Method for reporting channel state information and model management

Abstract

A method for reporting channel state information and model management is provided. In the method, a reporting configuration indicating a first physical resource set and a first reference signal set for reporting is received. A second reference signal set on a second physical resource set is received, the second physical resource set being a subset of the first physical resource set or the first reference signal set. First estimated channel information corresponding to the second reference signal set on the second physical resource set is determined. Second estimated channel information corresponding to the first reference signal set on the first physical resource set is determined by using the first estimated channel information on a prediction model. Channel state information associated with the second estimated channel information is reported. The prediction model is related to a mask ratio.

Description

Technical Field

[0001] This invention relates to a wireless communication system, and more specifically, to a method for reporting channel state information (CSI) and model management. Background Technology

[0002] In modern wireless communication systems, such as 5G New Radio (NR) and its follow-up technologies, accurate acquisition of Channel State Information (CSI) is crucial for enabling advanced multi-antenna technologies. Typically, network devices transmit Channel State Information Reference Signals (CSI-RS) to User Equipment (UE). The UE then measures these signals to estimate downlink channel characteristics and reports the estimated CSI back to the network device. These transmissions occur on a time-frequency grid, which consists of basic units called resource elements (REs), such as... Figure 1 As conceptually illustrated in the diagram, each RE represents a specific time and frequency resource within the communication channel.

[0003] Traditionally, the required number of CSI-RS resources is proportional to the number of transmit antenna ports on a network device, with each antenna port occupying one or more REs for its reference signal. As wireless communication evolves towards future versions (such as 6G), the number of antenna ports is expected to increase significantly to improve system capacity. This trend renders the traditional one-to-one mapping between antenna ports and CSI-RS resources unscalable. Transmitting a complete set of CSI-RS data for a large number of antenna ports would consume excessive REs, resulting in significant overhead and reducing the resources available for data transmission, thus limiting overall spectral efficiency. Summary of the Invention

[0004] This disclosure relates to methods for reporting channel state information and model management.

[0005] According to one or more example embodiments of this disclosure, a method for reporting channel state information performed by a user equipment (UE) is provided. The method includes: receiving a reporting configuration, wherein the reporting configuration indicates a first physical resource set and a first reference signal set for reporting; receiving a second reference signal set on a second physical resource set, wherein the second physical resource set is a subset of the first physical resource set or the second reference signal set is a subset of the first reference signal set; determining first estimated channel information corresponding to the second reference signal set on the second physical resource set; determining second estimated channel information corresponding to the first reference signal set on the first physical resource set by using the first estimated channel information on a prediction model, wherein the prediction model is a signal processing algorithm or trained by a machine learning algorithm; and reporting channel state information associated with the second estimated channel information.

[0006] According to one or more example embodiments of this disclosure, a method for predicting channel state information performed by a network device is provided. The method includes: transmitting a report configuration, wherein the report configuration indicates a second physical resource set and a second reference signal set for reporting; transmitting the second reference signal set on the second physical resource set; and receiving a channel state information report associated with first estimated channel information corresponding to the second reference signal set on the second physical resource set. The report is used on a prediction model to predict a second estimated channel information corresponding to the first reference signal set on the first physical resource set, wherein the second physical resource set is a subset of the first physical resource set or the second reference signal set is a subset of the first reference signal set, and the prediction model is a signal processing algorithm or trained via a machine learning algorithm.

[0007] According to one or more example embodiments of this disclosure, a model management method performed by a UE is provided. The method includes: associating a prediction model with a mask ratio, wherein the prediction model is trained by a signal processing algorithm or a machine learning algorithm, the prediction model recovers second estimated channel information corresponding to a first reference signal set having a first physical resource set by inputting first estimated channel information, the second reference signal set on the second physical resource set being used to determine the first estimated channel information, and the mask ratio being the ratio of a combination of the first reference signal set and the first physical resource set to a combination of the second reference signal set and the second physical resource set; and transmitting an indication of the prediction model. Attached Figure Description

[0008] Figure 1 This is a schematic diagram illustrating resource elements in a time-frequency grid according to an exemplary embodiment of this disclosure;

[0009] Figure 2 This is a schematic diagram illustrating a communication system based on an example embodiment of the present disclosure;

[0010] Figure 3 This is a flowchart illustrating a method for reporting channel state information performed by a user equipment (UE) according to an example embodiment of this disclosure;

[0011] Figure 4 This is a signaling diagram illustrating the channel state information reporting process;

[0012] Figure 5 This is a signaling diagram illustrating a channel state information reporting method with antenna domain reference signal reduction according to an example embodiment of this disclosure;

[0013] Figure 6 This is a signaling diagram illustrating a channel state information reporting method with frequency domain reference signal reduction according to an example embodiment of this disclosure;

[0014] Figure 7AThis is a signaling diagram illustrating a channel state information reporting method with combined antenna domain and frequency domain reference signal reduction according to an example embodiment of this disclosure;

[0015] Figure 7B This is a signaling diagram illustrating a channel state information reporting method with combined antenna domain and frequency domain reference signal reduction according to an example embodiment of this disclosure;

[0016] Figure 7C This is a schematic diagram illustrating resource allocation based on exemplary embodiments of this disclosure;

[0017] Figure 7D This is a signaling diagram illustrating a channel state information reporting method for a network-side model based on example embodiments of this disclosure;

[0018] Figure 8 This is a flowchart illustrating a channel state information reporting method performed by a network device according to an example embodiment of this disclosure;

[0019] Figure 9 This is a flowchart illustrating a model management method performed by a UE according to an example embodiment of this disclosure;

[0020] Figure 10 This is a signaling diagram illustrating the model management process according to an example embodiment of this disclosure;

[0021] Figure 11 This is a signaling diagram illustrating a network-initiated performance monitoring process according to an example embodiment of this disclosure;

[0022] Figure 12 This is a flowchart illustrating the process of calculating key performance indicators (KPIs) according to an example embodiment of this disclosure;

[0023] Figure 13 This is a signaling diagram illustrating the performance monitoring process initiated by the UE according to an example embodiment of this disclosure;

[0024] Figure 14 This is a signaling diagram illustrating the model retraining process according to an example embodiment of this disclosure;

[0025] Figure 15 This is a block diagram illustrating a communication device according to an example embodiment of the present disclosure. Detailed Implementation

[0026] Reference will now be made in detail to exemplary embodiments of the invention, examples of which are illustrated in the accompanying drawings. Wherever possible, the same element references are used in the drawings and description to denote the same or similar parts.

