Antenna downtilt angle optimization method and system based on traffic prediction and reinforcement learning

Abstract

The application discloses an antenna downtilt angle optimization method and system based on traffic prediction and reinforcement learning, and relates to the technical field of wireless communication. A space-time event cube data set of network traffic data of base stations in a target area is constructed; a traffic prediction model is obtained according to the space-time event cube data set; future traffic values of each base station in the target area are predicted through the traffic prediction model; and after clustering analysis of whether each base station participates in antenna downtilt angle adjustment is performed according to the future traffic values and three-dimensional position information of the base stations, the base stations that need to perform antenna downtilt angle adjustment are determined. After accurate network traffic prediction is obtained, clustering analysis of antenna downtilt angle adjustment is performed based on predicted traffic and geographical positions of users, and antenna downtilt angles among the base stations are collaboratively optimized and adjusted through a deep reinforcement learning algorithm, so that rational scheduling and utilization of mobile communication network resources are effectively realized.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to a method and system for optimizing antenna downtilt angles based on traffic prediction and reinforcement learning. Background Art

[0002] With the rapid development of mobile communication technology, users' demand for high-speed data transmission and seamless coverage continues to increase. Mobile services are dynamic in time and space, and various events can affect mobile data traffic, leading to tidal effects in base station network traffic. As a key component of mobile communication systems, antenna downtilt settings directly impact base station coverage, signal quality, and network capacity. A reasonable antenna downtilt not only optimizes signal coverage but also effectively reduces inter-cell interference, improving overall network performance. Antenna downtilt adjustment is a key research topic in wireless communications. It aims to expand or reduce cell coverage, improve communication service quality, and balance network performance between different base stations by rationally adjusting wireless network parameters. The goal is to improve base station resource utilization, achieve reliable transmission, and balance network capacity. This process requires comprehensive consideration of multiple factors, including user communication needs, the base station's radio environment, and the user's geographic location. Traditional antenna downtilt settings rely primarily on engineers' experience and field testing. However, with the expansion of network scale and increasing complexity, this approach has become inadequate for efficient and accurate network optimization. Therefore, scientific and automated optimization of antenna downtilt has become a key research topic in mobile communications.

[0003] Properly adjusting antenna downtilt not only improves the user experience but also effectively reduces operating costs, ensuring the stability and efficient operation of mobile communication networks. Furthermore, this optimization helps networks adapt to growing communication demands and the introduction of new technologies. Common strategies for antenna downtilt include simulation-based optimization, measurement-based optimization, machine learning-based optimization, multi-objective optimization, and dynamic optimization. Using these strategies individually or in combination, traffic can be dynamically allocated based on network scale, performance requirements, and application scenarios, ensuring efficient, stable, and reliable operation of mobile communication networks.

[0004] In mobile communication networks, various antenna downtilt optimization strategies are employed, all with the common goal of achieving an optimal balance between network performance, availability, and resource utilization. These strategies include simulation-based approaches, which leverage wireless propagation models and network simulation tools to simulate the impact of different downtilt angles on signal coverage and interference; measurement-based approaches, which analyze actual network conditions through drive testing and network performance metrics to guide downtilt adjustment; machine learning-based approaches, which build data-driven prediction models and apply intelligent optimization algorithms to find the optimal configuration; multi-objective optimization strategies, which balance coverage, system capacity, and interference management to find the overall optimal solution; and dynamic optimization methods, which dynamically adjust downtilt angles based on real-time user distribution and environmental changes, achieving adaptive network optimization. Applying these strategies individually or in combination can effectively improve mobile communication network performance and meet growing communication demands.

[0005] Currently, traffic prediction in base station networks relies primarily on simple historical data analysis, failing to fully consider the complex and ever-changing spatial and event characteristics of reality. The limitations of this approach lead to low traffic prediction accuracy. Furthermore, while methods for optimizing antenna downtilt angles based on reinforcement learning exist, these methods are less than ideal due to a lack of accurate traffic prediction support. Therefore, effectively optimizing antenna downtilt angles in mobile communication networks has become a critical and pressing issue. Summary of the Invention

[0006] In view of the defects in the prior art, the present invention solves the technical problem of how to effectively adjust the antenna downtilt angle of a mobile communication network.

[0007] To achieve the above objectives, in a first aspect, an embodiment of the present application provides an antenna downtilt angle optimization method based on traffic prediction and reinforcement learning, the method comprising the following steps:

[0008] Construct a spatiotemporal event cube dataset of network traffic data of base stations in the target area;

[0009] In the same dimension, the spatial, temporal, and event data in the spatiotemporal event cube dataset are fused; corresponding features are extracted in the temporal, spatial, and event domains, respectively. The extracted features are then cross-fused to achieve multi-domain information fusion and obtain a traffic prediction model;

[0010] The traffic prediction model predicts the future traffic value of each base station in the target area. Based on the future traffic value and three-dimensional location information of each base station, a cluster analysis is performed on the base stations to determine whether they will participate in antenna downtilt adjustment. The base stations that need antenna downtilt adjustment are then determined.

[0011] According to the future traffic value of the base station for which antenna downtilt adjustment is required, an antenna downtilt adjustment plan of the current base station is calculated to complete the adjustment of the antenna downtilt.

