A base station resource allocation method and device, electronic equipment and storage medium

Abstract

The application discloses a base station resource allocation method and device, electronic equipment and storage medium. The method comprises the following steps: acquiring a plurality of communication scene images photographed by a plurality of base stations; performing three-dimensional target detection processing on the communication scene images to obtain a detection frame of a plurality of target objects; constructing a target object channel space feature according to the detection frame of the plurality of target objects; inputting the target object channel space feature into a resource allocation model to perform resource allocation processing and obtaining a resource allocation result. The embodiment of the application can realize base station resource allocation assisted by computer vision images and neural networks, improve the base station resource allocation efficiency, and can be widely applied in the technical field of communication.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a base station resource allocation method, apparatus, electronic device and storage medium. Background Technology

[0002] With the widespread and dense deployment of 5G base stations and the rapid growth in the number of mobile terminal users, efficient resource scheduling and management in massive multi-antenna and multi-user scenarios has become an important future development direction. Base station scheduling and resource allocation methods for massive multi-base station and multi-user communication systems have become crucial means to save beam scanning energy consumption, reduce inter-link channel interference, and improve the overall transmission performance of the system. In related technologies, base station resource allocation methods require obtaining channel information between each base station and user, and calculating the optimal resource allocation scheme based on exhaustive search or iterative solution methods. However, this method requires high communication resource overhead, resulting in low resource scheduling efficiency of the communication system. In summary, the technical problems existing in related technologies need to be improved. Summary of the Invention

[0003] The main objective of this application is to provide a base station resource allocation method, apparatus, electronic device, and storage medium that can improve resource allocation efficiency.

[0004] To achieve the above objectives, one aspect of this application proposes a base station resource allocation method, the method comprising:

[0005] Acquire communication scene images captured by multiple base stations;

[0006] The communication scene image is subjected to three-dimensional target detection processing to obtain detection boxes of multiple target objects;

[0007] The target object channel space features are constructed based on the detection bounding boxes of the multiple target objects;

[0008] The target object's channel space characteristics are input into the resource allocation model for resource allocation processing to obtain the resource allocation result.

[0009] In some embodiments, acquiring communication scene images captured by multiple base stations includes the following steps:

[0010] Multiple base stations are set up in the communication system, and a corresponding camera is configured for each base station;

[0011] The communication scene image is obtained by performing joint view coverage processing on the service area of ​​the communication system using the camera.

[0012] In some embodiments, constructing the target object channel space features based on the detection bounding boxes of the plurality of target objects includes the following steps:

[0013] The minimum base station distance between each target object and the multiple base stations is determined based on the detection bounding boxes of the multiple target objects;

[0014] The target objects are sorted in ascending order based on the minimum base station distance to obtain a set of target objects.

[0015] Based on the set of target objects, the detection box is processed to construct a size matrix and a direction vector to obtain the channel space features of the target objects.

[0016] In some embodiments, the step of inputting the channel space characteristics of the target object into a resource allocation model for resource allocation processing to obtain a resource allocation result includes the following steps:

[0017] The target object's channel space characteristics are input into the resource allocation model, and the base station number matrix and base station energy matrix are output.

[0018] The base station number matrix and the base station energy matrix are decoded to obtain the resource allocation scheme between each target object and the corresponding base station;

[0019] The resource allocation scheme is used to allocate resources to multiple target objects to obtain the resource allocation result.

[0020] In some embodiments, inputting the channel space characteristics of the target object into the resource allocation model and outputting a base station number matrix and a base station energy matrix includes the following steps:

[0021] The size matrix and direction vector are obtained from the channel space features of the target object;

[0022] The size matrix and the direction vector are respectively input into the resource allocation model, which includes a first convolutional neural network and a second convolutional neural network;

[0023] The size matrix is ​​convolved using the first convolutional neural network to obtain size features;

[0024] The direction vector is convolved by the second convolutional neural network to obtain vector features;

[0025] The size feature and the vector feature are subjected to element-wise addition and convolution to obtain the base station number matrix and the base station energy matrix.

[0026] In some embodiments, decoding the base station number matrix and the base station energy matrix to obtain a resource allocation scheme for each target object and its corresponding base station includes the following steps:

[0027] The matrix containing the maximum element value is extracted from the base station number matrix, and the matrix containing the maximum element value is determined as the allocation matrix;

[0028] The target objects are assigned base stations according to the allocation matrix, and then returned to the matrix containing the maximum element value extracted from the base station number matrix, until all target objects are assigned base stations to obtain the base station allocation scheme.

[0029] Based on the base station allocation scheme and the base station energy matrix, average transmission energy allocation is performed on each base station to obtain the resource allocation scheme between each target object and the corresponding base station.

[0030] In some embodiments, before inputting the target object channel space features into the resource allocation model for resource allocation processing, the method further includes pre-training the resource allocation model, specifically including the following steps:

[0031] Acquire training images and perform 3D object detection and bounding box construction on the training images to obtain training samples;

[0032] The training target channel is collected, and the base station resource scheduling algorithm based on traversal search is used to perform resource scheduling processing on the training target channel to obtain the resource scheduling result;

[0033] The training samples are labeled according to the resource scheduling results to obtain training data.