[0027] Figure 2 This is a schematic diagram illustrating communication system 1 according to an exemplary embodiment of this disclosure. (See reference...) Figure 2Communication system 1 (e.g., a Long Term Evolution (LTE) system, an LTE-Advanced (LTE-A) system, an LTE-Advanced Pro system, a 5G NR Radio Access Network (RAN), or a 6G Radio Access Network) typically includes at least one network device (NW), at least one user equipment (UE), and one or more optional network elements providing connectivity to the network. The UE communicates with the network (e.g., a core network (CN), an evolved packet core (EPC) network, an evolved universal terrestrial radio access network (E-UTRAN), a 5G core (5GC), or the Internet) through the RAN established by one or more network devices.

[0028] It should be noted that, in this disclosure, the User Equipment (UE) may be, but is not limited to, a mobile station, a mobile terminal or device, or a user communication wireless terminal. For example, the UE may be a portable wireless device, including but not limited to mobile phones, tablet computers, wearable devices, sensors, vehicles, or personal digital assistants (PDAs) with wireless communication capabilities. The UE is configured to receive and transmit signals to one or more cells in the radio access network via an air interface.

[0029] Network equipment (NW, also known as a base station) may be configured to provide communication services according to at least one of the following radio access technologies (RATs): Global Microwave Access Interoperability (WiMAX), Global System for Mobile Communications (GSM, commonly referred to as 2G), GSM Evolution Enhanced Data Rate (EDGE) Radio Access Network (GERAN), General Packet Radio Service (GPRS), Universal Mobile Telecommunications System based on Basic Wideband Code Division Multiple Access (W-CDMA) (UMTS, commonly referred to as 3G), High-Speed ​​Packet Access (HSPA), LTE, LTE-A, eLTE (evolved LTE, such as LTE connected to 5GC), NR (commonly referred to as 5G), and / or LTE-A Pro. However, the scope of this disclosure should not be limited to the above-mentioned protocols.

[0030] Network equipment (NW) may include, but is not limited to, Node Bs (NBs) in UMTS, Evolved Node Bs (eNBs) in LTE or LTE-A, Radio Network Controllers (RNCs) in UMTS, Base Station Controllers (BSCs) in GSM / GSM Evolution Enhanced Data Rate (EDGE) Radio Access Networks (GERAN), Next-Generation eNBs (ng-eNBs) in Evolved Universal Terrestrial Radio Access (E-UTRA) base stations connected to 5GC, Next-Generation Node Bs (gNBs) in 5G Access Networks (5G-AN), and any other means capable of controlling wireless communications and managing intra-cell radio resources. Network equipment (NW) can be connected to serve one or more User Equipments (UEs) through a radio interface to the network.

[0031] A network device (NW) (or base station) is operable to provide radio coverage to a specific geographic area using multiple cells included in the RAN. The network device NW supports cell operation. Each cell is operable to provide service to at least one user equipment (UE) within its radio coverage area. Specifically, each cell (typically referred to as the serving cell) provides service to one or more UEs within its radio coverage area (e.g., each cell schedules downlink (DL) and optional uplink (UL) resources to at least one UE within its radio coverage area for DL ​​and optional UL packet transmissions). The network device NW can communicate with one or more UEs in a wireless communication system via multiple cells. It should be noted that for UL, the UE is the transmitter performing UL transmission, and the network device NW is the receiver performing UL reception. For DL, the UE is the receiver performing DL reception, and the network device NW is the transmitter performing DL transmission.

[0032] A base station NW may include a network node NN and one or more TRPs, such as TRP#1 and TRP#2.

[0033] It should be understood that the terms "system" and "network" used in this disclosure are generally used interchangeably. The term "and / or" in this disclosure describes the relationship between related objects only, meaning that there can be three relationships, for example, A and / or B, which can represent three cases: A exists alone, A and B exist simultaneously, or B exists alone. In addition, the character " / " in this disclosure generally indicates that related objects are in an "or" relationship.

[0034] To facilitate understanding of the technical solutions of the embodiments of this disclosure, the technical concepts related to the embodiments of this disclosure are described below.

[0035] Figure 3 This is a flowchart illustrating a method for reporting channel state information performed by a user equipment (UE) according to an exemplary embodiment of the present disclosure. Figure 3 This method can be implemented by communication equipment, such as the user equipment (UE) mentioned above. (See reference...) Figure 3 In step S310, the user equipment (UE) receives a report configuration from the network device (NW). Specifically, the report configuration indicates a first set of physical resources and a first set of reference signals for which the UE intends to report channel state information (CSI). The first set of resources represents the complete or all channel information that the network device (NW) intends to obtain.

[0036] In one embodiment, the reporting configuration may be a CSI reporting setting configuration. Therefore, the user equipment (UE) can receive a CSI resource setting configuration indicating a first set of physical resources and a first set of reference signals. In another embodiment, the reporting configuration may be another configuration for configuring CSI reporting. In one embodiment, channel state information includes at least one of a channel quality indicator (CQI), a precoding matrix indicator (PMI), explicit MIMO channel coefficients or their eigenvectors and eigenvalues, a rank indicator (RI), a layer indicator (LI), or a signal strength indicator. However, CSI can represent any channel quality-related information, such as received signal strength, signal quality, signal-to-noise ratio (SNR), and signal-to-interference-plus-noise ratio (SINR) for other communication technologies.

[0037] In one embodiment, the type of the first set of physical resources includes physical resource blocks (PRBs), frequency bands, and the number of antenna ports. For example, the reporting configuration indicates 52 physical resource blocks and 16 antenna ports for CSI reporting.

[0038] The reference signal is intended to be transmitted on the first physical resource set for channel estimation / measurement or CSI reporting. In one embodiment, the type of reference signal may be a Channel State Information Reference Signal (CSI-RS). For example, the CSI reporting settings may configure the User Equipment (UE) to report CSI corresponding to the first set of N=16 antenna ports and X=52 physical resource blocks (PRBs). In another embodiment, the type of reference signal may be a downlink (DL) reference signal, an uplink (UL) reference signal, a synchronization signal block (SSB), or a demodulation reference signal (DM-RS).

