Antenna feeder real-time adjustment method and system based on multi-dimensional data weighted fusion

Abstract

The invention relates to the technical field of communication, and discloses an antenna feeder real-time adjustment method and system based on multi-dimensional data weighted fusion, and the method comprises the steps: collecting multi-dimensional data and constructing a multi-dimensional data set based on various types of original data related to an antenna feeder system of a mobile communication base station; performing data cleaning and feature extraction on the multi-dimensional data set, and performing weighted fusion on multi-dimensional features to generate a comprehensive network state index; and based on the comprehensive network state index, deciding according to a scene strategy, and dynamically adjusting the antenna feeder system parameters of the mobile communication base station. According to the invention, a closed-loop feedback mechanism from data acquisition, fusion evaluation to parameter adjustment is constructed, the accuracy and real-time performance of network optimization are significantly improved, and the problems that a traditional antenna feed system depends on a single data source and adjustment is lagged are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication, and in particular to a real-time adjustment method and system for antennas and feeds based on multi-dimensional data weighted fusion. BACKGROUND

[0002] With the development of 5G networks, mobile communication base station antenna and feed systems face higher optimization needs. The traditional antenna and feed adjustment technology has the following limitations: 1) single data dimension: existing systems are mostly based on single-dimensional data (such as coverage quality or capacity indicators) for adjustment, which cannot fully reflect the network state; 2) insufficient real-time performance: manual adjustment or preset parameter methods are difficult to adapt to dynamic changes in the network environment, resulting in optimization lag; 3) difficulty in balancing energy saving and performance: the lack of effective inclusion of energy consumption indicators makes it difficult to ensure network performance while achieving green energy saving; 4) weak processing capacity for nonlinear relationships: traditional algorithms are difficult to capture complex nonlinear relationships between multi-dimensional data, resulting in inaccurate adjustment strategies.

[0003] CN116801277A discloses a method and device for updating antenna and feed parameters, a network device, and a storage medium. The method includes: obtaining an initial antenna and feed parameter vector of a target cell in the i-th iteration round, and a search direction set i corresponding to the i-th iteration round, i being a positive integer; starting from the initial antenna and feed parameter vector, performing iterative search on each search direction of the search direction set i to iteratively update the antenna and feed parameter vector; in response to the search directions in the search direction set i completing the search, updating the search direction set i to obtain a search direction set i+1 to perform the update of the antenna and feed parameter vector in the i+1-th iteration round. The present disclosure iteratively updates the antenna and feed parameters, realizes real-time adaptive optimization adjustment of the antenna and feed parameters, improves the utilization of resources, and enhances user perception.

[0004] CN111382755B discloses a common-mode antenna and feed optimization method and system based on adaboost weighting and third-order clustering. The method includes: fusing road test data, first data, and location information in any network, and performing grid clustering on the fused data to obtain a data set of any network; according to the initial weight corresponding to the data set of any network respectively set, the AdaBoost algorithm is used to determine the optimal weight corresponding to the data set of any network respectively, and the second data set is obtained according to the optimal weight corresponding to the data set of any network and the data set of any network; selecting a second sampling point in the second data set whose RSRP or SINR is less than a preset threshold value, and clustering each sampling point in the second sampling point with the location information to generate a difference point geographic location clustering set; according to the difference point geographic location clustering set and the antenna and feed values of the first common-mode cell, the angles between the antenna horizontal lobe normal and the vertical lobe normal of the cell and the difference point geographic location clustering set are calculated.

