Wireless communication network performance optimization method
Through deep learning algorithms, the signal strength data in the wireless communication network is analyzed, and the base station power optimization value is automatically recommended, which solves the problems of insufficient network adaptability and limitations in traditional methods, and achieves global network performance optimization and user experience improvement.
Patent Information
- Application Number
- CN202410351584.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-26
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-03-26
AI Technical Summary
Traditional wireless communication network performance optimization methods cannot be dynamically adjusted according to real-time network status and user needs, resulting in insufficient network adaptability and flexibility, limited to local optimization, making it difficult to globally coordinate the entire network performance, affecting the user experience.
By obtaining the signal strength data of each area of the cell to be optimized, using deep learning-based data processing and analysis algorithms to perform timing feature analysis and global signal correlation, the base station power optimization value is automatically recommended, and the accuracy and accuracy of optimization results are improved, and the global realm is achieved. Sex network performance optimization.
It improves the accuracy and efficiency of network performance optimization, meets users' needs for high-speed and stable connections, and reduces the complexity of the optimization process.
Smart Images

Figure CN118118929B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of intelligent optimization, and more specifically, to a method for optimizing performance of a wireless communication network. Background Art
[0002] In the city center, mobile communication networks face challenges such as high user density, obstruction from tall buildings, and signal interference. Users have high demands for data rates and network connection quality, and optimization is needed to improve network performance and user experience.
[0003] However, since traditional wireless communication network performance optimization methods are usually based on static network configuration and parameter adjustment, they cannot be dynamically adjusted according to real-time network status and user needs, resulting in insufficient network adaptability and flexibility. In addition, traditional methods are usually limited to local optimization, that is, only considering the performance optimization of a single network element or a part of the network area, and it is difficult to globally coordinate the performance of the entire network, thus affecting the user experience.
[0004] Therefore, an optimized wireless communication network performance optimization solution is desired. Summary of the invention
[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a method for optimizing the performance of a wireless communication network, which obtains the signal strength data of each area in the cell to be optimized, and uses a data processing and analysis algorithm based on deep learning to perform a timing feature analysis and global signal correlation of the signal strength data of each area in the cell to be optimized, thereby automatically recommending the optimization value of the base station power based on the timing correlation relationship of the signal strength between each area. In this way, the precision and accuracy of the optimization results can be improved, and the complexity of the optimization process can be reduced, so as to achieve global network performance optimization, thereby improving the efficiency and quality of the entire network, and meeting the user's needs for high-speed and stable connections.
[0006] According to one aspect of the present application, a method for optimizing performance of a wireless communication network is provided, which includes:
[0007] Acquire the time series of signal strength data of each area in the cell to be optimized to obtain a sequence of signal strength data time series;
[0008] Regularizing the sequence of the signal strength data time series according to the time dimension to obtain a sequence of local area signal strength time series input vectors;
[0009] The sequence of the local area signal strength time series input vectors is respectively passed through a signal strength time series pattern feature extractor based on a one-dimensional extended convolutional neural model to obtain a sequence of local area signal strength time series associated feature vectors;
[0010] Performing nonlinear correction on the sequence of local area signal strength time series correlation feature vectors to obtain a sequence of corrected local area signal strength time series correlation feature vectors;
[0011] Constructing a network space topology matrix based on the space topology structure of each area in the cell to be optimized;
[0012] Extracting spatial topological features from the network space topology matrix to obtain a network space topological feature matrix;
[0013] The sequence of the corrected local area signal strength time series correlation feature vectors and the network space topology feature matrix are passed through a graph neural network model to obtain the topology global signal strength time series correlation feature;
[0014] Based on the temporal correlation characteristics of the global signal strength of the topology, a base station power control instruction is obtained.
[0015] Compared with the prior art, the wireless communication network performance optimization method provided by the present application obtains the signal strength data of each area in the cell to be optimized, and uses the data processing and analysis algorithm based on deep learning to perform the time series feature analysis and global signal correlation of the signal strength data of each area in the cell to be optimized, so as to automatically recommend the optimization value of the base station power based on the time series correlation relationship of the signal strength between each area. In this way, the precision and accuracy of the optimization results can be improved, the complexity of the optimization process can be reduced, so as to achieve global network performance optimization, thereby improving the efficiency and quality of the entire network and meeting the user's demand for high-speed and stable connection. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.
