WO2023123062A1

WO2023123062A1 - Quality evaluation method for virtual channel sample, and device

Info

Publication number: WO2023123062A1
Application number: PCT/CN2021/142531
Authority: WO
Inventors: 刘文东
Original assignee: Oppo广东移动通信有限公司
Priority date: 2021-12-29
Filing date: 2021-12-29
Publication date: 2023-07-06

Abstract

A quality evaluation method for a virtual channel sample, and a device. The method comprises: a first device acquiring first input information and at least one piece of second input information, wherein the first input information is information of a real channel sample, the second input information is information of a virtual channel sample, the at least one piece of second input information corresponds to at least one generator model, and each piece of second input information is generated by a corresponding generator model; and the first device performing quality evaluation on the at least one piece of second input information and/or determining a target generator model from the at least one generator model according to the first input information and the at least one piece of second input information.

Description

Method and device for quality assessment of virtual channel samples

technical field

The embodiments of the present application relate to the communication field, and in particular to a method and device for evaluating the quality of virtual channel samples.

Background technique

For the collection of wireless channel samples, considering the difficulty and cost of collection, it is difficult to collect all scenes and features. Using Generative Adversarial Network (GAN) to realize the generation of virtual channel samples and the expansion of data sets is An efficient method for large-scale dataset acquisition. However, this approach also has significant disadvantages:

First, when the network design of the GAN generator and discriminator, loss function, etc. are not properly selected, and the training does not reach a stable convergence, it is difficult to fit the distribution of real channel samples, resulting in a relatively low similarity between virtual channel samples and real channel samples. poor, unusable;

Second, in order to achieve sufficient similarity, the GAN generator may tend to be conservative in order to fit the distribution of real channel samples during the training process, that is, the generated virtual channel samples are highly similar to real channel samples, but The similarity between virtual channel samples is also high, which is the so-called mode collapse. As a result, a large number of virtual channel samples do not actually cover more channel features, so there is no purpose of effectively increasing the size of the data set.

Therefore, how to design and implement a quality evaluation method for virtual channel samples to guide the training process of GAN, generate high-quality virtual channel samples for the construction of large-scale high-quality channel data sets, to support artificial intelligence and The research on the integration of wireless communication systems is a key problem to be solved urgently.

Contents of the invention

The present application provides a method and device for evaluating the quality of virtual channel samples, which can realize effective quality evaluation of virtual channel samples.

In a first aspect, a method for evaluating the quality of a virtual channel sample is provided, including: a first device acquires first input information and at least one second input information, wherein the first input information is information of a real channel sample, and the The second input information is information of virtual channel samples, the at least one second input information corresponds to at least one generator model, and each second input information is generated by a corresponding generator model;

The first device performs quality assessment on the at least one second input information and/or determines a target generator model in the at least one generator model according to the first input information and the at least one second input information.

In a second aspect, a method for evaluating the quality of a virtual channel sample is provided, including: first information sent by the second device to the first device, where the first information includes at least one of the following: first input information, which is Information about real channel samples; at least one second input information is information about virtual channel samples, the at least one second input information corresponds to at least one generator model, and each second input information is generated by a corresponding generator model; The index of at least one generator.

In a third aspect, a method for evaluating the quality of a virtual channel sample is provided, including: the first device determines the quality evaluation information of the second input information according to the first input information and the second input information; wherein, the first The input information is real channel sample information, the second input information is virtual channel sample information, and the quality evaluation information of the second input information is used to indicate the similarity between the virtual channel sample and the real channel sample and/or the virtual channel Sample dispersion.

In a fourth aspect, a wireless communication device is provided, configured to execute the method in the above first aspect or various implementations thereof.

Specifically, the first device includes a functional module configured to execute the method in the foregoing first aspect or each implementation manner thereof.

In a fifth aspect, a device for wireless communication is provided, configured to execute the method in the above second aspect or various implementations thereof.

Specifically, the second device includes a functional module configured to execute the method in the above second aspect or each implementation manner thereof.

In a sixth aspect, a device for wireless communication is provided, configured to perform the method in the above second aspect or various implementations thereof.

Specifically, the second device includes a functional module for executing the method in the above third aspect or each implementation manner thereof.

In a seventh aspect, a wireless communication device is provided, including a processor and a memory. The memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute the method in the above first aspect or its various implementations.

In an eighth aspect, a wireless communication device is provided, including a processor and a memory. The memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute the method in the above second aspect or its various implementations.

In a ninth aspect, a wireless communication device is provided, including a processor and a memory. The memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute the method in the above third aspect or its various implementations.

In a tenth aspect, a chip is provided, configured to implement any one of the foregoing first to third aspects or the method in each implementation manner thereof.

Specifically, the chip includes: a processor, configured to call and run a computer program from the memory, so that the device installed with the device executes any one of the above-mentioned first to third aspects or any of the implementations thereof. method.

In an eleventh aspect, there is provided a computer-readable storage medium for storing a computer program, and the computer program causes a computer to execute any one of the above-mentioned first to third aspects or the method in each implementation manner thereof.

In a twelfth aspect, a computer program product is provided, including computer program instructions, the computer program instructions causing a computer to execute any one of the above first to third aspects or the method in each implementation manner.

A thirteenth aspect provides a computer program, which, when running on a computer, causes the computer to execute any one of the above first to third aspects or the method in each implementation manner.

Through the above technical solution, the device can acquire first input information of real channel samples and at least one second input information of virtual channel samples, where the at least one second input information corresponds to at least one generator model. Further perform quality assessment on the virtual channel samples based on the first input information and the at least one second input information, and/or, the target generator model in the at least one generator model can support online virtual channel quality assessment and generator Model selection is conducive to meeting the data set expansion requirements such as model update and transfer learning in communication systems.

Description of drawings

FIG. 1 is a schematic diagram of a communication system architecture provided by an embodiment of the present application.

Fig. 2 is a schematic diagram of a CSI autoencoder model.

Fig. 3 is a schematic schematic diagram of generating an adversarial network.

Fig. 4 is a schematic diagram of a method for evaluating the quality of virtual channel samples according to an embodiment of the present application.

Fig. 5 is a schematic diagram of a time-domain channel sample provided by an embodiment of the present application.

Fig. 6 is a schematic diagram of a channel sample in the frequency domain provided by an embodiment of the present application.

Fig. 7 is a schematic diagram of an angle-domain channel sample provided by an embodiment of the present application.

Fig. 8 is a schematic interaction diagram of a method for evaluating the quality of a virtual channel sample provided by an embodiment of the present application.

Fig. 9 is a schematic interaction diagram of another method for evaluating the quality of virtual channel samples provided by an embodiment of the present application.

Fig. 10 is a schematic interaction diagram of another method for evaluating the quality of virtual channel samples provided by an embodiment of the present application.

Fig. 11 is a schematic interaction diagram of yet another method for evaluating the quality of virtual channel samples provided by an embodiment of the present application.

Fig. 12 is a working principle diagram of a method for evaluating the quality of a virtual channel sample according to an embodiment of the present application.

Fig. 13 is a schematic diagram of another method for evaluating the quality of virtual channel samples according to an embodiment of the present application.

Fig. 14 is a working principle diagram of a method for evaluating the quality of a virtual channel sample according to an embodiment of the present application.

Fig. 15 is a schematic block diagram of a wireless communication device provided according to an embodiment of the present application.

Fig. 16 is a schematic block diagram of another wireless communication device provided according to an embodiment of the present application.

Fig. 17 is a schematic block diagram of another wireless communication device provided according to an embodiment of the present application.

Fig. 18 is a schematic block diagram of a communication device provided according to an embodiment of the present application.

Fig. 19 is a schematic block diagram of a chip provided according to an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, but not all of the embodiments. With regard to the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the scope of protection of this application.

The technical solution of the embodiment of the present application can be applied to various communication systems, such as: Global System of Mobile communication (Global System of Mobile communication, GSM) system, code division multiple access (Code Division Multiple Access, CDMA) system, broadband code division multiple access (Wideband Code Division Multiple Access, WCDMA) system, General Packet Radio Service (GPRS), Long Term Evolution (LTE) system, Advanced long term evolution (LTE-A) system , New Radio (NR) system, evolution system of NR system, LTE (LTE-based access to unlicensed spectrum, LTE-U) system on unlicensed spectrum, NR (NR-based access to unlicensed spectrum) on unlicensed spectrum unlicensed spectrum (NR-U) system, Non-Terrestrial Networks (NTN) system, Universal Mobile Telecommunications System (UMTS), Wireless Local Area Networks (WLAN), Wireless Fidelity (Wireless Fidelity, WiFi), fifth-generation communication (5th-Generation, 5G) system or other communication systems, etc.

Generally speaking, the number of connections supported by traditional communication systems is limited and easy to implement. However, with the development of communication technology, mobile communication systems will not only support traditional communication, but also support, for example, Device to Device (Device to Device, D2D) communication, Machine to Machine (M2M) communication, Machine Type Communication (MTC), Vehicle to Vehicle (V2V) communication, or Vehicle to everything (V2X) communication, etc. , the embodiments of the present application may also be applied to these communication systems.

Optionally, the communication system in the embodiment of the present application may be applied to a carrier aggregation (Carrier Aggregation, CA) scenario, may also be applied to a dual connectivity (Dual Connectivity, DC) scenario, and may also be applied to an independent (Standalone, SA) deployment Web scene.

Optionally, the communication system in the embodiment of the present application may be applied to an unlicensed spectrum, where the unlicensed spectrum may also be considered as a shared spectrum; or, the communication system in the embodiment of the present application may also be applied to a licensed spectrum, where, Licensed spectrum can also be considered as non-shared spectrum.

The embodiments of the present application describe various embodiments in conjunction with network equipment and terminal equipment, wherein the terminal equipment may also be referred to as user equipment (User Equipment, UE), access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent or user device, etc.

The terminal device can be a station (STATION, ST) in a WLAN, a cellular phone, a cordless phone, a Session Initiation Protocol (Session Initiation Protocol, SIP) phone, a wireless local loop (Wireless Local Loop, WLL) station, a personal digital assistant (Personal Digital Assistant, PDA) devices, handheld devices with wireless communication functions, computing devices or other processing devices connected to wireless modems, vehicle-mounted devices, wearable devices, next-generation communication systems such as terminal devices in NR networks, or future Terminal equipment in the evolved public land mobile network (Public Land Mobile Network, PLMN) network, etc.

In the embodiment of this application, the terminal device can be deployed on land, including indoor or outdoor, handheld, wearable or vehicle-mounted; it can also be deployed on water (such as ships, etc.); it can also be deployed in the air (such as aircraft, balloons and satellites) superior).

In this embodiment of the application, the terminal device may be a mobile phone (Mobile Phone), a tablet computer (Pad), a computer with a wireless transceiver function, a virtual reality (Virtual Reality, VR) terminal device, an augmented reality (Augmented Reality, AR) terminal Equipment, wireless terminal equipment in industrial control, wireless terminal equipment in self driving, wireless terminal equipment in remote medical, wireless terminal equipment in smart grid , wireless terminal equipment in transportation safety, wireless terminal equipment in smart city, or wireless terminal equipment in smart home.

As an example but not a limitation, in this embodiment of the present application, the terminal device may also be a wearable device. Wearable devices can also be called wearable smart devices, which is a general term for the application of wearable technology to intelligently design daily wear and develop wearable devices, such as glasses, gloves, watches, clothing and shoes. A wearable device is a portable device that is worn directly on the body or integrated into the user's clothing or accessories. Wearable devices are not only a hardware device, but also achieve powerful functions through software support, data interaction, and cloud interaction. Generalized wearable smart devices include full-featured, large-sized, complete or partial functions without relying on smart phones, such as smart watches or smart glasses, etc., and only focus on a certain type of application functions, and need to cooperate with other devices such as smart phones Use, such as various smart bracelets and smart jewelry for physical sign monitoring.

In the embodiment of the present application, the network device may be a device for communicating with the mobile device, and the network device may be an access point (Access Point, AP) in WLAN, a base station (Base Transceiver Station, BTS) in GSM or CDMA , or a base station (NodeB, NB) in WCDMA, or an evolved base station (Evolutional Node B, eNB or eNodeB) in LTE, or a relay station or access point, or a vehicle-mounted device, a wearable device, and an NR network The network equipment (gNB) in the network or the network equipment in the future evolved PLMN network or the network equipment in the NTN network, etc.

As an example but not a limitation, in this embodiment of the present application, the network device may have a mobile feature, for example, the network device may be a mobile device. Optionally, the network equipment may be a satellite or a balloon station. For example, the satellite can be a low earth orbit (low earth orbit, LEO) satellite, a medium earth orbit (medium earth orbit, MEO) satellite, a geosynchronous earth orbit (geosynchronous earth orbit, GEO) satellite, a high elliptical orbit (High Elliptical Orbit, HEO) satellite. ) Satellite etc. Optionally, the network device may also be a base station installed on land, water, and other locations.

In this embodiment of the present application, the network device may provide services for a cell, and the terminal device communicates with the network device through the transmission resources (for example, frequency domain resources, or spectrum resources) used by the cell, and the cell may be a network device ( For example, a cell corresponding to a base station), the cell may belong to a macro base station, or may belong to a base station corresponding to a small cell (Small cell), and the small cell here may include: a metro cell (Metro cell), a micro cell (Micro cell), a pico cell ( Pico cell), Femto cell, etc. These small cells have the characteristics of small coverage and low transmission power, and are suitable for providing high-speed data transmission services.

Exemplarily, a communication system 100 applied in this embodiment of the application is shown in FIG. 1 . The communication system 100 may include a network device 110, and the network device 110 may be a device for communicating with a terminal device 120 (or called a communication terminal, terminal). The network device 110 can provide communication coverage for a specific geographical area, and can communicate with terminal devices located in the coverage area.

FIG. 1 exemplarily shows one network device and two terminal devices. Optionally, the communication system 100 may include multiple network devices and each network device may include other numbers of terminal devices within the coverage area. This application The embodiment does not limit this.

Optionally, the communication system 100 may further include other network entities such as a network controller and a mobility management entity, which is not limited in this embodiment of the present application.

