by the first device is the first model uniquely deployed by the first device, and / or the second model currently used by the second device is the second model uniquely deployed by the second device, and sample data is determined; Training the first model currently used by the first device and / or the second model currently used by the second device based on the sample data until the performance monitoring result of the trained first model and / or the trained second model meets the second preset condition, and then stopping the training; The model parameters of the trained first model are determined as the first model update parameters, and / or the model parameters of the trained second model are determined as the second model update parameters.

7. The method according to claim 6, characterized in that The first device performs a model switching update and sends indication information to the second device, including at least one of the following: The first device updates the model parameters of the first model currently in use to the first model update parameters; The first device sends indication information to the second device, where the indication information includes second model update parameters and / or first indication; the first indication is used to instruct the second device to update the model parameters of the second model currently in use to the second model update parameters.

8. The method according to any one of claims 1 to 5, characterized in that: The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter, including: The first device deploys at least two first models, and / or the second device deploys at least two second models, and at least one model group is determined, wherein the model group includes the first model and the second model corresponding to each other; Performing performance monitoring on the first model and / or the second model in each group of model groups; If the performance monitoring result of at least one of the model groups meets the second preset condition, the first model in the model group with the highest performance monitoring result is determined as the target first model, and the second model in the model group with the highest performance monitoring result is determined as the target second model.

9. The method according to claim 8, characterized in that The first device performs a model switching update and sends indication information to the second device, including at least one of the following: The first device deactivates the currently used first model and switches from the currently used first model to the target first model; The first device sends indication information to the second device, where the indication information includes a second indication and / or relevant information of a target second model; the second indication is used to indicate at least one of the following: The second device deactivates the second model currently in use; The second device switches from the currently used second model to the target second model.

10. The method according to any one of claims 1 to 5, characterized in that: The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter, including: The first device deploys at least two first models, and / or the second device deploys at least two second models, and at least one model group is determined, wherein the model group includes the first model and the second model corresponding to each other; Monitor the performance of each model group; If the performance monitoring results of each model group do not meet the second preset condition, the first model in the model group with the highest performance monitoring result is determined as the target first model, and the second model in the model group with the highest performance monitoring result is determined as the target second model; Determine sample data; Training the target first model and the target second model based on the sample data until the performance monitoring results of the trained first model and / or the trained second model meet a second preset condition, and then stopping the training; The model parameters of the trained target first model are determined as the first model update parameters, and / or the model parameters of the trained target second model are determined as the second model update parameters.

11. The method according to claim 10, characterized in that The first device performs a model switching update and sends indication information to the second device, including at least one of the following: The first device deactivates the currently used first model, switches from the currently used first model to the target first model, and updates the model parameters of the target first model to the first model update parameters; The first device sends indication information to the second device, where the indication information includes at least one of relevant information of the target second model, a second model update parameter, and a third indication, where the third indication is used to indicate at least one of the following: The second device deactivates the second model currently in use; The second device switches from the currently used second model to the target second model; The second device updates the model parameters of the target second model to the second model update parameters.

12. The method according to any one of claims 6 to 11, characterized in that: The second preset condition includes: the performance monitoring result is greater than or equal to the first parameter value.

13. The method according to any one of claims 1 to 12, characterized in that: After the first device sends the indication information to the second device, the method further includes: Receive confirmation information sent by the second device, where the confirmation information is used to indicate that the second device has completed the model switching update.

14. The method according to any one of claims 6 to 13, characterized in that: The first device is a terminal, and the second device is a network device; The first model is used to compress the CSI, and the second model is used to reconstruct the CSI.

15. The method according to claim 14, characterized in that The first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device, including: The first device uses the CSI data to monitor the performance of a first model currently used by the first device and / or a second model currently used by the second device.

16. The method according to claim 14 or 15, characterized in that After the first device sends the indication information to the second device, the method further includes: The first model after switching update is used for CSI compression.

17. The method according to any one of claims 6 to 13, characterized in that: The first device is a network device, and the second device is a terminal; The second model is used to compress the CSI, and the first model is used to reconstruct the CSI.

18. The method according to claim 17, characterized in that The first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device, including: The first device sends a first request to the second device, where the first request is used to request CSI data; The first device receives the CSI data sent by the second device; The first device uses the CSI data to monitor the performance of a first model currently used by the first device and / or a second model currently used by the second device.

19. The method according to claim 17 or 18, characterized in that The determining of sample data comprises: The first device sends a second request to the second device, where the second request is used to request the sample data; The first device receives the sample data sent by the second device.

20. The method according to any one of claims 17 to 19, characterized in that: After the first device sends the indication information to the second device, the method further includes: The CSI reconstruction is performed using the first model after switching updates.

21. The method according to any one of claims 15 to 20, characterized in that: The CSI data is obtained by the terminal through measuring the pilot signal sent by the network device and through channel estimation.

22. The method according to any one of claims 6 to 21, characterized in that: The sample data is: CSI data measured and estimated by the terminal within a first preset time period.

23. The method according to any one of claims 1 to 22, characterized in that: The performance monitoring result is used to reflect the CSI feedback accuracy of the first model and the second model.

