CN117641436A

CN117641436A - Channel information processing method, terminal and network equipment

Info

Publication number: CN117641436A
Application number: CN202210997811.8A
Authority: CN
Inventors: 温子睿; 李刚; 韩双锋
Original assignee: China Mobile Communications Group Co Ltd; China Mobile Communications Ltd Research Institute
Current assignee: China Mobile Communications Group Co Ltd; China Mobile Communications Ltd Research Institute
Priority date: 2022-08-19
Filing date: 2022-08-19
Publication date: 2024-03-01

Abstract

A processing method, terminal and network device of channel information, the method includes: the terminal measures a downlink reference signal to obtain first measurement information; the terminal compresses the first measurement information through an encoder to obtain first compressed information; and the terminal sends the first compressed information to a network according to the first model configuration information, or sends the first compressed information and the first scene information corresponding to the first measurement information to the network. The method and the device consider factors such as complexity of the network model, model performance, air interface transmission cost and the like, and reduce the influence on the performance of the network model while reducing the complexity of the network model.

Description

Channel information processing method, terminal and network equipment

Technical Field

The present invention relates to the field of mobile communications technologies, and in particular, to a method, a terminal, and a network device for processing channel information.

Background

In an actual communication environment, a network with limited complexity may not be responsible for channel state information (Channel State Information, CSI) compression feedback of a whole cell (e.g., cannot meet the requirement of-10 dB of reconstruction accuracy), but may process CSI of a local sub-area. The subarea may be smaller in range and single in scene (e.g., 20m×20m parking lot), and users in the subarea share a spatial propagation environment, so that CSI sampled in the subarea has a larger spatial correlation and a higher sample similarity. Thus, by sampling the region-partitioned dataset, the difficult-to-process CSI compression feedback reconstruction task can be divided into several easily managed sub-tasks. Based on this, the existing CSI compression feedback scheme with high generalization has two modes of S-to-S and M-to-M.

Wherein S-to-S is dedicated to designing a more complex network to manage CSI feedback for all scenarios of the whole cell, as in fig. 1. This approach enhances the generalization performance of the network for all scenarios by increasing the capacity of the encoder and decoder networks, requiring costly design, tuning and training costs. At the same time, too complex a network consumes significant amounts of device memory and power.

M-to-M feeds back CSI data sets by region division, i.e., using multiple relatively simple networks. Each network is responsible for feeding back CSI for local sub-area samples (single scenario), as in fig. 2. This approach, while avoiding the extra complexity network design cost of the S-to-S approach, requires the user and base station to store multiple sets of encoder network parameters, corresponding to sub-region (single scenario) CSI, respectively. When a scene is switched (e.g., the user moves across regions), the current scene is adapted by switching network parameters. Also, this approach is not user friendly and consumes more memory. Meanwhile, when switching networks, additional information needs to be fed back to the base station for judgment.

It can be seen that the S-to-S approach to model generalization through a single Artificial Intelligence (AI) model makes training of the model very difficult, and requires a high complexity of the model, which consumes more computing power of the user equipment, and the single generalization model is also relatively inefficient in performance compared to the dedicated model. The M-to-M mode can obtain better generalization by using a plurality of special simple models, and model training is relatively simple, but higher air interface transmission overhead is caused when the models are issued and the models are selected. Therefore, a method for processing channel information is needed to solve the problem that the complexity and performance of the network model for processing the channel information cannot be considered.

Disclosure of Invention

At least one embodiment of the application provides a processing method, a terminal and network equipment of channel information, which are used for solving the problem that the complexity and performance of a network model for processing the channel information cannot be considered.

In order to solve the technical problems, the application is realized as follows:

in a first aspect, an embodiment of the present application provides a method for processing channel information, including:

the terminal measures a downlink reference signal to obtain first measurement information;

the terminal compresses the first measurement information through an encoder to obtain first compressed information;

and the terminal sends the first compressed information to a network according to the first model configuration information, or sends the first compressed information and the first scene information corresponding to the first measurement information to the network.

Optionally, the method further comprises:

and receiving first model configuration information sent by a network, wherein the first model configuration information is used for configuring the encoder at a terminal or is used for configuring the encoder and a first judgment network at the terminal, and the first judgment network is used for identifying scene information.

Optionally, the method further comprises:

in the case that the first model configuration information is used to configure the encoder at a terminal, the terminal constructs the encoder according to the first model configuration information;

In the case that the first model configuration information is used to configure the encoder and the first decision network at the terminal, the terminal constructs the encoder and the first decision network according to the first model configuration information.

Optionally, in the case that the first model configuration information is used to configure the encoder at a terminal, the terminal sends the first compressed information to a network;

in the case that the first model configuration information is used to configure the encoder and the first decision network at the terminal, the terminal sends first compression information and first scene information corresponding to the first measurement information to the network.

Optionally, in the case that the first model configuration information is used to configure the encoder and the first decision network at a terminal, the terminal inputs the first measurement information to the first decision network to obtain the first scene information.

In a second aspect, an embodiment of the present application provides a method for processing channel information, including:

the network equipment sends a downlink reference signal;

the network equipment receives first compressed information of a downlink reference signal measurement result sent by a terminal;

the network device selects a first decoder from a plurality of decoders, wherein each decoder of the plurality of decoders corresponds to one of a plurality of scenes;

The network device inputs the first compressed information to the first decoder, and decompresses the first compressed information through the first decoder to obtain measurement information corresponding to the first compressed information.

Optionally, the method further comprises:

selecting a first decoder from a plurality of decoders based on second model configuration information at the network device,

wherein, in the case that the second model configuration information configures the plurality of decoders and a second decision network for identifying scene information, the first compression information is input to the second decision network, so as to obtain first scene information corresponding to the first compression information; selecting a decoder corresponding to first scene information from the plurality of decoders according to the first scene information as the first decoder;

receiving first scene information corresponding to the first measurement information sent by a terminal under the condition that the second model configuration information configures the plurality of decoders; and selecting a decoder corresponding to the first scene information from the plurality of decoders according to the first scene information as the first decoder.

