Mobile base station network automatic optimization method

By applying deep learning models to the mobile base station network for automated analysis and adjustment, the cumbersome optimization process in the existing technology is solved, and the optimization efficiency and speed are improved.

CN119450521APending Publication Date: 2025-02-14JIANGSU SO FINE COMM
View PDF 0 Cites 5 Cited by

Patent Information

Application Number
CN202411536452.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-31
Publication Date
2025-02-14

AI Technical Summary

Technical Problem

The existing mobile base station network optimization methods require the use of automated planning tools and manual data analysis, which leads to a cumbersome and long-term optimization process and low optimization efficiency.

Method used

The deep learning model is used to analyze the base station network data, automatically adjust network parameters, monitor optimization effects in real time, and update the model regularly to adapt to network changes.

Benefits of technology

Through automated deep learning model analysis and adjustment, the efficiency of mobile base station network optimization is significantly improved, manual intervention is reduced, and optimization time is shortened.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119450521A_ABST
    Figure CN119450521A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of mobile communication, in particular to an automatic optimization method for a mobile base station network, which comprises the steps of collecting data of network equipment, user equipment and other related sources, cleaning a collected data set to ensure data quality, extracting related characteristics in the cleaned data, selecting a deep learning model for training, and optimizing the mobile base station network. A training model trained in advance is utilized to analyze potential problems of a base station network to obtain an optimization decision, network parameters are automatically adjusted according to the optimization decision based on an automatic script, the adjusted network parameters are monitored in real time, comparison before and after optimization of the base station network is carried out, and the network optimization effect is ensured. Therefore, the problem that according to an existing mobile base station network optimization method, due to the fact that an automatic planning tool and manual data analysis are needed in the optimization process, the process is tedious, time consumption is long, and the optimization efficiency is low is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of mobile communication technology, and in particular to a method for automatic optimization of a mobile base station network. Background Art

[0002] In the 5G era of intelligent interconnection of all things, the intelligent and digital transformation and upgrading of vertical industry production methods will continue to accelerate. Traditional private network technology is difficult to meet the ever-changing information business needs of enterprise networks. Base station network optimization is a complex process aimed at improving network performance.

[0003] The prior art discloses a mobile base station network optimization method, which uses an automated planning tool to complete the neighboring area planning of 4G-5G and 5G-5G cells, and imports it into the wireless network management to enable the neighboring area self-configuration function. When the 5G site is powered on and the link is successfully established, the network management automatically configures the neighboring area for this site based on the imported neighboring area planning table to achieve network optimization.

[0004] However, using the above method, since the optimization process requires the use of automated planning tools and manual data analysis, the process is cumbersome and time-consuming, resulting in low optimization efficiency. Summary of the invention

[0005] The purpose of the present invention is to provide a method for automatic optimization of a mobile base station network, aiming to solve the problem of low optimization efficiency caused by the existing mobile base station network optimization method, which requires the use of automated planning tools and manual data analysis during the optimization process.

[0006] To achieve the above object, the present invention provides a method for automatic optimization of a mobile base station network, comprising:

[0007] Collect data from network devices, user devices, and other relevant sources;

[0008] Clean the collected data sets to ensure data quality;

[0009] Extract relevant features from the cleaned data;

[0010] Select a deep learning model for training;

[0011] The trained deep learning model analyzes potential problems in the base station network and makes optimization decisions;

[0012] Automatically adjust network parameters based on optimization decisions based on automated scripts;

[0013] Monitor the adjusted network parameters in real time and compare the base station network before and after optimization to ensure the network optimization effect;

[0014] Regularly inject new data into the trained model to maintain the model's accuracy and adaptability.

[0015] The data from other related sources include signal strength, user location, network traffic, call records and device status data.

[0016] The specific steps of extracting relevant features from the cleaned data include:

[0017] Extracting network performance-related features from the cleaned data, including signal-to-interference-and-noise ratio (SINR), cell load, and user mobility;

[0018] The most influential features for network performance prediction using machine learning algorithms;

[0019] Based on data conversion technology, reduce feature dimensions and improve model efficiency.

