WO2025011342A1 - Communication method and apparatus - Google Patents

Abstract

The present application relates to the technical field of communications, and discloses a communication method and an apparatus, used for supporting rapid matching of an encoding model of a terminal device and a decoding model of a network device, and applicable to wireless local area network systems supporting IEEE 802.11ax next-generation Wi-Fi protocols, such as 802.11be, WiFi 7 or EHT, and 802.11 series protocols such as 802.11be next-generation, WiFi 8, UHR, and WiFi AI, and also applicable to a UWB-based wireless personal local area network system, a sensing system and the like. The method comprises: a first device acquires at least one group of channel state information and channel state indication information that correspond to a second device; on the basis of the at least one group of channel state information and channel state indication information, the first device determines whether a first model of the first device matches a second model of the second device; and the first device sends first indication information to the second device, the first indication information being used for indicating whether the first model of the first device matches the second model of the second device.

Description

A communication method and device

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims priority to the Chinese patent application filed with the State Intellectual Property Office of China on July 7, 2023, with application number 202310835699.2 and application name “A Communication Method and Device”, the entire contents of which are incorporated by reference in this application.

Technical Field

The present application relates to the field of communication technology, and in particular to a communication method and device.

Background Art

Wireless communication is developing rapidly. The fifth-generation (5G) mobile communication system and the sixth-generation Wi-Fi (Wi-Fi 6) have been commercialized. The next-generation wireless technology and standardization are in full swing around the world. Wireless communication has penetrated into every aspect of daily life and work and has become an indispensable part. With the rapid growth of the number of smart terminal devices and the popularization of Internet of Things (IoT) devices, new wireless applications such as virtual reality, augmented reality and holographic imaging have emerged.

New wireless technologies, new terminal devices, and new wireless applications have made wireless networks unprecedentedly complex. It is foreseeable that wireless networks will become more and more complex in the future. In order to combat the high complexity trend of wireless networks, artificial intelligence (AI) has become a consensus in the industry as an effective tool for wireless network design and management.

In order to improve the communication rate, it is necessary to increase the bandwidth and the number of antennas, and adopt beamforming technology. Beamforming technology requires the terminal device to feedback channel state information (CSI). However, the increase in bandwidth and the number of antennas greatly increases the overhead of channel state information feedback. With the development of AI, the terminal device can use an AI-based coding model to encode and compress CSI, and send the coding compression result to the access point (AP). The AP uses a decoding model to decode and restore CSI, which can effectively save the feedback overhead of channel state information and thus improve the effective throughput. However, the decoding models used by different APs may be different, and the coding models used by different terminal devices may also be different. When the terminal device switches from one AP to another, how to judge whether the coding model of the terminal device matches the decoding model of another AP is related to the communication quality between the terminal device and the AP.

Summary of the invention

The embodiments of the present application provide a communication method and apparatus to support rapid matching of a coding model of a terminal device and a decoding model of a network device (such as an AP, etc.).

In the first aspect, the embodiment of the present application provides a communication method, which can be executed by a first device, or by a component of the first device (such as a processor, a chip, or a chip system, etc.), or by a logic module or software that can realize all or part of the functions of the first device. The following is an example of the first device executing the method, which includes: the first device obtains at least one set of channel state information and channel state indication information corresponding to the second device; the first device determines whether the first model of the first device matches the second model of the second device according to the at least one set of channel state information and channel state indication information, wherein the first model is used to process the channel state information to obtain the channel state indication information, and the second model is used to process the channel state indication information to obtain the channel state information, or the first model is used to process the channel state indication information to obtain the channel state information, and the second model is used to process the channel state indication information to obtain the channel state indication information; the first device sends the first indication information to the second device, and the first indication information is used to indicate whether the first model of the first device matches the second model of the second device.

In the embodiment of the present application, the first device may be, but not limited to, a network device, and the second device may be, but not limited to, a terminal device. In addition, the first device may also be a terminal device, and the second device may also be a network device.

Through this method, the first device can obtain at least one set of channel state information and channel state indication information corresponding to the second device, which is used to test the first model on the first device side, and can quickly determine whether the first model on the first device side and the second model on the second device side match, thereby achieving rapid matching of the first model on the first device side and the second model on the second device side.

In a possible design, the first device is a first network device, the second device is a first terminal device, the first model is used to process the channel state indication information to obtain the channel state information, and the second model is used to process the channel state information to obtain the channel state indication information; the first device obtains at least one set of channel state information and channel state indication information corresponding to the second device, including: the first network device sends a first request message to the first terminal device, and the first request message is used to request at least one set of channel state information input of the first terminal device and channel state indication information output; the first network device receives at least one set of channel state information input and channel state indication information output from the first terminal device.

In the above design, the first network device can obtain at least one set of channel state information input and channel state indication information output from the first terminal device for model matching judgment, supporting the first terminal device to roam from one network device to the first network device in a scenario, and quickly determine whether the first model on the first network device side and the second model on the first terminal device side match.

In one possible design, the first device is a first network device, the second device is a first terminal device, the first model is used to process channel state indication information to obtain channel state information, and the second model is used to process channel state information to obtain channel state indication information; the first network device sends a second request message to the second network device, the second request message is used to request the second network device to receive at least one set of channel state indication information input and channel state information output corresponding to the first terminal device, and the second network device is a network device that the first terminal device has accessed; the first network device receives at least one set of channel state indication information input and channel state information output from the second network device.

In the above design, the first network device can obtain at least one set of channel state indication information input and channel state information output corresponding to the first terminal device from other network device sides for model matching judgment, which can be used in multi-network device (such as multi-AP) collaborative scenarios to quickly determine whether the first model on the first network device side and the second model on the first terminal device side match.

In one possible design, the first device determines whether a first model of the first device matches a second model of the second device based on at least one set of channel state information and channel state indication information, including: the first network device inputs the channel state indication information into the first model for processing to obtain a channel state information test value; the first network device determines whether the first model of the first network device matches the second model of the first terminal device based on whether an error between the channel state information test value and the channel state information is less than or equal to a first threshold.

In one possible design, the method also includes: when the first model of the first network device does not match the second model of the first terminal device, the first network device sends a second target model to the first terminal device, where the second target model is a second model that matches the first model of the first network device.

In the above design, when the first model of the first network device does not match the second model of the first terminal device, the first network device can send a second target model that matches the first model of the first network device to the first terminal device, which is conducive to quickly adapting the first model on the first network device side and the second model on the first terminal device side, thereby improving the efficiency of channel state information feedback.

In one possible design, the second model includes a downsampling layer and at least one feature extraction layer or module, and the downsampling layer is used to downsample the frequency domain dimension of the channel state information.

The above design can simplify the structure of the first terminal device side model and the data volume of the channel state indication information generated based on the terminal device side model, thereby reducing the computational complexity of the terminal device side model and facilitating the deployment of the terminal device side model.

In one possible design, the first device is a first terminal device, the second device is a first network device, the first model is used to process channel state information to obtain channel state indication information, and the second model is used to process channel state indication information to obtain channel state information; at least one set of channel state information and channel state indication information is determined by the first network device based on at least one channel state indication information sent by the first terminal device to the second network device, and the second network device is a network device that the first terminal device has accessed; or, at least one set of channel state information and channel state indication information is determined by the first network device based on at least one channel state indication information from the second terminal device, and the second terminal device is a terminal device that has accessed the first network device.

The above design can support a scenario in which the first terminal device roams from one network device to the first network device, and the first terminal device can quickly determine whether the model on the first network device side matches the model on the first terminal device side.

In one possible design, the first device determines whether a first model of the first device matches a second model of the second device based on at least one set of channel state information and channel state indication information, including: the first terminal device inputs the channel state information into the first model for processing to obtain a channel state indication information test value; the first terminal device determines whether the first model of the first terminal device matches the second model of the first network device based on whether an error between the channel state indication information test value and the channel state indication information is less than or equal to a second threshold.

In one possible design, the method also includes: when the first model of the first terminal device does not match the second model of the first network device, the first terminal device receives a first target model from the first network device, where the first target model is a first model that matches the second model; and the first terminal device updates the first model of the first terminal device according to the first target model.

In the above design, when the second model on the first network device side does not match the first model on the first terminal device side, the first terminal device can update the first model on the first terminal device side according to the first target model from the network device side that matches the second model, which is conducive to quickly adapting the second model on the first network device side to the first model on the first terminal device side, thereby improving the efficiency of channel state information feedback.

In one possible design, the first model includes a downsampling layer and at least one feature extraction layer or module, and the downsampling layer is used to downsample the frequency domain dimension of the channel state information.

