Cooperative Beamforming Method and Apparatus, Electronic Device, and Storage Medium

Abstract

Embodiments of the present application provide a cooperative beamforming method and apparatus, an electronic device, and a storage medium. By inputting a first initial node feature vector, a second initial node feature vector, and an initial edge feature vector into a preset update network, a target feature vector is obtained. The target feature vector includes a target edge feature vector. A first constraint condition is determined according to the first initial node feature vector, and the target edge feature vector is subjected to a conversion process to obtain an initial cooperative beamforming vector that satisfies the first constraint condition. The function value of the objective function is calculated according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector. If the function value satisfies a preset second constraint condition, the initial cooperative beamforming vector is used as the target cooperative beamforming vector, and cooperative beamforming is performed according to the target cooperative beamforming vector, so that the cooperative beamforming method can be applicable to scenarios with complex connections between transmitters and receivers.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and in particular, to a cooperative beamforming method and apparatus, an electronic device, and a storage medium. Background Art

[0002] Currently, a wireless network is modeled as a graph, with transmitters and receivers as nodes in the graph. A mapping function from graph features to graph vectors is obtained based on a graph neural network, and the beamforming vector is solved through the mapping function. However, this method is only applicable to the case where a transmitter serves a single receiver in a wireless network and cannot be applied to complex scenarios such as cooperative beamforming design. When a transmitter serves multiple receivers, the graph neural network cannot handle the complex connections between the transmitter and multiple receivers, resulting in the inability to obtain the cooperative beamforming vector. Summary of the Invention

[0003] The main objective of the embodiments of the present application is to propose a cooperative beamforming method and apparatus, an electronic device, and a storage medium, which can handle the complex connections between transmitters and receivers in a wireless network when a transmitter serves multiple receivers to obtain the cooperative beamforming vector.

[0004] To achieve the above objective, a first aspect of the embodiments of the present application proposes a cooperative beamforming method, which is applied to a downlink multiple-input multiple-output system. The downlink multiple-input multiple-output system includes multiple base station nodes and multiple user nodes, and there is a communication channel between the base station nodes and the user nodes. The method includes:

[0005] Obtain a first initial node feature vector of the base station node and a second initial node feature vector of the user node, and use the communication channel from the base station node to the user node as the initial edge feature vector;

[0006] Input the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into a preset update network to obtain a target feature vector, where the target feature vector includes a target edge feature vector;

[0007] Determine a first constraint condition according to the first initial node feature vector;

[0008] Perform a conversion process on the target edge feature vector to obtain an initial cooperative beamforming vector that satisfies the first constraint condition;

[0009] Calculate the function value of the objective function according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector;

[0010] If the function value satisfies a preset second constraint condition, use the initial cooperative beamforming vector as the target cooperative beamforming vector, and perform cooperative beamforming according to the target cooperative beamforming vector.

[0011] In some embodiments, the update network includes a first update layer and a second update layer, and the target feature vector further includes a first target node feature vector and a second target node feature vector. The step of inputting the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into a preset update network to obtain a target feature vector includes:

[0012] Input the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into the first update layer to update the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector, and obtain a first intermediate node feature vector, a second intermediate node feature vector, and an intermediate edge feature vector;

[0013] Input the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector into the second update layer. Update the first intermediate node feature vector according to a preset first node update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the first target node feature vector; update the second intermediate node feature vector according to a preset second node update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the second target node feature vector; update the intermediate edge feature vector according to a preset edge update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the target edge feature vector.

[0014] In some embodiments, before inputting the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into the preset update network, the cooperative beamforming method further includes:

[0015] Input the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into a first multi-layer perceptron, and preprocess the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector respectively based on the first multi-layer perceptron to obtain a preprocessed first initial node feature vector, a preprocessed second initial node feature vector, and a preprocessed initial edge feature vector.

[0016] In some embodiments, the first node update rule includes a second multi-layer perceptron, a third multi-layer perceptron, and a first aggregation function. Updating the first intermediate node feature vector according to the preset first node update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the first target node feature vector includes:

[0017] Performing a non-linear transformation on the second intermediate node feature vector and the intermediate edge feature vector based on the second multi-layer perceptron to obtain a first intermediate feature vector;

[0018] Performing an aggregation process on the first intermediate feature vector according to the first aggregation function to obtain a second intermediate feature vector;

[0019] Performing a non-linear transformation on the first intermediate node feature vector and the second intermediate feature vector based on the third multi-layer perceptron to obtain the first target node feature vector.

[0020] In some embodiments, the second node update rule includes a fourth multi-layer perceptron, a fifth multi-layer perceptron, and a second aggregation function. Updating the second intermediate node feature vector according to the preset second node update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the second target node feature vector includes:

[0021] Performing a non-linear transformation on the first intermediate node feature vector and the intermediate edge feature vector based on the fourth multi-layer perceptron to obtain a third intermediate feature vector;

[0022] Performing an aggregation process on the third intermediate feature vector according to the second aggregation function to obtain a fourth intermediate feature vector;

[0023] Performing a non-linear transformation on the second intermediate node feature vector and the fourth intermediate feature vector based on the fifth multi-layer perceptron to obtain the second target node feature vector.

[0024] In some embodiments, the edge update rule includes a sixth multi-layer perceptron, a seventh multi-layer perceptron, an eighth multi-layer perceptron, and a third aggregation function. Updating the intermediate edge feature vector according to the preset edge update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the target edge feature vector includes:

[0025] Performing a non-linear transformation on the first intermediate node feature vector and the intermediate edge feature vector based on the sixth multi-layer perceptron to obtain a fifth intermediate feature vector;

[0026] Performing a non - linear transformation on the second intermediate node feature vector and the intermediate edge feature vector based on the seventh multi - layer perceptron to obtain a sixth intermediate feature vector;

[0027] Performing an aggregation process on the fifth intermediate feature vector and the sixth intermediate feature vector according to the third aggregation function to obtain a seventh intermediate feature vector;

[0028] Performing a non - linear transformation on the intermediate edge feature vector and the seventh intermediate feature vector based on the eighth multi - layer perceptron to obtain the target edge feature vector.