[0020] The specific steps of selecting a deep learning model for training include:

[0021] Select the corresponding deep learning model according to the base station network optimization direction;

[0022] Use the collected network datasets to train deep learning models and adjust model parameters to optimize performance;

[0023] Cross-validation is used to ensure the generalization ability of the model and avoid overfitting.

[0024] The specific steps of analyzing the potential problems of the base station network with the trained deep learning model to obtain an optimization decision include:

[0025] Use the trained deep learning model to predict uncollected data and identify potential network problems;

[0026] Make optimization decisions based on the results of identified network issues.

[0027] The specific steps of automatically adjusting network parameters based on the automation script according to the optimization decision include:

[0028] Develop automation scripts;

[0029] Leverage developed automation scripts to automatically adjust network parameters based on optimization decisions.

[0030] The specific steps of real-time monitoring of the adjusted network parameters and comparing the base station network before and after optimization to ensure the network optimization effect include:

[0031] Real-time monitoring of adjusted network performance indicators;

[0032] Evaluate the actual effect of the optimization measures and compare them with the performance before optimization to ensure the effectiveness of base station network optimization;

[0033] The specific steps of regularly injecting new data into the training model to maintain the accuracy and adaptability of the model include:

[0034] Collect new optimized data and inject it into the model for retraining;

[0035] Update the model based on new data and feedback to adapt to network changes and new optimization needs;

[0036] Ensure the accuracy and adaptability of the training model for base station network optimization.

[0037] The present invention discloses an automatic optimization method for a mobile base station network, which collects data from network equipment, user equipment and other related sources, cleans the collected data sets to ensure data quality, extracts relevant features from the cleaned data, selects a deep learning model for training, uses the pre-trained training model to analyze potential problems of the base station network to obtain optimization decisions, automatically adjusts network parameters according to the optimization decisions based on automated scripts, monitors the adjusted network parameters in real time, and compares the base station network before and after optimization to ensure the network optimization effect, thereby solving the problem of existing mobile base station network optimization methods. Since automated planning tools and manual data analysis are required during the optimization process, the process is cumbersome and time-consuming, resulting in low optimization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below.

[0039] Figure 1 It is a flow chart of a mobile base station network automatic optimization method of the present invention.

[0040] Figure 2 It is a flow chart of extracting relevant features from cleaned data according to the present invention.

[0041] Figure 3 It is a flowchart of selecting a deep learning model for training according to the present invention.

[0042] Figure 4 It is a flow chart of the trained deep learning model of the present invention analyzing potential problems of the base station network to obtain optimization decisions.

[0043] Figure 5 It is a flow chart of the present invention for automatically adjusting network parameters according to optimization decisions based on an automated script.

[0044] Figure 6It is a flow chart of the present invention for real-time monitoring of adjusted network parameters, performing comparison before and after base station network optimization, and ensuring network optimization effect.

[0045] Figure 7 It is a flow chart of the present invention for regularly injecting new data into the training model to maintain the accuracy and adaptability of the model. DETAILED DESCRIPTION

[0046] Embodiments of the present invention are described in detail below. Examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, but should not be construed as limiting the present invention.

[0047] See also Figure 1-Figure 7 ,in, Figure 1 is a flow chart of a mobile base station network automatic optimization method of the present invention, Figure 2 is a flow chart of the present invention for extracting relevant features from the cleaned data, Figure 3 is a flowchart of selecting a deep learning model for training according to the present invention, Figure 4 It is a flow chart of analyzing potential problems of base station network by the trained deep learning model of the present invention to obtain optimization decisions, Figure 5 is a flow chart of the present invention for automatically adjusting network parameters according to optimization decisions based on an automated script, Figure 6 It is a flowchart of the present invention for real-time monitoring of adjusted network parameters, performing comparison before and after base station network optimization, and ensuring network optimization effect. Figure 7 It is a flow chart of the present invention for regularly injecting new data into the training model to maintain the accuracy and adaptability of the model.

[0048] A mobile base station network automatic optimization method of the present invention comprises:

[0049] S1 collects data from network devices, user devices, and other related sources;

[0050] In an embodiment of the present invention, data collection tools such as NMS / OSS system, network probes, and log collectors are used, and data collection methods such as active collection and passive collection are selected. Performance counters are regularly extracted from network devices such as base stations and switches through the NMS / OSS system, and network probes are used to capture network traffic data. Then, all base station performance data in a certain area in the past month are collected from network devices, user devices, and other related sources, including data on signal strength, user location, network traffic, call records, and device status.