[0017] Figure 1 Flow chart of a method for optimizing wireless communication network performance according to an embodiment of the present application.
[0018] Figure 2 Schematic diagram of the architecture of a method for optimizing performance of a wireless communication network according to an embodiment of the present application.
[0019] Figure 3The present invention is a flowchart for training the signal strength timing pattern feature extractor based on the one-dimensional extended convolutional neural model, the spatial topology feature extractor based on the convolutional neural network model, the graph neural network model and the decoder-based base station power controller in the wireless communication network performance optimization method according to an embodiment of the present application.
[0020] Figure 4 The present invention is a flowchart of clustering and optimizing the training topology global signal strength timing correlation feature matrix in the wireless communication network performance optimization method according to an embodiment of the present application to obtain an optimized training topology global signal strength timing correlation feature matrix. DETAILED DESCRIPTION
[0021] The embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, but rather these embodiments are provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.
[0022] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.
[0023] In the description of the embodiments of the present disclosure, the term "including" and similar terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.
[0024] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".
[0025] In the city center, mobile communication networks face challenges such as high user density, obstruction from tall buildings, and signal interference. Users have high demands for data rates and network connection quality, and optimization is needed to improve network performance and user experience.
[0026] However, since traditional wireless communication network performance optimization methods are usually based on static network configuration and parameter adjustment, they cannot be dynamically adjusted according to real-time network status and user needs, resulting in insufficient network adaptability and flexibility. In addition, traditional methods are usually limited to local optimization, that is, only considering the performance optimization of a single network element or a part of the network area, and it is difficult to globally coordinate the performance of the entire network, thus affecting the user experience.
[0027] Therefore, in response to the above technical problems, the technical concept of this application is to obtain the signal strength data of each area in the cell to be optimized, and use the data processing and analysis algorithm based on deep learning to perform the time series feature analysis and global signal correlation of the signal strength data of each area in the cell to be optimized, so as to automatically recommend the optimization value of the base station power based on the time series correlation relationship of the signal strength between each area. In this way, the precision and accuracy of the optimization results can be improved, the complexity of the optimization process can be reduced, so as to achieve global network performance optimization, thereby improving the efficiency and quality of the entire network and meeting the user's demand for high-speed and stable connections.
[0028] Figure 1 Flow chart of a method for optimizing wireless communication network performance according to an embodiment of the present application. Figure 2 FIG. 1 is a schematic diagram of the architecture of a wireless communication network performance optimization method according to an embodiment of the present application. Figure 1 and Figure 2 As shown, according to the wireless communication network performance optimization method of the embodiment of the present application, the method includes: S110, obtaining the time series of signal strength data of each area in the cell to be optimized to obtain a sequence of signal strength data time series; S120, regularizing the sequence of signal strength data time series according to the time dimension to obtain a sequence of local area signal strength time series input vectors; S130, passing the sequence of local area signal strength time series input vectors through a signal strength time series pattern feature extractor based on a one-dimensional extended convolution neural model to obtain a sequence of local area signal strength time series correlation feature vectors; S140, The sequence of characteristic vectors is nonlinearly corrected to obtain a sequence of corrected local area signal strength timing correlation characteristic vectors; S150, a network space topology matrix is constructed based on the spatial topological structure of each area in the cell to be optimized; S160, spatial topological features are extracted from the network space topology matrix to obtain a network space topology feature matrix; S170, the sequence of corrected local area signal strength timing correlation characteristic vectors and the network space topology feature matrix are passed through a graph neural network model to obtain a topology global signal strength timing correlation feature; and, S180, a base station power control instruction is obtained based on the topology global signal strength timing correlation feature.