[0029] In some embodiments, calculating the function value of the objective function according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector includes:

[0030] Calculating the signal - to - interference - plus - noise ratio according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector;

[0031] Performing a logarithmic operation on the signal - to - interference - plus - noise ratio to obtain the function value of the objective function.

[0032] To achieve the above object, a second aspect of the embodiments of the present application proposes a cooperative beamforming device, which is applied to a downlink multiple - input multiple - output system. The downlink multiple - input multiple - output system includes multiple base station nodes and multiple user nodes, and there is a communication channel between the base station nodes and the user nodes. The device includes:

[0033] An acquisition module, configured to acquire a first initial node feature vector of a base station node and a second initial node feature vector of a user node, and use the communication channel from the base station node to the user node as the initial edge feature vector;

[0034] An update module, configured to input the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into a preset update network to obtain target feature vectors, where the target feature vectors include a target edge feature vector;

[0035] A first calculation module, configured to determine a first constraint condition according to the first initial node feature vector;

[0036] A conversion module, configured to perform a conversion process on the target edge feature vector to obtain an initial cooperative beamforming vector that satisfies the first constraint condition;

[0037] A second calculation module, configured to calculate the function value of the objective function according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector;

[0038] A cooperative beamforming module, configured to use the initial cooperative beamforming vector as the target cooperative beamforming vector and perform cooperative beamforming according to the target cooperative beamforming vector if the function value satisfies a preset second constraint condition.

[0039] To achieve the above object, a third aspect of the embodiments of the present application provides an electronic device, including a memory, a processor, a program stored on the memory and executable on the processor, and a data bus for realizing connection communication between the processor and the memory. When the program is executed by the processor, the method described in the first aspect above is implemented.

[0040] To achieve the above object, a fourth aspect of the embodiments of the present application provides a storage medium, which is a computer-readable storage medium for computer-readable storage. The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the method described in the first aspect above.

[0041] The cooperative beamforming method, device, electronic device and storage medium provided by the present application obtain the first initial node feature vector of the base station node and the second initial node feature vector of the user node, use the communication channel from the base station node to the user node as the initial edge feature vector, input the first initial node feature vector, the second initial node feature vector and the initial edge feature vector into a preset update network to obtain a target feature vector, where the target feature vector includes a target edge feature vector, determine a first constraint condition according to the first initial node feature vector, perform a transformation process on the target edge feature vector to obtain an initial cooperative beamforming vector that satisfies the first constraint condition, calculate the function value of the target function according to the second initial node feature vector, the initial edge feature vector and the initial cooperative beamforming vector, and if the function value satisfies a preset second constraint condition, use the initial cooperative beamforming vector as the target cooperative beamforming vector and perform cooperative beamforming according to the target cooperative beamforming vector. By adding an edge feature vector and obtaining a cooperative beamforming vector according to the edge feature vector, the embodiments of the present application enable the cooperative beamforming vector to be defined on the edge rather than on the node, enabling the update network to handle complex connections between the transmitter and multiple receivers during cooperative beamforming. Description of the Drawings

[0042] Figure 1 is a flowchart of the cooperative beamforming method provided by the embodiments of the present application;

[0043] Figure 2 is Figure 1 a flowchart of step S120 in

[0044] Figure 3 isFigure 2 The first flowchart of step S220 in

[0045] Figure 4 is Figure 2 The second flowchart of step S220 in

[0046] Figure 5 is Figure 3 The third flowchart of step S220 in

[0047] Figure 6 is Figure 1 The flowchart of step S150 in

[0048] Figure 7 is a schematic diagram of the cooperative beamforming method provided by an embodiment of the present application;

[0049] Figure 8 is the first result diagram of the cooperative beamforming method provided by an embodiment of the present application;

[0050] Figure 9 is the second result diagram of the cooperative beamforming method provided by an embodiment of the present application;

[0051] Figure 10 is the third result diagram of the cooperative beamforming method provided by an embodiment of the present application;

[0052] Figure 11 is the fourth result diagram of the cooperative beamforming method provided by an embodiment of the present application;

[0053] Figure 12 is a schematic structural diagram of the cooperative beamforming device provided by an embodiment of the present application;

[0054] Figure 13 is a schematic hardware structure diagram of the electronic device provided by an embodiment of the present application. Detailed implementation manners

[0055] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] It should be noted that although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order from the module division in the device or the order in the flowchart. Terms such as "first" and "second" in the specification, claims and the above-mentioned drawings are used to distinguish similar objects and do not necessarily need to describe a specific order or sequence.

[0057] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terms used herein are for the purpose of describing embodiments of this application only and are not intended to limit this application.

[0058] Cooperative beamforming is a technique in modern wireless communication systems that can meet the rapidly growing demand for wireless data traffic. In traditional methods, cooperative beamforming design is formulated as an optimization problem and solved iteratively on a case-by-case basis. Currently, by learning the mapping function from problem instances to corresponding solutions and obtaining the solution for cooperative beamforming design according to the mapping function, the real-time conversion requirements of problem instances can be met. Among various learning architectures, the Graph Neural Network (GNN) can effectively utilize the graph topology in wireless networks to achieve lower training complexity and better generalization ability, which is beneficial for wireless management. By modeling the wireless network as a graph, problem instances and corresponding solutions can be represented as graph features and graph variables, and based on GNN, the mapping function from graph features to graph variables can be learned. However, existing GNNs for communication networks are only equipped with node update mechanisms, such as the Message Passing Graph Neural Network (MPGNN) and the Permutation equivariance GNN (PGNN). When MPGNN is used for beamforming design in device-to-device networks, each transmitter in the communication network serves only a single receiver, and each transceiver pair is defined as a node of the graph, and the interference links between different transceiver pairs are defined as edges. Since MPGNN does not contain edge variables, the beamforming vector for each transceiver pair can only be defined on the corresponding node. When PGNN is used for power allocation in multi-cell systems, each transmitter serves multiple receivers in its cell. Since PGNN does not contain edge variables, each equivalent antenna is regarded as a separate node, and the transmit power is defined as a node variable. The GNN only equipped with a node update mechanism updates the node features to node variables through the node update mechanism. When performing cooperative beamforming design, the variables can only be defined on the nodes, resulting in the GNN architecture only with a node update mechanism not being able to easily scale to more complex scenarios, such as scenarios where a transmitter serves multiple receivers in cooperative beamforming and a receiver is served by multiple transmitters, and it cannot handle the complex connections between transmitters and receivers.