It should be understood that a device with a communication function in the network/system in the embodiment of the present application may be referred to as a communication device. Taking the communication system 100 shown in FIG. 1 as an example, the communication equipment may include a network equipment 110 and a terminal equipment 120 with communication functions. The network equipment 110 and the terminal equipment 120 may be the specific equipment described above, and will not be repeated here. The communication device may also include other devices in the communication system 100, such as network controllers, mobility management entities and other network entities, which are not limited in this embodiment of the present application.

It should be understood that the terms "system" and "network" are often used interchangeably herein. The term "and/or" in this article is just an association relationship describing associated objects, which means that there can be three relationships, for example, A and/or B can mean: A exists alone, A and B exist simultaneously, and there exists alone B these three situations. In addition, the character "/" in this article generally indicates that the contextual objects are an "or" relationship.

It should be understood that the "indication" mentioned in the embodiments of the present application may be a direct indication, may also be an indirect indication, and may also mean that there is an association relationship. For example, A indicates B, which can mean that A directly indicates B, for example, B can be obtained through A; it can also indicate that A indirectly indicates B, for example, A indicates C, and B can be obtained through C; it can also indicate that there is an association between A and B relation.

In the description of the embodiments of the present application, the term "corresponding" may indicate that there is a direct or indirect correspondence between the two, or that there is an association between the two, or that it indicates and is indicated, configuration and is configuration etc.

In the embodiment of this application, "predefinition" can be realized by pre-saving corresponding codes, tables or other methods that can be used to indicate related information in devices (for example, including terminal devices and network devices). The implementation method is not limited. For example, pre-defined may refer to defined in the protocol.

In the embodiment of the present application, the "protocol" may refer to a standard protocol in the communication field, for example, it may include the LTE protocol, the NR protocol, and related protocols applied in future communication systems, which is not limited in the present application.

In order to facilitate understanding of the embodiments of the present application, a physical layer solution related to artificial intelligence (Artificial Intelligence, AI) is described.

With the rapid development of AI, wireless communication combined with AI, especially the physical layer solution, has become an important evolution technology for the development of future wireless communication systems, and has attracted extensive attention from academia and industry. Generally speaking, AI-based solutions need to build high-quality data sets and design adapted deep neural networks (Deep Neural Network, DNN), convolutional neural networks (Convolutional Neural Network, CNN), and cyclic neural networks ( Recurrent Neural Network, RNN) and other deep learning network architectures. AI technology can mine the potential characteristics of large data sets and fit nonlinear mapping from input to output to better complete task goals. The AI-based wireless communication physical layer solution has significant advantages in nonlinear problems that are difficult to model, and can obtain obvious gains.

Taking AI-based Channel State Information (CSI) feedback as an example, by constructing a CSI autoencoder as shown in Figure 2, including an encoder on the terminal side and a decoder on the network side, the CSI input can be compressed into The feedback bit stream is fed back to the network side, and the decoder restores the original CSI. This scheme can realize the network architecture with less feedback bits by rationally designing the structure of the encoder and decoder, such as using the more classic residual neural network (Residual Neural Network, ResNet) structure, Transformer self-attention mechanism and other network architectures. , to achieve better CSI feedback performance. Wherein, the CSI input and the CSI output may use full channel information or channel feature vectors.

Taking the compressed feedback of channel eigenvectors as an example, under the configuration of link-level CDL-C300 channel and feedback bit 48bit, the cosine similarity square value of the restored CSI eigenvector of the enhanced Type II (eTypeII) codebook is specified to be 0.84, Based on the ResNet autoencoder architecture, it can achieve a CSI restoration degree with a cosine similarity square value of 0.94, which can significantly improve the CSI restoration accuracy compared to the existing classic eTypeII codebook scheme.

However, the AI-based CSI autoencoder scheme needs to construct a large number of channel data sets. The above CSI recovery degree is based on a data set containing 10 ⁵ channel samples constructed under a channel with 1000 terminals sampled and 100 time slot samples. Obtained from the training of the CSI autoencoder. The large-scale data set covers the sample characteristics of the channel in each dimension of space, time and frequency, which is beneficial to the deep network model to extract features and perform nonlinear fitting. However, when the data set size is insufficient, the channel sample characteristics cannot be effectively covered, and the AI-based CSI self-encoding scheme will form overfitting on a small-scale data set, and the generalization is poor, and the performance on the actual deployment test channel will be poor. There was a significant decline.

Therefore, in the development of future wireless communication systems, it is necessary to obtain large-scale, high-quality channel data sets to ensure the generalization of the model, thereby improving the performance of AI-based wireless communication physical layer solutions.

To facilitate understanding of the embodiments of the present application, channel sample collection and channel modeling related to the present application will be described.

In order to obtain large-scale, high-quality channel data sets, one of the most direct implementation methods is to collect actual wireless channels. For the actual cell environment, wireless channel information is obtained by deploying signal transmitters and signal receivers, or by collecting signals from third-party transmitters (such as base stations) through specific receivers to obtain wireless channel information. Through the above method, the propagation characteristics of the wireless channel can be obtained directly, thereby assisting the design of the wireless communication system.

For the collection of wireless channel samples, considering the difficulty and cost of collection, it is difficult to collect all scenes and features. Based on the collection of wireless channel samples, the traditional wireless channel modeling work can extract the statistical characteristics of the relevant transmission of a given channel on limited wireless channel samples, such as large-scale parameters and small-scale parameters. For example, multipath information, delay power spectral density, transmission emission angle or arrival angle, etc., and the channel can be modeled through this statistical feature. The channel modeling method can generate different wireless channel samples by introducing random terminal positions and corresponding small-scale parameters. However, with the development of wireless communication systems, the frequency band is gradually moving towards high frequency, and the scene is gradually moving towards more complex special environments such as space, space, earth and sea. The expansion of more scenarios such as special applications makes the wireless channel environment that the current wireless communication system needs to face more and more complex.

Under the above circumstances, it is very difficult to collect wireless channel samples, and the difficulties here include both technical difficulties and operational difficulties. At the same time, the mathematical modeling of the above-mentioned complex channels is also facing great challenges. The complexity of frequency bands, environments, and scenarios will lead to the complexity of channel modeling. Non-linear channel characteristics and difficult-to-fit channel propagation characteristics will affect the Traditional mathematical modeling to study channels brings difficulties and challenges. For example, high-frequency channel modeling is still an urgent problem to be solved. Furthermore, the difference between actual channel environment modeling and ideal channel modeling in complex scenarios and application environments will continue to increase with the complexity of the channel environment. It can be seen that in complex frequency bands, complex environments, and complex scenarios, it is a major challenge to obtain channel feature information through actual channel sampling and traditional mathematical modeling.

In order to facilitate the understanding of the embodiments of the present application, the Generative Adversarial Network (GAN) related to the present application will be described.

As the most promising branch of unsupervised learning methods in recent years, GAN has attracted widespread attention in the field of AI. The GAN model includes at least two modules, the Generator and the Discriminator, as shown in Figure 3. In GAN, the generator and the discriminator are trained simultaneously, where the generator simulates the distribution of real sample data, and the discriminator judges whether the input data comes from real samples or generated samples. The training process of the generator is to maximize the error probability of the discriminator, and the training process of the discriminator is to minimize the probability of judgment error under the premise of fixing the generator. Through the mutual game between the generator and the discriminator, under the stable convergence of the GAN model, the samples output by the generator can simulate real data samples. Based on the ideal GAN generator, it can generate fake pictures or voice data on the basis of real pictures or voice data, so as to achieve the effect of real ones. Furthermore, it is also possible to implement data supplementation to real samples of small samples to achieve larger-scale data sets.

Using GAN to realize the generation of virtual channel samples and data set expansion is an effective method for large-scale data set acquisition. However, this method also has significant disadvantages. Since the generation quality of virtual channel samples depends on the quality of the generator training in the GAN training process, and in the actual GAN training process, because the generator and the discriminator are in a mutual game state, it is relatively difficult to achieve a stable equilibrium convergence state. Difficult; at the same time, for the virtual channel samples generated by GAN, there is no explicit correspondence between its channel quality and the loss function of GAN.

In general, there may be two disadvantages of virtual channel samples generated based on GAN:

Therefore, currently, for AI-based wireless communication solutions, there is no direct virtual channel quality evaluation method to evaluate the performance of the virtual channel data set on a specific AI wireless communication task. It can only be determined through repeated deployment experiments, which further increases the actual deployment cost.

In summary, how to design and implement a quality evaluation method for virtual channel samples, and guide the training process of GAN, generate high-quality virtual channel samples for the construction of large-scale high-quality channel data sets to support artificial intelligence The research on integration with wireless communication system is a key problem to be solved urgently.

In order to facilitate understanding of the technical solutions of the embodiments of the present application, the technical solutions of the present application are described in detail below through specific examples. As optional solutions, the above related technologies may be combined with the technical solutions of the embodiments of the present application in any combination, and all of them belong to the protection scope of the embodiments of the present application. The embodiment of the present application includes at least part of the following contents.

FIG. 4 is a schematic interaction diagram of a method 200 for evaluating the quality of virtual channel samples according to an embodiment of the present application. As shown in FIG. 4 , the method 200 includes at least part of the following:

S210. The first device acquires first input information and at least one second input information, where the first input information is information about real channel samples, the second input information is information about virtual channel samples, and the at least one second input information corresponds to at least a generator model;

S220. According to the first input information and the at least one second input information, the first device evaluates the quality of the at least one second input information and/or determines a target generator model in the at least one generator model.

It should be understood that the technical solution of the present application can be used for the selection of the channel generator model during the offline training process, and can also be used for the selection of the channel generator during the online generation network training process. For example, when the network device side or the terminal device side uses the generation network to expand and generate the online data set, it is necessary to evaluate the quality of the generated virtual channel samples, then the generated virtual channel samples can be processed based on the technical solution of the embodiment of the present application. Quality assessment, further selecting a channel generator based on the assessment result.

In some embodiments of the present application, the first device may be a terminal device or a network device. That is, a terminal device or a network device may perform quality assessment on the virtual channel samples, and/or, target generator model selection.

In some embodiments of the present application, a channel may refer to a physical channel, for example, including a physical uplink channel and/or a physical downlink channel.

Optionally, the physical uplink channel may include but not limited to a physical uplink shared channel (Physical Uplink Shared Channel, PUSCH), and the physical downlink channel may include but not limited to a physical downlink shared channel (Physical Downlink Shared Channel, PDSCH).

In some embodiments of the present application, the first input information is information of real channel samples, for example, obtained by performing wireless channel acquisition, or obtained by measuring a reference signal.

It should be understood that, according to the difference between the sending end and the receiving end of the reference signal, the reference signal is also correspondingly different. For example, if the transmitting end of the reference signal is a terminal device and the receiving end of the reference signal is a network device, the reference signal may be a Sounding Reference Signal (SRS) or a Demodulation Reference Signal (DMRS). For another example, if the transmitting end of the reference signal is a network device and the receiving end of the reference signal is a terminal device, the reference signal may be a Synchronization Signal Block (SSB) or a Channel State Information Reference Signal (Channel State Information Reference Signal, CSI-RS) etc.

It should be understood that the embodiment of the present application does not limit the number of real channel samples included in the first input information, acquisition time and acquisition method, for example, the first input information may be acquired simultaneously, or may also be acquired in batches of.

In some embodiments of the present application, the first input information may include complete real channel sample information, or may also include partial real channel sample information, for example, the first input information is the information of the complete real channel sample cropped.

For example, the information of the real channel samples is the eigenvector or the eigenmatrix of the real channel samples, and the first input information may be the eigenvector or the eigenmatrix of the real channel samples, or it may be a subvector of the eigenvector of the real channel samples or the real A submatrix of the eigenmatrix for channel samples.

In some embodiments of the present application, the second input information is information of virtual channel samples. For example, the second input information is generated by a generator. Optionally, the generator may include, but is not limited to, a generator in a Generative Adversarial Network (GAN).

In some embodiments of the present application, the second input information may include complete virtual channel sample information, or may also include partial virtual channel sample information, for example, the second input information is the generated complete virtual channel sample information The information is obtained by clipping. For example, the information of the virtual channel sample is the eigenvector or the eigenmatrix of the virtual channel sample, and the second input information may be the eigenvector or the eigenmatrix of the virtual channel sample, or may also be a subvector or virtual A submatrix of the eigenmatrix for channel samples.

In some embodiments, the at least one second input information may correspond to at least one generator model, or there is a one-to-one correspondence between the at least one second input information and at least one generator model.

For example, the at least one second input information includes a plurality of second input information generated by different generator models. For each second input information, information of a plurality of virtual channel samples generated by the corresponding generator model may be included.

It should be understood that the embodiment of the present application does not limit the number of virtual channel samples included in each second input information, the acquisition time, and the acquisition manner. Optionally, the virtual channel samples included in the second input information may be generated once by the corresponding generator model, or may also be generated in batches by the corresponding generator model, as long as the samples corresponding to the same generator model The second input information may be generated by the same generator model, and the present application does not limit the generation method of each second input information.

Optionally, the number of virtual channel samples included in each piece of second input information may be the same, or may also be different.

In some embodiments, the at least one generator model may be a generator model of different rounds in the network training process.

That is, during the network training process, different rounds of generator models and virtual channel samples generated by the generator models may be saved for generating the at least one second input information.

Since the quality of the virtual channel samples in the second input information based on different generator models may vary, by evaluating the quality of the at least one second input information, the generator model corresponding to the virtual channel sample with the best quality is selected as The target generator model, and further, the expansion of the channel sample data set based on the target generator model is conducive to improving the quality of the channel sample data set.

In some embodiments of the present application, the real channel samples include but are not limited to at least one of the following:

Channel samples in time domain, channel samples in frequency domain, channel samples in angle domain.

In some embodiments of the present application, the virtual channel samples include but are not limited to at least one of the following:

It should be understood that, in this embodiment of the present application, the channel samples in the angle domain may also be referred to as: channel samples in the antenna domain, or channel samples in the space domain.

It should also be understood that in this embodiment of the present application, the real channel samples and the virtual channel samples are of the same type, for example, both are time-domain channel samples, or both are frequency-domain channel samples, or both are angle-domain channel samples, etc.

Hereinafter, specific implementations of the above three types of channel samples are described respectively, which are applicable to real channel samples and virtual channel samples.

