Wireless communication method and apparatus, and device

Abstract

The present application relates to the technical field of communications, and discloses a wireless communication method and apparatus, and a device. The wireless communication method in embodiments of the present application comprises: a first device acquires first buffer information, or a target AI unit acquires first buffer information, the first buffer information being used for acquiring output information associated with the target AI unit, or the first buffer information being used for acquiring first target CSI associated with the target AI unit, wherein the first buffer information comprises at least one of the following: CSI information of at least one historical time unit, at least one piece of CSI intermediate state information, and input information associated with the target AI unit on at least one time unit; CSI intermediate state information among the at least one piece of CSI intermediate state information is intermediate information acquired on the basis of the target AI unit; and the first buffer information is associated with at least one of the following: layer information, rank information, time information, a first timer, configuration information of an observation window, first information, and the target AI unit.

Description

Wireless communication methods, apparatus and equipment

[0001] Cross-reference to related applications

[0002] This application claims priority to Chinese Patent Application No. 202410795812.3, filed on June 19, 2024, entitled "Wireless Communication Method, Apparatus and Device", the entire contents of which are incorporated herein by reference. Technical Field

[0003] This application belongs to the field of communication technology, specifically relating to a wireless communication method, apparatus, and device. Background Technology

[0004] In New Radio (NR) systems, terminals need to feed back Channel State Information (CSI) to network-side equipment so that the network-side equipment can determine the precoding matrix for downlink transmission. Specifically, CSI feedback can utilize artificial intelligence (AI) units to extract and feed back channel features. For some cases (e.g., high-rank CSI feedback, such as rank ≥ 2), the high complexity of AI units and unclear interactions between different layers or time buffers make it difficult for the network and terminal sides to align their AI units, resulting in poor performance in CSI compression and decompression in the spatiotemporal frequency domains. Summary of the Invention

[0005] This application provides a wireless communication method, apparatus, and device that can solve the problem of poor performance in CSI compression and decompression in the space-time-frequency domain.

[0006] Firstly, a wireless communication method is provided, comprising:

[0007] The first device obtains the first cache information, or the target artificial intelligence (AI) unit obtains the first cache information. The first cache information is used to obtain the output information associated with the target AI unit, or the first cache information is used to obtain the first target CSI associated with the target AI unit.

[0008] The first cached information includes at least one of the following: CSI information of at least one historical time unit, at least one CSI intermediate state information, and input information associated with the target AI unit in at least one time unit;

[0009] Wherein, the CSI intermediate state information in the at least one CSI intermediate state information is intermediate information obtained based on the target AI unit;

[0010] Wherein, the first cache information is associated with at least one of the following: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and the target AI unit;

[0011] The time information includes at least one of the following: the at least one historical time unit, the time unit corresponding to the at least one CSI intermediate state information, and the at least one time unit;

[0012] In the event that the first timer expires, some or all of the information in the first cached information is cleared or reset;

[0013] The configuration information of the observation window includes at least one of the following: at least one of the time-domain interval, quantity, and start position corresponding to the CSI information of the at least one historical time unit; at least one of the time-domain interval, quantity, and start position corresponding to the at least one CSI intermediate state information; and at least one of the time-domain interval, quantity, and start position corresponding to the at least one time unit.

[0014] Wherein, the first information satisfies at least one of the following: instructing to clear or reset the first cache information, instructing to clear or reset the condition information of the first cache information, configuring the type of the first cache information, configuring the number of instances in the first cache information, instructing the association between the first cache information and the layer information, and instructing the association between the first cache information and the rank information.

[0015] Secondly, a wireless communication device is provided, comprising:

[0016] The processing module is used to obtain first cache information, wherein the first cache information is used to obtain output information associated with the target artificial intelligence (AI) unit, or the first cache information is used to obtain a first target CSI associated with the target AI unit;

[0017] The first cached information includes at least one of the following: CSI information of at least one historical time unit, at least one CSI intermediate state information, and input information associated with the target AI unit in at least one time unit;

[0018] Wherein, the CSI intermediate state information in the at least one CSI intermediate state information is intermediate information obtained based on the target AI unit;

[0019] Wherein, the first cache information is associated with at least one of the following: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and the target AI unit;

[0020] The time information includes at least one of the following: the at least one historical time unit, the time unit corresponding to the at least one CSI intermediate state information, and the at least one time unit;

[0021] In the event that the first timer expires, some or all of the information in the first cached information is cleared or reset;

[0022] The configuration information of the observation window includes at least one of the following: at least one of the time-domain interval, quantity, and start position corresponding to the CSI information of the at least one historical time unit; at least one of the time-domain interval, quantity, and start position corresponding to the at least one CSI intermediate state information; and at least one of the time-domain interval, quantity, and start position corresponding to the at least one time unit.

[0023] Wherein, the first information satisfies at least one of the following: instructing to clear or reset the first cache information, instructing to clear or reset the condition information of the first cache information, configuring the type of the first cache information, configuring the number of instances in the first cache information, instructing the association between the first cache information and the layer information, and instructing the association between the first cache information and the rank information.

[0024] Thirdly, a wireless communication device is provided, the device being configured to perform the steps of the method described in the first aspect.

[0025] Fourthly, a first device is provided, the first device including a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the method as described in the first aspect.

[0026] Fifthly, a first device is provided, including a processor and a communication interface;

[0027] The processor is used to obtain first cache information, which is used to obtain output information associated with the target AI unit, or the first cache information is used to obtain a first target CSI associated with the target AI unit.

[0028] The first cached information includes at least one of the following: CSI information of at least one historical time unit, at least one CSI intermediate state information, and input information associated with the target AI unit in at least one time unit;

[0029] Wherein, the CSI intermediate state information in the at least one CSI intermediate state information is intermediate information obtained based on the target AI unit;

[0030] Wherein, the first cache information is associated with at least one of the following: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and the target AI unit;

[0031] The time information includes at least one of the following: the at least one historical time unit, the time unit corresponding to the at least one CSI intermediate state information, and the at least one time unit;

[0032] In the event that the first timer expires, some or all of the information in the first cached information is cleared or reset;

[0033] The configuration information of the observation window includes at least one of the following: at least one of the time-domain interval, quantity, and start position corresponding to the CSI information of the at least one historical time unit; at least one of the time-domain interval, quantity, and start position corresponding to the at least one CSI intermediate state information; and at least one of the time-domain interval, quantity, and start position corresponding to the at least one time unit.

[0034] Wherein, the first information satisfies at least one of the following: instructing to clear or reset the first cache information, instructing to clear or reset the condition information of the first cache information, configuring the type of the first cache information, configuring the number of instances in the first cache information, instructing the association between the first cache information and the layer information, and instructing the association between the first cache information and the rank information.

[0035] In a sixth aspect, a target AI unit is provided, the target AI unit including a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the method as described in the first aspect.

[0036] Seventhly, a target AI unit is provided, including a processor and a communication interface;

[0037] The processor is used to acquire first cache information, which is used to acquire output information associated with the target AI unit, or the first cache information is used to acquire a first target CSI associated with the target AI unit.

[0038] The first cached information includes at least one of the following: CSI information of at least one historical time unit, at least one CSI intermediate state information, and input information associated with the target AI unit in at least one time unit;

[0039] Wherein, the CSI intermediate state information in the at least one CSI intermediate state information is intermediate information obtained based on the target AI unit;

[0040] Wherein, the first cache information is associated with at least one of the following: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and the target AI unit;

[0041] The time information includes at least one of the following: the at least one historical time unit, the time unit corresponding to the at least one CSI intermediate state information, and the at least one time unit;

[0042] In the event that the first timer expires, some or all of the information in the first cached information is cleared or reset;

[0043] The configuration information of the observation window includes at least one of the following: at least one of the time-domain interval, quantity, and start position corresponding to the CSI information of the at least one historical time unit; at least one of the time-domain interval, quantity, and start position corresponding to the at least one CSI intermediate state information; and at least one of the time-domain interval, quantity, and start position corresponding to the at least one time unit.

[0044] Wherein, the first information satisfies at least one of the following: instructing to clear or reset the first cache information, instructing to clear or reset the condition information of the first cache information, configuring the type of the first cache information, configuring the number of instances in the first cache information, instructing the association between the first cache information and the layer information, and instructing the association between the first cache information and the rank information.

[0045] In an eighth aspect, a readable storage medium is provided, on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0046] A ninth aspect provides a wireless communication system, comprising: a first device and a second device, wherein the first device is configured to perform the steps of the method described in the first aspect.

[0047] In a tenth aspect, a chip is provided, the chip including a processor and a communication interface coupled to the processor, the processor being used to run programs or instructions to implement the method as described in the first aspect.

[0048] Eleventhly, a computer program / program product is provided, which is stored in a storage medium and executed by at least one processor to implement the steps of the wireless communication method as described in the first aspect.

[0049] In this embodiment, a first device acquires first cache information, or a target AI unit acquires first cache information. The first cache information is used to acquire output information associated with the target AI unit, or to acquire a first target CSI associated with the target AI unit. Specifically, the first cache information is associated with at least one of the following: layer information, rank information, time information, a first timer, configuration information of the observation window, first information, and the target AI unit. The first cache information is introduced during the process of acquiring the output information associated with the target AI unit or the first target CSI associated with the target AI unit, thereby improving the accuracy of the target AI unit's prediction. Furthermore, the first cache information is associated with time information, the first timer, and the configuration information of the observation window, thereby enabling prediction and / or compression and decompression (such as CSI compression and CSI decompression) in the spatiotemporal frequency domain. Introducing time domain information allows the encoder and decoder to better align the model inference schemes at both ends, and / or can be applied to prediction and / or compression and decompression (such as CSI compression and CSI decompression) based on the target AI unit in high-rank (such as rank ≥ 2) cases, and improves the performance of prediction and / or compression and decompression in the spatiotemporal frequency domain (such as improving the performance of CSI compression and CSI decompression in the spatiotemporal frequency domain). Attached Figure Description

[0050] Figure 1 is a schematic diagram of a communication system architecture provided in an embodiment of this application.

[0051] Figure 2 is a schematic flowchart of a wireless communication method provided according to an embodiment of this application.

[0052] Figure 3 is a schematic diagram of a CSI compression and decompression scenario in the spatiotemporal frequency domain applicable to an embodiment of this application, provided according to an embodiment of this application.

[0053] Figure 4 is a schematic diagram of a CSI compression and decompression scenario in the spatiotemporal frequency domain applicable to another embodiment of this application provided by the present application.

[0054] Figure 5 is a schematic diagram of a Rank 4 model provided according to an embodiment of this application.

[0055] Figures 6 to 14 are schematic diagrams of the interaction flow of intermediate state information provided according to the embodiments of this application.

[0056] Figure 15 is a schematic block diagram of a wireless communication device according to an embodiment of this application.

[0057] Figure 16 is a schematic block diagram of a communication device provided according to an embodiment of this application.

[0058] Figure 17 is a schematic diagram of the hardware structure of a terminal according to an embodiment of this application.

[0059] Figure 18 is a schematic block diagram of a network-side device provided according to an embodiment of this application. Detailed Implementation

[0060] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0061] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, not limited in number; for example, the first object can be one or more. Furthermore, "or" in this application indicates at least one of the connected objects. For example, the scope of protection for "A or B" covers at least three scenarios: Scenario 1: including A but not B; Scenario 2: including B but not A; Scenario 3: including both A and B. In addition, the terms "A and / or B," "at least one of A and B," and "at least one of A or B" also cover at least the above three scenarios. The character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0062] The term "instruction" in this application can be either a direct instruction (or explicit instruction) or an indirect instruction (or implicit instruction). A direct instruction can be understood as one in which the sender explicitly informs the receiver of specific information, the operation to be performed, or the requested result, etc., in the instruction sent. An indirect instruction can be understood as one in which the receiver determines the corresponding information based on the instruction sent by the sender, or makes a judgment and determines the operation to be performed or the requested result, etc., based on the judgment result.

[0063] It is worth noting that the technologies described in this application are not limited to Long Term Evolution (LTE) / LTE-Advanced (LTE-A) systems, but can also be used in other wireless communication systems, such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single-carrier Frequency-Division Multiple Access (SC-FDMA), or other systems. The terms "system" and "network" in this application are often used interchangeably, and the described technologies can be used with the systems and radio technologies mentioned above, as well as with other systems and radio technologies. The following description describes New Radio (NR) systems for illustrative purposes, and the term NR is used in most of the following description; however, these technologies can also be applied to systems other than NR systems, such as 6th generation (6G) radio systems. th Generation 6G communication system.

[0064] Figure 1 shows a block diagram of a wireless communication system applicable to an embodiment of this application. Specifically, the wireless communication system includes a terminal 11 and a network-side device 12.

[0065] Among them, terminal 11 can be a mobile phone, tablet computer, laptop computer, notebook computer, personal digital assistant (PDA), handheld computer, netbook, ultra-mobile personal computer (UMPC), mobile internet device (MID), augmented reality (AR), virtual reality (VR) device, robot, wearable device, flight vehicle, vehicle user equipment (VUE), shipborne equipment, pedestrian user equipment (PUE), smart home (home devices with wireless communication functions, such as refrigerators, televisions, washing machines or furniture, etc.), game console, personal computer (PC), ATM or self-service machine, etc. Wearable devices include: smartwatches, smart bracelets, smart earphones, smart glasses, smart jewelry (smart bracelets, smart chains, smart rings, smart necklaces, smart anklets, smart anklets, etc.), smart wristbands, smart clothing, etc. Among these, in-vehicle devices can also be referred to as in-vehicle terminals, in-vehicle controllers, in-vehicle modules, in-vehicle components, in-vehicle chips, or in-vehicle units, etc. It should be noted that the specific type of terminal 11 is not limited in the embodiments of this application.

[0066] Among them, network-side equipment 12 may include access network equipment or core network equipment.

[0067] Alternatively, access network equipment may also be referred to as Radio Access Network (RAN) equipment, radio access network function, or radio access network unit. Access network equipment may include base stations, wireless local area network (WLAN) access points (APs), or wireless Fidelity (WiFi) nodes, etc. The term "base station" can be referred to as Node B (NB), Evolved Node B (eNB), Next Generation Node B (gNB), New Radio Node B (NR Node B), Access Point, Relay Base Station (RBS), Serving Base Station (SBS), Base Transceiver Station (BTS), Radio Base Station, Radio Transceiver, Basic Service Set (BSS), Extended Service Set (ESS), Home Node B (HNB), Home Evolved Node B, Transmit / Receive Point (TRP), or any other suitable term in the relevant field, as long as the same technical effect is achieved. The term "base station" is not limited to any specific technical terminology. It should be noted that this application embodiment only uses a base station in an NR system as an example for description and does not limit the specific type of base station.

[0068] Optionally, core network equipment may also be referred to as core network nodes, core network functions, or core network elements, and includes, but is not limited to, at least one of the following: Mobility Management Entity (MME), Access and Mobility Management Function (AMF), Session Management Function (SMF), User Plane Function (UPF), Policy Control Function (PCF), Policy and Charging Rules Function (PCRF), Edge Application Server Discovery Function (EASDF), Unified Data Management (UDM), Unified Data Repository (UDR), Home Subscriber Server (HSS), Centralized network configuration (CNC), Network Repository Function (NRF), Network Exposure Function (NEF), Local NEF (or L-NEF), and Binding Support. The core network functions include: BSF (Block Network Function), Application Function (AF), Location Management Function (LMF), Gateway Mobile Location Centre (GMLC), and Network Data Analytics Function (NWDAF). It should be noted that this application embodiment only uses core network equipment in the NR system as an example and does not limit the specific type of core network equipment. If the name of the core network equipment mentioned in this application embodiment changes in subsequent protocol versions (e.g., 6G), it will still be within the scope of protection of this application.