[0039] In step S320, the user equipment (UE) receives a second set of reference signals on the second set of physical resources. Specifically, the second set of resources is a subset of the first set of resources indicated in the report configuration. That is, the second set of physical resources is less than the first set of physical resources, and / or the second set of reference signals is less than the first set of reference signals. For example, the first set of physical resources consists of 52 physical resource blocks and 16 antenna ports, and the second set of physical resource resources consists of 12 physical resource blocks and 8 antenna ports. The first set of reference signals has... The second set of reference signals has one set of reference signals. One reference signal.

[0040] In one embodiment, the reporting configuration is a CSI reporting setting configuration, and the configuration of the second reference signal set on the second physical resource set is a CSI resource setting configuration. Therefore, the user equipment (UE) receives the CSI resource setting configuration, which indicates the second reference signal set and the second physical resource set from the network device (NW). In response to receiving the CSI resource setting configuration, the UE receives the second reference signal set on the second physical resource set indicated in the CSI resource setting configuration.

[0041] In step S330, the user equipment (UE) determines a first estimated channel information corresponding to a second set of reference signals on a second physical resource set. Specifically, the second set of reference signals transmitted on the second physical resource set is used for channel estimation / measurement. The result of channel estimation / measurement using the second set of reference signals (i.e., the first estimated channel information) relative to the first set of reference signals can represent a portion of the channel corresponding to the resource elements of the first set of reference signals. In channel estimation / measurement, the UE can determine the channel coefficients of these resource elements. These channel coefficients can be complex numbers representing the gain and phase shift introduced by the channel at specific times and frequencies corresponding to those resource elements. The channel coefficients can be obtained, for example, by dividing the symbol value of the received reference signal by the value of a known reference signal symbol, but are not limited thereto. A channel matrix can be generated based on these channel coefficients, and the channel matrix can be or be transformed into parameters of the estimated channel information (e.g., the first estimated channel information and another estimated channel information introduced later).

[0042] In step S340, the user equipment (UE) determines a second estimated channel information corresponding to a first reference signal set on a first physical resource set by using the first estimated channel information on the prediction model. Specifically, the second estimated channel information represents a reconstruction of the channel information of the complete radio resource set (i.e., the resource elements corresponding to the first reference signal set).

[0043] The prediction model is based on a signal processing algorithm or trained using a machine learning algorithm. The prediction model represents a mathematical model of a second estimated channel information based on a first estimated channel information. In one embodiment, the signal processing algorithm may be a CSI prediction algorithm. The CSI prediction algorithm predicts the second estimated channel information from the first estimated channel information. For example, the user equipment (UE) establishes an optimization problem or function based on the parameters of the second estimated channel information. By solving the optimization problem or function, the parameters of the second estimated channel information can be obtained.

[0044] As an example of a CSI prediction algorithm, a machine learning algorithm is used to learn the relationship between the input and output of a prediction model using training data. The input to the prediction model is a first estimated channel information, and the output of the prediction model is a second estimated channel information. In one embodiment, the machine learning algorithm can be a masked autoencoder (MAE). A MAE is a self-supervised learning model. By masking a portion of the input image, the MAE learns to predict the original image. Therefore, a MAE can be used to report the complete CSI (e.g., the second estimated channel information) using a reduced CSI-RS (i.e., a second set of reference signals). The MAE is trained to reconstruct the complete data structure (e.g., the parameters of the second estimated channel information) from partially masked or incomplete inputs (e.g., parameters of the first estimated channel information). For example, the training dataset is from cross-frequency domains (e.g., N... SB MIMO channel (subband) Channel matrix. During training, the input to the prediction model is the masked MIMO channel, which is... A portion of the MIMO channel is masked, unknown, or defined with predetermined values. One of the goals of the training process is to recover the channel coefficients of the mask. The loss function of the training process can be the mean squared error (MSE) of the channel coefficients of the mask. In some implementations, the prediction model can be trained on another device, and the trained model is deployed to the user equipment (UE).

[0045] In the embodiments described, the first estimated channel information is a masked version of the second estimated channel information. The first estimated channel information is input into the prediction model, and then the second estimated channel information is output. The prediction model processes the first estimated channel information based on its learned channel characteristics and correlations in the spatial and / or frequency domains, and performs interpolation and extrapolation to fill in missing channel information. In some embodiments, the CSI-RS pattern of the network device NW is matched with the mask ratio of the prediction model of the user equipment UE. The mask ratio will be described later.

[0046] In step S350, the User Equipment (UE) reports the channel state information associated with the second estimated channel information. Specifically, the channel coefficients or channel matrix of the second estimated channel information can be transformed into parameters of the channel state information, such as CQI, PMI, or RI. The UE can format the parameters of the channel state information into a CSI report message. The channel state information can be further compressed to reduce its size. Then, the channel state information associated with the second estimated channel information can be reported via the CSI report message.

[0047] Figure 4 This is a signaling diagram illustrating the channel state information reporting process. (Reference) Figure 4In step S410, the User Equipment (UE) receives a CSI report setting configuration and a CSI resource setting configuration from the Network Device (NW), both indicating N=16 antenna ports and X=52 physical resource blocks. In step S420, the UE receives reference signals on the resource elements corresponding to the 16 antenna ports and 52 physical resource blocks indicated in the CSI resource setting configuration. The UE uses these reference signals to perform channel estimation to obtain estimated channel information. Then, in step S430, the UE transmits a CSI report message carrying the CSI associated with the estimated channel information corresponding to the 16 antenna ports and 52 physical resource blocks indicated in the CSI resource setting configuration.

[0048] Figure 5 This is a signaling diagram illustrating a channel state information reporting method with antenna domain reference signal reduction according to exemplary embodiments of the present disclosure. (Reference) Figure 5 This introduces an antenna domain reduction scenario, where the CSI resource setting configuration (received in step S520) and the actually transmitted CSI-RS (received in step S530) correspond only to N=8 antenna ports (i.e., the second reference signal set on the second physical resource set), which is a subset of the N=16 antenna ports (i.e., the first set of reference signals on the first set of physical resources) in the CSI report setting configuration (received in step S520). The second reference signal set on the second physical resource set can be determined based on the capability report provided by the user equipment (UE) in step S510. The capability report can indicate the model identifier of the prediction model. Each model identifier is associated / configured with a mask ratio. The mask ratio is a hyperparameter. The mask ratio is the ratio of the combination of the first set of reference signals and the first set of physical resources, and the combination of the second reference signal set and the second physical resource set. For example, the first set of physical resources is 52 physical resource blocks and 16 antenna ports, and the second physical resource set is 52 physical resource blocks and 8 antenna ports. The combination of the first set of reference signals and the first set of physical resources is... The combination of the second reference signal set and the second physical resource set is The mask ratio of the prediction model can be at least 50% or 60%, and is greater than or equal to... .