[0012] In conjunction with the first aspect, in one embodiment, the process of constructing a spatiotemporal event cube dataset of network traffic data of base stations in a target area includes:

[0013] Set the target area and network traffic data collection period;

[0014] Collect network traffic data of base stations in the target area according to the collection period;

[0015] Based on the time, space, event dimensions and flow values ​​of network traffic data, a spatiotemporal event cube dataset of network traffic data is constructed.

[0016] In combination with the first aspect, in one embodiment, after the cube dataset is constructed, the following steps are further included: slicing the cube dataset in the time domain and classifying it through a classification algorithm to obtain a cube dataset that meets the requirements.

[0017] In conjunction with the first aspect, in one embodiment, the process of acquiring the traffic prediction model includes:

[0018] Perform full connection layer processing on the data in the spatiotemporal event cube data to achieve the fusion of space, time and event data in the same dimension;

[0019] In the spatial domain, the trend feature extraction module and node feature extraction module with attention mechanism are used to perform data correlation analysis to obtain spatial domain features.

[0020] In the time domain, we use the trend feature extraction module, period feature extraction module, and noise feature extraction module with an attention mechanism to perform correlation analysis on time domain data and obtain time domain features.

[0021] In the event domain, data association analysis is performed through the trend feature extraction module and node feature extraction module with attention mechanism to obtain event domain features;

[0022] By cross-fusion, the spatial domain features, time domain features and event domain features are integrated and fused to obtain the corresponding three vectors;

[0023] After the three vectors pass through two layers of identical multi-domain feature extraction and cross-fusion operations, they pass through a two-layer fully connected network to obtain the traffic prediction results.

[0024] In combination with the first aspect, in one embodiment, the process of obtaining a traffic prediction result by subjecting the three vectors to two layers of identical multi-domain feature extraction and cross-fusion operations, and then passing them through a two-layer fully connected network includes: subjecting the three vectors to two layers of identical multi-domain feature extraction and cross-fusion operations to obtain three fused vectors; flattening the three fused vectors, and fusing and extracting features through two layers of fully connected networks to obtain a traffic prediction result.

[0025] With reference to the first aspect, in one embodiment, the process of determining a base station requiring antenna downtilt adjustment includes:

[0026] Sort the base stations in a specified order according to their future traffic values, and determine a designated base station as a cluster center based on the sorting result;

[0027] The base stations are clustered and analyzed using a clustering algorithm. New cluster centers are calculated based on the average distance within the cluster. The algorithm is iterated continuously until convergence, and the base stations that require antenna downtilt adjustment are obtained.

[0028] In conjunction with the first aspect, in one embodiment, the average distance is defined in three-dimensional space, and the specific calculation formula is:

[0029]

[0030] in, Indicates the The three-dimensional space distance of each cell, Indicates the traffic volume of the current cell. Indicates the ratio of the number of users in the middle and lower layers of the current cell's three-dimensional space to the total number of users in the cell. Indicates the ratio of the number of users in the three-dimensional space of the current cell to the total number of users in the cell; and Obtained according to the corresponding three-dimensional position information.

[0031] In conjunction with the first aspect, in one embodiment, the state information of the deep reinforcement learning algorithm is:

[0032]

[0033] in, Represents the state information of the deep reinforcement learning algorithm, Indicates the number of cells in the area of ​​interest, Indicates the The predicted traffic volume of each cell, Indicates the The proportion of low- and middle-tier users in a cell to the total number of users, Indicates the The ratio of the number of users in the vicinity of a cell to the total number of users;

[0034] The actions of the deep reinforcement learning algorithm are:

[0035]

[0036] in, represents the action of the deep reinforcement learning algorithm, Indicates the The antenna downtilt angle value of each cell.

[0037] In combination with the first aspect, in one embodiment, the process of determining the three-dimensional position information includes: determining the three-dimensional coordinate position of the user according to the two-dimensional coordinate position of the user.

[0038] In a second aspect, an embodiment of the present application provides an antenna downtilt angle optimization system based on traffic prediction and reinforcement learning, which is used to implement the steps of the method described in the first aspect.

[0039] Compared with the prior art, the advantages of the present invention are:

[0040] The proposed spatial domain neural network, based on the attention mechanism, can accurately extract the spatial features of location information, breaking through geographical barriers and capturing subtle spatial correlations. The introduction of event domain feature extraction makes the model more realistic, accurately capturing the impact of events on traffic flow, and demonstrating exceptionally high accuracy when responding to temporary emergencies.

[0041] At the same time, the present invention combines cross fusion with the fully connected layer to perform fusion synchronously during the feature extraction process, avoiding the performance loss that may be caused by directly performing low-dimensional temporal feature processing after extracting high-dimensional spatial features, thereby improving the robustness of the model.

[0042] On this basis, after obtaining accurate network traffic prediction, the present invention performs cluster analysis on antenna downtilt adjustment based on the predicted traffic and the user's geographic location, and collaboratively optimizes and adjusts the antenna downtilt angles between base stations through a deep reinforcement learning algorithm, thereby effectively realizing the rational scheduling and utilization of mobile communication network resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0044] Figure 1This is a framework flow chart of an antenna downtilt angle optimization method based on traffic prediction and reinforcement learning in an embodiment of the present invention;

[0045] Figure 2 Detailed flowchart of the antenna downtilt angle optimization method based on traffic prediction and reinforcement learning in an embodiment of the present invention;

[0046] Figure 3 Schematic diagram of the hardware structure of the antenna downtilt optimization device based on traffic prediction and reinforcement learning in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0048] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, combined, or partially merged, so the actual execution order may vary depending on the actual situation.