The above design can simplify the structure of the first terminal device side model and the data volume of the channel state indication information generated based on the terminal device side model, thereby reducing the computational complexity of the terminal device side model and facilitating the deployment of the terminal device side model.

In one possible design, before the first terminal device inputs the channel state information into the first model for processing, the method also includes: the first terminal device fills the frequency domain dimension data of the channel state information with 0 when the channel state information is not an integer multiple of the reference value of the number of frequency domain dimension data; the first terminal device splits the channel state information into multiple sub-channel state information according to the reference value of the number of frequency domain dimension data.

The above design can achieve generalization of bandwidth and number of flows, that is, one model is suitable for inputs with different bandwidths and numbers of flows.

In a second aspect, an embodiment of the present application provides a communication device, which has the function of implementing the method in the first aspect, and the function can be implemented by hardware, or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above functions, such as an interface unit and a processing unit.

In one possible design, the device may be a chip or an integrated circuit.

In one possible design, the device includes a memory and a processor, the memory is used to store instructions executed by the processor, and when the instructions are executed by the processor, the device can execute the method of the first aspect above.

In one possible design, the apparatus may be a first device.

In a third aspect, an embodiment of the present application provides a communication device, which includes an interface circuit and a processor, and the processor and the interface circuit are coupled to each other. The processor is used to implement the method of the first aspect through a logic circuit or an execution instruction. The interface circuit is used to receive a signal from other communication devices outside the communication device and transmit it to the processor or send a signal from the processor to other communication devices outside the communication device. It is understandable that the interface circuit can be a transceiver or a transceiver or a transceiver or an input-output interface.

Optionally, the communication device may further include a memory for storing instructions executed by the processor or storing input data required by the processor to execute instructions or storing data generated after the processor executes instructions. The memory may be a physically independent unit or may be coupled to the processor, or the processor may include the memory.

In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a computer program or instructions are stored. When the computer program or instructions are executed by a processor, the method of the above-mentioned first aspect can be implemented.

In a fifth aspect, an embodiment of the present application further provides a computer program product, including a computer program or instructions, which, when executed by a processor, can implement the method of the first aspect described above.

In a sixth aspect, an embodiment of the present application further provides a chip, which is coupled to a memory and is used to read and execute programs or instructions stored in the memory to implement the method of the first aspect above.

The technical effects that can be achieved in the second to sixth aspects mentioned above can refer to the technical effects that can be achieved in the first aspect mentioned above, and will not be repeated here.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of a wireless network architecture and devices applicable to an embodiment of the present application;

FIG2 is a schematic diagram of a fully connected neural network provided in an embodiment of the present application;

FIG3 is a schematic diagram of a neuron calculating an output based on an input according to an embodiment of the present application;

FIG4 is a schematic diagram of an AI-based CSI feedback method provided in an embodiment of the present application;

FIG5 is a schematic diagram of a communication method according to an embodiment of the present application;

FIG6 is a schematic diagram of a terminal device and a network device providing an embodiment of the present application providing channel state feedback based on an AI model;

FIG7 is a second schematic diagram of a communication method provided in an embodiment of the present application;

FIG8 is a third schematic diagram of a communication method provided in an embodiment of the present application;

FIG9A is a fourth schematic diagram of a communication method provided in an embodiment of the present application;

FIG9B is a fifth schematic diagram of a communication method provided in an embodiment of the present application;

FIG10 is a schematic diagram of a neural network structure of an encoding model and a decoding model provided in an embodiment of the present application;

FIG11 is a schematic diagram of a coding model processing process according to an embodiment of the present application;

FIG12 is a second schematic diagram of a coding model processing process provided in an embodiment of the present application;

FIG13 is a third schematic diagram of the coding model processing process provided in an embodiment of the present application;

FIG14 is a schematic diagram of a signal-to-noise ratio-packet error rate curve provided in an embodiment of the present application;

FIG15 is a schematic diagram of a structure of a communication device according to an embodiment of the present application;

FIG. 16 is a second schematic diagram of the structure of the communication device provided in an embodiment of the present application.

DETAILED DESCRIPTION

The technical solution provided in the embodiment of the present application can be applied to wireless local area network (WLAN) systems, such as Wi-Fi, etc. For example, the embodiment of the present application can also be applied to the Institute of Electrical and Electronics Engineers (IEEE) 802.11 system standards, such as 802.11a/b/g protocol, 802.11n protocol, 802.11ac protocol, 802.11ax protocol, 802.11be protocol or next generation protocols, etc., which are not listed here one by one. Of course, the embodiments of the present application can also be applied to other types of communication systems, for example, cellular systems (including but not limited to: long term evolution (LTE) system, fifth-generation (5G) communication system, and new communication systems emerging in future communication development (such as 6G), etc.), Internet of Things (IoT) system, narrowband Internet of Things (NB-IoT) system, other short-range communication systems (including but not limited to: Bluetooth, ultra wide band (UWB)), etc.

In addition, the embodiments of the present application can be applied to scenarios where a node transmits data to one or more nodes. For example, uplink/downlink transmission of a single user, uplink/downlink transmission of multiple users, and device-to-device (D2D) transmission. Among them, any of the above nodes can be a communication device in a wireless communication system, that is, the method provided in the embodiments of the present application can be implemented by a communication device in a wireless communication system. For example, the communication device can be at least one of an access point (AP) or a station (STA). The communication devices (such as AP and STA) in the embodiments of the present application have certain artificial intelligence (AI) capabilities. For example, a neural network can be used for reasoning and decision-making.

In the embodiment of the present application, the WLAN system can provide high-speed and low-latency transmission. With the continuous evolution of WLAN application scenarios, the WLAN system will be applied to more scenarios or industries, such as the Internet of Things industry, the Internet of Vehicles industry or the banking industry, corporate offices, stadiums and exhibition halls, concert halls, hotel rooms, dormitories, wards, classrooms, supermarkets, squares, streets, production workshops and warehouses, etc. Of course, the devices (such as access points or sites) supporting WLAN communication or perception can be sensor nodes in smart cities (such as smart water meters, smart electric meters, smart air detection nodes), smart devices in smart homes (such as smart cameras, projectors, display screens, televisions, speakers, refrigerators, washing machines, etc.), nodes in the Internet of Things, entertainment terminals (such as wearable devices such as AR and VR), smart devices in smart offices (such as printers, projectors, loudspeakers, speakers, etc.), Internet of Vehicles devices in the Internet of Vehicles, infrastructure in daily life scenarios (such as vending machines, self-service navigation desks in supermarkets, self-service cashier devices, self-service ordering machines, etc.), and equipment in large sports and music venues, etc. For example, the access point and the station can be devices used in the Internet of Vehicles, IoT nodes and sensors in the Internet of Things (IoT), smart cameras in smart homes, smart remote controls, smart water meters and electricity meters, and sensors in smart cities. The specific forms of STA and AP are not limited in the embodiments of the present application, and are only exemplary.

While the embodiments of the present application are described using a network deployed on IEEE 802.11 as an example, various aspects of the present application can also be extended to other networks that adopt various standards or protocols, for example, high performance wireless LAN (HIPERLAN) (a wireless standard similar to the IEEE 802.11 standard, mainly used in Europe) and wide area network (WAN), wireless local area network (WLAN), personal area network (PAN) or other networks now known or developed later.

Referring to FIG. 1, a schematic diagram of a wireless network architecture and device applicable to an embodiment of the present application is shown. As shown in FIG. 1, in the network architecture, the device includes a wireless access point (AP) and three associated stations (STA), and each STA can communicate with the AP. The internal functional modules of the AP and the STA may include a central processing unit, a media access control (MAC), a transceiver, an antenna, and a neural network processing unit (NPU). The NPU includes a training module and an inference module. The trained neural network parameters (also referred to as model parameters) will be fed back to the inference module. The NPU can act on various other modules of the network node, including the central processing unit, MAC, transceiver, and antenna. The NPU can act on the decision-making tasks of each module, such as interacting with the transceiver, deciding the switch of the transceiver for energy saving, such as interacting with the antenna, controlling the direction of the antenna, such as interacting with the MAC, controlling channel access, channel selection, and spatial multiplexing decisions, etc.

FIG1 is only an example. Compared with the devices and the number of devices in FIG1, the system actually applied in the embodiment of the present application may include more or less. In addition, the embodiment of the present application can also be applied to communication between APs. For example, each AP can communicate with each other through a distributed system (DS). The embodiment of the present application can also be applied to communication between STAs.