[0049] In step S540, the prediction model is used to determine a second estimated channel information corresponding to the full set of CSI-RS with 52 physical resource blocks and 16 antenna ports using a reduced CSI-RS with 52 physical resource blocks and 8 antenna ports. Then, in step S550, a CSI report message carrying the CSI associated with the second estimated channel information is transmitted to the network device NW. The second estimated channel information can be further compressed before being transmitted to the network device NW to reduce the payload size.

[0050] Figure 6 This is a signaling diagram illustrating a channel state information reporting method with frequency domain reference signal reduction according to exemplary embodiments of the present disclosure. (Reference) Figure 6 A frequency domain reduction scenario is introduced, where the CSI resource setting configuration (received in step S620) and the actually transmitted CSI-RS (received in step S630) correspond only to X=12 physical resource blocks (i.e., the second reference signal set on the second physical resource set), which is a subset of the X=52 physical resource blocks (i.e., the first set of reference signals on the first set of physical resources) in the CSI report setting configuration (received in step S620). The second reference signal set on the second physical resource set can be determined based on the capability report provided by the user equipment (UE) in step S610. The capability report can indicate the model identifier of the prediction model. Each model identifier is associated / configured with a mask ratio. For example, the first set of physical resources is 52 physical resource blocks and 16 antenna ports, and the second set of physical resources is 12 physical resource blocks and 16 antenna ports. The combination of the first set of reference signals and the first set of physical resources is... The combination of the second reference signal set and the second physical resource set is The masking ratio of the prediction model can be at least 77% or 80%, which is greater than or equal to... .

[0051] In step S640, the prediction model is used to determine a second estimated channel information corresponding to the full set of CSI-RS with 52 physical resource blocks and 16 antenna ports using a reduced CSI-RS with 12 physical resource blocks and 16 antenna ports. Then, in step S650, a CSI report message carrying the CSI associated with the second estimated channel information is transmitted to the network device NW. The second estimated channel information can be further compressed before being transmitted to the network device NW to reduce the payload size.

[0052] Figure 7A This is a signaling diagram illustrating a channel state information reporting method with combined antenna domain and frequency domain reference signal reduction according to exemplary embodiments of the present disclosure. (Reference) Figure 7AThis introduces antenna domain and frequency domain reduction scenarios, where the CSI resource setting configuration (received in step S720) and the actually transmitted CSI-RS (received in step S730) correspond only to X=12 physical resource blocks and N=8 antenna ports (i.e., the second reference signal set on the second physical resource set), which is a subset of the X=52 physical resource blocks and N=16 antenna ports (i.e., the first set of reference signals on the first set of physical resources) in the CSI report setting configuration (received in step S720). The second reference signal set on the second physical resource set can be determined based on the capability report provided by the user equipment (UE) in step S710. The capability report can indicate the model identifier of the prediction model. Each model identifier is associated / configured with a mask ratio. For example, the first set of physical resources is 52 physical resource blocks and 16 antenna ports, and the second set of physical resources is 12 physical resource blocks and 8 antenna ports. The combination of the first set of reference signals and the first set of physical resources is... The combination of the second reference signal set and the second physical resource set is The masking ratio of the prediction model can be at least 89% or 90%, which is greater than or equal to... .

[0053] In step S740, the prediction model is used to determine a second estimated channel information corresponding to the full set of CSI-RS with 52 physical resource blocks and 16 antenna ports using a reduced CSI-RS with 12 physical resource blocks and 8 antenna ports. Then, in step S740, a CSI report message carrying the CSI associated with the second estimated channel information is transmitted to the network device NW. The second estimated channel information can be further compressed before transmission to the network device NW to reduce the payload size.

[0054] Figure 7B This is a signaling diagram illustrating a channel state information reporting method with combined antenna domain and frequency domain reference signal reduction according to exemplary embodiments of the present disclosure. Figure 7C This is a schematic diagram illustrating resource allocation according to exemplary embodiments of the present disclosure. (See reference) Figure 7B and Figure 7CThis introduces another antenna domain and frequency domain reduction scenario, where the CSI resource setting configuration (received in step S752) and the actually transmitted CSI-RS (received in step S753) correspond to a combination of X1=13 physical resource blocks and N1=4 antenna ports with X2=13 physical resource blocks and N2=4 antenna ports (i.e., the second reference signal set on the second physical resource set), which is a subset of X=52 physical resource blocks and N=16 antenna ports in the CSI report setting configuration (received in step S752) (i.e., the first set of reference signals on the first set of physical resources). The second reference signal set on the second physical resource set can be determined based on the capability report provided by the user equipment (UE) in step S751. The capability report can indicate the model identifier of the prediction model. Each model identifier is associated / configured with a mask ratio. For example, the first set of physical resources is 52 physical resource blocks and 16 antenna ports, and the second physical resource set is a combination of 13 physical resource blocks and 4 antenna ports with X2=13 physical resource blocks and 4 antenna ports.

[0055] In step S754, the prediction model is used to determine a second estimated channel information corresponding to the complete set of CSI-RS with 52 physical resource blocks and 16 antenna ports by using reduced CSI-RS with 4 antenna ports (e.g., port indices = 2, 4, 10, 12) on 13 physical resource blocks (every 4 physical resource blocks, starting from PRB1) and 4 antenna ports (e.g., port indices = 1, 7, 9, 15) on 13 physical resource blocks (every 4 physical resource blocks, starting from PRB3). Then, in step S755, a CSI report message carrying the CSI associated with the second estimated channel information is transmitted to the network device NW. The second estimated channel information may be further compressed before being transmitted to the network device NW to reduce the payload size.