Optionally, the method further comprises:

transmitting first model configuration information to the terminal,

the first model configuration information is used for configuring an encoder at a terminal under the condition that the plurality of decoders and the second decision network are configured by the second model configuration information, and the encoder is used for compressing measurement results of downlink reference signals measured by the terminal;

and in the case that the plurality of decoders are configured by the second model configuration information, the first model configuration information is used for configuring the encoder and a first decision network at the terminal, and the first decision network is used for identifying a corresponding scene based on the measurement result of the downlink reference signal measured by the terminal.

Optionally, the method further comprises:

performing multiple rounds of iterative training on the encoder and the plurality of decoders based on the pre-acquired second measurement information of the downlink reference signals in the multiple scenes until a preset training ending condition is met, so as to obtain the trained encoder and the trained plurality of decoders; wherein each round of iterative training comprises:

inputting second measurement information of the downlink reference signals under the various scenes to the encoder to obtain second compression information output by the encoder; the second compressed information is respectively input to each decoder and the temporary public decoder, and measurement information corresponding to the second compressed information is obtained; calculating a first loss value according to second measurement information of the downlink reference signals in the various scenes and measurement information corresponding to the second compression information, calculating a first gradient based on the first loss value, and transmitting the first gradient in a reverse path from a temporary public decoder to the encoder so as to adjust network parameters of the encoder;

Freezing network parameters of the encoder, and respectively inputting second measurement information of downlink reference signals under each scene to the encoder to obtain third compression information output by the encoder; respectively inputting the third compressed information to a decoder corresponding to the scene to obtain measurement information corresponding to the third compressed information; and calculating a second loss value according to second measurement information of the downlink reference signal under the scene and measurement information corresponding to the third compression information, calculating a second gradient based on the second loss value, and transmitting the second gradient in a reverse path from the decoder corresponding to the scene to the encoder so as to adjust network parameters of the decoder corresponding to the scene.

Optionally, the method further comprises:

a step of training a first decision network and/or a step of training a second decision network.

Optionally, the step of training the first decision network includes:

inputting second measurement information of downlink reference signals of each scene into a first judgment network to obtain second index information output by the first judgment network; inputting second measurement information of downlink reference signals of each scene to a trained encoder to obtain fourth compression information output by the encoder; respectively inputting the fourth compressed information to each trained decoder to obtain measurement information corresponding to the fourth compressed information output by each decoder; according to the second measurement information and the fourth compression information of the downlink reference signal in the scene, calculating to obtain a third loss value corresponding to each decoder, and determining third index information of the decoder corresponding to the minimum third loss value;

Calculating fourth loss values under the scenes according to the second index information and the third index information, counting the fourth loss values under all the scenes to obtain fifth loss values, calculating a third gradient based on the fifth loss values, and transmitting the third gradient in a reverse path from the first judgment network to the encoder to adjust network parameters of the first judgment network until preset training ending conditions are met, so as to obtain a trained first judgment network.

Optionally, the step of training the second decision network includes:

inputting second measurement information of downlink reference signals of each scene to a trained encoder to obtain fourth compression information output by the encoder; the fourth compressed information is respectively input into a second judgment network and each trained decoder to obtain fourth index information output by the second judgment network and measurement information corresponding to the fourth compressed information output by each decoder; according to the second measurement information and the fourth compression information of the downlink reference signal in the scene, calculating to obtain a third loss value corresponding to each decoder, and determining third index information of the decoder corresponding to the minimum third loss value;

And calculating a sixth loss value under the scene according to the fourth index information and the third index information, counting the sixth loss values under all scenes to obtain a seventh loss value, calculating a fourth gradient based on the seventh loss value, and transmitting the fourth gradient in a reverse path from the first judgment network to the encoder so as to adjust network parameters of the second judgment network until a preset training ending condition is met, thereby obtaining a trained second judgment network.

In a third aspect, embodiments of the present application provide a terminal comprising a transceiver and a processor, wherein,

the processor is used for measuring the downlink reference signal and obtaining first measurement information; compressing the first measurement information through an encoder to obtain first compressed information;

the transceiver is configured to send the first compressed information to a network according to current first model configuration information, or send the first compressed information and first scene information corresponding to the first measurement information to the network.

In a fourth aspect, an embodiment of the present application provides a terminal, including: a processor, a memory and a program stored on the memory and executable on the processor, which when executed by the processor implements the steps of the method as described in the first aspect.

In a fifth aspect, embodiments of the present application provide a network device comprising a transceiver and a processor, wherein,

the transceiver is used for sending downlink reference signals; receiving first compressed information of a downlink reference signal measurement result sent by a terminal;

the processor is used for selecting a first decoder from a plurality of decoders, wherein each decoder in the plurality of decoders corresponds to one scene in a plurality of scenes; and inputting the first compressed information to the first decoder, and decompressing the first compressed information by the first decoder to obtain measurement information corresponding to the first compressed information.

In a sixth aspect, embodiments of the present application provide a network device, including: a processor, a memory and a program stored on the memory and executable on the processor, which when executed by the processor implements the steps of the method as described in the second aspect.

In a seventh aspect, embodiments of the present application provide a computer-readable storage medium having stored thereon a program which, when executed by a processor, implements the steps of the method as described above.

Compared with the prior art, the channel information processing method, the terminal and the network equipment provided by the embodiment of the application consider factors such as complexity of a network model, model performance and air interface transmission cost, and influence on the performance of the network model is reduced while the complexity of the network model is reduced.

Drawings

Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the application. Also, like reference numerals are used to designate like parts throughout the figures. In the drawings:

FIG. 1 is a schematic diagram of a network architecture based on the S-to-S approach;

FIG. 2 is a schematic diagram of a network architecture based on the M-to-M scheme;

fig. 3 is a schematic view of an application scenario in an embodiment of the present application;

FIG. 4 is a schematic diagram of the overall structure of a network model according to an embodiment of the present application;

fig. 5 is a flowchart of a method for processing channel information according to an embodiment of the present application;

fig. 6 is another flowchart of a method for processing channel information according to an embodiment of the present application;

fig. 7 is a flowchart of an example of a method for processing channel information according to an embodiment of the present application;

Fig. 8 is another example flowchart of a method for processing channel information according to an embodiment of the present application;

FIG. 9 is a schematic diagram of each iteration of the codec training process in an embodiment of the present application;

FIG. 10 is a training schematic of an encoder in an embodiment of the present application;

FIG. 11 is a training schematic diagram of a subtask decoder in an embodiment of the present application;

FIG. 12 is a schematic diagram of performance comparison of a network model with other models in an embodiment of the present application;

fig. 13 is a schematic structural diagram of a terminal according to an embodiment of the present application;

fig. 14 is a schematic structural diagram of a terminal according to another embodiment of the present application;

fig. 15 is a schematic structural diagram of a network device according to an embodiment of the present application;

fig. 16 is a schematic structural diagram of a network device according to another embodiment of the present application;

fig. 17 is a schematic structural diagram of a terminal according to still another embodiment of the present application;

fig. 18 is a schematic structural diagram of a network device according to another embodiment of the present application.