[0059] Based on this, the embodiments of the present application provide a cooperative beamforming method and apparatus, an electronic device, and a storage medium, aiming to improve the applicability of GNN in complex scenarios, enable GNN to process complex connections between a transmitter and multiple receivers, obtain a cooperative beamforming vector, and perform cooperative beamforming of electromagnetic waves according to the cooperative beamforming vector, which can meet the wireless data traffic requirements and improve the communication quality of the communication network.

[0060] The cooperative beamforming method and apparatus, electronic device, and storage medium provided by the embodiments of the present application are specifically described through the following embodiments. First, the cooperative beamforming method in the embodiments of the present application is described.

[0061] The cooperative beamforming method provided by the embodiments of the present application relates to the field of wireless communication technologies. The cooperative beamforming method provided by the embodiments of the present application can be applied to a terminal, a server, or software running on a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application implementing the cooperative beamforming method, etc., but is not limited to the above forms.

[0062] The present application can be used in many general-purpose or special-purpose computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0063] Figure 1It is an alternative flowchart of the cooperative beamforming method provided by the embodiments of the present application, which is applied to a downlink multiple-input multiple-output system. The downlink multiple-input multiple-output system includes multiple base station nodes and multiple user nodes, and there is a communication channel between the base station nodes and the user nodes. Figure 1 The method in it may include but is not limited to steps S110 to S160.

[0064] Step S110: Obtain the first initial node feature vector of the base station node and the second initial node feature vector of the user node, and use the communication channel from the base station node to the user node as the initial edge feature vector.

[0065] Step S120: Input the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into a preset update network to obtain target feature vectors, where the target feature vectors include target edge feature vectors.

[0066] Step S130: Determine the first constraint condition according to the first initial node feature vector.

[0067] Step S140: Perform a conversion process on the target edge feature vector to obtain an initial cooperative beamforming vector that satisfies the first constraint condition.

[0068] Step S150: Calculate the function value of the objective function according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector.

[0069] Step S160: If the function value satisfies the preset second constraint condition, use the initial cooperative beamforming vector as the target cooperative beamforming vector, and perform cooperative beamforming according to the target cooperative beamforming vector.

[0070] In step S110 of some embodiments, the wireless network is modeled as a graph. The base stations in the wireless network are the base station nodes (i.e., BS nodes) in the graph, the users in the wireless network are the user nodes (i.e., UE nodes) in the graph, and the communication channel from the base station to the user is the edge in the graph. If the wireless network communication system is a downlink multiple-input multiple-output system (Multiple-Input Multiple-Output, MIMO system), where M base station nodes cooperate to serve K user nodes, each base station node is equipped with N antennas and serves all user nodes, and each user node is equipped with a single antenna and is served by all base station nodes, then the first initial node vector of the base station node is represented as f BS =[P1, P2, …, PM], where PM represents the node feature of the Mth base station node, and the node feature is the maximum transmission power. The node feature of the user node is defined by the noise standard deviation. The node features of the K user nodes constitute the second initial node feature vector, and the second initial node feature vector of the user node is represented as Since the noise of the user nodes follows a complex Gaussian distribution Therefore, is used to characterize the noise of the K-th user node. The communication channel is h m,k from the m-th base station node to the k-th user node. The communication channel can be an interference link or a communication link. The communication channel is represented as an initial edge feature vector E, where E ∈ C M×K×N .

[0071] In step S120 of some embodiments, the network is updated to a graph neural network, and the problem of solving the cooperative beamforming vector is transformed into the problem of the graph neural network learning a mapping function. By using the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector as the graph features of the wireless network, and inputting these graph features into the graph neural network, the node features in the first initial node feature vector and the second initial node feature vector are updated according to the node update rule of the update layer of the graph neural network, obtaining the first target node feature vector and the second target node feature vector. The first initial node feature vector, the second initial node feature vector, and the initial edge feature vector are respectively updated according to the update layer of the graph neural network to obtain the target feature vector, so that the graph neural network can learn the mapping function from the graph features to the graph variables. According to the mapping function, the graph features are transformed into the corresponding graph variables. Since the target feature vector includes the target edge feature vector, the beamforming vector can be defined on the edge. The mapping function is expressed as V = φ(f BS , f UE , E), where V is the graph variable and V (m,k,:) is the beamforming vector defined on the edge (m, k).

[0072] It should be noted that the mapping function has the property of permutation equivariance. Specifically, if the element order in the mapping function (f BS , f UE , E) is permuted, the mapping function should permute the element order in the output V correspondingly, so that the element order in V corresponds to the element order in the mapping function.

[0073] In step S130 of some embodiments, the first constraint condition is the maximum power limit. Specifically, the first constraint condition is shown in formula (1).

[0074]

[0075] where P m is the node feature of the first initial node feature vector, representing the initial node feature of the m-th base station node, that is, the maximum transmit power of the m-th base station node.

[0076] In step S140 of some embodiments, the target edge feature vector is transformed by a Multilayer Perceptron (MLP) of the graph neural network post-processing layer, that is, the target edge feature vector is standard normalized to obtain an initial cooperative beamforming vector that satisfies the first constraint condition.

[0077] In step S150 of some embodiments, the objective function is the sum rate of user nodes and can be expressed as The node feature of the corresponding user node in the second initial node feature vector, the corresponding edge feature in the initial edge feature vector, and the initial cooperative beamforming vector corresponding to the edge feature are used as input parameters of the objective function to calculate the function value of the objective function.