[0005] However, CN116801277A and CN111382755B only adjust the angle of the antenna feeder based on the single-dimensional or double-dimensional RSRP coverage quality dimension and the SINR interference level dimension, resulting in one-sided evaluation and decision bias, and often the evaluation result is not the optimal solution of the network. In addition, CN111382755B uses the AdaBoost algorithm for weighted fusion, which has weak model fitting ability and is difficult to capture the complex nonlinear relationship of multi-dimensional data such as coverage and capacity; the AdaBoost algorithm only limits the complexity of the weak learner to prevent overfitting, and when facing large-scale base stations and multi-dimensional data, it is easy to cause poor generalization ability due to overfitting; at the same time, the sample weight iteration mechanism of the AdaBoost algorithm is easy to amplify the base station data noise (such as misestimated values), interfere with the model learning, and reduce the reliability of the fusion result. SUMMARY

[0006] To solve the above problems, the present application proposes a kind of real-time adjustment method and system of antenna feeder based on multi-dimensional data weighted fusion, and closed-loop feedback mechanism is constructed from data acquisition, fusion evaluation to parameter adjustment, which can significantly improve the accuracy and real-time performance of network optimization, and solve the problem of traditional antenna feeder system relying on single data source and adjustment lag.

[0007] The technical scheme adopted by the present application is as follows: A real-time adjustment method of antenna feeder based on multi-dimensional data weighted fusion, comprising: Based on the multi-class original data related to the mobile communication base station antenna feeder system, multi-dimensional data sets are collected and constructed; Data cleaning and feature extraction are performed on the multi-dimensional data set, and multi-dimensional features are weighted and fused to generate a comprehensive network state index; Based on the comprehensive network state index, decision is made according to the scene strategy, and the parameters of the mobile communication base station antenna feeder system are dynamically adjusted.

[0008] Further, the data cleaning and feature extraction of the multi-dimensional data set, and the weighted fusion of the multi-dimensional features to generate a comprehensive network state index, comprise: Time alignment, dimensionless normalization and derivative feature construction are performed on the heterogeneous data, and the derivative features include dimension cross features and space-time features; Based on the feature importance analysis of XGBoost algorithm, the weights of each dimension feature are automatically adjusted according to the network scene; The comprehensive network state index is output as the core basis for adjusting the parameters of the mobile communication base station antenna feeder system.

[0009] Further, the decision based on the comprehensive network state index according to the scene strategy, and the dynamic adjustment of the parameters of the mobile communication base station antenna feeder system, comprise: According to the comprehensive network state index, three-level decisions of emergency adjustment, regular optimization and energy-saving mode are set; According to the scene strategy, corresponding decisions are issued to hardware devices to perform actual adjustment of the mobile communication base station antenna feeder system parameters; The mobile communication base station antenna feeder system parameter adjustment results are fed back to continuously iterate and form a closed loop optimization.

[0010] Further, the mobile communication base station antenna feeder system related multi-class raw data includes: base station measurement report data, terminal device measurement data, network management system log data, signaling monitoring data and base station energy efficiency management module data.

[0011] Further, the multi-dimensional data includes: coverage quality dimension data, capacity efficiency dimension data, connection reliability dimension data, interference level dimension data, user experience perception dimension data and energy saving efficiency dimension data.

[0012] Further, the evaluation indexes of the coverage quality dimension data include reference signal received power, reference signal received quality and radio resource control establishment success rate, and the raw data sources include base station measurement report data and terminal device measurement data; the evaluation indexes of the capacity efficiency dimension data include physical resource block utilization rate, user number density and uplink / downlink throughput, and the raw data sources include base station measurement report data.

[0013] Further, the evaluation indexes of the connection reliability dimension data include handover success rate, call drop rate, radio link failure rate and fault recovery time, and the raw data sources include network management system log data and signaling monitoring data.

[0014] Further, the evaluation indexes of the interference level dimension data include signal-to-interference-and-noise ratio and interference suppression ratio, and the raw data sources include base station measurement report data and terminal device measurement data; the evaluation indexes of the user experience perception dimension data include end-to-end delay, packet loss rate, video stall rate and voice quality score, and the raw data sources include base station measurement report data and terminal device measurement data.