[0069] Optionally, the core network equipment can be implemented by one or more functional modules in a single device, or by multiple devices working together; this application does not specifically limit this. It is understood that the aforementioned functional modules can be network elements in hardware devices, software functional modules running on dedicated hardware, or virtualized functional modules instantiated on a platform (e.g., a cloud platform).

[0070] To facilitate a better understanding of the embodiments of this application, artificial intelligence (AI) will be described.

[0071] Artificial intelligence (AI) has been widely applied in various fields. Integrating AI into wireless communication networks to significantly improve technical indicators such as throughput, latency, and user capacity is an important task for future wireless communication networks. AI modules can be implemented in various ways, such as neural networks, decision trees, support vector machines, and Bayesian classifiers. This application uses neural networks as an example for illustration, but it does not limit the specific type of AI module.

[0072] The parameters of a neural network are optimized using gradient optimization algorithms. Gradient optimization algorithms are a class of algorithms that minimize or maximize an objective function (sometimes called a loss function), which is often a mathematical combination of model parameters and data. For example, given data X and its corresponding label Y, we construct a neural network model f(.). With the model, we can obtain the predicted output f(x) based on the input x, and calculate the difference between the predicted value and the true value (f(x) - Y), which is the loss function. Our goal is to find suitable W and b that minimize the value of the above loss function. The smaller the loss value, the closer our model is to the reality.

[0073] Most common optimization algorithms are based on the error back propagation (BP) algorithm. The basic idea of ​​the BP algorithm is that the learning process consists of two parts: forward propagation of the signal and backward propagation of the error. During forward propagation, the input sample is introduced from the input layer, processed layer by layer by the hidden layers, and then propagated to the output layer. If the actual output of the output layer does not match the expected output, the process transitions to the error back propagation stage. Error back propagation involves propagating the output error back to the input layer layer by layer through the hidden layers, distributing the error to all units in each layer, thus obtaining the error signal of each unit. This error signal serves as the basis for adjusting the weights of each unit. This process of adjusting the weights through forward and backward propagation is repeated continuously. This continuous adjustment of weights is the learning and training process of the network. This process continues until the error of the network output is reduced to an acceptable level, or until the predetermined number of learning iterations is reached.

[0074] Common optimization algorithms include Gradient Descent, Stochastic Gradient Descent (SGD), mini-batch gradient descent, Momentum, Nesterov (specifically, stochastic gradient descent with momentum), Adaptive Gradient Descent (Adagrad), Adadelta, root mean square propagation (RMSprop), and Adaptive Moment Estimation (Adam). During error backpropagation, these algorithms calculate the gradient by taking the derivative / partial derivative of the error / loss obtained from the loss function with respect to the current neuron, adding the learning rate and the effects of previous gradients / derivatives / partial derivatives, and then propagating this gradient to the previous layer.

[0075] To facilitate a better understanding of the embodiments of this application, the AI ​​unit or AI model will be described.

[0076] The AI ​​model described in this application embodiment may also be referred to as an AI unit, AI model / AI unit, machine learning (ML) model, ML unit, AI structure, AI function, AI characteristic, neural network, neural network function, neural network functionality, etc. Alternatively, the AI ​​model described in this application may also refer to a processing unit capable of implementing specific algorithms, formulas, processing flows, capabilities, etc. related to AI. Alternatively, the AI ​​model described in this application may be a processing method, algorithm, function, module, or unit for a specific dataset. Alternatively, the AI ​​model described in this application may be a processing method, algorithm, function, module, or unit running on AI / ML related hardware such as a graphics processing unit (GPU), neural processing unit (NPU), tensor processing unit (TPU), or application-specific integrated circuit (ASIC). This application does not specifically limit this. Optionally, the specific dataset includes the input or output of the AI ​​model.

[0077] The AI ​​unit described in this application can be an AI model constructed with fully connected layers as the basic unit, or an AI model that includes fully connected layers as the basic unit, or an AI model constructed with CNNs as the basic unit, or an AI model constructed with Transformers as the basic unit, or an AI model that includes Transformers as the basic unit. Furthermore, any AI units subsequently evolved from this model should also be within the scope of protection of this application.

[0078] The identifier of the AI ​​model described in the embodiments of this application may be an AI unit identifier, an AI structure identifier, an AI algorithm identifier, or an identifier of a specific dataset associated with the AI ​​model described in this application, or an identifier of a specific scenario, environment, channel characteristics, or device related to the AI ​​model described in this application, or an identifier of a function, characteristic, capability, or module related to the AI ​​model described in this application. This application does not make any specific limitations on these.

[0079] To facilitate a better understanding of the embodiments of this application, the training collaboration types of the CSI compression model are explained.

[0080] CSI compression is a typical two-end model use case, meaning the complete CSI compression model needs to be deployed on different network nodes. Currently, most considerations involve deploying the encoder on the user interface (UE) side and the decoder on the network (NW) side. The (sub)models deployed on multiple nodes need to be paired with each other to function correctly. Considering the characteristics of two-end models, the 3rd Generation Partnership Project (3GPP) introduced the following basic training collaboration types:

[0081] 1) Joint training at a single entity (or type 1)

[0082] The training framework refers to training a complete encoder and decoder model on a network node (UE, NW, or a third-party server node, etc.), and then deploying the corresponding model modules to the target node through methods such as model transfer (e.g., transferring the encoder part to the UE and the decoder part to the NW).

[0083] 2) Joint training at multiple entities (or type 2)

[0084] The training framework refers to a process in which multiple nodes participate in the training process, with each node independently calculating the forward / backward propagation information required for its local model training and updating its own model parameters. Since the training process requires forward / backward propagation of the entire model (including the encoder and decoder), the participating nodes need to exchange the corresponding forward / backward propagation information. After training is complete, no further model transfer is required between nodes.

[0085] 3) Separate training (or type 3) on multiple nodes

[0086] The training framework refers to first training a reference model on a certain node, then sending the relevant information of the reference model to the target node, and finally, the target node training its own model based on this information, thus ensuring that the models of each node (sub-model) can be paired and used together. For example, the NW side first trains a complete encoder-decoder model and determines that the obtained decoder is the decoder to be actually used in the future. Then, the relevant information of the encoder corresponding to the decoder (usually the encoder's input and output data) is sent to the UE side, and the UE side trains its own encoder based on this information. This training framework can be further subdivided into two cases: UE-first training and NW-first training. UE-first training means that the complete model is trained on the UE side first, and then the information needed for the NW to train the matching model (usually the input and output data of the model to be trained on the NW side) is sent to the NW side. Conversely, NW-first training means that the complete model is trained on the NW side first, and then the information needed for the UE to train the matching model (usually the input and output data of the model to be trained on the UE side) is sent to the UE side.

[0087] The wireless communication method provided in this application will be described in detail below with reference to the accompanying drawings and through some embodiments and application scenarios.

[0088] Figure 2 is a schematic flowchart of a wireless communication method 200 according to an embodiment of this application. As shown in Figure 2, the wireless communication method 200 may include at least some of the following:

[0089] S210, the first device obtains first cache information, or the target AI unit obtains first cache information, the first cache information is used to obtain the output information associated with the target AI unit, or the first cache information is used to obtain the first target CSI associated with the target AI unit;

[0090] The first cached information includes at least one of the following: CSI information of at least one historical time unit, at least one CSI intermediate state information, and input information associated with the target AI unit in at least one time unit;

[0091] Wherein, the CSI intermediate state information in the at least one CSI intermediate state information is intermediate information obtained based on the target AI unit;

[0092] Wherein, the first cache information is associated with at least one of the following: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and the target AI unit;

[0093] The time information includes at least one of the following: the at least one historical time unit, the time unit corresponding to the at least one CSI intermediate state information, and the at least one time unit;

[0094] In the event that the first timer expires, some or all of the information in the first cached information is cleared or reset;

[0095] The configuration information of the observation window includes at least one of the following: at least one of the time-domain interval, quantity, and start position corresponding to the CSI information of the at least one historical time unit; at least one of the time-domain interval, quantity, and start position corresponding to the at least one CSI intermediate state information; and at least one of the time-domain interval, quantity, and start position corresponding to the at least one time unit.

[0096] Wherein, the first information satisfies at least one of the following: instructing to clear or reset the first cache information, instructing to clear or reset the condition information of the first cache information, configuring the type of the first cache information, configuring the number of instances in the first cache information, instructing the association between the first cache information and the layer information, and instructing the association between the first cache information and the rank information.

[0097] It should be understood that Figure 2 illustrates the steps or operations of the wireless communication method 200, but these steps or operations are merely examples, and other operations or variations of the various operations in Figure 2 may also be performed in this application.

[0098] In some embodiments, the target AI unit acquires first cached information, including:

[0099] The target AI unit obtains the first cache information from the registers, memory, high-level units, or other sub-modules of the first device.

[0100] In some embodiments, the target AI unit is deployed in the first device.

[0101] In some embodiments, the first cached information is already stored in the registers, memory, higher-level units, or other sub-modules of the first device before the output information associated with the target AI unit is obtained, or before the first target CSI associated with the target AI unit is obtained.

[0102] In some embodiments, the first device acquiring the first cache information includes: the first device acquiring the first cache information from the registers, memory, higher-level units, or other sub-modules of the first device, wherein the first cache information is used to acquire output information associated with the target AI unit, or the first cache information is used to acquire a first target CSI associated with the target AI unit; furthermore, it is understood that the first cache information used to acquire the output information associated with the target AI unit, or the first cache information used to acquire the first target CSI associated with the target AI unit, is already stored in the registers, memory, higher-level units, or other sub-modules of the first device before it is used to acquire the first target CSI associated with the target AI unit.

[0103] In some embodiments, the target AI unit obtaining the first cached information includes: the target AI unit obtaining the first cached information from the registers, memory, higher-level units, or other sub-modules of the first device, wherein the first cached information is used to obtain output information associated with the target AI unit, or the first cached information is used to obtain a first target CSI associated with the target AI unit; optionally, the target AI unit is deployed in the first device. Furthermore, it is understood that the first cached information used to obtain the output information associated with the target AI unit, or the first cached information used to obtain the first target CSI associated with the target AI unit, is already stored in the registers, memory, higher-level units, or other sub-modules of the first device before being used to obtain the first target CSI associated with the target AI unit.

[0104] In some embodiments, the first device maintains or stores the first cache information, which is used by the target AI unit to obtain output information associated with the target AI unit, or the first cache information is used to obtain a first target CSI associated with the target AI unit; optionally, the target AI unit is deployed in the first device.

[0105] In some embodiments, the higher layers or physical layers of the first device maintain or store the first cache information, which is used by the target AI unit to obtain output information associated with the target AI unit, or the first cache information is used to obtain a first target CSI associated with the target AI unit; optionally, the target AI unit is deployed in the first device.

[0106] In some embodiments, the first device or target AI unit may obtain the first cached information from other devices. For example, the first device or target AI unit may obtain the first cached information from a network-side device or a server. Optionally, the first device or target AI unit may first obtain the information from other devices into the registers or memory of the first device, and then extract the output information associated with the target AI unit from the registers or memory of the first device, or the first cached information may be used to obtain the first target CSI associated with the target AI unit.

[0107] In some embodiments, the first device or target AI unit obtaining the first cache information includes: obtaining the associated information of the first cache information; such as obtaining at least one of the following: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and target AI unit.

[0108] In some embodiments, the first device or the target AI unit may obtain the associated information of the first cache from other devices, such as at least one of the following from other devices: layer information, rank information, time information, first timer, configuration information of the observation window, first information, or the target AI unit.

[0109] In some embodiments, the first device or the target AI unit obtains first cache information from the registers, memory, higher-level units or other sub-modules of the first device, and obtains associated information of the first cache information from other devices, such as obtaining at least one of the following from other devices: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and target AI unit.

[0110] In some embodiments, the first device or the target AI unit may obtain the associated information of the first cached information from the protocol, such as obtaining at least one of the following from the protocol: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and target AI unit.

[0111] In some embodiments, the first device or target AI unit obtains first cached information from the registers, memory, higher-level units, or other sub-modules of the first device, and obtains associated information of the first cached information from a protocol, such as obtaining at least one of the following from the protocol: layer information, rank information, time information, first timer, configuration information of the observation window, first information, or the target AI unit.

[0112] In some embodiments, the first device or target AI unit obtains the association information of the first cached information from a higher layer of the first device, such as the first timer, which may be maintained by a higher layer. Optionally, some of the association information of the first cached information is obtained from other devices, and some of the first cached information is obtained from the first device; or, some of the association information of the first cached information is obtained from other devices, and some of the first cached information is obtained from a protocol.

[0113] In some embodiments, the first cached information is used to obtain a first target CSI associated with the target AI unit, including:

[0114] The first device or the target AI unit determines the first target CSI associated with the target AI unit based on the first cache information; or the first device inputs the first cache information into the target AI unit to obtain the first target CSI associated with the target AI unit.

[0115] Optionally, the first target CSI is information directly output by the target AI unit, or the first target CSI is information determined based on information directly output by the target AI unit.

[0116] Optionally, the association of the first target CSI with the target AI unit can be understood as either explicit or implicit. As long as the first target CSI is obtained in conjunction with the target AI unit, it is considered an associated first target CSI with the target AI unit. Explicitly, for example, this could be indicated in the first target CSI configuration information or reporting information, or the first target CSI report could include the relevant information of the target AI unit. Implicitly, for example, this could be when the first device actually uses the target AI unit.

[0117] Specifically, for example, the first cached information includes at least one of the following: CSI information for at least one historical time unit, and at least one CSI intermediate state information; in this case, the first device determines the first target CSI associated with the target AI unit based on the first cached information; or the first device inputs the first cached information into the target AI unit to obtain the first target CSI associated with the target AI unit, or the first device inputs the first cached information and the first CSI into the target AI unit to obtain the first target CSI associated with the target AI unit.

[0118] In some embodiments, the first cache information is used to obtain output information associated with the target AI unit, including:

[0119] The first device or the target AI unit determines the output information associated with the target AI unit based on the first cache information; or the first device inputs the first cache information into the target AI unit to obtain the output information associated with the target AI unit; or the first device inputs the first cache information and the first input information into the target AI unit to obtain the output information associated with the target AI unit.

[0120] Optionally, the output information associated with the target AI unit is information directly output by the target AI unit, or the output information associated with the target AI unit is information determined based on information directly output by the target AI unit.

[0121] It is understood that the output information associated with the target AI unit is information determined based on the information directly output by the target AI unit. This information can be obtained by quantizing the information directly output by the target AI unit, by selecting from the information directly output by the target AI unit, or by performing a first calculation on the information directly output by the target AI unit. Optionally, the first calculation can be encoding or mathematical operations.

[0122] For example, the first cache information includes the input information of the target AI unit at the at least one time unit. In this case, the first device or the target AI unit determines the output information associated with the target AI unit based on the first cache information, or the first device inputs the first cache information into the target AI unit to obtain the output information associated with the target AI unit.

[0123] In some embodiments, the inputs or outputs of the target AI unit may include at least one of the following:

[0124] Reference signal, channel-bearing signal, channel state information, beam information, channel prediction information, interference information, positioning information, trajectory information, codebook information, original channel information, specified prediction information and management information, and control signaling.

[0125] In some embodiments, the target AI unit may perform at least one of the following functions:

[0126] Reference signal processing includes signal detection, filtering, equalization, etc., wherein the reference signal includes at least one of the following: demodulation reference signal (DMRS), sounding reference signal (SRS), synchronization signal block (SSB), channel state information reference signal (CSI-RS), tracking reference signal (TRS), positioning reference signal (PRS), and phase tracking reference signal (PT-RS).

[0127] The transmission, reception, demodulation, or transmission of a channel or signal, wherein the channel includes the Physical Downlink Control Channel (PDCCH), Physical Downlink Shared Channel (PDSCH), Physical Uplink Control Channel (PUCCH), Physical Uplink Shared Channel (PUSCH), Physical Random Access Channel (PRACH), PBCH, etc.