[0034] The training data is input into the resource allocation model, and the training results are output.

[0035] Based on the training results, the parameters of the resource allocation model are adjusted to obtain the trained resource allocation model.

[0036] To achieve the above objectives, another aspect of this application provides a base station resource allocation apparatus, the apparatus comprising:

[0037] The first module is used to acquire communication scene images captured by multiple base stations;

[0038] The second module is used to perform three-dimensional target detection processing on the communication scene image to obtain detection boxes of multiple target objects;

[0039] The third module is used to construct the channel space features of the target objects based on the detection boxes of the multiple target objects;

[0040] The fourth module is used to input the channel space characteristics of the target object into the resource allocation model for resource allocation processing to obtain the resource allocation result.

[0041] In some embodiments, the first module is configured to acquire communication scene images captured by multiple base stations, including:

[0042] The first unit is used to set up multiple base stations in the communication system and configure a corresponding camera for each base station;

[0043] The second unit is used to perform joint view coverage processing on the service area of ​​the communication system through the camera, and capture images of the communication scene.

[0044] In some embodiments, the third module is configured to construct channel space features of the target objects based on the detection bounding boxes of the plurality of target objects, including:

[0045] The third unit is used to determine the minimum base station distance between each target object and the multiple base stations based on the detection frames of the multiple target objects;

[0046] The fourth unit is used to sort each target object in ascending order according to the minimum base station distance to obtain a set of target objects;

[0047] The fifth unit is used to construct the size matrix and direction vector of the detection box based on the target object set to obtain the channel space features of the target object.

[0048] In some embodiments, the fourth module is used to input the channel space characteristics of the target object into a resource allocation model for resource allocation processing to obtain a resource allocation result, including:

[0049] The sixth unit is used to input the channel space characteristics of the target object into the resource allocation model and output the base station number matrix and the base station energy matrix.

[0050] The seventh unit is used to decode the base station number matrix and the base station energy matrix to obtain the resource allocation scheme between each target object and the corresponding base station;

[0051] The eighth unit is used to allocate resources to multiple target objects according to the resource allocation scheme and obtain resource allocation results.

[0052] In some embodiments, the sixth unit is used to input the channel space characteristics of the target object into the resource allocation model and output a base station number matrix and a base station energy matrix, including:

[0053] The first sub-unit is used to obtain the size matrix and direction vector from the channel space features of the target object;

[0054] The second subunit is used to input the size matrix and the direction vector into the resource allocation model, which includes a first convolutional neural network and a second convolutional neural network.

[0055] The third subunit is used to perform convolution processing on the size matrix through the first convolutional neural network to obtain size features;

[0056] The fourth subunit is used to perform convolution processing on the direction vector through the second convolutional neural network to obtain vector features;

[0057] The fifth subunit is used to perform element-wise addition and convolution processing on the size features and the vector features to obtain the base station number matrix and the base station energy matrix.

[0058] In some embodiments, the seventh unit is configured to decode the base station numbering matrix and the base station energy matrix to obtain a resource allocation scheme for each target object and its corresponding base station, including:

[0059] The sixth subunit is used to extract the matrix containing the maximum element value from the base station number matrix and determine the matrix containing the maximum element value as the allocation matrix;

[0060] The seventh subunit is used to perform base station allocation processing on the target objects according to the allocation matrix, and return to the matrix containing the maximum element value extracted from the base station number matrix, until base station allocation processing is performed on all target objects to obtain the base station allocation scheme;

[0061] The eighth subunit is used to perform average transmission energy allocation processing on each base station according to the base station allocation scheme and the base station energy matrix to obtain the resource allocation scheme between each target object and the corresponding base station.

[0062] In some embodiments, before the fourth module inputs the target object channel space features into the resource allocation model for resource allocation processing, the apparatus further includes a fifth module for pre-training the resource allocation model, specifically including:

[0063] The ninth unit is used to acquire training images and perform 3D object detection and detection box construction on the training images to obtain training samples;

[0064] The tenth unit is used to collect the training target channel and perform resource scheduling processing on the training target channel based on the traversal search base station resource scheduling algorithm to obtain the resource scheduling result;

[0065] The eleventh unit is used to label the training samples according to the resource scheduling results to obtain training data;

[0066] The twelfth unit is used to input the training data into the resource allocation model and output the training results.

[0067] The thirteenth unit is used to adjust the parameters of the resource allocation model based on the training results, so as to obtain the trained resource allocation model.

[0068] To achieve the above objectives, another aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described above.

[0069] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods described above.