[0015] Further, the evaluation indexes of the energy saving efficiency dimension data include unit flow energy consumption, dynamic power consumption proportion and idle resource shutdown time length, and the raw data sources include base station energy efficiency management module data.

[0016] A real-time antenna feeder adjustment system based on multi-dimensional data weighted fusion, comprising: A data acquisition module configured to acquire multi-dimensional data and construct a multi-dimensional data set based on mobile communication base station antenna feeder system related multi-class raw data; The data fusion and processing module is configured to perform data cleaning and feature extraction on the multi-dimensional data set, and to perform weighted fusion on the multi-dimensional features to generate a comprehensive network state index. The intelligent adjustment module is configured to make decisions according to scene strategies based on the comprehensive network state index, and to dynamically adjust the parameters of the mobile communication base station antenna feeder system.

[0017] The present application has the following advantages: 1. The present application integrates multi-dimensional data such as coverage quality, capacity efficiency, connection reliability, interference level, user experience perception and energy efficiency, and uses the XGBoost algorithm to realize dynamic weighted fusion of multi-dimensional data to generate a comprehensive network state index (CNSI) as the basis for adjustment decisions. The present application constructs a closed-loop feedback mechanism from data collection, fusion evaluation to parameter adjustment, significantly improving the accuracy and real-time performance of network optimization, and solving the problem of lagging adjustment of traditional antenna feeder systems relying on a single data source.

[0018] 2. Compared with CN116801277A and CN111382755B, the present application constructs a full-scene network state evaluation model through the coverage quality dimension, interference level dimension, capacity efficiency dimension, connection reliability dimension, user experience perception dimension and energy efficiency dimension, which fundamentally solves the one-sidedness of data evaluation in the prior art.

[0019] 3. Compared with CN111382755B which uses the AdaBoost weighted fusion algorithm, the present application uses the XGBoost fusion data algorithm, which can better capture the complex nonlinear relationships between dimensions in the base station data, has stronger fitting ability, is less sensitive to noise and outliers, and can effectively avoid being disturbed by noise in the base station data. At the same time, the XGBoost algorithm supports parallel computing at the feature granularity, which can greatly shorten the training time when processing large-scale base station data, and the training efficiency is significantly improved compared with the serial training method of the AdaBoost algorithm. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 is a flow chart of a real-time adjustment method of an antenna feeder based on multi-dimensional data weighted fusion according to Embodiment 1 of the present application.

[0021] Figure 2 is a system architecture diagram of a real-time adjustment system of an antenna feeder based on multi-dimensional data weighted fusion according to Embodiment 2 of the present application.

[0022] Figure 3 is a flow chart of a real-time adjustment method of an antenna feeder based on multi-dimensional data weighted fusion according to Embodiment 3 of the present application. DETAILED DESCRIPTION

[0023] In order to make the technical features, objectives and effects of the present application clearer, the specific embodiments of the present application are described. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application, that is, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0024] Embodiment 1 As shown in Figure 1 The embodiment provides a real-time adjustment method for a mobile communication base station antenna and feeder system based on multi-dimensional data weighted fusion, which comprises the following steps: Based on the multi-type original data related to the mobile communication base station antenna and feeder system, multi-dimensional data is collected and a multi-dimensional data set is constructed; Data cleaning and feature extraction are performed on the multi-dimensional data set, and multi-dimensional features are weighted and fused to generate a comprehensive network state index; Based on the comprehensive network state index, a decision is made according to a scene strategy, and the parameters of the mobile communication base station antenna and feeder system are dynamically adjusted.

[0025] In the optimization of the mobile communication antenna and feeder system, the traditional scheme relies on single or small amount of dimensional data (such as only covering quality indicators), resulting in one-sided evaluation and decision bias. The embodiment breaks through the six-dimensional data system and realizes cross-dimensional weighted fusion through gradient boosting algorithm to construct a full-scene network state evaluation model, which fundamentally solves the data one-sidedness problem of the prior art.