[0029] In step S110, the time series of signal strength data of each area in the cell to be optimized is obtained to obtain a sequence of signal strength data time series. It should be understood that the signal strength data refers to the data used to represent the signal strength level in the communication network. Specifically, in a wireless communication system, signal strength generally refers to the received signal power level, which is an indicator of the strength of the wireless signal during the transmission process. The signal strength of different areas in the cell to be optimized may be different. Knowing the signal strength data of each area can help optimize network performance and improve communication quality and coverage. In particular, in a specific implementation of the present application, the signal strength data can be measured and collected in various ways, such as base station equipment, mobile equipment or special test equipment. Based on this, in the technical solution of the present application, the time series of signal strength data of each area in the cell to be optimized is obtained to obtain a sequence of signal strength data time series. That is, by obtaining the time series of signal strength data in each area, it is possible to analyze the signal strength change trends and temporal correlations between different areas, revealing the potential connections and influences between the various areas, thereby providing a more accurate basis for base station power optimization, so as to better plan and optimize the network, adjust base station deployment and power settings, and improve the efficiency and quality of wireless communication networks.
[0030] In step S120, the sequence of the signal strength data time series is data-regulated according to the time dimension to obtain a sequence of local area signal strength time series input vectors. Accordingly, considering that the signal strength data will fluctuate over time, that is, the signal strength data presents a dynamic fluctuation state over time. Therefore, in order to maintain the time series of the signal strength data and facilitate the subsequent analysis of the change trend of the signal strength data between different time points and the correlation of the signal strength between various regions, in the technical solution of the present application, the sequence of the signal strength data time series is data-regulated according to the time dimension to obtain a sequence of local area signal strength time series input vectors. In this way, the time series characteristics between the signal strength data can be revealed, which helps to understand the law of change of signal strength over time, so that the characteristic information of the signal strength in the time dimension can be more accurately analyzed, and a more effective reference basis can be provided for wireless communication network optimization and decision-making.
[0031] In step S130, the sequence of the local area signal strength time series input vector is respectively passed through the signal strength time series pattern feature extractor based on the one-dimensional extended convolution neural model to obtain a sequence of local area signal strength time series associated feature vectors. It should be understood that, considering that each local area signal strength time series input vector in the sequence of the local area signal strength time series input vector has local implicit feature information in the time dimension, and the one-dimensional convolution neural network (CNN) is a neural network structure specially designed for processing time series data, which can effectively capture local patterns and features in time series data. In addition, there is also a semantic association about signal strength between each local area signal strength local time series feature of different lengths in each local area signal strength time series input vector. However, since the one-dimensional convolution neural network cannot capture the time series feature association information under different time steps, based on this, in the technical solution of the present application, the sequence of the local area signal strength time series input vector is respectively passed through the signal strength time series pattern feature extractor based on the one-dimensional extended convolution neural model to obtain a sequence of local area signal strength time series associated feature vectors. It should be understandable that the one-dimensional expanded convolution neural model introduces a hyperparameter of the expansion rate to control the number of null values 0 in the convolution kernel, so that it can better extract the semantic correlation between the local time series of the local area signal intensity at different time spans for each local area signal intensity time series input vector in the sequence of the local area signal intensity time series input vector, thereby obtaining a more expressive sequence of local area signal intensity time series association feature vectors.
[0032] Specifically, in an embodiment of the present application, the sequence of the local area signal strength time series input vectors is respectively passed through a signal strength time series pattern feature extractor based on a one-dimensional extended convolution neural model to obtain a sequence of local area signal strength time series associated feature vectors, including: each local area signal strength time series input vector in the sequence of the local area signal strength time series input vector is respectively processed by the signal strength time series pattern feature extractor based on the one-dimensional extended convolution neural model using the following extended convolution formula to obtain the sequence of the local area signal strength time series associated feature vectors; wherein the extended convolution formula is:
[0033]
[0034] C i =f(ω·X i:i+j-l +b)
[0035] C=[C1,C2,...,C n-l+1 ]
[0036] Among them, X1,X2,...,X nrepresents each local area signal strength local time series input vector in each local area signal strength time series input vector, X 1:n A cascade vector representing the local time series input vectors of the signal strengths of the respective local areas, Indicates the connection operation, X i:i+j-l X in the local time series input vector representing the signal strength of each local area i ,X i+1 ,...,X i+j-1 , ω and b represent the weight matrix and bias vector respectively, f(·) represents the convolution operation, C i The local area signal strength local time series correlation feature vectors of each local area signal strength time series correlation feature vector in the sequence of the local area signal strength time series correlation feature vectors are represented, [·,·] represents vector splicing, and C is the local area signal strength time series correlation feature vector.