[0070] The embodiments of this application include at least the following beneficial effects: This application provides a base station resource allocation method, apparatus, electronic device, and storage medium. This scheme acquires communication scene images captured by multiple base stations, and can utilize computer vision image assistance to achieve base station resource allocation, thereby improving the efficiency of resource allocation. Furthermore, this scheme also performs three-dimensional target detection processing on the communication scene images to obtain detection boxes for multiple target objects, which can better detect the three-dimensional spatial coordinates and dimensions of the target objects, improving detection accuracy. Moreover, this scheme constructs the channel spatial features of the target objects based on the detection boxes of multiple target objects; inputting the channel spatial features of the target objects into the resource allocation model for resource allocation processing, and obtaining the resource allocation result, can directly output the optimal resource allocation scheme for the base station using a neural network, which has lower time overhead and improves the efficiency of base station resource allocation. Attached Figure Description

[0071] Figure 1 This is a flowchart of a base station resource allocation method provided in an embodiment of this application;

[0072] Figure 2 yes Figure 1 The flowchart of step S101 in the text;

[0073] Figure 3 This is a schematic diagram of a communication scenario provided in an embodiment of this application;

[0074] Figure 4 This is a schematic diagram of a detection frame provided in an embodiment of this application;

[0075] Figure 5 yes Figure 1 The flowchart of step S103 in the process;

[0076] Figure 6 yes Figure 1 The flowchart of step S104 in the process;

[0077] Figure 7 yes Figure 6 The flowchart of step S601 in the process;

[0078] Figure 8 This is a schematic diagram of the network structure of a resource allocation model provided in an embodiment of this application;

[0079] Figure 9 This is a schematic diagram of the structure of a base station resource allocation device provided in an embodiment of this application;

[0080] Figure 10 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0081] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0082] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various concepts, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if,” “when,” or “in response to a determination” as used herein may be interpreted as “when…” or “when…” or “in response to a determination.”

[0083] As used in this application, the terms "at least one", "multiple", "each", "any", etc., "at least one" includes one, two or more, "multiple" includes two or more, "each" refers to each of the corresponding multiples, and "any" refers to any one of the multiples.

[0084] 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 terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0085] Before providing a detailed description of the embodiments of this application, some of the nouns and terms involved in the embodiments of this application will be explained first. The nouns and terms involved in the embodiments of this application are subject to the following interpretations.

[0086] 1) 3D target detection: Using computer vision to identify and locate the three-dimensional bounding box of a specific target or object in an image, including the object's position, length, width, height, and azimuth information.

[0087] 2) Neural Network: A type of deep learning, consisting of a hierarchical structure of interconnected neuron nodes. Each connection has a weight, which can be adjusted through deep learning to learn complex mapping relationships between inputs and outputs.

[0088] 3) Base station resource allocation: The process of effectively managing and allocating parameters such as base station energy and the number of connected users in order to improve network performance and optimize the overall transmission rate of the communication system.

[0089] With the widespread and dense deployment of 5G base stations and the rapid growth in the number of mobile terminal users, high-frequency bands such as Massive MIMO antennas are being deployed in densely populated areas. Efficient resource scheduling and management in large-scale multi-antenna and multi-user scenarios is a crucial future development direction. Breakthroughs are needed in the configuration technology for resource allocation parameters such as high-level coverage, regional links, and energy. Beamforming and beam velocity positioning are required for high-level coverage and some future 5GtoB high-speed, low-latency applications. Therefore, base station scheduling and resource allocation methods for large-scale multi-base station, multi-user communication systems have become crucial for saving energy from beam scanning, reducing inter-link channel interference, and improving overall system transmission performance.

[0090] In related technologies, there are base station resource allocation methods that require obtaining channel information between each base station and the user, and calculating the optimal resource allocation scheme of the system based on exhaustive search or iterative solution methods. However, in practical applications, it has been found that this method has high time, spectrum, and computational resource overhead, resulting in low resource scheduling efficiency and overall link transmission rate of the communication system, thus affecting the efficiency of base station resource allocation.

[0091] In view of this, this application provides a base station resource allocation method, apparatus, electronic device, and storage medium. This solution acquires communication scene images captured by multiple base stations, enabling base station resource allocation using computer vision image assistance, thus improving resource allocation efficiency. Furthermore, this solution performs 3D target detection processing on the communication scene images to obtain detection boxes for multiple target objects, enabling better detection of the 3D spatial coordinates and dimensions of the target objects, thus improving detection accuracy. Moreover, this solution constructs channel space features of the target objects based on the detection boxes of multiple target objects; inputting these channel space features into a resource allocation model for resource allocation processing yields the resource allocation result. This allows for the direct output of the optimal base station resource allocation scheme using a neural network, resulting in lower time overhead and improved base station resource allocation efficiency.

[0092] The base station resource allocation method provided in this application relates to the field of communication technology. This method can be applied to a terminal, a server, or software running on a terminal or server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, or vehicle terminal, but is not limited thereto. The server can be configured as an independent physical server, a server cluster or 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 communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network. The software can be an application implementing the base station resource allocation method, but is not limited to the above forms.

[0093] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: 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, and distributed computing environments including any of the above systems or devices. This 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. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0094] It should be noted that in all specific embodiments of this application, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of this application require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirection to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data required for the proper functioning of these embodiments acquired.