Detailed Description

Exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

The terms first, second and the like in the description and in the claims, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that embodiments of the present application described herein may be capable of operation in sequences other than those illustrated or described herein, for example. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. "and/or" in the specification and claims means at least one of the connected objects.

The techniques described herein are not limited to NR systems and long term evolution (Long Time Evolution, LTE)/LTE evolution (LTE-Advanced, LTE-a) systems and may also be used for various wireless communication systems such as code division multiple access (Code Division Multiple Access, CDMA), time division multiple access (Time Division Multiple Access, TDMA), frequency division multiple access (Frequency Division Multiple Access, FDMA), orthogonal frequency division multiple access (Orthogonal Frequency Division Multiple Access, OFDMA), single-carrier frequency division multiple access (Single-carrier Frequency-Division Multiple Access, SC-FDMA), and other systems. The terms "system" and "network" are often used interchangeably. A CDMA system may implement radio technologies such as CDMA2000, universal terrestrial radio access (Universal Terrestrial Radio Access, UTRA), and the like. UTRA includes wideband CDMA (Wideband Code Division Multiple Access, WCDMA) and other CDMA variants. TDMA systems may implement radio technologies such as the global system for mobile communications (Global System for Mobile Communication, GSM). OFDMA systems may implement radio technologies such as ultra mobile broadband (UltraMobile Broadband, UMB), evolved UTRA (E-UTRA), IEEE 802.21 (Wi-Fi), IEEE802.16 (WiMAX), IEEE 802.20, flash-OFDM, and the like. UTRA and E-UTRA are parts of the universal mobile telecommunications system (Universal Mobile Telecommunications System, UMTS). LTE and higher LTE (e.g., LTE-a) are new UMTS releases that use E-UTRA. UTRA, E-UTRA, UMTS, LTE, LTE-a and GSM are described in the literature from an organization named "third generation partnership project" (3rd Generation Partnership Project,3GPP). CDMA2000 and UMB are described in the literature from an organization named "third generation partnership project 2" (3 GPP 2). The techniques described herein may be used for the systems and radio technologies mentioned above as well as for other systems and radio technologies. However, the following description describes an NR system for purposes of example, and NR terminology is used in much of the description below, although the techniques may also be applied to applications other than NR system applications.

The following description provides examples and does not limit the scope, applicability, or configuration as set forth in the claims. Changes may be made in the function and arrangement of elements discussed without departing from the spirit and scope of the disclosure. Various examples may omit, substitute, or add various procedures or components as appropriate. For example, the described methods may be performed in an order different than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

Referring to fig. 3, fig. 3 is a block diagram of a wireless communication system to which embodiments of the present application are applicable. The wireless communication system includes a terminal 11 and a network device 12. The terminal 11 may also be referred to as a User terminal or a User Equipment (UE), and the terminal 11 may be a terminal-side Device such as a mobile phone, a tablet Computer (Tablet Personal Computer), a Laptop (Laptop Computer), a personal digital assistant (Personal Digital Assistant, PDA), a mobile internet Device (Mobile Internet Device, MID), a Wearable Device (Wearable Device), or a vehicle-mounted Device, which is not limited to a specific type of the terminal 11 in the embodiments of the present application. The network device 12 may be a base station and/or a core network element, where the base station may be a 5G or later version base station (e.g., a gNB, a 5G NR NB, etc.), or a base station in another communication system (e.g., an eNB, a WLAN access point, or other access points, etc.), where the base station may be referred to as a node B, an evolved node B, an access point, a base transceiver station (Base Transceiver Station, a BTS), a radio base station, a radio transceiver, a basic service set (Basic Service Set, BSS), an extended service set (Extended Service Set, ESS), a node B, an evolved node B (eNB), a home node B, a home evolved node B, a WLAN access point, a WiFi node, or some other suitable terminology in the field, and the base station is not limited to a specific technical vocabulary, and in the embodiment of the present application, the base station in the NR system is merely an example, but is not limited to a specific type of the base station.

The base stations may communicate with the terminal 11 under the control of a base station controller, which may be part of the core network or some base stations in various examples. Some base stations may communicate control information or user data with the core network over a backhaul. In some examples, some of these base stations may communicate with each other directly or indirectly over a backhaul link, which may be a wired or wireless communication link. A wireless communication system may support operation on multiple carriers (waveform signals of different frequencies). A multicarrier transmitter may transmit modulated signals on the multiple carriers simultaneously. For example, each communication link may be a multicarrier signal modulated according to various radio technologies. Each modulated signal may be transmitted on a different carrier and may carry control information (e.g., reference signals, control channels, etc.), overhead information, data, and so on.

The base station may communicate wirelessly with the terminal 11 via one or more access point antennas. Each base station may provide communication coverage for a respective corresponding coverage area. The coverage area of an access point may be partitioned into sectors that form only a portion of that coverage area. A wireless communication system may include different types of base stations (e.g., macro base stations, micro base stations, or pico base stations). The base station may also utilize different radio technologies, such as cellular or WLAN radio access technologies. The base stations may be associated with the same or different access networks or operator deployments. The coverage areas of different base stations, including coverage areas of the same or different types of base stations, coverage areas utilizing the same or different radio technologies, or coverage areas belonging to the same or different access networks, may overlap.