[0128] Channel state information acquisition includes channel state information feedback and / or frequency division duplex (FDD) uplink and downlink reciprocity. Channel state information feedback includes channel-related information, channel matrix-related information, channel characteristic information, channel matrix characteristic information, precoding matrix indicator (PMI), rank indicator (RI), CSI-RS resource indicator (CRI), channel quality indicator (CQI), and layer indicator (LI). For FDD systems, based on partial reciprocity, the base station obtains angle and delay information from the uplink channel. This angle and delay information can be communicated to the UE via CSI-RS precoding or direct indication. The UE reports this information based on the base station's indication or selects and reports it within the indicated range, thereby reducing the UE's computational load and CSI reporting overhead.

[0129] Beam management includes beam measurement, beam reporting, beam prediction, beam failure detection, beam failure recovery, and new beam indication during beam failure recovery.

[0130] Channel prediction includes prediction of channel state information, beam prediction, etc.

[0131] Channel / source coding and decoding, such as channel coding, channel decoding, source coding, source decoding, joint source-channel coding, and joint source-channel decoding;

[0132] Interference suppression includes intra-cell interference, inter-cell interference, out-of-band interference, and intermodulation interference.

[0133] Positioning is the estimation of the UE's specific location (including horizontal and / or vertical position), location-related information, or possible future trajectory, or auxiliary location or trajectory estimation information, using measurement information from reference signals (such as SRS or PRS).

[0134] Forecasting and managing high-level business and parameters, including throughput, required data packet size, business demands, mobility, noise information, etc.

[0135] Parsing of control signaling, such as power control signaling and beam management signaling.

[0136] Optionally, in some embodiments, if the output information of the target AI unit is the positioning information of the first device, the first input information includes the location-related measurement information of the first device, such as delay profile (DP), power delay profile (PDP), or multipath measurement information, reference signal time difference (RSTD) measurement information, or time of arrival (TOA) measurement information, etc.

[0137] Optionally, in some embodiments, if the output information of the target AI unit is the beam information of the first device, the first input information includes the beam-related measurement information of the first device, such as beam indication information or beam energy information.

[0138] In this embodiment, a first device acquires first cache information, or a target AI unit acquires first cache information. The first cache information is used to acquire output information associated with the target AI unit, or to acquire a first target CSI associated with the target AI unit. Specifically, the first cache information is associated with at least one of the following: layer information, rank information, time information, a first timer, configuration information of the observation window, first information, and the target AI unit. The first cache information is introduced during the process of acquiring the output information associated with the target AI unit or the first target CSI associated with the target AI unit, thereby improving the accuracy of the target AI unit's prediction. Furthermore, the first cache information is associated with time information, the first timer, and the configuration information of the observation window, thereby enabling prediction and / or compression and decompression (such as CSI compression and CSI decompression) in the spatiotemporal frequency domain. Introducing time domain information allows the encoder and decoder to better align the model inference schemes at both ends, and / or can be applied to prediction and / or compression and decompression (such as CSI compression and CSI decompression) based on the target AI unit in high-rank (such as rank ≥ 2) cases, and improves the performance of prediction and / or compression and decompression in the spatiotemporal frequency domain (such as improving the performance of CSI compression and CSI decompression in the spatiotemporal frequency domain).

[0139] In some embodiments, the first cache information is associated with layer information and rank information, and is used to indicate the layer information and rank information associated with the first cache information, respectively.

[0140] In some embodiments, the higher layers or physical layers of the first device maintain or store the association between the first cache information and the layer information and rank information.

[0141] In some embodiments, the first cache information is associated with time information and a first timer, and the first timer is started, run, or terminated in conjunction with the time information.

[0142] In some embodiments, the first cache information is associated with the target AI unit and the first timer, and the first timer is activated or terminated in conjunction with the target AI unit.

[0143] In some embodiments, a higher layer or physical layer of the first device maintains the first cache information and the first timer, wherein the first timer is used to clear or associate the first cache information.

[0144] In some embodiments, a higher layer or physical layer of the first device maintains the first cache information and the first timer, and the first timer is restarted based on the first cache information. For example, the first timer is restarted when the first cache information is reset.

[0145] In some embodiments, the first cache information is associated with the configuration information of the observation window, and the first cache information is obtained according to the configuration information of the observation window.

[0146] In some embodiments, the first cache information is associated with the configuration information of the target AI unit and the observation window, wherein the observation window is associated with the target AI unit. Optionally, the configuration information of the target AI unit and the configuration information of the observation window are both included in the first target CSI report or configuration information, or the configuration information of the target AI unit includes the configuration information of the observation window.

[0147] In some embodiments, the first information may be configuration information and / or indication information, or the first information may be regarded as configuration information and / or indication information, or the first information may implement the functions of configuration information and / or indication information. It is understood that the first information is obtained from a network-side device or other devices. If the first cached information is also associated with other information, the associated information may be obtained from a sub-unit or higher layer of the first device, and the first information is obtained from a network-side device or other devices.

[0148] In some embodiments, the wireless communication method 200 includes:

[0149] The first device receives the first information, and the first device obtains the first cache information based on the first information.

[0150] In some embodiments, the first device may be a terminal, a network-side device, or a test device.

[0151] In some embodiments, the target AI unit includes at least one of the following:

[0152] The terminal is used to acquire the first AI unit of the first target CSI;

[0153] Network-side devices are used to acquire the second AI unit for reconstructing CSI;

[0154] The terminal is used to obtain the reference AI unit of the first target CSI;

[0155] Network-side devices are used to obtain reference AI units for reconstructing CSI;

[0156] The third AI unit used in testing by the terminal, network-side equipment, or testing equipment;

[0157] The terminal, network-side device, or test equipment is used to match the fifth AI unit of the fourth AI unit used in the test.

[0158] In some embodiments, the first device includes at least one of the following: a terminal, a network-side device, and a test device.

[0159] In some embodiments, the first device may also be a server, such as an over-the-top (OTT) server, an operation administration and maintenance (OAM) server, etc.

[0160] For example, the target AI unit includes a first AI unit used by the terminal to acquire a first target CSI. In this case, the first device can be the terminal. Optionally, in this case, the first AI unit can be an AI unit at the encoding end, and the first target CSI is a compressed CSI.

[0161] For example, the target AI unit includes a second AI unit used by a network-side device to acquire the reconstructed CSI. In this case, the first device can be a network-side device. Optionally, in this case, the second AI unit can be an AI unit at the decoding end, and the second target CSI is the decoded CSI.

[0162] For example, the target AI unit includes a first AI unit used by the terminal to acquire a first target CSI. In this case, the first device can be the terminal. Optionally, in this case, the first AI unit can be an AI unit at the encoding end, and the first target CSI is a compressed and predicted CSI. The first AI unit includes a compression function and a prediction function. The compression function and the prediction function can be a single AI subunit or two cascaded AI subunits.

[0163] For example, the target AI unit includes a second AI unit used by a network-side device to acquire the reconstructed CSI. In this case, the first device can be a network-side device. Optionally, in this case, the second AI unit can be an AI unit at the decoding end, and the second target CSI is the decompressed and predicted CSI. The first AI unit includes decompression and prediction functions. The decompression and prediction functions can be a single AI subunit or two cascaded AI subunits.

[0164] For example, the target AI unit includes a reference AI unit for the terminal to acquire a first target CSI, in which case the first device can be the terminal.

[0165] For example, the target AI unit includes a reference AI unit used by a network-side device to acquire reconstructed CSI, in which case the first device can be a network-side device.

[0166] For example, the target AI unit may include a third AI unit used by the terminal in the test, in which case the first device may be the terminal.

[0167] For example, the target AI unit may include a third AI unit used by the network-side device in the test, in which case the first device may be the network-side device.

[0168] For example, the target AI unit may include a third AI unit used by the test device in the test, in which case the first device may be the test device.

[0169] For example, the target AI unit includes a fifth AI unit used by the terminal to match the fourth AI unit used in the test. In this case, the first device can be the terminal.

[0170] For example, the target AI unit includes a fifth AI unit used by a network-side device to match a fourth AI unit used in the test. In this case, the first device can be a network-side device.

[0171] For example, the target AI unit includes a fifth AI unit that is matched with a fourth AI unit used in the test, in which case the first device can be the test device.

[0172] In some embodiments, the conditions for clearing or resetting the first cache information include, but are not limited to, at least one of the following: the target AI unit is invalidated, the target AI unit is deactivated, or the first timer times out or expires.

[0173] In some embodiments, the condition information for clearing or resetting the first cache information includes, but is not limited to, at least one of the following: a predefined first event, such as the first CSI not being acquired or measured at the current time K, or the first CSI not being acquired or measured for Z consecutive times.

[0174] The cached information described in this application embodiment can be understood as intermediate information of CSI processing, or intermediate state information of CSI processing at a historical time, or intermediate state information output by different layers, or intermediate state information output by different ranks, or CSI information or channel information acquired at a historical time.

[0175] In some embodiments, the first device stores at most R copies of the first cached information in a time unit;

[0176] Wherein, the R first cached information pieces are associated with one target AI unit, or the R first cached information pieces are associated with different target AI units, where R is a positive integer. For example, R ≥ 2.

[0177] In some embodiments, the first device stores at most R1 instances of the first cached information in a single time unit, where R1 is associated with one target AI unit, and the first device stores at most R2 instances of the first cached information in a single time unit. The R2 instances of the first cached information are associated with all target AI units, and R1 and R2 are positive integers. For example, R2 ≥ R1.

[0178] Optionally, the value of R is configured by the network side, or the value of R is agreed upon by the protocol, or the value of R is determined based on the capability information reported by the first device.

[0179] For example, R first cache information pieces are associated with a target AI unit. The R first cache information pieces may include, but are not limited to, at least one of the following: layer-specific first cache information, layer-common first cache information, rank-common first cache information, and rank-specific first cache information. Optionally, if the first cache information is associated with layer information, the R first cache information pieces correspond to different layers. For example, layer-specific first cache information, where each layer has one first cache information piece, and the corresponding first cache only stores the first cache information associated with the current layer. Optionally, if the first cache information is associated with rank information, the R first cache information pieces correspond to different ranks. For example, rank-specific first cache information, where each rank has one first cache information piece, and the corresponding first cache only stores the first cache information associated with the current rank. Optionally, if the first cache information is associated with rank information and layer information, the R first cache information pieces correspond to different ranks and / or layers. For example, the cache unit of first cache information with rank 2 and layer 1 is different from that of first cache information with rank 4 and layer 1.

[0180] The time unit described in the embodiments of this application includes, but is not limited to, at least one of the following:

[0181] Symbol, time slot, microtime slot, subframe, frame, millisecond, microsecond, second, minute, hour, specified or defined time, the time calculated by the target AI unit, the time run by the target AI unit.

[0182] In some implementations, the CSI compression and decompression scenarios in the space-time-frequency domain applicable to the embodiments of this application can be as shown in Figure 3. The encoder (e.g., an AI-based encoder) uses CSI 0 of time slot 0 to obtain intermediate state information 0 and encoded CSI 0, where intermediate state information 0 is the internal information output by the encoder before outputting encoded CSI 0. The decoder (e.g., an AI-based decoder) uses encoded CSI 0 to obtain intermediate state information 0' and CSI 0' of time slot 0, where intermediate state information 0' is the internal information output by the decoder before outputting CSI 0' of time slot 0. The encoder (e.g., an AI-based encoder) uses CSI 1 of time slot 1 and intermediate state information 0 to obtain intermediate state information 1 and encoded CSI 1, where intermediate state information 1 is the internal information output by the encoder before outputting encoded CSI 1. The encoder uses historical information (i.e., intermediate state information) to assist in encoding; the decoder (such as an AI-based decoder) uses the encoded CSI 1 and intermediate state information 0' to obtain intermediate state information 1' and CSI 1' of time slot 1. Intermediate state information 1' is the intermediate information output by the decoder before CSI 1' of output time slot 1. The encoder (such as an AI-based encoder) uses CSI 2 of time slot 2 and intermediate state information 1 to obtain intermediate state information 2 and encoded CSI 2. Intermediate state information 2 is the intermediate information output by the encoder before outputting encoded CSI 2. The decoder uses historical information (i.e., intermediate state information) to assist in encoding; the decoder (such as an AI-based decoder) uses the encoded CSI 2 and intermediate state information 1' to obtain intermediate state information 2' and CSI 2' of time slot 2. Intermediate state information 2' is the intermediate information output by the decoder before CSI 2 of output time slot 2. The intermediate information output before 2' is the historical information (i.e., intermediate state information) used by the decoder to assist in decoding; the encoder (such as an AI unit-based encoder) uses CSI 3 of time slot 3 and intermediate state information 2 to obtain the encoded CSI 3, that is, the encoder uses historical information (i.e., intermediate state information) to assist in encoding; the decoder (such as an AI unit-based decoder) uses the encoded CSI 3 and intermediate state information 2' to obtain CSI 3' of time slot 3, that is, the decoder uses historical information (i.e., intermediate state information) to assist in decoding.

[0183] It should be noted that time slots 0, 1, 2, and 3 are merely logical identifiers for the order of time slots and do not necessarily correspond to physical time slots 0, 1, 2, and 3. They can correspond to physical time slots N, N+K, N+2K, and N+3K, or physical time slots N, N+K1, N+K2, and N+K3.

[0184] It should be noted that, as shown in Figure 3, the intermediate state information of the encoder is generally only transmitted between encoders, and the intermediate state information of the decoder is generally only transmitted between decoders. Optionally, the associated information of the encoder's intermediate state information, such as layer information, rank information, time information, first timer, observation window configuration information, first information, or the target AI unit, can interact between the encoder and decoder so that the intermediate state information of the encoder used by the two ends (encoder and decoder) is consistent in time and / or configuration.

[0185] In some other implementations, the CSI compression and decompression scenarios applicable to the embodiments of this application in the spatiotemporal frequency domain can be as shown in Figure 4. The spatiotemporal frequency domain CSI compression uses the CSI information at time unit K and the CSI information at time unit K-1*P, time unit K-2*P, ..., time unit KQ*P. Specifically, as shown in Figure 4, the spatiotemporal frequency domain CSI compression uses PMI 3 at time unit K and PMI 2 at time unit K-1*P, PMI 1 at time unit K-2*P, and PMI 0 at time unit K-3*P. It should be understood that the CSI information output by the decoder can be the CSI information at time unit K, or it can be the CSI information at future times, such as the CSI information at time unit K+V, time unit K+V+D, ..., time unit K+V+Z*D. If the CSI information output by the decoder is the CSI information for time unit K, then the target AI unit is a CSI compression and decompression model. If the CSI information output by the decoder is for a future time, such as the CSI information for time unit K+V or time unit K+V+D, then the target AI unit is a CSI compression and prediction model. Here, K is an integer greater than or equal to 0, P and Q are both positive integers, and V, Z, and D are all positive integers. Optionally, the value of Z*D is the same as the value of Q*P. Of course, the value of Z*D can also be different from the value of Q*P; this embodiment of the application does not limit this.

[0186] Optionally, the CSI information can be a codebook or raw channel information. The codebook can be the raw channel information or information obtained by performing SVD decomposition on the signal measurement information, or it can be information corresponding to a codebook matrix, or information obtained by transforming or processing the codebook.

[0187] In some embodiments, the layer information includes at least one of the following: a layer identifier and layer type information; wherein the layer type includes at least one of the following: a shared layer (Layer common) and a specific layer (Layer specific). It is understood that the layer refers to the layer number corresponding to the CSI information.

[0188] The shared layer (rank common) described in this application refers to the CSI information of different layers corresponding to the same AI model or the same first cache information. For example, layer 0 and layer 4 correspond to the same model; or layer 1 uses the first cache information of layer 0, or the first cache information is obtained based on the cache information obtained from layer 0 and layer 1.

[0189] The Rank-specific layers described in this application are: one layer corresponds to one model, and different layers (rank) correspond to different models, for example, layer 0 and layer 4 correspond to different models; or, one layer corresponds to one first cache information, and different layers correspond to different first cache information.

