Communication method, apparatus, computer readable storage medium and program product

Abstract

This application provides a communication method, apparatus, computer-readable storage medium, and computer program product. In this method, the communication apparatus receives overlay data through a first resource, processes the overlay data through a third AI model based on first indication information or a second resource corresponding to a first artificial intelligence (AI) model, to obtain first data corresponding to the first AI model. The second resource is used to carry the output data of the first AI model. Based on second indication information or a third resource corresponding to a second AI model, the overlay data is processed through the third AI model to obtain second data corresponding to the second AI model. The third resource is used to carry the output data of the second AI model. Thus, in this application's embodiments, the efficiency of multiple access can be improved, and the receiving end does not need to deploy multiple decoders. It has good scalability in large-scale scenarios and facilitates cross-vendor collaboration.

Description

Technical Field

[0001] This application relates to the field of communications, and more specifically to a communication method, apparatus, computer-readable storage medium, and computer program product. Background Technology

[0002] Multiple access technology refers to techniques that allow multiple nodes (such as multiple terminals or multiple network devices) to access the same network or system for data transmission or interaction, enabling multiple nodes to simultaneously obtain communication services without signal interference between them. Common multiple access technologies include Frequency Division Multiple Access (FDMA), Time Division Multiple Access (TDMA), Code Division Multiple Access (CDMA), Orthogonal Frequency Division Multiple Access (OFDMA), and Non-Orthogonal Multiple Access (NOMA). However, there is still room for further optimization and improvement in multiple access technologies in the field of communication technology. Summary of the Invention

[0003] The embodiments of this application provide a communication method, apparatus, computer-readable storage medium, and computer program product that can improve the efficiency of multiple access, and the receiving end does not need to deploy multiple decoders. It has good scalability in large-scale scenarios and facilitates cross-vendor collaboration.

[0004] Firstly, embodiments of this application provide a communication method that can be applied to a data transmission receiving end, such as a network device or terminal, a communication module / processing module in the network device or terminal, circuits or chips responsible for communication functions in the network device or terminal (such as a modem chip, also known as a baseband chip, or a system-on-chip (SoC) chip containing a modem core, or a system-in-package (SIP) chip), and circuits or chips responsible for processing functions in the network device or terminal (such as a graphics processing unit (GPU), artificial intelligence (AI) processor, neural network processing unit (NPU), or application-specific integrated circuit (ASIC)). This method can also be implemented by a logical node, logical module, or software capable of implementing all or part of the receiving end's functions. For example, this method is applied to a network device. In this method, the network device receives superimposed data through a first resource. Furthermore, based on the first instruction information or the second resource corresponding to the first artificial intelligence (AI) model, the network device processes the overlay data through the third AI model to obtain the first data corresponding to the first AI model. The second resource is used to carry the output data of the first AI model. The network device also processes the overlay data through the third AI model based on the second instruction information or the third resource corresponding to the second AI model to obtain the second data corresponding to the second AI model. The third resource is used to carry the output data of the second AI model. Using this method, the receiving end can improve the efficiency of multiple access, and the receiving end does not need to deploy multiple decoders. It has good scalability in large-scale scenarios and facilitates cross-vendor collaboration.

[0005] In one possible design, the network device receives first indication information, which indicates the identifier, function, or one or more elements of the training dataset of the first AI model. In this way, the third AI model can separate the transmitting data corresponding to the first AI model from the overlay data, enabling multiple access for users and improving data transmission spectral efficiency and user access capabilities.

[0006] In one possible design, the network device receives second indication information, which indicates the identifier, function, or one or more elements of the training dataset of the second AI model. In this way, the third AI model can separate the transmitting data corresponding to the second AI model from the overlay data, enabling multiple access for users and improving data transmission spectral efficiency and user access capabilities.

[0007] In one possible design, during the process of the network device processing the superimposed data based on the second indication information or the third resource using a third AI model to obtain the second data, the network device can also obtain the second data based on the first data and the second indication information or the third resource, by processing the superimposed data using the third AI model. This approach can improve the data separation capability of the receiving end, further enhancing data transmission spectrum efficiency and user access capabilities.

[0008] In one possible design, the overlay data includes one or more tokens. This reduces communication overhead and improves the robustness of communication transmission.

[0009] In one possible design, the first data corresponds to the first node, and the second data corresponds to the second node. Alternatively, the first data and the second data correspond to the first node. This allows the receiver to improve transmission rates in different scenarios. By obtaining the superimposed data from multiple transmitting nodes in multiple access scenarios, multiple access for users can be achieved, improving data transmission spectrum efficiency and user access capabilities. Obtaining the superimposed data from the same transmitting node can increase data throughput, thereby improving the data transmission rate.

[0010] Secondly, embodiments of this application provide a communication method that can be applied to a data transmission transmitter, such as a network device or terminal, a communication module / processing module in the network device or terminal, a circuit or chip responsible for communication functions in the network device or terminal (such as a modem chip, also known as a baseband chip, or a system-on-a-chip containing a modem core or a system-in-package (SIP) chip), or a circuit or chip responsible for processing functions in the network device or terminal (such as a graphics processor, AI processor, neural network processor, or application-specific integrated circuit). This method can also be implemented by a logical node, logical module, or software capable of implementing all or part of the receiver's functions. For example, the transmitter may include multiple nodes. Taking the application of this method to a terminal as an example, the number of terminals can be multiple, and a terminal can be an example of a node. In this method, a first terminal (an example of a first node) at the transmitter processes the data of the first terminal through a first AI model to obtain third data, and then sends the third data through a first resource. A second terminal (an example of a second node) at the transmitter processes the data of the second terminal through a second AI model to obtain fourth data, and then sends the fourth data through the first resource. Since the third and fourth data are transmitted by different terminals using the same resources, they will overlap during transmission, forming superimposed data. The third data, along with the first indication information or second resource corresponding to the first AI model, serves as input data for the third AI model, enabling the recovery of the first data corresponding to the first AI model at the receiving end. Similarly, the fourth data, along with the second indication information or third resource corresponding to the second AI model, serves as input data for the third AI model, enabling the recovery of the second data corresponding to the second AI model at the receiving end. This method improves the efficiency of multiple access, simplifies receiving end components (e.g., eliminates the need to deploy multiple decoders), offers good scalability in large-scale scenarios, and facilitates cross-vendor collaboration.

[0011] Thirdly, this application provides a communication device that has the functions of the first aspect described above. For example, the communication device includes modules, units, or means that perform the operations involved in the first aspect. These modules, units, or means can be implemented by software, hardware, or a combination of software and hardware.

[0012] Fourthly, this application provides a communication device that has the functions of the second aspect above. For example, the communication device includes a module, unit, or means for performing the operations involved in the second aspect above. The module, unit, or means can be implemented by software, hardware, or a combination of software and hardware.

[0013] Fifthly, this application provides a communication device including an interface circuit and one or more processors. The one or more processors are coupled to a memory. The memory stores part or all of the necessary computer program or instructions for implementing the functions described in the first aspect. The one or more processors can execute the computer program or instructions, causing the communication device to implement the methods in any possible design or implementation of the first aspect. The interface circuit is used to implement the communication functions within the communication device and / or the communication functions between the communication device and other devices or components.

[0014] In one possible design, the processor is used to communicate with other devices or components through the interface circuit.

[0015] In one possible design, the communication device may also include the memory.

[0016] The aforementioned communication device may be a receiving end for data transmission, such as a network device or terminal, or a module (e.g., a circuit, chip, or chip system) in a network device or terminal, or a circuit or chip (e.g., a GPU, AI processor, NPU, or ASIC) in a network device or terminal responsible for communication and / or computing functions, or a logical node or logical module that can implement all or part of the functions of a network device or terminal.

[0017] Sixthly, this application provides a communication device including an interface circuit and one or more processors. The one or more processors are coupled to a memory. The memory stores part or all of the necessary computer program or instructions for implementing the functions described in the second aspect above. The one or more processors are executable to carry out the computer program or instructions, causing the communication device to implement the methods in any possible design or implementation of the second aspect above. The interface circuit is used to implement the communication functions within the communication device and / or the communication functions between the communication device and other devices or components.

[0018] In one possible design, the processor is used to communicate with other devices or components through the interface circuit.

[0019] In one possible design, the communication device may also include the memory.

[0020] The aforementioned communication device may be a terminal, or a communication and / or computing module in a terminal, or a chip in a terminal responsible for communication functions such as a modem chip (also known as a baseband chip) or a SoC or SIP chip containing a modem module, or a circuit or chip in a terminal responsible for communication and / or computing functions (such as a GPU, AI processor, NPU, or ASIC), or a logical node or logical module capable of implementing all or part of the terminal functions.

[0021] In a seventh aspect, this application provides a computer-readable storage medium storing computer-readable instructions that, when read and executed by a computer, cause the computer to perform any of the possible designs in the first to second aspects described above.