In some embodiments of the present application, the information of time-domain channel samples includes but not limited to information of at least one of the following dimensions:

The first dimension, such as the dimension of the number of transmitting antennas or the number of transmitting ports;

The second dimension, such as the dimension of the number of receiving antennas or the number of receiving ports;

The third dimension, such as the time-domain granularity length dimension.

In some embodiments, the time domain granularity length may include but not limited to at least one of the following:

The real multipath number of the channel, the number of sampling points sampled in the time domain according to the first sampling rate, and the number of time units for time domain sampling.

Optionally, the first sampling rate may be predefined, or configured by the network device.

Optionally, the time unit may be any time unit, for example including but not limited to: time slot, symbol and so on.

It should be understood that in this embodiment of the present application, for real time-domain channel samples, resources in the first dimension may include all resources of real time-domain channel samples in the first dimension, or may include real time-domain channel samples in the first dimension. Some resources of the first dimension. Similarly, the same is true for the second and third dimensions. That is, the first input information may correspond to the eigenvector or eigenmatrix of the real channel samples, or may also correspond to a subvector of the eigenvector of the real channel samples, or a submatrix of the eigenmatrix of the real channel samples.

Similarly, for the virtual time-domain channel samples, the resources in the first dimension may be all resources of the virtual time-domain channel samples in the first dimension, or may include some resources of the virtual time-domain channel samples in the first dimension. Similarly, the same is true for the second and third dimensions. That is, the second input information may correspond to the eigenvector or eigenmatrix of the virtual channel samples, or may also correspond to a subvector of the eigenvector of the virtual channel samples, or a submatrix of the eigenmatrix of the virtual channel samples.

It should be understood that, in the embodiment of the present application, when the resources of the first input information in a certain dimension include some resources of the real channel samples in the dimension, it may be to crop the real channel samples in the dimension (or in other words , intercept, select) obtained. Optionally, the clipping may refer to continuous clipping, or may also be discrete clipping. That is, the resources of the first input information in this dimension may be continuous resources, or may also be discrete resources.

For example, when the channel is a physical downlink channel or a physical uplink channel, only continuous physical resource blocks (Resource Blocks, RBs) that carry PDSCH or PUSCH can be tailored, or only for real RBs that carry reference signals such as DMRS, CSI-RS, and SRS, etc. The RBs of the physical channel are clipped, or only the RBs of the real physical channel that do not carry the reference signal are clipped.

Correspondingly, when the real channel samples are clipped to obtain the first input information, the virtual channel samples can also be clipped in a similar manner to obtain the second input information, wherein the subvector of the virtual channel in the second input information or The sub-matrixes are in one-to-one correspondence with the physical resources occupied by the sub-vectors or sub-matrices of the real channels in the first input information.

For the convenience of distinction and description, in the embodiment of this application, the real channel sample is denoted as h, and the virtual channel sample is denoted as

Fig. 5 is a schematic diagram of time-domain channel sample information provided by an embodiment of the present application. As shown in FIG. 5 , the complete real time-domain channel samples have a length of M in the first dimension, a length of N in the second dimension, and a length of D in the third dimension.

Optionally, the length of the first input information in the first dimension may be less than or equal to M. FIG. 5 takes the length of the first input information in the first dimension as m as an example, where m<M, but this application does not Not limited to this.

That is, the first input information on the first dimension is obtained by intercepting the complete real channel samples on the first dimension.

Correspondingly, the second input information on the first dimension may also be obtained by intercepting information of the same length (that is, length m) from the complete virtual channel sample on the first dimension.

Optionally, the real channel samples and the virtual channel samples may also be intercepted in the second dimension and/or the third dimension to obtain the first input information and the second input information in the second dimension and/or the third dimension, The embodiment of the present application does not limit the dimension, method and size of the clipped channel samples. Of course, the real channel samples and the virtual channel samples may not be clipped. In this case, the first input information may be the eigenvector or the real channel sample. A feature matrix, the second input information may be a feature vector or a feature matrix of a virtual channel sample.

In some embodiments of the present application, the information of the channel samples in the frequency domain includes, but is not limited to, information of at least one of the following dimensions:

The first dimension, for example, the dimension of the number of transmit antennas or the number of transmit ports;

The second dimension, for example, the dimension of the number of receiving antennas or the number of receiving ports;

The third dimension, for example, frequency domain granularity length dimension;

The fourth dimension, for example, the time-domain granularity length dimension.

In some embodiments, for the specific implementation of the time-domain granularity length, refer to the relevant description of the time-domain granularity length in the time-domain channel samples, which will not be repeated here.

In some embodiments, the frequency domain granularity length may be any frequency domain unit length, for example including but not limited to at least one of the following: the number of subcarriers, the number of resource blocks, and the number of subbands. That is, one frequency domain unit of frequency domain channel samples may be one or more subcarriers, or one or more RBs, or one or more subbands.

In some embodiments, the first dimension and the second dimension of the channel samples in the frequency domain may form the combined dimension of the number of transceiver antennas or the pair of transceiver ports. The present application does not limit the specific representation of the channel samples in the frequency domain.

It should be understood that in this embodiment of the present application, for real frequency domain channel samples, the resources on the first dimension may include all resources of the real frequency domain channel samples in the first dimension, or may include real frequency domain channel samples in the first dimension. Some of the resources for the first dimension, and similarly for the second, third, and fourth dimensions. That is, the first input information may correspond to the eigenvector or eigenmatrix of the real channel samples, or may also correspond to a subvector of the eigenvector of the real channel samples, or a submatrix of the eigenmatrix of the real channel samples.

Similarly, for the virtual frequency domain channel samples, the resources on the first dimension may include all the resources of the virtual frequency domain channel samples in the first dimension, or may include some resources of the virtual frequency domain channel samples in the first dimension, Similarly, the same is true for the second, third and fourth dimensions. That is, the second input information may correspond to the eigenvector or eigenmatrix of the virtual channel samples, or may also correspond to a subvector of the eigenvector of the virtual channel samples, or a submatrix of the eigenmatrix of the virtual channel samples.

It should be understood that, in this embodiment of the present application, when the first input information includes part of the resources of frequency-domain channel samples in this dimension in a certain dimension, it may be obtained by tailoring the frequency-domain channel samples in this dimension, specifically For the clipping method, refer to the relevant descriptions of the foregoing embodiments, which will not be repeated here.

Fig. 6 is a schematic diagram of channel sample information in the frequency domain provided by an embodiment of the present application. As shown in Figure 6, the length of the complete real frequency domain channel sample is M in the first dimension, N in the second dimension, B in the third dimension, and T in the fourth dimension.

In the example in Figure 6, the first dimension and the second dimension of the frequency domain channel samples constitute the joint dimension of the transceiver antenna pair, and its dimension size is MN, and each block in Figure 6 represents the frequency domain channel coefficient on a resource unit .

Optionally, for the first input information and the second input information, tailoring or selection can be performed on the third dimension and the fourth dimension, for example, only the first b<B subcarriers are selected on the third dimension, and on the fourth dimension Only the last t<T symbols are selected.

Optionally, the b subcarriers may be continuous. For example, the b subcarriers may be consecutive b subcarriers carrying physical channels, or the b subcarriers may also be discrete, for example, the b subcarriers may be subcarriers occupied by physical channels carrying reference signals carrier, or a subcarrier occupied by a physical channel that does not carry a reference signal.

Optionally, the t symbols may be continuous, for example, the t symbols may be consecutive t symbols carrying a physical channel, or the t symbols may also be discrete, for example, the t The symbols may be symbols occupied by physical channels carrying reference signals, or symbols occupied by physical channels not carrying reference signals.

It should be understood that the physical resources occupied by the sub-vector or sub-matrix of the virtual channel in the second input information correspond one-to-one to the physical resources occupied by the sub-vector or sub-matrix of the real channel in the first input information.

In some embodiments of the present application, the channel samples in the angle domain are obtained through Fourier transform of channels in the antenna domain. For example, for the case where there are multiple antennas at the transmitter or receiver, such as Single-input Multi-output (SIMO), Multi-input Single-output (MISO) and MIMO (Multi-input Multi-output; MIMO) communication system can collect angle domain channel samples.

In some embodiments of the present application, the information of the channel samples in the angle domain includes information of at least one of the following dimensions:

a first dimension, such as the emission angle dimension;

A second dimension, such as the angle-of-arrival dimension;

The third dimension, for example, the time domain granularity length or the frequency domain granularity length dimension.

Optionally, when the transmitting end or the receiving end is a single antenna, the first dimension or the second dimension may not exist.

In some embodiments, the first dimension may contain one or two sub-dimensions, eg, a horizontal emission angle dimension and/or a vertical emission angle dimension.

In some embodiments, the second dimension may include one or two sub-dimensions, for example, the dimension of the angle of arrival in the horizontal direction and/or the dimension of the angle of arrival in the vertical direction. Among them, the angle of arrival in the horizontal direction can be obtained by Fourier transform of the sub-channel formed by the horizontal antenna array corresponding to the transmitting or receiving antenna, and the vertical angle of arrival can be obtained by Fourier transform on the sub-channel formed by the vertical antenna array corresponding to the transmitting or receiving antenna. Leaf transformation is obtained.

In some embodiments, the number of sub-dimensions included in the first dimension and the number of sub-dimensions included in the second dimension may be the same, or may also be different. For example, the first dimension may include the horizontal emission angle dimension and the vertical emission angle dimension, and the second dimension may include the horizontal arrival angle dimension or the vertical arrival angle dimension, or the first dimension may include the horizontal emission angle dimension or the vertical In the direction emission angle dimension, the second dimension may include a horizontal direction arrival angle dimension and a vertical direction arrival angle dimension.

In some embodiments, the angular coverage of the first dimension and the angular coverage of the second dimension may be the same or different.

For example, both the first dimension and the second dimension may be an angle coverage of (-90, 90) degrees.

For another example, the first dimension is the angular coverage range of (-90, 90) degrees, and the second dimension is the angular coverage range of (0, 180) degrees.

In some embodiments, the angular granularity of the first dimension may be configured uniformly, that is, the angle value represented by each angular granularity is a fixed value, such as 2 degrees, 5 degrees or 10 degrees, etc., or may be non-uniformly configured , for example, divide evenly by the sine or cosine of the angle.

In some embodiments, the angular granularity of the second dimension may be uniformly configured, that is, the angle value represented by each angular granularity is a fixed value, such as 2 degrees, 5 degrees or 10 degrees, etc., or may be non-uniformly configured , for example, divide evenly by the sine or cosine of the angle.

In some embodiments, the angle granularity or angle division method of the first dimension and the second dimension may be the same, or may also be different.

Optionally, the third dimension may be a time domain granularity length dimension or a frequency domain granularity length dimension, that is, the angle domain channel may be in the time domain or in the frequency domain. For the specific implementation of the granular length in the time domain and the granular length in the frequency domain, refer to the related descriptions in the foregoing embodiments, and details are not repeated here.

It should be understood that, in this embodiment of the present application, for real angle domain channel samples, resources in the first dimension may include all resources of real angle domain channel samples in the first dimension, or may include real angle domain channel samples in the first dimension. Some resources for the first dimension, similarly for the second and third dimensions. That is, the first input information may correspond to the eigenvector or eigenmatrix of the real channel samples, or may also correspond to a subvector of the eigenvector of the real channel samples, or a submatrix of the eigenmatrix of the real channel samples.

Similarly, for the channel samples in the virtual angle domain, the resources on the first dimension may include all the resources of the channel samples in the virtual angle domain in the first dimension, and may also include some resources in the first dimension of the channel samples in the virtual angle domain, Similarly for the second dimension and the third dimension and the fourth dimension. That is, the second input information may correspond to the eigenvector or eigenmatrix of the virtual channel samples, or may also correspond to a subvector of the eigenvector of the virtual channel samples, or a submatrix of the eigenmatrix of the virtual channel samples.

It should be understood that, in the embodiment of the present application, when the first input information includes part of the resources of the channel samples in the angle domain in a certain dimension, it may be obtained by tailoring the channel samples in the angle domain in this dimension, specifically For the clipping method, refer to the relevant descriptions of the foregoing embodiments, which will not be repeated here.

Fig. 7 is a schematic diagram of angle-domain channel sample information provided by an embodiment of the present application. As shown in Figure 7, the first dimension of the channel samples in the angle domain is the dimension of the horizontal emission angle, the second dimension is the dimension of the horizontal arrival angle, and the third dimension is the dimension of the time-domain granularity length, where the time-domain granularity length is the channel Take the real multipath delay number of , as an example.

As an example, as shown in Figure 7, the first dimension contains the horizontal emission angle of (-90, 90) degrees, and the second dimension contains the horizontal arrival angle of (-90, 90) degrees, and each square represents the angle granularity size, the angle granularity can be uniformly configured on the horizontal angle, that is, the angle value represented by each angle granularity is fixed, such as 2 degrees, 5 degrees, and 10 degrees, or it can also be non-uniformly configured on the horizontal angle, For example, it is obtained by uniformly dividing the sine or cosine of the angle.

Optionally, the angle division granularity and division method of the first dimension and the second dimension may be the same or different.

Optionally, the angular coverage ranges of the first dimension and the second dimension may be the same or different.

It should be understood that, in the example in FIG. 7 , the channel samples in the angle domain may be clipped on at least one of the first dimension, the second dimension, and the third dimension, and the present application does not limit the specific clipping manner.

In some embodiments of the present application, the method 200 further includes:

The first device acquires first configuration information, where the first configuration information is used to configure at least one of the following:

A method of generating the first input information;

A method of generating the at least one second input information;

A quality assessment parameter of the at least one second input information.

For example, the first input information may be obtained according to the first configuration information and the collected real channel samples, and/or the second input information may be obtained according to the first configuration information and the virtual channel samples generated by the generator.

For another example, perform quality assessment on at least one piece of second input information according to the first configuration information.

In some embodiments, the first configuration information includes at least one of the following:

The number parameter L1 for real channel sample sampling;

The number parameter L2 for virtual channel sample sampling;

The number parameter K of similar channel samples;

Dimension number information of channel samples;

Dimension information of channel samples;

Configuration information for generating channel samples in the frequency domain dimension;

Configuration information for generating channel samples in the time domain dimension;

Configuration information for generating channel samples of antenna domain dimension.