[0056] In one embodiment, a network-side model may exist. Figure 7D This is a signaling diagram illustrating a channel state information reporting method for a network-side model according to exemplary embodiments of the present disclosure. (See reference...) Figure 7DThe User Equipment (UE) can be configured by the Network Device (NW) to report MIMO channels (step S771). The NW can transmit a reporting configuration indicating a second set of physical resources and a second set of reference signals, for example, X=12 physical resource blocks and N=8 antenna ports, for reporting. The actually transmitted CSI-RS (received in step S772) corresponds to X=12 physical resource blocks and N=8 antenna ports (i.e., the second set of reference signals on the second set of physical resources), as the same resources and reference signals configured in step S771. The UE can report compressed MIMO channels, for example, channels corresponding to 12 physical resource blocks and 8 antenna ports, to reduce feedback overhead (step S773). The report is used for a prediction model to predict a second estimated channel information corresponding to a first set of reference signals on a first set of physical resources, where the second set of physical resources is a subset of the first set of physical resources or the second set of reference signals is a subset of the first reference signals. In this case, the network device NW first decompresses the MIMO channel, and then estimates the channel corresponding to the complete set CSI-RS by using a prediction model with the compressed MIMO channel corresponding to 12 physical resource blocks and 8 antenna ports (step S774), for example, with 52 physical resource blocks and 16 antenna ports (i.e., the first set reference signal on the first set of physical resources).

[0057] Figure 8 This is a flowchart illustrating a channel state information reporting method performed by a network device NW according to an exemplary embodiment of the present disclosure. (See also...) Figure 8 In step S810, the network device NW transmits a report configuration. The report configuration indicates a first physical resource set and a first reference signal set for reporting. In step S820, the network device NW transmits a second reference signal set on a second physical resource set. The second physical resource set is less than the first physical resource set, or the second reference signal set is less than the first reference signal set. The second reference signal set on the second physical resource set is used to determine a first estimated channel information. The first estimated channel information is used to predict a model to determine a second estimated channel information corresponding to the first reference signal set on the first physical resource set, and the prediction model is a signal processing algorithm or trained by a machine learning algorithm. In step S830, the network device NW receives a channel state information report associated with the second estimated channel information.

[0058] In one embodiment, the types of physical resources in the first and second sets include physical resource blocks (PRBs), frequency bands, and the number of antenna ports.

[0059] In one embodiment, the first and second set reference signals are of the type of channel state information reference signals (CSI-RS).

[0060] In one embodiment, the channel state information includes at least one of the following: channel quality indicator (CQI), precoding matrix indicator (PMI), explicit MIMO channel coefficients or their eigenvectors and eigenvalues, rank indicator (RI), layer indicator (LI), or signal strength indicator.

[0061] In one embodiment, the reporting configuration is a CSI reporting setting configuration, and the configuration of the second reference signal set on the second physical resource set is a CSI resource setting configuration. The network device NW transmits the CSI resource setting configuration.

[0062] In one embodiment, the prediction model is configured to have a masking ratio, and the masking ratio is the ratio of the combination of a first set of reference signals and a first set of physical resources to the combination of a second set of reference signals and a second set of physical resources. The masking ratio can be implicitly inferred at the user equipment from information of the first set of reference signals and the second set of physical resources, or explicitly signaled from the network device to the user equipment or from the user equipment to the network device.

[0063] Figure 9 This is a flowchart illustrating a model management method performed by a user equipment (UE) according to an exemplary embodiment of the present disclosure. (See reference...) Figure 9 In step S910, the user equipment (UE) associates the prediction model with the mask ratio. Specifically, the prediction model is a signal processing algorithm or trained by a machine learning algorithm. The prediction model recovers a second estimated channel information corresponding to a first reference signal set with a first physical resource set by inputting the first estimated channel information. The second reference signal set on the second physical resource set is used to determine the first estimated channel information, and the mask ratio is the ratio of the combination of the first reference signal set and the first physical resource set to the combination of the second reference signal set and the second physical resource set. The prediction model and mask ratio have been described above, and repeated descriptions will be omitted.

[0064] The mask ratio can be the minimum ratio by which a prediction model can recover the complete channel corresponding to the first reference signal set and the first physical resource set from a portion of the channel corresponding to the second set of reference signals and the second set of physical resources, wherein the mask ratio is greater than or equal to the ratio of the number of mask resource elements to the number of resource elements of the complete channel. Multiple prediction models may exist. Each prediction model is configured to / associate its mask ratio in order to associate the prediction model with its mask ratio.

[0065] In step S920, the user equipment (UE) transmits an indication of the prediction model to the network device (NW).

[0066] In one embodiment, the indication of the prediction model is a capability report. For example, a user equipment capability report. The capability report indicates the prediction models supported by the user equipment (UE) with mask ratios. The capability report may specify the mapping relationship between the model identifiers of one or more prediction models and their mask ratios. For example, Table (1) shows the mapping relationship between m model identifiers of prediction models and their mask ratios.

[0067] Table (1)

[0068] ,

[0069] Where m is a positive integer, and N_1 to N_m are positive values. The capability report may include a model identifier. The network device NW can identify the corresponding prediction model and its mask ratio by receiving the model identifier and the indicated model identifier. Then, the network device NW can transmit the corresponding reference signal, such as CSI-RS, based on the prediction model and / or mask ratio supported by the user equipment UE. That is, the second set of reference signals transmitted on the second physical resource set can be determined according to the capability report, which reports the prediction model and / or mask ratio supported by the user equipment UE.

[0070] In one embodiment, the capability report may include an RRC information element, the

[0071] (i) Enumerate the modelSetId(s) supported by the user equipment for model-assisted CSI compression / reporting and their corresponding mask ratio configuration files; and

[0072] (ii) Declare the CSI processing unit for computational / memory and inference latency capabilities used to perform predictive models on CSI reports.

[0073] In one embodiment, the User Equipment (UE) can report how many CSI processing units a model occupies when processing CSI. The UE can report a single processing unit shared by all model identifiers, or report it separately for each model identifier (i.e., model identifier). The CSI processing unit is related to the computation / memory and inference latency capabilities used to execute predictive models on the CSI report. The UE can also report a maximum CSI processing unit, which is the number of CSI reports it can process simultaneously; that is, the maximum number of CSI processing units that the UE is not expected to simultaneously configure to occupy that number of CSI reports, where the total number of CSI processing units does not exceed its capacity.