[0026] Preferably, the multi-type original data related to the mobile communication base station antenna and feeder system includes base station measurement report data, terminal device measurement data, network management system log data, signaling monitoring data and base station energy efficiency management module data.

[0027] Preferably, the multi-dimensional data collected in the embodiment includes coverage quality dimensional data, capacity efficiency dimensional data, connection reliability dimensional data, interference level dimensional data, user experience perception dimensional data and energy saving efficiency dimensional data.

[0028] More preferably, the evaluation indexes and original data sources of the multi-dimensional data in the embodiment are shown in Table 1.

[0029] Table 1-Evaluation indexes and original data sources of multi-dimensional data

[0030] The evaluation indexes of the coverage quality dimensional data include reference signal received power, reference signal received quality and radio resource control establishment success rate, and the original data sources include base station measurement report data and terminal device measurement data.

[0031] Evaluation indicators of capacity efficiency dimension data include physical resource block utilization, user number density, and uplink / downlink throughput. The original data sources include base station measurement report data.

[0032] Evaluation indicators of connection reliability dimension data include handover success rate, call drop rate, radio link failure rate, and fault recovery time. The original data sources include network management system log data and signaling monitoring data.

[0033] Evaluation indicators of interference level dimension data include signal-to-interference-and-noise ratio and interference suppression ratio. The original data sources include base station measurement report data and terminal device measurement data.

[0034] Evaluation indicators of user experience perception dimension data include end-to-end delay, packet loss rate, video stutter rate, and voice quality score. The original data sources include base station measurement report data and terminal device measurement data.

[0035] Evaluation indicators of energy saving efficiency dimension data include unit flow energy consumption, dynamic power consumption proportion, and idle resource shutdown duration. The original data sources include base station energy efficiency management module data.

[0036] Preferably, data cleaning and feature extraction are performed on the multi-dimensional data set, and the multi-dimensional features are weighted and fused to generate a comprehensive network status index, including: Time alignment, dimensionless normalization, and derivative feature construction are performed on the heterogeneous data. The derivative features include dimension cross features and spatio-temporal features. Based on the XGBoost algorithm, the feature importance analysis is performed to automatically adjust the weights of the dimensional features according to the network scenario. The comprehensive network status index is output as the core basis for mobile communication base station antenna feeder system parameter adjustment.

[0037] More preferably, time alignment of heterogeneous data includes: through time window sliding (such as 5-minute aggregation period), the time granularity of each dimension data is unified to ensure the spatio-temporal consistency of coverage quality (second-level real-time data) and energy saving efficiency (minute-level statistical data).

[0038] More preferably, dimensionless normalization includes: for positive indicators (such as RSRP, throughput), 0-1 standardization is adopted ( ), and for negative indicators (such as call drop rate, delay), reverse normalization is adopted ( ) to be uniformly mapped to the [0, 1] interval.

[0039] More preferably, the derived feature construction includes: introducing dimension cross features (such as "coverage quality x capacity efficiency" reflecting the coverage stability under high load), spatio-temporal features (such as "regional user density x time period factor" depicting the difference between morning and evening peak), enhancing the model's representation ability for complex scenarios.

[0040] More preferably, based on the feature importance analysis of the XGBoost algorithm, the weight of each dimension feature is automatically adjusted according to the network scenario, including: taking the comprehensive network state index (CNSI, Comprehensive Network State Index) as the output target (0-100 points), constructing a regression model containing a regularization term:

[0041] wherein, is the expert score marked by artificial, is the model prediction value, is the weight of the th tree, and is the overfitting suppression parameter.