In some embodiments, the quantity parameters L1 and L2 may be equal or unequal.

Optionally, when the first configuration information includes a number parameter L, the first input information and the second input information use the same number parameter by default, that is, L1=L2=L.

Optionally, when the first configuration information does not include the number parameter, the default number parameter of the first input information is equal to all real channel samples included in the first input information, and the number parameter of the second input information is equal to all real channel samples included in the second input information. Virtual channel samples.

It should be understood that the present application does not limit the sampling method of virtual channel samples and real channel samples. The sampling method can be, for example, random sampling, uniform sampling, or continuous sampling, for example, taking the first L1 or L2 channel samples Or take the middle L1 or L2 channel samples, etc.

In some embodiments, the parameter K of the number of similar channel samples may be used to determine the K real channel samples with the highest similarity to the virtual channel samples when evaluating the quality of the virtual channel samples.

In some embodiments, the dimension information of the channel samples may be the dimension information of the foregoing examples.

For example, for time-domain channel samples, it may include but not limited to at least one of the following dimensions:

The third dimension, such as the time-domain granularity length dimension.

For another example, for frequency domain channel samples, it may include but not limited to at least one of the following dimensions:

For another example, for angle-domain channel samples, at least one of the following dimensions is included:

a first dimension, such as the emission angle dimension;

A second dimension, such as the angle-of-arrival dimension;

In some embodiments, the configuration information for generating channel samples in the frequency domain dimension includes at least one of the following:

Frequency domain granularity configuration, resource tailoring method of frequency domain dimension, indication information of target frequency domain resources.

In some embodiments, frequency domain granularity configuration may be used to configure frequency domain granularity (or frequency domain unit) of channel samples, for example, it may be subcarrier granularity, resource bank granularity, or subband granularity, etc. As an example, the frequency domain granularity configuration can be 2 bits, the 2 bits are 00, indicating that the frequency domain granularity is one subcarrier, the 2 bits are 01, indicating that the frequency domain granularity is one RB, and the 2 bits are 10, indicating that the frequency domain granularity is one Subband.

Optionally, the number of sub-frequency domain units included in a frequency domain unit may be fixed or predefined, or may also be configured through frequency domain granularity configuration. For example, when the frequency domain granularity is configured as one subband, the number of RBs included in one subband may be further configured. In this case, the frequency domain granularity configuration may include 3 bits, the first 2 bits are used to indicate the frequency domain granularity or the frequency domain unit, and the last bit is used to indicate the number of sub-frequency domain units included in the frequency domain unit. For example, when the value of these 3 bits is 100, it means that one subband includes 4 resource blocks; when the value of these 3 bits is 101, it means that one subband includes 8 resource blocks.

In some embodiments, the resource tailoring manner of the frequency domain dimension may include a continuous frequency domain resource tailoring manner and a discrete frequency domain resource tailoring manner. Optionally, the resource tailoring mode of the frequency domain dimension may be 1 bit, for example, if the 1 bit is 0, it indicates a continuous frequency domain resource tailoring mode, and if it is 1, it indicates a discrete frequency domain resource tailoring mode.

In some embodiments, when the frequency domain dimension of the channel samples needs to be tailored, the configuration information for generating the channel samples of the frequency domain dimension further includes indication information of target frequency domain resources to be tailored.

Optionally, when the resource tailoring mode of the frequency domain dimension is a continuous frequency domain resource tailoring mode, the indication information of the target frequency domain resource may include the index and length of the starting frequency domain resource (that is, the starting clipping point) (that is, the clipping point length).

Optionally, when the resource tailoring method of the frequency domain dimension is a discrete frequency domain resource tailoring method, the indication information of the target frequency domain resource may include the index of the starting frequency domain resource (that is, the starting clipping point) and the frequency domain resource interval (that is, the clipping interval), wherein the frequency domain resource interval is the interval between two adjacent target frequency domain resources.

In some embodiments, the configuration information for generating channel samples of the time domain dimension includes at least one of the following:

Time-domain granularity configuration, resource clipping mode of time-domain dimension, and indication information of target time-domain resources.

In some embodiments, the time-domain granularity configuration may be used to configure the time-domain granularity (or in other words, time-domain units) of channel samples. For example, the time domain granularity may be one or more slot granularities, or one or more symbol granularities, and so on.

As an example, the time domain granularity configuration can be 2 bits, the 2 bits are 00, indicating that the time domain granularity is one slot, the 2 bits are 01, indicating that the time domain granularity is one symbol, and the 2 bits are 10, indicating that the time domain granularity is multiple time slots, where 2 he is 11 means that the time domain granularity is multiple symbols.

In some embodiments, the resource tailoring manner of the time domain dimension may include a continuous time domain resource tailoring manner and a discrete time domain resource tailoring manner. Optionally, the resource tailoring mode of the time domain dimension may be 1 bit, for example, if the 1 bit is 0, it indicates a continuous frequency domain resource tailoring mode, and if it is 1, it indicates a discrete frequency domain resource tailoring mode.

In some embodiments, when the time-domain dimension of the channel samples needs to be tailored, the configuration information for generating the channel samples of the time-domain dimension further includes indication information of target time-domain resources to be tailored.

Optionally, when the resource clipping mode of the time domain dimension is a continuous time domain resource clipping mode, the indication information of the target time domain resource may include the index and length (that is, the clipping point) of the starting time domain resource (that is, the starting clipping point). length).

Optionally, when the resource clipping mode of the time domain dimension is a discrete time domain resource clipping mode, the indication information of the target time domain resource may include the index of the starting time domain resource (that is, the starting clipping point) and the time domain resource interval (that is, the clipping interval), wherein the time-domain resource interval is the interval between two adjacent target time-domain resources.

In some embodiments, the configuration information for generating channel samples of antenna domain dimensions includes at least one of the following:

Antenna domain granularity configuration, resource tailoring mode of antenna domain dimension, indication information of target antenna domain resources.

It should be understood that, in the embodiment of the present application, the antenna domain may also be replaced with an angle domain or a space domain.

In some embodiments, the resource tailoring manner of the antenna domain dimension may include a continuous antenna domain resource tailoring manner and a discrete antenna domain resource tailoring manner. Optionally, the resource tailoring mode of the antenna domain dimension may be 1 bit, for example, if the 1 bit is 0, it indicates a continuous frequency domain resource tailoring mode, and if it is 1, it indicates a discrete frequency domain resource tailoring mode.

In some embodiments, when the antenna domain dimension of the channel sample needs to be tailored, the configuration information for generating the channel sample of the antenna domain dimension further includes indication information of the tailored antenna domain resource.

Optionally, when the resource tailoring mode of the antenna domain dimension is a continuous antenna domain resource tailoring mode, the indication information of the target antenna domain resource may include the index and length of the starting antenna domain resource (that is, the starting clipping point) (that is, the clipping length).

Optionally, when the resource tailoring mode of the antenna domain dimension is a discrete antenna domain resource tailoring mode, the indication information of the target antenna domain resource may include the index of the initial antenna domain resource (that is, the initial tailoring point) and the antenna domain resource interval (that is, the clipping interval), wherein the antenna domain resource interval is the interval between two adjacent target antenna domain resources.

Hereinafter, manners for acquiring the first input information and the at least one second input information will be described in conjunction with specific embodiments.

Example 1: first input information

Embodiment 1-1: the first input information is obtained according to real channel samples collected by the first device.

For example, the first device may collect or measure a reference signal to obtain real channel samples, and further obtain first input information based on the first configuration information.

Embodiment 1-2: The first input information may be acquired by the first device from the second device.

For example, the second device may collect or measure a reference signal to obtain real channel samples, and further obtain the first input information based on the first configuration information.

In some embodiments, the second device may send the first input information to the first device based on the indication of the first device.

For example, the first device sends first indication information to the second device, where the first indication information is used to instruct the second device to send the first input information to the first device. The second device may send the first input information to the first device after receiving the first indication information.

In other embodiments, the second device may send the collected real channel samples to the first device, and the first device further obtains first input information according to the first configuration information and the real channel samples.

The first device sends first configuration information to the second device. Further, the second device may generate the first input information according to the first configuration information and real channel samples.

Embodiment 2: the at least one second input information

Embodiment 2-1: the at least one second input information is obtained by the first device according to the virtual channel samples generated by the generator deployed on the first device.

For example, the first device may generate at least one set of virtual channel samples according to at least one generator model, and further obtain the at least one second input information based on the first configuration information and the at least one set of virtual channel samples.

Embodiment 2-2: the at least one second input information is acquired by the first device from the second device.

For example, the second device may generate at least one set of virtual channel samples according to at least one generator model, and further obtain the at least one second input information based on the first configuration information and the at least one set of virtual channel samples.

In some embodiments, the second device may send the at least one second input information to the first device based on the instruction of the first device.

For example, the first device sends second indication information to the second device, where the second indication information is used to instruct the second device to send the at least one piece of second input information to the first device. The second device may send the at least one second input information to the first device in a case of receiving the second indication information.

The first device receives the at least one second input information sent by the second device and generator indexes respectively corresponding to the at least one second input information.

The first device sends first configuration information to the second device. Further, the second device may generate second input information according to the first configuration information and the virtual channel samples.

In some embodiments, the method 200 also includes:

The first device sends third indication information to the second device, where the third indication information is used to indicate the target generator index determined by the first device.

In some embodiments, the first device is a terminal device, and the second device is a network device.

In other embodiments, the first device is a network device, and the second device is a terminal device.

The method for evaluating the quality of a virtual channel according to an embodiment of the present application will be described below in combination with specific embodiments.

Case 1: Both the first input information and the second input information are generated by the network device.

For example, the network device may generate the first input information according to real channel samples, where the real channel samples may be obtained by collecting or measuring a reference signal by the network device.

For example, the generator is deployed on the network device, and the network device can generate at least one set of virtual channel samples through at least one generator model, and further obtain at least one second input information based on the at least one set of virtual channel samples.

Specifically, for example, the network device may sequentially save generator models of different rounds and corresponding virtual channel samples during the generator training process. The at least one second input information is further generated based on virtual channel samples generated by different generator models.

In this case 1, the network device may evaluate the quality of the at least one second input information according to the first input information and the at least one second input information. Further, the target generator model may be selected according to the quality of the at least one second input information. For example, the generator model corresponding to the second input information with the best quality is selected as the target generator model. Further, virtual channel samples can be generated based on the target generator model for augmentation of the channel sample data set.

Case 2: Both the first input information and the second input information are generated by the terminal device.

For example, the terminal device may generate the first input information according to real channel samples, where the real channel samples may be obtained by the terminal device collecting or measuring a reference signal.

For example, the generator is deployed on the terminal device, and the terminal device can generate at least one set of virtual channel samples through at least one generator model, and further obtain at least one second input information based on the at least one set of virtual channel samples.

Specifically, for example, the terminal device may sequentially save generator models of different rounds and corresponding virtual channel samples during the generator training process. The at least one second input information is further generated based on virtual channel samples generated by different generator models.

In this case 2, the terminal device may evaluate the quality of the at least one second input information according to the first input information and the at least one second input information. Further, the target generator model may be selected according to the quality of the at least one second input information. For example, the generator model corresponding to the second input information with the best quality is selected as the target generator model. Further, virtual channel samples can be generated based on the target generator model for augmentation of the channel sample data set.

Case 3: the first input information is generated by a network device, and the at least one second input information is generated by a terminal device.

For the specific ways of generating the first input information and the second input information, refer to the relevant descriptions of case 1 and case 2, which will not be repeated here.

Case 3-1: Quality assessment of virtual channel samples is performed by network equipment.

In this case, the network device needs to acquire the at least one piece of second input information from the terminal device.

For example, a specific signaling interaction process may be as shown in FIG. 8 .

S231. The network device sends first configuration information to the terminal device, so that the terminal device generates the at least one piece of second input information.

Optionally, in some embodiments, the terminal device may also directly send the generated virtual channel samples to the network device, and in this case, the first configuration information may not be required. Alternatively, the terminal device may also generate the at least one piece of second input information based on the second configuration information. The second configuration information may be default configuration information.

Optionally, for the parameters included in the second configuration information, refer to the parameters included in the first configuration information in the foregoing embodiments, and details are not repeated here for brevity. It should be understood that the parameters included in the second configuration information and the first configuration information may be the same or different, and/or the values of the parameters included in the second configuration information and the first configuration information may be the same, or, It can also be different.

Optionally, when the first configuration information and the second configuration information include different parameters, and/or include different values of the same parameters, the network device sends the first configuration information to the terminal device.

Optionally, the first configuration information may be carried by any downlink signaling, such as radio resource control (Radio Resource Control, RRC) signaling, media access control (Media Access Control, MAC) signaling or downlink control information (Downlink Control Information, DCI) and so on.

Optionally, the first configuration information may be carried in dedicated downlink signaling for supporting AI capabilities.

S232. The network device sends second indication information to the terminal device, where the second indication information is used to instruct the terminal device to send the at least one piece of second input information to the network device.

Optionally, the second indication information may be carried by any downlink signaling, such as RRC signaling, MAC signaling, or DCI.

Optionally, the second indication information may be carried in dedicated downlink signaling for supporting AI capability.

It should be understood that the first configuration information and the second indication information may be sent through the same signaling, or may also be sent through different signaling. When the first configuration information and the second indication information are sent through different signaling, this application does not A specific sequence is not limited.

S233. The terminal device sends at least one piece of second input information and an index of the generator model corresponding to the at least one piece of second input information to the network device.

S234. The network device evaluates the quality of the virtual channel sample based on the first input information and at least one second input information.

Further, the target generator model may be selected according to the quality evaluation result of the virtual channel sample corresponding to the at least one second input information, for example, the generator model corresponding to the virtual channel sample with the best quality is selected as the target generator model.

S235. The network device sends third indication information to the terminal device, where the third indication information is used to indicate the index of the target generator model selected by the network device.

Case 3-2: The quality assessment of the virtual channel samples is performed by the terminal device.

In this case, the terminal device needs to obtain the first input information from the network device.

For example, a specific signaling interaction process may be as shown in FIG. 9 .

S241. The terminal device sends first configuration information to the network device, so that the network device generates the first input information.