[0037] In step S140, the sequence of the local area signal strength time series correlation feature vectors is nonlinearly corrected to obtain a sequence of the corrected local area signal strength time series correlation feature vectors. Accordingly, considering that each local area signal strength data may be affected by various nonlinear factors during the acquisition and transmission process, such as multipath effect, nonlinear amplifier and channel distortion, thereby generating a variety of interferences and noises, and these noises may cause uncertainty and fluctuations in the signal strength of each local area. Based on this, in the technical solution of the present application, the sequence of the local area signal strength time series correlation feature vectors is nonlinearly corrected to obtain a sequence of the corrected local area signal strength time series correlation feature vectors. Specifically, by performing nonlinear correction on each local area signal strength time series correlation feature vector, the characteristics of the low signal-to-noise ratio signal strength data can be enhanced, and the influence of noise on the characteristics can be reduced, so as to better adapt to the nonlinear relationship in each local area signal strength data, and capture the complex signal strength time series correlation characteristics in each local area signal strength time series correlation feature vector, so as to improve the expression ability and discrimination of each local area signal strength time series correlation feature vector after correction.
[0038] Specifically, in an embodiment of the present application, performing nonlinear correction on the sequence of local area signal strength time series correlation feature vectors to obtain a corrected sequence of local area signal strength time series correlation feature vectors includes: performing nonlinear correction on the sequence of local area signal strength time series correlation feature vectors using the following correction formula to obtain the corrected sequence of local area signal strength time series correlation feature vectors; wherein the correction formula is:
[0039]
[0040] Among them, v i is the eigenvalue of the ith position of each local area signal strength time series correlation feature vector in the sequence of local area signal strength time series correlation feature vectors, L i is the eigenvalue of the ith position of each corrected local area signal strength time series correlation feature vector in the sequence of the corrected local area signal strength time series correlation feature vectors, and A, B, C and D are corresponding adjustment parameters.
[0041] In step S150, a network space topology matrix is constructed based on the space topology structure of each area in the cell to be optimized. It should be understood that, considering that the network space topology matrix is a data structure, it can be used to describe the space topology structure between each area in the cell to be optimized, that is, through the elements and structures in the matrix, the connection relationship, distance relationship and communication path between different areas in the cell to be optimized can be clearly represented to help understand the spatial layout of each area in the network. Specifically, based on the network space topology matrix, decision-making for network planning, deployment and optimization can be made for the cell to be optimized, helping to select the optimal base station power value, thereby improving the overall operation efficiency of the network.
[0042] In step S160, the network space topology matrix is subjected to spatial topology feature extraction to obtain a network space topology feature matrix. Specifically, in an embodiment of the present application, the network space topology matrix is subjected to spatial topology feature extraction to obtain a network space topology feature matrix, including: passing the network space topology matrix through a spatial topology feature extractor based on a convolutional neural network model to obtain the network space topology feature matrix. Accordingly, considering that the network space topology matrix has implicit feature information in space, and the convolutional neural network performs well in processing spatial data and can effectively extract implicit features from the data. Therefore, in the technical solution of the present application, the network space topology matrix is passed through a spatial topology feature extractor based on a convolutional neural network model to capture the complex spatial features in the network space topology matrix, such as connectivity, distance relationships and topological structures between regions, thereby obtaining a network space topology feature matrix.
[0043] In step S170, the sequence of the corrected local area signal strength timing correlation feature vectors and the network space topology feature matrix are passed through a graph neural network model to obtain the topology global signal strength timing correlation feature. It should be understood that, considering that the sequence of the corrected local area signal strength timing correlation feature vectors represents the signal strength timing characteristics of each local area, the network space topology feature matrix represents the distance between each two areas. Therefore, in order to comprehensively integrate the timing correlation features of the signal strength of each area in the cell to be optimized to obtain a comprehensive global signal strength, so as to make the best recommendation for the base station power of the cell to be optimized later, in the technical solution of the present application, the sequence of the corrected local area signal strength timing correlation feature vectors and the network space topology feature matrix are passed through a graph neural network model to obtain the topology global signal strength timing correlation feature matrix. It is worth mentioning that the graph neural network is a deep learning model specially designed for processing graph structure data. Specifically, the sequence of the corrected local area signal strength time series correlation feature vectors reflects the signal strength characteristics of the nodes, and the network space topology feature matrix reflects the feature correlation between nodes, which can be regarded as a graph structure. By integrating the local area signal strength time series correlation characteristics and the network space topology characteristics into GNN, a more comprehensive and integrated feature representation can be obtained, so as to better capture the interaction information between nodes and obtain more accurate topology global signal strength time series correlation characteristics, thereby optimizing the cell network performance and improving the base station power prediction accuracy.
