Communication method and apparatus

Abstract

The application provides a communication method and device. The method comprises: receiving, by a terminal device, first signaling from a network device, the first signaling being used for configuring performance monitoring of a first channel state information (CSI) feedback compression model, the first signaling at least comprising a trigger mode of second signaling; determining, based on the first signaling and a performance index estimation model, a first estimated value of a performance index of the first CSI feedback compression model; and sending, based on the trigger mode, the second signaling to the network device, the second signaling being used for requesting the network device to perform the performance monitoring of the first CSI feedback compression model. In this way, effective model performance monitoring can be achieved while reducing complexity and saving air interface overhead.

Description

Technical Field

[0001] This application relates to the field of communications, and more specifically, to communication methods and communication apparatus. Background Technology

[0002] In mobile communication systems, terminal devices need to feed back channel state information (CSI) to network devices to help the network side select beams and improve communication quality. Due to the large amount of data fed back in CSI, feedback compression is typically employed. Compressed and decompressed models are deployed at both the terminal and network devices to reduce the complexity of the CSI feedback module and air interface overhead, and to some extent, improve system performance. However, the models deployed at either the network or terminal devices may experience performance degradation due to changes in the channel environment or the distribution of input data. Therefore, while deploying the model, a complete and reliable model performance monitoring scheme needs to be designed to ensure the correct operation of the model and to provide reference information for model lifecycle management (LCM). Summary of the Invention

[0003] This application provides a communication scheme that enables effective model performance monitoring while reducing complexity and saving air interface overhead.

[0004] In a first aspect of this application, a communication method is provided. The method includes: a terminal device receiving a first signaling from a network device, the first signaling being configured to perform performance monitoring of a first CSI feedback compression model, the first signaling including at least a triggering mode for a second signaling; determining a first estimated value of a performance index of the first CSI feedback compression model based on the first signaling and a performance index estimation model; and sending a second signaling to the network device based on the triggering mode to request the network device to perform performance monitoring of the first CSI feedback compression model.

[0005] In a second aspect of this application, a communication method is provided. The method includes: a network device sending a first signaling message to a terminal device, the first signaling message configuring performance monitoring of a first CSI feedback compression model at the terminal device, the first signaling message including at least a triggering mode of a second signaling message; and receiving a second signaling message from the terminal device, the second signaling message requesting the network device to perform performance monitoring of the first CSI feedback compression model.

[0006] In a third aspect of this application, a communication device is provided, which has the function of implementing the behavior in the method example of the first aspect described above. The function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described function. The description of the method of the first aspect also applies to the communication device of the third aspect; that is, the communication device of the third aspect can refer to the communication device itself or to components within the communication device (e.g., processor, chip, or chip system, etc.). In one possible design, the communication device includes a unit that performs the method of the first aspect or its implementation. The beneficial effects of the communication device provided by the third aspect can be found in the description of the first aspect, and will not be repeated here.

[0007] In a fourth aspect of this application, a communication device is provided, which has the function of implementing the behavior in the method example of the second aspect described above. The function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described function. The description of the method in the second aspect also applies to the communication device of the fourth aspect; that is, the communication device of the fourth aspect can refer to the communication device itself or to components within the communication device (e.g., processor, chip, or chip system, etc.). In one possible design, the communication device includes a unit that performs the method of the second aspect or its implementation. The beneficial effects of the communication device provided in the fourth aspect can be found in the description of the second aspect, and will not be repeated here.

[0008] In a fifth aspect of this application, an apparatus is provided, including a processor and a memory storing a computer program or instructions, which, when executed by an electronic device, cause the electronic device to perform any of the methods described in the first or second aspect and their implementations.

[0009] In a sixth aspect of this application, a computer-readable storage medium is provided, which stores a computer program or instructions that, when executed by an electronic device, cause the electronic device to perform the methods described in the first or second aspect above.

[0010] In a seventh aspect of this application, a computer program or computer program product is provided, the computer program or computer program product including computer program instructions, which, when executed by an electronic device, cause the electronic device to perform the methods in the first or second aspect described above.

[0011] In an eighth aspect of this application, embodiments of this disclosure provide a chip system including a processor for implementing the functions of the apparatus in the methods of the first or second aspect described above. In one possible design, the chip system further includes a memory for storing computer programs or instructions and / or data. This chip system may be composed of chips or may include chips and other discrete devices. Attached Figure Description

[0012] Figure 1 A schematic block diagram of a communication system in which one possible implementation of this application may be carried out is shown;

[0013] Figures 2A to 2E Signaling interaction diagrams of the communication process according to some possible implementations of this application are shown respectively;

[0014] Figure 3 A signaling interaction diagram of a communication process according to one possible implementation of this application is shown;

[0015] Figure 4 A signaling interaction diagram of a communication process according to one possible implementation of this application is shown;

[0016] Figure 5A An example of a user-side performance monitoring cycle according to one possible implementation of this application is shown;

[0017] Figure 5B A schematic diagram of a possible implementation of GCS according to this application is shown;

[0018] Figure 6A , 6B Figures 6 and 6C respectively illustrate schematic diagrams of triggering modes of the second signaling according to some possible implementations of this application;

[0019] Figure 7A and Figure 7B Signaling interaction diagrams of the communication process according to some possible implementations of this application are shown respectively;

[0020] Figure 8 A signaling interaction diagram of a communication process according to one possible implementation of this application is shown;

[0021] Figure 9 and Figure 10 Flowcharts illustrating the implementation of some possible implementations of this application at the communication device are shown respectively;

[0022] Figure 11 and Figure 12 Schematic diagrams of the apparatus according to some embodiments of the present disclosure are shown respectively;

[0023] Figure 13 A block diagram is shown that can be used to implement some embodiments of the present disclosure.

[0024] In the various figures, the same or similar reference numerals represent the same or similar elements. Detailed Implementation

[0025] Possible implementations of this application will now be described in more detail with reference to the accompanying drawings.

[0026] In the description of possible implementations of this application, the term "comprising" and similar terms should be understood as open inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one possible implementation" or "the possible implementation" should be understood as "at least one possible implementation". The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below. Expressions such as "at least one of A, B, and C" or "at least one of A, B, or C" should be understood as any of the following: at least one A; at least one B; at least one C; at least one A and at least one B; at least one A and at least one C; at least one B and at least one C; at least one A, at least one B, and at least one C. The above examples use three elements, A, B, and C, for illustration. When there are more elements in the expression, the meaning of the expression can be obtained according to the aforementioned rules.

[0027] Figure 1 An example of a communication system 100 in which one possible implementation of this application can be carried out is shown. As shown, the communication system 100 may include a terminal device 110 and a network device 120. The terminal device 110 is located within the coverage area 122 of the network device 120.

[0028] Terminal device 110 is a device with wireless transceiver capabilities. Terminal device 110 can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; it can also be deployed on water (such as on ships); and it can also be deployed in the air (such as on airplanes, balloons, and satellites). Terminal device 110 can be a mobile phone, tablet, computer with wireless transceiver capabilities, virtual reality (VR) terminal device, augmented reality (AR) terminal device, wireless terminal device in industrial control, wireless terminal device in self-driving, wireless terminal device in remote medical care, wireless terminal device in smart grid, wireless terminal device in transportation safety, wireless terminal device in smart city, wireless terminal device in smart home, or user equipment (UE), etc.