[0049] On the first aspect, an embodiment of the present application provides an antenna downtilt angle optimization method based on traffic prediction and reinforcement learning, which belongs to the cross-technology of the Chinese invention patent "A load balancing optimization method and device based on traffic prediction and reinforcement learning" with publication number CN117479231A.

[0050] See also Figure 1 As shown, the method includes the following steps:

[0051] Step A: Obtain the location information and periodic network traffic data of base stations in the target area, and then construct a spatiotemporal event cube dataset of the network traffic data of base stations in the target area.

[0052] Step B: Within the same dimension, the space, time, and event data in the spatiotemporal event cube dataset are fused; corresponding features are extracted in the time domain, space domain, and event domain respectively, and the extracted features are cross-fused to achieve the fusion of multi-domain information and obtain the traffic prediction model.

[0053] Step C: The future traffic value of each base station in the target area is predicted using the traffic prediction model. Based on the future traffic value of each base station and the three-dimensional location information of the user, a cluster analysis is performed on the base stations to determine whether they will participate in antenna downtilt adjustment, and the base stations that need to undergo antenna downtilt adjustment are determined.

[0054] Step D: Based on the future traffic value of the base station for which antenna downtilt adjustment is required, a deep reinforcement learning algorithm (such as DQN, DDPG, PPO, A3C, SAC) is used to calculate the antenna downtilt adjustment plan for the current base station to complete the antenna downtilt adjustment.

[0055] In one embodiment, the process of step A includes:

[0056] Step A01: Set the target area (area of ​​interest or designated area) and the collection period of network traffic data (e.g., 15 seconds).

[0057] Step A02: Obtain event data of a specific location in the future through an information website; for example, relevant event information of a location at a specific time point in the future can be obtained through a weather website, a map service, or a ticket purchasing website.

[0058] Step A03: Collect network traffic data of base stations in the target area according to a collection period, and save it through a time series database and a storage device.

[0059] Step A04: Construct a spatiotemporal event cube dataset of the network traffic data based on the time, space, event dimensions and traffic value of the network traffic data.

[0060] Furthermore, after constructing the cube dataset, step A04 also includes the following steps: slicing the spatiotemporal event cube dataset in the time domain and classifying the dataset using the SVM (support vector machine) classification algorithm; and removing contaminated data that does not meet the classification requirements through the data clipping method, thereby obtaining a "pure" cube dataset.

[0061] Specifically, the data will be sliced ​​by time domain, such as weekly units, and then the SVM classification algorithm will be used to divide the data into two categories: pure data and contaminated data. The contaminated data that does not meet the requirements (such as pulsed data or severely abnormal data) will be deleted through trimming methods.

[0062] In one embodiment, the process of step B includes:

[0063] Step B01: Perform full connection layer processing on the data in the spatiotemporal event cube data to achieve the fusion of space, time and event data in the same dimension.

[0064] This paper analyzes and predicts network traffic data using a multi-domain feature extraction and cross-fusion approach. Clean data is first processed through a fully connected layer, initially fusing data from different time points to form a four-dimensional feature space of time, space, events, and traffic. Specifically, a spatiotemporal event cube dataset records traffic values ​​corresponding to a specific time, spatial location, and event information.

[0065] Step B02: In the spatial domain, use the trend feature extraction module and node feature extraction module with an attention mechanism to perform data correlation analysis to obtain spatial domain features. For example, the traffic characteristics of the CBD and Zhongguancun areas during rush hour may be similar, and the module with an attention mechanism can help capture these correlations.

[0066] The trend feature extraction module of the attention mechanism enhances the importance of data with fixed trends by assigning higher weights to them. For example, when predicting traffic flow in a certain community at 10:00 AM, traffic flow between 8:00 AM and 10:00 AM, traffic flow during the same time period yesterday, and traffic flow during the same time period last week are given greater weight. In the four-dimensional feature space, the geographic location is first fixed, and the other three-dimensional information is formed into an index table and input into the attention module for efficient feature extraction. Spatial domain information is then further extracted using a graph convolutional neural network and a convolutional network. Without the attention mechanism, the large amount of redundant information may slow the learning process and be unable to effectively process very long time series data. The extracted spatial domain features are then fed into the time and event domains for fusion. Without cross-domain fusion, the integration of low-dimensional information is insufficient, potentially resulting in reduced prediction accuracy.

[0067] Step B03: In the time domain, by using the trend feature extraction module, period feature extraction module and noise feature extraction module with attention mechanism, correlation analysis of time domain data is performed to obtain time domain features.

[0068] Similar to the trend feature extraction module in the spatial domain, trend data in the temporal domain is also enhanced through the attention mechanism. However, network traffic in the temporal domain often exhibits periodicity, necessitating the introduction of a periodic feature extraction module (for example, data from Monday to Friday constitutes a period), along with a noise feature extraction module to remove the effects of periodicity. Information is then further extracted through graph convolutional neural networks and convolutional networks. The extracted temporal features are then passed as input to the spatial and event domains. Without the periodic attention mechanism, the periodic characteristics of network traffic may not be effectively captured, resulting in reduced prediction accuracy.

[0069] Step B04: In the event domain, data association analysis is performed through the trend feature extraction module and the node feature extraction module with the attention mechanism to obtain event domain features.