The communication link in the wireless communication system may include an Uplink for carrying Uplink (UL) transmissions (e.g., from the terminal 11 to the network device 12) or a Downlink for carrying Downlink (DL) transmissions (e.g., from the network device 12 to the terminal 11). UL transmissions may also be referred to as reverse link transmissions, while DL transmissions may also be referred to as forward link transmissions. Downlink transmissions may be made using licensed bands, unlicensed bands, or both. Similarly, uplink transmissions may be made using licensed bands, unlicensed bands, or both.

As described in the background, the encoder/decoder schemes of the prior art S-to-S or M-to-M channel information have difficulty balancing the complexity of the network model with the performance of the model. In order to solve at least one of the above problems, the embodiment of the present application provides a method for processing channel information, which can compromise factors such as complexity of a network model, model performance, overhead of air interface transmission, and the like, and reduce the influence on the performance of the network model while reducing the complexity of the network model.

Referring to fig. 4, an overall structure diagram of a network model for compressing and decompressing channel information provided in an embodiment of the present application is shown in fig. 4, where the network model includes an encoder, a decision network, and a plurality of decoders, where each decoder corresponds to a scenario. The scenes can be obtained by dividing according to preset conditions, for example, an area with the same or similar space propagation environment is used as one scene, so that various scenes can be obtained, and data acquired by a terminal in the same area is used as data of a scene corresponding to the area. For another example, the division is performed according to the service type of the terminal, the service of the same or similar QoS index is divided into the same scene, and the data collected by the terminal executing the service type of the same or similar QoS index is used as the data of the scene. Each scenario is also referred to herein as a subtask. In FIG. 4, subtask 1-k represents scene 1-k, where k is typically an integer greater than or equal to 2. In this way, the encoder may be configured to perform compression processing on measurement information of the downlink reference signal acquired in any one of the multiple scenarios to obtain compression information, where each scenario corresponds to one decoder. The compressed information is used for being provided for a decoder corresponding to any scene to decompress so as to obtain measurement information corresponding to the compressed information.

In the embodiment of the application, a network device sends a downlink reference signal; and the terminal measures a downlink reference signal sent by the network to obtain measurement information. The encoder is used for compressing measurement information obtained by measuring the terminal, obtaining compressed information and sending the compressed information to the network side. That is, the terminal may compress measurement information measured in any scene using the encoder.

The decision network is used for identifying the current scene measured by the terminal. The network side decompresses the compressed information sent by the terminal by utilizing a decoder corresponding to the current scene, thereby obtaining measurement information corresponding to the compressed information. The output of the decision network may specifically be Index information (Index) corresponding to scenes, where each scene has a preset Index information. And selecting a decoder for decompression according to the index information, and decompressing the compressed information sent by the terminal.

In particular, the decoder in the above network model is provided at the network device, the encoder is provided at each terminal, and the decision network may be provided at the terminal side or at the network device. For convenience of explanation, when the decision network is set on the terminal side, the decision network is referred to as a first decision network in the embodiment of the present application; when the decision network is set on the network device side, the decision network is referred to as a second decision network in the embodiment of the present application. The input of the first decision network may be measurement information of a downlink reference signal measured by the terminal, and the output is specific scene indication information. The input of the second decision network may be compressed information sent by the terminal to the network, and the output is specific scene indication information. Of course, the input of the first decision network may also be compressed information sent by the terminal to the network. Thus, in the embodiment of the present application, there are two network structures, the first network structure including an encoder, a first decision network, and respective decoders; the second network structure comprises an encoder, a second decision network and respective decoders. The two network architectures differ only in the decision network.

Referring to fig. 5, a method for processing channel information provided in an embodiment of the present application is applied to a terminal, and includes:

step 51, the terminal measures the downlink reference signal to obtain the first measurement information.

Here, the network transmits a downlink reference signal, such as a channel state information reference signal (CSI Reference Signal, CSI-RS), and the terminal receives and measures the downlink reference signal to obtain first measurement information, such as channel state information (Channel State Information, CSI).

And step 52, the terminal compresses the first measurement information through an encoder to obtain first compressed information.

The terminal uses a preconfigured encoder to compress the first measurement information to obtain first compressed information, so that the data volume of the measurement information transmitted to the network is reduced, and the resource cost required by information transmission is saved.

And step 53, the terminal sends the first compressed information to the network according to the first model configuration information, or sends the first compressed information and the first scene information corresponding to the first measurement information to the network.

Here, the first model configuration information is used to configure the encoder at the terminal, or to configure the encoder and a first decision network for identifying scene information at the terminal. That is, the first model configuration information is used to configure the network model structure of the terminal side, and for example, only the network structure of the encoder of the terminal side may be configured, and the compression processing of the channel measurement information may be performed on the terminal side.

In this embodiment of the present application, in a case where the first model configuration information is used to configure the encoder at a terminal (a first decision network is not configured), the terminal sends the first compression information to a network; and in case the first model configuration information is used to configure the encoder and the first decision network at the terminal, the terminal sends first compression information and first scene information corresponding to the first measurement information to the network.

Through the steps, the embodiment of the application realizes that the first measurement information is compressed by the encoder at the terminal side, and the first compressed information is decoded by the first decoder corresponding to the first scene information at the network side, wherein the scene corresponding to the first measurement information is the first scene when the first scene information is generated, that is, the first measurement information is obtained by measuring the downlink reference signal under the first scene. According to the embodiment of the application, the plurality of decoders corresponding to each scene are deployed on the network side, and the corresponding decoders are obtained by training aiming at different scenes, so that the capacity of a decoder network can be reduced, and the cost of designing, parameter adjusting and training of the decoder network is reduced; meanwhile, the embodiment of the application adopts a unified encoder to compress the measurement information at the terminal side, so that generalization of the encoder network can be improved, and compared with an S-to-S mode, air interface transmission overhead when a model is issued to the terminal and the model is selected can be reduced. It can be seen that the embodiment of the application takes factors such as complexity of the network model, model performance, overhead of air interface transmission and the like into consideration, and reduces the influence on the performance of the network model while reducing the complexity of the network model.

The terminal may also receive first model configuration information sent by the network before the step 51, and then locally configure the relevant network model according to the first model configuration information.

For example, in the case where the first model configuration information is used to configure the encoder at a terminal, the terminal constructs the encoder according to the first model configuration information.