[0029] Network device 120 is a radio access network (RAN) device that connects terminal device 110 to a wireless network. RAN devices may include, but are not limited to: Next-generation Node B (gNB), Transmission Reception Point (TRP), Evolved Node B (eNB), Radio Network Controller (RNC), Node B (NB), Base Station Controller (BSC), Base Transceiver Station (BTS), femtocell (e.g., femtocell evolved Node B, or femtocell Node B), Base Band Unit (BBU), or Wireless Fidelity (WiFi) Access Point (AP), Integrated Access and Backhaul (IAB) node, etc. In one possible implementation, network device 120 may include a Centralized Unit (CU), a Distributed Unit (DU), or both CU and DU.

[0030] Network device 120 can communicate with terminal device 110 via a channel such as a wireless communication channel. Communication in a multi-carrier network can conform to any suitable standard, including but not limited to Global System for Mobile Communication (GSM), Long Term Evolution (LTE), LTE Evolution, LTE-Advanced (LTE-A), Wideband Code Division Multiple Access (WCDMA), Code Division Multiple Access (CDMA), GSM Edge Radio Access Network (GERAN), Machine Type Communication (MTC), etc. Furthermore, communication can be performed according to any generation of communication protocols currently known or to be developed in the future. Examples of communication protocols include, but are not limited to, First Generation (1G), Second Generation (2G), 2.5G, 2.75G, Third Generation (3G), Fourth Generation (4G), 4.5G, Fifth Generation (5G), and Sixth Generation (6G) communication protocols.

[0031] Understandable. Figure 1 The number of terminal devices and network devices shown is merely an example, and this application does not impose any limitation on it. Depending on actual needs, the communication system 100 can deploy any appropriate number of terminal devices and network devices.

[0032] In 5G and future 6G systems, obtaining accurate Channel Component Index (CSI) is crucial for achieving key technologies such as Massive MIMO and beamforming. However, with the continuous increase in antenna size, the dimensionality of the CSI matrix increases significantly, leading to a sharp increase in uplink feedback overhead, which becomes a bottleneck for system performance improvement. Traditional compression methods mostly rely on the sparsity of the channel in the angular delay domain, but in scattering-rich non-line-of-sight (NLOS) scenarios such as Sub-6GHz, path sparsity is weak, limiting their effectiveness. In recent years, AI-based CSI compression technology has shown significant advantages. Its core idea is to use deep neural networks to learn the inherent low-dimensional structure of the channel matrix to achieve efficient dimensionality reduction and reconstruction.

[0033] In fact, although the CSI matrix exhibits a high-dimensional, dense form in the spatial domain, the limited number of dominant paths in the wireless propagation environment means that channel energy is often concentrated in a few spatial directions, resulting in a significantly low-rank characteristic of the channel matrix. Let the downlink channel matrix be... Its singular value decomposition (SVD) can be expressed as: ,

[0034] in It is a left singular vector matrix. Represents a singular value matrix. Describes a right singular vector matrix. , is a singular value. The effective rank is represented by . This low-rank structure shows that CSI can be precisely approximated by a small number of dominant singular vectors and singular values, providing a theoretical basis for compression.

[0035] Based on this, existing AI / ML models achieve efficient compression by mapping the channel matrix to a low-dimensional latent space through end-to-end training. Typical architectures include CsiNet, CNN-LSTM, or transformer-based encoder-decoder structures, where the encoder maps the original channel matrix to a low-dimensional latent space. Compressed into low-dimensional feature vectors The decoder then from Reconstructing the approximate channel The network is trained by minimizing reconstruction error, implicitly learning the statistical properties and low-dimensional manifold structure of the channel. Compared to traditional sparse representation methods, artificial intelligence / machine learning (AI / ML) models do not require explicit modeling of propagation paths and can adapt to complex and variable channel environments. They maintain high reconstruction accuracy even with low bit feedback. The AI-driven CSI compression framework fully utilizes the low-rank prior of the channel, significantly reducing feedback overhead and providing strong support for future intelligent wireless communication systems.

[0036] In the AI / ML-based CSI compression method, the terminal device deploys an AI / ML-based compression model to compress the CSI and then sends the compressed CSI to the network device via the air interface. The network device, in turn, deploys an AI / ML-based decompression model to decompress the compressed CSI. This significantly reduces the amount of CSI feedback while maintaining high restoration accuracy, thereby saving wireless resources and improving system efficiency. The key to CSI feedback compression lies in the compression of the Precoding Matrix Indicator (PMI) in the feedback information. Unlike traditional methods that use different codebooks and parameters for transmission in different scenarios and situations, AI / ML-based compression can potentially achieve optimized compression of the PMI in multiple scenarios through a unified two-sided model, achieving good generalization ability while maintaining good compression performance.

[0037] However, the performance of AI / ML-based CSI compression methods is affected by various factors, such as data drift, training errors, and compilation errors. Currently, AI / ML-based CSI compression methods generally lack effective performance monitoring mechanisms, making it difficult to detect and address performance degradation of AI / ML models in a timely manner. Once the performance of the compressed model deteriorates, it will affect the accuracy and reliability of CSI feedback, leading to a decline in system performance or network quality. Furthermore, there is the problem of not being able to clearly indicate the specific model causing the performance degradation.

[0038] As mentioned earlier, models deployed on network devices or terminal devices may experience performance degradation due to changes in the channel environment or the distribution of input data. Therefore, while deploying the model, it is also necessary to design a complete and reliable model performance monitoring scheme to ensure the correct operation of the model and provide reference information for the model's LCM.

[0039] For CSI feedback compression two-sided models, ensuring reliable model performance monitoring requires simultaneous monitoring of the decompression / compression models on both the network and user sides. However, the intermediate performance indicator (KPI) used for monitoring—Squared Generalized Cosine Similarity (SGCS)—requires both the actual CSI value before compression and the decompressed CSI value. These two parameters are generated by the terminal device and the network device, respectively. This means that monitoring the model performance on either side requires either deploying an additional model or directly exchanging information over the air interface. Therefore, rationally configuring the performance monitoring scheme to minimize complexity and save air interface overhead becomes a pressing issue.

[0040] To address the aforementioned issues, a model monitoring approach combining network device-side and terminal device-side monitoring can be adopted. This will refer to... Figures 2A to 2E Describe it.

[0041] Figures 2A to 2E Signaling interaction diagrams for process 200A to 200E used for model monitoring on the network device side and terminal device side are shown respectively. For the purposes of discussion, the following will refer to... Figure 1 The elements shown are used to describe processes 200A to 200E. However, it should be understood that processes 200A to 200E can also be performed between network devices and terminal devices in any other communication scenario.

[0042] like Figure 2A As shown, in process 200A, network device 120 first sends control signaling to terminal device 110. This control signaling configures performance monitoring of the AI / ML-based CSI compression model. Network device 120 sends a channel state information reference signal (CSI-RS) to terminal device 110. Terminal device 110 performs measurements based on the CSI-RS to obtain the raw CSI, i.e., the true value of the CSI, and sends a CSI report including compressed CSI to network device 120 via the CSI feedback module. Terminal device 110 periodically sends the true value of the CSI to network device 120. Network device 120 decodes the received compressed CSI using a decoder and calculates the SGCS based on the decoded information and the true value of the CSI. Then, network device 120 performs self-management of the model based on the SGCS. Therefore, by introducing the reporting of the true value of CSI on top of conventional CSI feedback, performance monitoring and management of the CSI feedback decompression model can be achieved on the network device side.