In an embodiment of the present application, an access point may be an access point for a terminal device (such as a mobile phone) to enter a wired (or wireless) network. It may also be called a network device, which is mainly deployed in homes, buildings, and campuses. The typical coverage radius is tens to hundreds of meters. Of course, it can also be deployed outdoors. An access point is equivalent to a bridge connecting a wired network and a wireless network. Its main function is to connect various wireless network clients together and then connect the wireless network to the Ethernet. Specifically, an access point may be a terminal device (such as a mobile phone) or a network device (such as a router) with a Wi-Fi chip. An access point may be a device that supports the 802.11be standard. An access point may also be a device that supports 802.11ax, Devices of various wireless local area networks (WLAN) standards of the 802.11 family, such as 802.11ac, 802.11ad, 802.11ay, 802.11n, 802.11g, 802.11b, 802.11a and the next generation of 802.11be. The access point in this application may also be an extremely high throughput (EHT) AP, an ultra-high reliability (UHR) AP, and an access point applicable to a future generation of Wi-Fi standards.

The site can be a wireless communication chip, a wireless sensor or a wireless communication terminal, etc., and can also be called a user or terminal device. For example, the site can be a mobile phone that supports Wi-Fi communication function, a tablet computer that supports Wi-Fi communication function, a set-top box that supports Wi-Fi communication function, a smart TV that supports Wi-Fi communication function, a smart wearable device that supports Wi-Fi communication function, a vehicle-mounted communication device that supports Wi-Fi communication function, and a computer that supports Wi-Fi communication function, etc. Optionally, the site can support the 802.11be standard. The site can also support multiple wireless local area networks (WLAN) standards of the 802.11 family, such as 802.11ax, 802.11ac, 802.11n, 802.11g, 802.11b, 802.11a, and 802.11be next generation.

The station in this application can also be an extremely high throughput (EHT) STA, or a STA applicable to a future generation of Wi-Fi standards.

For example, access points and sites can be devices used in the Internet of Vehicles, IoT nodes and sensors in the IoT, smart cameras and smart remote controls in smart homes, smart water and electricity meters, and sensors in smart cities.

The AP and STA involved in the embodiments of the present application may be AP and STA applicable to the IEEE 802.11 system standard. AP is a device deployed in a wireless communication network to provide wireless communication functions for its associated STA. The AP can be used as the hub of the communication system, and is usually a network-side product that supports the MAC and physical layer (physical, PHY) of the 802.11 system standard. For example, it may be a base station, a router, a gateway, a repeater, a communication server, a switch or a bridge and other communication equipment, wherein the base station may include various forms of macro base stations, micro base stations, relay stations, etc. Here, for the sake of convenience of description, the above-mentioned devices are collectively referred to as AP. STA is usually a terminal product that supports the MAC and PHY of the 802.11 system standard, such as a mobile phone, a laptop computer, etc.

In order to better understand the embodiments of the present application, the names and technical features involved in the embodiments of the present application are explained below. It should be noted that these explanations are intended to make the embodiments of the present application easier to understand and should not be regarded as limiting the scope of protection claimed by the present application.

1) Neural network (NN) is a machine learning technology that simulates the human brain neural network in order to achieve artificial intelligence-like. A neural network can include an input layer, an intermediate layer (also called a hidden layer), and an output layer. Taking the simplest neural network as an example, its internal structure and implementation are explained. See Figure 2, which is a schematic diagram of a fully connected neural network containing three layers, where each circle represents a neuron. As shown in Figure 2, the neural network includes three layers, namely the input layer, the hidden layer, and the output layer, where the input layer has three neurons, the hidden layer has four neurons, and the output layer has two neurons, and each layer of neurons is fully connected to the neurons in the next layer. Each line between neurons corresponds to a weight, which can be updated through training. Each neuron in the hidden layer and the output layer can also correspond to a bias, which can be updated through training. Updating a neural network means updating these weights and biases. After determining the structure of the neural network, that is, the number of neurons contained in each layer and how the output of the previous neuron is input to the subsequent neuron (that is, the connection relationship between neurons), and adding the parameters of the neural network, that is, the weights and biases, all the information of the neural network can be determined.

As shown in Figure 2, each neuron may have multiple input connections, and each neuron can calculate output based on the input. Figure 3 is a schematic diagram of a neuron calculating output based on input. As shown in Figure 3, a neuron contains 3 inputs, 1 output, and 2 calculation functions. The calculation formula for the output can be expressed as:

Output = activation function (input 1 * weight 1 + input 2 * weight 2 + input 3 * weight 3 + bias) Formula 1

The symbol “*” represents the mathematical operation “multiplication” or “times”, and the same applies to the symbols shown below.

In addition, each neuron may also have multiple output connections, and the output of one neuron can be used as the input of the next neuron. It should be understood that the input layer only has output connections, and each neuron in the input layer is the value of the input neural network, and the output value of each neuron is directly used as the input of all output connections. The output layer only has input connections, and the output is calculated using the calculation method of formula 1 above. Optionally, the output layer can have no activation function calculation, then the above formula 1 can be transformed into: Output = Input 1 * Weight 1 + Input 2 * Weight 2 + Input 3 * Weight 3 + Bias.

For example, a k-layer neural network can be expressed as:

y＝fk(fk-1(…(f1(w1*x+b1)))                      Formula 2

Among them, x represents the input of the neural network, y represents the output of the neural network, wi represents the weight of the i-th layer of the neural network, bi represents the bias of the i-th layer of the neural network, and fi represents the activation function of the i-th layer of the neural network. i＝1,2,…,k.

The CSI feedback method currently used in the standard is that the network device (such as AP) first sends a null data packet announcement (NDPA) and then sends a null data packet (NDP). The terminal device performs channel estimation on the NDP sent by the network device based on the NDPA to obtain the channel state information H, performs singular value decomposition (SVD) on H to obtain the precoding matrix V, performs a Givens rotation operation on V to convert it into two types of angles, φ and ψ, and then quantizes and feeds back these angles to the network device. The network device recovers the CSI (such as the precoding matrix) from the received angles. Used to perform operations such as beamforming on data or information sent to a terminal device.

In wireless networks, AI's advantages are reflected in four aspects: 1. Solving complex network problems without mathematical models; 2. Solving wireless network management problems with large search spaces; 3. Global optimization at the cross-layer and cross-node network level; 4. Actively optimizing wireless network parameters through AI's predictive capabilities. Wireless resource allocation is crucial to the performance of wireless networks, and wireless resource allocation includes channel access, channel allocation, rate configuration, and power configuration. Therefore, an AI-based coding model can be used to encode and compress CSI.

As shown in Figure 4, a schematic diagram of an AI-based CSI feedback method provided in an embodiment of the present application includes a training phase and an inference phase. In the training phase, the terminal device (taking STA as an example in Figure 4) performs CSI feedback normally, such as performing channel estimation, SVD and other processing, obtaining a precoding matrix V or two types of angles of φ and ψ, and sending it to the network device (taking AP as an example in Figure 4). The AP uses CSI to train an autoencoder (including an encoder, a decoder, and a dictionary), wherein the output of the encoder is used as the input of the decoder, the input of the encoder is CSI, and the expected output of the decoder is CSI. The dictionary can be used to quantize the result of the encoder output and dequantize the quantized result. After the network device completes the training, it sends the encoder and the dictionary (if any) to the terminal device. In the inference phase, after the terminal device estimates the channel, it can input the CSI (such as the precoding matrix) V into the encoder to obtain the output, and after quantization (optional, it can be based on the dictionary or use simple uniform quantization), the channel state indication information (such as index information) is fed back to the network device, and the network device uses the decoder to recover the CSI It is used to perform operations such as beamforming on data or information sent to a terminal device. It should be understood that the encoder and decoder may also be referred to as encoding models and decoding models, and the dictionary may also be referred to as a codebook.

However, the decoding models used by different network devices (such as APs) may be different, and the encoding models used by different terminal devices may also be different. When a terminal device switches from one network device to another, how to determine whether the encoding model of the terminal device matches the decoding model of another network device is related to the communication quality between the terminal device and the network device.

Based on this, the present application provides a communication method and apparatus to support fast matching of the encoding model of the terminal device and the decoding model of the network device. The embodiments of the present application will be described in detail below in conjunction with the accompanying drawings, wherein the dotted lines in the drawings represent optional steps or components.

In addition, it should be understood that the ordinal numbers such as "first" and "second" mentioned in the embodiments of the present application are used to distinguish multiple objects, and are not used to limit the size, content, order, timing, priority or importance of multiple objects. For example, a first network device and a second network device do not mean that the priorities or importance of the two network devices are different.