Optionally, in some embodiments, the network device may also directly send the obtained real channel samples to the terminal device. In this case, the first configuration information may not be required. Alternatively, the network device may also generate the first input information based on the second configuration information. The second configuration information may be default configuration information.

Optionally, for the parameters included in the second configuration information, refer to the parameters included in the first configuration information in the foregoing embodiments, and details are not repeated here for brevity.

It should be understood that the parameters included in the second configuration information and the first configuration information may be the same or different, and/or the values of the parameters included in the second configuration information and the first configuration information may be the same, or, It can also be different.

Optionally, when the first configuration information and the second configuration information include different parameters, and/or include different values of the same parameters, the terminal device sends the first configuration information to the network device.

Optionally, in this case 3-2, the first configuration information may also be configured by the network device for the terminal device, as long as the network device and the terminal device have the same understanding of how the first input information and the second input information are generated. Yes, the present application does not limit the specific configuration manner of the first configuration information.

Optionally, the first configuration information may be carried by any uplink signaling, such as RRC signaling or MAC signaling.

Optionally, the first configuration information may be carried by dedicated uplink signaling for supporting AI capabilities.

S242. The terminal device sends first indication information to the network device, where the first indication information is used to instruct the network device to send the first input information to the terminal device.

Optionally, the first indication information may be carried by any uplink signaling, such as RRC signaling, MAC signaling and so on.

Optionally, the first indication information may be carried by dedicated uplink signaling for supporting AI capabilities.

It should be understood that the first configuration information and the first indication information may be sent through the same signaling, or may also be sent through different signaling. When the first configuration information and the first indication information are sent through different signaling, this application does not A specific sequence is not limited.

S243. The network device sends the first input information to the terminal device.

S244. The terminal device evaluates the quality of the virtual channel sample based on the first input information and at least one second input information.

Case 4: the first input information is generated by a terminal device, and the at least one second input information is generated by a network device.

Case 4-1: Quality assessment of virtual channel samples is performed by network equipment.

In this case, the network device needs to obtain the first input information from the terminal device.

For example, a specific signaling interaction process may be as shown in FIG. 10 .

S251. The network device sends first configuration information to the terminal device, so that the terminal device generates the first input information.

Optionally, in some embodiments, the terminal device may also directly send the obtained real channel samples to the network device, and in this case, the first configuration information may not be required. Alternatively, the terminal device may also generate the first input information based on the second configuration information. The second configuration information may be default configuration information.

Optionally, the first configuration information may be carried by any downlink signaling, such as RRC signaling or MAC signaling.

S252. The network device sends first indication information to the terminal device, where the first indication information is used to instruct the terminal device to send the first input information to the network device.

Optionally, the first indication information may be carried by any downlink signaling, such as RRC signaling, MAC signaling and so on.

Optionally, the first indication information may be carried in dedicated downlink signaling for supporting AI capabilities.

S253. The terminal device sends the first input information to the network device.

S254. The network device evaluates the quality of the virtual channel sample based on the first input information and at least one second input information.

Case 4-2: The quality assessment of the virtual channel samples is performed by the terminal device.

In this case, the terminal device needs to acquire the at least one piece of second input information from the network device.

For example, a specific signaling interaction process may be as shown in FIG. 11 .

S261. The terminal device sends first configuration information to the network device, so that the network device generates the at least one piece of second input information.

Optionally, in some embodiments, the network device may also directly send the generated virtual channel samples to the terminal device. In this case, the first configuration information may not be required. Alternatively, the network device may also generate the at least one piece of input information based on the second configuration information. The second configuration information may be default configuration information.

Optionally, when the first configuration information and the second configuration information include different parameters, and/or include different values of the same parameters, the terminal device sends the first configuration information to the network device. Optionally, in this case 4-2, the first configuration information may also be configured by the network device for the terminal device, as long as the network device and the terminal device have the same understanding of how the first input information and the second input information are generated. Yes, the present application does not limit the specific configuration manner of the first configuration information.

Optionally, the first configuration information may be carried by any uplink signaling, such as RRC signaling, MAC signaling and so on.

S262. The terminal device sends second indication information to the network device, where the second indication information is used to instruct the network device to send the at least one second input information to the terminal device.

Optionally, the second indication information may be carried by any uplink signaling, such as RRC signaling or MAC signaling.

Optionally, the second indication information may be carried by dedicated uplink signaling used to support the AI capability.

S263. The network device sends at least one piece of second input information and an index of a generator model corresponding to the at least one piece of second input information to the terminal device.

S264. The terminal device evaluates the quality of the virtual channel sample based on the first input information and at least one second input information.

S265. The terminal device sends third indication information to the network device, where the third indication information is used to indicate the index of the target generator model selected by the network device.

It should be understood that the embodiment of the present application does not limit the specific manner in which the network device or the terminal device evaluates the quality of the virtual channel sample according to the first input information and the second input information. Hereinafter, the method for evaluating the quality of virtual channel samples provided by the present application will be described in conjunction with specific embodiments.

In some embodiments of the present application, S220 may include:

According to the first input information and each second input information in the at least one second input information, determine the quality evaluation information corresponding to each second input information, and determine the quality assessment information corresponding to each second input information The quality assessment information is used to indicate the similarity between the real channel samples and the virtual channel samples and/or the dispersion of the virtual channel samples.

In some embodiments, the quality evaluation information of the second input information includes a first index and a second index, and the first index is used to indicate the similarity between the virtual channel sample and the real channel sample, that is, the first index is similar A degree index, the second index is used to indicate the dispersion degree of the virtual channel samples, and the second index is a dispersion degree index. Among them, the higher the first index, the closer the virtual channel sample is to the real channel sample, and the lower the second index, the better the diversity of the virtual channel sample, that is, the higher the dispersion. Based on the virtual channel sample expansion data set, there is It is beneficial to improve the generalization performance of the network.

In some embodiments, the quality evaluation information of the second input information includes a third index, and the third index is generated according to the first index and the second index. For example, the third index is the ratio of the first index to the second index. The larger the third index, the higher the generation quality of the virtual channel.

In some embodiments, the first device generates quality evaluation information of the second input information according to the first input information and the second input information, including:

The quality evaluation information of the second input information is generated according to the first input information, the second input information and the third input information.

Optionally, the third input information may be an evaluation parameter of the virtual channel sample, for example, the third input information may include the evaluation parameter of the virtual channel sample in the first configuration information.

In some embodiments, the third input information includes at least one of the following information:

The number parameter L1 for real channel sample sampling;

The number parameter L2 for virtual channel sample sampling;

The maximum number of similar channel samples parameter K.

For example, the first device may determine a target similarity between virtual channel samples and real channel samples and/or a target dispersion of virtual channel samples according to the first input information and the second input information.

In some embodiments of the present application, the first device determines the target similarity between virtual channel samples and real channel samples according to the first input information and the second input information, including the following steps:

Step 1: The first device may sample the first input information and the second input information according to the third input information to obtain a real channel sample set and a virtual channel sample set.

For example, L1 real channel samples are extracted from the first input information according to the quantity parameter L1 to form a real channel sample set, and L2 virtual channel samples are extracted from the second input information according to the quantity parameter L2 to form a virtual channel sample set.

Step 2: Determine the K target real channel samples corresponding to each virtual channel sample in the virtual channel sample set (that is, K-similarity), wherein the K target real channel samples corresponding to each virtual channel sample are the K real channel samples with the highest similarity to the virtual channel samples in the set of real channel samples.

For example, for the i-th virtual channel sample in the set of virtual channel samples

Determine L1 real channel samples and virtual channel samples in the real channel sample set

The K target real channel samples with the highest similarity.

Step 3: According to the K target real channel samples corresponding to each virtual channel sample, determine the similarity information between each virtual channel sample and the real channel sample.

For example, the virtual channel sample

The average value of the similarity with the corresponding K target real channel samples is used as the virtual channel sample

Similarity to real channel samples.

Step 4: Determine the target similarity between the virtual channel sample and the real channel sample according to the similarity information between each virtual channel sample and the real channel sample in the virtual channel sample set.

For example, the average value of the similarity information of each virtual channel sample and real channel sample in the virtual channel sample set is used as the target similarity between the virtual channel sample and the real channel sample.

In some embodiments, the similarity between the real channel samples and the virtual channel samples can be measured according to the distance between the real channel samples and the virtual channel samples. For the convenience of representation, before determining the similarity between the real channel samples and the virtual channel samples, the dimensions of the real channel samples and the virtual channel samples have been vectorized, that is, h represents the vector of the real channel,

A vector representing a virtual channel. For example, the distance function of real channel samples and virtual channel samples can be defined as

In some embodiments, cosine similarity can be used to determine the similarity between real channel samples and virtual channel samples.

As an example, the distance function of real channel samples and virtual channel samples

It can be calculated using the following formula:

Among them, the greater the cosine similarity, the real channel sample h of the first input information and the virtual channel sample h of the second input information

The higher the degree of similarity.

In other embodiments, the distance function

It can also be measured by the vector x-norm, for example, as shown in the following formula:

Wherein, x can take the norm of 0, 1, 2, etc., which is not limited in this application. At this time, the smaller the norm value is, the real channel sample h of the first input information and the virtual channel sample h of the second input information

The higher the degree of similarity.

It should be understood that the above distance measurement method between the virtual channel sample and the real channel sample is only an example, and other distance measurement methods may also be used in the present application, and the present application is not limited thereto.

Optionally, taking the cosine similarity distance to measure the similarity between virtual channel samples and real channel samples as an example, the cosine similarity index K _sim of virtual channel samples and real channel samples is defined as follows:

That is, the cosine similarity index K _sim may be an implementation of the aforementioned first index, and of course other similarity indexes may also be used to represent the similarity between virtual channel samples and real channel samples, and the present application is not limited thereto.

Specifically, for the i-th virtual channel sample in the set of virtual channel samples

Among the L1 real channel samples in the set of real channel samples, find the same

The K target real channel samples with the highest similarity, and the average of the similarities between all virtual channel samples and their corresponding target real channel samples is taken as the target similarity K _sim between the virtual channel samples and the real channel samples.

K-similarity measures the similarity between real channel samples and virtual channel samples. The higher the K-similarity, the closer the distribution of generated virtual channel samples is to the distribution of real channel samples, indicating the quality of the generated virtual channel samples. higher.

In order to further ensure the diversity of the generated virtual channel samples and avoid the problem that when the virtual channel samples are used as a data set to train the network, the network will overfit on the data set and affect the generalization of the network. The dispersion of the virtual channel samples can be further evaluated.

In some embodiments of the present application, the determining the target dispersion of virtual channel samples according to the first input information and the second input information includes the following steps:

Step 1: Sampling the first input information and the second input information according to the third input information to obtain a real channel sample set and a virtual channel sample set.

Step 2: Determine K target real channel samples corresponding to each virtual channel sample in the virtual channel sample set, wherein the K target real channel samples corresponding to each virtual channel sample are the real channel sample set The K real channel samples with the highest similarity with the virtual channel samples.

The K target real channel samples with the highest similarity.

Step 3: Determine the number of target virtual channel samples corresponding to each real channel sample, wherein the real channel samples belong to K target real channel samples corresponding to the corresponding target virtual channel samples.

For example, a counter is maintained for each real channel sample in the real channel sample set, or, for the real channel sample set including L1 real channel samples, a sequence I of length L1 is maintained, and the sequence I includes L1 counts value, corresponding to the L1 real channel samples, each count value is used to record the number of times the corresponding real channel sample is selected as the target real channel sample based on the K-similarity, or in other words, satisfy the K-similarity with the real channel sample The number of virtual channel samples required by the degree. For example, in step 2, for each virtual channel sample in the virtual channel sample set, K target real channel samples can be found, and at this time, add one to the corresponding count value of the target real channel sample in the sequence I.

Step 4: Determine the target dispersion of the virtual channel samples according to the number of target virtual channel samples corresponding to each real channel sample.

For example, the target dispersion of the virtual channel samples is determined according to the standard deviation of the number of target virtual channel samples corresponding to each real channel sample.

That is, the standard deviation K _dis can be calculated for the L1 count values in the sequence I, and the target dispersion of the virtual channel samples can be represented by the standard deviation index, as shown in the following formula:

That is, the standard deviation index K _sim may be an implementation of the aforementioned second index, and of course other indexes may also be used to represent the dispersion of the virtual channel samples, and the present application is not limited thereto.

It should be understood that in the embodiment of the present application, when the standard deviation is used to represent the dispersion of virtual channel samples, the higher the value of the standard deviation, the lower the dispersion of the virtual channel samples, and the lower the value of the standard deviation, it means that the virtual channel The higher the dispersion of the sample is.

Further, the target generator model in the at least one generator model may be determined according to the target similarity between the virtual channel samples and the real channel samples and the target dispersion of the virtual channel samples.

For example, among the virtual channel samples with the highest target similarity (for example, the first index has the highest value), select the generator model corresponding to the virtual channel sample with the highest target dispersion (for example, the second index has the lowest value) as the target generation device model.

For another example, among the virtual channel samples with the highest target dispersion (for example, the value of the second index is the lowest), select the generator model corresponding to the virtual channel sample with the highest target similarity (for example, the highest value of the first index) as the target generator model.

For another example, the generator model corresponding to the virtual channel sample with the largest ratio of the target similarity to the target dispersion (for example, the value of the third index is the largest) is selected as the target generator model.

Therefore, in the embodiment of the present application, the quality evaluation of the virtual channel samples is performed based on the similarity between the virtual channel samples and the real channel samples and the dispersion of the virtual channel samples, which is beneficial to take into account the similarity between the virtual channel samples and the real channel samples, and The diversity of virtual channel samples can effectively expand the data set and ensure the generalization performance of deep neural network training under AI tasks.

Moreover, in the embodiment of the present application, flexible input formats of virtual channel samples and real channel samples can be supported, specifically, before performing quality evaluation on virtual channel samples, for any input format of real channel samples and virtual channel samples, Both can generate first input information and second input information in a unified format based on the first configuration information. At the same time, when the virtual channel samples and the real channel samples are not on the same device, a dedicated signaling interaction method is designed, which can support online virtual channel quality evaluation and generator model selection, which is conducive to meeting the model update and migration in the communication system Data set expansion requirements such as learning.