[0051] S2 cleans the collected data sets to ensure data quality;

[0052] In the embodiment of the present invention, the data is cleaned to remove noise and outliers, and data standardization and normalization are performed to ensure data quality.

[0053] S3 extracts relevant features from the cleaned data;

[0054] The specific steps include:

[0055] S31 extracts features related to network performance from the cleaned data, the features including signal to interference and noise ratio (SINR), cell load, and user mobility;

[0056] S32 uses machine learning algorithms to predict the most influential features of network performance;

[0057] S33 is based on data conversion technology to reduce feature dimensions and improve model efficiency;

[0058] In an embodiment of the present invention, features related to network performance, such as base station location, user density, timestamp, weather conditions, signal-to-interference-and-noise ratio (SINR), cell load, and user mobility, are extracted from the cleaned data, statistical methods or machine learning algorithms are used to select the features that have the greatest influence on network performance prediction, key features that affect network performance are determined, and data transformation techniques, such as principal component analysis (PCA), are applied to reduce feature dimensionality and improve model efficiency.

[0059] S4 selects a deep learning model for training;

[0060] The specific steps include:

[0061] S41 selects a corresponding deep learning model according to the base station network optimization direction;

[0062] S42 uses the collected network dataset to train the deep learning model and adjusts the model parameters to optimize performance;

[0063] S43 ensures the generalization ability of the model through cross-validation to avoid overfitting;

[0064] In an embodiment of the present invention, a deep learning model, such as a convolutional neural network (CNN), a recurrent neural network (RNN), a random forest, a support vector machine (SVM), and deep learning, is selected according to an optimization goal such as improving network coverage, reducing call drop rate, and optimizing resource allocation. After selecting a suitable deep learning model, model parameters are set, such as a learning rate, number of iterations, and number of hidden layer nodes. A loss function and an optimization algorithm, such as mean square error (MSE) and gradient descent, are selected. Historical data is input to train the model, and changes in the loss function are observed to ensure that the model is converging. A validation set is used to monitor the generalization ability of the model. Regularization and early stopping techniques are used when necessary. The validation set is used to evaluate model performance, and model parameters are adjusted. Evaluation indicators may include accuracy, recall, F1 score, and mean square error. The model is adjusted according to the evaluation results, such as adding data features, modifying the model structure, and adjusting parameters. The final model is tested using a test set to verify its performance on unknown data, ensuring that the test results meet predetermined performance standards, and the trained model is deployed in a production environment for actual network optimization. The key factors affecting network performance are identified, and the generalization ability of the model is ensured through cross-validation to avoid overfitting.

[0065] The deep learning model trained by S5 analyzes potential problems in the base station network and makes optimization decisions;

[0066] The specific steps include:

[0067] S51 uses the trained deep learning model to predict uncollected data and identify potential network problems;

[0068] S52 makes optimization decisions based on the identified network problem results;

[0069] In an embodiment of the present invention, data for prediction are injected into a trained model. These data should have the same characteristics and format as the training data. The trained model is used to predict future network performance. For regression problems, the model will output continuous values; for classification problems, the model will output category labels, ensuring that the input data undergoes the same preprocessing steps as the training data, identifying potential network problems to obtain prediction results, analyzing the prediction output of the model, identifying potential problems and optimization points in the network. For regression models, it may be necessary to analyze the difference between predicted values ​​and actual values. For classification models, it is necessary to pay attention to cases of prediction errors, and then formulate optimization strategies, such as adjusting antenna direction, power level and frequency allocation, clarifying optimization goals, such as minimizing call drop rate and maximizing network capacity, and selecting the most appropriate optimization strategy based on the prediction results.

[0070] S6 automatically adjusts network parameters based on optimization decisions based on automated scripts.

