Artificial Intelligence-Based Communication Base Station Monitoring Method and System

Abstract

This application provides an artificial intelligence-based communication base station monitoring method and system, which acquires operation and maintenance data of various components related to a first target base station; calculates the target fault degree; acquires a historical fault type set; acquires a first fault parameter from a first training set; acquires a second fault parameter from a second training set; performs fault detection on various components of the first target base station based on the first fault parameter; acquires a second target base station; fuses the first and second fault parameters to obtain a third fault parameter; performs fault detection on various components of the second target base station based on the third fault parameter; and calculates the predicted fault probability of each component of the second target base station. This method not only improves the efficiency of data processing during monitoring, ensuring that operation and maintenance personnel can perform timely maintenance on communication base stations, but also generates accurate fault probability prediction results by combining the dynamic changes of environmental variables, which helps to improve the level of operation and maintenance.

Description

Technical Field

[0001] This application relates to the field of communication base station technology, specifically to a communication base station monitoring method and system based on artificial intelligence. Background Technology

[0002] With the rapid development of my country's telecommunications industry and the gradual improvement of network coverage, the role of communication base stations has become increasingly prominent. They possess functions such as communication coverage, engineering maintenance, transmission conversion, system expansion, and network access, and are responsible for providing better call services to customers. In the operation of base stations, the most important task is daily maintenance, as its proper functioning directly affects the smooth operation of services. To achieve real-time monitoring of each communication base station, management rooms are typically set up at each base station to monitor the parameters and status of each station.

[0003] In existing technologies, sensors are first used to acquire parameters related to various components of a communication base station, which are then sent to a management room for technical personnel to review. However, this process presents at least the following problems: Firstly, due to the high redundancy of the parameters, maintenance personnel in the management room prioritize handling urgent parameters while delaying the processing of general parameters. This results in low response efficiency for general parameters and an inability to promptly detect potential faults. Consequently, the operational status of components deteriorates further, increasing the failure rate. If maintenance personnel fail to repair components in a timely manner, the risk of communication base station failure increases. Secondly, maintenance personnel typically rely solely on their own experience to analyze parameter changes and struggle to cope with dynamic changes in complex environments, such as sudden and periodic environmental shifts. This hinders reliable and accurate prediction of potential faults, leading to low maintenance efficiency and high maintenance difficulty.

[0004] Therefore, developing an artificial intelligence-based communication base station monitoring method and system has significant practical implications and broad application prospects. Summary of the Invention

[0005] The purpose of this application is to provide an artificial intelligence-based communication base station monitoring method and system, so as to improve the efficiency of data processing during the monitoring of communication base stations, cope with the dynamic changes in complex environments, and enhance the accuracy and interpretability of fault prediction.

[0006] In a first aspect, embodiments of this application provide a communication base station monitoring method based on artificial intelligence, including:

[0007] Obtain operation and maintenance data for each component related to the first target base station;

[0008] Identify the corresponding feature factor set from the operation and maintenance data. The feature factor set includes the fault feature factor set and the stability feature factor set.

[0009] The fault types of each component related to the first target base station are obtained based on the fault feature factor set;

[0010] Calculate the target fault degree based on the fault characteristic factor set and the stability characteristic factor set;

[0011] When the target fault degree reaches the fault degree threshold, the historical fault type set related to the first target base station is obtained. The historical fault type set includes a first training set and a second training set. The first training set includes a plurality of first historical fault types, and the second training set includes a plurality of second historical fault types.

[0012] Obtain first fault parameters corresponding to each first historical fault type from the first training set, with each first fault parameter corresponding to a first historical fault type;

[0013] Obtain the second fault parameters corresponding to each second historical fault type from the second training set, with each second fault parameter corresponding to a second historical fault type.

[0014] When the matching degree between the fault type and a first historical fault type reaches a first matching degree threshold, fault detection is performed on each component of the first target base station according to the first fault parameters.

[0015] When the matching degree between the fault type and a second historical fault type reaches the second matching degree threshold, a second target base station in the same preset space as the first target base station is obtained, and there is at least one second target base station in the preset space.

[0016] By combining the first fault parameters and the second fault parameters, the third fault parameters are obtained;

[0017] Based on the third fault parameter, fault detection is performed on each component of the second target base station to obtain the fault probability and environmental variables:

[0018] Based on the failure probability and environmental variables, obtain the time-series data of failure probability and environmental variable of each component of the second target base station within a certain time period;

[0019] After preprocessing the time series data of failure probability and environmental variables, the data is input into a pre-trained prediction model to obtain the predicted failure probability of each component of the second target base station within a set time period in the future.

[0020] The prediction model is obtained by training a neural network based on high- and low-frequency cycles and environmental variables using historical fault probability time-series data samples of various components of the second target base station and historical environmental variable time-series data samples.