[0095] Figure 1 This is an optional flowchart of a base station resource allocation method provided in an embodiment of this application. Figure 1 The method may include, but is not limited to, steps S101 to S104.

[0096] Step S101: Acquire communication scene images captured by multiple base stations;

[0097] Step S102: Perform three-dimensional target detection processing on the communication scene image to obtain detection boxes of multiple target objects;

[0098] Step S103: Construct the channel space features of the target objects based on the detection boxes of the multiple target objects;

[0099] Step S104: Input the channel space characteristics of the target object into the resource allocation model for resource allocation processing to obtain the resource allocation result.

[0100] Steps S101 to S104 of this embodiment involve acquiring communication scene images captured by multiple base stations. The communication system has multiple base stations, which can capture real-time communication scene images containing communication objects. This embodiment, by capturing communication scene images, can combine computer vision technology to assist base station resource allocation, thereby improving resource allocation efficiency. Then, a 3D target detection algorithm is used to perform three-dimensional target detection processing on the communication scene images, resulting in detection boxes for multiple target objects. These target objects are communication objects or communication users, i.e., objects requiring communication resource scheduling by the communication system. They can be smartphones or vehicle terminals, without limitation. The detection boxes include the target object's position coordinates, length, width, height, and azimuth parameters. This application embodiment also constructs target object channel space features based on detection boxes of multiple target objects. These target object channel space features include the size matrix and orientation vector of the target object. Then, the constructed target object channel space features are input into a resource allocation model composed of a multi-layer two-dimensional convolutional neural network. Resource allocation is performed through the resource allocation model to obtain resource allocation results. These results represent the resource scheduling results between each target object and the base station allocated in the communication system. This application embodiment can directly output the optimal base station resource allocation scheme using a neural network, resulting in lower time overhead and improved base station resource allocation efficiency.

[0101] Please see Figure 2 In step S101 of some embodiments, acquiring communication scene images captured by multiple base stations includes the following steps:

[0102] Step S201: Set up multiple base stations in the communication system and configure a corresponding camera for each base station;

[0103] Step S202: The camera performs joint view coverage processing on the service area of ​​the communication system to capture the communication scene image.

[0104] In step S201 of some embodiments, this application is applied to a large-scale multi-base station, multi-user communication system, wherein multiple base stations are deployed in the communication system. This application, by configuring a corresponding camera for each base station, can acquire images of the communication scene through the camera, thereby realizing integrated communication sensing technology based on the combination of computer vision and communication. By adding camera-assisted communication functions to operator base stations, the resource allocation efficiency of base stations can be improved.

[0105] In step S202 of some embodiments, the cameras configured at each base station can jointly cover the service area of ​​the entire communication system, enabling the horizontal and vertical fields of view of all cameras to jointly cover the service area of ​​the communication system, thereby capturing a communication scene image. Embodiments of this application can perform image fusion processing on the images captured by each base station according to the shooting angle to obtain a communication scene image. Please refer to... Figure 3 This application embodiment can be applied to a traffic communication scenario where multiple vehicle-mounted terminals communicate with multiple base stations via wireless links at an intersection. This application embodiment deploys a camera on each base station to acquire images of the communication scene, providing a data foundation for subsequent resource allocation.

[0106] In step S102 of some embodiments, the communication scene image undergoes three-dimensional target detection processing to obtain multiple detection boxes for target objects. Each detection box includes the target object's position coordinates, length, width, height, and azimuth parameters. Please refer to [link to relevant documentation]. Figure 4 In traffic communication scenarios, vehicle-mounted terminals or communication terminals within vehicles communicate with base stations. This application embodiment uses a 3D target detection algorithm to mark vehicles in the communication scene image, obtaining detection boxes. This application embodiment can perform three-dimensional target detection processing on the communication scene image using a 3D target detection algorithm, obtaining the position, size, and category of key targets in 3D space, such as motor vehicles, pedestrians, and cyclists. Compared to 2D target detection, which only generates 2D bounding boxes on the image and ignores the actual distance information between the target and the vehicle, 3D target detection focuses on the localization and recognition of targets in the real-world 3D coordinate system. This application embodiment uses 3D target recognition technology to obtain the 3D spatial coordinates and dimensions of environmental objects from the image. Compared to current visual-assisted communication methods based on 2D target recognition, this application embodiment has higher accuracy in detecting environmental scattering objects and can achieve better auxiliary performance.

[0107] Please see Figure 5 In step S103 of some embodiments, constructing the target object channel space features based on the detection boxes of the plurality of target objects includes the following steps:

[0108] Step S501: Determine the minimum base station distance between each target object and the multiple base stations based on the detection frames of the multiple target objects;

[0109] Step S502: Sort each target object in ascending order according to the minimum base station distance to obtain a set of target objects;

[0110] Step S503: Based on the target object set, the detection box is processed to construct a size matrix and a direction vector to obtain the target object channel space features.

[0111] In step S501 of some embodiments, multiple detection boxes of target objects are obtained. Each detection box includes the position coordinates of each target object. Based on the position coordinates, the distance between each target object and each base station in the communication system can be determined, thereby determining the minimum base station distance between the target object and the base station.