[0078] In step S160 of some embodiments, the second constraint condition is that the sum rate of user nodes reaches the maximum value. When the function value satisfies the second constraint condition, that is, the sum rate of all user nodes reaches the maximum value, it indicates that the initial cooperative beamforming vector is the optimal beamforming vector and can achieve the purpose of maximizing the sum rate in the communication scenario. The initial cooperative beamforming vector is used as the target cooperative beamforming vector, and cooperative beamforming is performed according to the target cooperative beamforming vector.

[0079] Steps S110 to S160 illustrated in the embodiments of the present application obtain the first initial node feature vector of the base station node and the second initial node feature vector of the user node, and use the communication channel from the base station node to the user node as the initial edge feature vector. The first initial node feature vector, the second initial node feature vector, and the initial edge feature vector are input into a preset update network to obtain a target feature vector. The target feature vector includes a target edge feature vector. The target edge feature vector is transformed to obtain an initial cooperative beamforming vector that satisfies the first constraint condition. The function value of the objective function is calculated according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector. If the function value satisfies the preset second constraint condition, the initial cooperative beamforming vector is used as the target cooperative beamforming vector, and cooperative beamforming is performed according to the target cooperative beamforming vector. In the embodiments of the present application, by defining cooperative beamforming on the edge rather than on the node, the update network can process the complex connection between the transmitter and multiple receivers to obtain the cooperative beamforming vector, and perform cooperative beamforming according to the cooperative beamforming vector.

[0080] Please refer to Figure 2 , in some embodiments, the update network includes a first update layer and a second update layer. The target feature vector further includes a first target node feature vector and a second target node feature vector. Step S120 may include but is not limited to steps S210 to S220:

[0081] Step S210: Input the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into the first update layer to update the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector, and obtain the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector;

[0082] Step S220: Input the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector into the second update layer. Update the first intermediate node feature vector to obtain the first target node feature vector according to the preset first node update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector; Update the second intermediate node feature vector to obtain the second target node feature vector according to the preset second node update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector; Update the intermediate edge feature vector to obtain the target edge feature vector according to the preset edge update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector.

[0083] Before step S210, input the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into the first multi-layer perceptron, and preprocess the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector respectively based on the first multi-layer perceptron to obtain the preprocessed first initial node feature vector, the preprocessed second initial node feature vector, and the preprocessed initial edge feature vector. Specifically, input the first initial node feature vector f BS 、the second initial node feature vector f UE and the initial edge feature vector E into the preprocessing layer of the graph neural network. The preprocessing layer includes three first multi-layer perceptrons, and each first multi-layer perceptron corresponds to an input feature. Perform a non-linear transformation on the first initial node feature vector f BS to obtain the preprocessed first initial node feature vector The preprocessed first initial node feature vector includes the node representation of the base station node, and d BS is the dimension of the node representation of the base station node; Perform a non-linear transformation on the second initial node feature vector f UE to obtain the preprocessed second initial node feature vector The preprocessed second initial node feature vector includes the node representation of the user node, and d UE is the dimension of the node representation of the user node; Perform a non-linear transformation on the initial edge feature vector E to obtain the preprocessed initial edge feature vector The initial edge feature vector after preprocessing includes the edge representation between the base station node and the user node, d E is the dimension of the edge representation.

[0084] In steps S210 to S220 of some embodiments, the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector are input into the first update layer. According to the first node update rule, the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector of the first update layer, the first initial node feature vector is updated to obtain the first intermediate node feature vector. According to the second node update rule, the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector of the first update layer, the second initial node feature vector is updated to obtain the second intermediate node feature vector. According to the edge update rule, the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector of the first update layer, the initial edge feature vector is updated to obtain the intermediate edge feature vector. The first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector are input into the second update layer. According to the first intermediate node feature vector, the second intermediate node feature vector, the intermediate edge feature vector, and the first node update rule of the second update layer, the first intermediate node feature vector is updated to obtain the first target node feature vector. According to the second node update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector of the second update layer, the second intermediate node feature vector is updated to obtain the second target node feature vector. According to the edge update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector of the second update layer, the intermediate edge feature vector is updated to obtain the target edge feature vector.

[0085] It should be noted that the three feature vectors input into the first update layer can be the preprocessed first initial node feature vector, the preprocessed second initial node feature vector, and the preprocessed initial edge feature vector, or the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector that have not been preprocessed.

[0086] Please refer to Figure 3 , in some embodiments, the first node update rule includes a second multi-layer perceptron, a third multi-layer perceptron, and a first aggregation function. Step S220 may include but is not limited to steps S310 to S330:

[0087] Step S310, based on the second multi-layer perceptron, performs a non-linear transformation on the second intermediate node feature vector and the intermediate edge feature vector to obtain the first intermediate feature vector;

[0088] Step S320: Aggregate the first intermediate feature vector according to the first aggregation function to obtain a second intermediate feature vector;

[0089] Step S330: Perform a non-linear transformation on the first intermediate node feature vector and the second intermediate feature vector based on the third multi-layer perceptron to obtain a first target node feature vector.

[0090] Specifically, the first node update rule is shown in Equation (2).

[0091]

[0092] where, is the node representation of the m-th base station node output by the l-th update layer; is the second multi-layer perceptron of the l-th update layer; is the third multi-layer perceptron of the l-th update layer; is the first aggregation function of the l-th update layer; is the set of user nodes connected to the m-th base station node; is the node representation of the k-th user node in the set of user nodes output by the (l - 1)-th update layer; is the edge feature representation defined on the edge connecting the m-th base station node and the k-th user node.