For another example, in the case where the first model configuration information is used to configure the encoder and the first decision network at a terminal, the terminal constructs the encoder and the first decision network according to the first model configuration information.

Specifically, in the embodiment of the present application, network parameters of the encoder, the first decision network, and the second decision network may be configured to the terminal in advance. Then, the encoder is activated by the above-described first model configuration information, or the encoder and the first decision network are activated by the second model configuration information.

In addition, in the case where the first model configuration information is used to configure the encoder and the first decision network at a terminal, at this time, the terminal inputs the first measurement information to the first decision network, thereby obtaining the first scene information.

Referring to fig. 6, a method for processing channel information provided in an embodiment of the present application is applied to a network device (such as a base station), and includes:

in step 61, the network device sends a downlink reference signal.

Here, the network device, such as a base station, may transmit a downlink reference signal, such as CSI-RS, to the terminal.

Step 62, the network device receives first compressed information of the downlink reference signal measurement result sent by the terminal.

The terminal receives and measures a downlink reference signal sent by the network, obtains measurement information, compresses the measurement information by using an encoder configured by the terminal to obtain first compression information, and then sends the first compression information to the network equipment. And the network equipment receives the first compressed information of the downlink reference signal measurement result sent by the terminal.

Step 63, the network device selects a first decoder from a plurality of decoders, wherein each decoder of the plurality of decoders corresponds to a respective one of a plurality of scenes.

In step 64, the network device inputs the first compressed information to the first decoder, and decompresses the first compressed information by the first decoder to obtain measurement information corresponding to the first compressed information.

Here, the network device selects one decoder (referred to herein as a first decoder) from among a plurality of decoders, and decompresses the first compressed information using the first decoder, thereby obtaining measurement information corresponding to the first compressed information. The subsequent network can perform channel precoding and other processing according to the obtained measurement information.

Through the steps, the embodiment of the application realizes the processing such as compression and decompression of the channel measurement information by using the single-encoder multi-decoder structure, and compared with the single-encoder decoder structure, the embodiment of the application reduces the training difficulty of an AI model and can obtain better performance; in addition, compared with a multi-encoder multi-decoder structure, the embodiment of the application reduces the data overhead when the network side issues the encoder, and when the decision network is deployed on the network side, the embodiment of the application can also reduce the air interface data overhead when the terminal reports the encoder selection decision information (such as scene information) and transmits the decision network. Through high-level configuration, the use mode of the network model designed in the embodiment of the application can be flexibly selected so as to cope with challenges in more severe channel environments.

In step 63 described above, the network device may select the first decoder from the plurality of decoders according to the second model configuration information. The second model configuration information may be configured by the network manager to the network device or sent by the core network to the network device.

For example, in the case that the second model configuration information configures the plurality of decoders and a second decision network for identifying scene information locally at a network device, the first compression information is input to the second decision network to obtain first scene information corresponding to the first compression information; then, according to the first scene information, a decoder corresponding to the first scene information is selected from the plurality of decoders as the first decoder.

For another example, in the case where the second model configuration information configures the plurality of decoders in the network device (the second decision network is not configured), the network device may receive first scene information corresponding to the first measurement information sent by the terminal; and selecting a decoder corresponding to the first scene information from the plurality of decoders according to the first scene information as the first decoder.

Here, the above-described first scene information may be an Index (Index) indicating the first scene.

In this embodiment of the present application, the network device may further send first model configuration information to the terminal. Specific: the first model configuration information is used for configuring an encoder at the terminal under the condition that the plurality of decoders and the second decision network are configured by the second model configuration information, and the encoder is used for compressing the measurement result of the downlink reference signal measured by the terminal; and in case the plurality of decoders are configured by the second model configuration information, the first model configuration information is used for configuring the encoder and a first decision network at the terminal, and the first decision network is used for identifying a corresponding scene based on the measurement result of the downlink reference signal measured by the terminal.

The training of the network model in the embodiments of the present application is described below. The training of the network model may be performed at the network device or at the server side. The following description will take training on the network device side as an example.

In this embodiment of the present application, the network device may perform multiple rounds of iterative training on the encoder and the multiple decoders based on the second measurement information of the downlink reference signals in the multiple scenarios acquired in advance, until a preset training end condition is met, so as to obtain the trained encoder and multiple decoders; wherein each round of iterative training comprises:

Through the above training, a trained encoder and a plurality of decoders can be obtained.

Further, the embodiment of the application can train the first decision network and/or train the second decision network.

The training of the first decision network specifically includes:

(1) Inputting second measurement information of downlink reference signals of each scene into a first judgment network to obtain second index information output by the first judgment network; inputting second measurement information of downlink reference signals of each scene to a trained encoder to obtain fourth compression information output by the encoder; respectively inputting the fourth compressed information to each trained decoder to obtain measurement information corresponding to the fourth compressed information output by each decoder; and calculating to obtain a third loss value corresponding to each decoder according to the second measurement information and the fourth compression information of the downlink reference signal in the scene, and determining third index information of the decoder corresponding to the minimum third loss value.

(2) Calculating fourth loss values under the scenes according to the second index information and the third index information, counting the fourth loss values under all the scenes to obtain fifth loss values, calculating a third gradient based on the fifth loss values, and transmitting the third gradient in a reverse path from the first judgment network to the encoder to adjust network parameters of the first judgment network until preset training ending conditions are met, so as to obtain a trained first judgment network.

Training a second decision network, comprising in particular:

(1) Inputting second measurement information of downlink reference signals of each scene to a trained encoder to obtain fourth compression information output by the encoder; the fourth compressed information is respectively input into a second judgment network and each trained decoder to obtain fourth index information output by the second judgment network and measurement information corresponding to the fourth compressed information output by each decoder; and calculating to obtain a third loss value corresponding to each decoder according to the second measurement information and the fourth compression information of the downlink reference signal in the scene, and determining third index information of the decoder corresponding to the minimum third loss value.

(2) And calculating a sixth loss value under the scene according to the fourth index information and the third index information, counting the sixth loss values under all scenes to obtain a seventh loss value, calculating a fourth gradient based on the seventh loss value, and transmitting the fourth gradient in a reverse path from the first judgment network to the encoder so as to adjust network parameters of the second judgment network until a preset training ending condition is met, thereby obtaining a trained second judgment network.