[0070] Event domain feature extraction focuses on analyzing factors such as weather conditions, traffic flow, large-scale events like concerts, and holidays. Attention mechanisms are used to enhance the extraction of trend and node features, similar to the operations in the spatial domain. After further processing by graph convolutional neural networks and convolutional networks, event domain features are input into the spatial and temporal domains.

[0071] Step B05: Integrate and fuse the spatial domain features, temporal domain features, and event domain features through cross-fusion to obtain the corresponding three vectors (spatial domain vector, temporal domain vector, and event domain vector).

[0072] Step B06: After the three vectors pass through two layers of the same multi-domain feature extraction and cross-fusion operations, they are passed through a two-layer fully connected network to obtain the traffic prediction results. At this time, the training of the traffic prediction model is completed and the traffic prediction model is obtained.

[0073] Specifically, the process of step B06 includes:

[0074] The three vectors are subjected to two layers of the same multi-domain feature extraction and cross-fusion operations to obtain three fused vectors: that is, the first layer performs multi-domain feature extraction and cross-fusion operations on the three vectors to obtain three new vectors, and the second layer performs multi-domain feature extraction and cross-fusion operations on the three new vectors to obtain three fused vectors (the spatial domain fusion vector corresponding to the spatial domain vector, the time domain fusion vector corresponding to the time domain vector, and the event domain fusion vector corresponding to the event domain vector).

[0075] The three fusion vectors are flattened and fused through a two-layer fully connected network to extract features to obtain the traffic prediction result.

[0076] It is important to note that after training the spatiotemporal event cross-fusion network feature extraction model, the prediction model is trained. It needs to be repeatedly verified on the test set to ensure that the average prediction accuracy of the prediction model is below the preset threshold. The model needs to be tested on the test set to ensure that it meets generalization requirements. When the model's prediction error is less than the specified accuracy standard (for example, the root mean square error is below the set value), the model is saved for future use.

[0077] In one embodiment, the process of determining the three-dimensional location information in step C includes: locating the user's two-dimensional coordinate position through TDOA (Time Difference of Arrival), GPS, and Wi-Fi assistance technologies, and determining the user's three-dimensional coordinate position through high-precision three-dimensional building maps and base station signal beam data, thereby forming a high-, medium-, and low-level distribution cognition of the user.

[0078] In one embodiment, the process of step C includes:

[0079] Step C01: sorting the base stations in a specified order (from high to low) based on the predicted future traffic values ​​of the base stations in the target area;

[0080] Step C02: Select the first and last cells in the sorting results as initial cluster centers;

[0081] Step C03: Apply the K-Means clustering algorithm to perform cluster analysis on the base stations. Then, calculate the new cluster center based on the average distance within the cluster. Continue iterating until the algorithm converges to determine the base stations that need antenna downtilt adjustment and those that do not. The above average distance is defined in three-dimensional space and is calculated as follows:

[0082]

[0083] in, Indicates the The three-dimensional space distance of each cell, Indicates the traffic volume of the current cell. Indicates the ratio of the number of users in the middle and lower layers of the current cell's three-dimensional space to the total number of users in the cell. Indicates the ratio of the number of users in the three-dimensional space of the current cell to the total number of users in the cell; and Obtained according to the corresponding three-dimensional position information.

[0084] In one embodiment, the state information of the deep reinforcement learning algorithm in step D is:

[0085]

[0086] in, Represents the state information of the deep reinforcement learning algorithm, Indicates the number of cells in the area of ​​interest, Indicates the The predicted traffic volume of each cell, Indicates the The proportion of low- and middle-tier users in a cell to the total number of users, Indicates the The ratio of the number of nearby users in a cell to the total number of users.

[0087] The actions of the deep reinforcement learning algorithm are:

[0088]

[0089] in, represents the action of the deep reinforcement learning algorithm, Indicates the Antenna downtilt angle value of each cell;

[0090] The reward of the deep reinforcement learning algorithm is the total system throughput in the target area.

[0091] On this basis, the specific process of using the deep reinforcement learning algorithm in step D to calculate the antenna downtilt angle adjustment plan of the current base station includes: according to the location ratio of the cell terminal equipment (obtained according to the geographical location) and the future traffic value, the antenna downtilt angle adjustment parameters of the current base station are obtained as actions through the deep reinforcement learning algorithm. The final goal is to achieve the optimal system throughput.

[0092] At the same time, before step D is calculated through the deep reinforcement learning algorithm, it also includes the following steps: establishing a handshake communication with the current base station. A successful handshake communication means that the current base station can perform antenna downtilt operation.

[0093] In summary, this paper proposes an antenna downtilt optimization method based on traffic prediction and reinforcement learning. This method fully considers multidimensional factors such as time, space, and events, employs more efficient feature extraction and fusion techniques, and further optimizes the antenna downtilt strategy based on traffic prediction. This method effectively addresses the current challenge of resource optimization in mobile communication networks, improving energy efficiency and saving costs while ensuring transmission accuracy.

[0094] Preferably, fully connected operations are first performed in the time domain, space domain, and event domain to fuse their respective trend information. Subsequently, in the spatial domain, feature extraction is performed using a trend feature extraction module and a node feature extraction module with an attention mechanism. Then, spatiotemporal features are extracted using a graph convolutional network and a convolutional neural network. After completion, the results are input into the time domain and event domain modules for cross-fusion.