[0112] In step S502 of some embodiments, a sequence number is assigned to each target object according to the minimum base station distance of each target object. All target objects are sorted in ascending order of minimum base station distance and assigned sequence numbers to obtain a set of target objects. This set includes all target objects and the sequence number corresponding to each target object.

[0113] In step S503 of some embodiments, after determining the sequence number of each target object or communication user, channel space feature design is performed on the detection box based on the target object set. This channel space feature design includes the construction of a size matrix and a direction vector, wherein the size matrix has a size of N. max ×3×2,N max This represents the total number of target objects in the communication scenario, and the first layer of the size matrix is ​​set to [l n ,w n ,h n ], where l n w n with h n These represent the length, width, and height of the 3D detection box corresponding to the nth user, where n is a positive integer. The second layer of this size matrix is ​​set to [x]. n ,y n ,z n ], where x n y n With z n These are the center coordinates of the 3D bounding box corresponding to the nth user, and the remaining rows of this size matrix are set to zero vectors. The nth element of the direction vector is set to the orientation angle d of the nth user. n The other elements of the direction vector d are set to zero. This embodiment of the application constructs channel space features, providing a data foundation for subsequent resource allocation through a neural network model.

[0114] Please see Figure 6 In step S104 of some embodiments, the step of inputting the channel space characteristics of the target object into the resource allocation model for resource allocation processing to obtain the resource allocation result includes the following steps:

[0115] Step S601: Input the channel space characteristics of the target object into the resource allocation model, and output the base station number matrix and the base station energy matrix;

[0116] Step S602: Decode the base station number matrix and the base station energy matrix to obtain the resource allocation scheme of each target object and the corresponding base station;

[0117] Step S603: Allocate resources to multiple target objects according to the resource allocation scheme to obtain resource allocation results.

[0118] In this embodiment, the channel spatial characteristics of the target object are input into a pre-trained resource allocation model. This model outputs a base station number matrix and a base station energy matrix. After obtaining these matrices, decoding is required to determine the specific base station number and energy level required for each user to connect to—that is, the resource allocation scheme for each target object and its corresponding base station. Based on this scheme, multiple target objects are allocated resources, thus completing the base station resource allocation. This embodiment obtains the base station number and corresponding energy level required for each user to connect to, maximizing the overall system transmission rate and improving base station resource allocation efficiency.

[0119] Please see Figure 7 In step S601 of some embodiments, the step of inputting the channel space characteristics of the target object into the resource allocation model and outputting the base station number matrix and the base station energy matrix includes the following steps:

[0120] Step S701: Obtain the size matrix and direction vector from the channel space features of the target object;

[0121] Step S702: Input the size matrix and the direction vector into the resource allocation model, which includes a first convolutional neural network and a second convolutional neural network.

[0122] Step S703: Perform convolution processing on the size matrix using the first convolutional neural network to obtain size features;

[0123] Step S704: The direction vector is convolved by the second convolutional neural network to obtain vector features;

[0124] Step S705: Perform element-wise addition and convolution on the size feature and the vector feature to obtain the base station number matrix and the base station energy matrix.

[0125] In this embodiment, the resource allocation model is constructed using a multi-layer two-dimensional convolutional neural network. Please refer to [link / reference]. Figure 8 In this embodiment, the user size matrix F and the user direction vector d are used as inputs to a resource allocation neural network. These inputs are fed into two multi-layer two-dimensional convolutional neural networks, namely the first convolutional neural network and the second convolutional neural network. Through convolution processing, size features and vector features are obtained respectively. Then, through element-wise addition and convolution processing, the final output is a matrix I1, I2, ..., I of the base stations that the user needs to connect to. V With the corresponding base station energy coefficients P1, P2, ..., P V Among them, the base station numbering matrix I v The dimension is N max ×1, v=1,2,…,V, where V represents the number of base stations, N max This represents the total number of users in the communication scenario. If the optimal base station number that the nth user needs to connect to is... Then let The nth element is 1, and the remaining elements are set to zero, and it is necessary to ensure that I v The sum of all elements is less than the maximum number of users N that the v-th base station can serve. v At the same time, let the base station energy coefficient P v The average value of the optimal transmit energy required to serve all users of the v-th base station. In this embodiment, by inputting the channel spatial characteristics of the target object into the resource allocation model, the optimal resource allocation scheme for the base station can be directly output using a neural network, thereby improving the efficiency of base station resource allocation.

[0126] In step S602 of some embodiments, decoding the base station number matrix and the base station energy matrix to obtain a resource allocation scheme for each target object and its corresponding base station includes the following steps:

[0127] The matrix containing the maximum element value is extracted from the base station number matrix, and the matrix containing the maximum element value is determined as the allocation matrix;

[0128] The target objects are assigned base stations according to the allocation matrix, and then returned to the matrix containing the maximum element value extracted from the base station number matrix, until all target objects are assigned base stations to obtain the base station allocation scheme.