[0022] Eighthly, this application provides a computer program product that, when read and executed by a computer, causes the computer to perform any of the possible designs in the first to second aspects described above.

[0023] The beneficial effects of the third to eighth aspects mentioned above can be found in the descriptions of the first and second aspects. Attached Figure Description

[0024] Figure 1A A schematic diagram of a communication system according to some embodiments of this application is shown.

[0025] Figure 1B and Figure 1C A schematic diagram illustrating an example application framework of a communication system according to some embodiments of this application is shown.

[0026] Figures 2A-2B A schematic diagram of the communication flow of some embodiments of this application is shown.

[0027] Figure 3 , Figure 6 , Figure 8 , Figure 16 A schematic diagram of a multiple access process according to some embodiments of this application is shown.

[0028] Figure 4 , Figure 13 , Figure 17 A schematic diagram illustrating the time-frequency resource usage of multiple access in some embodiments of this application is shown.

[0029] Figure 5A – Figure 5B , Figures 9-12 , Figures 14-15 A schematic diagram of signal detection at the receiver end of some embodiments of this application is shown.

[0030] Figure 7 The diagram illustrates the signaling interaction of multiple access according to some embodiments of this application.

[0031] Figure 18 This is a schematic diagram of the structure of an apparatus according to some embodiments of this application.

[0032] Figure 19 This is a schematic diagram of the terminal structure according to this application. Detailed Implementation

[0033] Embodiments of this application will now be described in more detail with reference to the accompanying drawings. While some embodiments of this application are shown in the drawings, it should be understood that embodiments of this application can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this application. It should be understood that the accompanying drawings and embodiments are for illustrative purposes only and are not intended to limit the scope of protection of this application.

[0034] In the description of embodiments of this application, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0035] The embodiments of this application can be implemented according to any suitable communication protocol, including but not limited to cellular communication protocols such as fourth-generation (4G), fifth-generation (5G), and future communication protocols, and wireless local area network communication protocols such as IEEE 802.11. The technical solutions provided in this application can also be applied to device-to-device (D2D) communication, vehicle-to-everything (V2X) communication, machine-to-machine (M2M) communication, machine-type communication (MTC), and Internet of Things (IoT) communication systems or other communication systems.

[0036] Figure 1A A schematic diagram of a communication system according to some embodiments of this application is shown. Figure 1A This is a schematic diagram illustrating one possible, non-limiting system. For example... Figure 1A As shown, the communication system 10 includes a radio access network (RAN) 100 and a core network (CN) 200. RAN 100 includes at least one RAN node (e.g., ...). Figure 1A 110a and 110b (collectively referred to as 110) and at least one terminal (such as Figure 1ARAN 100, denoted as RAN 120a-120j, is collectively referred to as RAN 120. RAN 100 may also include other RAN nodes, such as wireless relay equipment and / or wireless backhaul equipment. Figure 1A (Not shown in the image). Terminal 120 is connected to RAN node 110 wirelessly. RAN node 110 is connected to core network 200 wirelessly or via wired connection. The core network elements in core network 200 and RAN node 110 in RAN 100 can be different physical devices, or they can be the same physical device integrating core network logical functions and radio access network logical functions.

[0037] RAN 100 can be a cellular system related to the 3rd Generation Partnership Project (3GPP), such as 4G, 5G mobile communication systems, or future-oriented evolution systems. RAN 100 can also be an open access network, cloud radio access network (CRAN), virtualized radio access network (vRAN), artificial intelligence radio access network (AI RAN), or wireless fidelity (WiFi) system. RAN 100 can also be a communication system that integrates two or more of the above systems.

[0038] RAN node 110, sometimes also referred to as access network equipment, RAN entity, or access node, constitutes part of the communication system and is used to help terminals achieve wireless access. Multiple RAN nodes 110 in communication system 10 can be of the same type or different types. In some scenarios, the roles of RAN node 110 and terminal 120 are relative, for example... Figure 1A Network element 120i can be a helicopter or a drone, and it can be configured as a mobile base station. For terminals 120j that access RAN 100 through network element 120i, network element 120i is a base station; however, for base station 110a, network element 120i is a terminal. RAN node 110 and terminal 120 are sometimes referred to as communication devices, for example... Figure 1A Network elements 110a and 110b can be understood as communication devices with base station functions, while network elements 120a-120j can be understood as communication devices with terminal functions.

[0039] In one possible scenario, a RAN node can be a base station, an evolved NodeB (eNodeB, eNB), an access point (AP), a transmission reception point (TRP), a next-generation NodeB (gNB), a base station in a future mobile communication system, or an access node in a WiFi system, etc. A RAN node can also be a macro base station (such as...) Figure 1A 110a), micro base stations (also known as small stations) or indoor stations (such as...) Figure 1A The RAN node can be a relay station, relay node, or donor node (as described in section 110b), or a wireless controller in a CRAN scenario. Optionally, the RAN node can also be a server, wearable device, vehicle, or in-vehicle equipment. For example, the access network equipment in vehicle-to-everything (V2X) technology can be a roadside unit (RSU). All or part of the functions of the RAN node in this application can also be implemented through software functions running on hardware, or through virtualization functions instantiated on a platform (e.g., a cloud platform). The RAN node can also be equipped with communication modules, circuits, or chips that perform corresponding communication functions. The RAN node can also be configured with program instructions for performing corresponding communication functions and corresponding program instructions. The RAN node in this application can also be a logical node, logical module, or software that can implement all or part of the access node functions, or a circuit or chip (such as a GPU, AI processor, NPU, or ASIC) responsible for communication and / or computing functions in the access node.

[0040] In another possible scenario, multiple RAN nodes collaborate to assist terminals in achieving wireless access, with different RAN nodes each implementing some of the base station's functions. For example, RAN nodes can be central units (CUs), distributed units (DUs), CU-control plane (CPs), CU-user plane (UPs), or radio units (RUs), etc. CUs and DUs can be set up separately or included in the same network element, such as a baseband unit (BBU). RUs can be included in radio frequency equipment or radio frequency units, such as remote radio units (RRUs), active antenna units (AAUs), or remote radio heads (RRHs). Furthermore, RAN nodes can also be computing units, providing computing power for tasks (such as model inference and / or model training), and can also be used to implement one or more of the following: task partitioning, scheduling, and orchestration. The functionality of a computing unit can be implemented by a separate module independent of other units (e.g., CU, DU, RU), or by one or more other units (e.g., one or more of CU, DU, RU).

[0041] In different systems, CU (or CU-CP and CU-UP), DU, computing unit, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an ORAN system, CU can also be called O-CU (open CU), DU can also be called O-DU, CU-CP can also be called O-CU-CP, CU-UP can also be called O-CU-UP, and RU can also be called O-RU. For ease of description, this application uses CU, CU-CP, CU-UP, DU, computing unit, and RU as examples. Any of the units among CU (or CU-CP, CU-UP), DU, computing unit, and RU in this application can be implemented through software modules, hardware modules, or a combination of software modules and hardware modules.

[0042] A terminal can be a device or module that accesses the aforementioned communication system and has corresponding communication functions. A terminal can also be called a terminal device, user equipment (UE), mobile station, mobile terminal, etc. Terminals can be widely used in various scenarios, such as device-to-device (D2D), V2X communication, MTC, IoT, virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, smart transportation, smart cities, etc. A terminal can be a mobile phone, tablet computer, computer with wireless transceiver capabilities, wearable device, vehicle, drone, helicopter, airplane, ship, robot, robotic arm, smart home device, transportation vehicle with wireless communication capabilities, communication module, etc. The embodiments of this application do not limit the device form of the terminal. A terminal typically contains a communication module, circuit, or chip that performs corresponding communication functions. Furthermore, it can also contain modules, circuits, or chips (such as GPUs, AI processors, NPUs, or ASICs) that perform corresponding communication and / or computing functions. The terminal can also be configured with program instructions for performing corresponding communication and / or computing functions.

[0043] To support artificial intelligence (AI) technology in wireless networks, AI nodes may be introduced into the network. AI nodes can be deployed in one or more of the following locations within the communication system: access network nodes (RAN nodes), terminal devices, or core network elements, or they can be deployed independently, for example, in a location other than any of the above devices, such as a host in an over-the-top (OTT) system or a cloud server. AI nodes can communicate with other devices in the communication system, which may be one or more of the following: network devices, terminal devices, or core network elements. It is understood that this application does not limit the number of AI nodes. For example, when there are multiple AI nodes, they can be divided based on function, such as different AI nodes being responsible for different functions. It is also understood that AI nodes can be independent devices, integrated into the same device to implement different functions, or they can be network elements in hardware devices, software functions running on dedicated hardware, or virtualized functions instantiated on a platform (e.g., a cloud platform). This application does not limit the specific form of the AI ​​nodes. AI nodes can be AI network elements or AI modules.