[0042] It should be noted that the XGBoost algorithm has high efficiency, nonlinear modeling ability, parallel training advantage and feature importance evaluation mechanism, so the method uses the XGBoost algorithm to realize dynamic weighted fusion of multi-dimensional data. XGBoost algorithm can automatically learn the dynamic weight of each dimension feature (such as coverage 25%, experience 20%) by integrating multiple decision trees, and supports incremental update to adapt to network changes. Experiments show that compared with traditional methods, XGBoost algorithm can improve the network state evaluation accuracy by 15%-20%, significantly enhancing the system's adaptive ability to complex scenarios.

[0043] Preferably, based on the comprehensive network state index, the scene strategy is used for decision-making, and the parameters of the mobile communication base station antenna feeder system are dynamically adjusted, including: Setting three-level decision-making of emergency adjustment, regular optimization and energy saving mode according to the comprehensive network state index; According to the scene strategy, the corresponding decision is sent to the hardware device to execute the actual adjustment of the parameters of the mobile communication base station antenna feeder system; The adjustment results of the parameters of the mobile communication base station antenna feeder system are fed back to form a closed loop optimization.

[0044] More preferably, the three-level decision-making of emergency adjustment, regular optimization and energy saving mode is set according to the comprehensive network state index (CNSI), including: If CNSI<60 (reference value), "emergency adjustment" is triggered to preferentially solve the problems of high call drop rate and coverage hole; If 60≤CNSI<80 (reference value), then “normal optimization” is initiated to balance capacity and interference (such as adjusting the downtilt angle to expand coverage). If CNSI ≥ 80 (reference value), maintain the current parameters and activate the power saving mode (such as dynamically shutting down idle RRUs).

[0045] Furthermore, this embodiment also provides a real-time antenna and feeder adjustment system based on multi-dimensional data weighted fusion, including: The data acquisition module is configured to collect multi-dimensional data and construct a multi-dimensional dataset based on various types of raw data related to the mobile communication base station antenna feeder system. The data fusion and processing module is configured to perform data cleaning and feature extraction on multi-dimensional datasets, and to perform weighted fusion of multi-dimensional features to generate a comprehensive network state index. The intelligent adjustment module is configured to make decisions based on the comprehensive network status index and scenario-based strategies, and dynamically adjust the parameters of the mobile communication base station antenna feeder system.

[0046] Example 2 like Figure 2 As shown, this embodiment provides a real-time antenna and feeder adjustment system based on multi-dimensional data weighted fusion, which is divided into five layers from bottom to top, and the functions of each layer are as follows: Data source layer: Provides various types of raw data, including base station measurement report data, terminal equipment measurement data, network management logs, and base station special integration module data, providing basic input for the system.

[0047] Data Acquisition Layer: Responsible for raw data acquisition, unifying and aggregating data from underlying data sources to prepare for subsequent processing.

[0048] Data Fusion and Processing Layer: 1) Data Cleaning: Remove noise and handle missing values ​​to improve data quality. 2) Feature Extraction: Extract key features (such as coverage, capacity, and other dimensional features) from the raw data. 3) XGBoost Weighted Fusion: Use the XGBoost algorithm to weighted fuse multi-dimensional features to generate a comprehensive network state index (such as CNSI), realizing intelligent data processing and fusion.

[0049] Intelligent adjustment layer: Based on the fused data, it makes intelligent decisions according to scenario strategies (coverage priority, experience priority, other alarm analysis), dynamically adjusts network parameters (such as antenna and feeder configuration), and adapts to different business needs (such as coverage optimization, experience guarantee, alarm handling).

[0050] Execution and Monitoring Layer: 1) Hardware Execution: Implementing adjustment strategies on hardware devices such as base stations to achieve actual adjustments to network parameters. 2) Closed-Loop Feedback: Sending execution results (such as adjusted data) back to the data processing layer to form a closed-loop optimization, continuously iterating to improve system performance.

[0051] Embodiment 3 As Figure 3 shown, the embodiment provides a real-time adjustment method for antenna and feeder based on multi-dimensional data weighted fusion, including data preprocessing stage, XGBoost model stage, result output stage, which are described in detail as follows.