In the embodiments of the present application, the number of nouns, unless otherwise specified, means "singular noun or plural noun", that is, "one or more". "At least one" means one or more, and "plural" means two or more. "And/or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and/or B can mean: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The character "/" generally indicates that the previous and next associated objects are in an "or" relationship. For example, A/B means: A or B. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c means: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, c can be single or multiple.

FIG5 is a schematic diagram of a communication method provided in an embodiment of the present application, which can be executed by a first device (or a second device), or by a component of the first device (or a second device) (such as a processor, a chip, or a chip system, etc.), or a device used in conjunction with the first device (or a second device), etc. The method includes:

S501: The first device obtains at least one set of channel state information and channel state indication information corresponding to the second device.

S502: The first device determines whether a first model of the first device matches a second model of the second device according to at least one set of channel state information and channel state indication information.

S503: The first device sends first indication information to the second device, where the first indication information is used to indicate whether a first model of the first device matches a second model of the second device.

In the embodiment of the present application, the first device may be a network device (such as an AP, etc.), the second device may be a terminal device (such as a STA, etc.), or the first device may be a terminal device (such as a STA, etc.), the second device may be a network device (such as an AP, etc.). The terminal device and the network device may provide feedback of channel state information based on an AI model.

Referring to FIG. 6 , a schematic diagram of a terminal device and a network device performing channel state feedback based on an AI model is shown, wherein the terminal device includes The terminal device includes a coding model for processing the channel state information to obtain the channel state indication information, and the network device includes a decoding model for processing the channel state indication information to obtain the channel state information. After the terminal device obtains the channel state information V through operations such as channel estimation, the channel state information V can be input into the coding model of the terminal device for processing to obtain the channel state indication information m output, and the obtained channel state indication information m output can be sent to the network device. After the network device receives the channel state indication information m from the terminal device, the channel state indication information m can be input into the decoding model of the network device for processing to obtain the channel state information output by the decoding model. Channel status indication information It can be used by network equipment to perform operations such as beamforming on signals sent to terminal devices.

In some implementations, the terminal device may further quantize the channel state indication information m before sending it (e.g., quantize it according to a preconfigured dictionary); the network device may dequantize the received channel state indication information m before inputting the channel state indication information m into the decoding model for processing (e.g., quantize it according to a preconfigured dictionary), etc., so as to improve the reliability of the transmission of the channel state indication information m between the terminal device and the network device and reduce the signaling overhead of the transmission of the channel state indication information m.

However, when the terminal device switches (or roams) from one network device to another network device, such as switching (or roaming) from the second network device (such as AP2) to the first network device (such as AP1), AP1 and AP2 may use different decoding models, such as different decoding model neural network structures, and/or different decoding model neural network parameters. Therefore, when the terminal device switches (or roams) from one network device to another network device, it is necessary to determine whether the coding model of the terminal device matches the decoding model of the other network device to which it switches (or roams), so as to determine whether the coding model of the terminal device is adapted to the decoding model of the other network device to which it switches (or roams), and whether the coding model of the terminal needs to be updated on the other network device to which it switches (or roams).

In an embodiment of the present application, after a terminal device switches (or roams) from one network device to another network device, the terminal device may obtain at least one set of channel state information and channel state indication information corresponding to the network device to which it switches (or roams), and the terminal device tests the coding model of the terminal device by switching (or roaming) to at least one set of channel state information and channel state indication information corresponding to the network device to which it switches (or roams), so as to determine whether the coding model of the terminal device matches the decoding model of the network device to which it switches (or roams). The network device to which it switches (or roams) may also obtain at least one set of channel state information and channel state indication information corresponding to the terminal device, and the network device tests the decoding model of the network device according to at least one set of channel state information and channel state indication information corresponding to the terminal device, so as to determine whether the coding model of the terminal device matches the decoding model of the network device.

It should be understood that, in the case where the terminal device determines whether the coding model of the terminal device matches the decoding model of the network device switched (or roamed), the first device is the terminal device, the second device is the network device switched (or roamed) to, the first model is a coding model for processing the channel state information to obtain the channel state indication information, and the second model is a decoding model for processing the channel state indication information to obtain the channel state information. In the case where the terminal device switches (or roams) to the network device to determine whether the coding model of the terminal device matches the decoding model of the network device switched (or roamed), the first device is the network device switched (or roamed) to, the second device is the terminal device, the first model is a decoding model for processing the channel state indication information to obtain the channel state information, and the second model is a coding model for processing the channel state information to obtain the channel state indication information.

Below, taking the example of the first terminal device switching (or roaming) from the second network device to the first network device, the communication method of the present application is described in detail in combination with the situations in which the first terminal device determines whether the encoding model of the first terminal device matches the decoding model of the first network device in different scenarios, or the first network device determines whether the encoding model of the first terminal device matches the decoding model of the first network device.

Scenario 1: A first network device (i.e., the first device) obtains at least one set of channel state information input and channel state indication information output of a first terminal device (i.e., the second device) from the first terminal device to determine whether a decoding model 1 (i.e., the first model) of the first network device matches an encoding model 1 (i.e., the second model) of the first terminal device.

As shown in FIG. 7 , a communication method applicable to scenario 1 is provided in an embodiment of the present application, wherein AP1 in FIG. 7 represents a first network device, STA1 represents a first terminal device, and AP2 represents a second network device, wherein the second network device is a network device that the first terminal device has accessed, such as a network device that the first terminal device accessed before switching (or roaming) to the first network device. The method includes:

S701: The first network device sends a first request message to the first terminal device, and correspondingly, the first terminal device receives the first request message.

The first request message is used to request at least one set of channel state information V input and channel state indication information m output of the first terminal device.

S702: The first terminal device sends at least one set of channel state information V input and channel state indication information m output to the first network device, and correspondingly, the first network device receives at least one set of channel state information V input and channel state indication information m output.

In a possible implementation, when the first terminal device accesses (or associates) with the second network device, the first terminal device may store at least one set of input and output when performing channel measurement with the second network device, that is, at least one set of channel state information V input and channel state indication information m output of coding model 1 (i.e., the second model). The first terminal device stores at least one set of channel state information V input and channel state indication information m output, which may be performed autonomously by the first terminal device when accessing the second network device, or may be triggered by the second network device (or the first network device) through signaling, and this application does not limit this.

As an example: when the first terminal device accesses the second network device, it can perform channel estimation and SVD processing on the NDP from the second network device to obtain channel state information V, input the channel state information V into its own coding model 1 (i.e., the second model) for processing, obtain channel state indication information m output, and store the channel state information V input and the channel state indication information m output.

After the first terminal device switches (or roams) to the first network device, the first network device may send a first request message to the first terminal device, requesting at least one set of channel state information V input and channel state indication information m output of the first terminal device. After receiving the first request message, the first terminal device may send the stored at least one set of channel state information V input and channel state indication information m output to the first network device, that is, report the stored at least one set of {V, m} to the first network device.

In some implementations, the first network device may send the first request message to the first terminal device by sending a request frame or the like. The present application does not limit the manner in which the first network device sends the first request message to the first terminal device.

S703: The first network device inputs the channel state indication information m into the decoding model 1 (i.e., the first model) for processing to obtain a channel state information test value V’.

S704: The first network device determines whether the first model of the first network device matches the second model of the first terminal device based on whether the error between the channel state information test value V’ and the channel state information V is less than or equal to the first threshold.

After the first network device obtains at least one set of channel state information V input and channel state indication information m output of the first terminal device, the obtained channel state indication information m can be used as the input of the decoding model 1 (i.e., the first model) of the first network device to obtain the channel state information test value V’ output by the decoding model 1, and calculate the error between the channel state information test value V’ and the obtained channel state information V of the first terminal device.

In a possible implementation, the error between the channel state information test value V' and the acquired first terminal device channel state information V may be a cosine similarity, such as It can also be mean square error (mese), such as mse=(V-V') ^2, etc. Of course, it can also be a combination of cosine similarity and mean square error, such as cosine similarity+β*mse, etc., where β is a preset proportionality coefficient.

It should be understood that the above-mentioned methods of calculating cosine similarity, calculating mean square error, and combining cosine similarity and mean square error are only examples of calculating cosine similarity, calculating mean square error, and combining cosine similarity and mean square error. The present application does not limit the methods of calculating cosine similarity, calculating mean square error, and combining cosine similarity and mean square error. In addition, the present application is not limited to calculating errors by using cosine similarity, mean square error, etc. For example, the error can also be determined by calculating the difference, calculating the mean square error, etc.

When the error between the channel state information test value V’ and the acquired channel state information V of the first terminal device is less than or equal to the first threshold, the first network device can determine that the decoding model 1 (i.e., the first model) of the first network device matches the encoding model 1 (i.e., the second model) of the first terminal device; when the error is greater than the first threshold, it can be determined that the decoding model 1 (i.e., the first model) of the first network device does not match the encoding model 1 (i.e., the second model) of the first terminal device.