[0190] In some embodiments, the rank information includes at least one of the following: a rank identifier, and rank type information; wherein the rank type includes at least one of the following: shared rank (rank common), and rank specific (rank specific). It is understood that the rank refers to the rank number corresponding to the CSI information.

[0191] The shared rank (rank common) mentioned in this application embodiment refers to CSI information of different ranks corresponding to the same AI model or the same first cache information, or it refers to CSI information of different ranks or the same layer corresponding to the same AI model or the same first cache information.

[0192] The specific rank described in this application means that one rank corresponds to one model, and different ranks correspond to different models. For example, rank 1, rank 2, rank 3, and rank 4 correspond to different models; or, one rank corresponds to one first cache information, and different ranks correspond to different first cache information.

[0193] The model's input corresponds to all rank-related layers. Optionally, for a single layer, its channel or CSI information passes through the corresponding sub-model within the entire rank model. For example, the channel or CSI information of Rank 2 Layer 0 and Rank 3 Layer 0 pass through different models. Alternatively, layers of different ranks may be associated with the same target AI model, but each input corresponds to a rank-related layer.

[0194] It should be understood that the model or sub-model in the relevant description of shared rank can be the target AI unit described in this application.

[0195] The shared layer described in the embodiments of this application:

[0196] For different ranks, the same layer corresponds to the same model. For example, the channel or CSI information of Layer 0, Layer 1, Layer 2, and Layer 3 of Rank 4 corresponds to different models, but the channel or CSI information of Layer 0 of Rank 2 and Layer 0 of Rank 3 correspond to the same model. Alternatively, for the same rank, the same layer corresponds to the same model. For example, the channel or CSI information of Layer 0 of Rank 2 and Layer 0 of Rank 2 correspond to the same model, but the channel or CSI information of Layer 0 of Rank 4 and Layer 0 of Rank 2 correspond to different models.

[0197] The input of the model corresponds to a Layer;

[0198] For a layer, its channel or CSI information passes through the complete model of that layer, or the sub-model of the corresponding layer in the entire rank model.

[0199] It should be understood that the model or sub-model in the relevant description of the shared layer can be the target AI unit described in this application.

[0200] The shared rank and shared layer (common) described in the embodiments of this application are as follows:

[0201] Different ranks and different layers correspond to the same model. For example, Rank 4 Layer 0, Rank 4 Layer 1, Rank 4 Layer 2, and Rank 4 Layer 3 correspond to the same model, and the channel or CSI information of Rank 2 Layer 0 and Rank 3 Layer 2 correspond to the same model.

[0202] The model's input is a single layer; for each layer, its channel or CSI information passes through the complete model of that layer. Alternatively, the model's input is all layers corresponding to a given rank; the channels or codebooks corresponding to all layers pass through a single complete model.

[0203] It should be understood that the model or sub-model in the relevant descriptions of shared rank and shared layer can be the target AI unit described in this application.

[0204] In some implementations, a Rank 4 model (a single encoder or a single decoder) can be as shown in Figure 5.

[0205] As shown in Figure 5, in scheme 0, it consists of a maximum of 4 single-layer models, namely the model of layer 0 + the model of layer 1 + the model of layer 2 + the model of layer 3.

[0206] As shown in Figure 5, in Scheme 1, a model with one shared layer (layer common) is applicable to multiple layers, but each layer has its own dedicated buffer.

[0207] As shown in Figure 5, in Scheme 2, there is a buffer shared by multiple layers. The buffer can be obtained through the buffer of the encoder of one of the layers, or it can be the buffer information obtained by the joint operation of multiple layers.

[0208] In some embodiments, the wireless communication method 200 further includes: the first device inputting the first cached information and the first CSI into the target AI unit to obtain the first target CSI. That is, in this embodiment, the first CSI is used to determine the first target CSI.

[0209] For example, the first device shown is an encoding device, the first target CSI is the information obtained after encoding the first CSI, and after obtaining the first target CSI, the first device feeds back the first target CSI to the decoding device, so that the decoding device can decode the first CSI based on the first target CSI.

[0210] For example, the first device shown is a decoding device, and the first CSI is the encoded information fed back by the encoding device. The first device inputs the first buffer information and the first CSI into the target AI unit to obtain the first target CSI. That is, the first target CSI is the information obtained after decoding the first CSI.

[0211] For example, the first device shown is an encoding end device, and the first CSI is raw channel information or codebook information.

[0212] In some embodiments, the first cache information includes CSI intermediate state information of time unit KN; and / or,

[0213] The first cached information includes the CSI intermediate state information of time unit KL;

[0214] Wherein, the first target CSI corresponds to time unit K, or the first CSI corresponds to time unit K;

[0215] Where K is an integer greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

[0216] Optionally, in some embodiments, the first cache information includes multiple of the following: CSI intermediate state information for time unit KN, CSI intermediate state information for time unit K-2N, CSI intermediate state information for time unit K-3N, and CSI intermediate state information for time unit K-4N.

[0217] Optionally, in some embodiments, the first cached information includes CSI intermediate state information for time unit K-L1, CSI intermediate state information for time unit K-L2, etc.; wherein L1 and L2 are integer multiples of N.

[0218] In this embodiment, the first target CSI corresponds to time unit K, or the first CSI corresponds to time unit K, and the first cache information includes the intermediate state information of the CSI in time unit KN, and / or the first cache information includes the intermediate state information of the CSI in time unit KL, so that the first device can determine the content of the first cache information corresponding to the time unit based on the time unit corresponding to the first target CSI or the time unit corresponding to the first CSI.

[0219] In some embodiments, if the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cache information includes the intermediate state information of the CSI corresponding to layer X; wherein X is an integer greater than or equal to 0. Optionally, the first cache information includes intermediate state information of the CSI corresponding to at least two time units, and the intermediate state information of the CSI corresponding to the at least two time units is associated with the same layer.

[0220] In this embodiment, the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, and the first cache information includes the intermediate state information of the CSI corresponding to layer X, so that the first device can determine the content of the first cache information based on the CSI information of the layer corresponding to the first target CSI or the CSI information of the layer corresponding to the first CSI; or, the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, and the first cache information includes the intermediate state information of the CSI corresponding to layer X, so that the first device can obtain the first target CSI of layer X based on the intermediate state information of the CSI corresponding to layer X.

[0221] In some embodiments, if the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cached information includes the intermediate CSI state information corresponding to layer Y; wherein X and Y are both integers greater than or equal to 0. Optionally, the first cached information includes the intermediate CSI state information corresponding to at least two time units, wherein the intermediate CSI state information corresponding to the at least two time units is associated with different layers. Alternatively, the first cached information includes the intermediate CSI state information corresponding to at least two layers.

[0222] In this embodiment, the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, and the first cache information includes the intermediate state information of the CSI corresponding to layer Y. Therefore, the first device can determine the content of the first cache information based on the CSI information of the layer corresponding to the first target CSI or the CSI information of the layer corresponding to the first CSI. Alternatively, the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, and the first cache information includes the intermediate state information of the CSI corresponding to layer Y. Therefore, the first device can obtain the first target CSI of layer X based on the intermediate state information of the CSI corresponding to layer Y. Optionally, the first device can obtain the first target CSI and the intermediate state information of the CSI corresponding to layer Y based on the intermediate state information of the CSI corresponding to layer Y and the first CSI of layer Y. The first device can obtain the first target CSI and the intermediate state information of the CSI corresponding to layer Y based on the intermediate state information of the CSI corresponding to layer Y and the first CSI of layer Y / layer X.

[0223] In some embodiments, the CSI intermediate state information corresponding to layer Y is at least one of the following:

[0224] The CSI intermediate state information corresponding to layer Y obtained from historical time;

[0225] The CSI intermediate state information corresponding to layer Y obtained at the current time;

[0226] Wherein, layer Y is layer 0, or layer Y is the highest layer, or layer Y is a layer configured by the network side or specified by the protocol;

[0227] The historical time includes at least one of the following: time unit KN, time unit KL;

[0228] Wherein, the current time is time unit K;

[0229] Wherein, the first target CSI corresponds to the time unit K, or the first CSI corresponds to the time unit K;

[0230] Where K is an integer greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

[0231] For example, the CSI intermediate state information corresponding to layer Y includes CSI intermediate state information corresponding to layer Y obtained at multiple historical times.

[0232] For example, the CSI intermediate state information corresponding to layer Y obtained from multiple historical times includes at least one of the following: associated layer information, rank information, time information, first timer, observation window configuration information, first information, and the target AI unit.

[0233] Optionally, the plurality of historical time intervals are associated with the configuration information of the observation window, such as being consistent with the period or interval of the observation window.

[0234] Optionally, at least one of the plurality of historical times is included in the time information, such as the most recent historical time being the time information.

[0235] Optionally, if layer Y is configured on the network side, then the first information includes indication information for layer Y.

[0236] For example, if the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cached information includes the intermediate CSI state information corresponding to layer Y obtained in historical time; wherein, layer Y is layer 0 among all layers in the historical time, for example, layer 0, layer 1, layer 2, and layer 3 in the historical time all use the intermediate CSI state information corresponding to layer 0; or, layer Y is the highest layer among all layers in the historical time, for example, layer 0, layer 1, layer 2, and layer 3 in the historical time all use the intermediate CSI state information corresponding to layer 3.

[0237] For example, if the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cached information includes the intermediate CSI state information corresponding to layer Y obtained at the current time; wherein, layer Y is layer 0 among all layers at the current time, for example, layer 0, layer 1, layer 2, and layer 3 at the current time all use the intermediate CSI state information corresponding to layer 0; or, layer Y is the highest layer among all layers at the current time, for example, layer 0, layer 1, layer 2, and layer 3 at the current time all use the intermediate CSI state information corresponding to layer 3.

[0238] For example, if the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cache information includes the intermediate CSI state information of layer Y obtained at the current time; wherein, layer Y is the layer information obtained by mixing all layers at the current time. For example, layer 0, layer 1, layer 2, and layer 3 at the current time all use the mixed CSI intermediate state information of the intermediate CSI state information corresponding to layer 0, layer 1, layer 2, and layer 3.

[0239] For example, if the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cache information includes the intermediate CSI state information of layer Y obtained at the current time; wherein, layer Y is the layer information obtained by mixing some layers among all layers at the current time. For example, layer 0, layer 1, layer 2, and layer 3 at the current time all use the mixed CSI intermediate state information of the intermediate CSI information corresponding to layer 0 and layer 1.

[0240] In some embodiments, if the rank of the first CSI of time unit K is less than or equal to the rank of time unit KN, or the rank of the first CSI of time unit K is less than or equal to the rank of time unit KL, the first cache information includes the intermediate state information of the CSI corresponding to layer X on time unit KN and / or the intermediate state information of the CSI corresponding to layer X on time unit KL; or,

[0241] If the rank of the first CSI of time unit K is inconsistent with the rank of time unit KN, or the rank of the first CSI of time unit K is inconsistent with the rank of time unit KL, the first cache information includes the intermediate state information of CSI corresponding to layer 0 on time unit KN and / or the intermediate state information of CSI corresponding to layer 0 on time unit KL; or, the first cache information includes the intermediate state information of CSI corresponding to a specific layer on time unit KN and / or the intermediate state information of CSI corresponding to a specific layer on time unit KL.

[0242] Wherein, the first target CSI corresponds to the time unit K and the layer X, or the first CSI corresponds to the time unit K and the layer X;

[0243] Wherein, the specific layer is the lowest layer among all layers in the corresponding time unit, or the specific layer is the highest layer among all layers in the corresponding time unit, and the specific layer is a layer configured on the network side or agreed upon by the protocol;

[0244] Where K and X are both integers greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

[0245] In this embodiment, the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X. If the rank of the first CSI in time unit K is less than or equal to the rank of time unit KN, or the rank of the first CSI in time unit K is less than or equal to the rank of time unit KL, the first cache information includes the intermediate state information of the CSI corresponding to layer X in time unit KN and / or the intermediate state information of the CSI corresponding to layer X in time unit KL. Thus, when the rank of the first CSI in time unit K is less than or equal to the rank of time unit KN, or the rank of the first CSI in time unit K is less than or equal to the rank of time unit KL, the first device can determine the content of the first cache information.

[0246] In this embodiment, the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X. If the rank of the first CSI in time unit K is inconsistent with the rank of time unit KN, or the rank of the first CSI in time unit K is inconsistent with the rank of time unit KL, the first cache information includes the intermediate state information of the CSI corresponding to layer 0 in time unit KN and / or the intermediate state information of the CSI corresponding to layer 0 in time unit KL. Alternatively, the first cache information includes the intermediate state information of the CSI corresponding to a specific layer in time unit KN and / or the intermediate state information of the CSI corresponding to a specific layer in time unit KL. Thus, when the rank of the first CSI in time unit K is inconsistent with the rank of time unit KN, or the rank of the first CSI in time unit K is inconsistent with the rank of time unit KL, the first device can determine the content of the first cache information.

[0247] In some embodiments, if the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by the first S layers on time unit K is the CSI intermediate state information corresponding to layer X of the first S layers on time unit KN and / or the CSI intermediate state information corresponding to layer X of the first S layers on time unit KL; and / or,

[0248] If the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, then the CSI intermediate state information used by the last W layers of time unit K is the CSI intermediate state information corresponding to at least one of the first W layers of time unit KN, or the CSI intermediate state information used by the last W layers of time unit K is the CSI intermediate state information corresponding to at least one of the first W layers of time unit KL; or...

[0249] If the rank of time unit K is inconsistent with the rank of time unit KN, or if the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by the last W layers of time unit K is empty; or,

[0250] If the rank of time unit K is inconsistent with the rank of time unit KN, or if the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by each layer of time unit K shall be configured by the network layer or agreed upon by the protocol.

[0251] Wherein, the first target CSI corresponds to the time unit K and the layer X, or the first CSI corresponds to the time unit K and the layer X;

[0252] Where X and K are both integers greater than or equal to 0, L, N, S, and W are all positive integers, and L is an integer multiple of N.

[0253] For example, if the rank of time unit K is 2 and the rank of time unit KN is 4, then layer 0 of time unit K uses the CSI intermediate state information corresponding to layer 0 of time unit K as the first cache information, and / or layer 1 of time unit K uses the CSI intermediate state information corresponding to layer 1 of time unit KN as the first cache information.

[0254] For example, if the rank of time unit K is 2 and the rank of time unit KN is 2, then layer 0 of time unit K uses the CSI intermediate state information corresponding to layer 0 of time unit K as the first cache information, and / or layer 1 of time unit K uses the CSI intermediate state information corresponding to layer 1 of time unit K as the first cache information.

[0255] In this embodiment, the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X. If the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the intermediate CSI state information used by the first S layers on time unit K is the intermediate CSI state information corresponding to layer X of the first S layers on time unit KN and / or the intermediate CSI state information corresponding to layer X of the first S layers on time unit KL. Thus, when the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the first device can determine the content of the first cache information.

[0256] In this embodiment, the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X. If the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the intermediate CSI state information used by the last W layers of time unit K is the intermediate CSI state information corresponding to at least one of the first W layers of time unit KN, or the intermediate CSI state information used by the last W layers of time unit K is the intermediate CSI state information corresponding to at least one of the first W layers of time unit KL. Thus, when the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the first device can determine the content of the first cache information.

[0257] For example, if the rank of time unit K is 4 and the rank of time unit KN is 2, then layer 0 of time unit K uses the CSI intermediate state information corresponding to layer 0 of time unit KN as the first cache information, layer 1 of time unit K uses the CSI intermediate state information corresponding to layer 1 of time unit KN as the first cache information, layer 2 of time unit K uses the CSI intermediate state information corresponding to layer 0 or layer 1 of time unit KN as the first cache information, and layer 3 of time unit K uses the CSI intermediate state information corresponding to layer 0 or layer 1 of time unit KN as the first cache information.