[0095] Similarly, in the time domain, the trend feature extraction module, period feature extraction module, and noise feature extraction module with attention mechanism are used for feature extraction, and then enter the graph convolutional network and convolutional neural network for spatiotemporal feature extraction. After completion, they are passed to the spatial domain and event domain modules for cross-fusion.

[0096] Next, in the event domain, feature extraction is performed using a trend feature extraction module and a node feature extraction module with an attention mechanism. A graph convolutional network and a convolutional neural network are then used to extract spatiotemporal features. The results are then fed into the time domain and space domain modules for cross-fusion. This completes a single layer of time-space-event feature cross-fusion. The same process is then repeated for two more layers. Finally, the features are flattened and connected to a two-layer fully connected network for final information fusion, yielding the prediction result.

[0097] It should be noted that relying solely on the attention mechanism may overlook subtle correlations between multiple domains. Current feature extraction methods typically extract temporal features first, followed by spatial features, lacking the cross-integration of multi-domain features between high-dimensional and low-dimensional domains.

[0098] Leaky ReLU (a type of activation function) is used in all neural network activation functions to avoid the "dead neuron" problem. The root mean squared error (RMSE) is used as the loss function to accurately assess the degree of prediction deviation.

[0099] After the model meets accuracy requirements, it is deployed and the predicted cell traffic is calculated. The communication scenario considered is a heavily loaded cell, where antenna downtilt adjustment is performed to improve coverage. Based on the predicted traffic, the K-Means clustering algorithm is used to cluster cells into those that require antenna downtilt adjustment and those that do not. A "handshake" query is performed to determine the final list of cells participating in antenna downtilt adjustment.

[0100] Based on the cell list, traffic prediction results, user terminal device location and other environmental information, a deep reinforcement learning algorithm is used to determine the antenna downtilt angle value of the cell, enabling antenna downtilt angle adjustment to dynamically meet wireless network traffic requirements.

[0101] The following combination Figure 2 , the antenna downtilt angle optimization method based on traffic prediction and reinforcement learning of the present invention is specifically described:

[0102] S101: Receive the target area and traffic collection time period input by the user, and proceed to S102;

[0103] The central server prepares to adjust the antenna downtilt angle according to the area specified by the user and sets the time interval for traffic data collection (for example: 1 second / 10 seconds / 1 minute / 30 minutes, etc.).

[0104] S102: The central server establishes communication with the base station. The base station uploads its network traffic data to the central server within a set time period, forming a spatiotemporal event cube dataset. Simultaneously, the central server integrates internet data (e.g., weekday and weekend event information from calendars, meteorological event information from weather websites, traffic flow event information from map websites, and group activity information from concert websites) to construct a complete spatiotemporal event cube dataset, and then proceeds to S103.

[0105] S103: Use the SVM classification algorithm to perform binary classification on the collected data, distinguishing between clean data and contaminated data, and then determine whether the spatiotemporal event cube dataset is a clean dataset. If it is, proceed to S105; if not, proceed to S104. Contaminated data that does not meet the requirements (for example, pulsed data, contaminated data, etc.) is removed through clipping. The expression for SVM binary classification is as follows:

[0106]

[0107] in, Represents the input sample to be classified; Indicates the training samples; Indicates the The label of the training sample is +1 or -1; represents the Lagrange multiplier, which represents the importance of the sample; Represents the kernel function, which is used to calculate the similarity between samples; represents the bias term; Represents the symbolic function, which determines the output category; Indicates the number of training samples.

[0108] S104: After confirming that the data of interest is contaminated, a pure spatiotemporal event cube dataset is obtained through a clipping method, and then proceeds to S105;

[0109] S105: After obtaining a clean data set, first perform a full connection operation, then perform feature extraction in the spatial domain, and then enter S106. Preferably, a node feature extraction module with an attention mechanism is used to perform correlation analysis in the spatial domain.

[0110] Subsequently, the spatial features are extracted using the frequency domain graph convolution method, which replaces the traditional convolution operator by linearly diagonalizing the operator in the Fourier domain. At the same time, a convolutional neural network is used to extract temporal features.

[0111] Similarly, the temporal feature extraction process includes a trend feature extraction module with an attention mechanism, a period feature extraction module, and a noise feature extraction module. The attention mechanism, multi-head attention feature extraction formula, and graph convolution feature extraction formula of these modules are the same as those in the spatial domain.

[0112] Similarly, feature extraction in the event domain includes a trend feature extraction module and a node feature extraction module with an attention mechanism. The attention mechanism, multi-head attention feature extraction formula, and graph convolution feature extraction formula are also consistent with those in the spatial domain.

[0113] S106: Traffic prediction is performed using the extracted features. After feature extraction in the spatial domain, feature extraction continues in the time and event domains. Similarly, after feature extraction in the time domain, further feature extraction is performed in the spatial and event domains. After feature extraction in the event domain, feature extraction is performed again in the time and spatial domains. This cycle achieves information fusion of multi-domain features at different high and low dimensions. Finally, a two-layer fully connected network is connected to complete global information fusion, and the process proceeds to S107.