[0043] In addition, network device 120 can also request auxiliary information from terminal device 110 for updating, retraining or fine-tuning the model on the network device 120 side.

[0044] like Figure 2BAs shown, in process 200B, network device 120 first sends control signaling to terminal device 110. This control signaling configures performance monitoring of the ML-based CSI compression model. Network device 120 sends a CSI-RS to terminal device 110. Terminal device 110 performs raw CSI calculation based on the CSI-RS to obtain the raw CSI, and then converts the raw CSI into compressed CSI based on the encoder. Terminal device 110 sends a CSI report including the compressed CSI to network device 120. Network device 120 decodes the compressed CSI to obtain the CSI decompression value and returns the decompression value to terminal device 110. Terminal device 110 calculates the SGCS based on the raw CSI and the decompressed CSI value. Subsequently, terminal device 110 feeds back the monitoring results to network device 120. Therefore, by calculating the SGCS on the terminal device 110 side, performance monitoring and management of the CSI feedback decompression model can be achieved on the terminal device 110 side, and a model performance reference can be provided to the network device side.

[0045] like Figure 2C As shown, in process 200C, network device 120 first sends control signaling to terminal device 110. This control signaling configures performance monitoring of the ML-based CSI compression model. Network device 120 sends a CSI-RS to terminal device 110. Terminal device 110 performs measurements based on the CSI-RS to obtain the raw CSI, and converts the raw CSI into compressed CSI based on an encoder. Terminal device 110 sends a CSI report including the compressed CSI to network device 120. Network device 120 decompresses the compressed CSI to obtain the decompressed CSI value, and returns the decompressed CSI value to terminal device 110 via a precoded reference signal (precoded RS). Then, terminal device 110 calculates the SGCS based on the raw CSI and the decompressed CSI value to obtain the model performance index, and feeds back the monitoring results to network device 120. Therefore, by performing SGCS calculation on the terminal device 110 side and combining it with the precoded reference signal, performance monitoring of the CSI compression and decompression process can be achieved on the terminal device side, providing a model performance reference for the network device side.

[0046] like Figure 2DAs shown, in process 200D, network device 120 first sends control signaling to terminal device 110. This control signaling configures performance monitoring of the ML-based CSI compression model and simultaneously sends CSI-RS to terminal device 110. Terminal device 110 performs measurements based on CSI-RS to obtain the true value of the CSI and generates compressed CSI using an encoder. Terminal device 110 decompresses the compressed CSI using a locally deployed proxy decoder to obtain the decompressed CSI value. This decompressed CSI value and the true value of the CSI are then input into the SGCS estimation model to obtain the SGCS, and the monitoring results are fed back to network device 120. Therefore, by deploying a proxy decoder and performing SGCS estimation on the terminal device side, local performance monitoring of the CSI compression model can be achieved on the terminal device side, thereby assisting network device 120 in model performance management.

[0047] like Figure 2E As shown, in process 200E, network device 120 first sends control signaling to terminal device 110. This control signaling configures performance monitoring of the ML-based CSI compression model and simultaneously sends CSI-RS to terminal device 110. Terminal device 110 performs measurements based on CSI-RS to obtain the CSI and generates compressed CSI through an encoder. Simultaneously, terminal device 110 inputs the compressed CSI and the original CSI into a locally deployed SGCS estimation model to obtain the SGCS (i.e., the SGCS estimate), and feeds the monitoring results back to network device 120. Thus, by deploying the SGCS estimation model on the terminal device 110 side, performance monitoring of the CSI compression model is achieved, thereby assisting network device 120 in model performance management.

[0048] To address at least the aforementioned problems and other potential related issues, this application proposes a communication method. The method includes: a terminal device receiving a first signaling message from a network device, the first signaling message being used to configure performance monitoring of a first CSI feedback compression model, the first signaling message including at least a triggering mode for a second signaling message; determining a first estimated value of the performance index of the first CSI feedback compression model based on the first signaling message and a performance index estimation model; and sending a second signaling message to the network device based on the triggering mode, requesting the network device to perform performance monitoring of the first CSI feedback compression model. This method can achieve effective model performance monitoring while reducing complexity and saving air interface overhead.

[0049] Figure 3 A signaling interaction diagram of a communication process 300 according to some possible implementations of this application is shown. For illustrative purposes, reference will be made to... Figure 1 The elements shown are used to describe process 300. However, it should be understood that process 300 can also be executed between network devices and terminal devices in any other communication scenario.

[0050] like Figure 3 As shown, in communication process 300, at 301, terminal device 110 receives a first signaling from network device 120. The first signaling is used to configure performance monitoring of a first CSI feedback compression model. The first signaling includes at least the triggering mode of a second signaling.

[0051] In one possible implementation, when terminal device 110 completes initial access or needs to perform RRC reconfiguration, network device 120 can send a first signaling to terminal device 110 to configure the performance monitoring of the CSI feedback compression model.

[0052] In one possible implementation, while network device 120 configures CSI-RS and CSI feedback at terminal device 110, network device 120 can simultaneously configure performance monitoring of the CSI feedback compression model at terminal device 110. Network device 120 can send the first signaling when sending CSI feedback configuration information.

[0053] In one possible implementation, the triggering mode of the second signaling may include a threshold triggering mode. In the threshold triggering mode, if the terminal device 110 detects a first estimated value of a performance indicator as lower than a first threshold in a single instance or a consecutive first count, then the second signaling is sent.

[0054] Alternatively, in one possible implementation, the triggering mode of the second signaling may include a periodic triggering mode. In the periodic triggering mode, the terminal device 110 sends the second signaling at predetermined intervals.

[0055] Alternatively, in one possible implementation, the triggering mode of the second signaling may include a hybrid triggering mode. In the hybrid triggering mode, if the terminal device 110 detects that the first estimated value of the performance indicator is lower than the first threshold once or multiple times consecutively within a predetermined period, then the second signaling is sent; if the terminal device 110 does not detect that the first estimated value of the performance indicator is lower than the first threshold within the predetermined period, then the second signaling is sent when the predetermined period expires, and the predetermined period is reset after the second signaling is sent.

[0056] In one possible implementation, the three different triggering modes of the second signaling can be set for different monitoring scenarios.

[0057] Therefore, by selecting different triggering modes for the second signaling for different monitoring scenarios, flexible monitoring of the performance of CSI-related models deployed locally on terminal device 110 can be achieved, and a balance can be struck between air interface overhead, monitoring latency, and stability for different devices and network environments.

[0058] The implementation method of the triggering mode for the second signaling will be referred to below. Figure 6A , 6B It is described in conjunction with 6C.

[0059] At 303, terminal device 110 determines a first estimated value of the performance index of the first CSI feedback compression model based on the first signaling and performance index estimation model.

[0060] At 305, terminal device 110 sends a second signaling message to network device 120 based on trigger mode to request network device 120 to perform performance monitoring of the first CSI feedback compression model.

[0061] In this way, terminal device 110 can adopt different second signaling trigger modes in different scenarios according to the instructions of network device 120, thereby realizing flexible monitoring of the performance of CSI feedback compression model.

[0062] The following sections will describe some possible implementations of this application.

[0063] Figure 4 A signaling interaction diagram of a communication process 400 according to one possible implementation of this application is shown. Process 400 can be understood as an example implementation of process 300. For purposes of discussion, reference will be made to... Figure 1 The elements shown in the diagram describe process 400.