[0071] The specific steps include:

[0072] S61 develops automation scripts;

[0073] S62 uses developed automation scripts to automatically adjust network parameters based on optimization decisions;

[0074] In an embodiment of the present invention, through an automated workflow, each link from prediction to implementation is clarified, and the conditions for triggering automation are determined. For example, based on time, events or performance indicators, automated scripts or software are developed to automatically adjust network parameters according to optimization decisions.

[0075] S7 monitors the adjusted network parameters in real time and compares the base station network before and after optimization to ensure the network optimization effect;

[0076] The specific steps include:

[0077] S71 monitors the adjusted network performance indicators in real time;

[0078] S72 evaluates the actual effect of the optimization measures and compares it with the performance before optimization to ensure the effect of the base station network optimization;

[0079] In the embodiment of the present invention, network performance indicators such as throughput, delay and call drop rate are continuously monitored, the actual effect of the optimization measures is evaluated, and the performance is compared with that before optimization to ensure the effect of network optimization.

[0080] S8 regularly injects new data into the training model to maintain the accuracy and adaptability of the model;

[0081] The specific steps include:

[0082] S81 collects new optimized data and injects it into the model for retraining;

[0083] S82 updates the model based on new data and feedback to adapt to network changes and new optimization requirements;

[0084] S83 ensures the accuracy and adaptability of the training model for base station network optimization;

[0085] In the embodiment of the present invention, new optimized data is collected for retraining and improvement of the model, and the model is updated according to the new data and feedback to adapt to changes in the network and new optimization requirements.

[0086] In using the mobile base station network automatic optimization method described in the present invention, a data collection tool such as an NMS / OSS system, a network probe, and a log collector are used to select a data collection method such as active collection and passive collection. Performance counters are regularly extracted from network devices such as base stations and switches through the NMS / OSS system. Network traffic data is captured using a network probe, and then all base station performance data in a certain area in the past month are collected from network devices, user devices, and other related sources, including data on signal strength, user location, network traffic, call records, and device status. The collected data is cleaned to remove noise and outliers, and data standardization and normalization are performed to ensure data quality. Features related to network performance, such as base station location, user density, timestamp, weather conditions, signal-to-interference-plus-noise ratio (SINR), cell load, and user mobility, are extracted from the cleaned data. The features that have the greatest influence on network performance prediction are selected using statistical methods or machine learning algorithms, and key features that affect network performance are determined. Data conversion techniques, such as principal component analysis (PCA), are applied to reduce feature dimensions and improve model efficiency. A deep learning model, such as a convolutional neural network, is selected according to optimization objectives such as improving network coverage, reducing call drop rate, and optimizing resource allocation. (CNN), recurrent neural network (RNN), random forest, support vector machine (SVM) and deep learning. After selecting a suitable deep learning model, set the model parameters, such as learning rate, number of iterations and number of hidden layer nodes. Select loss function and optimization algorithm, such as mean square error (MSE) and gradient descent. Input historical data to train the model. Observe the change of loss function to ensure that the model is converging. Use validation set to monitor the generalization ability of the model. If necessary, use regularization and early stopping techniques. Use validation set to evaluate model performance and adjust model parameters. Evaluation indicators may include accuracy, recall, F1 score and mean square error. According to the evaluation Adjust the model based on the estimation results, such as adding data features, modifying the model structure and adjusting parameters. Use the test set to test the final model to verify its performance on unknown data and ensure that the test results meet the predetermined performance standards. Deploy the trained model to the production environment for actual network optimization and identify the key factors affecting network performance. Ensure the generalization ability of the model through cross-validation to avoid overfitting. Inject the data used for prediction into the trained model. These data should have the same characteristics and format as the training data. Use the trained model to predict future network performance. For regression problems, the model will output continuous values.For classification problems, the model will output category labels. Ensure that the input data undergoes the same preprocessing steps as the training data, identify potential network problems and obtain prediction results, analyze the model's prediction output, identify potential problems and optimization points in the network. For regression models, you may need to analyze the difference between the predicted value and the actual value. For classification models, you need to pay attention to cases of prediction errors, and then develop optimization strategies, such as adjusting antenna direction, power level, and frequency allocation. Clarify optimization goals, such as minimizing dropped calls and maximizing network capacity. Select the most appropriate optimization strategy based on the prediction results. Through automated workflows, clarify each link from prediction to implementation, determine the conditions that trigger automation, such as based on time, events, or performance indicators, and develop automation scripts or software. The software automatically adjusts network parameters according to optimization decisions, establishes a closed-loop system, monitors network performance in real time and automatically adjusts optimization measures according to feedback, continuously monitors network performance indicators such as throughput, delay and call drop rate, evaluates the actual effect of optimization measures, and compares them with the performance before optimization to ensure the effect of network optimization, collects new optimized data for retraining and improvement of the model, and updates the model according to new data and feedback to adapt to network changes and new optimization needs, thereby solving the problem of low optimization efficiency caused by the cumbersome and time-consuming process of the existing mobile base station network optimization method due to the need to use automated planning tools and manual data analysis during the optimization process. ;