[0052] I. Data preprocessing stage Six-dimensional original data input: Collect six-dimensional data covering coverage, capacity, reliability, interference, experience, and energy saving, and construct multi-source data set.

[0053] Data cleaning: Eliminate noise, fill missing values, and handle abnormal data to improve data quality.

[0054] Time window alignment: Unify different frequency data to 5-minute time granularity to ensure spatio-temporal consistency.

[0055] Data standardization: Eliminate dimensional differences and use different normalization methods for positive / negative indicators.

[0056] Feature engineering: Construct cross-features (such as "coverage x capacity") and spatio-temporal features to enhance data representation capability.

[0057] II. XGBoost model stage Training / test set segmentation: Divide data in 80% / 20% ratio for model training and independent verification.

[0058] Model initialization: Set max_depth=6, learning_rate=0.1, etc. parameters to build the basic model architecture.

[0059] Model training: Learn the nonlinear relationship of six-dimensional data through gradient boosting iteration, and output feature importance.

[0060] Model evaluation: Use MSE (Mean Squared Error), RMSE (Root Mean Squared Error), etc. indicators to verify model accuracy and ensure reliable fusion results.

[0061] Online model update: Based on new data incremental learning, regularly fine-tune parameters to adapt to network changes.

[0062] III. Result output stage Dimension weight calculation: Generate dynamic weights (such as coverage 25%, experience 20%) based on feature importance to support scenario adaptation.

[0063] CNSI calculation: Weighted fusion of six-dimensional scores, output 0-100 points of comprehensive network status index.

[0064] Adjustment decision output: generate the antenna feeder adjustment strategy according to the CNSI classification (<60 emergency / 60-80 regular / ≥80 energy saving).

[0065] Feedback to model update: use the adjusted data as new samples to optimize the model weight distribution in a closed loop.

[0066] Embodiment 4 This embodiment is based on embodiment 1: This embodiment provides a computer device, comprising a memory and a processor, the memory stores a computer program, and the processor implements the method of embodiment 1 when executing the computer program. The computer program can be in the form of source code, object code, executable file or some intermediate form, etc.

[0067] Embodiment 5 This embodiment is based on embodiment 1: This embodiment provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the method of embodiment 1. The computer program can be in the form of source code, object code, executable file or some intermediate form, etc. The storage medium includes any entity or device capable of carrying computer program code, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content of the storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the storage medium does not include electrical carrier signals and telecommunication signals.

[0068] The above only describes the preferred embodiments of the present application, and it should be understood that the present application is not limited to the forms disclosed herein, and should not be considered as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concepts described herein, by the above-mentioned teaching or related art or knowledge. The modifications and changes made by those skilled in the art without departing from the spirit and scope of the present application shall be within the scope of protection of the appended claims of the present application.

[0069] It is apparent that, for the method embodiments described previously, the steps of the methods have been described as being arranged in a particular order. However, it is to be appreciated that this is merely one example, and that the steps of the methods can be arranged in other orders or performed contemporaneously. Furthermore, it is to be appreciated that the embodiments described in the specification are merely preferred embodiments, and that the steps of the methods need not be performed in the order described.

Claims

1. A real-time antenna feeder adjustment method based on multi-dimensional data weighted fusion, characterized in that, include: Based on various types of raw data related to mobile communication base station antenna feeder systems, multi-dimensional data are collected and multi-dimensional datasets are constructed. Data cleaning and feature extraction are performed on multi-dimensional datasets, and the multi-dimensional features are weighted and fused to generate a comprehensive network state index. Based on the comprehensive network status index, decisions are made according to scenario strategies, and the parameters of the mobile communication base station antenna feeder system are dynamically adjusted.