24. A model switching update method, characterized in that: The method comprises: The second device receives indication information sent by the first device, where the indication information indicates a target second model and / or a second model update parameter; wherein the target second model is a target model to be switched by the second device; The second device performs model switching update based on the indication information.

25. The method of claim 24, wherein: A first model is deployed on the first device, and a second model is deployed on the second device; the first model and the second model correspond to each other, and the first model and the second model are used to realize compression and reconstruction of CSI.

26. The method according to claim 24 or 25, characterized in that Before the second device receives the indication information sent by the first device, the method further includes: Sending model-related information corresponding to the second device to the first device; the model-related information includes at least one of the following: a second model currently used by the second device; at least one second model deployed by the second device; The model structure corresponding to at least one second model deployed by the second device.

27. The method according to claim 24 or 25, characterized in that Before the second device receives the indication information sent by the first device, the method further includes: A first parameter value is sent to the first device, where the first parameter value is used to determine whether a model switching update is required.

28. The method according to any one of claims 24 to 27, characterized in that: The indication information includes a second model update parameter and / or a first indication; the first indication is used to instruct the second device to update the model parameters of the second model currently in use to the second model update parameter; The performing model switching update based on the indication information includes: The second device updates the model parameters of the second model currently in use to the second model update parameters.

29. The method according to any one of claims 24 to 27, characterized in that: The indication information includes the second indication and / or relevant information of the target second model; the second indication is used to indicate at least one of the following: the second device deactivates the currently used second model, the second device switches from the currently used second model to the target second model; The performing model switching update based on the indication information includes: The second device deactivates the currently used second model and switches from the currently used second model to the target second model.

30. The method according to any one of claims 24 to 27, characterized in that: The indication information includes at least one of relevant information of the target second model, a second model update parameter, and a third indication, wherein the third indication is used to indicate at least one of the following: the second device deactivates the currently used second model, the second device switches from the currently used second model to the target second model, and the second device updates the model parameters of the target second model to the second model update parameters; The performing model switching update based on the indication information includes: The second device deactivates the currently used second model, switches from the currently used second model to the target second model, and updates the model parameters of the target second model to the second model update parameters.

31. The method according to any one of claims 24 to 30, characterized in that: After the second device performs model switching update based on the indication information, the method further includes: Sending a confirmation message to the first device, where the confirmation message is used to indicate that the second device has completed the model switching update.

32. The method according to any one of claims 24 to 31, characterized in that: The first device is a terminal, and the second device is a network device; The first model is used to compress the CSI, and the second model is used to reconstruct the CSI.

33. The method of claim 32, wherein: After the second device performs model switching update based on the indication information, the method further includes: The CSI reconstruction is performed using the second model after switching updates.

34. The method according to any one of claims 24 to 31, characterized in that: The first device is a network device, and the second device is a terminal; The second model is used to compress the CSI, and the first model is used to reconstruct the CSI.

35. The method of claim 34, wherein: After the second device performs model switching update based on the indication information, the method further includes: The second model after switching and updating is used for CSI compression.

36. A model switching update method, characterized in that: Used in a communication system, the communication system comprising a first device and a second device, the method comprising at least one of the following: The first device performs performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device; the first model and the second model correspond to each other, and the first model and the second model are used to implement compression and reconstruction of channel state information CSI; The performance monitoring result meets the first preset condition, and the first device determines that a model switching update is required; The first device determines at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter; wherein the target first model is a target model to be switched by the first device, and the target second model is a target model to be switched by the second device; The first device performs a model switching update and sends indication information to the second device, where the indication information indicates the target second model and / or an update parameter of the second model; The second device receives the indication information sent by the first device; The second device performs model switching update based on the indication information.

37. A first device, comprising: a processing module, configured to perform performance monitoring on a first model currently used by the first device and / or a second model currently used by the second device; The first model and the second model correspond to each other, and the first model and the second model are used to realize compression and reconstruction of channel state information CSI; The processing module is further configured to determine that a model switching update is required when a performance monitoring result meets a first preset condition; The processing module is further used to determine at least one of a target first model, a first model update parameter, a target second model, and a second model update parameter; The sending module is used to perform a model switching update and send an indication message to the second device, wherein the indication message indicates that the target second model model and / or update parameters of the second model.

38. A second device, comprising: A receiving module, configured for a second device to receive indication information sent by a first device, wherein the indication information indicates a target second model and / or a second model update parameter; wherein the target second model is a target model to be switched by the second device; A processing module is used for the second device to perform model switching update based on the indication information.

39. A communication device, characterized in that: include: One or more processors; The processor is used to call instructions to enable the communication device to execute the model switching update method described in any one of claims 1-23 and 24-35.

40. A communication system, characterized in that: It includes a first device and a second device, wherein the first device is configured to implement the model switching update method described in any one of claims 1-23, and the second device is configured to implement the model switching update method described in any one of claims 24-35.

41. A storage medium storing instructions, characterized in that: When the instruction is executed on the communication device, the communication device executes the model switching update method according to any one of claims 1-23 and 24-35.

Images