In this embodiment of the present application, the training end condition of each network model may be that training reaches a preset number of rounds, or meets a preset convergence condition, or the like.

The above methods of embodiments of the present application are further described below by way of several examples.

Example 1:

in example 1, the encoder and the first decision network are disposed on a terminal (UE) side, and the plurality of decoders are disposed on a network device (gNB) side. As shown in fig. 7, the processing method of channel information in this example includes the steps of:

step 70, performing offline training on the encoder and the decoder corresponding to different subtasks, and loading the trained model to a network side;

step 71, configuring a channel compression coding and decoding mode used by the network high-level RRC signaling;

Step 72, the network side transmits the encoder and the first decision network (GateNet) to the terminal;

step 73, the network side transmits the CSI-RS to the terminal;

step 74, the terminal obtains downlink CSI according to the received reference signal measurement;

step 75, compressing the measured CSI by the terminal through an encoder;

step 76, the terminal inputs the measured CSI into a first decision network, and outputs an index (index) of a scene, namely, an index required by a decoder is selected, namely, the index of the decoder;

step 77, the terminal sends the CSI compression result and index to the network side;

step 78, the network side selects the corresponding decoder according to the index;

step 79, the network side decompresses the CSI by using the selected decoder;

example 2:

in example 2, the encoder is disposed at the terminal side, and the second decision network and the plurality of decoders are disposed at the network device side. . As shown in fig. 8, the processing method of channel information in this example includes the steps of:

step 80, performing offline training on the encoder and the decoders corresponding to different subtasks, and loading the trained model to a network side;

step 81, configuring a channel compression coding and decoding mode used by the network high-level RRC signaling;

step 82, the network side transmits the encoder to the terminal;

Step 83, the network side sends CSI-RS to the terminal;

step 84, the terminal obtains downlink CSI according to the received reference signal measurement;

step 85, the terminal compresses the measured CSI through an encoder;

step 86, the terminal sends the CSI compression result to the network side;

step 87, the network side inputs the received compression result into GateNet, and outputs the index (index) of the scene, namely, selects the index required by the decoder, namely, the index of the decoder;

step 88, the network side selects a corresponding decoder according to the index;

step 89. Network side decompresses CSI using selected decoder

The model training process of the embodiment of the present application is illustrated below.

The embodiment of the application can obtain the encoder and the decoder which can adapt to a series of subtasks of different scenes through offline training.

Fig. 9 is a schematic diagram of each iteration process in the training process of the codec. Fig. 10 is a schematic diagram of the training of the encoder, and fig. 11 is a schematic diagram of the training of the subtask decoder.

Wherein, as shown in FIG. 10, in the offline training process, the subtask data sets collected under each scene need to be mixed into a larger data set, the mixed data set is propagated forward through the encoder, the decoder corresponding to each subtask and the temporary common decoder, and in the back propagation, the Loss value (Loss) used is calculated by the following formula, wherein L _MTL Representing the loss value, T is the number of subtasks, N _k For the number of data samples in the kth subtask, enc (·) represents the encoder network, Φ represents the network parameters of the encoder, dec _k (. Cndot.) represents the decoder k network, ψ _k Representing the network parameters of the decoder k,representing the channel matrix corresponding to the encoder k network structure. The calculated Loss is used to counter-propagate along the path of the temporary common decoder and encoder to update the model parameters.

As shown in fig. 11, for training of each subtask decoder, the parameters of the encoder need to be frozen, each subtask data is propagated forward along the encoder and the corresponding subtask decoder, the respective Loss value is calculated, and then the Loss value is propagated backward along the original path to adjust the parameters of each subtask decoder, so as to finish one round of offline training iteration, and the flow of each round of training iteration in the codec training process is shown in fig. 9.

The present example obtains an AI model, i.e., a decision network, capable of determining a subtask decoder selection strategy based on a compressed feedback result (or CSI actual measurement result) through offline training. After the encoder and the decoder of each subtask are trained, selecting one subtask data to be input into the encoder to obtain a compression result of the subtask data, and taking the compression result as a training set of a second judgment network (the subtask data is directly used as the training set for a first judgment network); the compressed result of the subtask continues to forward propagate, so that the compressed result passes through all subtask decoders, and the forward propagation result of all subtask decoders on the subtask data can be obtained. By comparing the output of each subtask decoder with the Loss value of the subtask data, the subtask decoder with the minimum Loss value is selected as the actually used decoder, and the judgment network outputs the label corresponding to the subtask decoder. Based on the training of a large amount of data in various different subtasks, a decision network can be obtained that can determine how to select a subtask decoder by compressing intermediate results (or CSI actual measurement results).

It can be seen that two network architectures are available in the embodiments of the present application, the first network comprising an encoder, a first decision network and respective decoders, and the second network comprising an encoder, a second decision network and respective decoders.

In addition, in the embodiment of the present application, the network side may configure what kind of generalized channel compression method with low transmission overhead is adopted, that is, the first or second network structure is adopted.

For example, embodiments of the present application may consider modifying codebook configuration (codebook) in RRC. After the method of the embodiment of the application is applied, for the codebook mode parameter, a mode tag can be added, for example, the mode tag takes a value of 0 when a first network is adopted, and takes a value of 1 when a second network structure is adopted. The value of the n1-n2-codebook subsetreference parameter needs to take into account the performance of the model itself used. The n1-n2 parameters may be used to describe the size of the channel measured matrix to be compressed. While numberOfBeams, paramCombination, portSelectionSamplingSize, parameters related to numberOfPMI-subepidsps-subend, etc. only to the conventional type-I and type ii may not be configured.

In the on-site reasoning, after receiving the channel compression intermediate result, the network side inputs the intermediate result into a second decision network stored in the network side to obtain a subtask encoder selection tag (equivalent to index information of a scene), or the terminal directly inputs the CSI into a first decision network received by the terminal to obtain the subtask encoder selection tag and feeds the subtask encoder selection tag back to the network side. When the decision network infers and acquires the sub-task encoder selection tag, the terminal feedback CSI compression result (or CSI actual measurement result) is taken as the input of the decision network, the corresponding sub-task encoder number is output, and considering that the characteristics of the channel actual measurement data are greatly different in different scenes, in the embodiment of the application, different sub-tasks can be classified through different scenes, and an example of dividing the sub-tasks (scenes) according to regions is shown in a table 1.