[0064] Overall, process 400 involves first-stage performance monitoring and second-stage performance monitoring of the model at terminal device 110. In the first-stage performance monitoring, terminal device 110 calculates or estimates the performance metrics of the first CSI feedback compression model and the SGCS estimation model. Network device 120 performs second-stage performance monitoring of the first CSI feedback compression model after receiving the second signaling.

[0065] like Figure 4 As shown, network device 120 sends a 401 first signaling message to terminal device 110. The first signaling message is used to configure performance monitoring of the first CSI feedback compression model. The first signaling message includes at least the triggering mode of the second signaling message.

[0066] In one possible implementation, the first signaling may further include at least one of the following: an identifier of the CSI feedback decompression model at network device 120, an identifier of the first CSI feedback compression model at terminal device 110, a first indication, and a performance monitoring period for the first CSI feedback compression model (also referred to as the "user-side performance monitoring period," determined by T). moni_UE (represented by P), the first threshold for the performance metrics of the first CSI feedback compression model (by P) thr_1 (represented), the first number (by N) con_UE(indicated by T), the transmission period or maximum interval of the second signaling (by T). m2 express).

[0067] In one possible implementation, the first indication is used to indicate whether to activate or deactivate the terminal device 110 for performance monitoring of the first CSI feedback compression model.

[0068] In one possible implementation, if the terminal device 110 performs a single or consecutive first time (i.e., N) con_UE (time) The first estimated value of the performance index of the first CSI feedback compression model was detected to be lower than the first threshold (P). thr_1 If the threshold (P) is reached, then the performance degradation of the first CSI feedback compression model is determined. Therefore, in the following text, the first threshold (P) is used to determine the performance degradation of the first CSI feedback compression model. thr_1 This is also known as the "user-side performance degradation threshold," the first number (N) con_UE This is also known as the "continuous error trigger threshold".

[0069] In one possible implementation, the transmission period or maximum interval of the second signaling is used. When the transmission mode of the second signaling is a periodic trigger mode, the transmission period or maximum interval of the second signaling is used to configure the transmission period of the second signaling; when the transmission mode of the second signaling is a mixed trigger mode, it is used to configure the maximum transmission interval of the second signaling.

[0070] Network device 120 sends 403 CSI-RS to terminal device 110.

[0071] Subsequently, terminal device 110 can perform first-stage performance monitoring of the model at terminal device 110 based on the first signaling and the received CSI-RS. The first-stage performance monitoring may involve actions 405 to 411.

[0072] In one possible implementation, terminal device 110 may activate or deactivate performance monitoring of the first CSI feedback compression model (i.e., first-stage performance monitoring) based on a first instruction.

[0073] After activation, the terminal device 110 can periodically perform the first-stage performance monitoring of the first CSI feedback compression model.

[0074] Figure 5A The first-stage performance monitoring period (i.e., the "user-side performance monitoring period" T) of the first CSI feedback compression model is shown. moni_UE Examples of ).

[0075] In one possible implementation, network device 120 may send CSI-RS to terminal device 110 according to a first cycle, and terminal device 110 may perform first-phase performance monitoring according to a user-side performance monitoring cycle. The user-side performance monitoring cycle may be an integer multiple of the first cycle, such as three times.

[0076] In one possible implementation, the user-side performance monitoring period can be determined by the network device 120. The first signaling sent by the network device 120 to the terminal device 110 can indicate the user-side performance monitoring period. The network device 120 can configure the user-side performance monitoring period according to its own processing capabilities, the reporting capabilities of the terminal device 120, and channel conditions. For example, when the user-side performance monitoring period is longer, the computational overhead of the terminal device 110 can be reduced, but the corresponding performance monitoring latency will increase; when the user-side performance monitoring period is shorter, low-latency performance monitoring can be achieved, but the computational overhead of the terminal device 110 will increase relatively.

[0077] Return to Figure 4 The terminal device 110 performs CSI calculation based on the received CSI-RS to determine the true value of 405 CSI.

[0078] Terminal device 110 determines the compression value of 407 CSI based on the first CSI feedback compression model and the actual value of CSI.

[0079] In one possible implementation, the first CSI feedback compression model can be implemented as an encoder. In this possible implementation, the terminal device 110 can convert the actual CSI value into a compressed CSI value using the encoder.

[0080] Terminal device 110 calculates or estimates the performance index of the first CSI feedback compression model 409 based on the true value of CSI, the compressed value of CSI, and the performance index estimation model.

[0081] In one possible implementation, the performance metric of the first CSI feedback compression model may include generalized cosine similarity (GCS). In this possible implementation, the performance metric estimation model can be a GCS estimation model.

[0082] Generalized cosine similarity (GCS) is a classic metric for measuring the directional similarity between two vectors. It is defined as the ratio of the dot product of the vectors to the product of their magnitudes, expressed by the formula:

[0083] Where A and B represent vectors (e.g., ... Figure 5B(As shown). The value range of GCS is between [-1, 1], where 1 indicates completely in the same direction, -1 indicates completely in opposite directions, and 0 indicates that the vectors are orthogonal. The core of GCS is that it eliminates the influence of vector length and purely reflects the consistency of direction. In this application, A and B can be the true value and compressed value of CSI represented by the eigenvectors of the channel matrix, respectively, and GCS can represent the similarity of the vector directions of the true value and the compressed value of CSI.

[0084] In another possible implementation, the performance metric of the first CSI feedback compression model may include squared generalized cosine similarity (SGCS). In this implementation, the performance metric estimation model can be an SGCS estimation model. The performance metric of the first CSI feedback compression model will be described below using SGCS as an example.

[0085] SGCS is obtained by squaring GCS, and its expression is:

[0086] At this point, the range of SGCS values ​​is limited to the [0,1] interval, which brings three important characteristics: First, the squaring operation amplifies the discriminative power of high similarity values, making similarity vectors more prominent; second, it eliminates negative correlation information, making it suitable for scenarios where only positive correlations need to be considered; and finally, mathematically, it is equivalent to the square ratio of the projection lengths, and is directly related to energy calculation. These characteristics make SGCS particularly suitable for application scenarios that require strengthening salient features.

[0087] In practical applications, SGCS exhibits unique advantages over GCS. During feature selection, SGCS automatically strengthens the weights of important features through squaring operations, achieving more effective feature filtering. In machine learning techniques, SGCS is often used as a fundamental component for constructing nonlinear kernel functions. It is worth noting that the choice of SGCS requires a trade-off: while it enhances the ability to distinguish high similarity, it also loses directional information, which may lead to information loss in scenarios requiring the differentiation of positive and negative correlations. Therefore, a reasonable choice between GCS and SGCS needs to be made based on the specific requirements of the problem.

[0088] Terminal device 110 can send 411 second signaling to network device 120 according to the triggering mode of the second signaling indicated in the first signaling.

[0089] Network device 120 performs Phase 2 performance monitoring based on the second signaling. Possible implementation methods for Phase 2 performance monitoring will be referenced below. Figure 7B Describe it.

[0090] After performing the second phase of performance monitoring, network device 120 can send 415 third signaling to terminal device 110. Possible implementations of the third signaling will be referenced below. Figure 7B Describe it.

[0091] Furthermore, terminal device 110 can perform 417 model management based on performance monitoring results. This will be referenced below. Figure 7B Describe it.