[0114] The activation function of the neural network uses the Leaky ReLU function, which is expressed as follows:

[0115]

[0116] in, is a decimal less than 1, usually set to 0.01. When it is greater than 0, the Leaky ReLU function is the same as the ordinary ReLU function, both are linear functions; when When the value is less than 0, the Leaky ReLU function is a nonlinear function. It addresses the "neuron death" issue in the standard ReLU function, making neural network training more robust. Fully connected layers can capture subtle connections between different domains.

[0117] S107: Check whether the prediction accuracy meets the requirements; if so, proceed to S108; if not, return to S105. After fusing multi-domain features, the root mean square error (RMSE) is used as the loss function to accurately evaluate the degree of prediction deviation. Its expression is as follows:

[0118]

[0119] in, It is The predicted value of the sample, It is The true value of the sample, is the number of samples. The units of RMSE are the same as those of the predicted and actual values. A smaller RMSE value indicates a smaller discrepancy between the predicted and true values, and a higher prediction accuracy. By calculating the RMSE and determining whether it is below a preset threshold, we determine whether the prediction accuracy meets the requirements. If not, we continue multi-domain feature extraction and cross-fusion calculations until the required prediction accuracy is achieved.

[0120] S108: After the central processor has deployed the traffic prediction model, it calculates the predicted traffic of each cell and proceeds to S109.

[0121] S109: After the CPU obtains the predicted traffic data, it uses a two-stage clustering algorithm based on these prediction results to divide the cells into two categories: cells that need to participate in antenna downtilt adjustment and cells that do not need to participate in antenna downtilt adjustment, and then enters S110. The two-stage clustering algorithm includes: 1) calculating the network traffic of each cell and sorting them from high to low according to traffic; 2) using the K-Means clustering algorithm to perform cluster analysis on the cells. Distance is a three-dimensional space. The distance defined in .

[0122] S110: After obtaining the traffic prediction data and the antenna downtilt angle list, the central processor communicates with each cell base station again to confirm whether the antenna downtilt angle adjustment operation can be performed; if it can, it enters S111; if not, it returns to S101 to avoid being unable to adjust resources in the current period.

[0123] S111: The central processing unit calculates an adjustment plan for the antenna downtilt angle based on the predicted network traffic and the user's three-dimensional location information through a deep reinforcement learning algorithm, and executes the adjustment of the antenna downtilt angle.

[0124] The state information of the deep reinforcement learning algorithm is:

[0125]

[0126] in, Represents the state information of the deep reinforcement learning algorithm, Indicates the number of cells in the area of ​​interest, Indicates the The predicted traffic volume of each cell, Indicates the The proportion of low- and middle-tier users in a cell to the total number of users, Indicates the The ratio of the number of users in the vicinity of a cell to the total number of users;

[0127] The actions of the deep reinforcement learning algorithm are:

[0128]

[0129] in, represents the action of the deep reinforcement learning algorithm, Indicates the Antenna downtilt angle value of each cell;

[0130] The reward of the deep reinforcement learning algorithm is the total system throughput in the region of interest. Five classic deep reinforcement learning algorithms, DQN, DDPG, PPO, A3C, and SAC, are used to assign antenna downtilt parameters to each cell as actions, with the ultimate goal of achieving optimal system throughput.

[0131] In a second aspect, an embodiment of the present invention further provides an antenna downtilt angle optimization system based on traffic prediction and reinforcement learning, which is used to implement the steps of the above-mentioned antenna downtilt angle optimization method.

[0132] Specifically, the system includes:

[0133] The base station module is used to upload the base station's network traffic data and perform antenna downtilt adjustment operations;

[0134] The data construction module is used to execute the process of step A in the above method. Specifically, the module includes:

[0135] The data collection module is used to collect network traffic data of base stations in the target area according to a set fixed period;

[0136] The data preprocessing module is used to filter out contaminated network traffic data using the SVM classification algorithm to obtain a clean data set.

[0137] The traffic prediction model training module is used to execute the process of step B in the above method. Specifically, the module includes:

[0138] The spatiotemporal event feature extraction module is used to extract the features of network traffic data from three dimensions: time, space, and events.

[0139] The cross-fusion and traffic prediction module is used to fuse multi-domain features and generate traffic prediction results.

[0140] The antenna downtilt angle adjustment cell clustering module is used to: execute step C in the above method; that is, perform cluster analysis on the cells to determine whether antenna downtilt angle adjustment is required based on traffic prediction results and location information.

[0141] The antenna downtilt angle adjustment reinforcement learning module is used to: execute step D in the above method; that is, use a deep reinforcement learning algorithm to calculate the antenna downtilt angle adjustment plan for the cell, and send the plan to the base station module to perform antenna downtilt angle adjustment.

[0142] The temporal relationship between each module includes: the base station module sends base station traffic data, base station location, user three-dimensional location and other information to the data acquisition module; the data acquisition module passes the collected traffic data to the data preprocessing module; the preprocessed data is sent to the spatiotemporal event feature extraction module; then, the data enters the multi-domain feature cross-fusion and traffic prediction module for traffic prediction; the prediction results are passed to the antenna downtilt angle adjustment cell clustering module to cluster the antenna downtilt angles of the cells; finally, the antenna downtilt angle adjustment reinforcement learning module calculates the antenna downtilt angle optimization plan based on the deep reinforcement learning algorithm, and sends it to each base station module to achieve dynamic optimization and utilization of base station resources in the region.