[0087] What is disclosed above is only one or more preferred embodiments of the present application, and cannot be used to limit the scope of rights of the present application. Ordinary technicians in this field can understand that all or part of the processes of implementing the above embodiments and equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. A method for automatic optimization of a mobile base station network, characterized in that: include: Collect data from network devices, user devices, and other relevant sources; Clean the collected data sets to ensure data quality; Extract relevant features from the cleaned data; Select a deep learning model for training; The trained deep learning model analyzes potential problems in the base station network and makes optimization decisions; Automatically adjust network parameters based on optimization decisions based on automated scripts; Monitor the adjusted network parameters in real time and compare the base station network before and after optimization to ensure the network optimization effect; Regularly inject new data into the trained model to maintain the model's accuracy and adaptability.

2. A mobile base station network automatic optimization method as claimed in claim 1, characterized in that: The data from other relevant sources include signal strength, user location, network traffic, call records and device status data.

3. A mobile base station network automatic optimization method as claimed in claim 2, characterized in that: The specific steps of extracting relevant features from the cleaned data include: Extracting network performance-related features from the cleaned data, including signal-to-interference-and-noise ratio, cell load, and user mobility; The most influential features for network performance prediction using machine learning algorithms; Based on data conversion technology, reduce feature dimensions and improve model efficiency.

4. A mobile base station network automatic optimization method as claimed in claim 3, characterized in that: The specific steps of selecting a deep learning model for training include: Select the corresponding deep learning model according to the base station network optimization direction; Use the collected network datasets to train deep learning models and adjust model parameters to optimize performance; Cross-validation is used to ensure the generalization ability of the model and avoid overfitting.

5. A mobile base station network automatic optimization method as claimed in claim 4, characterized in that: The specific steps of analyzing the potential problems of the base station network with the trained deep learning model to obtain an optimization decision include: Use the trained deep learning model to predict uncollected data and identify potential network problems; Make optimization decisions based on the results of identified network issues.

6. A mobile base station network automatic optimization method as claimed in claim 5, characterized in that: The specific steps of automatically adjusting network parameters according to optimization decisions based on the automation script include: Develop automation scripts; Leverage developed automation scripts to automatically adjust network parameters based on optimization decisions.

7. A mobile base station network automatic optimization method as claimed in claim 6, characterized in that: The specific steps of real-time monitoring of the adjusted network parameters and comparing the base station network before and after optimization to ensure the network optimization effect include: Real-time monitoring of adjusted network performance indicators; Evaluate the actual effect of the optimization measures and compare them with the performance before optimization to ensure the effectiveness of base station network optimization.

8. A mobile base station network automatic optimization method as claimed in claim 7, characterized in that: The specific steps of regularly injecting new data into the training model to maintain the accuracy and adaptability of the model include: Collect new optimized data and inject it into the model for retraining; Update the model based on new data and feedback to adapt to network changes and new optimization needs; Ensure the accuracy and adaptability of the training model for base station network optimization.

Citation Information

Cited By

  • Network request analysis method and related device

    CN119892656A

  • Key parameter determination method and operation method of base station, storage medium and base station

    CN119997066A

  • Multi-modal data source probe state monitoring and failover method and multi-modal data source probe state monitoring and failover system

    CN120529344A

  • A multi-modal data source probe state monitoring and failover method and system

    CN120529344B

  • Method and device for adjusting azimuth angle of base station antenna and electronic equipment

    CN120639137A