S705: The first network device sends first indication information to the first terminal device, and correspondingly, the first terminal device receives the first indication information.

The first indication information indicates whether the decoding model 1 (ie, the first model) of the first network device matches the encoding model 1 (ie, the second model) of the first terminal device.

In some implementations, the first network device may also send first indication information to the first terminal device only when the decoding model 1 (i.e., the first model) of the first network device matches the encoding model 1 (i.e., the second model) of the first terminal device, and the first indication information is used to indicate that the decoding model 1 (i.e., the first model) of the first network device matches the encoding model 1 (i.e., the second model) of the first terminal device.

When the decoding model 1 (ie, the first model) of the first network device matches the encoding model 1 (ie, the second model) of the first terminal device, the first terminal device may continue to use the current encoding model 1 (ie, the second model) to process the channel state information.

When the decoding model 1 (i.e., the first model) of the first network device does not match the encoding model 1 (i.e., the second model) of the first terminal device, the first network device may also send a target encoding model (i.e., the second target model) that matches its own decoding model 1 (i.e., the first model) to the first terminal device for updating the encoding model 1 (i.e., the second model) on the first terminal device side.

In a possible implementation, if the terminal devices all use a model with the same neural network structure, that is, the target coding model (i.e., the second target model) has the same neural network structure as the coding model 1 (i.e., the second model) of the first terminal device, the first network device may only send the neural network parameters (such as weights and/or biases) of the target coding model (i.e., the second target model) to the first terminal device for updating the neural network parameters of the coding model 1 (i.e., the second model) on the first terminal device side. If the target coding model (i.e., the second target model) has a different neural network structure than the coding model 1 (i.e., the second model) of the first terminal device, the first network device may send the neural network structure and neural network parameters of the target coding model (i.e., the second target model) to the first terminal device for updating the neural network structure and neural network parameters of the coding model 1 (i.e., the second model) of the first terminal device.

Scenario 2: The first network device (i.e., the first device) obtains at least one set of channel state indication information input and channel state information output of the second network device corresponding to the first terminal device (i.e., the second device), and determines whether the decoding model 1 (i.e., the first model) of the first network device matches the encoding model 1 (i.e., the second model) of the first terminal device. The second network device is a network device that the first terminal device has accessed, such as a network device that the first terminal device accessed before switching (or roaming) to the first network device.

As shown in FIG8 , a communication method applicable to scenario 2 is provided in an embodiment of the present application, wherein AP1 in FIG8 represents a first network device, STA1 represents a first terminal device, and AP2 represents a second network device, which can be applicable to the situation where multiple APs collaborate in an enterprise network. The method includes:

S801: The first network device sends a second request message to the second network device, and correspondingly, the second network device receives the second request message.

The second request message is used to request the second network device to input at least one set of channel state indication information m and channel state information corresponding to the first terminal device. Output.

S802: The second network device sends at least one set of channel state indication information input m and channel state information corresponding to the first terminal device to the first network device. Output.

In a possible implementation, when the first terminal device accesses (or associates) with the second network device, the second network device may store at least one set of channel state indication information m input and channel state information corresponding to the first terminal device of its own decoding model 2. Output.

As an example: when the first terminal device accesses the second network device, it can perform channel estimation and SVD processing on the NDP from the second network device to obtain channel state information V, input the channel state information V into its own coding model 1 (i.e., the second model) for processing, obtain channel state indication information m output, and send it to the second network device. The second network device uses the channel state indication information m as the input of its own decoding model 2, processes the channel state indication information m, and obtains the channel state information output by the decoding model 2. The second network device may store at least one set of channel state indication information m input and channel state information of the decoding model 2. Output.

After the first terminal device switches (or roams) to the first network device, the first network device may send a second request message to the second network device, requesting the second network device to input at least one set of channel state indication information m and channel state information corresponding to the first terminal device. After receiving the second request message from the first network device, the second network device may send the stored at least one set of channel state indication information m input and channel state information corresponding to the first terminal device to the second network device. Output.

S803: The first network device inputs the channel state indication information m into the decoding model 1 (i.e., the first model) for processing to obtain a channel state information test value V’.

S804: The first network device compares the channel state information test value V' with the channel state information is less than or equal to a first threshold, determining whether a decoding model 1 (ie, a first model) of the first network device matches an encoding model 1 (ie, a second model) of the first terminal device.

The first network device obtains at least one set of channel state indication information m input and channel state information corresponding to the first terminal device from the second network device. After the output, the obtained channel state indication information m can be used as the input of the decoding model 1 (i.e., the first model) of the first network device to obtain the channel state information test value V' output by the decoding model 1 (i.e., the first model), and the channel state information test value V' and the obtained channel state information are calculated. of error.

In a possible implementation, the channel state information test value V' is related to the acquired channel state information The error can be cosine similarity, such as It can also be the mean square error (mese), such as Etc. Of course, it can also be a combination of cosine similarity and mean square error, such as cosine similarity+β*mse, where β is a preset proportional coefficient.

The first network device compares the channel state information test value V' with the acquired channel state information When the error is less than or equal to the first threshold, it can be determined that the decoding model 1 (i.e., the first model) of the first network device matches the encoding model 1 (i.e., the second model) of the first terminal device; when the error is greater than the first threshold, it can be determined that the decoding model 1 (i.e., the first model) of the first network device does not match the encoding model 1 (i.e., the second model) of the first terminal device.

S805: The first network device sends first indication information to the first terminal device, and correspondingly, the first terminal device receives the first indication information.

The first indication information indicates whether the decoding model 1 (ie, the first model) of the first network device matches the encoding model 1 (ie, the second model) of the first terminal device.

In some implementations, the first network device may also send first indication information to the first terminal device only when the decoding model 1 (i.e., the first model) 1 of the first network device matches the encoding model 1 (i.e., the second model) of the first terminal device, and the first indication information is used to indicate that the decoding model 1 (i.e., the first model) 1 of the first network device matches the encoding model 1 (i.e., the second model) of the first terminal device.

When the decoding model 1 (ie, the first model) of the first network device matches the encoding model 1 (ie, the second model) of the first terminal device, the first terminal device may continue to use the current encoding model 1 (ie, the second model) to process the channel state information.

When the decoding model 1 (i.e., the first model) of the first network device does not match the encoding model 1 (i.e., the second model) of the first terminal device, the first network device can also send a target encoding model (i.e., the second target model) that matches its own decoding model 1 (i.e., the first model) to the first terminal device, for updating the encoding model 1 (i.e., the second model) on the first terminal device side.

In a possible implementation, if the terminal devices all use a model with the same neural network structure, that is, the target coding model (i.e., the second target model) has the same neural network structure as the coding model 1 (i.e., the second model) of the first terminal device, the first network device may only send the neural network parameters (such as weights and/or biases) of the target coding model (i.e., the second target model) to the first terminal device for updating the neural network parameters of the coding model 1 (i.e., the second model) on the first terminal device side. If the target coding model (i.e., the second target model) has a different neural network structure than the coding model 1 (i.e., the second model) of the first terminal device, the first network device may send the neural network structure and neural network parameters of the target coding model (i.e., the second target model) to the first terminal device for updating the neural network structure and neural network parameters of the coding model 1 (i.e., the second model) on the first terminal device side.

Scenario three: The first terminal device (i.e., the first device) obtains at least one set of channel state information and channel state indication information from the first network device (i.e., the second device) to determine whether the encoding model 1 (i.e., the first model) of the first terminal device matches the decoding model 1 (i.e., the second model) of the first network device.

As shown in FIG9A , a communication method applicable to scenario 3 is provided in an embodiment of the present application, wherein AP1 in FIG9A represents a first network device, STA1 represents a first terminal device, and AP2 represents a second network device, which can be applicable to the situation where multiple APs collaborate in an enterprise network. The second network device is a network device that the first terminal device has accessed, such as a network device that the first terminal device accessed before switching (or roaming) to the first network device. The method includes:

S901: The first terminal device sends a third request message to the first network device, and the first network device receives the third request message accordingly.

The third request message is used to request at least one set of channel state indication information m input and channel state information of the first network device. Output.

S902: The first terminal device receives at least one set of channel state indication information m input and channel state information from the first network device. Output.

In an embodiment of the present application, when the first terminal device accesses (or associates) with the second network device, the second network device may store at least one channel state indication information m reported by the first terminal device. After the first terminal device switches (or roams) to the first network device, the first network device may request the at least one channel state indication information m from the second network device.