[0093] In steps S310 to S330 of some embodiments, the first intermediate node feature vector the second intermediate node feature vector and the intermediate edge feature vector E( l-1 ) are input into the l-th update layer. Based on the second multi-layer perceptron perform a non-linear mapping on the second intermediate node feature vector and the intermediate edge feature vector E (l-1) to obtain a first intermediate feature vector. Aggregate the first intermediate feature vector according to the first aggregation function to obtain a second intermediate feature vector. Perform a non-linear transformation on the first intermediate node feature vector and the second intermediate feature vector based on the third multi-layer perceptron to obtain a first target node feature vector

[0094] It should be noted that the first intermediate feature vector is subjected to max aggregation processing according to the first aggregation function, that is, the maximum value of each dimension of the first intermediate feature vector is taken to obtain the second intermediate feature vector. For example, if there are two vectors [1, 3, 4] and [6, 1, 9] respectively, after max aggregation processing, the obtained vector is [6, 3, 9].

[0095] Please refer to Figure 4 , in some embodiments, the second node update rule includes a fourth multi-layer perceptron, a fifth multi-layer perceptron, and a second aggregation function. Step S220 may further include but is not limited to steps S410 to S430:

[0096] Step S410, perform a non-linear transformation on the first intermediate node feature vector and the intermediate edge feature vector based on the fourth multi-layer perceptron to obtain a third intermediate feature vector;

[0097] Step S420, perform an aggregation process on the third intermediate feature vector according to the second aggregation function to obtain a fourth intermediate feature vector;

[0098] Step S430, perform a non-linear transformation on the second intermediate node feature vector and the fourth intermediate feature vector based on the fifth multi-layer perceptron to obtain a second target node feature vector.

[0099] Specifically, the second node update rule is shown in formula (3).

[0100]

[0101] Wherein, is the node representation of the k-th user node output by the l-th update layer; is the fourth multi-layer perceptron of the l-th update layer; is the fifth multi-layer perceptron of the l-th update layer; is the second aggregation function of the l-th update layer; is the set of base station nodes connected to the k-th user node; is the node representation of the k-th user node in the user node set output by the (l - 1)-th update layer; is the edge feature representation defined on the edge connecting the m-th base station node and the k-th user node.

[0102] In steps S410 to S430 of some embodiments, the first intermediate node feature vector output by the (l - 1)-th update layer the second intermediate node feature vector and the intermediate edge feature vector E (l-1) are input into the l-th update layer. Based on the fourth multi-layer perceptron perform a non-linear mapping on the first intermediate node feature vector and the intermediate edge feature vector E (l-1) to obtain a third intermediate feature vector. According to the second aggregation function perform an aggregation process on the third intermediate feature vector to obtain a fourth intermediate feature vector. Based on the fifth multi-layer perceptron For the second intermediate node feature vector and the fourth intermediate feature vector are subjected to a non-linear transformation to obtain a second target node feature vector

[0103] Please refer to Figure 5 , in some embodiments, the edge update rule includes a sixth multi-layer perceptron, a seventh multi-layer perceptron, an eighth multi-layer perceptron, and a third aggregation function. Step S240 includes but is not limited to steps S510 to S540:

[0104] Step S510, based on the sixth multi-layer perceptron, performs a non-linear transformation on the first intermediate node feature vector and the intermediate edge feature vector to obtain a fifth intermediate feature vector;

[0105] Step S520, based on the seventh multi-layer perceptron, performs a non-linear transformation on the second intermediate node feature vector and the intermediate edge feature vector to obtain a sixth intermediate feature vector;

[0106] Step S530, according to the third aggregation function, aggregates the fifth intermediate feature vector and the sixth intermediate feature vector to obtain a seventh intermediate feature vector;

[0107] Step S540, based on the eighth multi-layer perceptron, performs a non-linear transformation on the intermediate edge feature vector and the seventh intermediate feature vector to obtain a target edge feature vector.

[0108] Specifically, the edge update rule is shown in formula (4).

[0109]

[0110] Among them, is the edge feature representation defined on the edge connecting the m-th base station node and the k-th user node output by the l-th update layer; is the sixth multi-layer perceptron of the l-th update layer; is the seventh multi-layer perceptron of the l-th update layer; is the eighth multi-layer perceptron of the l-th update layer; is the third aggregation function of the l-th update layer; is the set of user nodes connected to the m-th base station node; is the set of base station nodes connected to the k-th user node.

[0111] In steps S510 to S540 of some embodiments, in order to obtain the edge representation defined on the edge connecting the m-th base station node to the k-th user node, based on the sixth multi-layer perceptron Perform a non - linear transformation on the edge representation defined on edge (m, k1) of the output of the (l - 1)-th update layer and the node representation of the m-th base station node to obtain a fifth intermediate feature vector, where edge (m, k1) is the edge formed by connecting the m-th base station node with the k1-th user node in the user node set. Based on the seventh multi - layer perceptron Perform a non - linear transformation on the edge representation defined on edge (m1, k) of the output of the (l - 1)-th update layer and the node representation of the k-th user node to obtain a sixth intermediate feature vector, where edge (m1, k) is the edge formed by connecting the m1-th base station node in the base station node set with the k-th user node. Perform a max - aggregation process on the fifth intermediate feature vector and the sixth intermediate feature vector according to the third aggregation function to obtain a seventh intermediate feature vector. Perform a non - linear transformation on the edge representation defined on edge (m, k) and the seventh intermediate feature vector based on the eighth multi - layer perceptron to obtain the edge representation defined on edge (m, k).

[0112] Please refer to Figure 6 , in some embodiments, step S150 may include but is not limited to steps S610 to S620:

[0113] Step S610, calculate the signal - to - interference - plus - noise ratio (SINR) according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector;

[0114] Step S620, perform a logarithmic operation on the SINR to obtain the function value of the objective function.

[0115] In step S610 of some embodiments, if the signal sent by the BS node to the k-th UE node in the communication network is s k , due to the existence of interference and noise, the signal y received by the k-th UE node k is as shown in formula (5).

[0116]

[0117] where y k includes three terms. The first term is the signal term, the second term is the interference term, which is used to characterize the interference from the BS node, and the third term is the noise term, which is used to characterize the noise from the UE node. The noise follows a complex Gaussian distribution

[0118] The SINR of the k-th UE node is as shown in formula (6).