[0044] Figure 1B A schematic diagram illustrating an example application framework of a communication system according to some embodiments of this application is shown. For example... Figure 1BAs shown, network elements in a communication system are connected via interfaces (e.g., NG, Xn) or air interfaces. These network element nodes, such as core network elements (or core network equipment), access network nodes (RAN nodes), terminals, or one or more devices in operations administration and maintenance (OAM), are equipped with one or more AI modules (for clarity, ...). Figure 1B (Only one is shown in the image). An access network node can be a single RAN node or can include multiple RAN nodes, such as CU and DU. CU and / or DU can also be configured with one or more AI modules. CU can also be split into CU-CP and CU-UP, with one or more AI modules configured in CU-CP and / or CU-UP.

[0045] The AI ​​modules deployed in different network elements can be the same or different. The models of the AI ​​modules can achieve different functions depending on the parameter configurations. The models of the AI ​​modules can be configured based on one or more of the following parameters: structural parameters (e.g., at least one of the following: number of neural network layers, neural network width, inter-layer connections, neuron weights, neuron activation function, or biases in the activation function), input parameters (e.g., the type and / or dimension of the input parameters), or output parameters (e.g., the type and / or dimension of the output parameters). The biases in the activation function can also be referred to as the biases of the neural network.

[0046] In one example, the neural network mentioned above can be a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), or a generative adversarial network (GAN).

[0047] Deep Neural Networks (DNNs) are artificial neural network architectures with multiple layers of nonlinear transformation units stacked in a hierarchical structure to form deep computational models. Compared to shallow neural networks, deep neural networks have more hidden layers, allowing the network model to capture more complex data structures and higher-level abstract features.

[0048] A CNN is a deep neural network with a convolutional structure. A CNN contains a feature extractor consisting of convolutional layers and subsampling layers. This feature extractor can be viewed as a filter, and the convolution process can be seen as performing convolution between a trainable filter and an input image or a convolutional feature map.

[0049] RNN is a type of recursive neural network that takes sequence data as input, recursively moves along the direction of sequence evolution, and connects all nodes (recurrent units) in a chain-like manner.

[0050] GAN is a deep learning model. It consists of a generator and a discriminator, and is trained through adversarial learning. Its purpose is to estimate the potential distribution of data samples and generate new data samples.

[0051] An AI module can have one or more models. A model can infer an output, which includes one or more parameters. The learning, training, or inference processes of different models can be deployed on different nodes or devices, or they can be deployed on the same node or device.

[0052] Figure 1C A schematic diagram illustrating an example application framework of a communication system according to some embodiments of this application is shown. For example... Figure 1C As shown, the communication system includes a RAN intelligent controller (RIC). For example, the RIC can be an AI module used to implement AI-related functions. RICs include near-real-time RICs (near-RT RICs) and non-real-time RICs (non-RT RICs). Non-real-time RICs primarily process non-real-time information, such as data that is not sensitive to latency, with latency in the order of seconds. Real-time RICs primarily process near-real-time information, such as data that is relatively sensitive to latency, with latency in the order of tens of milliseconds.

[0053] Near real-time (NRT) RICs are used for model training and inference. For example, they are used to train AI models (sometimes referred to as AI functions) and then use those models for inference. NRT RICs can obtain network-side and / or terminal-side information from RAN nodes (e.g., CUs, CU-CPs, CU-UPs, DUs, compute nodes, and / or RUs) and / or terminals. This information can be used as training data or inference data. NRT RICs can deliver inference results to RAN nodes and / or terminals. Inference results can be exchanged between CUs and DUs, and / or between DUs and RUs. For example, a NRT RIC delivers an inference result to a DU, which then sends it to an RU. Non-real-time (NRT) RICs are also used for model training and inference. For example, they are used to train AI models and then use those models for inference. Non-real-time (NRT) RICs can obtain network-side and / or terminal-side information from RAN nodes (e.g., CUs, CU-CPs, CU-UPs, DUs, compute nodes, and / or RUs) and / or terminals. This information can be used as training data or inference data, and inference results can be delivered to RAN nodes and / or terminals. Inference results can be exchanged between CU and DU, and / or between DU and RU. For example, a non-real-time RIC can submit its inference result to a DU, which then forwards it to a RU. Near-real-time and non-real-time RICs can also be configured as separate network elements. Furthermore, near-real-time and non-real-time RICs can be integrated into other devices. For instance, near-real-time RICs can be located in RAN nodes (e.g., CU, DU, compute nodes), while non-real-time RICs can be located in OAM, cloud servers, core network elements, or other network devices.

[0054] In this application, "sending information" can be understood as one device sending information to another device, or it can also be understood as one logic module within a device sending information to another logic module. For example, "sending device sending information" can be understood as the sending device sending information to another device (such as the receiving device), or it can be understood as logic module 1 in the sending device sending information to logic module 2 in the sending device.

[0055] In this application, "receiving information" can be understood as one device receiving information from another device, or it can also be understood as a logical module within a device receiving information from another logical module. For example, "receiving device receiving information" can be understood as the receiving device receiving information from another device (such as the sending device), or it can be understood as logical module 1 in the receiving device receiving information from logical module 2 in the receiving device.

[0056] In this application, phrases such as "sending information to... (e.g., a network device)" or related illustrations in the accompanying drawings can be understood as the destination of the information being a terminal. This can include sending information directly or indirectly to a terminal. Similarly, phrases such as "receiving information from... (e.g., a terminal)," "receiving information from... (e.g., a terminal)," or "receiving information sent by (e.g., a terminal)," or related illustrations in the accompanying drawings, can be understood as the source of the information being a terminal. This can include receiving information directly or indirectly from a terminal. Information may undergo necessary processing between the source and destination, such as format changes, but the destination can understand the valid information from the source. Similar expressions in this application can be interpreted similarly and will not be elaborated further here.

[0057] The communication system 10 in this application embodiment includes, but is not limited to: narrow band-internet of things (NB-IoT), long term evolution (LTE), long term evolution advanced (LTE-A), and 5G mobile communication system, etc.

[0058] This application's embodiments can be applied to 5G or future communication systems, satellite communication, and short-range wireless communication systems. Wireless communication systems can consist of cells, each containing a network-side device (such as a base station), which provides services to multiple terminals. Wireless communication systems can also perform point-to-point communication, such as communication between multiple terminals. A network-side device refers to a device deployed in a wireless access network that provides wireless communication functions for a mobile station (MS), such as a base station. In systems employing different wireless access technologies, the names of devices with base station functions may differ. The MS involved in this application's embodiments can include various handheld devices, vehicle-mounted devices, wearable devices, computing devices, or other processing devices connected to a wireless modem with wireless communication functions. An MS can also be called a terminal, and can also be a subscriber unit, cellular phone, smartphone, wireless data card, personal digital assistant (PDA) computer, tablet computer, wireless modem, handset, laptop computer, MTC terminal, etc.

[0059] The network device in this application embodiment can be the aforementioned RAN node (or access network device, radio access network device). The network device can also be a core network element or a device connected to the core network. Core network elements can be used to provide core network services to access network devices and terminal devices. Core network elements can correspond to different devices in different systems. For example, in 4G, core network elements can correspond to a mobility management entity (MME) and / or a serving gateway (S-GW). In 5G, core network elements can correspond to access and mobility management functions (AMF), session management functions (SMF), or user plane functions (UPF).

[0060] Network devices and terminal devices can be deployed on the ground, including indoors or outdoors, handheld or vehicle-mounted; they can also be deployed on water; and they can also be deployed in the air on airplanes, balloons, and satellites. This application does not limit the scenario in which the network devices and terminal devices are located.

[0061] The communication method and apparatus of the present application embodiments will be further described below with reference to the accompanying drawings. It is understood that the present application uses a sending end (including at least a first node and a second node) and a receiving end as examples of the execution entities in the interaction illustration, but the present application does not limit the execution entities in the interaction illustration. For example, the method executed by the receiving end in the present application can also be implemented by a module (e.g., a circuit, chip, or chip system) in the receiving end, or by a logic node, logic module, or software that can implement all or part of the receiving end's functions; the method executed by the sending end in the present application can also be implemented by a communication / processing module in the sending end or by a circuit or chip responsible for communication / processing functions in the terminal (such as a modem chip (also known as a baseband chip), or a SoC chip / SIP chip containing a modem core, or a GPU / NPU / AI processor / ASIC), or by a logic node, logic module, or software that can implement all or part of the sending end's functions.

[0062] Current multiple access schemes are based on, for example, modular division multiple access (MDMA) and semantic feature multiple access (SMA). Personalized semantics in MDMA rely on orthogonal access. Both MDMA and SMA require the deployment of multiple decoders on the network (NW) side, leading to scalability issues in large-scale scenarios and increasing the difficulty of cross-vendor collaboration. Therefore, some embodiments of this application propose a communication method for data transmission scenarios in multiple access, addressing the problems of current multiple access schemes. This method leverages the orthogonality of the model domain to achieve user multiple access and utilizes AI models to further improve the efficiency of multiple access.