[0129] Based on the base station allocation scheme and the base station energy matrix, average transmission energy allocation is performed on each base station to obtain the resource allocation scheme between each target object and the corresponding base station.

[0130] In this embodiment of the application, the optimal user base station number matrix I1,I2,...,I ​​output by the neural network is obtained. V With the corresponding base station energy matrix P1, P2, ..., P V After that, it is necessary to further process the matrices I1, I2, ..., IV With matrices P1, P2, ..., P V Decoding is performed to obtain the specific base station numbers and base station power required for each user to connect to. The specific decoding steps are as follows: First, extract the matrix containing the largest element value from the base station number matrix. For example, in the base station number matrix I1, I2, ..., I... V The matrix containing the largest element is If the maximum element value is... Let N be the total number of users in the communication scenario, then let The optimal connection base station number for each user is: Next, the second-largest element value in the base station numbering matrix is ​​obtained, and the above steps are repeated until all users are assigned the optimal base station connection number. During the repeated base station number allocation process, if the number of connected users to a base station exceeds its maximum serviceable user count, that user is assigned to the next second-largest element value corresponding to the second-optimal base station. The base station energy matrix P1, P2, ..., P... V Decoding then makes the total energy of all user base stations of the v-th base station P. v If the energy exceeds the maximum energy of the corresponding base station, then only the transmission energy of that base station is set to the maximum. This embodiment of the application can obtain the base station number and corresponding base station energy that each user needs to connect to through a neural network, thereby maximizing the overall system transmission rate and improving the efficiency of base station resource allocation.

[0131] In some embodiments, before inputting the target object channel space features into the resource allocation model for resource allocation processing, the method further includes pre-training the resource allocation model, specifically including the following steps:

[0132] Acquire training images and perform 3D object detection and bounding box construction on the training images to obtain training samples;

[0133] The training target channel is collected, and the base station resource scheduling algorithm based on traversal search is used to perform resource scheduling processing on the training target channel to obtain the resource scheduling result;

[0134] The training samples are labeled according to the resource scheduling results to obtain training data.

[0135] The training data is input into the resource allocation model, and the training results are output.

[0136] Based on the training results, the parameters of the resource allocation model are adjusted to obtain the trained resource allocation model.

[0137] In this embodiment, firstly, images of the communication scene captured during training, user channel spatial features, and corresponding optimal user connection base station numbers and energy are collected to construct training data for the resource allocation model. Multi-view scene images captured by cameras at multiple times are used as training images, and user channel spatial features at corresponding times are constructed using a 3D object detection algorithm, serving as input features for the resource allocation neural network to obtain training samples. Then, at each image capture time, user channels at that time are collected, and the optimal connection base station number and energy for each user at that time are obtained using a traditional traversal-search-based base station resource scheduling algorithm, serving as training labels for the resource allocation neural network. Training data is obtained by labeling the training samples. Specifically, the traversal-search-based base station resource scheduling algorithm first traverses all possible link connection combinations between the base station and the user, then calculates the optimal energy and achievable transmission rate for each link combination to serve the user, selecting the link combination with the maximum achievable transmission rate, thus obtaining the optimal connection base station number and energy for each user. The labeled training data is then input into the resource allocation model, and the training results are obtained by convolutional processing of the training data using the model's convolutional neural network. The similarity between the training results and the labels of the training data is calculated using the loss function of the resource allocation model. Based on the similarity, the loss function of the resource allocation model is optimized. The model loss of the loss function is backpropagated, and the model parameters are continuously adjusted until the similarity is greater than or equal to the similarity threshold. At this point, the optimization of the resource allocation model is stopped, and the trained resource allocation model is obtained.

[0138] The solutions of this application embodiment will be described in detail and explained below with reference to specific application examples:

[0139] This application's embodiments can be applied to technologies such as mobile communication and artificial intelligence. They can be applied to mobile network optimization in the communication field, significantly improving wireless network feedback optimization efficiency, reducing optimization costs, ensuring smooth network operation, and enhancing network operational efficiency. In one feasible embodiment, this application is applied to a traffic communication scenario. In complex traffic environments, multiple vehicle-mounted terminals communicate with multiple base stations in the communication system, requiring resource scheduling of the base stations through the communication system. This application's embodiments utilize cameras added to the active antenna unit of the base station and leverage computer vision image assistance on the base station side to achieve base station resource allocation. This eliminates the need for frequent and real-time collaborative scheduling among multiple base stations. The optimal transmission energy of the base station and the identities of the users requiring connection can be directly estimated from image information captured by multiple cameras, effectively reducing the transmission rate decrease caused by base station scheduling resource overhead. Specifically, this embodiment utilizes a target detection algorithm from the computer's perspective at the base station to obtain the position, length, width, height, and azimuth information of various vehicle-mounted terminals and other target objects within a traffic communication scene from images captured by the base station. Since vehicle-mounted terminals are difficult to identify, image recognition detection can be performed on the corresponding vehicles to obtain detection boxes for each vehicle. The user channel spatial features are designed using the obtained 3D detection boxes of multiple target objects. These user channel spatial features can fully characterize the channel correlation of each user under the influence of surrounding scattering objects, and can thus be used to effectively estimate the overall optimal connection and energy allocation scheme for the system's base stations. Finally, by inputting the user channel spatial features into the resource allocation model, the base station number that each user needs to connect to and the corresponding base station transmission energy are obtained to maximize the overall system transmission rate, thus completing the base station resource allocation. This embodiment utilizes visual image prediction to determine the optimal energy and beam resource scheduling scheme in a complex multi-base station, multi-user communication system. Compared to resource allocation methods based on exhaustive search or iterative methods, it can significantly reduce the system's computational and radio frequency resource overhead with only a small loss in rate performance, thereby improving the efficiency of base station resource allocation.