[0119]

[0120] Take the second initial node feature vector the initial edge feature vector h m,k and the initial cooperative beamforming vector v m,kInput the signal-to-interference-plus-noise ratio formula shown in formula (6) to obtain the signal-to-interference-plus-noise ratio.

[0121] In step S620 of some embodiments, perform a logarithmic operation on the signal-to-interference-plus-noise ratio to obtain the function value of the objective function, where the objective function is shown in formula (7).

[0122]

[0123] Refer to Figure 7 , the embodiment of the present application performs cooperative beamforming based on the graph neural network Edge-GNN with a node update mechanism and an edge update mechanism. The graph neural network includes a preprocessing layer, L update layers, and a postprocessing layer. Input the node feature f of the BS node BS , the node feature f of the UE node UE , and the edge feature E defined on the edge connecting the BS node and the UE node into the graph neural network. According to the preprocessing layer, preprocess the node features f BS , f UE and the edge feature E to obtain the initial node representation and the initial edge representation E (0) . Input the initial node representation and the initial edge representation into the first update layer. Update the initial base station node representation according to the first node update rule to obtain the intermediate base station node representation output by the first update layer Update the initial user node representation according to the second node update rule to obtain the intermediate user node representation output by the first update layer Update the initial edge representation E (0 ) according to the edge update rule to obtain the intermediate edge representation E( 1 ). Input the intermediate base station node representation, the intermediate user node representation, and the intermediate edge representation output by the first update layer into the second update layer. Update the intermediate base station node representation according to the first node update rule of the second update layer, update the intermediate user node representation according to the second node update rule of the second update layer, and update the intermediate edge representation according to the edge update rule of the second update layer. By analogy, input the intermediate base station node representation, the intermediate user node representation, and the intermediate edge representation output by the previous update layer into the next update layer for update until the target base station node representation the target user node representation and the target edge representation E( L) Note that the first node update rule, the second node update rule, and the edge update rule of each update layer are the same. The first node update rule is shown in Equation (2), the second node update rule is shown in Equation (3), and the edge update rule is shown in Equation (4). The dimension of the representation does not change in the update layer, and the cooperative beamforming vector is obtained by transforming the target edge representation according to the post-processing layer.

[0124] In a downlink MIMO system, within a 2×2 km 2 area, M BSs cooperate to serve K UEs. The BSs are uniformly distributed in this area, and the minimum distance between BSs is 500 meters. All UE nodes are uniformly distributed between 50 meters and 250 meters away from the BS nodes. Each BS is equipped with 2 antennas, with a maximum transmit power of 33 dBm, and the path loss is 30.5 + 36.7 log 10 d, in dB, where d is the distance in meters. The small-scale channel follows Rayleigh fading, and the noise power of each UE is -99 dBm. An Edge-GNN with L = 2 update layers is adopted. The aggregation function AGG is implemented by a max aggregator, which returns the maximum value of each dimension of the input vector. The dimensions d BS 、d UE and d E are all 64. All MLPs in the BS node update mechanism, UE node update mechanism, and edge update mechanism are implemented by 3 linear layers, with 64, 256, and 64 neurons respectively, and each linear layer is followed by a ReLU activation function.

[0125] During the training process of the graph neural network, the number of iterations is set to 500, and each iteration contains 256 training samples. The learning rate is γ = 10 -4 . In each training sample, the positions of the BS nodes and UE nodes and the small-scale channel are randomly generated. The number of BS nodes is set to 5, and the number of UE nodes is set to 2. The parameters of the Edge-GNN are updated through the RMSProp optimizer. After training, the average performance of the Edge-GNN is tested on 100 samples. This average performance includes the performance of the Edge-GNN in terms of sum rate and calculation time. The performance of the Edge-GNN in terms of sum rate under different numbers of BS nodes is shown in Figure 8 , and the performance of the Edge-GNN in terms of calculation time under different numbers of BS nodes is shown in Figure 9 , the performance of the Edge-GNN in terms of sum rate under different numbers of UE nodes is shown in Figure 10 , and the performance of the Edge-GNN in terms of calculation time under different numbers of UE nodes is shown in Figure 11 .

[0126] Note that the power of the UE received signal is obtained by subtracting the path loss from the transmit power of the BS.

[0127] It should be further noted that the maximum transmit power of the m-th base station node is 33 dBm. 33 dBm is 0 dB, and 0 dB is 1 V, so P m is 1.

[0128] By testing the sum rate of the trained Edge-GNN under 5 to 8 BSs, the generalization ability of Edge-GNN for different numbers of BSs is demonstrated. As Figure 8 shown, Edge-GNN achieves a higher sum rate than GP and WMMSE. As the number of BSs increases, Edge-GNN still achieves a higher sum rate than GP and WMMSE. This superiority is because Edge-GNN learns a general mapping function from edge features (channel state), BS node features (power budget), and UE node features (noise power) to the corresponding beamforming vectors, and this mapping function is independent of the number of BSs. In addition, the number of parameters in Edge-GNN is independent of the number of BSs, making the trained Edge-GNN applicable to different numbers of BSs. Since Edge-GNN has the permutation equivariance property, i.e., the PE property, many unnecessary permutation training samples are avoided, improving the generalization ability of the graph neural network.

[0129] As Figure 9 shown, compared with GP and WMMSE, the computing time of Edge-GNN is significantly shortened. The computing time of Edge-GNN is even shortened by more than 1000 times compared with WMMSE. Edge-GNN runs fast and is applicable to real-time communication scenarios.

[0130] By testing the sum rate of the trained Edge-GNN under 2 to 8 UEs, the generalization ability of Edge-GNN for different numbers of UEs is demonstrated. As Figure 10 shown, as the number of UEs increases, Edge-GNN is superior to GP and WMMSE in terms of sum rate. As Figure 11 shown, under different numbers of UEs, the computing time of Edge-GNN is also shorter than that of GP and WMMSE.