Fig. 12 is a schematic block diagram of a method for evaluating the quality of virtual channel samples according to an embodiment of the present application. Wherein, the quality evaluation method may be executed by an evaluator on the first device, and the evaluator may be implemented as a processor.

Specifically, the generator may generate virtual channel samples based on the input information of the generator, and further, generate second input information based on the virtual channel samples. In this step, a plurality of second input information can be obtained based on virtual channel samples generated by different generator models. For example, the corresponding virtual channel samples are generated based on the generator model of different rounds in the network training process.

Further, the first input information, the second input information and the third input information may be input into the evaluator to obtain the quality evaluation information of the second input information, for example, the aforementioned first index and second index, or the third indicators etc.

In some embodiments, when the virtual channel samples generated by GAN or traditional channel modeling are used for data set expansion, the technical solution of the present application can be used to evaluate the quality of the generated virtual channel samples. For example, evaluate the generation quality of virtual channel samples from two perspectives of K-similarity and K-dispersion, and support flexible input dimensions and methods of real channel samples and virtual channel samples.

Taking AI-based CSI eigenvector feedback as an example, 4000 real CDL-C300 channel samples are used as input. When the 4000 real channel samples are directly used in the training set, the number of samples in the training set is too small, and the neural network is trained in this way. It is easy to produce overfitting on the set, and the training effect is poor, so it is necessary to generate virtual channel samples to expand the training set.

In the following, 20,000 virtual channel samples are respectively generated for the two methods of generating virtual channel samples as an example, and the quality of the generated virtual channel samples is compared. The comparison results are shown in Table 1.

Table 1

Wherein, mode 1 corresponds to the virtual channel samples generated by the traditional virtual channel generation mode, and mode 2 corresponds to the virtual channel samples generated based on the quality assessment method of the embodiment of the present application. Wherein, in mode 2, K=1, that is, 1-similarity and 1-dispersion. The third and fourth columns of the index in Table 1 indicate the recovery degree of the CSI feature vector, and the larger the value, the better the recovery performance of the CSI feature vector.

It can be seen from Table 1 that the virtual channel samples generated based on the method 2 are better than the virtual channel samples generated based on the method 1 in terms of similarity index. Moreover, the virtual channel samples generated based on the method 2 are better than the virtual channel samples generated based on the method 1 in terms of dispersion index. Therefore, when the virtual channel samples generated by the two methods are used as the training set to train the CSI feedback model respectively and applied to the real channel samples, the CSI model obtained based on method 2 is also better than the CSI model obtained based on method 1 (0.784> 0.762).

And when the trained CSI model is used to test two groups of virtual channel samples, from the third column index in Table 1, we can see the restoration of the virtual channel samples generated based on method 2 relative to the virtual channel samples generated based on method 1 Higher performance. That is, the virtual channel samples generated based on method 2 are closer to the real channel samples (0.807>0.699).

Therefore, the quality assessment method of virtual channel samples based on K-similarity and K-dispersion can select better virtual channel samples and use them for data set expansion, thereby improving the accuracy of data sets under AI-based deep learning tasks. Sufficient problem to improve deep neural network generalization.

In some embodiments, the quality evaluation index based on K-similarity and K-dispersion is used to guide the selection of the generator model, which is beneficial to improve the quality of the virtual channel generated by the generator.

Continuing the above ideas, build a class of convolutional channel generators for the generation of virtual channel samples. In the training process, two generator models are included, and the above indicators are used to test respectively. The comparison results of the indicators of the two generators are shown in Table 2.

Table 1

It can be seen from Table 2 that the virtual channel samples generated by generator model 2 are better than those generated by generator model 1 in terms of similarity index. Therefore, when using the trained CSI model to test two groups of virtual channel samples, it can be seen from the last column of Table 2 that the generator model 2 is better than the generator model 1 in this indicator, that is, the generator model 2 generated The virtual channel samples of are closer to the real channel samples (0.807>0.802).

At the same time, it can be seen from Table 2 that the virtual channel samples generated by generator model 2 are better than those generated by generator model 1 in terms of dispersion index, so the virtual channel samples generated by these two generator models are When the CSI feedback model is trained separately as a training set and applied to real channel samples, the generalization of the network model obtained by generator model 2 is also better than generator model 1 (0.784>0.760).

Therefore, the method for evaluating the quality of virtual channel samples based on the embodiment of the present application can generate high-quality virtual channel samples, which is conducive to expanding the scale of the data set and ensuring the training effect and generalization performance of the model.

It should be understood that, in this embodiment of the present application, the channel quality assessment method may also be extended to the generation and evaluation of single-stream feature vectors and multi-stream feature vectors extracted from channels. At this time, the real feature vector samples can be used as the first input information, and the virtual feature vector samples can be used as the second input information, and the technical solution of the present application is also applicable.

FIG. 13 is a schematic diagram of a method for evaluating the quality of virtual channel samples according to other embodiments of the present application. As shown in FIG. 13 , the method 300 may include at least part of the following:

S310. The first device determines quality evaluation information of the second input information according to the first input information and the second input information;

Wherein, the first input information is information of real channel samples, the second input information is information of virtual channel samples, and the quality evaluation information of the second input information is used to indicate the similarity between virtual channel samples and real channel samples degree and/or dispersion of virtual channel samples.

It should be understood that the specific implementation of the quality assessment method in the method 300 is similar to the quality assessment method in the method 200, for detailed description refer to the method 200, and for the sake of brevity, details are not repeated here.

In some embodiments, the first device is a terminal device or a network device.

In some embodiments of the present application, the real channel samples include at least one of the following:

In some embodiments of the present application, the virtual channel samples include at least one of the following:

The information of the time-domain channel samples includes information of at least one of the following dimensions:

The number of transmit antennas or transmit ports; the number of receive antennas or receive ports; the length of time domain granularity.

In some embodiments of the present application, the information of the channel samples in the frequency domain includes information of at least one of the following dimensions:

The number of transmit antennas or transmit ports; the number of receive antennas or receive ports; the length of frequency domain granularity; the length of time domain granularity.

Emission angle; Arrival angle; Time domain granularity length or Frequency domain granularity length.

In some embodiments of the present application, the quality evaluation information of the second input information includes a first index and a second index, the first index is used to indicate the similarity between the virtual channel sample and the real channel sample, and the second The indicator is used to indicate the dispersion of the virtual channel samples; or, the quality assessment information of the second input information includes a third indicator, and the third indicator is generated according to the first indicator and the second indicator.

In some embodiments of the present application, the first device generates quality evaluation information of the second input information according to the first input information and the second input information, including:

According to the first input information and the second input information, a target similarity between virtual channel samples and real channel samples and/or a target dispersion of virtual channel samples are determined.

In some embodiments of the present application, the determining the target similarity between virtual channel samples and real channel samples and/or the target dispersion of virtual channel samples according to the first input information and the second input information includes:

According to the first input information, the second input information and the third input information, determine the target similarity between the virtual channel samples and the real channel samples and/or the target dispersion of the virtual channel samples, wherein the third input information includes At least one of the following information:

The number parameter L1 for real channel sample sampling;

The number parameter L2 for virtual channel sample sampling;

The maximum number of similar channel samples parameter K.

In some embodiments of the present application, the determining the target similarity between the virtual channel samples and the real channel samples and/or the target dispersion of the virtual channel samples according to the first input information, the second input information and the third input information includes :

Sampling the first input information and the second input information according to the third input information to obtain a real channel sample set and a virtual channel sample set;

determining K target real channel samples corresponding to each virtual channel sample in the set of virtual channel samples, wherein the K target real channel samples corresponding to each virtual channel sample are in the set of real channel samples corresponding to the set The K real channel samples with the highest similarity to the virtual channel samples;

According to K target real channel samples corresponding to each virtual channel sample, determine the similarity information of each virtual channel sample and the real channel sample;

According to the similarity information between each virtual channel sample and the real channel sample in the virtual channel sample set, determine the target similarity between the virtual channel sample and the real channel sample.

Determine the number of target virtual channel samples corresponding to each real channel sample, wherein the real channel samples belong to K target real channel samples corresponding to the corresponding target virtual channel samples;

Determine the target dispersion of the virtual channel samples according to the number of target virtual channel samples corresponding to each real channel sample.

In some embodiments of the present application, the determining the target dispersion of virtual channel samples according to the number of target virtual channel samples corresponding to each real channel sample includes:

The target dispersion of the virtual channel samples is determined according to the standard deviation of the number of target virtual channel samples corresponding to each real channel sample.

For example, among the virtual channel samples with the highest target similarity, the generator model corresponding to the virtual channel sample with the highest target dispersion is selected as the target generator model.

For another example, among the virtual channel samples with the highest target dispersion, the generator model corresponding to the virtual channel sample with the highest target similarity is selected as the target generator model.

For another example, the generator model corresponding to the virtual channel sample with the largest ratio of the target similarity to the target dispersion is selected as the target generator model.

Fig. 14 is a schematic block diagram of a method for evaluating the quality of virtual channel samples according to an embodiment of the present application. Wherein, the quality evaluation method may be executed by an evaluator on the first device, and the evaluator may be implemented as a processor.

Specifically, the first input information, the second input information and the third input information may be input into the evaluator to obtain the quality evaluation information of the second input information, for example, the aforementioned first index and second index, or, is the third indicator etc.

Moreover, in the embodiment of the present application, flexible input formats of virtual channel samples and real channel samples can be supported, specifically, before performing quality evaluation on virtual channel samples, for any input format of real channel samples and virtual channel samples, Both can generate first input information and second input information in a unified format based on the first configuration information.

The method embodiment of the present application is described in detail above in conjunction with FIG. 4 to FIG. 14 , and the device embodiment of the present application is described in detail below in conjunction with FIG. 15 to FIG. 19 . It should be understood that the device embodiment and the method embodiment correspond to each other, similar to The description can refer to the method embodiment.

Fig. 15 shows a schematic block diagram of a wireless communication device 400 according to an embodiment of the present application. As shown in FIG. 15 , the device 400 includes: a processing module 410, configured to acquire first input information and at least one second input information, wherein the first input information is information of real channel samples, and the second input information the information is information of virtual channel samples, and the at least one second input information corresponds to at least one generator model; and

Performing quality assessment on the at least one second input information and/or determining a target generator model in the at least one generator model based on the first input information and the at least one second input information.

In some embodiments, the processing module 410 is also used for:

Obtain first configuration information, where the first configuration information is used to configure at least one of the following:

A method of generating the first input information;

A method of generating the at least one second input information;

A quality assessment parameter of the at least one second input information.

Dimension number information of channel samples;

Dimension information of channel samples;

Configuration information for generating channel samples of antenna domain dimensions;

The number parameter L1 for real channel sample sampling;

The number parameter L2 for virtual channel sample sampling;

The number of similar channel samples parameter K.

In some embodiments, the indication information of the target frequency domain resource includes: index and length of the starting frequency domain resource; or,

The indication information of the target frequency domain resource includes: an index of a starting frequency domain resource and a frequency domain resource interval, wherein the frequency domain resource interval is an interval between two adjacent target frequency domain resources.

In some embodiments, the indication information of the target time domain resource includes: index and length of the starting time domain resource; or,

The indication information of the target time-domain resource includes: an index of a starting time-domain resource and a time-domain resource interval, wherein the time-domain resource interval is an interval between two adjacent target time-domain resources.

In some embodiments, the indication information of the target antenna domain resource includes: the index and the number of antennas of the starting antenna, or the index and the number of ports of the starting antenna; or,

The indication information of the target antenna domain resource includes: the index and antenna interval of the starting antenna, or the index and port interval of the starting antenna, wherein the antenna interval or port interval is between two adjacent antenna domain resources interval.

In some embodiments, the first input information is obtained according to real channel samples collected by the device.

In some embodiments, the first input information is obtained from a second device.

In some embodiments, the device 400 further includes: a communication module, configured to send first indication information to the second device, where the first indication information is used to instruct the second device to send the the first input information.

In some embodiments, the device 400 further includes: a communication module, configured to send first configuration information to the second device, the first input information is generated according to the first configuration information.

In some embodiments, the at least one second input information is obtained according to virtual channel samples generated by a generator deployed on the device.

In some embodiments, the at least one second input information is obtained from a second device.

In some embodiments, the device 400 further includes: a communication module, configured to send second indication information to the second device, where the second indication information is used to instruct the second device to send the at least one second input information.

In some embodiments, the device 400 further includes: a communication module, configured to send first configuration information to the second device, and the at least one second input information is generated according to the first configuration information.

In some embodiments, the device 400 further includes: a communication module, configured to receive the at least one second input information sent by the second device and generator indexes respectively corresponding to the at least one second input information.

In some embodiments, the device 400 further includes: a communication module, configured to send third indication information to the second device, where the third indication information is used to indicate the target generator index determined by the device.

In some embodiments, the device is a terminal device, and the second device is a network device; or,

The device is a network device, and the second device is a terminal device.

In some embodiments, the real channel samples include at least one of the following:

In some embodiments, the virtual channel samples include at least one of the following:

In some embodiments, the information of the time-domain channel samples includes information of at least one of the following dimensions:

In some embodiments, the information of the frequency domain channel samples includes information of at least one of the following dimensions:

In some embodiments, the information of the channel samples in the angle domain includes information of at least one of the following dimensions:

Angle of launch; Angle of arrival; length of granularity in time domain or length of frequency domain granularity.

In some embodiments, the time-domain granularity length is one of the following:

In some embodiments, the frequency domain granularity length is one of the following:

Number of subcarriers, number of resource blocks, number of subbands.

In some embodiments, the processing module 410 is also used for:

According to the first input information and each second input information in the at least one second input information, determine the quality evaluation information corresponding to each second input information, and determine the quality assessment information corresponding to each second input information The quality assessment information is used to indicate the similarity between the virtual channel samples and the real channel samples and/or the dispersion of the virtual channel samples.

In some embodiments, the processing module 410 is also used for:

According to the first input information, each second input information and third input information, determine the quality evaluation information corresponding to each second input information, wherein the third input information includes the following information at least one of:

The number parameter L1 for real channel sample sampling;

The number parameter L2 for virtual channel sample sampling;

The maximum number of similar channel samples parameter K.

In some embodiments, the processing module 410 is further configured to: determine a target generator model in the at least one generator model according to the quality evaluation information corresponding to each second input information.