[0063] Figure 2A The diagram illustrates the communication flow of some embodiments of this application. For example... Figure 2A As shown, process 201 involves a receiver 210 and a sender. The sender may include one or more nodes. Figure 2A Exemplary examples show a first node 220 and a second node 230 at the sending end. In different examples, the data transmission from the sending node to the receiving end 210 can be uplink data transmission, downlink data transmission, or sidelink data transmission. The communication flow between the receiving end 210 and the sending end is illustrated below using examples of the receiving end 210 communicating with the first node 220 and the second node 230 respectively. It is understood that the implementation of flow 201 is not limited to the example shown, and the sending node is not limited to the nodes shown. Furthermore, it should be noted that the first node 220 and the second node 230 can be the same node or different nodes. Figure 2A Let's take an example where the first node 220 and the second node 230 are different nodes. For an example where the first node 220 and the second node 230 are the same node, then... Figure 2A The two data transmission actions shown are both performed by the same node (e.g., the first node 220).

[0064] At point 202, the first node 220 and the second node 230 send data through the first resource. The data sent by the first node 220 can be referred to as third data 205 (represented as X' in the following example), and the data sent by the second node 230 can be referred to as fourth data 215 (represented as Y' in the following example). Since the third data 205 and the fourth data 215 are sent through the same resource (i.e., the first resource), they will be superimposed during transmission, thus forming superimposed data (represented as X'+Y' in the following example, where "+" indicates superposition). Therefore, at the receiving end 210, the receiving end 210 can receive this superimposed data through the first resource.

[0065] In some examples, before sending data through the first resource, the first node 220 and the second node 230 process the data using their respective AI models. That is, the third data 205 and the fourth data 215 are obtained by processing the data using the AI ​​models of the first node 220 and the second node 230, respectively. For example, the data from the first node 220 (referred to as X in the following example) is processed by the AI ​​model at the first node 220 (referred to as the first AI model) to obtain the third data 205 (e.g., X'). The data from the second node 230 (referred to as Y in the following example) is processed by the AI ​​model at the second node 230 (referred to as the second AI model) to obtain the fourth data 215 (e.g., Y').

[0066] At 204, receiver 210, based on the first indication information or second resource corresponding to the first AI model, processes the superimposed data through the third AI model to obtain the first data corresponding to the first AI model (which may be represented as X” in some examples). In some examples, receiver 210 may receive the first indication information from first node 220. In some examples, the first indication information corresponding to the first AI model may be used to indicate one or more of the identifier, function, or training dataset of the first AI model. The second resource may be used to carry the output data of the first AI model, and the output data of the first AI model carried may be, for example, the data X mentioned above obtained by processing the data X through the first AI model. The third data 205 (i.e., data X') can be data obtained by processing other data from the first node 220, different from the aforementioned data X, through the first AI model (which can be represented as Z'). Due to the influence of various conditions such as channel transmission, the data obtained by the receiver through the third AI model is not exactly the same as the data input to the first AI model by the transmitter. Therefore, the data obtained by the receiver through further processing of the data processed by the first AI model after receiving the data can be called the first data corresponding to the first AI model. In some examples, the first data can be an estimate of the data input to the first AI model by the transmitter (e.g., data X), or it can be called the recovered data of the data input by the transmitter (e.g., data X).

[0067] In some examples, such as channel state information (CSI) compression feedback, different encoders are deployed at the transmitting end and decoders are deployed at the receiving end. The data from the transmitting end is processed by the first AI model, encoded, modulated, and then sent to the receiving end. At the receiving end, it is demodulated, decoded, etc., and processed by the third AI model to obtain the aforementioned first data corresponding to the data from the transmitting end.

[0068] In some examples of implementing step 204, the receiving end 210 may use the first indication information and the superimposed data as input to the third AI model to process the superimposed data and obtain the first data (e.g., data X”) from the output of the third AI model.

[0069] In some other examples of implementing step 204, the receiving end 210 can receive the output data (e.g., data X' or data Z') of the first AI model through the second resource. Further, the receiving end 210 can input the output data (e.g., data X' or data Z') of the first AI model along with superimposed data into the third AI model, and obtain the first data (e.g., data X”) from the output of the third AI model. It should be noted that the output data of the first AI model may include the direct output result of the first AI model (i.e., without further processing), or it may include the result of further processing of the direct output result of the first AI model. For example, after processing the input data through the first AI model, the transmitting end can perform processing such as channel coding and modulation before transmitting the data through the second resource.

[0070] At 206, the receiving end 210, based on the second indication information or the third resource corresponding to the second AI model, processes the superimposed data through the third AI model to obtain the second data corresponding to the second AI model (which may be represented as Y” in some examples). In some examples, the receiving end 210 can receive the second indication information from the second node 230. In some examples, the second indication information is used to indicate one or more of the identifier, function, or training dataset of the second AI model. In some examples, the third resource is used to carry the output data of the second AI model, and the output data of the second AI model carried may be, for example, the fourth data 215 (i.e., data obtained by processing the aforementioned data Y through the second AI model). According to Y'), or it can be data obtained by processing other data different from the aforementioned data Y from the second node 230 through the second AI model (which can be represented as V'). Due to the influence of various conditions such as channel transmission, the data obtained by the receiver through the third AI model is not exactly the same as the data input to the second AI model by the transmitter. Therefore, the data obtained by further processing the data processed by the second AI model after the receiver receives the data can be called the second data corresponding to the second AI model. In some examples, the second data can be an estimate of the data input to the second AI model by the transmitter (e.g., data Y), or it can be called the recovered data of the data (e.g., data Y) of the transmitter.

[0071] Examples of the first resource mentioned above can be found below. Figure 4 or Figure 13 As shown below. Examples of the second and third resources mentioned above can be found in the following text. Figure 13As shown in the figure. In some examples, the first resource, the second resource, and / or the third resource mentioned above may include one or more of the following: time domain resources, frequency domain resources, spatial domain resources, or code domain resources.

[0072] Some example processes for implementing step 206 are similar to those for implementing step 204 above. For example, receiver 210 can input the second indication information and superimposed data into the third AI model to process the superimposed data and obtain the second data (e.g., Y”) from the output of the third AI model. Alternatively, receiver 210 can receive the output data of the second AI model (e.g., data Y' or data V') through a third resource and input the output data of the second AI model and superimposed data into the third AI model to obtain the second data (e.g., Y”) from the output of the third AI model. The output data of the second AI model may include the direct output result of the second AI model (i.e., without further processing) or the result of further processing of the direct output result of the second AI model (e.g., channel coding, modulation, etc.).

[0073] In some other example processes implementing step 206, the receiving end 210 can process the superimposed data through a third AI model based on the first data (e.g., X”) and the second indication information or a third resource to obtain the second data (e.g., Y”). For example, the receiving end 210 can input the first data (e.g., X”), the second indication information, and the superimposed data into the third AI model, and obtain the second data (e.g., Y”) from the output of the third AI model. Alternatively, the receiving end 210 can receive the output data of the second AI model (e.g., data Y' or data V') through a third resource. Further, the receiving end 210 can input the first data (e.g., X”), the received output data of the second AI model, and the superimposed data into the third AI model, and obtain the second data (e.g., Y”) from the output of the third AI model. In such example processes, the third AI model used by the receiving end can be called a generative model, and the data processing of the generative model can be further explained below. Figure 9 or Figure 12 The example shown.

[0074] In some examples, the overlay data may include one or more tokens. For example, the sender and receiver may communicate using tokens, where a token can be understood as a basic unit of text, data, or features in the processing of a model, such as an AI model. For example, text, data, or features in the AI ​​model processing may be represented as tokens for processing and transmission. For example, the input and output of an AI model may contain a series of numbers or vectors, which may be called tokens, or these numbers or vectors may be represented as tokens. Optionally, a token may also be called a mark, term, or token, etc., and the naming of the token is not limited in this application embodiment.

[0075] As mentioned above, the first node 220 and the second node 230 can be different nodes. In this case, the first data corresponds to the first node 220, and the second data corresponds to the second node 230. Alternatively, the first node 220 and the second node 230 can be the same node (e.g., the first node 220). In this case, the first data and the second data mentioned above correspond to the same node, for example, both correspond to the first node 220.

[0076] Figure 2B The following are schematic diagrams illustrating the communication flow of some other embodiments of this application. For example... Figure 2B As shown, in process 211, a receiver 210 and a sender are involved. The sender may include one or more nodes, and... Figure 2A similar, Figure 2B The first node 220 and the second node 230 are illustrated exemplarily. The first node 220 and the second node 230 can be the same node or different nodes. For a detailed description of the receiving end 210, the sending end and their nodes (e.g., the first node 220 and the second node 230), please refer to the above-described process 201. Steps 216, 218 and 222 in process 211 correspond to the same steps 202, 204 and 206 in process 201, respectively. These steps can be found in the description of process 201.