[0140] Please see Figure 9 This application also provides a base station resource allocation apparatus that can implement the above-described base station resource allocation method. The apparatus includes:

[0141] The first module 901 is used to acquire communication scene images captured by multiple base stations;

[0142] The second module 902 is used to perform three-dimensional target detection processing on the communication scene image to obtain detection boxes of multiple target objects;

[0143] The third module 903 is used to construct the channel space features of the target objects based on the detection boxes of the multiple target objects;

[0144] The fourth module 904 is used to input the channel space characteristics of the target object into the resource allocation model for resource allocation processing to obtain the resource allocation result.

[0145] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0146] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned base station resource allocation method. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0147] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0148] Please see Figure 10 , Figure 10 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes:

[0149] The processor 1001 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0150] The memory 1002 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 1002 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 1002 and is called and executed by the processor 1001 using the base station resource allocation method of the embodiments of this application.

[0151] Input / output interface 1003 is used to implement information input and output;

[0152] The communication interface 1004 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0153] Bus 1005 transmits information between various components of the device (e.g., processor 1001, memory 1002, input / output interface 1003, and communication interface 1004);

[0154] The processor 1001, memory 1002, input / output interface 1003 and communication interface 1004 are connected to each other within the device via bus 1005.

[0155] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described base station resource allocation method.

[0156] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0157] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0158] This application provides a base station resource allocation method, apparatus, electronic device, and storage medium. This solution acquires communication scene images captured by multiple base stations, enabling base station resource allocation using computer vision image assistance, thus improving allocation efficiency. Furthermore, the solution performs 3D target detection processing on the communication scene images to obtain multiple target object detection boxes, allowing for better detection of the 3D spatial coordinates and dimensions of the target objects, improving detection accuracy. Moreover, the solution constructs target object channel spatial features based on the multiple target object detection boxes; inputting these features into a resource allocation model for resource allocation processing yields the allocation result. This allows for direct output of the optimal base station resource allocation scheme using a neural network, resulting in lower time overhead and improved base station resource allocation efficiency.

[0159] This application embodiment utilizes a camera added to the active antenna unit of the base station and leverages computer vision image assistance on the base station side to achieve base station resource allocation. Compared to resource allocation methods based on communication channel estimation and detection, this application embodiment eliminates the need for frequent and real-time collaborative scheduling among multiple base stations. It can directly estimate the optimal transmission energy of the base station and the identities of users requiring connection through image information captured by multiple cameras, effectively reducing the transmission rate decrease caused by base station scheduling resource overhead. Furthermore, this application embodiment directly outputs the optimal base station resource allocation scheme using a neural network, resulting in lower time overhead compared to resource allocation methods based on exhaustive search or iteration. Moreover, this application embodiment utilizes base station-side computer vision 3D target recognition technology to obtain the 3D spatial coordinates and dimensions of environmental objects from images. Compared to visual-assisted communication methods based on 2D target recognition, this application embodiment achieves higher accuracy in detecting environmental scattering objects and provides better auxiliary performance.