[0143] This method not only integrates information from both the temporal and spatial domains but also takes into account the impact of different events on traffic prediction. Furthermore, it cross-integrates features from different domains at the intermediate layer, enhancing the ability to capture temporal, spatial, and event correlations across high and low dimensions. By designing a graph convolutional neural network, it effectively extracts location-related features between any two base stations.

[0144] This invention also performs a cluster analysis of antenna downtilt angles on cells, identifying those that require and are qualified to adjust antenna downtilt angles, thus avoiding excessive handovers. Leveraging deep reinforcement learning, this system learns from the environment and adjusts the antenna downtilt parameters of each cell, maximizing system throughput and reducing costs. Preemptive adjustment of antenna downtilt angles based on traffic predictions enables faster response to demand compared to solutions based solely on reinforcement learning.

[0145] In addition, the present invention adopts an antenna downtilt adjustment mechanism that combines traffic prediction and reinforcement learning to predict network traffic in real time and perform antenna downtilt adjustment operations in real time, which is more adapted to the actual needs of base stations to intelligently perceive the environment and dynamically adjust, thereby improving resource utilization efficiency.

[0146] On the third aspect, an embodiment of the present application provides an antenna downtilt angle optimization device based on traffic prediction and reinforcement learning. The antenna downtilt angle optimization device based on traffic prediction and reinforcement learning can be a personal computer (PC), a laptop computer, a server, and other devices with data processing functions.

[0147] Reference Figure 3 , Figure 3 Schematic diagram of the hardware structure of the antenna downtilt optimization device based on traffic prediction and reinforcement learning involved in the embodiment of the present application. In the embodiment of the present application, the antenna downtilt optimization device based on traffic prediction and reinforcement learning may include a processor, a memory, a communication interface, and a communication bus.

[0148] The communication bus may be of any type and is used to interconnect the processor, memory, and communication interface.

[0149] Communication interfaces include input / output (I / O), physical, and logical interfaces, which interconnect components within the device and other devices (such as other computing devices or user equipment). Physical interfaces can include Ethernet, fiber, or ATM interfaces; user equipment can include displays or keyboards.

[0150] The memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0151] The processor may be a general-purpose processor that can invoke a traffic prediction and reinforcement learning-based antenna downtilt optimization program stored in a memory and execute the traffic prediction and reinforcement learning-based antenna downtilt optimization method provided in the embodiments of the present application. For example, the general-purpose processor may be a central processing unit (CPU). The method executed when the traffic prediction and reinforcement learning-based antenna downtilt optimization program is invoked can be referenced from the various embodiments of the traffic prediction and reinforcement learning-based antenna downtilt optimization method of the present application and will not be further described here.

[0152] Those skilled in the art will understand that Figure 3 The hardware structure shown in the figure does not constitute a limitation to the present application and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0153] In a fourth aspect, an embodiment of the present application also provides a computer-readable storage medium.

[0154] The computer-readable storage medium of the present application stores an antenna downtilt optimization program based on traffic prediction and reinforcement learning, wherein when the antenna downtilt optimization program based on traffic prediction and reinforcement learning is executed by a processor, the steps of the antenna downtilt optimization method based on traffic prediction and reinforcement learning as described above are implemented.

[0155] Among them, the method implemented when the antenna downtilt angle optimization program based on traffic prediction and reinforcement learning is executed can refer to the various embodiments of the antenna downtilt angle optimization method based on traffic prediction and reinforcement learning in this application, and will not be repeated here.

[0156] It should be noted that the serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0157] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, or the part that contributes to the existing technology, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above and includes a number of instructions for enabling a terminal device to execute the methods described in each embodiment of this application.

[0158] The terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices. The terms "first", "second" and "third" are used to distinguish different objects, etc., and do not represent a sequence, nor do they limit the "first", "second" and "third" to different types.

[0159] In the description of the embodiments of this application, the words "exemplary," "for example," or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary," "for example," or "for example" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary," "for example," or "for example" is intended to present the relevant concepts in a concrete manner.

[0160] In the description of the embodiments of the present application, unless otherwise specified, “ / ” means or, for example, A / B can mean A or B; “and / or” in the text is merely a description of the association relationship of associated objects, indicating that three relationships may exist, for example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiments of the present application, “multiple” refers to two or more than two.

[0161] In some processes described in the embodiments of the present application, multiple operations or steps are included that appear in a specific order. However, it should be understood that these operations or steps may not be performed in the order in which they appear in the embodiments of the present application or may be performed in parallel. The sequence numbers of the operations are only used to distinguish between different operations, and the sequence numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations or steps may be performed in sequence or in parallel, and these operations or steps may be combined.

[0162] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, or the part that contributes to the existing technology, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above and includes a number of instructions for enabling a terminal device to execute the methods described in each embodiment of this application.