[0077] In some examples, the first and second indication information upon which the overlay data is processed by the third AI model are received by the receiver 210 from the first node 220 and the second node 230, respectively, see steps 212 and 214 in process 211. Specifically, at 212, the first node 220 sends the first indication information 225a to the receiver 210. Accordingly, the receiver 210 receives the first indication information 225a. At 214, the second node 230 sends the second indication information 225b to the receiver 210. Accordingly, the receiver 210 receives the second indication information 225b.

[0078] The local AI models deployed on different nodes are typically different and can be personalized local models. For example, the first node 220 and the second node 230 can obtain local AI models by acquiring data and training them, or they can be configured via the network. In some examples, the first AI model and the second AI model in the above process can be referred to as node-specific AI models, node-specific AI models, node-personalized AI models, node-local AI models, or node-local personalized models, etc. For example, in some examples where the node is a UE, according to the above naming rules, the first AI model deployed on UE1 can be called the UE1-specific AI model, the second AI model deployed on UE2 can be called the UE2-specific AI model, and so on. A node can have one or more local personalized models. Local personalized models can be distinguished by the indication information corresponding to the local personalized model (such as the first indication information 225a or the second indication information 225b mentioned above). The indication information corresponding to the local personalized model can be used to indicate the identifier, function, or one or more items in the training dataset of the local personalized model. For example, in an example where multiple AI models are deployed on the first node 220, the receiver 210 can determine that the first node 220 is currently using the first AI model through the first indication information 225a reported by the first node 220. The third AI model in the above process can be referred to as the common model in some examples. The common model is used to extract data sent by different nodes from multi-access data. In the examples below, the first AI model can be represented by W1, the second AI model by W2, and the third AI model by W.

[0079] In some examples, the AI ​​models deployed at the transmitting and receiving ends in this application embodiment can refer to the low-rank adaptation of large language models (LoRA) technology to achieve source encoding and decoding for multi-access communication transmission. LoRA is a low-rank fine-tuning method. Its core idea is to freeze the weights of the pre-trained model and then inject the trainable low-rank decomposition matrix into the layers of the transformer architecture to reduce the number of trainable parameters for downstream tasks. By using different LoRA parameters, the same pre-trained model can be fine-tuned for different downstream tasks. In such examples, the multiple access in this application embodiment can be called LoRA multiple access, and the AI ​​model can be called a low-rank model accordingly. For example, in an uplink multiple access example where the transmitting end is a UE and the receiving end is a NW, a personalized low-rank model (i.e., different LoRA parameters) is deployed on the UE side. The same common model can be combined with different LoRA parameters to complete different tasks, thereby enabling multiple UEs to achieve uplink multiple access through different LoRA parameters. Alternatively, the above-mentioned LoRA multiple access can also be applied to downlink multiple access or sidelink communication scenarios. In uplink multiple access (MLA) processes, the transmitting end is the terminal side (e.g., the UE side) and the receiving end is the NW side. In downlink multiple access (MLA) processes, the transmitting end is the NW side and the receiving end is the terminal side (e.g., the UE side). In sidelink communication, both the transmitting end and the receiving end are terminal sides (e.g., the UE). Although the multiple access processes illustrated in the following examples are described using uplink multiple access, they can also be applied to downlink multiple access or sidelink communication.

[0080] Figure 3 A schematic diagram of a multiple access process according to some embodiments of this application is shown. This multiple access process may involve multiple nodes, with a first node (e.g., the first node 220 in the example above) and a second node (e.g., the second node 230 in the example above) exemplarily shown. W1 and W2 are local personalized models of the nodes, which can be examples of a first AI model and a second AI model, respectively. A common model (denoted by W) is deployed on the receiving side; the common model W can be an example of a third AI model. Taking uplink multiple access as an example, the first node is, for example, UE1, the second node is, for example, UE2, and the receiving side is, for example, the NW side.

[0081] The first node processes data X using a local personalized model W1, and the second node processes data Y using a local personalized model W2. Data X processed by personalized model W1 results in data X', which is an example of the third data. Data Y processed by personalized model W2 results in data Y', which is an example of the fourth data. The first and second nodes respectively send their results (i.e., data X' and data Y') processed by their local personalized models to the sending end. The input to a node's local personalized model is the data that the node needs to upload to the sending end, such as X or Y, and its output is a high-dimensional vector representation of the corresponding data projected into a high-dimensional space, such as data X' or Y'. The vector representations of data from different nodes after their respective personalized processing will be orthogonal in the high-dimensional space.

[0082] In some examples, multiple nodes can reuse the same time-frequency resource (e.g., the first resource) to send their respective local personalized model output data to the NW, and the occupancy of their time-frequency resources is as follows: Figure 4 As shown, the first resource represents the time-frequency resource shared by the first node, such as UE1, and the second node, such as UE2.

[0083] In some examples, data can also be referred to as signals. The receiver receives signals from nodes, which are data X' and Y' superimposed together. The superimposed data X' and Y' can be called superimposed data, represented as X'+Y'. Superimposed data is, for example, the superposition result of high-dimensional orthogonal high-dimensional vector representations of data from multiple nodes. Superimposed data can also be understood as data formed by waveform aliasing during air interface transmission of multiple data (such as X' and Y' mentioned above) transmitted on the same resource. The receiver inputs the received superimposed data into a common model W to complete signal detection. The output of the common model W is an estimate of the data reported by the UE to the NW, such as data X” and Y”. Signal detection is the process by which the receiver obtains data X” and Y” from the superimposed data through the common model W.

[0084] Figure 5A and Figure 5B The diagram illustrates signal detection at the receiver of some embodiments of this application. As mentioned in some examples above, first data can be obtained by processing superimposed data through a third AI model based on first indication information corresponding to a first AI model, and second data can be obtained by processing superimposed data through a third AI model based on second indication information corresponding to a second AI model. Figure 5A and Figure 5BIn the example, the first or second indication information can be a LoRA ID. Specifically, LoRA1 ID is used as an example of the first indication information to indicate at least the identifier, function, and / or training dataset of the first AI model, and LoRA2 ID is used as an example of the second indication information to indicate the identifier, function, and / or training dataset of the second AI model. During signal detection, the receiver inputs the received superimposed data X'+Y' and LoRA IDs (e.g., LoRA1 ID or LoRA2 ID) into the common model W to obtain the output data X" and Y" of the common model W, thus completing signal detection. The receiver can run the common model W multiple times to obtain data X" and data Y" sequentially, or the receiver can combine the superimposed data X'+Y' from the node with a combination of LoRA IDs (e.g., a combination of LoRA1 ID and LoRA2 ID) into a batch input and input it into the common model for batch processing. See the following descriptions of signal detection based on multiple processing and batch processing methods for details.

[0085] In some examples, the LoRA ID can be an N-bit binary string sent from the transmitting node to the receiving node. Taking the transmitting node as a UE as an example, it can be carried through UE assistance information (UAI) or uplink control information (UCI). In some examples, the LoRA ID can be transmitted after being scrambled by a radio network temporary identifier (RNTI). LoRA ID can indicate the identifier, function, and / or training dataset of the corresponding AI model. In some examples, the LoRA ID can be used as input data in the model calculation process. Additionally or alternatively, the LoRA ID can be used to indicate the activation of some parameters in the common model W, and the LoRA1 ID can be used to activate some parameters in the common model W corresponding to the first AI model. In examples where the common model is based on a mixture of experts (MOE) architecture (containing several sub-models, or expert models), the LoRA ID can be used as an indication to implement routing in the MOE model, i.e., to activate the corresponding sub-model or expert model.

[0086] Figure 5AThe signal detection at the receiver shown is based on a multi-processing approach. Specifically, when processing the overlay data X'+Y' using the common model W, firstly, based on the LoRA1 ID, the overlay data X'+Y' is processed using the common model W to obtain data X'' corresponding to the local personalized model W1 of the first node, such as UE1. This data X''' is, for example, an estimate of the data X'' from the first node, or the recovered data of data X''. Then, based on the LoRA2 ID, the overlay data X'+Y'' is processed using the common model W to obtain data Y'' corresponding to the local personalized model W2 of the second node, such as UE2. This data Y''' is, for example, an estimate of the data Y'' from UE2, or the recovered data of data Y''. In this signal detection method, data X''' and data Y''' are obtained through serial processing.

[0087] Figure 5B The signal detection at the receiver shown is based on batch processing. Specifically, the operation of processing superimposed data X'+Y' using the LoRA1 ID and the operation of processing superimposed data X'+Y' using the LoRA2 ID and the common model W are executed as part of the same batch. In this signal detection method, data X' and data Y' are obtained through parallel processing.