After the first network device obtains at least one channel state indication information m from the second network device, the channel state indication information m can be input into its own decoding model 1 (i.e., the second model) for processing to obtain the channel state information Output and store channel state indication information m input and channel state information Output

After the first network device receives the third request message from the first network device, the channel state indication information m corresponding to the first terminal device can be input and the channel state information The output is sent to the first terminal device.

S903: The first terminal device transmits the channel state information The signal is input to coding model 1 (ie, the first model) for processing to obtain a channel state indication information test value m'.

S904: The first terminal device determines whether the encoding model 1 (i.e., the first model) of the first terminal device matches the decoding model 1 (i.e., the second model) 1 of the first network device based on whether the error between the channel state indication information test value m’ and the channel state indication information m is less than or equal to the second threshold.

The first terminal device obtains at least one set of channel state indication information m input and channel state information corresponding to the first terminal device from the first network device. After output, the acquired channel status information can be As the input of the coding model 1 (ie, the first model) of the first terminal device, the channel state indication information test value m' output by the coding model 1 (ie, the first model) of the first terminal device is obtained, and the error between the channel state indication information test value m' and the obtained channel state indication information m is calculated.

In a possible implementation, the error between the channel state indication information test value m' and the acquired channel state indication information m may be a cosine similarity, such as It can also be mean square error (mese), such as mse=(m-m') ^2, etc. Of course, it can also be a combination of cosine similarity and mean square error, such as cosine similarity+β*mse, etc., where β is a preset proportional coefficient.

When the error between the channel state indication information test value m' and the acquired channel state indication information m of the first terminal device is less than or equal to the second threshold, it can be determined that the encoding model 1 (i.e., the first model) of the first terminal device matches the decoding model 1 (i.e., the second model) of the first network device; when the error is greater than the second threshold, it can be determined that the encoding model 1 (i.e., the first model) of the first terminal device does not match the decoding model 1 (i.e., the second model) obtained by the first network device.

S905: The first terminal device sends first indication information to the first network device, and correspondingly, the first network device receives the first indication information.

The first indication information indicates whether the encoding model 1 (ie, the first model) of the first terminal device matches the decoding model 1 (ie, the second model) of the first network device.

In some implementations, the first terminal device may also send first indication information to the first network device only when the encoding model 1 (i.e., the first model) of the first terminal device matches the decoding model 1 (i.e., the second model) of the first network device, and the first indication information is used to indicate that the encoding model 1 (i.e., the first model) of the first terminal device matches the decoding model 1 (i.e., the second model) of the first network device. When the encoding model 1 (i.e., the first model) of the first terminal device matches the decoding model 1 (i.e., the second model) of the first network device, the first terminal device may continue to use the current encoding model 1 (i.e., the first model) to process the channel state information.

When the first indication information received indicates that the encoding model 1 (i.e., the first model) of the first terminal device does not match the decoding model 1 (i.e., the second model) of the first network device, or when the first indication information from the first terminal device is not received, the first network device can determine that the encoding model 1 (i.e., the first model) of the first terminal device does not match the decoding model 1 (i.e., the second model) of the first network device. The first network device can also send a target encoding model (i.e., the first target model) that matches its own decoding model 1 (i.e., the second model) to the first terminal device for updating the encoding model 1 (i.e., the first model) on the first terminal device side.

As shown in Figure 9B, another communication method suitable for scenario three is provided for an embodiment of the present application. Different from the method in Figure 9A, in which the first network device obtains at least one set of channel state information and channel state indication information through the second network device that the first terminal device has previously accessed, in Figure 9B, the first network device can also obtain at least one set of channel state information and channel state indication information through the second terminal device that has accessed (or has accessed) the first network device.

As an example: when the second terminal device accesses (or associates) with the first network device, the first network device can store at least one set of channel state indication information m input and channel state information corresponding to the second terminal device of its own decoding model 1 (i.e., the second model). After the first terminal device is connected to (or associated with) the first network device, the first network device can store at least one set of channel state indication information m input and channel state information The output is sent to the first terminal device for determining whether the encoding model 1 (ie, the first model) of the first terminal device matches the decoding model 1 (ie, the second model) of the first network device.

In addition, in order to further reduce the problems of complex coding model structure, large number of parameters, high reasoning complexity and high power consumption on the terminal device side, the embodiment of the present application also provides a standardized and minimalist coding model structure, which coding model may include a downsampling layer and at least one feature extraction layer or module, wherein the downsampling layer is used to downsample the frequency domain dimension of the channel state information, in order to reduce the complexity of the coding model structure, as well as the reasoning complexity and power consumption of the coding model on the terminal device side.

10 is a schematic diagram of a neural network structure of an encoding model and a decoding model provided in an embodiment of the present application. On the terminal device side, the terminal device can pre-process the channel state information and then input it into the encoding model for processing, wherein the encoding model may include a down sampling layer, a fully connected layer (FC), a convolutional layer with a convolution kernel size of 1*1 ((Conv 1*1)) and a 4-layer structure of FC, wherein the down sampling layer can be used to downsample the frequency domain dimension of the channel state information, FC, convolution layer and FC can be used to extract features from channel state information to obtain channel state indication information. On the network device side, the channel state indication information output by the coding model on the terminal device side can be processed to restore the channel state information, wherein the decoding model can also adopt a 4-layer structure, such as FC, a convolution layer with a convolution kernel size of 1*1, a transformer layer and FC for restoring the channel state information. This application does not limit the specific neural network structure of the decoding model adopted on the network device side. For the coding model, in some implementations, the coding model of each terminal device can adopt the same neural network structure to improve the applicability of the coding model on the terminal device side.

In some implementations, the terminal device side may also quantize the channel state indication information output by the coding model, and the network device may Before inputting the channel state indication information into the decoding model for processing, the received channel state indication information can also be dequantized to improve the reliability of the transmission of the channel state indication information between the terminal device and the network device and reduce the signaling overhead of the channel state indication information transmission.

In one possible implementation, the above-mentioned preprocessing of the channel state information may include filling the frequency domain dimension data of the channel state information with 0 when the channel state information is not an integer multiple of the reference value of the number of frequency domain dimension data; and splitting the channel state information into multiple sub-channel state information according to the reference value of the number of frequency domain dimension data to suit inputs of different bandwidths (such as 40/80/160MHz).

Taking the reference value of the number of frequency domain dimension data as 64 as an example, the original channel state information can be represented by a complex matrix of dimension (Ndata, Nsc, Ntx, Nss), where Ndata represents the number of data, Nsc represents the number of subcarriers, Ntx represents the number of transmitting antennas, and Nss represents the number of spatial streams. The terminal device can first reshape each spatial stream data to the Ndata dimension, and splice the real and imaginary parts of the reshaped data to the Ntx dimension, then the data becomes (Ndata*Nss, Nsc, Ntx*2); secondly, if the Nsc dimension is not a multiple of 64, it is necessary to fill in the dimension. 0, the data becomes (Ndata*Nss, N*64, Ntx*2), Indicates rounding up. After preprocessing, the data is input into the coding model N times. Using the above preprocessing method, taking 80MHz data as an example, the original data Nsc=250, then it is necessary to fill 6 zeros in this dimension, and then input the coding model 4 times in units of 64, and then feedback is obtained after the coding model output. The method of zero filling can be 3 zeros on the left and right, or 6 zeros at the end, etc., and this application does not limit this.

Take the channel state information (B, 64, Ntx*2) of the minimum bandwidth of Wi-Fi 20MHz as an example, where B represents the number of data (Ndata)*number of spatial streams (Nss), 64 represents the number of 64 subcarriers, Ntx represents the number of transmitting antennas, and 2 represents the real and imaginary parts of the data. Referring to the coding model processing process shown in Figure 11, the down sampling layer can first downsample the subcarrier dimension of the channel state information (B, 64, Ntx*2) by 2 times, and the subcarrier dimension is reduced from 64 to 32, (B, 64, Ntx*2) becomes (B, 32, Ntx*2); after another layer of FC, the layer can have 256 neurons, so the last dimension of (B, 32, Ntx*2) becomes 256, (B , 32, Ntx*2) becomes (B, 32, 256); then it passes through a convolution layer with a convolution kernel size of 1*1 ((Conv 1*1)), the number of convolution kernels is 20, so the dimension of 32 is reduced to 20, (B, 32, 256) becomes (B, 20, 256); finally it passes through a fully connected layer with 32 neurons, so the last dimension of (B, 20, 256) becomes 32, and (B, 20, 256) becomes (B, 20, 32).