[0131] Please refer to Figure 12 , this embodiment of the present application also provides a cooperative beamforming device, which can implement the above cooperative beamforming method and is applied to a downlink multiple-input multiple-output system. The downlink multiple-input multiple-output system includes multiple base station nodes and multiple user nodes, and there is a communication channel between the base station nodes and the user nodes. The device includes:

[0132] An acquisition module 1210, configured to acquire a first initial node feature vector of the base station node and a second initial node feature vector of the user node, and use the communication channel from the base station node to the user node as the initial edge feature vector;

[0133] An update module 1220, configured to input a first initial node feature vector, a second initial node feature vector, and an initial edge feature vector into a preset update network to obtain a target feature vector, where the target feature vector includes a target edge feature vector;

[0134] A first calculation module 1230, configured to determine a first constraint condition according to the first initial node feature vector;

[0135] A conversion module 1240, configured to perform a conversion process on the target edge feature vector to obtain an initial cooperative beamforming vector that satisfies the first constraint condition;

[0136] A second calculation module 1250, configured to calculate a function value of an objective function according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector;

[0137] A cooperative beamforming module 1260, configured to, if the function value satisfies a preset second constraint condition, use the initial cooperative beamforming vector as a target cooperative beamforming vector, and perform cooperative beamforming according to the target cooperative beamforming vector.

[0138] The specific implementation manner of this cooperative beamforming device is basically the same as the specific embodiments of the above cooperative beamforming method, and will not be elaborated here.

[0139] An embodiment of this application further provides an electronic device, including: a memory, a processor, a program stored on the memory and executable on the processor, and a data bus for implementing connection communication between the processor and the memory. When the program is executed by the processor, the above cooperative beamforming method is implemented. The electronic device may be any intelligent terminal including a tablet computer, an in-vehicle computer, etc.

[0140] Please refer to Figure 13 , Figure 13 which schematically shows the hardware structure of an electronic device in another embodiment. The electronic device includes:

[0141] A processor 1310, which may be implemented by using a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is configured to execute relevant programs to implement the technical solutions provided by the embodiments of this application;

[0142] The memory 1320 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 1320 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 1320 and are called by the processor 1310 to execute the cooperative beamforming method of the embodiments of this application;

[0143] The input / output interface 1330 is used to implement information input and output;

[0144] The communication interface 1340 is used to implement communication interaction between this device and other devices. It can achieve communication through a wired manner (such as USB, network cable, etc.) or through a wireless manner (such as mobile network, WIFI, Bluetooth, etc.);

[0145] The bus 1350 transmits information between the various components of the device (such as the processor 1310, the memory 1320, the input / output interface 1330, and the communication interface 1340);

[0146] Among them, the processor 1310, the memory 1320, the input / output interface 1330, and the communication interface 1340 achieve communication connections with each other inside the device through the bus 1350.

[0147] The embodiments of this application also provide a storage medium. The storage medium is a computer-readable storage medium for computer-readable storage. The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the above-mentioned cooperative beamforming method.

[0148] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory can include high-speed random access memory, and can also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory optionally includes a memory remotely set relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above networks include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0149] The cooperative beamforming method, cooperative beamforming device, electronic device, and storage medium provided by the embodiments of the present application obtain a first initial node feature vector of a base station node and a second initial node feature vector of a user node, use the communication channel from the base station node to the user node as the initial edge feature vector, input the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into a preset update network to obtain target feature vectors, where the target feature vectors include target edge feature vectors, perform a transformation process on the target edge feature vectors to obtain an initial cooperative beamforming vector that satisfies the first constraint condition, calculate the function value of the objective function according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector, and if the function value satisfies the preset second constraint condition, use the initial cooperative beamforming vector as the target cooperative beamforming vector, and perform cooperative beamforming according to the target cooperative beamforming vector. By defining cooperative beamforming on edges rather than on nodes, the update network in the embodiments of the present application can process complex connections between transmitters and multiple receivers to obtain a cooperative beamforming vector, and perform cooperative beamforming according to the cooperative beamforming vector.

[0150] The embodiments described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation to the technical solutions provided by the embodiments of the present application. Those skilled in the art know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0151] Those skilled in the art can understand that Figure 1-11 the technical solutions shown do not constitute a limitation to the embodiments of the present application, and may include more or fewer steps than those shown in the figures, or combine certain steps, or different steps.

[0152] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0153] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and their appropriate combinations.

[0154] In the description of the present application and the above-mentioned drawings, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0155] It should be understood that in the present application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that three relationships may exist. For example, "A and / or B" may mean: only A exists, only B exists, and both A and B exist at the same time. Among them, A and B can be singular or plural. The character " / " generally means that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or a similar expression means any combination of these items, including any combination of single items (ones) or plural items (ones). For example, at least one (one) of a, b, or c can mean: 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 plural.

[0156] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the above-mentioned unit division is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of devices or units can be in an electrical, mechanical or other form.

[0157] The units described above as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0158] In addition, in each embodiment of the present application, each functional unit can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0159] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of the present application. The aforementioned storage medium includes: various media that can store programs, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0160] The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the rights of the embodiments of the present application. Any modifications, equivalent replacements, and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall fall within the scope of the rights of the embodiments of the present application.

Claims

1. A cooperative beamforming method, characterized in that, Applied to a downlink multiple-input multiple-output system, the downlink multiple-input multiple-output system includes a plurality of base station nodes and a plurality of user nodes, and there is a communication channel between the base station nodes and the user nodes. The method includes: Obtain a first initial node feature vector of the base station node and a second initial node feature vector of the user node, and use the communication channel from the base station node to the user node as the initial edge feature vector; Input the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into a preset update network to obtain a target feature vector, where the target feature vector includes a target edge feature vector; Determine a first constraint condition according to the first initial node feature vector; the first constraint condition is a maximum power limit, and the first constraint condition is expressed as: Among them, m represents the m-th base station node, and P m is the node feature of the first initial node feature vector of the m-th base station node, V( m,k,: ) is the beamforming vector defined on the edge (m, k), k represents the k-th user node, M is the number of base station nodes, and K is the number of user nodes; Perform a conversion process on the target edge feature vector to obtain an initial cooperative beamforming vector that satisfies the first constraint condition; Calculate the function value of the objective function according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector; If the function value satisfies a preset second constraint condition, use the initial cooperative beamforming vector as the target cooperative beamforming vector, and perform cooperative beamforming according to the target cooperative beamforming vector; the preset second constraint condition is that the sum rate of the user nodes reaches the maximum value.