[0163] The above are only specific implementations of the embodiments of the present invention, but the scope of protection of the embodiments of the present invention is not limited to them. Any person skilled in the art can easily conceive of various equivalent modifications or replacements within the technical scope disclosed in the embodiments of the present invention, and such modifications or replacements should be included in the scope of protection of the embodiments of the present invention. Therefore, the scope of protection of the embodiments of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for optimizing antenna downtilt angle based on traffic prediction and reinforcement learning, characterized in that: The method comprises the following steps: Construct a spatiotemporal event cube dataset of network traffic data of base stations in the target area; In the same dimension, the spatial, temporal, and event data in the spatiotemporal event cube dataset are fused; corresponding features are extracted in the temporal, spatial, and event domains, respectively. The extracted features are then cross-fused to achieve multi-domain information fusion and obtain a traffic prediction model; The traffic prediction model predicts the future traffic value of each base station in the target area. Based on the future traffic value and three-dimensional location information of each base station, a cluster analysis is performed on the base stations to determine whether they will participate in antenna downtilt adjustment. The base stations that need antenna downtilt adjustment are then determined. Based on the future traffic value of the base station for which antenna downtilt adjustment is required, the antenna downtilt adjustment plan of the current base station is calculated to complete the antenna downtilt adjustment; The state information of the reinforcement learning algorithm is: in, Represents the state information of the reinforcement learning algorithm, Indicates the number of cells in the area of ​​interest, Indicates the The predicted traffic volume of each cell, Indicates the The proportion of low- and middle-tier users in a cell to the total number of users, Indicates the The ratio of the number of users in the vicinity of a cell to the total number of users; The actions of the reinforcement learning algorithm are: in, represents the action of the reinforcement learning algorithm, Indicates the Antenna downtilt angle value of each cell; The reward of the reinforcement learning algorithm is the total system throughput in the area of ​​interest. The antenna downtilt parameters of each cell are given as actions through the reinforcement learning algorithm, and the final goal is to optimize the system throughput.

2. The antenna downtilt angle optimization method based on traffic prediction and reinforcement learning according to claim 1, characterized in that: The process of constructing a spatiotemporal event cube dataset of network traffic data of base stations in the target area includes: Set the target area and network traffic data collection period; Collect network traffic data of base stations in the target area according to the collection period; Based on the time, space, event dimensions and flow values ​​of network traffic data, a spatiotemporal event cube dataset of network traffic data is constructed.

3. The antenna downtilt angle optimization method based on traffic prediction and reinforcement learning according to claim 2, characterized in that: After the cube dataset is constructed, the following steps are further included: slicing the cube dataset in the time domain and classifying it using a classification algorithm to obtain a cube dataset that meets the requirements.

4. The antenna downtilt angle optimization method based on traffic prediction and reinforcement learning according to claim 1, wherein The process of obtaining the traffic prediction model includes: Perform full connection layer processing on the data in the spatiotemporal event cube data to achieve the fusion of space, time and event data in the same dimension; In the spatial domain, the trend feature extraction module and node feature extraction module with attention mechanism are used to perform data correlation analysis to obtain spatial domain features. In the time domain, we use the trend feature extraction module, period feature extraction module, and noise feature extraction module with an attention mechanism to perform correlation analysis on time domain data and obtain time domain features. In the event domain, data association analysis is performed through the trend feature extraction module and node feature extraction module with attention mechanism to obtain event domain features; By cross-fusion, the spatial domain features, time domain features and event domain features are integrated and fused to obtain the corresponding three vectors; After the three vectors pass through two layers of identical multi-domain feature extraction and cross-fusion operations, they pass through a two-layer fully connected network to obtain the traffic prediction results.

5. The antenna downtilt angle optimization method based on traffic prediction and reinforcement learning according to claim 4, characterized in that: The process of obtaining a traffic prediction result by subjecting the three vectors to two layers of identical multi-domain feature extraction and cross-fusion operations, and then passing them through a two-layer fully connected network includes: performing two layers of identical multi-domain feature extraction and cross-fusion operations on the three vectors to obtain three fused vectors; flattening the three fused vectors, and fusing and extracting features through a two-layer fully connected network to obtain a traffic prediction result.

6. The antenna downtilt angle optimization method based on traffic prediction and reinforcement learning according to claim 1, characterized in that The process of determining a base station requiring antenna downtilt adjustment includes: Sort the base stations in a specified order according to their future traffic values, and determine a designated base station as a cluster center based on the sorting result; The base stations are clustered and analyzed using a clustering algorithm. New cluster centers are calculated based on the average distance within the cluster. The algorithm is iterated continuously until convergence, and the base stations that require antenna downtilt adjustment are obtained.

7. The antenna downtilt angle optimization method based on traffic prediction and reinforcement learning according to claim 6, characterized in that: The average distance is defined in three-dimensional space, and the specific calculation formula is: in, Indicates the The three-dimensional space distance of each cell, Indicates the traffic volume of the current cell. Indicates the ratio of the number of users in the middle and lower layers of the current cell's three-dimensional space to the total number of users in the cell. Indicates the ratio of the number of users in the three-dimensional space of the current cell to the total number of users in the cell; and Obtained according to the corresponding three-dimensional position information.

8. The antenna downtilt angle optimization method based on traffic prediction and reinforcement learning according to any one of claims 1 to 7, characterized in that: The process of determining the three-dimensional position information includes: determining the three-dimensional coordinate position of the user according to the two-dimensional coordinate position of the user. 9.An antenna downtilt angle optimization device based on traffic prediction and reinforcement learning, characterized in that: The antenna downtilt angle optimization device based on traffic prediction and reinforcement learning includes a processor, a memory, and an antenna downtilt angle optimization program based on traffic prediction and reinforcement learning stored on the memory and executable by the processor, wherein when the antenna downtilt angle optimization program based on traffic prediction and reinforcement learning is executed by the processor, the steps of the antenna downtilt angle optimization method based on traffic prediction and reinforcement learning as described in any one of claims 1 to 7 are implemented.

Images