[0092] In the following text, reference will be made to Figure 6A , 6B The second signaling triggering mode is described in section 6C.

[0093] In one possible implementation, the triggering mode of the second signaling may include a threshold triggering mode.

[0094] In threshold-triggered mode, if terminal device 110 detects a first estimated value of a performance metric below a first threshold in a single instance or for the first consecutive number of times, a second signaling is sent. For example, if terminal device 110 detects a first estimated value of a performance metric below a first threshold in a single instance or for the first consecutive N times... con_UE The calculated or estimated model performance metric SGCS was found to be below the performance degradation threshold P. thr_1 Then, a second signaling message is sent to network device 120 to request network device 120 to perform the second phase of performance monitoring.

[0095] The SGCS estimate may experience a single or brief decrease under special circumstances, such as when the channel environment changes drastically. Therefore, network device 120 can adjust the parameter N. con_UE To control terminal device 110 when it detects continuous N con_UE The sub-performance index is below the threshold P. thr_1 Then a second signaling message is sent, thus avoiding frequent retransmission of the second signaling message due to a single or brief drop in the indicator.

[0096] Figure 6A A schematic diagram of a threshold-triggered mode for second signaling according to one possible implementation of this application is shown. Figure 6A As shown, if the terminal device 110 detects three consecutive times that the calculated or estimated model performance index SGCS is lower than the performance degradation threshold P, thr_1 Then, a second signaling message is sent to network device 120 to request network device 120 to perform the second phase of performance monitoring.

[0097] Alternatively, in one possible implementation, the triggering mode of the second signaling may include a periodic triggering mode. In the periodic triggering mode, the terminal device 110 sends the second signaling at predetermined intervals.

[0098] In one possible implementation, network device 120 may configure the second signaling transmission period to terminal device 110.

[0099] In one possible implementation, the primary purpose of configuring the periodic triggering mode for network device 120 may be to periodically verify the performance of the user-side model. Alternatively or additionally, the primary purpose of configuring the periodic triggering mode for network device 120 may be to enable terminal device 110 to periodically send the actual value and compressed value of CSI to network device 120, and to monitor the performance of the model on the network device 120 side by multiplexing this signaling.

[0100] Figure 6B A schematic diagram of a periodic triggering mode for second signaling according to one possible implementation of this application is shown. Figure 6B As shown, the terminal device 110 sends the second signaling every 5 seconds. That is, if the terminal device 110 detects that the performance indicators of the first CSI feedback compression model are normal for 5 consecutive times, it sends the second signaling to the network device 120.

[0101] Alternatively, in one possible implementation, the triggering mode of the second signaling may include a hybrid triggering mode. In the hybrid triggering mode, if the terminal device 110 detects a performance indicator lower than the first threshold once or multiple times consecutively within a predetermined period, it sends the second signaling; if the terminal device 110 does not detect a performance indicator lower than the first threshold within the predetermined period, it sends the second signaling when the predetermined period expires, and resets the predetermined period after sending the second signaling.

[0102] The hybrid triggering mode combines the features of the two triggering modes mentioned above. By setting the period (i.e., the maximum interval) for sending the second signaling, network device 120 can periodically perform network-side model monitoring and verify the monitoring performed by terminal device 110. Simultaneously, terminal device 110 can also trigger the sending of the second signaling through a threshold, requesting network device 120 to perform a second-stage performance monitoring and obtain an indication of the erroneous model when model performance degradation is detected. Terminal device 110 resets the sending period after sending the second signaling.

[0103] Figure 6C A schematic diagram of a hybrid triggering mode for second signaling according to one possible implementation of this application is shown. Figure 6C As shown, the period (i.e., maximum interval) for terminal device 110 to send the second signaling is 5. If terminal device 110 detects twice consecutively that the calculated or estimated model performance index SGCS is lower than the performance degradation threshold P... thr_1 If the terminal device 110 sends a second signaling message to network device 120, it requests network device 120 to perform the second phase of performance monitoring. After sending the second signaling message, terminal device 110 resets its transmission cycle.

[0104] If terminal device 110 detects that the performance indicators of the first CSI feedback compression model are normal five times consecutively, it sends a second signaling message to network device 120. After sending the second signaling message, terminal device 110 resets the transmission cycle.

[0105] Figure 7A A signaling interaction diagram of a communication process 700 according to one possible implementation of this application is shown. Process 700A can be understood as an example implementation of process 300. For purposes of discussion, reference will be made to... Figure 1 The elements shown in the diagram describe process 700A.

[0106] Overall, process 700A involves the first phase of model performance monitoring at terminal device 110, but not the second phase. After network device 120 sends the first signaling to configure and activate model performance monitoring at terminal device 110, terminal device 110 enters the first phase of model performance monitoring. When terminal device 110 detects that the model performance metric SGCS is greater than the decay threshold or fails to trigger the transmission of the second signaling, terminal device 110 does not need to send the second signaling for performance monitoring to network device 120, but instead sends CSI feedback.

[0107] Specifically, actions 701, 703, 705, 707, and 709 in process 700A are similar to actions 401, 403, 405, 404, and 409 in process 400, and therefore will not be described again. The difference between process 700A and process 400 lies in actions 711 and 713.

[0108] like Figure 7A As shown, if the terminal device 110 detects that the first estimated value of the performance index of the first CSI feedback compression model is higher than the first threshold (P... thr_1 If the second signaling is not triggered, the terminal device 110 may send a 711 CSI feedback to the network device 120. The CSI feedback may include the compressed value of the CSI.

[0109] Subsequently, network device 120 determines the 713 CSI output value based on the CSI feedback decompression model and the CSI compression value.

[0110] In one possible implementation, the CSI feedback decompression model can be implemented as a decompressor. Network device 120 can use the decompressor to decompress the compressed CSI value to obtain the CSI output value.

[0111] In another possible implementation, the CSI feedback decompression model can be implemented as a decoder. Network device 120 can use the decoder to convert the compressed CSI value into a CSI output value.

[0112] Figure 7B A signaling interaction diagram of a communication process 700 according to one possible implementation of this application is shown. Process 700B can be understood as an example implementation of process 300. For purposes of discussion, reference will be made to... Figure 1 The elements shown in the diagram describe process 700B.

[0113] In general, process 700B involves first-stage and second-stage performance monitoring of the model at terminal device 110. After network device 120 sends a first signaling instruction to configure and activate model performance monitoring at terminal device 110, terminal device 110 enters the first-stage performance monitoring phase. When terminal device 110 detects that the model performance metric SGCS is below the decay threshold or triggers the transmission of a second signaling instruction, terminal device 110 sends a second signaling instruction to network device 120 for performance monitoring.

[0114] Specifically, actions 721, 723, 725, 727, 729, 731, and 741 in process 700B are similar to actions 401, 403, 405, 407, 409, 411, and 417 in process 400, and therefore will not be described again. The difference between process 700B and process 400 lies in actions 733, 735, 737, and 739.

[0115] like Figure 7B As shown, if the terminal device 110 detects that the first estimated value of the performance index of the first CSI feedback compression model is lower than the first threshold (P... thr_1 When the second signaling is triggered, the terminal device 110 may send the second signaling 731 to the network device 120.