[0088] Figure 6 A schematic diagram of a multiple access process according to some embodiments of this application is shown. Figure 6 In the multiple access process shown, on the transmitting end side, the data transmission process includes first / second AI model processing, channel coding, modulation, and channel transmission. First AI model processing involves processing data X from the first node using the first AI model, and processing data Y from the second node using the second AI model. Taking a UE as an example, the transmitting end includes the UE's local personalized model (i.e., source coding model, such as W1 for UE1 and W2 for UE2), a channel coding module, and a modulation module. After the above processing by the transmitting end, the resulting data X' and Y' are transmitted in the channel through the same resource (i.e., the first resource), meaning the transmitted data is superimposed data X'+Y'. Correspondingly, the data received at the receiving end is superimposed data X'+Y'. In some examples, the transmitting end's local personalized model can integrate the aforementioned modules for channel coding and modulation. In such examples, the channel coding and modulation operations on the transmitting end side are completed within the transmitting end's local personalized model.

[0089] At the receiving end, processes such as channel equalization, demodulation, channel decoding, and common model processing can be performed. Common model processing includes the signal detection process described above; for details, please refer to the section above. Figure 5A and Figure 5BThe example shown is an introduction. In some examples, the common model W can integrate the aforementioned processing modules for channel equalization, demodulation, or channel decoding, so that channel equalization, demodulation, or channel decoding operations will not need to be performed outside of the common model W at the receiver side during the multiple access process, such as... Figure 6 As shown. In other examples, operations such as channel equalization, demodulation, or channel decoding can be performed outside the common model W at the receiver. That is, the received superimposed data is first subjected to operations such as channel equalization, demodulation, or channel decoding before being input into the common model W.

[0090] Figure 7 The diagram illustrates the signaling interaction of multiple access according to some embodiments of this application. Figure 7 The process 700 shown involves a receiver 210 and a transmitter node, wherein the transmitter node is shown as an example of a first node 220 and a second node 230.

[0091] At 702, the first node 220 reports LoRA1 ID 705 to the receiving end 210, where LoRA1 ID 705 is an example of the first indication information. At 704, the second node 230 reports LoRA2 ID 715 to the receiving end 210, where LoRA2 ID 715 is an example of the second indication information. At 706a and 706b, the first node 220 and the second node 230 respectively perform local personalized model processing. For details of the local personalized model processing, please refer to the process of processing data X through the first AI model (e.g., personalized model W1) and the process of processing data Y through the second AI model (e.g., personalized model W2) in some examples above. At 708, the first node 220 and the second node 230 respectively return their respective local personalized model processing results (i.e., the first processing result 725, such as data X', and the second processing result 735, such as data Y') to the receiving end 210 through the same resource (i.e., the first resource). The local personalized model processing results returned by the two nodes form the superimposed data X'+Y'. At 712, the common model (e.g., common model W) of the receiver 210 performs signal detection through multiple processing or batch processing to obtain data X (an example of first data) corresponding to the local personalized model of the first node 220 and data Y (an example of second data) corresponding to the local personalized model of the second node 230.

[0092] As mentioned in some examples above, in the process of obtaining the second data by processing the superimposed data through a third AI model based on the second indication information at the receiving end, it can be specifically implemented as follows: the receiving end processes the superimposed data through a third AI model based on the first data and the second indication information to obtain the second data. The following will combine... Figure 8 A detailed explanation is provided. Figure 8In the multiple access process illustrated, the receiving end can employ a generative model (such as a decoder-only architecture). The receiving end can sequentially detect the signals of the transmitting end's nodes. After detecting the signal of a certain node, the recovered signal is input again into the third AI model to facilitate the detection of the signal of the next transmitting node. The detected signal here refers to the data obtained after processing by the third AI model, such as the first data or the second data. In some examples, the way the receiving end performs signal detection based on a generative model can be called autoregressive processing or generative processing.

[0093] for Figure 8 The signal detection diagram at the receiving end of the multiple access process shown is as follows: Figure 9 As shown. Taking a sending node consisting of a first node and a second node as an example, the third AI model (e.g., the common model W) takes the received overlay data (X'+Y') and the LoRA1 ID used by the first node as input, and outputs the recovered data X” of the first node's data X, i.e., the first data. Then, the recovered data (X”) of the first node's data X is input again into the third AI model. The third AI model takes the received overlay data (X'+Y'), the LoRA2 ID used by the second node, and the recovered data X” of the first node's data X as input, and outputs the recovered data Y” of the second node's data Y, i.e., the second data.

[0094] Figure 8 The operations on the receiving end side during the multiple access process include the above. Figure 9 For the signal detection process and receiver operation, please refer to [link / reference]. Figure 9 The description. In Figure 8 The operation of the multiple access procedure on the transmitting end side shown is similar to... Figure 6 The operations on the transmitting end side are the same, including first / second AI model processing, channel coding, modulation, and channel transmission, etc. See details for the section on... Figure 6 The examples provided will not be repeated here.

[0095] As mentioned in some examples above, the receiving end can process the overlay data using a third AI model based on a second resource to obtain first data corresponding to the first AI model, while the second resource is used to carry the output data of the first AI model. Similarly, based on the third resource, the overlay data can be processed using a third AI model to obtain second data corresponding to the second AI model, while the third resource is used to carry the output data of the second AI model. For detailed implementation details, please refer to [link to relevant documentation]. Figures 10 to 12The signal detection process at the receiving end is shown below. Taking the receiving end as NW and the transmitting end node as UE as an example, the NW allocates non-orthogonal time-frequency resources to each UE (the same resource element (RE) can be allocated to multiple UEs). Specifically, of the time-frequency resources allocated by the NW to the UE, some time-frequency resources are unique to the UE, while others are shared with other UEs. The time-frequency resources shared by all UEs can be called common resources. In this case, the NW can complete signal detection based on multiple processing / batch processing / generative processing methods. The following describes each of these methods.

[0096] exist Figures 10 to 12 In the example shown, the output data of the first AI model received through the second resource can be data X' or data Z', and the output data of the second AI model received through the third resource can be data Y' or data V'. For details on data X' or data Z', and data Y' or data V', please refer to the description of the example above. The superimposed data received through the first resource is represented by X'+Y'. The first resource can be a common resource of all nodes at the sending end, while the second and third resources can be unique resources of different nodes at the sending end. An example of the third AI model is the common model W. An example of the first data corresponding to the first AI model is represented by X”, and an example of the second data corresponding to the second AI model is represented by Y”.

[0097] Figure 10 The signal detection method shown is a multi-processing method, meaning that NW can run the common model W multiple times to obtain the first data (X”) and the second data (Y”) sequentially. The specific implementation process is as follows... Figure 5A The signal detection processes shown are similar, the difference being that... Figure 5A In the example shown, the indicator information (first indicator information such as LoRA1 ID, or second indicator information such as LoRA2 ID) and the overlay data (X'+Y') are used as inputs to the common model W. Figure 10 The example shown uses the output data of the sender's local personalized model received through the sender node's unique resources, along with overlay data (X'+Y'), as input to the common model W. For instance, the output data X' or Z' of UE1's local personalized model (an example of the first AI model) received through the sender node's (e.g., UE1's) unique resources (second resources), along with overlay data (X'+Y'), is used as input to the common model W. After processing by the common model W, data X" (an example of the first data) is output. Regarding... Figure 10 The implementation process of the example can be found in the above text. Figure 5A An example of this is provided.

[0098] Figure 11The signal detection method shown is a batch processing method, and its signal detection process is similar to... Figure 5B The signal detection process shown is similar to... Figure 5B The difference is that, Figure 11 The signal detection process involves the receiver combining the superimposed data from the UE with data Z' (or X') received through the second resource and data V' (or Y') received through the third resource to form a batch of input data, which is then processed by the common model W. The second and third resources are unique resources of the transmitting node. (Regarding...) Figure 11 The implementation process of the example can be found in the above text. Figure 5B An example of this is provided.

[0099] Figure 12 The signal detection method shown is based on generative processing, also known as autoregressive processing, and its signal detection process is similar to... Figure 9 The signal detection process shown is similar to... Figure 9 The difference is that, Figure 11 In the signal detection process, the data Z' (or data X') that serves as the input data of the third AI model (e.g., the common model W) at the receiving end replaces... Figure 9 In the example, the LoRA1 ID is replaced by data V' (or data Y') as another input data of the third AI model at the receiving end. Figure 9 The LoRA2 ID in the example. About Figure 12 The implementation process of the example can be found in the above text. Figure 9 An example of this is provided.