2. The method for real-time adjustment of antenna feeders based on multi-dimensional data weighted fusion according to claim 1, characterized in that, The process of cleaning and extracting features from a multi-dimensional dataset, and then weighting and fusing these features to generate a comprehensive network state index, includes: Heterogeneous data is time-aligned, dimensionally normalized, and derivation features are constructed, including dimensional cross features and spatiotemporal features. Feature importance analysis based on XGBoost algorithm, automatically adjusting feature weights of each dimension according to network scenario; The comprehensive network status index is output as the core basis for adjusting the parameters of the mobile communication base station antenna feeder system.

3. The method for real-time adjustment of antenna feeders based on multi-dimensional data weighted fusion according to claim 1, characterized in that, The process of making decisions based on a comprehensive network state index and scenario-based strategies, and dynamically adjusting the parameters of the mobile communication base station antenna feeder system, includes: Based on the comprehensive network status index, a three-level decision-making system is set up, consisting of emergency adjustment, routine optimization, and energy-saving mode. Based on the scenario strategy, the corresponding decisions are sent to the hardware devices to execute the actual adjustment of the parameters of the mobile communication base station antenna feeder system. The results of parameter adjustments for the mobile communication base station antenna feeder system are transmitted back, and continuous iteration is used to form a closed-loop optimization.

4. The method for real-time adjustment of antenna feeders based on multi-dimensional data weighted fusion according to claim 1, characterized in that, The various types of raw data related to the mobile communication base station antenna feeder system include: base station measurement report data, terminal equipment measurement data, network management system log data, signaling monitoring data, and base station energy efficiency management module data.

5. The method for real-time adjustment of antenna feeders based on multi-dimensional data weighted fusion according to claim 1, characterized in that, The multi-dimensional data includes: coverage quality data, capacity efficiency data, connection reliability data, interference level data, user experience perception data, and energy efficiency data.

6. The method for real-time adjustment of antenna feeders based on multi-dimensional data weighted fusion according to claim 5, characterized in that, The evaluation metrics for the coverage quality dimension data include reference signal received power, reference signal received quality, and radio resource control establishment success rate. The original data sources include base station measurement report data and terminal device measurement data. The evaluation metrics for the capacity efficiency dimension data include physical resource block utilization, user density, and uplink / downlink throughput. The original data sources include base station measurement report data.

7. The method for real-time adjustment of antenna feeders based on multi-dimensional data weighted fusion according to claim 5, characterized in that, The evaluation metrics for the connection reliability dimension data include handover success rate, call drop rate, wireless link failure rate, and fault recovery time. The raw data sources include network management system log data and signaling monitoring data.

8. The method for real-time adjustment of antenna feeders based on multi-dimensional data weighted fusion according to claim 5, characterized in that, The evaluation metrics for the interference level dimension data include signal-to-interference-plus-noise ratio (SINR) and interference suppression ratio (ISNR), with raw data sources including base station measurement report data and terminal device measurement data. The evaluation metrics for the user experience perception dimension data include end-to-end latency, packet loss rate, video stuttering rate, and voice quality score, with raw data sources including base station measurement report data and terminal device measurement data.

9. The method for real-time adjustment of antenna feeders based on multi-dimensional data weighted fusion according to claim 5, characterized in that, The evaluation indicators for the energy efficiency dimension data include energy consumption per unit flow, dynamic power consumption ratio, and idle resource shutdown time. The raw data source includes data from the base station energy efficiency management module.

10. A real-time antenna feeder adjustment system based on multi-dimensional data weighted fusion, characterized in that, include: The data acquisition module is configured to collect multi-dimensional data and construct a multi-dimensional dataset based on various types of raw data related to the mobile communication base station antenna feeder system. The data fusion and processing module is configured to perform data cleaning and feature extraction on multi-dimensional datasets, and to perform weighted fusion of multi-dimensional features to generate a comprehensive network state index. The intelligent adjustment module is configured to make decisions based on the comprehensive network status index and scenario-based strategies, and dynamically adjust the parameters of the mobile communication base station antenna feeder system.

Images