[0074] For example, Figure 10 This is a signaling diagram illustrating a model management procedure according to exemplary embodiments of this disclosure. (See reference...) Figure 10In step S1010, the User Equipment (UE) transmits a UE Capability Report to the Network Device (NW). The UE Capability Report indicates ID#1, ID#2, and ID#3 of the prediction model. In step S1020, the Network Device (NW) transmits a CSI Report Setting Configuration and a CSI Resource Setting Configuration to the UE. The CSI Report Setting Configuration indicates N=16 antenna ports and X=52 physical resource blocks, which is a first set of reference signals on a first set of physical resources used for the CSI report. The CSI Resource Setting Configuration indicates N=8 antenna ports and X=52 physical resource blocks, which is a second set of reference signals on a second set of physical resources. In step S1030, the Network Device (NW) transmits CSI-RS#1, which is a set of CSI-RS transmitted on each resource element corresponding to 8 antenna ports in the 52 physical resource blocks.

[0075] CSI-RS#1 corresponds to ID#1 of the prediction model. That is, ID#1 matches CSI-RS#1. In step S1035, the user equipment (UE) uses ID#1 of the prediction model to determine the second estimated channel information corresponding to N=16 antenna ports and X=52 physical resource blocks by using the first estimated channel information corresponding to N=8 antenna ports and X=52 physical resource blocks. In step S1040, the UE transmits a complete CSI report to the network device (NW) associated with the second estimated channel information corresponding to the complete set of CSI-RS with 52 physical resource blocks and 16 antenna ports.

[0076] In one embodiment, the User Equipment (UE) may transmit a suggestion indication to suggest to the Network Equipment (NW) what reference signal to transmit. In this case, for example, during initial model activation, model switching, or fallback, the UE transmits the suggestion indication. For instance, the UE may detect a channel change and transmit the suggestion indication.

[0077] refer to Figure 10 In step S1050, the UE transmits a suggestion indication to suggest what reference signals, such as CSI-RS, the network device NW can transmit. For example, it suggests ID#2 and ID#3 of the prediction model. In step S1060, the network device NW transmits another CSI report setting configuration and another CSI resource setting configuration to the user equipment UE. The CSI report setting configuration indicates N=16 antenna ports and X=52 physical resource blocks, which is a first set of reference signals on a first set of physical resources used for CSI reporting. However, the CSI resource setting configuration indicates N=8 antenna ports and X=16 physical resource blocks, which is a second set of reference signals on a second set of physical resources. In step S1070, the network device NW transmits CSI-RS#2, which is a set of CSI-RS transmitted on resource elements corresponding to 8 antenna ports in each of the 16 physical resource blocks.

[0078] CSI-RS#2 corresponds to prediction model ID#3. In step S1075, the user equipment (UE) uses prediction model ID#3 to determine second estimated channel information corresponding to N=16 antenna ports and X=52 physical resource blocks by using first estimated channel information corresponding to N=8 antenna ports and X=16 physical resource blocks. In step S1080, the UE transmits a complete CSI report to the network device (NW) associated with the second estimated channel information corresponding to the complete set of CSI-RS with 52 physical resource blocks and 16 antenna ports.

[0079] In one embodiment, the predictive model is indicated by monitoring results. The monitoring results can indicate a performance metric between the second estimated channel information and the true information, and the true information is determined by a first set of reference signals on a first set of physical resources. The performance metric (or key performance indicator (KPI)) can be, for example, an average error (e.g., normalized root mean square error (NMSE)), an error distribution (e.g., NMSE per monitoring instance), or an event (e.g., more than N consecutive errors, where N is a positive integer). The true information is channel state information determined using a complete and comprehensive set of reference signals (e.g., the first set of reference signals on the first set of physical resources) and serves as the gold standard for evaluating the estimated channel information.

[0080] In one embodiment, the monitoring results may (further) indicate one or more suggested model identifiers for model switching.

[0081] For example, Figure 11 This is a signaling diagram illustrating a network-initiated performance monitoring procedure according to exemplary embodiments of this disclosure. (See also:) Figure 11 In step S1110, the User Equipment (UE) receives a monitoring report (or monitoring configuration) from the Network Device (NW) to configure one or more performance metrics. The monitoring configuration indicates the type of performance metric, and the type of performance metric is at least one of average error, error distribution, or indicates whether a predefined event is met. In step S1120, the UE receives a monitoring command (e.g., a monitoring report trigger command) from the Network Device (NW). The monitoring command may indicate the performance metrics indicated in the monitoring report. That is, the monitoring command is used to trigger the UE to generate performance metrics. The monitoring command may indicate a legacy / complete CSI-RS resource, which is a resource element of a first reference signal set on a first physical resource set. Furthermore, step S1110 may be an RRC configuration, and step S1120 may be a DCI message. Steps S1110 and S1120 may be combined into a single message; in this case, the message may be a DCI message.

[0082] In step S1130, the network device NW transmits (traditional / complete) CSI-RS to the user equipment UE. The traditional / complete CSI-RS is a first set of reference signals on a first physical resource set. The user equipment UE receives the first set of reference signals on the first physical resource set.

[0083] In one embodiment, the user equipment (UE) may receive a model identifier (or the aforementioned model identifier) ​​associated with a performance metric, if available. The monitoring command may (further) include the model identifier of the predictive model.

[0084] In one embodiment, the monitoring command may (further) include the format of the monitoring report. The format of the monitoring report is the content to be reported. For example, the type of performance metric in the monitoring report.

[0085] In one embodiment, the user equipment (UE) may receive predefined events, if any. The monitoring command may (further) include the type of one or more predefined events that trigger a monitoring report. That is, the monitoring report is event-based. An event may be, for example, N consecutive inferences where the normalized squared error of the monitoring is greater than a threshold. In some embodiments, the user equipment (UE) may not send or disable sending monitoring reports when (only when or in response to) a predefined event is not met.

[0086] In step S1140, the user equipment (UE) evaluates performance metrics. For example, Figure 12 This is a block diagram illustrating the process of calculating Key Performance Indicators (KPIs) according to exemplary embodiments of this disclosure. (See reference...) Figure 12 The User Equipment (UE) estimates the channel using (traditional / full) CSI-RS on N=16 antenna ports and X=52 physical resource blocks. In step S1220, the UE calculates the true CSI (i.e., true information) using (traditional / full) CSI-RS.