[0100] Figure 13 This illustration shows a time-frequency resource usage diagram for multiple access in some embodiments of this application. For example... Figure 13The diagram illustrates time-frequency resources shared by multiple nodes at the transmitting end, as well as time-frequency resources unique to each node. The shared time-frequency resources can be considered an example of a first resource. The time-frequency resources corresponding to the first node (i.e., the unique time-frequency resources of the first node) can be considered an example of a second resource. The time-frequency resources corresponding to the second node (i.e., the unique time-frequency resources of the second node) can be considered an example of a third resource. Both the first and second nodes can transmit data through the first resource; for example, the first node transmits data X' through the first resource, and the second node transmits data Y' through the first resource. The first node can also transmit data Z' (or data X') through the second resource, and the second node can also transmit data V' (or data Y') through the third resource. The aforementioned data Z' and data X' transmitted by the first node through the first resource can be obtained by the first node processing different data using the same personalized model (i.e., a first AI model). The aforementioned data V' and data Y' transmitted by the second node through the first resource can be obtained by the second node processing different data using the same personalized model (i.e., a second AI model).

[0101] In some examples, the NW can allocate orthogonal time-frequency resources to each UE. The NW then receives signals (e.g., data X' and data Y') from each UE (e.g., UE1 and UE2) using the orthogonal time-frequency resources, and then performs multiple processing or batch processing using a third AI model (e.g., common model W) to complete signal detection. Figure 14 and Figure 15 The signal detection process is illustrated below, where Figure 14 The example shown illustrates a signal detection process that involves multiple processing steps. For instance, X' is first input into a common model W to output X', and then Y' is input into the common model W to output Y'. Figure 15 The example shown illustrates a signal detection process using batch processing, where X' and Y' are input as the same batch of data into a common model W to output X" and Y".

[0102] For cases where the transmitter sends data to the receiver using orthogonal time-frequency resources, the corresponding multiple access procedure can be referenced. Figure 16 The operations of the sending end and Figure 6 or Figure 8 The multiple access process is the same as described above; please refer to the relevant descriptions in the examples above for details. Data from the sending end is transmitted through the channel to the receiving end, and the receiving end's operations include those described above. Figure 14 or Figure 15 The signal detection process also includes operations such as channel equalization, demodulation, or channel decoding. In such an example, the data input to the common model W in the above signal detection process is the data after channel equalization, demodulation, or channel decoding. In other examples, Figure 16 The common model W on the receiver side can integrate the aforementioned processing modules for channel equalization, demodulation, or channel decoding. Accordingly, channel equalization, demodulation, or channel decoding will not need to be performed outside the common model W on the receiver side of the multiple access process. Figure 17 A time-frequency resource occupancy diagram applicable to the above-described multiple access process is shown, wherein the time-frequency resource (fourth resource) corresponding to UE1 and the time-frequency resource (fifth resource) corresponding to UE2 are shown by way of example. The two time-frequency resources are orthogonal.

[0103] Figure 18 A possible exemplary block diagram of the communication device involved in an embodiment of this application is shown. For example... Figure 18 As shown, the communication device 1800 may include modules or units for implementing the methods described in the embodiments above. In one possible design, the communication device 1800 includes a processing unit 1802 and a communication unit 1803. Optionally, the communication device 1800 may further include a storage unit 1801 for storing device program code and / or data.

[0104] The communication device 1800 can be the receiving end device or apparatus in the above embodiments. For example, the receiving end can be a network device or terminal, or a communication module / processing module in a network device or terminal, or a circuit or chip in a network device or terminal responsible for communication functions, etc.

[0105] For example, in one embodiment, the communication unit 1803 is used to: receive overlay data through a first resource.

[0106] In one embodiment, processing unit 1802 is configured to process overlaid data through a third AI model based on first instruction information or a second resource corresponding to a first AI model, to obtain first data corresponding to the first AI model, wherein the second resource is used to carry the output data of the first AI model. Processing unit 1802 is also configured to process overlaid data through a third AI model based on second instruction information or a third resource corresponding to a second AI model, to obtain second data corresponding to the second AI model, wherein the third resource is used to carry the output data of the second AI model.

[0107] In one possible design, the communication unit 1803 is specifically used to receive first indication information, which indicates one or more of the identifier, function, or training dataset of the first AI model.

[0108] In one possible design, the communication unit 1803 is specifically used to receive second indication information, which indicates one or more of the identifier, function, or training dataset of the second AI model.

[0109] In one possible design, the processing unit 1802 is used to process the overlay data through a third AI model based on the second instruction information or the third resource to obtain the second data, including: the processing unit 1802 processes the overlay data through a third AI model based on the first data and the second instruction information or the third resource to obtain the second data.

[0110] In one possible design, the overlay data includes one or more tokens.

[0111] In one possible design, the first data corresponds to the first node, and the second data corresponds to the second node; or the first data and the second data correspond to the first node.

[0112] In one possible design, when the communication device 1800 is a communication module in a terminal or network device, the function of the processing unit 1802 can be implemented by one or more processors. Specifically, the processor may include a modem chip, or a system-on-a-chip (SoC) chip or a SIP chip containing a modem core. The function of the communication unit 1803 can be implemented by transceiver circuitry.

[0113] In one possible design, when the communication device 1800 is a circuit or chip responsible for communication functions in a terminal or network device, such as a modem chip or a system-on-a-chip (SoC) or SIP chip containing a modem core, the function of the processing unit 1802 can be implemented by a circuit system in the aforementioned chip that includes one or more processors or processor cores. The function of the communication unit 1803 can be implemented by interface circuitry or data transceiver circuitry on the aforementioned chip.

[0114] In one possible design, when the communication device 1800 is a communication and / or computing module in a terminal or network device, the functionality of the processing unit 1802 can be implemented by one or more processors. Specifically, the processor may include a GPU, or a system-on-a-chip (SoC) or SIP chip containing a GPU. Alternatively, the processor may include an AI processor, or a SoC or SIP chip containing an AI processor. Alternatively, the processor may include an ASIC, or a SoC or SIP chip containing an ASIC. Alternatively, the processor may include an NPU, or a SoC or SIP chip containing an NPU. The functionality of the communication unit 1803 can be implemented by transceiver circuitry.

[0115] In one possible design, when the communication device 1800 is a circuit or chip in a terminal or network device responsible for communication and / or computing functions, such as a GPU or a system-on-a-chip (SoC) or SIP chip containing a GPU, an AI processor or a SoC or SIP chip containing an AI processor, an NPU or a SoC or SIP chip containing an NPU, or an ASIC or a SoC or SIP chip containing an ASIC, the function of the processing unit 1802 can be implemented by a circuit system in the aforementioned chip that includes one or more processors or processor cores. The function of the communication unit 1803 can be implemented by interface circuits or data transceiver circuits on the aforementioned chip.

[0116] It is understood that the division of units in the above-described device is merely a logical functional division. One function can correspond to one functional unit, or two or more functions can be integrated into one functional unit. In actual implementation, all or some units can be integrated onto a single physical entity, or distributed across different physical entities. Furthermore, the aforementioned functional units can be implemented in hardware, software, or a combination of both. Whether a function is executed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for specific applications, but such implementations should not be considered beyond the scope of this application.

[0117] In one example, the functional unit in any of the above devices may be one or more integrated circuits configured to implement the above methods, such as: one or more application-specific integrated circuits, or one or more CPUs, one or more microcontroller units (MCUs), one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.

[0118] In one example, storage unit 1801 may include random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory and / or registers, etc.

[0119] See Figure 19 This is a schematic diagram of the structure of a terminal 1900 provided in an embodiment of this application. The terminal 1900 can correspond to the sending or receiving end shown in the aforementioned embodiments, such as a first node or a second node. Figure 19 As shown, the terminal includes: one or more antennas 1910, a radio frequency processing system 1920, and a processor system 1930.

[0120] In the downlink or sidelink direction, the RF processing system 1920 receives RF signals through the antenna 1910 and sends the RF-processed signals to the processor system 1930 for further processing. In the uplink or sidelink direction, the processor system 1930 processes the terminal-side information and sends it to the RF processing system 1920, which then processes the signal and transmits it through the antenna 1910.

[0121] In one example, the radio frequency (RF) processing system 1920 serves as the communication interface for external communication of the terminal and may include an RF front end (RFFE) 1921 and an RF transceiver 1922. The RFFE 1921 is primarily used for one or more processing operations, such as shaping, passband selection, or gain adjustment, on the RF signals received by the antenna or those to be transmitted through the antenna. It may include one or more components such as RF switches, duplexers, filters, power amplifiers, antenna tuners, and low-noise amplifiers. The RFFE 1921 can be a circuit system composed of multiple discrete components or integrated into one or more chips. The RF transceiver 1922 processes the RF signals received by the RFFE into baseband / IF signals for further processing by the processor system 1930, and processes the baseband / IF signals provided by the processor system 1930 into RF signals for transmission to the RFFE 1921. The baseband / IF signals transmitted between the RF transceiver 1922 and the processor system 1930 can be digital or analog signals. The RF transceiver 1922 can be implemented by one or more chips, which are commonly referred to as RF ICs.