[0044] Specifically, in an embodiment of the present application, the sequence of the corrected local area signal strength timing correlation feature vectors and the network space topology feature matrix are passed through a graph neural network model to obtain a topology-wide signal strength timing correlation feature, including: the graph neural network processes the sequence of the corrected local area signal strength timing correlation feature vectors and the network space topology feature matrix through learnable neural network parameters to obtain a topology-wide signal strength timing correlation feature matrix containing local area network space topology features and each local area signal strength timing feature information as the topology-wide signal strength timing correlation feature.
[0045] In step S180, the monitoring result is determined based on the essential semantic pattern feature. Specifically, in an embodiment of the present application, based on the topology global signal strength timing correlation feature, a base station power control instruction is obtained, including: the topology global signal strength timing correlation feature matrix is passed through a decoder-based base station power controller to obtain a base station power control instruction, and the base station power control instruction is used to represent the optimized recommended value of the base station power. That is, the topology global signal strength timing correlation feature obtained by processing the graph neural network model is used to perform decoding regression, so as to automatically recommend the optimized value of the base station power based on the timing correlation of the global signal strength data based on each area in the cell to be optimized. In this way, the precision and accuracy of the optimization results can be improved, and the complexity of the optimization process can be reduced, so as to achieve global network performance optimization, thereby improving the efficiency and quality of the entire network and meeting the user's needs for high speed and stable connection.
[0046] It is worth mentioning that those skilled in the art should be aware that before applying a deep neural network model for inference, the deep neural network model must be trained so that the deep neural network can implement specific functional capabilities.
[0047] Specifically, in an embodiment of the present application, a training step is also included: for training the signal strength timing pattern feature extractor based on the one-dimensional extended convolutional neural model, the spatial topology feature extractor based on the convolutional neural network model, the graph neural network model and the decoder-based base station power controller.
[0048] Figure 3 The present invention is a flowchart of training the signal strength timing pattern feature extractor based on the one-dimensional extended convolution neural model, the spatial topology feature extractor based on the convolution neural network model, the graph neural network model and the decoder-based base station power controller in the wireless communication network performance optimization method according to an embodiment of the present application. Figure 3As shown, the training step includes: S210, obtaining training data, the training data including a sequence of training signal strength data time series in the cell to be optimized, and the true value of the optimization recommendation of the base station power; S220, regularizing the sequence of the training signal strength data time series according to the time dimension to obtain a sequence of training local area signal strength timing input vectors; S230, respectively passing the sequence of training local area signal strength timing input vectors through the signal strength timing pattern feature extractor based on the one-dimensional extended convolution neural model to obtain a sequence of training local area signal strength timing correlation feature vectors; S240, performing nonlinear correction on the sequence of training local area signal strength timing correlation feature vectors to obtain a sequence of corrected training local area signal strength timing correlation feature vectors; S250, constructing a training network space topology matrix based on the spatial topological structure of each area in the cell to be optimized; S260, passing the training network space topology matrix through the A spatial topology feature extractor based on a convolutional neural network model is used to obtain a training network spatial topology feature matrix; S270, the sequence of the corrected training local area signal strength timing correlation feature vectors and the training network spatial topology feature matrix are passed through the graph neural network model to obtain a training topology global signal strength timing correlation feature matrix; S280, clustering optimization is performed on the training topology global signal strength timing correlation feature matrix to obtain an optimized training topology global signal strength timing correlation feature matrix; S290, the optimized training topology global signal strength timing correlation feature matrix is passed through the decoder-based base station power controller to obtain a decoding loss function value; and, S300, based on the decoding loss function value and through gradient descent back propagation, the signal strength timing pattern feature extractor based on the one-dimensional extended convolutional neural model, the spatial topology feature extractor based on the convolutional neural network model, the graph neural network model and the decoder-based base station power controller are trained.