2. The collaborative beamforming method according to claim 1, characterized in that, The update network includes a first update layer and a second update layer. The target feature vector further includes a first target node feature vector and a second target node feature vector. The step of inputting the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into a preset update network to obtain a target feature vector includes: Input the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into the first update layer to update the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector to obtain a first intermediate node feature vector, a second intermediate node feature vector, and an intermediate edge feature vector; Input the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector into the second update layer. Update the first intermediate node feature vector according to a preset first node update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the first target node feature vector; update the second intermediate node feature vector according to a preset second node update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the second target node feature vector; update the intermediate edge feature vector according to a preset edge update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the target edge feature vector.

3. The collaborative beamforming method according to claim 1, wherein Before inputting the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into a preset update network, the cooperative beamforming method further includes: Inputting the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into a first multi-layer perceptron, and preprocessing the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector respectively based on the first multi-layer perceptron to obtain a preprocessed first initial node feature vector, a preprocessed second initial node feature vector, and a preprocessed initial edge feature vector.

4. The collaborative beamforming method according to claim 2, wherein The first node update rule includes a second multi-layer perceptron, a third multi-layer perceptron, and a first aggregation function. Updating the first intermediate node feature vector according to the preset first node update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the first target node feature vector includes: Performing a non-linear transformation on the second intermediate node feature vector and the intermediate edge feature vector based on the second multi-layer perceptron to obtain a first intermediate feature vector; Performing an aggregation process on the first intermediate feature vector according to the first aggregation function to obtain a second intermediate feature vector; Performing a non-linear transformation on the first intermediate node feature vector and the second intermediate feature vector based on the third multi-layer perceptron to obtain the first target node feature vector.

5. The collaborative beamforming method according to claim 2, wherein The second node update rule includes a fourth multi-layer perceptron, a fifth multi-layer perceptron, and a second aggregation function. Updating the second intermediate node feature vector according to the preset second node update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the second target node feature vector includes: Performing a non-linear transformation on the first intermediate node feature vector and the intermediate edge feature vector based on the fourth multi-layer perceptron to obtain a third intermediate feature vector; Performing an aggregation process on the third intermediate feature vector according to the second aggregation function to obtain a fourth intermediate feature vector; Performing a non-linear transformation on the second intermediate node feature vector and the fourth intermediate feature vector based on the fifth multi-layer perceptron to obtain the second target node feature vector.

6. The collaborative beamforming method according to claim 2, wherein The edge update rule includes a sixth multi-layer perceptron, a seventh multi-layer perceptron, an eighth multi-layer perceptron, and a third aggregation function. Updating the intermediate edge feature vector according to the preset edge update rule, the first intermediate node feature vector, the second intermediate node feature vector, and the intermediate edge feature vector to obtain the target edge feature vector includes: Performing a non-linear transformation on the first intermediate node feature vector and the intermediate edge feature vector based on the sixth multi-layer perceptron to obtain a fifth intermediate feature vector; Performing a non-linear transformation on the second intermediate node feature vector and the intermediate edge feature vector based on the seventh multi-layer perceptron to obtain a sixth intermediate feature vector; Aggregate the fifth intermediate feature vector and the sixth intermediate feature vector according to the third aggregation function to obtain a seventh intermediate feature vector; Based on the eighth multi-layer perceptron, perform a non-linear transformation on the intermediate edge feature vector and the seventh intermediate feature vector to obtain the target edge feature vector.

7. The collaborative beamforming method according to any one of claims 1 to 6, characterized in that, The calculating the function value of the objective function according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector includes: Calculate the signal-to-interference-plus-noise ratio according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector; Perform a logarithmic operation on the signal-to-interference-plus-noise ratio to obtain the function value of the objective function.

8. A cooperative beamforming device, characterized in that, Applied to a downlink multiple-input multiple-output system, the downlink multiple-input multiple-output system includes a plurality of base station nodes and a plurality of user nodes, and there is a communication channel between the base station nodes and the user nodes. The apparatus includes: An acquisition module, configured to acquire a first initial node feature vector of a base station node and a second initial node feature vector of a user node, and use the communication channel from the base station node to the user node as an initial edge feature vector; An update module, configured to input the first initial node feature vector, the second initial node feature vector, and the initial edge feature vector into a preset update network to obtain target feature vectors, where the target feature vectors include target edge feature vectors; A first calculation module, configured to determine a first constraint condition according to the first initial node feature vector; the first constraint condition is a maximum power limit, and the first constraint condition is expressed as: Among them, m represents the m-th base station node, and P m is the node feature of the first initial node feature vector of the m-th base station node, V( m,k,: ) is the beamforming vector defined on the edge (m, k), k represents the k-th user node, M is the number of base station nodes, and K is the number of user nodes; A conversion module, configured to perform a conversion process on the target edge feature vector to obtain an initial cooperative beamforming vector that satisfies the first constraint condition; A second calculation module, configured to calculate the function value of the objective function according to the second initial node feature vector, the initial edge feature vector, and the initial cooperative beamforming vector; A cooperative beamforming module, configured to, if the function value satisfies a preset second constraint condition, use the initial cooperative beamforming vector as a target cooperative beamforming vector, and perform cooperative beamforming according to the target cooperative beamforming vector; the preset second constraint condition is that the sum rate of the user nodes reaches the maximum value.

9. An electronic device, characterized in that, The electronic device includes a memory, a processor, a program stored on the memory and executable on the processor, and a data bus for realizing the connection and communication between the processor and the memory. When the program is executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A storage medium, which is a computer-readable storage medium for computer-readable storage, characterized in that, The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the method according to any one of claims 1 to 7.

Images