In some implementations, at least one feature extraction layer or module of the coding model may also include 4 depthwise (DW) separable convolution modules (DWBlock), where the DW separable convolution module may also be referred to as a DW module. Take the channel state information (B = 1, C = 2, H = 250, W = 16) input to the coding model for processing as an example, where B represents the number of data (Ndata), C represents the real and imaginary parts of the data, H represents the number of subcarriers, and W represents the number of spatial dimensions, which can be the number of antennas (Ntx) * the number of spatial streams (Nss). Referring to the schematic diagram of the processing process of the depthwise separable convolution coding model shown in Figure 12, the input of the coding model is ① (B = 1, C = 2, H = 250, W = 16). The first layer of the coding model (down-sampling layer) will downsample the third dimension (250) of the input, and the downsampling rate is 4, so the output is reduced to ② (1, 2, 63, 16), which can reduce the amount of subsequent calculations. Then, four depthwise separable convolution modules (DWBlocks) are used for feature extraction. Each module consists of a depth-wise convolution module and a point-wise convolution module. The first DWBlock will double the number of channels of the vector, and the 2nd to 4th DWBlocks will not change the dimension of the vector. After obtaining the feature vector ⑤(1,4,63,16) extracted by DWBlock, two 2-dimensional (2d) convolutions can be used to reduce its dimension to obtain ⑦(1,32,2,10). Finally, vector ⑦ is reshaped and the last three dimensions are merged to obtain ⑧(1,640). In the subsequent uniform quantization, each value in ⑧ can be represented by 2 bits, so the number of compressed bits of this coding model = 1*640*2 = 1280 bits.

In some implementations, at least one feature extraction layer or module of the coding model may also include 4 convolution modules (Blocks), each of which may be composed of a residual module (ResBlock) and a convolution module (ConvBlock). Take the channel state information (B=1, C=2, H=250, W=16) input to the coding model for processing as an example, where B represents the number of data (Ndata), C represents the real and imaginary parts of the data, H represents the number of subcarriers, and W represents the number of spatial dimensions, which can be the number of antennas (Ntx)*the number of spatial streams (Nss). Referring to the schematic diagram of the coding model processing process using convolution and residual networks shown in FIG13, the input of the coding model is ① (B=1, C=2, H=250, W=16). The first layer of the encoding model (down-sampling layer) downsamples the third dimension (250) of the input with a downsampling rate of 2, so the output is reduced to ②(1,2,125,16). This step can reduce the amount of subsequent calculations. Then, 4 convolutional modules (Blocks) can be used for feature extraction. Each module consists of a residual module (ResBlock) and a convolutional module (ConvBlock). The 1st to 3rd blocks do not change the vector dimension, and the 4th block doubles the number of channels of the vector. After obtaining the feature vector ⑤(1,4,125,16) extracted by Block, the second and fourth dimensions can be merged to reshape the vector into (1,125,4*16=64), and then divided into five parts along the second dimension (thereby reducing the number of parameters and calculations of the subsequent convolutional network), and then spliced in the first dimension to obtain (1*5=5,125/5=25,64), and then two 1-dimensional (1d) convolutions are used to reduce the dimension to obtain ⑦(5,32,4). Finally, the vector ⑦ is reshaped and the four dimensions are merged to obtain ⑧(1,640). In the subsequent uniform quantization, each value in ⑧ can be represented by 2 bits, so the number of compressed bits of this coding model = 1*640*2=1280 bits.

Referring to the performance indicators shown in Table 1, Table 1 shows the performance indicators of the number of feedback bits (bits) and effective throughput (goodput) of the standard scheme for directly feeding back the precoding matrix V or two types of angles of φ and ψ; the performance indicators of the number of feedback bits, effective throughput, the number of coding model parameters, and the coding model reasoning complexity that the prior art does not adopt the above-mentioned coding model of this application; and the performance indicators of the number of feedback bits, actual throughput, the number of coding model parameters, and the coding model reasoning complexity that this application proposes to adopt the above-mentioned coding model (such as the coding model shown in Figure 11). Referring to Table 1, it can be seen that compared with the Standard scheme, this scheme can reduce the feedback overhead from 32500 bits to 1280 bits and increase the effective throughput from 5.07Mbps to 16.00Mbps. Compared with the prior art solution, this application adopts a minimalist coding model structure, which can reduce the number of parameters of the coding model from 5M to 4.7k and the computational complexity of the coding model from 2.2G multiplication and accumulation operations to 1.2M multiplication and accumulation operations while ensuring the same performance and feedback overhead.

Table 1

In addition, referring to the signal to interference plus noise ratio (SNR) - packet error rate (PER) curve diagram of Standard, prior art and the proposed solution as shown in Figure 14, it can be seen that at 80MHz, 8 transmitting antennas, 2 receiving antennas (8x2), and 2 spatial streams, the performance of the proposed solution in SNR and PER does not decrease compared with Standard and prior art.

The method provided by the embodiment of the present application is introduced above in combination with the accompanying drawings, and the device provided by the embodiment of the present application is introduced below in combination with the accompanying drawings.

Based on the same technical concept, an embodiment of the present application provides a communication device, which includes a module/unit/means for executing the method executed by any device in the above method embodiment. The module/unit/means can be implemented by software, or by hardware, or the corresponding software can be implemented by hardware.

Please refer to Figure 15, which is a schematic diagram of the structure of the communication device of the embodiment of the present application. The communication device may include units or modules corresponding to all or part of the steps in the above method embodiment, and can be used to execute the steps executed by the first device in the above embodiment. For details, please refer to the relevant introduction in the above method embodiment.

As shown in Fig. 15, the communication device 1500 includes a processing unit 1510 and an interface unit 1520, wherein the processing unit 1510 may be a processor or a processing circuit, and the interface unit 1520 may also be a transceiver unit or an input/output interface. The communication device 1500 may be used to implement the steps performed by the first device in the above embodiment.

When the communication device 1500 is used to implement the steps performed by the first device in the above embodiment:

The interface unit 1520 is used to obtain at least one set of channel state information and channel state indication information corresponding to the second device;

A processing unit 1510 is used to determine whether a first model of the communication device matches a second model of the second device according to at least one set of channel state information and channel state indication information, wherein the first model is used to process the channel state information to obtain the channel state indication information, and the second model is used to process the channel state indication information to obtain the channel state information, or the first model is used to process the channel state indication information to obtain the channel state information, and the second model is used to process the channel state indication information to obtain the channel state indication information;

The interface unit 1520 is further used to send first indication information to the second device, where the first indication information is used to indicate whether the first model of the communication device matches the second model of the second device.

For other implementation methods, please refer to the relevant introduction of the first device in the aforementioned embodiment, which will not be repeated here.

As shown in FIG16 , the present application further provides a communication device 1600, including a processor 1610, and may further include a communication interface 1620. The processor 1610 and the communication interface 1620 are coupled to each other. It is understandable that the communication interface 1620 may be a transceiver, an input-output interface, an input interface, an output interface, an interface circuit, etc. Optionally, the communication device 1600 may further include a memory 1630 for storing instructions executed by the processor 1610 or storing input data required for the processor 1610 to run instructions or storing data generated after the processor 1610 runs instructions. Among them, the memory 1630 may be a physically independent unit, or may be coupled to the processor 1610, or the processor 1610 may include the memory 1630.

When the communication device 1600 is used to implement the steps executed by the first device in the above embodiment, the processor 1610 can be used to implement the functions of the above processing unit 1510, and the communication interface 1620 can be used to implement the functions of the above interface unit 1520.

It should be understood that the processor mentioned in the embodiments of the present application can be implemented by hardware or by software. When implemented by hardware, the processor can be a logic circuit, an integrated circuit, etc. When implemented by software, the processor can be a general-purpose processor implemented by reading software code stored in a memory.

Exemplarily, the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

It should be understood that the memory mentioned in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (Double Data Eate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (Enhanced SDRAM, ESDRAM), synchronous link dynamic random access memory (Synchlink DRAM, SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DR RAM).

It should be noted that when the processor is a general-purpose processor, DSP, ASIC, FPGA or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, the memory (storage module) can be integrated into the processor.

It should be noted that the memory described herein is intended to include, but is not limited to, these and any other suitable types of memory.

Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

The present application is described with reference to the flowchart and/or block diagram of the method, device (system), and computer program product according to the present application. It should be understood that each process and/or box in the flowchart and/or block diagram, as well as the combination of the process and/or box in the flowchart and/or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing the function specified in one process or multiple processes in the flowchart and/or one box or multiple boxes in the block diagram.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims (24)

A communication method, characterized by comprising: The first device obtains at least one set of channel state information and channel state indication information corresponding to the second device; The first device determines, according to the at least one set of channel state information and channel state indication information, whether a first model of the first device matches a second model of the second device, wherein the first model is used to process the channel state information to obtain the channel state indication information, and the second model is used to process the channel state indication information to obtain the channel state information, or the first model is used to process the channel state indication information to obtain the channel state information, and the second model is used to process the channel state indication information to obtain the channel state indication information; The first device sends first indication information to the second device, where the first indication information is used to indicate whether a first model of the first device matches a second model of the second device. The method according to claim 1, wherein the first device is a first network device, the second device is a first terminal device, the first model is used to process the channel state indication information to obtain the channel state information, and the second model is used to process the channel state information to obtain the channel state indication information; The first device obtains at least one set of channel state information and channel state indication information corresponding to the second device, including: The first network device sends a first request message to the first terminal device, where the first request message is used to request at least one set of channel state information input and channel state indication information output of the first terminal device; The first network device receives the at least one set of channel state information input and channel state indication information output from the first terminal device. The method according to claim 1 or 2, characterized in that the first device is a first network device, the second device is a first terminal device, the first model is used to process the channel state indication information to obtain the channel state information, and the second model is used to process the channel state information to obtain the channel state indication information; The first network device sends a second request message to a second network device, where the second request message is used to request the second network device to input at least one set of channel state indication information and output channel state information corresponding to the first terminal device, and the second network device is a network device that the first terminal device has accessed; The first network device receives the at least one set of channel state indication information input and channel state information output from the second network device. The method according to claim 2 or 3, characterized in that, the first device determines whether the first model of the first device matches the second model of the second device according to the at least one set of channel state information and channel state indication information, comprising: The first network device inputs the channel state indication information into the first model for processing to obtain a channel state information test value; The first network device determines whether the first model of the first network device matches the second model of the first terminal device according to whether the error between the channel state information test value and the channel state information is less than or equal to a first threshold. The method according to any one of claims 2 to 4, characterized in that the method further comprises: When the first model of the first network device does not match the second model of the first terminal device, the first network device sends a second target model to the first terminal device, where the second target model is a second model that matches the first model of the first network device. The method according to claim 1, wherein the first device is a first terminal device, the second device is a first network device, the first model is used to process the channel state information to obtain the channel state indication information, and the second model is used to process the channel state indication information to obtain the channel state information; The at least one set of channel state information and channel state indication information is determined by the first network device according to at least one channel state indication information sent by the first terminal device to a second network device, and the second network device is a network device that the first terminal device has accessed; or The at least one set of channel state information and channel state indication information is determined by the first network device according to at least one channel state indication information from a second terminal device, and the second terminal device is a terminal device that has accessed the first network device. The method of claim 6, wherein the first device determines whether the first model of the first device matches the second model of the second device according to the at least one set of channel state information and channel state indication information, comprising: The first terminal device inputs the channel state information into the first model for processing to obtain a channel state indication information test value; The first terminal device determines whether the first model of the first terminal device matches the second model of the first network device according to whether the error between the channel state indication information test value and the channel state indication information is less than or equal to a second threshold. The method according to claim 6 or 7, characterized in that the method further comprises: The first terminal device receives a first target model from the first network device when the first model of the first terminal device does not match the second model of the first network device, the first target model being a first model that matches the second model; The first terminal device updates the first model of the first terminal device according to the first target model. The method as described in any one of claims 6-8 is characterized in that the first model includes a downsampling layer and at least one feature extraction layer or module, and the downsampling layer is used to downsample the frequency domain dimension of the channel state information. The method according to any one of claims 6 to 9, characterized in that before the first terminal device inputs the channel state information into the first model for processing, the method further comprises: The first terminal device fills the frequency domain dimension data of the channel state information with 0 when the channel state information is not an integral multiple of the reference value of the number of frequency domain dimension data; The first terminal device splits the channel state information into multiple sub-channel state information according to the reference value of the number of frequency domain dimension data. The method as described in any one of claims 2-5 is characterized in that the second model includes a downsampling layer and at least one feature extraction layer or module, and the downsampling layer is used to downsample the frequency domain dimension of the channel state information. A communication device, characterized in that it comprises an interface unit and a processing unit; The interface unit is used to obtain at least one set of channel state information and channel state indication information corresponding to the second device; The processing unit is used to determine whether a first model of the communication device matches a second model of the second device according to the at least one set of channel state information and channel state indication information, wherein the first model is used to process the channel state information to obtain the channel state indication information, and the second model is used to process the channel state indication information to obtain the channel state information, or the first model is used to process the channel state indication information to obtain the channel state information, and the second model is used to process the channel state indication information to obtain the channel state indication information; The interface unit is further used to send first indication information to the second device, where the first indication information is used to indicate whether the first model of the communication device matches the second model of the second device. The device according to claim 12, characterized in that the communication device is a first network device, the second device is a first terminal device, the first model is used to process the channel state indication information to obtain the channel state information, and the second model is used to process the channel state information to obtain the channel state indication information; When the interface unit obtains at least one set of channel state information and channel state indication information corresponding to the second device, it is specifically used to send a first request message to the first terminal device, wherein the first request message is used to request at least one set of channel state information input and channel state indication information output of the first terminal device; and receive the at least one set of channel state information input and channel state indication information output from the first terminal device. The device according to claim 12 or 13, characterized in that the communication device is a first network device, the second device is a first terminal device, the first model is used to process the channel state indication information to obtain the channel state information, and the second model is used to process the channel state information to obtain the channel state indication information; When the interface unit obtains at least one set of channel state information and channel state indication information corresponding to the second device, it is specifically used to send a second request message to the second network device, the second request message is used to request the second network device to input at least one set of channel state indication information and output of channel state information corresponding to the first terminal device, the second network device is a network device that the first terminal device has accessed; and receive the at least one set of channel state indication information input and channel state information output from the second network device. The device as described in claim 13 or 14 is characterized in that when the processing unit determines whether the first model of the communication device matches the second model of the second device based on the at least one set of channel state information and channel state indication information, it is specifically used to input the channel state indication information into the first model for processing to obtain a channel state information test value; and determine whether the first model of the first network device matches the second model of the first terminal device based on whether the error between the channel state information test value and the channel state information is less than or equal to a first threshold. The device as described in any one of claims 13-15 is characterized in that the interface unit is also used to send a second target model to the first terminal device when the first model of the first network device does not match the second model of the first terminal device, and the second target model is a second model that matches the first model of the first network device. The device according to claim 12, characterized in that the communication device is a first terminal device, the second device is a first network device, the first model is used to process the channel state information to obtain the channel state indication information, and the second model is used to process the channel state indication information to obtain the channel state information; The at least one set of channel state information and channel state indication information is determined by the first network device according to at least one channel state indication information sent by the first terminal device to a second network device, and the second network device is a network device that the first terminal device has accessed; or The at least one set of channel state information and channel state indication information is determined by the first network device according to at least one channel state indication information from a second terminal device, and the second terminal device is a terminal device that has accessed the first network device. The device as described in claim 17 is characterized in that when the processing unit determines whether the first model of the communication device matches the second model of the second device based on the at least one set of channel state information and channel state indication information, it is specifically used to input the channel state information into the first model for processing to obtain a channel state indication information test value; and determine whether the first model of the first terminal device matches the second model of the first network device based on whether the error between the channel state indication information test value and the channel state indication information is less than or equal to a second threshold. The device as described in claim 17 or 18 is characterized in that the interface unit is also used to receive a first target model from the first network device when the first model of the first terminal device does not match the second model of the first network device, and the first target model is a first model that matches the second model; and update the first model of the first terminal device according to the first target model. The device as described in any one of claims 17-19 is characterized in that the first model includes a downsampling layer and at least one feature extraction layer or module, and the downsampling layer is used to downsample the frequency domain dimension of the channel state information. The device as described in any one of claims 17-20 is characterized in that the processing unit is also used to fill the frequency domain dimension data of the channel state information with 0 before inputting the channel state information into the first model for processing, when the channel state information is not an integer multiple of the reference value of the number of frequency domain dimension data; and split the channel state information into multiple sub-channel state information according to the reference value of the number of frequency domain dimension data. The device as described in any one of claims 13-16 is characterized in that the second model includes a downsampling layer and at least one feature extraction layer or module, and the downsampling layer is used to downsample the frequency domain dimension of the channel state information. A chip, characterized in that the chip is used to implement the method according to any one of claims 1 to 11. A computer-readable storage medium, characterized in that a computer program or instruction is stored in the storage medium, and when the computer program or instruction is executed by a processor, the method according to any one of claims 1 to 11 is implemented.