[0049] Specifically, in step S280, cluster optimization is performed on the training topology global signal strength temporal correlation feature matrix to obtain an optimized training topology global signal strength temporal correlation feature matrix. Figure 4The present invention is a flowchart of clustering and optimizing the training topology global signal strength timing correlation feature matrix in the wireless communication network performance optimization method according to an embodiment of the present application to obtain an optimized training topology global signal strength timing correlation feature matrix. It should be understood that in the above technical solution, each corrected training local area signal strength time-series correlation feature vector in the sequence of corrected training local area signal strength time-series correlation feature vectors respectively expresses the time-series correlation characteristics of the signal strength data of the corresponding area through nonlinear correction. In this way, after the sequence of corrected training local area signal strength time-series correlation feature vectors and the training network spatial topology feature matrix pass through the graph neural network model, the signal strength time-series correlation characteristics of each area can be topologically correlated based on the spatial topology of each area. However, considering the relative independence of the feature representations of each training topology global signal strength time-series correlation feature vector corresponding to the corrected training local area signal strength time-series correlation feature vector of the training topology global signal strength time-series correlation feature matrix, such as the row feature vectors, when the training topology global signal strength time-series correlation feature matrix is classified as a whole through a classifier, the local feature distribution discreteness of the training topology global signal strength time-series correlation feature matrix may lead to difficulties in convergence of regression values when decoding regression is performed through a decoder, thereby affecting the training speed of the decoder and the accuracy of the decoded value finally obtained. Based on this, in the technical solution of the present application, cluster optimization is performed on the training topology global signal strength temporal correlation feature matrix to obtain an optimized training topology global signal strength temporal correlation feature matrix.
[0050] Specifically, in the embodiments of the present application, Figure 4 As shown, the training topology global signal strength timing correlation feature matrix is clustered and optimized to obtain an optimized training topology global signal strength timing correlation feature matrix, including: S310, clustering each eigenvalue in the training topology global signal strength timing correlation feature matrix based on the distance between eigenvalues to obtain a cluster feature set; and, S320, based on the cluster feature set, optimizing the training topology global signal strength timing correlation feature matrix based on the in-class and out-of-class representations after clustering to obtain the optimized training topology global signal strength timing correlation feature matrix.
[0051] More specifically, in an embodiment of the present application, based on the clustering feature set, the training topology global signal strength timing correlation feature matrix is optimized based on the in-class and out-of-class characterization after clustering to obtain the optimized training topology global signal strength timing correlation feature matrix, including: based on the clustering feature set, the training topology global signal strength timing correlation feature matrix is optimized based on the in-class and out-of-class characterization after clustering to obtain the optimized training topology global signal strength timing correlation feature matrix according to the following optimization formula; wherein the optimization formula is:
[0052]
[0053] Among them, f is the eigenvalues of the training topology global signal strength timing correlation feature matrix, n is the number of feature sets corresponding to the training topology global signal strength timing correlation feature matrix, k is the number of cluster features, C represents the cluster feature set, and f' is the eigenvalues of the optimized training topology global signal strength timing correlation feature matrix.
[0054] Specifically, by taking the in-class features and out-of-class features of the training topology global signal strength timing correlation feature matrix as different instance roles to perform class instance description based on clustering ratio distribution, and introducing clustering response history based on in-class and out-of-class dynamic context, a coordinated global perspective is maintained for the in-class distribution and out-of-class distribution of the overall features of the training topology global signal strength timing correlation feature matrix, so that the optimized feature clustering operation of the training topology global signal strength timing correlation feature matrix can maintain a consistent response of the in-class and out-of-class features, so that the regression convergence path based on feature clustering remains consistent during the decoding regression process, and the convergence effect of the regression value of the training topology global signal strength timing correlation feature matrix is improved. In this way, the precision and accuracy of the optimization results can be improved, the complexity of the optimization process can be reduced, so as to achieve global network performance optimization, thereby improving the efficiency and quality of the entire network and meeting the user's needs for high-speed and stable connections.