[0258] For example, if the rank of time unit K is 4 and the rank of time unit KN is 2, then layer 0 / layer 1 / layer 2 / layer 3 of time unit K uses the CSI intermediate state information corresponding to layer 0 of time unit KN as the first cache information; or, layer 0 / layer 1 / layer 2 / layer 3 of time unit K uses the CSI intermediate state information corresponding to layer 1 of time unit KN as the first cache information; or, layer 0 / layer 1 / layer 2 / layer 3 of time unit K uses the mixed CSI intermediate state information of layer 0 and layer 1 of time unit KN as the first cache information.

[0259] In this embodiment, the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X. If the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the intermediate state information of CSI used by the last W layers of time unit K is empty. Therefore, in the case that the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the first device can determine the content of the first cache information.

[0260] In this embodiment, the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X. If the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the intermediate state information of CSI used by each layer of time unit K is configured by the network layer or agreed by the protocol. Therefore, when the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the first device can determine the content of the first cache information.

[0261] In some embodiments, if the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cached information includes intermediate CSI state information corresponding to layer Y on the time unit K; wherein K, X, and Y are all integers greater than or equal to 0. Optionally, layer Y is layer 0 on the time unit K, or layer Y is the lowest layer among all layers on the time unit K. This embodiment can solve the problem of how to obtain the intermediate CSI state information of layer X when layer X does not exist in the time unit before time unit K.

[0262] In this embodiment, if the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, and the first cache information includes the intermediate state information of the CSI corresponding to layer Y on time unit K, then the first cache information can be determined when the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X. That is, this embodiment can help find the corresponding buffer when the rank is different at different times. For example, if the rank of time unit K is 2 (meaning time unit K includes the intermediate state information of CSI for layer 0 and the intermediate state information of CSI for layer 1), and the rank of time unit KN is 1 (meaning there is only one layer, layer 0), then the buffer does not contain the buffer for layer 1 of time unit KN, which will lead to a mismatch problem.

[0263] For example, taking four layers (layer 0, layer 1, layer 2 and layer 3) and two time units as an example, the intermediate state information transmission process provided in this application embodiment can be as shown in Figure 6. The intermediate state information generated by layer 0 in time unit T is transmitted to all layers in time unit T+1.

[0264] For example, taking four layers (layer 0, layer 1, layer 2, and layer 3) and three time units as an example, the intermediate state information transmission process provided in this application embodiment can be as shown in Figure 7. The intermediate state information generated by layer 0 above time unit T is transmitted to layer 1 above time unit T; the intermediate state information generated by layer 1 above time unit T is transmitted to layer 2 above time unit T; the intermediate state information generated by layer 2 above time unit T is transmitted to layer 3 above time unit T; the intermediate state information generated by layer 3 above time unit T is transmitted to layer 0 above time unit T+1; and the intermediate state information generated by layer 0 above time unit T+1 is transmitted to the time unit... Intermediate state information generated by upper layer 1 of time unit T+1 is passed to upper layer 2 of time unit T+1. Intermediate state information generated by upper layer 2 of time unit T+1 is passed to upper layer 3 of time unit T+1. Intermediate state information generated by upper layer 3 of time unit T+1 is passed to upper layer 0 of time unit T+2. Intermediate state information generated by upper layer 0 of time unit T+2 is passed to upper layer 1 of time unit T+2. Intermediate state information generated by upper layer 1 of time unit T+2 is passed to upper layer 2 of time unit T+2. Intermediate state information generated by upper layer 2 of time unit T+2 is passed to upper layer 3 of time unit T+2.

[0265] For example, taking four layers (layer 0, layer 1, layer 2, and layer 3) and three time units as an example, the intermediate state information transmission process provided in this application embodiment can be as shown in Figure 8. The intermediate state information generated by layer 0 above time unit T is transmitted to layer 0 above time unit T+1. The intermediate state information generated by layer 0 above time unit T+1 is transmitted to layer 0 above time unit T+2. The intermediate state information generated by layer 1 above time unit T is transmitted to layer 1 above time unit T+1. The intermediate state information generated by layer 1 above time unit T+1 is transmitted to layer 1 above time unit T+2. The intermediate state information generated by layer 2 above time unit T is transmitted to layer 2 above time unit T+1. The intermediate state information generated by layer 2 above time unit T is transmitted to layer 2 above time unit T+1. The intermediate state information generated by layer 2 above time unit T+1 is transmitted to layer 2 above time unit T+2. The intermediate state information generated by layer 3 above time unit T is transmitted to layer 3 above time unit T+1. The intermediate state information generated by layer 3 above time unit T+1 is transmitted to layer 3 above time unit T+2.

[0266] For example, taking four layers (layer 0, layer 1, layer 2, and layer 3) and three time units as an example, the intermediate state information transmission process provided in this application embodiment can be as shown in Figure 9. The intermediate state information generated by layer 3 above time unit T is transmitted to layer 2 above time unit T. The intermediate state information generated by layer 2 above time unit T is transmitted to layer 1 above time unit T. The intermediate state information generated by layer 1 above time unit T is transmitted to layer 0 above time unit T. The intermediate state information generated by layer 0 above time unit T is transmitted to layer 3 above time unit T+1. The intermediate state information generated by layer 3 above time unit T+1 is transmitted to time unit T+2. Intermediate state information generated by upper layer 2 of time unit T+1 is passed to upper layer 1 of time unit T+1. Intermediate state information generated by upper layer 1 of time unit T+1 is passed to upper layer 0 of time unit T+1. Intermediate state information generated by upper layer 0 of time unit T+1 is passed to upper layer 3 of time unit T+2. Intermediate state information generated by upper layer 3 of time unit T+2 is passed to upper layer 2 of time unit T+2. Intermediate state information generated by upper layer 2 of time unit T+2 is passed to upper layer 1 of time unit T+2. Intermediate state information generated by upper layer 1 of time unit T+2 is passed to upper layer 0 of time unit T+2.

[0267] For example, taking four layers (layer 0, layer 1, layer 2, and layer 3) and three time units as an example, the intermediate state information transmission process provided in this application embodiment can be as shown in Figure 10. The intermediate state information generated by layer 0 above time unit T is transmitted to layer 1 above time unit T. The intermediate state information generated by layer 1 above time unit T is transmitted to layer 0 above time unit T+1. The intermediate state information generated by layer 0 above time unit T+1 is transmitted to layer 1 above time unit T+1. The intermediate state information generated by layer 1 above time unit T+1 is transmitted to layer 0 above time unit T+2. The intermediate state information generated by layer 0 above time unit T+2 is transmitted to layer 1 above time unit T+2. The intermediate state information generated by layer 2 above time unit T is transmitted to layer 2 above time unit T+1. The intermediate state information generated by layer 2 above time unit T+1 is transmitted to layer 2 above time unit T+2. The intermediate state information generated by layer 3 above time unit T is transmitted to layer 3 above time unit T+1. The intermediate state information generated by layer 3 above time unit T+1 is transmitted to layer 3 above time unit T+2.

[0268] For example, taking four layers (layer 0, layer 1, layer 2, and layer 3) and three time units as an example, the intermediate state information transmission process provided in this application embodiment can be as shown in Figure 11: the intermediate state information generated by layer 0 above time unit T is transmitted to layer 1 above time unit T; the intermediate state information generated by layer 1 above time unit T is transmitted to layer 2 above time unit T; the intermediate state information generated by layer 2 above time unit T is transmitted to layer 0 above time unit T+1; the intermediate state information generated by layer 0 above time unit T+1 is transmitted to layer 1 above time unit T+1; and the intermediate state information generated by layer 1 above time unit T+1 is transmitted to layer 0 above time unit T+1. Intermediate state information generated by upper layer 2 of time unit T+1 is passed to upper layer 0 of time unit T+2. Intermediate state information generated by upper layer 0 of time unit T+2 is passed to upper layer 1 of time unit T+2. Intermediate state information generated by upper layer 1 of time unit T+2 is passed to upper layer 2 of time unit T+2. Intermediate state information generated by upper layer 2 of time unit T+2 is passed to upper layer 3 of time unit T+2. Intermediate state information generated by upper layer 3 of time unit T+1 is passed to upper layer 3 of time unit T+1. Intermediate state information generated by upper layer 3 of time unit T+1 is passed to upper layer 3 of time unit T+2.

[0269] For example, taking four layers (layer 0, layer 1, layer 2, and layer 3) and three time units as an example, the intermediate state information transmission process provided in this application embodiment can be as shown in Figure 12. The intermediate state information generated by layer 0 above time unit T is transmitted to layer 0 above time unit T+1. The intermediate state information generated by layer 0 above time unit T+1 is transmitted to layer 0 above time unit T+2. The intermediate state information generated by layer 1 above time unit T is transmitted to layer 1 above time unit T+1. The intermediate state information generated by layer 1 above time unit T+1 is transmitted to layer 1 above time unit T+2. The intermediate state information generated by layer 2 above time unit T is transmitted to layer 3 above time unit T. The intermediate state information generated by layer 3 above time unit T is transmitted to layer 2 above time unit T+1. The intermediate state information generated by layer 2 above time unit T+1 is transmitted to layer 3 above time unit T+1. The intermediate state information generated by layer 3 above time unit T is transmitted to layer 2 above time unit T+1. The intermediate state information generated by layer 2 above time unit T+1 is transmitted to layer 3 above time unit T+1. The intermediate state information generated by layer 3 above time unit T+1 is transmitted to layer 2 above time unit T+2. The intermediate state information generated by layer 2 above time unit T+2 is transmitted to layer 3 above time unit T+2.

[0270] For example, taking four layers (layer 0, layer 1, layer 2, and layer 3) and three time units as an example, the intermediate state information transmission process provided in this application embodiment can be as shown in Figure 13. The intermediate state information generated by layer 0 of time unit T is transmitted to layer 0 of time unit T+1 and layer 1 of time unit T; the intermediate state information generated by layer 0 of time unit T+1 is transmitted to layer 0 of time unit T+2 and layer 1 of time unit T+1; the intermediate state information generated by layer 0 of time unit T+2 is transmitted to layer 1 of time unit T+2; the intermediate state information generated by layer 1 of time unit T is transmitted to layer 1 of time unit T+1 and layer 2 of time unit T; the intermediate state information generated by layer 1 of time unit T+1... The intermediate state information generated by the upper layer 1 of time unit T+2 is passed to the upper layer 2 of time unit T+1. The intermediate state information generated by the upper layer 2 of time unit T is passed to the upper layer 2 of time unit T+2 and the upper layer 3 of time unit T+1. The intermediate state information generated by the upper layer 2 of time unit T+1 is passed to the upper layer 2 of time unit T+2 and the upper layer 3 of time unit T+1. The intermediate state information generated by the upper layer 2 of time unit T+2 is passed to the upper layer 3 of time unit T+2. The intermediate state information generated by the upper layer 3 of time unit T is passed to the upper layer 3 of time unit T+1. The intermediate state information generated by the upper layer 3 of time unit T+1 is passed to the upper layer 3 of time unit T+2.

[0271] For example, taking four layers (layer 0, layer 1, layer 2, and layer 3) and three time units as an example, the intermediate state information transmission process provided in this application embodiment can be as shown in Figure 14. The intermediate state information generated by layer 0 of time unit T is transmitted to layer 0 of time unit T+1 and layer 1 of time unit T; the intermediate state information generated by layer 0 of time unit T+1 is transmitted to layer 0 of time unit T+2 and layer 1 of time unit T+1; the intermediate state information generated by layer 0 of time unit T+2 is transmitted to layer 1 of time unit T+2; and the intermediate state information generated by layer 1 of time unit T... The information is transmitted to the upper layer 1 of time unit T+1 and the upper layer 2 of time unit T. The intermediate state information generated by the upper layer 1 of time unit T+1 is transmitted to the upper layer 1 of time unit T+2 and the upper layer 2 of time unit T+1. The intermediate state information generated by the upper layer 1 of time unit T+2 is transmitted to the upper layer 2 of time unit T+2. The intermediate state information generated by the upper layer 2 of time unit T is transmitted to the upper layer 3 of time unit T. The intermediate state information generated by the upper layer 2 of time unit T+1 is transmitted to the upper layer 3 of time unit T+1. The intermediate state information generated by the upper layer 2 of time unit T+2 is transmitted to the upper layer 3 of time unit T+2.

[0272] In some embodiments, if the first target CSI corresponds to time unit K, or the first CSI corresponds to time unit K, the first cached information includes CSI information at time unit K-1*P, CSI information at time unit K-2*P, ..., CSI information at time unit KQ*P; wherein K is an integer greater than or equal to 0, and P and Q are both positive integers. In this embodiment, the first cached information includes CSI information at time unit K-1*P, CSI information at time unit K-2*P, ..., CSI information at time unit KQ*P, in order to predict CSI information at one or more future time units.

[0273] In some embodiments, if the first target CSI corresponds to time unit K, or the first CSI corresponds to time unit K, the first cached information includes the first CSI information at time unit K-1*P, the first CSI information at time unit K-2*P, ..., the first CSI information at time unit KQ*P; wherein K is an integer greater than or equal to 0, and P and Q are both positive integers. In this embodiment, the first cached information includes the first CSI information at time unit K-1*P, the first CSI information at time unit K-2*P, ..., the first CSI information at time unit KQ*P, in order to predict CSI information at one or more future time units.

[0274] In some embodiments, the first cached information includes at least one of the following

[0275] First CSI information obtained from historical time;

[0276] The first CSI information obtained at the current time;

[0277] The historical time includes at least one of the following: time unit KN, time unit KL;

[0278] Wherein, the current time is time unit K;

[0279] Wherein, the first target CSI corresponds to the time unit K, or the first CSI corresponds to the time unit K;

[0280] Where K is an integer greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

[0281] In some embodiments, the first cached information includes first CSI information acquired at multiple historical times.

[0282] For example, the first CSI information acquired from multiple historical times includes association layer information, rank information, time information, first timer, observation window configuration information, first information, and at least one of the target AI unit.

[0283] Optionally, the plurality of historical time intervals are associated with the configuration information of the observation window, such as being consistent with the period or interval of the observation window.

[0284] Optionally, at least one of the plurality of historical times is included in the time information, such as the most recent historical time being the time information.

[0285] Optionally, the first CSI information obtained from the multiple historical times may be associated with the same layer information.

[0286] Optionally, the first CSI information acquired at the multiple historical times may be associated with different layer information.

[0287] Optionally, the first CSI information acquired at the multiple historical times may be associated with different rank information.

[0288] Optionally, the first CSI information acquired at multiple historical times may be associated with the same rank information.

[0289] In some embodiments, if the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cached information includes the CSI information of layer X at time unit K-1*P, the CSI information of layer X at time unit K-2*P, ..., the CSI information of layer X at time unit KQ*P; wherein X and K are both integers greater than or equal to 0, and P and Q are both positive integers. In this embodiment, the first cached information includes the CSI information of layer X at time unit K-1*P, the CSI information of layer X at time unit K-2*P, ..., the CSI information of layer X at time unit KQ*P, so as to predict the CSI information of layer X at one or more future time units.

[0290] In some embodiments, the first cached information includes at least one of the following

[0291] First CSI information of layer X obtained from historical time;

[0292] The first CSI information of layer X obtained at the current time;

[0293] The historical time includes at least one of the following: time unit KN, time unit KL;

[0294] Wherein, the current time is time unit K;

[0295] Wherein, the first target CSI corresponds to the time unit K, or the first CSI corresponds to the time unit K;

[0296] Where K is an integer greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

[0297] In some embodiments, the first cache information includes first CSI information of layer X obtained at multiple historical times.

[0298] For example, the first CSI information acquired from multiple historical times includes association layer information, rank information, time information, first timer, observation window configuration information, first information, and at least one of the target AI unit.

[0299] Optionally, the plurality of historical time intervals are associated with the configuration information of the observation window, such as being consistent with the period or interval of the observation window.