[0116] In one possible implementation, the second signaling may include at least one of the following: an identifier of the first CSI feedback compression model at terminal device 110, an identifier of the performance indicator estimation model at terminal device 110, the true value of CSI, the compressed value of CSI, and a second indication. The true value of CSI is, for example, the true value of CSI determined by terminal device 110 through action 725. The compressed value of CSI is, for example, the compressed value of CSI determined by terminal device 110 through action 727. The second indication is used to instruct network device 120 to activate performance monitoring of the first CSI feedback compression model.

[0117] Subsequently, network device 120 determines the 733 CSI output value based on the CSI feedback decompression model and the CSI compression value.

[0118] In one possible implementation, the CSI feedback decompression model can be implemented as a decompressor. Network device 120 can use the decompressor to decompress the compressed CSI value to obtain the CSI output value.

[0119] In another possible implementation, the CSI feedback decompression model can be implemented as a decoder. Network device 120 can use the decoder to convert the compressed CSI value into a CSI output value.

[0120] Furthermore, network device 120 can calculate a second estimate of the 735 performance index SGCS based on the output value of CSI and the actual value of CSI.

[0121] Furthermore, network device 120 can base its performance threshold P on the second estimate of SGCS and the CSI feedback decompression model. thr_2 The comparison was used to determine the model that caused the performance degradation.

[0122] For example, if the second estimate of SGCS is greater than the performance threshold P of the CSI feedback decompression model at network device 120 thr_2 That is, network device 120 determines that the performance of the CSI feedback compression model and the CSI feedback decompression model are normal, while the performance of the SGCS estimation model at terminal device 110 degrades (i.e., is abnormal).

[0123] If the second estimate of SGCS is less than the performance threshold P of the CSI feedback decompression model at network device 120 thr_2 That is, network device 120 determined that the performance of both the CSI feedback compression model and the CSI feedback decompression model degraded.

[0124] Furthermore, network device 120 sends 739 third signaling to terminal device 110.

[0125] In one possible implementation, the third signaling may include at least one of the following: an identifier of the first CSI feedback compression model, an identifier of the performance index estimation model (e.g., the SGCS estimation model), a second estimate of the performance index of the first CSI feedback compression model, an identifier of the model that causes performance degradation, and instructions for managing the first CSI feedback compression model.

[0126] In one possible implementation, the model that causes performance degradation may include one of the following: a first CSI feedback compression model, a performance metric estimation model, and a CSI feedback decompression model at network device 120.

[0127] After receiving the third signaling, terminal device 110 can perform model management operations based on the identifier of the model causing performance degradation indicated by network device 120 and the instructions for managing the model causing performance degradation, thereby restoring network performance. For example, terminal device 110 can update, switch, or disable the model causing performance degradation.

[0128] Figure 8 A signaling interaction diagram of a communication process 800 according to another possible implementation of this application is shown. Process 800 can be understood as an example implementation of process 400. For purposes of discussion, reference will be made to... Figure 1 The elements shown in the diagram describe process 800.

[0129] Overall, process 800 involves first-stage and second-stage performance monitoring of the model at terminal device 110, and also involves performance monitoring of the model at network device 120. Upon receiving a second signaling message from terminal device 110, network device 120 performs second-stage performance monitoring while simultaneously multiplexing the second signaling message for performance monitoring of the model at network device 120.

[0130] Specifically, actions 801, 803, 805, 807, 809, 811, 821, and 823 in process 800 are similar to actions 721, 723, 725, 727, 729, 731, 739, and 741 in process 700B, and therefore will not be described in detail. The difference between process 800 and process 700B lies in actions 813, 815, 817, and 819.

[0131] like Figure 8 As shown, after receiving the second signaling from the terminal device 110, the network device 120 determines the compressed value of 813 CSI based on the true value of CSI in the second signaling and the second CSI feedback compression model.

[0132] Furthermore, the network device 120 determines the 815 CSI output value (i.e., the CSI decompression value) based on the CSI feedback decompression model deployed on the network device 120 and the CSI compression value.

[0133] Furthermore, based on the CSI output value and the actual CSI value, network device 120 calculates a second estimate of the performance metric (e.g., SGCS) of the CSI feedback decompression model at network device 120.

[0134] Furthermore, the second estimate of network device 120 based on SGCS is compared with the performance metric threshold P. thr_3 The comparison was used to determine whether the performance of the CSI feedback decompression model at point 120 of the 819 network device had degraded.

[0135] For example, when the second estimate of SGCS is greater than the performance metric threshold P thr_3 At that time, network device 120 determined that the performance of the CSI feedback decompression model was normal; when the second estimate of SGCS was less than P thr_3At that time, network device 120 determines the performance degradation of the CSI feedback decompression model. Performance monitoring of the model at network device 120 is achieved through multiplexing of the second signaling. If performance degradation occurs in the model at network device 120, network device 120 can perform subsequent model management operations and, if so, indicate the performance degradation of the model at network device 120 to terminal device 110 via the identifier of the degraded model in the third signaling.

[0136] Figure 9 A flowchart of a communication method 900 according to one possible implementation of this application is shown. In one possible implementation, method 900 may be implemented by a terminal device 110 in the communication system 100, for example, by a processor or processing unit of the terminal device 110 in conjunction with other components (e.g., a transceiver).

[0137] At 910, terminal device 110 receives a first signaling, which is used to configure the performance monitoring of the first CSI feedback compression model. The first signaling includes at least the triggering mode of the second signaling.

[0138] At 920, terminal device 110 determines a first estimated value of the performance index of the first CSI feedback compression model based on the first signaling and performance index estimation model.

[0139] At 930, terminal device 110 sends a second signaling message to request network device to perform performance monitoring of the first CSI feedback compression model.

[0140] In one possible implementation, the triggering mode of the second signaling includes one of the following: a threshold triggering mode, wherein the second signaling is sent if the terminal device detects a first estimated value of the performance indicator as lower than a first threshold once or consecutively for the first time; a periodic triggering mode, wherein the terminal device sends the second signaling at a predetermined period; and a hybrid triggering mode, wherein the second signaling is sent if the terminal device detects a first estimated value of the performance indicator as lower than the first threshold once or consecutively multiple times within the predetermined period, and if the terminal device does not detect a first estimated value of the performance indicator as lower than the first threshold within the predetermined period, the second signaling is sent at the end of the predetermined period, and the predetermined period is reset after the second signaling is sent.

[0141] In one possible implementation, the first signaling further includes at least one of the following: an identifier of a first CSI feedback compression model, an identifier of a CSI feedback decompression model at the network device, a first indication indicating whether to activate or deactivate the terminal device's performance monitoring of the first CSI feedback compression model, a period for performance monitoring of the first CSI feedback compression model, a first threshold for performance indicators, a first count, and a transmission period for the second signaling.

[0142] In one possible implementation, the second signaling includes at least one of the following: an identifier of a first CSI feedback compression model, an identifier of a performance index estimation model, the true value of CSI, the compressed value of CSI, and a second indication that instructs the network device to activate performance monitoring of the first CSI feedback compression model.

[0143] In one possible implementation, method 900 further includes: receiving third signaling from a network device, the third signaling including at least one of the following: an identifier of a first CSI feedback compression model, an identifier of a performance metric estimation model, a second estimate of the performance metric of the first CSI feedback compression model, an identifier of a model causing performance degradation, the model causing performance degradation including one of the following: the first CSI feedback compression model, the performance metric estimation model and the CSI feedback decompression model at the network device, and instructions for managing the model causing performance degradation.