[0087] On the other hand, in step S1230, the user equipment UE masks the estimated channel so that the unmasked channel corresponds to N=8 antenna ports and X=52 physical resource blocks. Then, in step S1240, the user equipment UE estimates the channel corresponding to the full / recovered CSI-RS with N=16 antenna ports and X=52 physical resource blocks by using a prediction model and the unmasked channel to generate estimated channel information. In one embodiment, the user equipment UE may receive a model identifier and use the corresponding prediction model (associated with the received model identifier) ​​to recover the second estimated channel information corresponding to the channel associated with the full / recovered CSI-RS. In step S1250, the user equipment UE calculates the performance KPI (e.g., the NMSE between the actual CSI and the estimated channel information).

[0088] refer to Figure 11 The user equipment (UE) generates monitoring results, including performance metrics. In step S1150, the UE reports the monitoring to the network device (NW) (i.e., transmits the monitoring results).

[0089] In one embodiment, the user equipment (UE) can determine that a predefined event has been met. The UE can then transmit monitoring results in response to (if or only if) determining that the predefined event has been met.

[0090] Figure 13 This is a signaling diagram illustrating a performance monitoring process initiated by a UE according to exemplary embodiments of this disclosure. (See reference...) Figure 13 In step S1310, the user equipment (UE) receives a configuration from the network device (NW) to configure a monitoring report (or monitoring configuration), which indicates one or more performance metrics. The monitoring configuration indicates the type of performance metric, which is at least one of average error, error distribution, or indicates whether a predefined event is met. In step S1320, when the UE detects, for example, a change in channel statistics, the UE requests performance monitoring, for example, a deviation of the mean and / or variance of the channel estimation based on CSI-RS from the specified range.

[0091] In step S1330, the network device NW may trigger a monitoring report, but this can be omitted in some implementations. In step S1340, the network device NW transmits (traditional / complete) CSI-RS to the user equipment UE. The traditional / complete CSI-RS is the first set of reference signals on the first set of physical resources. The user equipment UE receives the first set of reference signals on the first set of physical resources.

[0092] In step S1350, the user equipment (UE) is as described in step S1140 and... Figure 12 The performance metrics described herein are evaluated. Then, in step S1360, the user equipment (UE) reports the monitoring to the network device (NW) (i.e., transmits the monitoring results). In this embodiment, a faster response to model failures can be achieved, and the load on the network device (NW) can be transferred to the user equipment (UE).

[0093] In one embodiment, the indicator of the prediction model indicates a retraining request for the prediction model, the retraining request being used to request retraining of the prediction model. Retraining the prediction model can, for example, be fine-tuning the prediction model.

[0094] For example, Figure 14 This is a signaling diagram illustrating the model retraining process according to exemplary embodiments of this disclosure. (See also:) Figure 14The User Equipment (UE) determines that its prediction model needs retraining (e.g., fine-tuning). Situations requiring retraining may include, for example, changes in channel statistics or monitoring KPIs failing to meet thresholds. In step S1410, the UE requests data collection from the Network Device (NW) (e.g., transmitting a retraining request). The retraining request may include a model identifier associated with a specific mask ratio to be fine-tuned. The retraining request may suggest that the Network Device (NW) switch to another model identifier (i.e., a different model identifier).

[0095] In step S1420, the network device NW may transmit a data collection trigger command (or retraining command) to the user equipment UE. The retraining command may indicate the configuration of a third set of reference signals with a third set of physical resources. For example, (traditional / complete) CSI-RS. In some embodiments, the retraining command may (further) indicate the model identifier of the prediction model currently used by the user equipment UE to be disabled. In step S1430, the network device NW transmits (traditional / complete) CSI-RS to the user equipment UE. The traditional / complete CSI-RS is a third set of reference signals on a third set of physical resources. The user equipment UE receives the third set of reference signals on the third set of physical resources according to the configuration of the third set of reference signals with the third set of physical resources.

[0096] In step S1440, the User Equipment (UE) performs data collection. For example, the UE uses (traditional / complete) CSI-RS to calculate the true CSI (i.e., true information). In step S1450, the UE retrains the prediction model (e.g., model fine-tuning) using third estimated channel information determined by using a third set of reference signals on a third set of physical resources to generate a retrained prediction model. For example, the UE masks the estimated channel corresponding to the third set of reference signals on the third set of physical resources. The UE then estimates the channel corresponding to the complete / recovered CSI-RS using the prediction model and the unmasked channel to generate third estimated channel information. A loss function between the third estimated channel information and the true information is used to fine-tune the prediction model. Parameters of the prediction model, such as weights and offsets, can be changed.

[0097] In step S1460, the user equipment (UE) notifies that the fine-tuning was successful (i.e., the results of the retrained prediction model are transmitted). In step S1470, the network device (NW) can activate the retrained prediction model. The UE can then use the retrained prediction model to estimate the channel.

[0098] Figure 15 This is a block diagram illustrating a communication device 1500 according to exemplary embodiments of the present disclosure. (See also...) Figure 15The communication device 1500 can be a UE or a network device. The communication device 1500 may include, but is not limited to, a processor 1510. The processor 1510 (e.g., having processing circuitry) may include a CSI processing unit (CPU). In some implementations, the processor 1510 may include a smart hardware device, such as a central processing unit (CPU), a microcontroller, an ASIC, etc. The processor 1510 can call and run computer programs from memory to implement the methods in the embodiments of this disclosure.

[0099] Since the program code stored in the communication device 1500 adopts all the technical solutions of all the foregoing embodiments when executed by the processor 1510, it has at least all the beneficial effects brought by all the technical solutions of all the foregoing embodiments, which will not be further described here.

[0100] Optionally, such as Figure 15 As shown, the communication device 1500 may further include a memory 1520. The memory 1520 may include a computer storage medium in the form of volatile and / or non-volatile memory. The memory 1520 may be removable, non-removable, or a combination thereof. Exemplary memories include solid-state memory, hard disk drives, optical disk drives, etc. The processor 1510 may call and run computer programs from the memory 1520 to implement the methods in the embodiments of this disclosure.

[0101] The memory 1520 can be a separate device independent of the processor 1510, or it can be integrated into the processor 1510.

[0102] Optionally, such as Figure 15 As shown, the communication device 1500 may further include a transceiver 1530, and the processor 1510 may control the transceiver 1530 to communicate with other devices. The transceiver 1530 has a transmitter (e.g., transmit / transfer circuitry) and a receiver (e.g., receive / receive circuitry), and can be configured to transmit and / or receive time and / or frequency resource segmentation information. In some implementations, the transceiver 1530 may be configured to transmit in different types of subframes and time slots, including but not limited to available, unavailable, and flexibly available subframe and time slot formats. The transceiver 1530 may be configured to receive data and control channels.