[0300] Optionally, at least one of the plurality of historical times is included in the time information, such as the most recent historical time being the time information.

[0301] Optionally, the information of the first CSI information obtained from the plurality of historical times is associated with the information of layer X.

[0302] In some embodiments, the first cached information includes first CSI information of layer X obtained at multiple historical times and first CSI information of layer X obtained at the current time.

[0303] In some embodiments, if the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cached information includes the CSI information of layer Y on time unit K-1*P, the CSI information of layer Y on time unit K-2*P, ..., the CSI information of layer Y on time unit KQ*P; wherein X, Y, and K are all integers greater than or equal to 0, and P and Q are both positive integers. In this embodiment, the first cached information includes the CSI information of layer Y on time unit K-1*P, the CSI information of layer Y on time unit K-2*P, ..., the CSI information of layer Y on time unit KQ*P, so as to predict the CSI information of layer Y on one or more future time units.

[0304] In some embodiments, if the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cached information includes CSI information at time unit K-1*P, CSI information at time unit K-2*P, ..., CSI information at time unit KQ*P; wherein X and K are integers greater than or equal to 0, and P and Q are positive integers. In this embodiment, the first cached information includes CSI information at time unit K-1*P, CSI information at time unit K-2*P, ..., CSI information at time unit KQ*P, in order to predict CSI information at one or more future time units.

[0305] In some embodiments, the first cached information includes CSI information at least two time units, wherein the CSI information at least two time units is associated with the same layer, or the first cached information includes layer information corresponding to the CSI information at least two time units.

[0306] In some embodiments, if a first time unit exists among the at least two time units and the CSI information on the first time unit is not cached, the first cached information includes only the CSI information on the time units preceding the first time unit among the at least two time units, or the first cached information includes only the CSI information on the time units other than the first time unit among the at least two time units.

[0307] For example, if CSI information is not cached on time unit K-2*P, among time units K-1*P, K-2*P, ..., KQ*P, the first cached information only includes the CSI information on time unit K-1*P, or the first cached information only includes the CSI information on time unit K-1*P, the CSI information on time unit K-3*P, ..., KQ*P.

[0308] In some embodiments, if a first time unit exists among the at least two time units, and the CSI information in the first time unit is not cached, the wireless communication method 200 further includes:

[0309] The first device sends second information to the second device, wherein the second information is used to indicate that no CSI information is cached on the first time unit. That is, in this embodiment, the first device sends default information of the first cached information to the second device.

[0310] Alternatively, the first device sends a second message to the second device, wherein the second message instructs the first cached information to restart caching. It can be understood that the cached information prior to K-2*P is cleared.

[0311] Optionally, the first time unit may include one or more time units.

[0312] In some embodiments, the first device is an encoding device and the second device is a decoding device; or, the first device is a decoding device and the second device is an encoding device.

[0313] In this embodiment, the first device sends second information to the second device. This second information indicates that no CSI information is cached on the first time unit. Thus, the second device can be aware that no CSI information is cached on the first time unit, ensuring consistency in understanding of the cached information between the first and second devices. In other words, if CSI information is missing on one or more time units adjacent to the time unit K, the prediction result of the target AI unit has performance uncertainty, which needs to be indicated to the second device.

[0314] In some embodiments, if a first time unit exists among the at least two time units, and the CSI information in the first time unit is not cached, the wireless communication method 200 further includes:

[0315] The first device clears or resets the CSI information on the at least two time units, or the first device determines that the first cached information has expired or become invalid.

[0316] In some embodiments, the wireless communication method 200 further includes:

[0317] The first device sends a third message to the second device;

[0318] The third information includes, but is not limited to, at least one of the following:

[0319] The first instruction information is used to instruct the clearing or resetting of the first cache information;

[0320] The second indication information is used to indicate that the first timer has expired;

[0321] The type information of the first cached information;

[0322] The time information associated with the first cached information;

[0323] The number of instances in the first cache information.

[0324] For example, if the number of instances in the first cache information does not meet the number of instances indicated by the network side, the number of instances in the first cache information is reported.

[0325] In this embodiment, the second device can learn from the first indication information that the first device has cleared or reset the first cache information, or the second device can learn from the first indication information that the first timer has expired, or the second device can determine the type information of the first cache information based on the third information, or the second device can determine the time information associated with the first cache information based on the third information, or the second device can determine the number of instances in the first cache information based on the third information. Thus, the second device can learn from the relevant information of the first cache information so that the first device and the second device have a consistent understanding of the first cache information.

[0326] In some embodiments, before the first device inputs the first CSI and the first cached information into the first AI model, the wireless communication method 200 further includes:

[0327] The first device sends a fourth message to the second device;

[0328] The fourth piece of information includes, but is not limited to, at least one of the following:

[0329] The maximum number of instances in the cache supported by the first device;

[0330] The cache type information supported by the first device;

[0331] The maximum load value of one instance in the cache supported by the first device;

[0332] Does the first device support dedicated layer caching?

[0333] Does the first device support dedicated rank caching?

[0334] For example, the maximum number of instances in the cache supported by the first device can be understood as the total number of instances in the cache supported by the first device for all target AI units.

[0335] For example, the maximum load value of an instance in the cache supported by the first device can be understood as the maximum number of time units of CSI information that can be included, or the maximum number of time units of CSI intermediate state information that can be included. For example, 4 time units, 6 time units, etc.

[0336] For example, the maximum load value of an instance in the cache supported by the first device can be understood as the maximum number of CSI information entries that can be included, or the maximum number of CSI intermediate state information entries that can be included.

[0337] For example, the maximum load value of an instance in the cache supported by the first device can be understood as the maximum number of layers of CSI information that can be included, or the maximum number of layers of CSI intermediate state information that can be included. For example, 4 layers, 2 layers, 1 layer, etc.

[0338] For example, the maximum load value of an instance in the cache supported by the first device can be understood as the maximum number of instances that a target AI unit or a CSI can report.

[0339] In this embodiment, the first device can report cached capability information to the second device, thereby the second device can determine or configure the parameters of the first cache information based on the cached capability information reported by the first device.

[0340] This application embodiment may involve the following three types of intermediate state information interaction:

[0341] Type 1: The intermediate state information transmission process should be part of the model information or model description information. The intermediate state information is exchanged between the two ends (i.e., the encoder and the decoder). If the protocol defines a reference encoder and / or a reference decoder (i.e., the protocol directly defines the encoder and / or decoder), then the intermediate state information exchanged between the encoder and the decoder needs to be defined.

[0342] Type 2: The intermediate state information transmission process may be part of the implementation and may not be defined or interacted with; or the protocol may define the intermediate state information transmission process, or the two parties may interact to transmit intermediate state information to facilitate joint training, improve the joint training effect, or speed up the iteration speed.

[0343] Type 3 involves the intermediate state information transmission process, which aims to prevent the design of the intermediate state information transmission process between the NW side and the UE side from being too different, which may lead to the inability of separate training to achieve good learning results.

[0344] It should be understood that the model involved in the above intermediate state information interaction types can be the target AI unit described in this application.

[0345] Therefore, in this embodiment, the first device acquires first cache information, or the target AI unit acquires first cache information. The first cache information is used to acquire output information associated with the target AI unit, or the first cache information is used to acquire a first target CSI associated with the target AI unit. Specifically, the first cache information is associated with at least one of the following: layer information, rank information, time information, a first timer, configuration information of the observation window, first information, and the target AI unit. The first cache information is introduced in the process of acquiring the output information associated with the target AI unit or the first target CSI associated with the target AI unit, thereby improving the accuracy of the target AI unit's prediction. Furthermore, the first cache information is associated with time information, the first timer, and the configuration information of the observation window, thereby enabling prediction and / or compression and decompression (such as CSI compression and CSI decompression) in the spatiotemporal frequency domain. Introducing time domain information allows the encoder and decoder to better align the model inference schemes at both ends, and / or can be applied to prediction and / or compression and decompression (such as CSI compression and CSI decompression) based on the target AI unit in high-rank (such as rank ≥ 2) cases, and improves the performance of prediction and / or compression and decompression in the spatiotemporal frequency domain (such as improving the performance of CSI compression and CSI decompression in the spatiotemporal frequency domain).

[0346] The wireless communication method provided in this application can be executed by a wireless communication device. This application uses an example of a wireless communication device executing the wireless communication method to illustrate the wireless communication device provided in this application.

[0347] This application provides a wireless communication device. As an example, the wireless communication device can be a communication equipment or a component within a communication equipment, such as a chip. The communication equipment can be a terminal, a network-side device, an AI unit, or a server, etc. Exemplarily, the terminal can be, but is not limited to, the type of terminal 11 listed above, and the network-side device can be, but is not limited to, the type of network-side device 12 listed above. This application does not impose specific limitations on these aspects.

[0348] The wireless communication device includes a receiving module, a transmitting module, and a processing module. These modules can be implemented in software or hardware. When implemented in hardware, the processing module can be implemented by a processor. For example, the processor can include general-purpose processors, special-purpose processors, such as a Central Processing Unit (CPU), microprocessor, Digital Signal Processor (DSP), Artificial Intelligence (AI) processor, Graphics Processing Unit (GPU), Application Specific Integrated Circuit (ASIC), Network Processor (NP), Field Programmable Gate Array (FPGA), or other programmable logic devices, gate circuits, transistors, discrete hardware components, etc. The receiving and transmitting modules can be implemented by a communication interface, which can include one or more of the following: transceiver, pins, circuits, bus, radio frequency unit, etc.

[0349] Referring to Figure 15, when the wireless communication device is a first device or a component of the first device, or when the wireless communication device is a target AI unit, the wireless communication device 300 includes:

[0350] The processing module 301 is used to obtain first cache information, wherein the first cache information is used to obtain output information associated with the target AI unit, or the first cache information is used to obtain the first target CSI associated with the target AI unit;

[0351] The first cached information includes at least one of the following: CSI information of at least one historical time unit, at least one CSI intermediate state information, and input information associated with the target AI unit in at least one time unit;

[0352] Wherein, the CSI intermediate state information in the at least one CSI intermediate state information is intermediate information obtained based on the target AI unit;

[0353] Wherein, the first cache information is associated with at least one of the following: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and the target AI unit;

[0354] The time information includes at least one of the following: the at least one historical time unit, the time unit corresponding to the at least one CSI intermediate state information, and the at least one time unit;

[0355] In the event that the first timer expires, some or all of the information in the first cached information is cleared or reset;

[0356] The configuration information of the observation window includes at least one of the following: at least one of the time-domain interval, quantity, and start position corresponding to the CSI information of the at least one historical time unit; at least one of the time-domain interval, quantity, and start position corresponding to the at least one CSI intermediate state information; and at least one of the time-domain interval, quantity, and start position corresponding to the at least one time unit.

[0357] Wherein, the first information satisfies at least one of the following: instructing to clear or reset the first cache information, instructing to clear or reset the condition information of the first cache information, configuring the type of the first cache information, configuring the number of instances in the first cache information, instructing the association between the first cache information and the layer information, and instructing the association between the first cache information and the rank information.

[0358] In some embodiments, the layer information includes at least one of the following: a layer identifier, layer type information; and / or,

[0359] The rank information includes at least one of the following: the rank identifier, and the rank type information;

[0360] The types of layers include at least one of the following: shared layer, dedicated layer;

[0361] The types of rank include at least one of the following: shared rank, exclusive rank.

[0362] In some embodiments, the processing module 301 is further configured to input the first cached information and the first CSI into the target AI unit to obtain the first target CSI.

[0363] In some embodiments, the first cache information includes CSI intermediate state information of time unit KN; and / or,

[0364] The first cached information includes the CSI intermediate state information of time unit KL;

[0365] Wherein, the first target CSI corresponds to time unit K, or the first CSI corresponds to time unit K, and the first CSI is used to determine the first target CSI;

[0366] Where K is an integer greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

[0367] In some embodiments, if the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cache information includes the intermediate state information of the CSI corresponding to layer X.

[0368] Wherein, the first CSI is used to determine the first target CSI, and X is an integer greater than or equal to 0.

[0369] In some embodiments, the first cache information includes CSI intermediate state information corresponding to at least two time units, and the CSI intermediate state information corresponding to the at least two time units is associated with the same layer.

[0370] In some embodiments, if the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cache information includes the intermediate state information of the CSI corresponding to layer Y.

[0371] Wherein, the first CSI is used to determine the first target CSI, and X and Y are both integers greater than or equal to 0.

[0372] In some embodiments, the first cached information includes CSI intermediate state information corresponding to at least two time units, wherein the CSI intermediate state information corresponding to the at least two time units is associated with different layers.

[0373] In some embodiments, the CSI intermediate state information corresponding to layer Y is at least one of the following:

[0374] The CSI intermediate state information corresponding to layer Y obtained from historical time;

[0375] The CSI intermediate state information corresponding to layer Y obtained at the current time;

[0376] Wherein, layer Y is layer 0, or layer Y is the highest layer, or layer Y is a layer configured by the network side or specified by the protocol;

[0377] The historical time includes at least one of the following: time unit KN, time unit KL;

[0378] Wherein, the current time is time unit K;

[0379] Wherein, the first target CSI corresponds to the time unit K, or the first CSI corresponds to the time unit K, and the first CSI is used to determine the first target CSI;

[0380] Where K is an integer greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

[0381] In some embodiments, if the rank of the first CSI of time unit K is less than or equal to the rank of time unit KN, or the rank of the first CSI of time unit K is less than or equal to the rank of time unit KL, the first cache information includes the intermediate state information of the CSI corresponding to layer X on time unit KN and / or the intermediate state information of the CSI corresponding to layer X on time unit KL; or,

[0382] If the rank of the first CSI of time unit K is inconsistent with the rank of time unit KN, or the rank of the first CSI of time unit K is inconsistent with the rank of time unit KL, the first cache information includes the intermediate state information of CSI corresponding to layer 0 on time unit KN and / or the intermediate state information of CSI corresponding to layer 0 on time unit KL; or, the first cache information includes the intermediate state information of CSI corresponding to a specific layer on time unit KN and / or the intermediate state information of CSI corresponding to a specific layer on time unit KL.

[0383] Wherein, the first target CSI corresponds to the time unit K and the layer X, or the first CSI corresponds to the time unit K and the layer X, and the first CSI is used to determine the first target CSI;

[0384] Wherein, the specific layer is the lowest layer among all layers in the corresponding time unit, or the specific layer is the highest layer among all layers in the corresponding time unit, and the specific layer is a layer configured on the network side or agreed upon by the protocol;

[0385] Where K and X are both integers greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

[0386] In some embodiments, if the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by the first S layers on time unit K is the CSI intermediate state information corresponding to layer X of the first S layers on time unit KN and / or the CSI intermediate state information corresponding to layer X of the first S layers on time unit KL; and / or,

[0387] If the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, then the CSI intermediate state information used by the last W layers of time unit K is the CSI intermediate state information corresponding to at least one of the first W layers of time unit KN, or the CSI intermediate state information used by the last W layers of time unit K is the CSI intermediate state information corresponding to at least one of the first W layers of time unit KL; or...

[0388] If the rank of time unit K is inconsistent with the rank of time unit KN, or if the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by the last W layers of time unit K is empty; or,

[0389] If the rank of time unit K is inconsistent with the rank of time unit KN, or if the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by each layer of time unit K shall be configured by the network layer or agreed upon by the protocol.

[0390] Wherein, the first target CSI corresponds to the time unit K and the layer X, or the first CSI corresponds to the time unit K and the layer X, and the first CSI is used to determine the first target CSI;

[0391] Where X and K are both integers greater than or equal to 0, L, N, S, and W are all positive integers, and L is an integer multiple of N.

[0392] In some embodiments, if the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cache information includes the intermediate state information of the CSI corresponding to layer Y on the time unit K;

[0393] Wherein, the first CSI is used to determine the first target CSI, and K, X and Y are all integers greater than or equal to 0.