[0055] In summary, the wireless communication network performance optimization method based on the embodiment of the present application is explained, which obtains the signal strength data of each area in the cell to be optimized, and uses the data processing and analysis algorithm based on deep learning to perform the timing feature analysis and global signal correlation of the signal strength data of each area in the cell to be optimized, so as to automatically recommend the optimization value of the base station power based on the timing correlation relationship of the signal strength between each area. In this way, the precision and accuracy of the optimization results can be improved, and the complexity of the optimization process can be reduced, so as to achieve global network performance optimization, thereby improving the efficiency and quality of the entire network and meeting the user's needs for high-speed and stable connections.
[0056] The foregoing is merely an example of the principles of the present disclosure, and various modifications may be made by those skilled in the art without departing from the scope of the present disclosure. The above embodiments are presented for the purpose of illustration and not limitation. The present disclosure may also take many forms other than those explicitly described herein. Therefore, it is emphasized that the present disclosure is not limited to the explicitly disclosed methods and apparatus, but is intended to include variations and modifications within the spirit and scope of the appended claims.
Claims
1. A method for optimizing wireless communication network performance, characterized in that: include: Acquire the time series of signal strength data of each area in the cell to be optimized to obtain a sequence of signal strength data time series; Regularizing the sequence of the signal strength data time series according to the time dimension to obtain a sequence of local area signal strength time series input vectors; The sequence of the local area signal strength time series input vectors is respectively passed through a signal strength time series pattern feature extractor based on a one-dimensional extended convolutional neural model to obtain a sequence of local area signal strength time series associated feature vectors; Performing nonlinear correction on the sequence of local area signal strength time series correlation feature vectors to obtain a sequence of corrected local area signal strength time series correlation feature vectors; Constructing a network space topology matrix based on the space topology structure of each area in the cell to be optimized; Extracting spatial topological features from the network space topology matrix to obtain a network space topological feature matrix; The sequence of the corrected local area signal strength time series correlation feature vectors and the network space topology feature matrix are passed through a graph neural network model to obtain the topology global signal strength time series correlation feature; Based on the temporal correlation characteristics of the global signal strength of the topology, a base station power control instruction is obtained; The sequence of the local area signal strength time series input vectors is respectively passed through a signal strength time series pattern feature extractor based on a one-dimensional extended convolution neural model to obtain a sequence of local area signal strength time series associated feature vectors, including: each local area signal strength time series input vector in the sequence of the local area signal strength time series input vector is respectively processed by the signal strength time series pattern feature extractor based on a one-dimensional extended convolution neural model using the following extended convolution formula to obtain a sequence of the local area signal strength time series associated feature vectors; Wherein, the extended convolution formula is: C i =f(ω·X i:i+j-l +b) C=[C1,C2,...,C n-l+1 ] Among them, X1,X2,...,X n Respectively represent each local area signal strength local time series input vector in each of the local area signal strength time series input vectors, X1: n A cascade vector representing the local time series input vectors of the signal strengths of the respective local areas, Indicates the connection operation, X i:i+j-l X in the local time series input vector representing the signal strength of each local area i ,X i+1 ,...,X i+j-1 , ω and b represent the weight matrix and bias vector respectively, f(·) represents the convolution operation, C i represents each local area signal strength local time series correlation feature vector of each local area signal strength time series correlation feature vector in the sequence of the local area signal strength time series correlation feature vector, [·,·] represents vector concatenation, and C is the local area signal strength time series correlation feature vector; The sequence of local area signal strength time series correlation feature vectors is subjected to nonlinear correction to obtain a sequence of corrected local area signal strength time series correlation feature vectors, comprising: performing nonlinear correction on the sequence of local area signal strength time series correlation feature vectors using the following correction formula to obtain a sequence of corrected local area signal strength time series correlation feature vectors; Wherein, the correction formula is: Among them, v i is the eigenvalue of the ith position of each local area signal strength time series correlation feature vector in the sequence of local area signal strength time series correlation feature vectors, L i is the eigenvalue of the ith position of each corrected local area signal strength time series correlation feature vector in the sequence of the corrected local area signal strength time series correlation feature vectors, and A, B, C and D are corresponding adjustment parameters.