Optionally, in some embodiments, the above-mentioned communication module may be a communication interface or a transceiver, or an input-output interface of a communication chip or a system on a chip. The above-mentioned processing module and encoding module may be one or more processors.

It should be understood that the device 400 according to the embodiment of the present application may correspond to the first device in the method embodiment of the present application, and the above-mentioned and other operations and/or functions of each unit in the device 400 are respectively in order to realize the The corresponding process of the first device in the method 200 is shown, and for the sake of brevity, details are not repeated here.

Fig. 16 is a schematic block diagram of a wireless communication device 500 according to an embodiment of the present application. The device 500 of Figure 16 includes:

A communication module 510, configured to send first information to the first device, where the first information includes at least one of the following:

The first input information is information of real channel samples;

At least one second input information is information of virtual channel samples, the at least one second input information corresponds to at least one generator model, and each second input information is generated by a corresponding generator model;

an index of the at least one generator.

In some embodiments, the communication module 510 is also used for:

receiving first indication information sent by the first device, where the first indication information is used to instruct the second device to send the first input information to the first device.

In some embodiments, the communication module 510 is also used for:

Sending second indication information to the second device, where the second indication information is used to instruct the second device to send the at least one piece of second input information to the first device.

In some embodiments, the communication module 510 is also used for:

Receive first configuration information sent by the first device, where the first configuration information is used to configure at least one of the following:

A method of generating the first input information;

A method of generating the at least one second input information;

A quality assessment parameter of the at least one second input information.

The number parameter L1 for real channel sample sampling;

The number parameter L2 for virtual channel sample sampling;

The maximum number of similar channel samples parameter K;

Dimension number information of channel samples;

configuration information for generating the channel sample size of the time domain dimension;

Number of subcarriers, number of resource blocks, number of subbands.

Optionally, in some embodiments, the above-mentioned communication module may be a communication interface or a transceiver, or an input-output interface of a communication chip or a system-on-chip. The above-mentioned processing module and encoding module may be one or more processors.

It should be understood that the device 500 according to the embodiment of the present application may correspond to the second device in the method embodiment of the present application, and the above-mentioned and other operations and/or functions of each unit in the device 500 are respectively in order to realize the The corresponding process of the second device in the method 200 is shown, and for the sake of brevity, details are not repeated here.

Fig. 17 is a schematic block diagram of a wireless communication device 800 according to an embodiment of the present application. The device 800 of Figure 17 includes:

A processing module 810, configured to determine quality evaluation information of the second input information according to the first input information and the second input information;

In some embodiments, the quality evaluation information of the second input information includes a first index and a second index, the first index is used to indicate the similarity between the virtual channel sample and the real channel sample, and the second index is used to indicate the dispersion of virtual channel samples; or,

The quality evaluation information of the second input information includes a third index, and the third index is generated according to the first index and the second index.

In some embodiments, the processing module 810 is also used for:

According to the first input information, the second input information and the third input information, determine the target similarity between the virtual channel samples and the real channel samples and/or the target dispersion of the virtual channel samples, wherein the third input information includes the following information At least one of:

The number parameter L1 for real channel sample sampling;

The number parameter L2 for virtual channel sample sampling;

The maximum number of similar channel samples parameter K.

In some embodiments, the processing module 810 is also used for:

Determine the similarity information between each virtual channel sample and the real channel sample according to the K target real channel samples corresponding to each virtual channel sample;

In some embodiments, the processing module 810 is also used for:

In some embodiments, the processing module 810 is further configured to: determine the target dispersion of the virtual channel samples according to the standard deviation of the number of target virtual channel samples corresponding to each real channel sample.

In some embodiments, the device is a terminal device or a network device.

It should be understood that the device 800 for wireless communication according to the embodiment of the present application may correspond to the first device in the method embodiment of the present application, and the above-mentioned and other operations and/or functions of each unit in the device 500 are respectively in order to realize the For the sake of brevity, the corresponding process of the first device in the method 300 shown in FIG. 14 will not be repeated here.

Fig. 18 is a schematic structural diagram of a communication device 600 provided by an embodiment of the present application. The communication device 600 shown in FIG. 18 includes a processor 610, and the processor 610 can call and run a computer program from a memory, so as to implement the method in the embodiment of the present application.

Optionally, as shown in FIG. 18 , the communication device 600 may further include a memory 620 . Wherein, the processor 610 can invoke and run a computer program from the memory 620, so as to implement the method in the embodiment of the present application.

Wherein, the memory 620 may be an independent device independent of the processor 610 , or may be integrated in the processor 610 .

Optionally, as shown in FIG. 6, the communication device 600 may further include a transceiver 630, and the processor 610 may control the transceiver 630 to communicate with other devices, specifically, to send information or data to other devices, or receive other Information or data sent by the device.

Wherein, the transceiver 630 may include a transmitter and a receiver. The transceiver 630 may further include antennas, and the number of antennas may be one or more.

Optionally, the communication device 600 may specifically be the first device in the embodiment of the present application, and the communication device 600 may implement the corresponding processes implemented by the first device in each method of the embodiment of the present application. Let me repeat.

Optionally, the communication device 600 may specifically be the second device in the embodiment of the present application, and the communication device 600 may implement the corresponding processes implemented by the second device in each method of the embodiment of the present application. For the sake of brevity, the Let me repeat.

FIG. 19 is a schematic structural diagram of a chip according to an embodiment of the present application. The chip 700 shown in FIG. 19 includes a processor 710, and the processor 710 can call and run a computer program from a memory, so as to implement the method in the embodiment of the present application.

Optionally, as shown in FIG. 19 , the chip 700 may further include a memory 720 . Wherein, the processor 710 can invoke and run a computer program from the memory 720, so as to implement the method in the embodiment of the present application.

Wherein, the memory 720 may be an independent device independent of the processor 710 , or may be integrated in the processor 710 .

Optionally, the chip 700 may also include an input interface 730 . Wherein, the processor 710 may control the input interface 730 to communicate with other devices or chips, specifically, may obtain information or data sent by other devices or chips.

Optionally, the chip 700 may also include an output interface 740 . Wherein, the processor 710 can control the output interface 740 to communicate with other devices or chips, specifically, can output information or data to other devices or chips.

Optionally, the chip can be applied to the first device in the embodiments of the present application, and the chip can implement the corresponding processes implemented by the first device in the various methods of the embodiments of the present application. For the sake of brevity, details are not repeated here.

Optionally, the chip can be applied to the second device in the embodiments of the present application, and the chip can implement the corresponding processes implemented by the second device in the various methods of the embodiments of the present application. For the sake of brevity, details are not repeated here.

It should be understood that the chip mentioned in the embodiment of the present application may also be called a system-on-chip, a system-on-chip, a system-on-a-chip, or a system-on-a-chip.

It should be understood that the processor in the embodiment of the present application may be an integrated circuit chip, which has a signal processing capability. In the implementation process, each step of the above-mentioned method embodiments may be completed by an integrated logic circuit of hardware in a processor or instructions in the form of software. The above-mentioned processor can be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), an off-the-shelf programmable gate array (Field Programmable Gate Array, FPGA) or other available Program logic devices, discrete gate or transistor logic devices, discrete hardware components. Various methods, steps, and logic block diagrams disclosed in the embodiments of the present application may be implemented or executed. A general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like. The steps of the method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the field such as random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, register. The storage medium is located in the memory, and the processor reads the information in the memory, and completes the steps of the above method in combination with its hardware.

It can be understood that the memory in the embodiments of the present application may be a volatile memory or a nonvolatile memory, or may include both volatile and nonvolatile memories. Among them, the non-volatile memory can be read-only memory (Read-Only Memory, ROM), programmable read-only memory (Programmable ROM, PROM), erasable programmable read-only memory (Erasable PROM, EPROM), electronically programmable Erase Programmable Read-Only Memory (Electrically EPROM, EEPROM) or Flash. The volatile memory can be Random Access Memory (RAM), which acts as external cache memory. By way of illustration and not limitation, many forms of RAM are available, such as Static Random Access Memory (Static RAM, SRAM), Dynamic Random Access Memory (Dynamic RAM, DRAM), Synchronous Dynamic Random Access Memory (Synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (Double Data Rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (Enhanced SDRAM, ESDRAM), synchronous connection dynamic random access memory (Synchlink DRAM, SLDRAM ) and Direct Memory Bus Random Access Memory (Direct Rambus RAM, DR RAM). It should be noted that the memory of the systems and methods described herein is intended to include, but not be limited to, these and any other suitable types of memory.

It should be understood that the above-mentioned memory is illustrative but not restrictive. For example, the memory in the embodiment of the present application may also be a static random access memory (static RAM, SRAM), a dynamic random access memory (dynamic RAM, DRAM), Synchronous dynamic random access memory (synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (double data rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (enhanced SDRAM, ESDRAM), synchronous connection Dynamic random access memory (synch link DRAM, SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DR RAM), etc. That is, the memory in the embodiments of the present application is intended to include, but not be limited to, these and any other suitable types of memory.

The embodiment of the present application also provides a computer-readable storage medium for storing computer programs.

Optionally, the computer-readable storage medium may be applied to the first device in the embodiments of the present application, and the computer program causes the computer to execute the corresponding processes implemented by the first device in the methods of the embodiments of the present application. For brevity, I won't repeat them here.

Optionally, the computer-readable storage medium may be applied to the second device in the embodiment of the present application, and the computer program causes the computer to execute the corresponding process implemented by the second device in each method of the embodiment of the present application. For brevity, I won't repeat them here.

The embodiment of the present application also provides a computer program product, including computer program instructions.

Optionally, the computer program product can be applied to the first device in the embodiments of the present application, and the computer program instructions cause the computer to execute the corresponding processes implemented by the first device in the methods of the embodiments of the present application. For brevity, the This will not be repeated here.

Optionally, the computer program product can be applied to the second device in the embodiment of the present application, and the computer program instructions cause the computer to execute the corresponding process implemented by the second device in each method of the embodiment of the present application. For brevity, in This will not be repeated here.

The embodiment of the present application also provides a computer program.

Optionally, the computer program may be applied to the first device in the embodiment of the present application, and when the computer program is run on the computer, the computer executes the corresponding process implemented by the first device in each method of the embodiment of the present application, For the sake of brevity, details are not repeated here.

Optionally, the computer program can be applied to the second device in the embodiment of the present application, and when the computer program is run on the computer, the computer executes the corresponding process implemented by the second device in each method of the embodiment of the present application, For the sake of brevity, details are not repeated here.

Those skilled in the art can appreciate that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed by hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans may use different methods to implement the described functions for each specific application, but such implementation should not be regarded as exceeding the scope of the present application.

Those skilled in the art can clearly understand that for the convenience and brevity of the description, the specific working process of the above-described system, device and unit can refer to the corresponding process in the foregoing method embodiment, which will not be repeated here.

In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods may be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or May be integrated into another system, or some features may be ignored, or not implemented. In another point, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.

The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing module, each unit may exist separately physically, or two or more units may be integrated into one unit.

If the functions described above are realized in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including Several instructions are used to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disk or optical disc and other media that can store program codes. .

The above is only a specific implementation of the application, but the scope of protection of the application is not limited thereto. Anyone familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the application. Should be covered within the protection scope of this application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

A method for evaluating the quality of virtual channel samples, comprising:

The first device acquires first input information and at least one second input information, wherein the first input information is information of real channel samples, the second input information is information of virtual channel samples, and the at least one second input information The input information corresponds to at least one generator model;

The first device performs quality assessment on the at least one second input information and/or determines a target generator model in the at least one generator model according to the first input information and the at least one second input information.
The method according to claim 1, further comprising:

The first device acquires first configuration information, where the first configuration information is used to configure at least one of the following:

A method of generating the first input information;

A method of generating the at least one second input information;

A quality assessment parameter of the at least one second input information.
The method according to claim 2, wherein the first configuration information includes at least one of the following:

Dimension number information of channel samples;

Dimension information of channel samples;

Configuration information for generating channel samples in the frequency domain dimension;

Configuration information for generating channel samples in the time domain dimension;

Configuration information for generating channel samples of antenna domain dimensions;

The number parameter L1 for real channel sample sampling;

The number parameter L2 for virtual channel sample sampling;

The number of similar channel samples parameter K.
The method according to claim 3, wherein the configuration information for generating channel samples in the frequency domain dimension includes at least one of the following:

Frequency domain granularity configuration, resource tailoring method of frequency domain dimension, indication information of target frequency domain resources.
The method according to claim 4, wherein the indication information of the target frequency domain resource includes: the index and length of the starting frequency domain resource; or,

The indication information of the target frequency domain resource includes: an index of a starting frequency domain resource and a frequency domain resource interval, wherein the frequency domain resource interval is an interval between two adjacent target frequency domain resources.
The method according to any one of claims 3-5, wherein the configuration information for generating the channel samples of the time domain dimension includes at least one of the following:

Time-domain granularity configuration, resource clipping mode of time-domain dimension, and indication information of target time-domain resources.
The method according to claim 6, wherein the indication information of the target time domain resource includes: the index and length of the initial time domain resource; or,

The indication information of the target time-domain resource includes: an index of a starting time-domain resource and a time-domain resource interval, wherein the time-domain resource interval is an interval between two adjacent target time-domain resources.
The method according to any one of claims 3-7, wherein the configuration information for generating channel samples of antenna domain dimensions includes at least one of the following:

Antenna domain granularity configuration, resource tailoring mode of antenna domain dimension, indication information of target antenna domain resources.
The method according to claim 8, wherein the indication information of the target antenna domain resource includes: the index and the number of antennas of the starting antenna, or the index and the number of ports of the starting antenna; or,

The indication information of the target antenna domain resource includes: the index and antenna interval of the starting antenna, or the index and port interval of the starting antenna, wherein the antenna interval or port interval is between two adjacent antenna domain resources interval.
The method according to any one of claims 1-9, wherein the first input information is obtained according to real channel samples collected by the first device.
The method according to any one of claims 1-9, wherein the first input information is acquired from a second device.
The method according to claim 11, characterized in that the method further comprises:

The first device sends first indication information to the second device, where the first indication information is used to instruct the second device to send the first input information to the first device.
The method according to claim 11 or 12, characterized in that the method further comprises:

The first device sends first configuration information to the second device, and the first input information is generated according to the first configuration information.
The method according to any one of claims 1-13, wherein the at least one second input information is obtained by the first device according to a virtual channel sample generated by a generator deployed on the first device of.
The method according to any one of claims 1-13, wherein the at least one second input information is acquired by the first device from a second device.
The method according to claim 15, further comprising:

The first device sends second indication information to the second device, where the second indication information is used to instruct the second device to send the at least one piece of second input information to the first device.
The method according to claim 15 or 16, wherein the method further comprises:

The first device sends first configuration information to the second device, and the at least one second input information is generated according to the first configuration information.
The method according to any one of claims 15-17, further comprising:

The first device receives the at least one second input information sent by the second device and generator indexes respectively corresponding to the at least one second input information.
The method according to claim 18, further comprising:

The first device sends third indication information to the second device, where the third indication information is used to indicate the target generator index determined by the first device.
The method according to any one of claims 11-13, 15-19, characterized in that,

The first device is a terminal device, and the second device is a network device; or,

The first device is a network device, and the second device is a terminal device.
The method according to any one of claims 1-20, wherein the real channel samples comprise at least one of the following:

Channel samples in time domain, channel samples in frequency domain, channel samples in angle domain.
The method according to any one of claims 1-21, wherein the virtual channel samples include at least one of the following:

Channel samples in time domain, channel samples in frequency domain, channel samples in angle domain.
The method according to claim 21 or 22, wherein the information of the time-domain channel samples includes information of at least one of the following dimensions:

Number of transmit antennas or transmit ports;

Number of receiving antennas or receiving ports;

Time domain granularity length.
The method according to any one of claims 21-23, wherein the information of the frequency domain channel samples includes information of at least one of the following dimensions:

Number of transmit antennas or transmit ports;

Number of receiving antennas or receiving ports;

frequency domain granularity length;

Time domain granularity length.
The method according to any one of claims 21-24, wherein the information of the channel samples in the angle domain includes information of at least one of the following dimensions:

launch angle;

angle of arrival;

Time Domain Granular Length or Frequency Domain Granular Length.
The method according to any one of claims 23-25, wherein the time-domain granularity length is one of the following:

The real multipath number of the channel, the number of sampling points sampled in the time domain according to the first sampling rate, and the number of time units for time domain sampling.
The method according to claim 24 or 25, wherein the frequency-domain granularity length is one of the following:

Number of subcarriers, number of resource blocks, number of subbands.
The method according to any one of claims 1-27, wherein the first device, according to the first input information and the at least one second input information, performing a quality assessment and/or determining a target generator model in said at least one generator model comprising:

According to the first input information and each second input information in the at least one second input information, determine the quality evaluation information corresponding to each second input information, and determine the quality assessment information corresponding to each second input information The quality assessment information is used to indicate the similarity between the virtual channel samples and the real channel samples and/or the dispersion of the virtual channel samples.
The method according to claim 28, wherein, according to the first input information and each second input information in the at least one second input information, it is determined that each second input information corresponds to quality assessment information, including:

According to the first input information, each second input information and third input information, determine the quality evaluation information corresponding to each second input information, wherein the third input information includes the following information at least one of:

The number parameter L1 for real channel sample sampling;

The number parameter L2 for virtual channel sample sampling;

The maximum number of similar channel samples parameter K.
The method according to claim 28 or 29, wherein the first device performs quality assessment and /or determining a target generator model in said at least one generator model, comprising:

A target generator model in the at least one generator model is determined according to the quality assessment information corresponding to each second input information.
A method for evaluating the quality of virtual channel samples, comprising:

The first information sent by the second device to the first device, where the first information includes at least one of the following:

The first input information is information of real channel samples;

At least one second input information is information of virtual channel samples, the at least one second input information corresponds to at least one generator model, and each second input information is generated by a corresponding generator model;

an index of the at least one generator.
The method according to claim 31, further comprising:

The second device receives first indication information sent by the first device, where the first indication information is used to instruct the second device to send the first input information to the first device.
The method according to claim 31 or 32, further comprising:

The first device sends second indication information to the second device, where the second indication information is used to instruct the second device to send the at least one piece of second input information to the first device.
The method according to any one of claims 31-33, further comprising:

The second device receives first configuration information sent by the first device, where the first configuration information is used to configure at least one of the following:

A method of generating the first input information;

A method of generating the at least one second input information;

A quality assessment parameter of the at least one second input information.
The method according to claim 34, wherein the first configuration information includes at least one of the following:

The number parameter L1 for real channel sample sampling;

The number parameter L2 for virtual channel sample sampling;

The maximum number of similar channel samples parameter K;

Dimension number information of channel samples;

Configuration information for generating channel samples in the frequency domain dimension;

configuration information for generating the channel sample size of the time domain dimension;

Configuration information for generating channel samples of antenna domain dimension.
The method according to claim 35, wherein the configuration information for generating channel samples in the frequency domain dimension includes at least one of the following:

Frequency domain granularity configuration, resource tailoring method of frequency domain dimension, indication information of target frequency domain resources.
The method according to claim 36, wherein the indication information of the target frequency domain resource includes: the index and length of the starting frequency domain resource; or,

The indication information of the target frequency domain resource includes: an index of a starting frequency domain resource and a frequency domain resource interval, wherein the frequency domain resource interval is an interval between two adjacent target frequency domain resources.
The method according to any one of claims 35-37, wherein the configuration information for generating channel samples of the time domain dimension includes at least one of the following:

Time-domain granularity configuration, resource clipping mode of time-domain dimension, and indication information of target time-domain resources.
The method according to claim 38, wherein the indication information of the target time-domain resource includes: the index and length of the starting time-domain resource; or,

The indication information of the target time-domain resource includes: an index of a starting time-domain resource and a time-domain resource interval, wherein the time-domain resource interval is an interval between two adjacent target time-domain resources.
The method according to any one of claims 35-39, wherein the configuration information for generating the channel samples of the antenna domain dimension comprises at least one of the following:

Antenna domain granularity configuration, resource tailoring mode of antenna domain dimension, indication information of target antenna domain resources.
The method according to claim 40, wherein the indication information of the target antenna domain resource includes: the index and the number of antennas of the starting antenna, or the index and the number of ports of the starting antenna; or,

The indication information of the target antenna domain resource includes: the index and antenna interval of the starting antenna, or the index and port interval of the starting antenna, wherein the antenna interval or port interval is between two adjacent antenna domain resources interval.
The method according to any one of claims 31-41, wherein the real channel samples include at least one of the following:

Channel samples in time domain, channel samples in frequency domain, channel samples in angle domain.
The method according to any one of claims 31-42, wherein the virtual channel samples include at least one of the following:

Channel samples in time domain, channel samples in frequency domain, channel samples in angle domain.
The method according to claim 42 or 43, wherein the information of the time-domain channel samples includes information of at least one of the following dimensions:

Number of transmit antennas or transmit ports;

Number of receiving antennas or receiving ports;

Time domain granularity length.
The method according to any one of claims 42-44, wherein the information of the frequency domain channel samples includes information of at least one of the following dimensions:

Number of transmit antennas or transmit ports;

Number of receiving antennas or receiving ports;

frequency domain granularity length;

Time domain granularity length.
The method according to any one of claims 42-45, wherein the information of the channel samples in the angle domain includes information of at least one of the following dimensions:

launch angle;

angle of arrival;

Time Domain Granular Length or Frequency Domain Granular Length.
The method according to any one of claims 44-46, wherein the time-domain granularity length is one of the following:

The real multipath number of the channel, the number of sampling points sampled in the time domain according to the first sampling rate, and the number of time units for time domain sampling.
The method according to claim 45 or 46, wherein the frequency-domain granularity length is one of the following:

Number of subcarriers, number of resource blocks, number of subbands.
A method for evaluating the quality of virtual channel samples, comprising:

The first device determines quality evaluation information of the second input information according to the first input information and the second input information;

Wherein, the first input information is information of real channel samples, the second input information is information of virtual channel samples, and the quality evaluation information of the second input information is used to indicate the similarity between virtual channel samples and real channel samples degree and/or dispersion of virtual channel samples.
The method according to claim 49, wherein the real channel samples comprise at least one of the following:

Channel samples in time domain, channel samples in frequency domain, channel samples in angle domain.
The method according to claim 49 or 50, wherein the virtual channel samples comprise at least one of the following:

Channel samples in time domain, channel samples in frequency domain, channel samples in angle domain.
The method according to claim 50 or 51, wherein the information of the time-domain channel samples includes information of at least one of the following dimensions:

Number of transmit antennas or transmit ports;

Number of receiving antennas or receiving ports;

Time domain granularity length.
The method according to any one of claims 50-52, wherein the information of the frequency domain channel samples includes information of at least one of the following dimensions:

Number of transmit antennas or transmit ports;

Number of receiving antennas or receiving ports;

frequency domain granularity length;

Time domain granularity length.
The method according to any one of claims 50-53, wherein the information of the channel samples in the angle domain includes information of at least one of the following dimensions:

launch angle;

angle of arrival;

Time Domain Granular Length or Frequency Domain Granular Length.
The method according to any one of claims 49-54, wherein the quality evaluation information of the second input information includes a first index and a second index, and the first index is used to indicate the virtual channel samples and The similarity of real channel samples, the second indicator is used to indicate the dispersion of virtual channel samples; or,

The quality evaluation information of the second input information includes a third index, and the third index is generated according to the first index and the second index.
The method according to claim 55, wherein the first device generates the quality evaluation information of the second input information according to the first input information and the second input information, comprising:

According to the first input information and the second input information, a target similarity between virtual channel samples and real channel samples and/or a target dispersion of virtual channel samples are determined.
The method according to claim 56, wherein, according to the first input information and the second input information, determining the target similarity between the virtual channel sample and the real channel sample and/or the target of the virtual channel sample Dispersion, including:

According to the first input information, the second input information and the third input information, determine the target similarity between the virtual channel samples and the real channel samples and/or the target dispersion of the virtual channel samples, wherein the third input information includes At least one of the following information:

The number parameter L1 for real channel sample sampling;

The number parameter L2 for virtual channel sample sampling;

The maximum number of similar channel samples parameter K.
The method according to claim 57, wherein, according to the first input information, the second input information and the third input information, determining the target similarity between the virtual channel sample and the real channel sample and/or the target similarity of the virtual channel sample Target dispersion, including:

Sampling the first input information and the second input information according to the third input information to obtain a real channel sample set and a virtual channel sample set;

determining K target real channel samples corresponding to each virtual channel sample in the set of virtual channel samples, wherein the K target real channel samples corresponding to each virtual channel sample are in the set of real channel samples corresponding to the set The K real channel samples with the highest similarity to the virtual channel samples;

Determine the similarity information between each virtual channel sample and the real channel sample according to the K target real channel samples corresponding to each virtual channel sample;

According to the similarity information between each virtual channel sample and the real channel sample in the virtual channel sample set, determine the target similarity between the virtual channel sample and the real channel sample.
The method according to claim 57 or 58, characterized in that, according to the first input information, the second input information and the third input information, determining the target similarity between the virtual channel sample and the real channel sample and/or the virtual channel The target dispersion of the sample, including:

Sampling the first input information and the second input information according to the third input information to obtain a real channel sample set and a virtual channel sample set;

determining K target real channel samples corresponding to each virtual channel sample in the set of virtual channel samples, wherein the K target real channel samples corresponding to each virtual channel sample are in the set of real channel samples corresponding to the set The K real channel samples with the highest similarity to the virtual channel samples;

Determine the number of target virtual channel samples corresponding to each real channel sample, wherein the real channel samples belong to K target real channel samples corresponding to the corresponding target virtual channel samples;

Determine the target dispersion of the virtual channel samples according to the number of target virtual channel samples corresponding to each real channel sample.
The method according to claim 59, wherein said determining the target dispersion of virtual channel samples according to the number of target virtual channel samples corresponding to each real channel sample comprises:

The target dispersion of the virtual channel samples is determined according to the standard deviation of the number of target virtual channel samples corresponding to each real channel sample.
The method according to any one of claims 49-60, wherein the first device is a terminal device or a network device.
A wireless communication device, characterized in that it includes:

A processing module, configured to obtain first input information and at least one second input information, wherein the first input information is information of real channel samples, the second input information is information of virtual channel samples, and the at least one the second input information corresponds to at least one generator model; and

Performing quality assessment on the at least one second input information and/or determining a target generator model in the at least one generator model based on the first input information and the at least one second input information.
A wireless communication device, characterized in that it includes:

A communication module, configured to send first information to the first device, where the first information includes at least one of the following:

The first input information is information of real channel samples;

At least one second input information is information of virtual channel samples, the at least one second input information corresponds to at least one generator model, and each second input information is generated by a corresponding generator model;

an index of the at least one generator.
A wireless communication device, characterized in that it includes:

A processing module, configured to determine quality evaluation information of the second input information according to the first input information and the second input information;

Wherein, the first input information is information of real channel samples, the second input information is information of virtual channel samples, and the quality assessment information of the second input information is used to indicate the similarity between virtual channel samples and real channel samples degree and/or dispersion of virtual channel samples.
A wireless communication device, characterized by comprising: a processor and a memory, the memory is used to store a computer program, the processor is used to invoke and run the computer program stored in the memory, and perform the tasks described in claims 1 to 30 The method of any one of the above, or the method of any one of claims 31 to 48, or the method of any one of claims 49-61.
A chip, characterized by comprising: a processor, configured to call and run a computer program from a memory, so that a device equipped with the chip executes the method according to any one of claims 1 to 30, or as described in A method as claimed in any one of claims 31 to 48, or a method as claimed in any one of claims 49-61.
A computer-readable storage medium, characterized in that it is used to store a computer program, the computer program causes the computer to perform the method according to any one of claims 1 to 30, or any one of claims 31 to 48 The method described in claim 49, or the method described in any one of claims 49-61.
A computer program product, characterized in that it includes computer program instructions, the computer program instructions cause the computer to perform the method as described in any one of claims 1 to 30, or as described in any one of claims 31 to 48 , or a method as claimed in any one of claims 49-61.
A computer program, characterized in that the computer program causes a computer to execute the method according to any one of claims 1 to 30, or the method according to any one of claims 31 to 48, or the method according to any one of claims The method of any one of claims 49-61.