[0394] In some embodiments, layer Y is layer 0 on the time unit K, or layer Y is the lowest layer among all layers on the time unit K.

[0395] In some embodiments, if the first target CSI corresponds to time unit K, or the first CSI corresponds to time unit K, the first cached information includes CSI information on time unit K-1*P, CSI information on time unit K-2*P, ..., CSI information on time unit KQ*P;

[0396] Wherein, the first CSI is used to determine the first target CSI, K is an integer greater than or equal to 0, and P and Q are both positive integers.

[0397] In some embodiments, if the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cached information includes the CSI information of layer X at time unit K-1*P, the CSI information of layer X at time unit K-2*P, ..., the CSI information of layer X at time unit KQ*P; or,

[0398] If the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cached information includes the CSI information of layer Y at time unit K-1*P, the CSI information of layer Y at time unit K-2*P, ..., the CSI information of layer Y at time unit KQ*P; or,

[0399] If the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cached information includes CSI information on time unit K-1*P, CSI information on time unit K-2*P, ..., CSI information on time unit KQ*P;

[0400] Wherein, the first CSI is used to determine the first target CSI, X, Y, and K are all integers greater than or equal to 0, and P and Q are both positive integers.

[0401] In some embodiments, the first cached information includes CSI information at least two time units, wherein the CSI information at least two time units is associated with the same layer, or the first cached information includes layer information corresponding to the CSI information at least two time units.

[0402] In some embodiments, if a first time unit exists among the at least two time units and the CSI information on the first time unit is not cached, the first cached information includes only the CSI information on the time units preceding the first time unit among the at least two time units, or the first cached information includes only the CSI information on the time units other than the first time unit among the at least two time units.

[0403] In some embodiments, if a first time unit exists among the at least two time units, and the CSI information in the first time unit is not cached, the wireless communication device 300 further includes:

[0404] The sending module 302 is used to send second information to the second device, wherein the second information is used to indicate that CSI information is not cached on the first time unit.

[0405] In some embodiments, the wireless communication device 300 further includes:

[0406] The sending module 302 is used to send third information to the second device;

[0407] The third information includes at least one of the following:

[0408] The first instruction information is used to instruct the clearing or resetting of the first cache information;

[0409] The second indication information is used to indicate that the first timer has expired;

[0410] The type information of the first cached information;

[0411] The time information associated with the first cached information;

[0412] The number of instances in the first cache information.

[0413] In some embodiments, before the wireless communication device 300 acquires the first cached information, the wireless communication device 300 further includes:

[0414] The sending module 302 is used to send fourth information to the second device;

[0415] The fourth piece of information includes at least one of the following:

[0416] The maximum number of instances in the cache supported by the wireless communication device 300;

[0417] The wireless communication device 300 supports cache type information;

[0418] The maximum load value of an instance in the cache supported by the wireless communication device 300;

[0419] Does the wireless communication device 300 support dedicated layer caching?

[0420] Does the wireless communication device 300 support dedicated rank caching?

[0421] In some embodiments, the wireless communication device 300 stores at most R copies of the first cached information in a time unit;

[0422] Wherein, the R first cache information pieces are associated with one target AI unit, or the R first cache information pieces are associated with different target AI units, and R is a positive integer.

[0423] In some embodiments, the target AI unit obtains the first cached information from at least one of the following:

[0424] The registers of the wireless communication device 300, the memory of the wireless communication device 300, and the high-level units of the wireless communication device 300.

[0425] In some embodiments, the target AI unit is deployed in the wireless communication device 300.

[0426] In some embodiments, the target AI unit includes at least one of the following:

[0427] The terminal is used to acquire the first AI unit of the first target CSI;

[0428] Network-side devices are used to acquire the second AI unit for reconstructing CSI;

[0429] The terminal is used to obtain the reference AI unit of the first target CSI;

[0430] Network-side devices are used to obtain reference AI units for reconstructing CSI;

[0431] The third AI unit used in testing by the terminal, network-side equipment, or testing equipment;

[0432] The terminal, network-side device, or test equipment is used to match the fifth AI unit of the fourth AI unit used in the test.

[0433] In some embodiments, the wireless communication device 300 includes at least one of the following: a terminal, a network-side device, and a test device.

[0434] Therefore, in this embodiment, the first device acquires first cache information, or the target AI unit acquires first cache information. The first cache information is used to acquire output information associated with the target AI unit, or the first cache information is used to acquire a first target CSI associated with the target AI unit. Specifically, the first cache information is associated with at least one of the following: layer information, rank information, time information, a first timer, configuration information of the observation window, first information, and the target AI unit. The first cache information is introduced in the process of acquiring the output information associated with the target AI unit or the first target CSI associated with the target AI unit, thereby improving the accuracy of the target AI unit's prediction. Furthermore, the first cache information is associated with time information, the first timer, and the configuration information of the observation window, thereby enabling prediction and / or compression and decompression (such as CSI compression and CSI decompression) in the spatiotemporal frequency domain. Introducing time domain information allows the encoder and decoder to better align the model inference schemes at both ends, and / or can be applied to prediction and / or compression and decompression (such as CSI compression and CSI decompression) based on the target AI unit in high-rank (such as rank ≥ 2) cases, and improves the performance of prediction and / or compression and decompression in the spatiotemporal frequency domain (such as improving the performance of CSI compression and CSI decompression in the spatiotemporal frequency domain).

[0435] The wireless communication device provided in this application embodiment can implement the various processes implemented in the method embodiments of Figures 2 to 14 and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0436] As shown in Figure 16, this application embodiment also provides a communication device 400, including a processor 401 and a memory 402. The memory 402 stores a program or instructions that can be executed on the processor 401. For example, when the communication device 400 is a first device, when the program or instructions are executed by the processor 401, they implement the various steps of the above-described wireless communication method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0437] This application also provides a terminal, including a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the steps in the method embodiment shown in FIG2. This terminal embodiment corresponds to the first device-side method embodiment described above, and all implementation processes and methods of the above method embodiments can be applied to this terminal embodiment and can achieve the same technical effect. The terminal may be the wireless communication device 300 shown in FIG15. Specifically, FIG17 is a schematic diagram of the hardware structure of a terminal implementing an embodiment of this application.

[0438] The terminal 500 includes, but is not limited to, at least some of the following components: radio frequency unit 501, network module 502, audio output unit 503, input unit 504, sensor 505, display unit 506, user input unit 507, interface unit 508, memory 509, and processor 510.

[0439] Those skilled in the art will understand that terminal 500 may also include a power supply (such as a battery) for powering various components. The power supply can be logically connected to processor 510 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The terminal structure shown in Figure 17 does not constitute a limitation on the terminal. The terminal may include more or fewer components than shown, or combine certain components, or have different component arrangements, which will not be elaborated here.

[0440] It should be understood that, in this embodiment, the input unit 504 may include a graphics processor 5041 and a microphone 5042. The graphics processor 5041 processes image data of still images or videos obtained by an image capture device (such as a camera) in video capture mode or image capture mode. The display unit 506 may include a display panel 5061, which may be configured in the form of a liquid crystal display, an organic light-emitting diode, or the like. The user input unit 507 includes at least one of a touch panel 5071 and other input devices 5072. The touch panel 5071 is also called a touch screen. The touch panel 5071 may include two parts: a touch detection device and a touch controller. Other input devices 5072 may include, but are not limited to, physical keyboards, function keys (such as volume control buttons, power buttons, etc.), trackballs, mice, and joysticks, which will not be described in detail here.

[0441] In this embodiment, after receiving downlink data from the network-side device, the radio frequency unit 501 can transmit it to the processor 510 for processing; in addition, the radio frequency unit 501 can send uplink data to the network-side device. Typically, the radio frequency unit 501 includes, but is not limited to, antennas, amplifiers, transceivers, couplers, low-noise amplifiers, duplexers, etc.

[0442] The memory 509 can be used to store software programs or instructions, as well as various data. The memory 509 may primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area may store the operating system, application programs or instructions required for at least one function (such as sound playback, image playback, etc.). Furthermore, the memory 509 may include volatile memory or non-volatile memory. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DRRAM). The memory 509 in this embodiment includes, but is not limited to, these and any other suitable types of memory.

[0443] Processor 510 may include one or more processing units; optionally, processor 510 integrates an application processor and a modem processor, wherein the application processor mainly handles operations involving the operating system, user interface, and applications, and the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into processor 510.

[0444] The processor 510 is used to obtain first cache information, which is used to obtain output information associated with the target AI unit, or the first cache information is used to obtain a first target CSI associated with the target AI unit.

[0445] The first cached information includes at least one of the following: CSI information of at least one historical time unit, at least one CSI intermediate state information, and input information associated with the target AI unit in at least one time unit;

[0446] Wherein, the CSI intermediate state information in the at least one CSI intermediate state information is intermediate information obtained based on the target AI unit;

[0447] Wherein, the first cache information is associated with at least one of the following: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and the target AI unit;

[0448] The time information includes at least one of the following: the at least one historical time unit, the time unit corresponding to the at least one CSI intermediate state information, and the at least one time unit;

[0449] In the event that the first timer expires, some or all of the information in the first cached information is cleared or reset;

[0450] The configuration information of the observation window includes at least one of the following: at least one of the time-domain interval, quantity, and start position corresponding to the CSI information of the at least one historical time unit; at least one of the time-domain interval, quantity, and start position corresponding to the at least one CSI intermediate state information; and at least one of the time-domain interval, quantity, and start position corresponding to the at least one time unit.

[0451] Wherein, the first information satisfies at least one of the following: instructing to clear or reset the first cache information, instructing to clear or reset the condition information of the first cache information, configuring the type of the first cache information, configuring the number of instances in the first cache information, instructing the association between the first cache information and the layer information, and instructing the association between the first cache information and the rank information.

[0452] Therefore, in this embodiment, the terminal obtains first cache information, which is used to obtain output information associated with the target AI unit, or to obtain a first target CSI associated with the target AI unit. Specifically, the first cache information is associated with at least one of the following: layer information, rank information, time information, a first timer, configuration information of the observation window, first information, and the target AI unit. The first cache information is introduced in the process of obtaining the output information associated with the target AI unit or the first target CSI associated with the target AI unit, thereby improving the accuracy of the target AI unit's prediction. Furthermore, the first cache information is associated with time information, the first timer, and the configuration information of the observation window, thereby enabling prediction and / or compression and decompression (such as CSI compression and CSI decompression) in the spatiotemporal frequency domain. Introducing time domain information allows the encoder and decoder to better align the model inference schemes at both ends, and / or can be applied to prediction and / or compression and decompression (such as CSI compression and CSI decompression) based on the target AI unit in high-rank (such as rank ≥ 2) cases, and improves the performance of prediction and / or compression and decompression in the spatiotemporal frequency domain (such as improving the performance of CSI compression and CSI decompression in the spatiotemporal frequency domain).

[0453] It is understood that the implementation process of each implementation method mentioned in this embodiment can refer to the relevant description of method embodiment XXX and achieve the same or corresponding technical effects. To avoid repetition, it will not be described again here.

[0454] This application also provides a network-side device, including a processor and a communication interface. The communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the steps of the method embodiment shown in FIG2. This network-side device embodiment corresponds to the method embodiment executed by the first device described above. All implementation processes and methods of the above method embodiments can be applied to this network-side device embodiment and can achieve the same technical effect.

[0455] Specifically, this application embodiment also provides a network-side device, which may be the wireless communication device 300 shown in FIG15. As shown in FIG18, the network-side device 600 includes: an antenna 61, a radio frequency device 62, a baseband device 63, a processor 64, and a memory 65. The antenna 61 is connected to the radio frequency device 62. In the uplink direction, the radio frequency device 62 receives information through the antenna 61 and sends the received information to the baseband device 63 for processing. In the downlink direction, the baseband device 63 processes the information to be transmitted and sends it to the radio frequency device 62. The radio frequency device 62 processes the received information and transmits it through the antenna 61.

[0456] The method executed by the first device in the above embodiments can be implemented in the baseband device 63, which includes a baseband processor.

[0457] The baseband device 63 may include at least one baseband board, on which multiple chips are disposed, as shown in FIG18. One of the chips is, for example, a baseband processor, which is connected to the memory 65 via a bus interface to call the program in the memory 65 and execute the operations performed by the first device shown in the above method embodiment.

[0458] The network-side device may also include a network interface 66, such as a Common Public Radio Interface (CPRI).

[0459] Specifically, the network-side device 600 in this application embodiment further includes: instructions or programs stored in memory 65 and executable on processor 64. Processor 64 calls the instructions or programs in memory 65 to execute the methods executed by each module shown in FIG15 and achieve the same technical effect. To avoid repetition, it will not be described in detail here.

[0460] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described wireless communication method embodiments and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0461] The processor is the processor in the first device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk. In some examples, the readable storage medium may be a non-transient readable storage medium.

[0462] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described wireless communication method embodiments and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0463] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0464] This application also provides a computer program / program product, which is stored in a storage medium and executed by at least one processor to implement the various processes of the above-described wireless communication method embodiments, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0465] This application also provides a communication system, including a first device and a second device, wherein the first device can be used to perform the steps of the wireless communication method described above.

[0466] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0467] From the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of computer software products plus necessary general-purpose hardware platforms, and of course, they can also be implemented by hardware. The computer software product is stored in a storage medium (such as ROM, RAM, magnetic disk, optical disk, etc.) and includes several instructions to cause the terminal or network-side device to execute the methods described in the various embodiments of this application.

[0468] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other implementations under the guidance of this application without departing from the spirit and scope of the claims. All of these implementations are within the protection scope of this application.

Claims

1. A wireless communication method, comprising: The first device obtains the first cache information, or the target artificial intelligence (AI) unit obtains the first cache information. The first cache information is used to obtain the output information associated with the target AI unit, or the first cache information is used to obtain the first target channel state information (CSI) associated with the target AI unit. The first cached information includes at least one of the following: CSI information of at least one historical time unit, at least one CSI intermediate state information, and input information associated with the target AI unit in at least one time unit; Wherein, the CSI intermediate state information in the at least one CSI intermediate state information is intermediate information obtained based on the target AI unit; Wherein, the first cache information is associated with at least one of the following: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and the target AI unit; The time information includes at least one of the following: the at least one historical time unit, the time unit corresponding to the at least one CSI intermediate state information, and the at least one time unit; In the event that the first timer expires, some or all of the information in the first cached information is cleared or reset; The configuration information of the observation window includes at least one of the following: at least one of the time-domain interval, quantity, and start position corresponding to the CSI information of the at least one historical time unit; at least one of the time-domain interval, quantity, and start position corresponding to the at least one CSI intermediate state information; and at least one of the time-domain interval, quantity, and start position corresponding to the at least one time unit. Wherein, the first information satisfies at least one of the following: instructing to clear or reset the first cache information, instructing to clear or reset the condition information of the first cache information, configuring the type of the first cache information, configuring the number of instances in the first cache information, instructing the association between the first cache information and the layer information, and instructing the association between the first cache information and the rank information.

2. The method according to claim 1, wherein, The layer information includes at least one of the following: layer identifier, layer type information; and / or, The rank information includes at least one of the following: the rank identifier, and the rank type information; The types of layers include at least one of the following: shared layer, dedicated layer; The types of rank include at least one of the following: shared rank, exclusive rank.

3. The method according to claim 1 or 2, wherein, The method further includes: The first device inputs the first cached information and the first CSI into the target AI unit to obtain the first target CSI.

4. The method according to any one of claims 1 to 3, wherein, The first cached information includes the CSI intermediate state information of time unit KN; and / or, The first cached information includes the CSI intermediate state information of time unit KL; Wherein, the first target CSI corresponds to time unit K, or the first CSI corresponds to time unit K, and the first CSI is used to determine the first target CSI; Where K is an integer greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

5. The method according to any one of claims 1 to 4, wherein, If the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cache information includes the intermediate state information of the CSI corresponding to layer X; Wherein, the first CSI is used to determine the first target CSI, and X is an integer greater than or equal to 0.