TABLE 1

Referring to the nomenclature of S-to-S and M-to-M in the prior art, the self-encoder structure in the embodiments of the present application may be referred to as S-to-M, and the pair of performances of S-to-S and M-to-M are shown in FIG. 12. The three subgraphs from left to right in fig. 12 correspond to three neural networks, with increasing complexity from left to right. In each sub-graph, the X-axis represents compression rate, decreasing in sequence from left to right; the broken line represents NMSE performance (averaged across subtask results) corresponding to the left Y-axis, while the bar graph (right Y-axis) represents NMSE performance improvement for S-to-M versus S-to-S, i.e., NMSE (S-to-S) -NMSE (S-to-M) [ dB ]. The higher bar shows that S-to-M improves NMSE performance more strongly than S-to-S. It can be seen that the S-to-M and M-to-M are significantly more powerful than S-to-S, while S-to-M is only marginally inferior to M-to-M, so that the S-to-M structured self-encoder can be considered as a low-overhead and highly generalized precoder.

From the above, it can be seen that, by using a single encoder and multiple decoder structure, the embodiment of the present application reduces the training difficulty of the AI model and can obtain better performance compared with a single encoder and decoder structure; compared with a multi-codec structure, the method can reduce the data overhead when the network side issues the encoder, and can avoid reporting the air interface data overhead of the encoder selection decision and transmission decision network when selecting the model structure of the second decision network. In addition, the embodiment of the application can flexibly select the use mode of the self-encoder through high-level configuration so as to cope with challenges in more severe channel environments.

Various methods of embodiments of the present application are described above. An apparatus for carrying out the above method is further provided below.

Referring to fig. 13, an embodiment of the present application further provides a terminal 1300, including:

a measurement module 1301, configured to measure a downlink reference signal, and obtain first measurement information;

a compression module 1302, configured to perform compression processing on the first measurement information by using an encoder to obtain first compression information;

the sending module 1303 is configured to send the first compressed information to a network according to the current first model configuration information, or send the first compressed information and the first scene information corresponding to the first measurement information to the network.

Through the modules, the embodiment of the application can solve the problem that the complexity and the performance of the network model for channel information processing cannot be considered.

Optionally, the terminal further includes:

the system comprises a receiving module, a processing module and a processing module, wherein the receiving module is used for receiving first model configuration information sent by a network, the first model configuration information is used for configuring the encoder at a terminal, or is used for configuring the encoder and a first judgment network at the terminal, and the first judgment network is used for identifying scene information.

Optionally, the terminal further includes:

A construction module for constructing the encoder according to the first model configuration information in a case where the first model configuration information is used to configure the encoder at a terminal; in case the first model configuration information is used to configure the encoder and the first decision network at the terminal, the encoder and the first decision network are constructed according to the first model configuration information.

Optionally, the sending module is further configured to send the first compressed information to a network in a case where the first model configuration information is used to configure the encoder at a terminal; and sending first compression information and first scene information corresponding to the first measurement information to a network under the condition that the first model configuration information is used for configuring the encoder and the first judgment network at a terminal.

Optionally, the terminal further includes:

and the determining module is used for inputting the first measurement information into the first judgment network to obtain the first scene information under the condition that the first model configuration information is used for configuring the encoder and the first judgment network at the terminal.

The device in this embodiment corresponds to the method applied to the terminal, and the implementation manner in each embodiment is applicable to the embodiment of the device, so that the same technical effects can be achieved. The above device provided in this embodiment of the present application can implement all the method steps implemented in the above method embodiment, and can achieve the same technical effects, and detailed descriptions of the same parts and beneficial effects as those in the method embodiment in this embodiment are omitted herein.

Referring to fig. 14, an embodiment of the present application further provides a terminal 1400, including: a transceiver 1401 and a processor 1402;

the processor 1402 is configured to measure a downlink reference signal to obtain first measurement information; compressing the first measurement information through an encoder to obtain first compressed information;

the transceiver 1401 is configured to send the first compressed information to a network according to the first model configuration information, or send the first compressed information and the first scene information corresponding to the first measurement information to the network.

Optionally, the transceiver is further configured to receive first model configuration information sent by a network, where the first model configuration information is used to configure the encoder at the terminal, or is used to configure the encoder and a first decision network at the terminal, and the first decision network is used to identify scene information.

Optionally, the processor is further configured to construct the encoder according to the first model configuration information if the first model configuration information is used to configure the encoder at a terminal; in case the first model configuration information is used to configure the encoder and the first decision network at the terminal, the encoder and the first decision network are constructed according to the first model configuration information.

Optionally, the transceiver is further configured to send the first compressed information to a network if the first model configuration information is used to configure the encoder at a terminal; and sending first compression information and first scene information corresponding to the first measurement information to a network under the condition that the first model configuration information is used for configuring the encoder and the first judgment network at a terminal.

Optionally, the processor is further configured to, in a case where the first model configuration information is used to configure the encoder and the first decision network at a terminal, input the first measurement information to the first decision network to obtain the first scene information.

Referring to fig. 15, an embodiment of the present application further provides a network device 1500, including:

a transmitting module 1501, configured to transmit a downlink reference signal;

a receiving module 1502, configured to receive first compressed information of a downlink reference signal measurement result sent by a terminal;

a selection module 1503 for selecting a first decoder from a plurality of decoders, wherein each decoder of the plurality of decoders corresponds to one of a plurality of scenes;

the decompression module 1504 is configured to input the first compressed information to the first decoder, decompress the first compressed information by the first decoder, and obtain measurement information corresponding to the first compressed information.