[0103] Specifically, transceiver 1530 can send information or data to other devices, or receive information or data sent by other devices.

[0104] Specifically, transceiver 1530 may include a transmitter and a receiver. Transceiver 1530 may also include antennas, and the number of antennas may be one or more.

[0105] Optionally, the communication device 1500 may specifically be a network device in the embodiments of this disclosure, and the communication device 1500 may implement the corresponding processes implemented by the network device in the various methods of the embodiments of this disclosure. For the sake of brevity, relevant descriptions are omitted.

[0106] Optionally, the communication device 1500 may specifically be a mobile terminal, terminal device, or UE in the embodiments of this disclosure, and the communication device 1500 may implement the corresponding processes implemented by the mobile terminal, terminal device, or UE in the various methods of the embodiments of this disclosure. For the sake of brevity, relevant descriptions are omitted.

[0107] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for reporting channel state information, performed by a user equipment, the method comprising: Receive report configuration, wherein the report configuration indicates a first set of physical resources and a first set of reference signals for reporting; Receive a second set of reference signals on a second set of physical resources, wherein the second set of physical resources is less than the first set of physical resources or the second set of reference signals is less than the first set of reference signals. Determine the first estimated channel information corresponding to the second reference signal set on the second physical resource set; The channel information corresponding to the first reference signal set on the first physical resource set is determined by using the first estimated channel information on a prediction model, wherein the prediction model is a signal processing algorithm or trained by a machine learning algorithm; and The report includes channel state information associated with the second estimated channel information.

2. The method according to claim 1, wherein the type of the first and second reference signal sets is a channel state information reference signal.

3. The method according to claim 1, wherein the channel state information includes at least one of a channel quality indicator, a precoding matrix indicator, a rank indicator, a layer indicator, or a signal strength indicator.

4. The method according to claim 1, wherein the report configuration is a channel state information report setting configuration, the configuration of the second reference signal set on the second physical resource set is a channel state information resource setting configuration, and the method further comprises: Receive the channel state information resource settings configuration.

5. The method according to claim 1, further comprising: A dataset of estimated channel information is collected by receiving reference signals corresponding to the first reference signal set on the first physical resource set; Transfer the dataset to the model training entity; as well as The model is trained from the model training entity.

6. The method of claim 1, wherein the prediction model is configured with a mask ratio, and the mask ratio is the ratio of the combination of the first reference signal set and the first physical resource set to the combination of the second reference signal set and the second physical resource set.

7. A method for reporting channel state information, performed by a network device, the method comprising: Transmission report configuration, wherein the report configuration indicates a first set of physical resources and a first set of reference signals for reporting; A second set of reference signals is transmitted on a second set of physical resources, wherein the second set of physical resources is less than the first set of physical resources or the second set of reference signals is less than the first set of reference signals. The second set of reference signals on the second set of physical resources is used to determine a first estimated channel information, and the first estimated channel information is used on a prediction model to determine a second estimated channel information corresponding to the first set of reference signals on the first set of physical resources. The prediction model is a signal processing algorithm or trained by a machine learning algorithm. Receive a channel state information report associated with the second estimated channel information.

8. The method of claim 7, wherein the type of the first and second reference signal sets is a channel state information reference signal.

9. The method of claim 7, wherein the channel state information comprises at least one of a channel quality indicator, a precoding matrix indicator, a rank indicator, a layer indicator, or a signal strength indicator.

10. The method of claim 7, wherein the report configuration is a channel state information report setting configuration, the configuration of the second reference signal set on the second physical resource set is a channel state information resource setting configuration, and the method further comprises: Configure the channel state information resources for transmission.

11. The method of claim 7, wherein the prediction model is configured with a mask ratio, and the mask ratio is the ratio of the combination of the first reference signal set and the first physical resource set to the combination of the second reference signal set and the second physical resource set.

12. A method for model management, executed by a user equipment, the method comprising: The prediction model is associated with a masking ratio, wherein the prediction model is a signal processing algorithm or trained by a machine learning algorithm, and the prediction model is used to recover second estimated channel information corresponding to a first estimated channel information on a first reference signal set having a first physical resource set, the second reference signal set on the second physical resource set being used to determine the first estimated channel information, and the masking ratio is the ratio of the combination of the first reference signal set and the first physical resource set to the combination of the second reference signal set and the second physical resource set; and Transmit instructions for the prediction model.

13. The method of claim 12, wherein the indication of the prediction model is a capability report, and the capability report indicates the prediction model supported by the user equipment having the mask ratio.

14. The method of claim 12, further comprising: Receive the model identifier and use the corresponding prediction model to recover the second estimated channel information.

15. The method of claim 12, wherein the indication of the prediction model is a monitoring result, the monitoring result indicating a performance metric between the second estimated channel information and the true information, and the true information is determined by the first reference signal set on the first physical resource set.

16. The method of claim 15, further comprising: Receive a monitoring command, wherein the monitoring command instructs the generation of the performance metrics; Receive the model identifier associated with the performance metric, if any; Receive predefined events, if any; Receive the first reference signal set on the first physical resource set; as well as Generate the monitoring results that include the performance metrics.

17. The method of claim 16, further comprising: Receive monitoring configuration, wherein the monitoring configuration indicates the type of the performance metric, and the type of the performance metric is at least one of average error, error distribution, or indicates whether a predefined event is met.

18. The method of claim 15, wherein the instruction to transmit the prediction model comprises: Determine if a predefined event is satisfied; and The monitoring results are transmitted in response to the determination that the predefined event is met.

19. The method of claim 12, wherein the indication of the prediction model indicates a retraining request for the prediction model, and the retraining request is used to request retraining of the prediction model.

20. The method of claim 12, further comprising: Receive a retraining command, wherein the retraining command indicates the configuration of a third reference signal set having a third physical resource set; According to the configuration of the third reference signal set having the third physical resource set, the third reference signal set on the third physical resource set is received; The prediction model is retrained using the third estimated channel information determined by using the third reference signal set on the third physical resource set; as well as The results of retraining the prediction model are transmitted.

Images