6. The method according to claim 5, wherein, The first cached information includes CSI intermediate state information corresponding to at least two time units, and the CSI intermediate state information corresponding to the at least two time units is associated with the same layer.

7. The method according to any one of claims 1 to 4, wherein, If the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cache information includes the intermediate state information of the CSI corresponding to layer Y; Wherein, the first CSI is used to determine the first target CSI, and X and Y are both integers greater than or equal to 0.

8. The method according to claim 7, wherein, The first cached information includes CSI intermediate state information corresponding to at least two time units, and the CSI intermediate state information corresponding to the at least two time units is associated with different layers.

9. The method according to claim 7 or 8, wherein, The CSI intermediate state information corresponding to layer Y is at least one of the following: The CSI intermediate state information corresponding to layer Y obtained from historical time; The CSI intermediate state information corresponding to layer Y obtained at the current time; Wherein, layer Y is layer 0, or layer Y is the highest layer, or layer Y is a layer configured by the network side or specified by the protocol; The historical time includes at least one of the following: time unit KN, time unit KL; Wherein, the current time is time unit K; Wherein, the first target CSI corresponds to the time unit K, or the first CSI corresponds to the time unit K, and the first CSI is used to determine the first target CSI; Where K is an integer greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

10. The method according to any one of claims 1 to 4, wherein, If the rank of the first CSI of time unit K is less than or equal to the rank of time unit KN, or the rank of the first CSI of time unit K is less than or equal to the rank of time unit KL, the first cache information includes the intermediate state information of CSI corresponding to layer X on time unit KN and / or the intermediate state information of CSI corresponding to layer X on time unit KL. or, If the rank of the first CSI of time unit K is inconsistent with the rank of time unit KN, or the rank of the first CSI of time unit K is inconsistent with the rank of time unit KL, the first cache information includes the intermediate state information of CSI corresponding to layer 0 on time unit KN and / or the intermediate state information of CSI corresponding to layer 0 on time unit KL; or, the first cache information includes the intermediate state information of CSI corresponding to a specific layer on time unit KN and / or the intermediate state information of CSI corresponding to a specific layer on time unit KL. Wherein, the first target CSI corresponds to the time unit K and the layer X, or the first CSI corresponds to the time unit K and the layer X, and the first CSI is used to determine the first target CSI; Wherein, the specific layer is the lowest layer among all layers in the corresponding time unit, or the specific layer is the highest layer among all layers in the corresponding time unit, and the specific layer is a layer configured on the network side or agreed upon by the protocol; Where K and X are both integers greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

11. The method according to any one of claims 1 to 4, wherein, If the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by the first S layers on time unit K is the CSI intermediate state information corresponding to layer X of the first S layers on time unit KN and / or the CSI intermediate state information corresponding to layer X of the first S layers on time unit KL; and / or, If the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by the last W layers of time unit K is the CSI intermediate state information corresponding to at least one of the first W layers of time unit KN, or the CSI intermediate state information used by the last W layers of time unit K is the CSI intermediate state information corresponding to at least one of the first W layers of time unit KL. or, If the rank of time unit K is inconsistent with the rank of time unit KN, or if the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by the last W layers of time unit K is empty; or, If the rank of time unit K is inconsistent with the rank of time unit KN, or if the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by each layer of time unit K shall be configured by the network layer or agreed upon by the protocol. Wherein, the first target CSI corresponds to the time unit K and the layer X, or the first CSI corresponds to the time unit K and the layer X, and the first CSI is used to determine the first target CSI; Where X and K are both integers greater than or equal to 0, L, N, S, and W are all positive integers, and L is an integer multiple of N.

12. The method according to any one of claims 1 to 3, wherein, If the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cache information includes the intermediate state information of the CSI corresponding to layer Y on the time unit K; Wherein, the first CSI is used to determine the first target CSI, and K, X and Y are all integers greater than or equal to 0.

13. The method according to claim 12, wherein, The layer Y is layer 0 on the time unit K, or the layer Y is the lowest layer among all layers on the time unit K.

14. The method according to any one of claims 1 to 3, wherein, If the first target CSI corresponds to time unit K, or the first CSI corresponds to time unit K, the first cached information includes CSI information on time unit K-1*P, CSI information on time unit K-2*P, ..., CSI information on time unit KQ*P; Wherein, the first CSI is used to determine the first target CSI, K is an integer greater than or equal to 0, and P and Q are both positive integers.

15. The method according to any one of claims 1 to 3, 14, wherein, If the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cached information includes the CSI information of layer X on time unit K-1*P, the CSI information of layer X on time unit K-2*P, ..., the CSI information of layer X on time unit KQ*P; or, If the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cache information includes the CSI information of layer Y on time unit K-1*P, the CSI information of layer Y on time unit K-2*P, ..., the CSI information of layer Y on time unit KQ*P; or, If the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cached information includes CSI information on time unit K-1*P, CSI information on time unit K-2*P, ..., CSI information on time unit KQ*P; Wherein, the first CSI is used to determine the first target CSI, X, Y, and K are all integers greater than or equal to 0, and P and Q are both positive integers.

16. The method according to claim 15, wherein, The first cached information includes CSI information at least two time units, wherein the CSI information at least two time units is associated with the same layer, or the first cached information includes layer information corresponding to the CSI information at least two time units.

17. The method according to claim 16, wherein, If a first time unit exists among the at least two time units, and the CSI information on the first time unit is not cached, the first cached information only includes the CSI information on the time units preceding the first time unit among the at least two time units, or the first cached information only includes the CSI information on the time units other than the first time unit among the at least two time units.

18. The method according to claim 16, wherein, If a first time unit exists among the at least two time units, and the CSI information in the first time unit is not cached, the method further includes: The first device sends a second message to the second device, wherein the second message is used to indicate that no CSI information is cached on the first time unit.

19. The method according to any one of claims 1 to 18, wherein, The method further includes: The first device sends a third message to the second device; The third information includes at least one of the following: The first instruction information is used to instruct the clearing or resetting of the first cache information; The second indication information is used to indicate that the first timer has expired; The type information of the first cached information; The time information associated with the first cached information; The number of instances in the first cache information.

20. The method according to any one of claims 1 to 19, wherein, Before the first device acquires the first cached information, or before the target AI unit acquires the first cached information, the method further includes: The first device sends a fourth message to the second device; The fourth piece of information includes at least one of the following: The maximum number of instances in the cache supported by the first device; The cache type information supported by the first device; The maximum load value of one instance in the cache supported by the first device; Does the first device support dedicated layer caching? Does the first device support dedicated rank caching? 21. The method according to any one of claims 1 to 20, wherein, The first device stores at most R copies of the first cache information in a single time unit; Wherein, the R first cache information pieces are associated with one target AI unit, or the R first cache information pieces are associated with different target AI units, and R is a positive integer.

22. The method according to any one of claims 1 to 21, wherein, The target AI unit obtains the first cached information, including: The target AI unit obtains the first cached information from at least one of the following: The registers of the first device, the memory of the first device, and the high-level units of the first device.

23. The method according to any one of claims 1 to 22, wherein, The target AI unit is deployed in the first device.

24. The method according to any one of claims 1 to 23, wherein, The target AI unit includes at least one of the following: The terminal is used to acquire the first AI unit of the first target CSI; Network-side devices are used to acquire the second AI unit for reconstructing CSI; The terminal is used to obtain the reference AI unit of the first target CSI; Network-side devices are used to obtain reference AI units for reconstructing CSI; The third AI unit used in testing by the terminal, network-side equipment, or testing equipment; The terminal, network-side device, or test equipment is used to match the fifth AI unit of the fourth AI unit used in the test; And / or, The first device includes at least one of the following: a terminal, a network-side device, and a test device.

25. A wireless communication device, comprising: The processing module is used to obtain first cache information, which is used to obtain output information associated with the target artificial intelligence (AI) unit, or the first cache information is used to obtain first target channel state information (CSI) associated with the target AI unit. The first cached information includes at least one of the following: CSI information of at least one historical time unit, at least one CSI intermediate state information, and input information associated with the target AI unit in at least one time unit; Wherein, the CSI intermediate state information in the at least one CSI intermediate state information is intermediate information obtained based on the target AI unit; Wherein, the first cache information is associated with at least one of the following: layer information, rank information, time information, first timer, configuration information of the observation window, first information, and the target AI unit; The time information includes at least one of the following: the at least one historical time unit, the time unit corresponding to the at least one CSI intermediate state information, and the at least one time unit; In the event that the first timer expires, some or all of the information in the first cached information is cleared or reset; The configuration information of the observation window includes at least one of the following: at least one of the time-domain interval, quantity, and start position corresponding to the CSI information of the at least one historical time unit; at least one of the time-domain interval, quantity, and start position corresponding to the at least one CSI intermediate state information; and at least one of the time-domain interval, quantity, and start position corresponding to the at least one time unit. Wherein, the first information satisfies at least one of the following: instructing to clear or reset the first cache information, instructing to clear or reset the condition information of the first cache information, configuring the type of the first cache information, configuring the number of instances in the first cache information, instructing the association between the first cache information and the layer information, and instructing the association between the first cache information and the rank information.

26. The apparatus according to claim 25, wherein, The first cached information includes the CSI intermediate state information of time unit KN; and / or; The first cached information includes the CSI intermediate state information of time unit KL; Wherein, the first target CSI corresponds to time unit K, or the first CSI corresponds to time unit K, and the first CSI is used to determine the first target CSI; Where K is an integer greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

27. The apparatus according to claim 25 or 26, wherein, If the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cache information includes the intermediate state information of the CSI corresponding to layer X; Wherein, the first CSI is used to determine the first target CSI, and X is an integer greater than or equal to 0.

28. The apparatus according to claim 25 or 26, wherein, If the first target CSI corresponds to the CSI information of layer X, or the first CSI corresponds to the CSI information of layer X, the first cache information includes the intermediate state information of the CSI corresponding to layer Y; Wherein, the first CSI is used to determine the first target CSI, and X and Y are both integers greater than or equal to 0.

29. The apparatus according to claim 25 or 26, wherein, If the rank of the first CSI of time unit K is less than or equal to the rank of time unit KN, or the rank of the first CSI of time unit K is less than or equal to the rank of time unit KL, the first cache information includes the intermediate state information of CSI corresponding to layer X on time unit KN and / or the intermediate state information of CSI corresponding to layer X on time unit KL. or, If the rank of the first CSI of time unit K is inconsistent with the rank of time unit KN, or the rank of the first CSI of time unit K is inconsistent with the rank of time unit KL, the first cache information includes the intermediate state information of CSI corresponding to layer 0 on time unit KN and / or the intermediate state information of CSI corresponding to layer 0 on time unit KL; or, the first cache information includes the intermediate state information of CSI corresponding to a specific layer on time unit KN and / or the intermediate state information of CSI corresponding to a specific layer on time unit KL. Wherein, the first target CSI corresponds to the time unit K and the layer X, or the first CSI corresponds to the time unit K and the layer X, and the first CSI is used to determine the first target CSI; Wherein, the specific layer is the lowest layer among all layers in the corresponding time unit, or the specific layer is the highest layer among all layers in the corresponding time unit, and the specific layer is a layer configured on the network side or agreed upon by the protocol; Where K and X are both integers greater than or equal to 0, L and N are both positive integers, and L is an integer multiple of N.

30. The apparatus according to claim 25 or 26, wherein, If the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by the first S layers on time unit K is the CSI intermediate state information corresponding to layer X of the first S layers on time unit KN and / or the CSI intermediate state information corresponding to layer X of the first S layers on time unit KL; and / or, If the rank of time unit K is inconsistent with the rank of time unit KN, or the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by the last W layers of time unit K is the CSI intermediate state information corresponding to at least one of the first W layers of time unit KN, or the CSI intermediate state information used by the last W layers of time unit K is the CSI intermediate state information corresponding to at least one of the first W layers of time unit KL. or, If the rank of time unit K is inconsistent with the rank of time unit KN, or if the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by the last W layers of time unit K is empty; or, If the rank of time unit K is inconsistent with the rank of time unit KN, or if the rank of time unit K is inconsistent with the rank of time unit KL, the CSI intermediate state information used by each layer of time unit K shall be configured by the network layer or agreed upon by the protocol. Wherein, the first target CSI corresponds to the time unit K and the layer X, or the first CSI corresponds to the time unit K and the layer X, and the first CSI is used to determine the first target CSI; Where X and K are both integers greater than or equal to 0, L, N, S, and W are all positive integers, and L is an integer multiple of N.

31. The apparatus according to claim 25, wherein, If the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cache information includes the intermediate state information of the CSI corresponding to layer Y on the time unit K; Wherein, the first CSI is used to determine the first target CSI, and K, X and Y are all integers greater than or equal to 0.

32. The apparatus according to claim 31, wherein, The layer Y is layer 0 on the time unit K, or the layer Y is the lowest layer among all layers on the time unit K.

33. The apparatus according to claim 25, wherein, If the first target CSI corresponds to time unit K, or the first CSI corresponds to time unit K, the first cached information includes CSI information on time unit K-1*P, CSI information on time unit K-2*P, ..., CSI information on time unit KQ*P; Wherein, the first CSI is used to determine the first target CSI, K is an integer greater than or equal to 0, and P and Q are both positive integers.

34. The apparatus according to claim 25 or 33, wherein, If the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cached information includes the CSI information of layer X on time unit K-1*P, the CSI information of layer X on time unit K-2*P, ..., the CSI information of layer X on time unit KQ*P; or, If the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cache information includes the CSI information of layer Y on time unit K-1*P, the CSI information of layer Y on time unit K-2*P, ..., the CSI information of layer Y on time unit KQ*P; or, If the first target CSI corresponds to time unit K and layer X, or the first CSI corresponds to time unit K and layer X, the first cached information includes CSI information on time unit K-1*P, CSI information on time unit K-2*P, ..., CSI information on time unit KQ*P; Wherein, the first CSI is used to determine the first target CSI, X, Y, and K are all integers greater than or equal to 0, and P and Q are both positive integers.

35. The apparatus according to claim 34, wherein, The first cached information includes CSI information at least two time units, wherein the CSI information at least two time units is associated with the same layer, or the first cached information includes layer information corresponding to the CSI information at least two time units.

36. The apparatus according to claim 35, wherein, If a first time unit exists among the at least two time units, and the CSI information on the first time unit is not cached, the first cached information only includes the CSI information on the time units preceding the first time unit among the at least two time units, or the first cached information only includes the CSI information on the time units other than the first time unit among the at least two time units.

37. The apparatus according to claim 35, wherein, If a first time unit exists among the at least two time units, and the CSI information in the first time unit is not cached, the wireless communication device further includes: The sending module is used to send second information to the second device, wherein the second information is used to indicate that no CSI information is cached on the first time unit.

38. The apparatus according to any one of claims 25 to 37, wherein, The wireless communication device further includes: The sending module is used to send third information to the second device; The third information includes at least one of the following: The first instruction information is used to instruct the clearing or resetting of the first cache information; The second indication information is used to indicate that the first timer has expired; The type information of the first cached information; The time information associated with the first cached information; The number of instances in the first cache information.

39. The apparatus according to any one of claims 25 to 38, wherein, Before the wireless communication device acquires the first cached information, the wireless communication device further includes: The sending module is used to send the fourth information to the second device; The fourth piece of information includes at least one of the following: The maximum number of instances in the cache supported by the wireless communication device; The wireless communication device supports cache type information; The maximum load value of an instance in the cache supported by the wireless communication device; Does the wireless communication device support dedicated layer caching? Does the wireless communication device support dedicated rank caching? 40. A first device comprising a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the wireless communication method as claimed in any one of claims 1 to 24.

41. A readable storage medium storing a program or instructions that, when executed by a processor, implement the wireless communication method as described in any one of claims 1 to 24.

Images