[0144] In some embodiments, method 900 may further include combining the embodiments of this disclosure. Figures 3 to 8 Other operations performed at terminal device 110 as described in one or more examples.

[0145] Figure 10 A flowchart of a communication method 1000 according to one possible implementation of this application is shown. In one possible implementation, method 1000 may be implemented by a network device 120 in a communication system 100, for example, by a processor or processing unit of the network device 120 in conjunction with other components (e.g., a transceiver).

[0146] At 1010, network device 120 sends a first signaling message, which is used to configure performance monitoring of a first CSI feedback compression model at the terminal device. The first signaling message includes at least the triggering mode of the second signaling message.

[0147] At 1020, network device 120 receives a second signaling message, which is used to request the network device to perform performance monitoring of the first CSI feedback compression model.

[0148] In one possible implementation, the triggering mode of the second signaling includes one of the following: a threshold triggering mode, wherein the second signaling is sent if the terminal device detects a first estimated value of the performance indicator as lower than a first threshold once or consecutively for the first time; a periodic triggering mode, wherein the terminal device sends the second signaling at a predetermined period; and a hybrid triggering mode, wherein the second signaling is sent if the terminal device detects a first estimated value of the performance indicator as lower than the first threshold once or consecutively multiple times within the predetermined period, and if the terminal device does not detect a first estimated value of the performance indicator as lower than the first threshold within the predetermined period, the second signaling is sent at the end of the predetermined period, and the predetermined period is reset after the second signaling is sent.

[0149] In one possible implementation, the first signaling further includes at least one of the following: an identifier of a first CSI feedback compression model, an identifier of a CSI feedback decompression model at the network device, a first indication indicating whether to activate or deactivate the terminal device's performance monitoring of the first CSI feedback compression model, a period for performance monitoring of the first CSI feedback compression model, a first threshold for performance indicators, a first count, and a transmission period for the second signaling.

[0150] In one possible implementation, the second signaling includes at least one of the following: an identifier of a first CSI feedback compression model, an identifier of a performance index estimation model, the true value of CSI, the compressed value of CSI, and a second indication that instructs the network device to activate performance monitoring of the first CSI feedback compression model.

[0151] In one possible implementation, method 1000 further includes: receiving the true value of CSI from a terminal device; determining the compressed value of CSI based on the true value of CSI and a second CSI feedback compression model; determining the CSI output value based on the compressed value of CSI and a CSI feedback decompression model; and determining a second estimate of the performance index of the first CSI feedback compression model based on the CSI output value and the true value of CSI.

[0152] In one possible implementation, method 1000 further includes: sending a third signaling to the terminal device, the third signaling including at least one of the following: an identifier of a first CSI feedback compression model, an identifier of a performance metric estimation model, a second estimate of the performance metric of the first CSI feedback compression model, an identifier of a model causing performance degradation, the model causing performance degradation including one of the following: the first CSI feedback compression model, the performance metric estimation model and the CSI feedback decompression model at the network device, and instructions for managing the model causing performance degradation.

[0153] In some embodiments, method 1000 may further include combining the embodiments of this disclosure. Figures 3 to 8 Other operations performed at network device 120 as described in one or more examples.

[0154] Figure 11 A schematic block diagram of a communication device 1100 according to one possible implementation of this application is shown. The communication device 1100 can be implemented as a device or a chip within a device, and the scope of this application is not limited in this respect. The communication device 1100 may include multiple units for performing functions such as... Figure 9 The process discussed in section 900. Communication device 1100 can be implemented as follows: Figure 1 The terminal device 110 or a part of the terminal device 110 shown in the figure.

[0155] like Figure 11 As shown, the communication device 1100 includes a transceiver unit 1102 and a determination unit 1104.

[0156] The transceiver unit 1102 is configured to receive a first signaling, which is used to configure the performance monitoring of the first CSI feedback compression model, and the first signaling includes at least the triggering mode of the second signaling.

[0157] The determining unit 1104 is configured to: determine a first estimated value of the performance index of the first CSI feedback compression model based on the first signaling and the performance index estimation model.

[0158] The transceiver unit 1102 is also configured to send a second signaling message to request the network device to perform performance monitoring of the first CSI feedback compression model.

[0159] In some other embodiments, apparatus 1100 may include various other units or modules that can be configured to perform the various operations or functions described with respect to the foregoing method embodiments. Specific details can be obtained by referring to the detailed description of the foregoing method embodiments, and will not be repeated here.

[0160] Figure 12 A schematic block diagram of a communication device 1200 according to one possible implementation of this application is shown. The communication device 1200 can be implemented as a device or a chip within a device, and the scope of this application is not limited in this respect. The communication device 1200 may include multiple units for performing functions such as... Figure 10 The process 1000 discussed herein. The communication device 1200 can be implemented as follows: Figure 1 The network device 120 or a part of the network device 120 shown in the figure.

[0161] like Figure 12 As shown, the communication device 1200 includes a transceiver unit 1202.

[0162] The transceiver unit 1202 is configured to send a first signaling, which is used to configure the performance monitoring of a first CSI feedback compression model at the terminal device, and the first signaling includes at least the triggering mode of the second signaling.

[0163] Transceiver unit 1202 is also configured to receive a second signaling message, which is used to request the network device to perform performance monitoring of the first CSI feedback compression model.

[0164] In some other embodiments, the apparatus 1200 may include various other units or modules that can be configured to perform the various operations or functions described with respect to the foregoing method embodiments. Specific details can be obtained by referring to the detailed description of the foregoing method embodiments, and will not be repeated here.

[0165] Figure 13 This is a simplified block diagram of an example device 1300 suitable for implementing possible implementations of this application. Device 1300 can be used to implement, for example... Figure 1 The terminal device 110 or network device 120 shown. As shown, device 1300 includes one or more processors (or processing units) 1310, and may also include one or more memories 1320 coupled to processor 1310, and may also include a communication interface 1340 coupled to processor 1310.

[0166] Communication interface 1340 can be used to communicate with other devices or apparatuses, such as transmitting or receiving data and / or signals. Communication interface 1340 may have at least one communication interface for communication. The communication interface may include any interface necessary for communicating with other devices. Exemplarily, the communication interface may be a transceiver, circuit, bus, module, or other type of communication interface.

[0167] Processor 1310 may include, but is not limited to, one or more of the following: a general-purpose computer, a special-purpose computer, a microcontroller, a digital signal processor (DSP), or a controller-based multi-core controller architecture. Device 1300 may have multiple processors, such as application-specific integrated circuit chips, which are time-dependent on a clock synchronized with the main processor.

[0168] Memory 1320 may include one or more non-volatile memories and one or more volatile memories. Examples of non-volatile memories include, but are not limited to, at least one of the following: Read-Only Memory (ROM) 1324, Electrically Programmable Read-Only Memory (EPROM), flash memory, hard disk, Compact Disc (CD), Digital Video Disk (DVD), or other magnetic and / or optical storage. Examples of volatile memories include, but are not limited to, at least one of the following: Random Access Memory (RAM) 1322, or other volatile memories that do not persist during power-off periods.

[0169] Computer program 1330 includes computer-executable instructions that are executed by associated processor 1310. Program 1330 may be stored in ROM 1320. Processor 1310 may perform any suitable actions and processes by loading program 1330 into RAM 1320.

[0170] The possible implementations of this application can be achieved using program 1330, enabling device 1300 to perform the various operations or functions described in the foregoing method embodiments. The possible implementations of this application can also be achieved through hardware or a combination of software and hardware.