2. The wireless communication network performance optimization method according to claim 1, characterized in that: Performing spatial topological feature extraction on the network space topology matrix to obtain the network space topology feature matrix includes: passing the network space topology matrix through a spatial topological feature extractor based on a convolutional neural network model to obtain the network space topology feature matrix.
3. The method for optimizing wireless communication network performance according to claim 2, characterized in that: The sequence of the corrected local area signal strength timing correlation feature vectors and the network space topology feature matrix are passed through a graph neural network model to obtain a topology-wide signal strength timing correlation feature, including: the graph neural network processes the sequence of the corrected local area signal strength timing correlation feature vectors and the network space topology feature matrix through learnable neural network parameters to obtain a topology-wide signal strength timing correlation feature matrix containing local area network space topology features and each local area signal strength timing feature information as the topology-wide signal strength timing correlation feature.
4. The method for optimizing wireless communication network performance according to claim 3, characterized in that: Based on the topology global signal strength timing correlation characteristics, a base station power control instruction is obtained, including: passing the topology global signal strength timing correlation characteristic matrix through a decoder-based base station power controller to obtain a base station power control instruction, and the base station power control instruction is used to represent the optimized recommended value of the base station power.
5. The method for optimizing wireless communication network performance according to claim 4, characterized in that: It also includes a training step: for training the signal strength timing pattern feature extractor based on the one-dimensional extended convolutional neural model, the spatial topology feature extractor based on the convolutional neural network model, the graph neural network model and the decoder-based base station power controller.
6. The method for optimizing wireless communication network performance according to claim 5, characterized in that: The training step comprises: Acquire training data, the training data including a sequence of training signal strength data time series in the cell to be optimized, and a true value of the optimization recommendation of the base station power; Regularizing the sequence of the training signal strength data time series according to the time dimension to obtain a sequence of training local area signal strength time series input vectors; Passing the sequence of the training local area signal strength time series input vectors through the signal strength time series pattern feature extractor based on the one-dimensional extended convolution neural model to obtain a sequence of training local area signal strength time series associated feature vectors; Performing nonlinear correction on the sequence of training local area signal strength time series associated feature vectors to obtain a sequence of corrected training local area signal strength time series associated feature vectors; Constructing a training network spatial topology matrix based on the spatial topology structure of each area in the cell to be optimized; Passing the training network spatial topology matrix through the spatial topology feature extractor based on the convolutional neural network model to obtain a training network spatial topology feature matrix; The sequence of the corrected training local area signal strength time series correlation feature vectors and the training network space topology feature matrix are passed through the graph neural network model to obtain a training topology global signal strength time series correlation feature matrix; Performing cluster optimization on the training topology global signal strength temporal correlation feature matrix to obtain an optimized training topology global signal strength temporal correlation feature matrix; Passing the optimized training topology global signal strength time series correlation feature matrix through the decoder-based base station power controller to obtain a decoding loss function value; The signal strength timing pattern feature extractor based on the one-dimensional extended convolutional neural model, the spatial topology feature extractor based on the convolutional neural network model, the graph neural network model and the decoder-based base station power controller are trained based on the decoding loss function value and through back propagation of gradient descent.
7. The method for optimizing wireless communication network performance according to claim 6, characterized in that: Cluster optimization is performed on the training topology global signal strength time series correlation feature matrix to obtain an optimized training topology global signal strength time series correlation feature matrix, including: Clustering each eigenvalue in the training topology global signal strength time series correlation feature matrix based on the distance between eigenvalues to obtain a cluster feature set; Based on the clustering feature set, the training topology global signal strength temporal correlation feature matrix is optimized based on in-class and out-of-class representations after clustering to obtain the optimized training topology global signal strength temporal correlation feature matrix.
Citation Information
Patent Citations
Mine prospecting target area optimization system and method based on graph neural network
CN116665067A
Flow management method and system for Internet of Things
CN117201410A
Communication network recovery method and system based on unmanned helicopter
CN117459981A
Park network management system and method based on optical network technology
CN117499821A
Mobile terminal communication stability test method and system
CN117580090A