[0160] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0161] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0162] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0163] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0164] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0165] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0166] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0167] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0168] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0169] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the 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 to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0170] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A base station resource allocation method, characterized in that, The method includes the following steps: Acquire communication scene images captured by multiple base stations; The communication scene image is subjected to three-dimensional target detection processing to obtain detection boxes of multiple target objects; The target object channel space features are constructed based on the detection bounding boxes of the multiple target objects; The target object's channel space characteristics are input into the resource allocation model for resource allocation processing to obtain the resource allocation result; The step of inputting the channel space characteristics of the target object into the resource allocation model for resource allocation processing to obtain the resource allocation result includes the following steps: The target object's channel space characteristics are input into the resource allocation model, and the output is a base station number matrix and a base station energy matrix. The base station number matrix is ​​used to represent the number of the base station that the target object needs to connect to, and the base station energy matrix is ​​used to represent the base station energy coefficient corresponding to the base station number. The base station energy coefficient represents the average value of the optimal transmission energy required by the base station to serve all its target objects. The base station number matrix and the base station energy matrix are decoded to obtain the resource allocation scheme between each target object and its corresponding base station; The resource allocation is performed on multiple target objects according to the resource allocation scheme to obtain the resource allocation result; The step of inputting the channel space characteristics of the target object into the resource allocation model and outputting the base station number matrix and the base station energy matrix includes the following steps: The size matrix and direction vector are obtained from the channel space features of the target object; The size matrix and the direction vector are respectively input into the resource allocation model, which includes a first convolutional neural network and a second convolutional neural network; The size matrix is ​​convolved using the first convolutional neural network to obtain size features; The direction vector is convolved by the second convolutional neural network to obtain vector features; The size features and the vector features are subjected to element-wise addition and convolution to obtain the base station number matrix and the base station energy matrix; The process of decoding the base station number matrix and the base station energy matrix to obtain the resource allocation scheme for each target object and its corresponding base station includes the following steps: The matrix containing the maximum element value is extracted from the base station number matrix, and the matrix containing the maximum element value is determined as the allocation matrix; Based on the allocation matrix, base station allocation is performed on the target objects, and the process returns to the matrix containing the maximum element value extracted from the base station number matrix, until base station allocation is performed on all target objects to obtain the base station allocation scheme; Based on the base station allocation scheme and the base station energy matrix, average transmission energy allocation is performed on each base station to obtain the resource allocation scheme between each target object and the corresponding base station.

2. The method according to claim 1, characterized in that, The acquisition of communication scene images captured by multiple base stations includes the following steps: Multiple base stations are set up in the communication system, and a corresponding camera is configured for each base station; The communication scene image is obtained by performing joint view coverage processing on the service area of ​​the communication system using the camera.

3. The method according to claim 1, characterized in that, The process of constructing the target object channel space features based on the detection bounding boxes of the multiple target objects includes the following steps: The minimum base station distance between each target object and the multiple base stations is determined based on the detection bounding boxes of the multiple target objects; The target objects are sorted in ascending order based on the minimum base station distance to obtain a set of target objects; Based on the set of target objects, the detection box is processed to construct a size matrix and a direction vector to obtain the channel space features of the target objects.

4. The method according to any one of claims 1 to 3, characterized in that, Before inputting the channel space features of the target object into the resource allocation model for resource allocation processing, the method further includes pre-training the resource allocation model, specifically including the following steps: Acquire training images, and perform 3D object detection and bounding box construction on the training images to obtain training samples; The training target channel is collected, and the base station resource scheduling algorithm based on traversal search is used to perform resource scheduling processing on the training target channel to obtain the resource scheduling result; The training samples are labeled according to the resource scheduling results to obtain training data. The training data is input into the resource allocation model, and the training results are output. Based on the training results, the parameters of the resource allocation model are adjusted to obtain the trained resource allocation model.

5. A base station resource allocation device, characterized in that, The device includes: The first module is used to acquire communication scene images captured by multiple base stations; The second module is used to perform three-dimensional target detection processing on the communication scene image to obtain detection boxes of multiple target objects; The third module is used to construct the channel space features of the target objects based on the detection boxes of the multiple target objects; The fourth module is used to input the channel space characteristics of the target object into the resource allocation model for resource allocation processing to obtain the resource allocation result; The step of inputting the channel space characteristics of the target object into the resource allocation model for resource allocation processing to obtain the resource allocation result includes: The target object's channel space characteristics are input into the resource allocation model, and the output is a base station number matrix and a base station energy matrix. The base station number matrix is ​​used to represent the number of the base station that the target object needs to connect to, and the base station energy matrix is ​​used to represent the base station energy coefficient corresponding to the base station number. The base station energy coefficient represents the average value of the optimal transmission energy required by the base station to serve all its target objects. The base station number matrix and the base station energy matrix are decoded to obtain the resource allocation scheme between each target object and its corresponding base station; The resource allocation is performed on multiple target objects according to the resource allocation scheme to obtain the resource allocation result; The step of inputting the channel space characteristics of the target object into the resource allocation model and outputting the base station number matrix and the base station energy matrix includes: The size matrix and direction vector are obtained from the channel space features of the target object; The size matrix and the direction vector are respectively input into the resource allocation model, which includes a first convolutional neural network and a second convolutional neural network; The size matrix is ​​convolved using the first convolutional neural network to obtain size features; The direction vector is convolved by the second convolutional neural network to obtain vector features; The size features and the vector features are subjected to element-wise addition and convolution to obtain the base station number matrix and the base station energy matrix; The process of decoding the base station number matrix and the base station energy matrix to obtain the resource allocation scheme for each target object and its corresponding base station includes: The matrix containing the maximum element value is extracted from the base station number matrix, and the matrix containing the maximum element value is determined as the allocation matrix; Based on the allocation matrix, base station allocation is performed on the target objects, and the process returns to the matrix containing the maximum element value extracted from the base station number matrix, until base station allocation is performed on all target objects to obtain the base station allocation scheme; Based on the base station allocation scheme and the base station energy matrix, average transmission energy allocation is performed on each base station to obtain the resource allocation scheme between each target object and the corresponding base station.

6. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method according to any one of claims 1 to 4.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 4.

Images