[0171] In some implementations, program 1330 may be tangibly contained in a computer-readable medium, which may include in device 1300 (such as in memory 1320) or other storage devices accessible by device 1300. Program 1330 may be loaded from the computer-readable medium into RAM 1322 for execution. The computer-readable medium may include any type of tangible non-volatile memory, such as ROM, EPROM, flash memory, hard disk, CD, DVD, etc.

[0172] This application also provides a communication system. The communication system includes the communication device 1100 and the communication device 1200 as described above.

[0173] Generally, various possible implementations of this application can be implemented in hardware or special-purpose circuitry, software, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software, which can be executed by a controller, microprocessor, or other computing device. Although various aspects of possible implementations of this application are shown and described as block diagrams, flowcharts, or represented using some other illustrations, it should be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented as, as non-limiting examples, in hardware, software, firmware, special-purpose circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0174] This application also provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer-executable instructions, such as instructions included in program modules, which execute in a device on a target's real or virtual processor to perform the above-mentioned... Figures 3 to 10 The process / method described herein. Typically, a program module includes routines, programs, libraries, objects, classes, components, data structures, etc., that perform specific tasks or implement specific abstract data types. In various possible implementations, the functionality of program modules can be combined or divided as needed. The machine-executable instructions used in a program module can execute locally or on a distributed device. In a distributed device, program modules can reside on both local and remote storage media.

[0175] The program code used to implement the methods of this application may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0176] In the context of this application, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0177] Furthermore, although the operations are described in a specific order, this should be understood as requiring that such operations be performed in the specific order shown or in sequential order, or requiring that all illustrated operations be performed to achieve the desired result. In certain environments, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this application. Certain features described in the context of individual possible implementations can also be implemented in combination in a single implementation. Conversely, various features described in the context of a single implementation can also be implemented individually or in any suitable sub-combination in multiple implementations.

[0178] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A communication method, comprising: The terminal device receives a first signaling from the network device. The first signaling is used to configure the performance monitoring of the first CSI feedback compression model of channel state information. The first signaling includes at least the triggering mode of the second signaling. Based on the first signaling and performance index estimation model, a first estimated value of the performance index of the first CSI feedback compression model is determined. as well as Based on the triggering mode, the second signaling is sent to the network device to request the network device to perform performance monitoring of the first CSI feedback compression model.

2. The method according to claim 1, wherein the triggering mode of the second signaling includes one of the following: A threshold-triggered mode, wherein if the terminal device detects for the first time, in a single instance or consecutively, that the first estimated value of the performance indicator is lower than a first threshold, then the second signaling is sent. Periodic triggering mode, wherein the terminal device sends the second signaling at a predetermined period; and In a hybrid triggering mode, if the terminal device detects that the first estimated value of the performance indicator is lower than the first threshold once or multiple times within the predetermined period, it sends the second signaling; if the terminal device does not detect that the first estimated value of the performance indicator is lower than the first threshold within the predetermined period, it sends the second signaling when the predetermined period expires, and resets the predetermined period after sending the second signaling.

3. The method of claim 2, wherein the first signaling further comprises at least one of the following: The identifier of the first CSI feedback compression model, The identifier of the CSI feedback decompression model at the network device. The first instruction indicates whether to activate or deactivate the performance monitoring of the first CSI feedback compression model by the terminal device. The performance monitoring period of the first CSI feedback compression model. Regarding the first threshold of the performance metric, The first number, and The transmission period of the second signaling.

4. The method of claim 1, wherein the second signaling comprises at least one of the following: The identifier of the first CSI feedback compression model, The identifier of the performance index estimation model, The true value of CSI The compressed value of CSI, and The second instruction instructs the activation of the network device's performance monitoring of the first CSI feedback compression model.

5. The method according to claim 1, further comprising: Receive third signaling from the network device, the third signaling including at least one of the following: The identifier of the first CSI feedback compression model, The identifier of the performance index estimation model, The second estimate of the performance index of the first CSI feedback compression model. The identification of the model causing performance degradation, wherein the model causing performance degradation includes one of the following: the first CSI feedback compression model, the performance metric estimation model, and the CSI feedback decompression model at the network device, and Instructions for managing the model that causes performance degradation.

6. A communication method, comprising: The network device sends a first signaling message to the terminal device. The first signaling message is used to configure the performance monitoring of the first channel state information (CSI) feedback compression model at the terminal device. The first signaling message includes at least the triggering mode of the second signaling message. as well as The terminal device receives a second signaling message, which is used to request the network device to perform performance monitoring of the first CSI feedback compression model.

7. The method of claim 6, wherein the triggering mode of the second signaling includes one of the following: A threshold-triggered mode, wherein if the terminal device detects a first estimated value of the performance indicator as lower than a first threshold in a single instance or for the first consecutive time, the second signaling is sent. Periodic triggering mode, wherein the terminal device sends the second signaling at a predetermined period; and In a hybrid triggering mode, if the terminal device detects that the first estimated value of the performance indicator is lower than the first threshold once or multiple times within the predetermined period, it sends the second signaling; if the terminal device does not detect that the first estimated value of the performance indicator is lower than the first threshold within the predetermined period, it sends the second signaling when the predetermined period expires, and resets the predetermined period after sending the second signaling.

8. The method of claim 7, wherein the first signaling further comprises at least one of the following: The identifier of the first CSI feedback compression model, The identifier of the CSI feedback decompression model at the network device. The first instruction indicates whether to activate or deactivate the performance monitoring of the first CSI feedback compression model by the terminal device. The performance monitoring period of the first CSI feedback compression model. Regarding the first threshold of the performance metric, The first number, and The transmission period of the second signaling.

9. The method of claim 6, wherein the second signaling comprises at least one of the following: The identifier of the first CSI feedback compression model, The identifier of the performance index estimation model, The true value of CSI The compressed value of CSI, and The second instruction instructs the activation of the network device's performance monitoring of the first CSI feedback compression model.

10. The method of claim 6, further comprising: Receive the actual value of CSI from the terminal device; Based on the true value of CSI and the second CSI feedback compression model, the compressed value of CSI is determined; Based on the compressed value of CSI and the CSI feedback decompression model, the CSI output value is determined; as well as Based on the CSI output value and the actual CSI value, a second estimated value of the performance index of the first CSI feedback compression model is determined.

11. The method of claim 10, further comprising: Send a third signaling message to the terminal device, the third signaling message including at least one of the following: The identifier of the first CSI feedback compression model, The identifier of the performance index estimation model, The second estimate of the performance index of the first CSI feedback compression model. The identification of the model causing performance degradation, wherein the model causing performance degradation includes one of the following: the first CSI feedback compression model, the performance metric estimation model, and the CSI feedback decompression model at the network device, and Instructions for managing the model that causes performance degradation.

12. A communication device, comprising: Units or modules for performing the method according to any one of claims 1 to 5 or 6 to 11.

13. A communication device, comprising: The processor is configured to perform the method according to any one of claims 1 to 5 or claims 6 to 11.

14. A computer-readable storage medium storing a computer program or instructions that, when executed, cause the method according to any one of claims 1 to 5 or claims 6 to 11 to be performed.

15. A computer program product comprising a computer program or instructions, which, when executed, cause the method according to any one of claims 1 to 5 or claims 6 to 11 to be performed.

Images