Optionally, the selection module is further configured to select a first decoder from a plurality of decoders based on second model configuration information at the network device,

Optionally, the sending module is further configured to send first model configuration information to the terminal,

Optionally, the network device further includes:

the first training module is used for performing multi-round iterative training on the encoder and the plurality of decoders based on the pre-acquired second measurement information of the downlink reference signals in the various scenes until a preset training ending condition is met, so that the trained encoder and the trained plurality of decoders are obtained; wherein each round of iterative training comprises:

Optionally, the network device further includes:

the second training module is used for training the first judgment network, and/or the third training module is used for training the second judgment network.

Optionally, the second training module is specifically configured to:

Optionally, the third training module is specifically configured to:

It should be noted that, the device in this embodiment corresponds to the method applied to the network side, and the implementation manners in the foregoing embodiments are all applicable to the embodiment of the device, so that the same technical effects can be achieved. The above device provided in this embodiment of the present application can implement all the method steps implemented in the above method embodiment, and can achieve the same technical effects, and detailed descriptions of the same parts and beneficial effects as those in the method embodiment in this embodiment are omitted herein.

Referring to fig. 16, an embodiment of the present application further provides a network device 1600, including: a transceiver 1601 and a processor 1602;

the transceiver 1601 is configured to send a downlink reference signal; receiving first compressed information of a downlink reference signal measurement result sent by a terminal;

the processor 1602 is configured to select a first decoder from a plurality of decoders, wherein each decoder of the plurality of decoders corresponds to a respective one of a plurality of scenes; and inputting the first compressed information to the first decoder, and decompressing the first compressed information by the first decoder to obtain measurement information corresponding to the first compressed information.

Optionally, the processor is further configured to select a first decoder from a plurality of decoders based on second model configuration information at the network device,

Optionally, the transceiver is further configured to send first model configuration information to the terminal,

Optionally, the processor is further configured to perform multiple rounds of iterative training on the encoder and the multiple decoders based on the pre-acquired second measurement information of the downlink reference signals in the multiple scenes until a preset training end condition is met, so as to obtain the trained encoder and the multiple decoders; wherein each round of iterative training comprises:

Optionally, the processor is further configured to train the first decision network, and/or train the second decision network.

Optionally, the processor is further configured to input second measurement information of the downlink reference signal of each scene to the first decision network, so as to obtain second index information output by the first decision network; inputting second measurement information of downlink reference signals of each scene to a trained encoder to obtain fourth compression information output by the encoder; respectively inputting the fourth compressed information to each trained decoder to obtain measurement information corresponding to the fourth compressed information output by each decoder; according to the second measurement information and the fourth compression information of the downlink reference signal in the scene, calculating to obtain a third loss value corresponding to each decoder, and determining third index information of the decoder corresponding to the minimum third loss value;

Optionally, the processor is further configured to input second measurement information of the downlink reference signal of each scene to the trained encoder, to obtain fourth compression information output by the encoder; the fourth compressed information is respectively input into a second judgment network and each trained decoder to obtain fourth index information output by the second judgment network and measurement information corresponding to the fourth compressed information output by each decoder; according to the second measurement information and the fourth compression information of the downlink reference signal in the scene, calculating to obtain a third loss value corresponding to each decoder, and determining third index information of the decoder corresponding to the minimum third loss value;

Referring to fig. 17, the embodiment of the present application further provides a terminal 1760, including a processor 1761, a memory 1762, and a computer program stored in the memory 1762 and capable of running on the processor 1761, where the computer program when executed by the processor 1761 implements each process of the above embodiment of the method for processing channel information executed by the terminal, and can achieve the same technical effects, so that repetition is avoided and redundant description is omitted herein.

Referring to fig. 18, the embodiment of the present application further provides a network device 1800, which includes a processor 1801, a memory 1802, and a computer program stored in the memory 1802 and capable of running on the processor 1801, where the computer program when executed by the processor 1801 implements the processes of the foregoing embodiment of the method for processing channel information executed by the network device, and the same technical effects can be achieved, so that repetition is avoided and redundant description is omitted herein.

The embodiment of the application also provides a computer readable storage medium, on which a computer program is stored, where the computer program when executed by a processor implements each process of the above-mentioned channel information processing method embodiment, and the same technical effects can be achieved, so that repetition is avoided, and no further description is given here. Wherein the computer readable storage medium is selected from Read-Only Memory (ROM), random access Memory (Random Access Memory, RAM), magnetic disk or optical disk.

It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk), including several instructions for causing a terminal (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method described in the embodiments of the present application.

The embodiments of the present application have been described above with reference to the accompanying drawings, but the present application is not limited to the above-described embodiments, which are merely illustrative and not restrictive, and many forms may be made by those of ordinary skill in the art without departing from the spirit of the present application and the scope of the claims, which are also within the protection of the present application.

Claims

1. A method for processing channel information, comprising:

2. The method as recited in claim 1, further comprising:

3. The method of claim 2, wherein,

in the case that the first model configuration information is used to configure the encoder at a terminal, the terminal transmits the first compressed information to a network; and/or the number of the groups of groups,

4. The method of claim 3, wherein,

in case the first model configuration information is used to configure the encoder and the first decision network at a terminal, the terminal inputs the first measurement information to the first decision network, obtaining the first scene information.

5. A method for processing channel information, comprising:

the network equipment sends a downlink reference signal;

6. The method as recited in claim 5, further comprising:

Receiving first scene information corresponding to first measurement information sent by a terminal under the condition that the second model configuration information configures the plurality of decoders; and selecting a decoder corresponding to the first scene information from the plurality of decoders according to the first scene information as the first decoder.

7. The method as recited in claim 6, further comprising:

transmitting first model configuration information to the terminal,

8. The method as recited in claim 7, further comprising:

9. The method as recited in claim 8, further comprising:

10. The method of claim 9, wherein the step of training the first decision network comprises:

11. The method of claim 9, wherein the step of training the second decision network comprises:

12. A terminal comprising a transceiver and a processor, wherein,

13. A terminal, comprising: a processor, a memory and a program stored on the memory and executable on the processor, which when executed by the processor, performs the steps of the method according to any one of claims 1 to 4.

14. A network device comprising a transceiver and a processor, wherein,

15. A network device, comprising: a processor, a memory and a program stored on the memory and executable on the processor, which when executed by the processor, performs the steps of the method according to any one of claims 5 to 11.

16. A computer-readable storage medium, on which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 11.