[0122] In one example, processor system 1930 may include one or more processors for processing signals and executing one or more communication protocols. Optionally, processor system 1930 may also include memory 1936. In one example, the one or more processors include at least one baseband processor 1931 (also known as a modem processor). Memory 1936 is used to store data and / or computer program instructions. Optionally, processor system 1930 may also include one or more application processors 1932 for implementing processing of the terminal operating system and application layer. Application processor 1932 may include, for example, a GPU, NPU, AI processor, or ASIC. Optionally, processor system 1930 may also include one or more of a voice subsystem 1933, a multimedia subsystem 1934, or an interface circuit 1935. The voice subsystem 1933 is used to process voice signals, the multimedia subsystem 1934 is used to handle multimedia-related operations, such as video encoding / decoding, image processing, etc., and the interface circuit 1935 is used to implement communication with other terminal components, such as a display 1940, an input device 1950, memory 1960, etc. The aforementioned components in the processor system 1930 can communicate with each other via a bus or communication interface circuit.

[0123] In one example, the processor system 1930 can be packaged as a single processor chip, such as a SoC chip or a SIP chip. In another example, the processor system 1930 can be a system composed of multiple chips; for example, the baseband processor 1931 can be packaged as a single chip, or packaged with part or all of the circuitry of the radio frequency processing system into a single chip.

[0124] In one example, memory 1936 can be on-chip memory, i.e., located on the processor system 1930 chip. In another example, memory 1960 can be off-chip memory, i.e., located outside the processor system 1930 chip.

[0125] In one example, the baseband processor 1931 may include one or more processor cores 19311 and interface circuitry 19314. The one or more processor cores 19311 are used to process signals and execute one or more communication protocols. Optionally, the baseband processor 1931 may also include a memory 19312 for storing at least a portion of the corresponding computer program instructions and / or data. In one example, the one or more processor cores 19311 execute the computer program instructions stored in the memory 19312 to perform the relevant operations in the above method embodiments (such as determining a first transmission parameter based on experience quality parameters). In this disclosure, memory 19312 is used to store corresponding computer program instructions and / or data. This can mean that memory 19312 stores all corresponding computer program instructions and / or data for execution by processor core 19311; or it can mean that memory 19312 stores a portion of corresponding computer program instructions and / or data, including the computer program instructions and / or data currently required to be executed by processor core 19311. Memory 19312 can store different portions of computer program instructions and / or data multiple times for execution by processor core 19311 to implement the relevant operations in the above method embodiments. Interface circuit 19314 serves as a communication interface for communication with other components, such as transmitting signals with radio frequency processing system 1920, communicating with other subsystems and related components of processor system 1930 via a bus, such as transmitting data control signals with application processor 1932, and transmitting data or computer program instructions with memory 1936 or memory 1960. Optionally, in order to reduce the load on the processor core, a baseband signal processing circuit 19313 can be set to perform at least some of the baseband signal processing work, including one or more of signal demodulation, modulation, encoding or decoding.

[0126] In one example, the communication device provided in this application may be a terminal 1900, a communication module including a processor system 1930 and a radio frequency system 1920, or a baseband processor 1931.

[0127] The processor, processor system, application processor, baseband processor, processor circuit or processor core mentioned above can be collectively referred to as a processor. The processor may include one or more of the following: CPU, DSP, microprocessor unit (MPU), MCU, GPU, FPGA, application-specific integrated circuit, artificial intelligence processor (AI processor) or neural network processor.

[0128] The aforementioned memory may include one or more of the following storage media: random access memory (RAM), static random access memory (SRAM), DRAM, phase-change memory (PCM), resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), cache, register, ROM, flash memory, erasable programmable read-only memory (EPROM), hard disk, etc. In one example, computer program instructions for executing the above embodiments may be stored in non-volatile memory, such as at least a portion of the aforementioned memory 1960 (e.g., one or more of ROM, flash memory, EPROM, or hard disk). When the terminal is running, the corresponding computer program instructions may be partially or wholly loaded onto a memory with a faster transfer speed than the processor, such as at least a portion of memory 1936 and / or memory 19312 (e.g., one or more of RAM, SRAM, DRAM, PCM, RERAM, MRAM, FRAM, cache, or register), for the processor to execute in order to implement the steps in the above method embodiments.

[0129] In one example, the RF transceiver 1922 and the RF front-end 1921 can also be packaged in a single chip. In another example, the RF transceiver 1922, the RF front-end 1921, and the baseband processor 1931 can also be packaged in a single chip.

[0130] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, or all or part of it, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, random access memory (RAM), magnetic disks, or optical disks.

[0131] Based on the above embodiments, this application also provides a computer program that, when run on a computer, causes the computer to execute any of the methods provided in the above embodiments.

[0132] Based on the above embodiments, this application also provides a computer-readable storage medium storing a computer program. When executed by a computer, the computer program causes the computer to perform any of the methods provided in the above embodiments. The storage medium can be any available medium that can be accessed by a computer. By way of example, but not limited to, a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code having the form of instructions or data structures and that can be accessed by a computer.

[0133] Based on the above embodiments, this application also provides a chip for reading a computer program stored in a memory and implementing any of the methods provided in the above embodiments.

[0134] Based on the above embodiments, this application provides a chip system including a processor for supporting a computer device in implementing the functions involved in the communication devices in the above embodiments. In one possible design, the chip system further includes a memory for storing necessary programs and data of the computer device.

[0135] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0136] The terms "system" and "network" in this application embodiment are used interchangeably. "At least one" refers to one or more, and "multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, "at least one of A, B, or C" includes A, B, C, AB, AC, BC, or ABC; "at least one of A, B, and C" can also be understood as including A, B, C, AB, AC, BC, or ABC. Furthermore, unless otherwise specified, the ordinal numbers such as "first" and "second" mentioned in this application embodiment are used to distinguish multiple objects and are not used to limit the order, sequence, priority, or importance of multiple objects.

[0137] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, etc.) containing computer-usable program code.

[0138] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0139] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0140] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

Claims

1. A communication method, characterized in that, include: Receive overlay data through the first resource; Based on the first instruction information or the second resource corresponding to the first artificial intelligence (AI) model, the superimposed data is processed by the third AI model to obtain the first data corresponding to the first AI model. The second resource is used to carry the output data of the first AI model. Based on the second instruction information or the third resource corresponding to the second AI model, the superimposed data is processed by the third AI model to obtain the second data corresponding to the second AI model. The third resource is used to carry the output data of the second AI model.

2. The method according to claim 1, characterized in that, Also includes: The first indication information is received, which is used to indicate the identifier, function, or one or more items in the training dataset of the first AI model.

3. The method according to claim 1 or 2, characterized in that, Also includes: Receive the second instruction information, which is used to indicate one or more of the identifier, function, or training dataset of the second AI model.

4. The method according to any one of claims 1 to 3, characterized in that, Based on the second indication information or the third resource, the overlay data is processed by the third AI model to obtain the second data, including: Based on the first data, and the second indication information or the third resource, the overlay data is processed by the third AI model to obtain the second data.

5. The method according to any one of claims 1 to 4, characterized in that, The overlay data includes one or more tokens.

6. The method according to any one of claims 1 to 5, characterized in that, The first data corresponds to the first node, and the second data corresponds to the second node; or The first data and the second data correspond to the first node.

7. A communication device, characterized in that, include: A communication unit, wherein the communication unit is configured to receive superimposed data through a first resource; The processing unit is used to process the superimposed data through a third AI model based on a first instruction information or a second resource corresponding to the first AI model, to obtain first data corresponding to the first AI model. The second resource is used to carry the output data of the first AI model. The processing unit is further configured to process the superimposed data through the third AI model based on the second instruction information or the third resource corresponding to the second AI model, to obtain the second data corresponding to the second AI model, wherein the third resource is used to carry the output data of the second AI model.

8. The apparatus according to claim 7, characterized in that, The communication unit is also used for: The first indication information is received, which is used to indicate the identifier, function, or one or more items in the training dataset of the first AI model.

9. The apparatus according to claim 7 or 8, characterized in that, The communication unit is also used for: Receive the second instruction information, which is used to indicate one or more of the identifier, function, or training dataset of the second AI model.

10. The apparatus according to any one of claims 7 to 9, characterized in that, The processing unit is used to process the overlay data based on the second indication information or the third resource through the third AI model to obtain the second data, including: The processing unit is used to process the overlay data based on the first data and the second indication information or the third resource through the third AI model to obtain the second data.

11. The apparatus according to any one of claims 7 to 10, characterized in that, The overlay data includes one or more tokens.

12. The apparatus according to any one of claims 7 to 11, characterized in that, The first data corresponds to the first node, and the second data corresponds to the second node; or The first data and the second data correspond to the first node.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program or instructions that, when executed, cause the method according to any one of claims 1 to 6 to be performed.

14. A computer program product, characterized in that, It includes a computer program or instructions that, when executed, cause the method according to any one of claims 1 to 6 to be performed.

Images