[0021] Optionally, the step of identifying the corresponding feature factor set from the operation and maintenance data, the feature factor set including a fault feature factor set and a stability feature factor set, specifically includes:

[0022] Time-domain processing is performed on each piece of operational data to obtain a sequence set;

[0023] Obtain the fault characteristic factor set and the stability characteristic factor set from the sequence set.

[0024] Optionally, the method further includes: performing fault detection on each component of the second target base station based on the third fault parameters, and obtaining a fault detection report;

[0025] The fault detection report is transmitted to the operation and maintenance center for policy analysis and optimization to obtain regional joint operation and maintenance policy parameters. Based on the regional joint operation and maintenance policy parameters, fault operation and maintenance management is carried out in the regions of the first target base station and the second target base station.

[0026] Optionally, the method involves pre-training the prediction model in the following manner:

[0027] Acquire historical fault probability time-series data samples and historical environmental variable time-series data samples for each component of the second target base station;

[0028] After preprocessing the historical failure probability time series data samples and the historical environmental variable time series data samples, the historical failure probability time series data samples and the historical environmental variable time series data samples are transformed into supervised failure probability data and supervised environmental variable data, respectively.

[0029] Supervised data on failure probabilities and supervised data on environmental variables are input into a neural network for training. A prediction model is obtained after the preset training cutoff condition is met.

[0030] Optionally, the neural network based on high- and low-frequency recurrent and environmental variable synergy includes a data input module, a data preprocessing module, a high-frequency feature extraction module, a low-frequency feature extraction module, an environmental variable processing module, a feature fusion module, a recurrent neural network module, and an output module connected in sequence.

[0031] Optionally, when the matching degree between the fault type and the first historical fault type reaches a first matching degree threshold, fault detection is performed on each component of the first target base station according to the first fault parameters, specifically including:

[0032] Calculate the matching degree between the fault type and the first historical fault type based on the matching degree measurement algorithm;

[0033] When the matching degree reaches the first matching degree threshold, fault detection is performed on each target component based on the first fault parameter.

[0034] Optionally, calculating the matching degree between the fault type and the first historical fault type according to the matching degree measurement algorithm specifically includes:

[0035] Calculate the first state matrix for the fault type;

[0036] Calculate the second state matrix for the first historical fault type;

[0037] Calculate the matching degree between the first state matrix and the second state matrix.

[0038] Optionally, the method further includes: calculating a first fault-related factor set related to the first target base station based on a set of historical fault types, wherein the first fault-related factor set includes a plurality of first key fault-related factors and a plurality of second key fault-related factors;

[0039] Obtain the first weight parameters corresponding to each first critical fault-related factor from the first training set, wherein each first weight parameter corresponds to one first critical fault-related factor;

[0040] Obtain the second weight parameters corresponding to each second critical fault-related factor from the second training set, wherein each second weight parameter corresponds to a second critical fault-related factor;

[0041] Calculate the critical parameters of the first critical failure corresponding to the factors related to the first critical failure;

[0042] Calculate the critical parameters of the second critical failure corresponding to the factors related to the second critical failure;

[0043] Calculate the first degree of fusion for each critical parameter of the first fault and the critical parameter of the second fault;

[0044] The component whose first degree of integration reaches the integration threshold is identified as the first key component of the first target base station.

[0045] Optionally, the method further includes: performing fault detection on the first target base station when the first fault critical parameter is equal to 0 and the second fault critical parameter is equal to 0.

[0046] Secondly, embodiments of this application provide an artificial intelligence-based communication base station monitoring system, comprising:

[0047] The first acquisition module is used to acquire operation and maintenance data of various components related to the first target base station;

[0048] The identification module is used to identify the corresponding feature factor set from the operation and maintenance data. The feature factor set includes the fault feature factor set and the stability feature factor set.

[0049] The fault type acquisition module is used to obtain the fault type of each component related to the first target base station based on the fault feature factor set;

[0050] The calculation module is used to calculate the target fault degree based on the fault characteristic factor set and the stability characteristic factor set;

[0051] The second acquisition module is used to acquire a set of historical fault types related to the first target base station when the target fault degree reaches the fault degree threshold. The set of historical fault types includes a first training set and a second training set. The first training set includes a plurality of first historical fault types, and the second training set includes a plurality of second historical fault types.

[0052] The second acquisition module is used to acquire first fault parameters corresponding to each first historical fault type from the first training set, and each first fault parameter corresponds to a first historical fault type.

[0053] The third acquisition module is used to acquire second fault parameters corresponding to each second historical fault type from the second training set, and each second fault parameter corresponds to a second historical fault type.

[0054] The first fault detection module is used to perform fault detection on each component of the first target base station according to the first fault parameters when the matching degree between the fault type and a first historical fault type reaches a first matching degree threshold.

[0055] The fourth acquisition module is used to acquire a second target base station in the same preset space as the first target base station when the matching degree between the fault type and a second historical fault type reaches a second matching degree threshold. There is at least one second target base station in the preset space.

[0056] The fusion module is used to fuse the first fault parameters and the second fault parameters to obtain the third fault parameters.

[0057] The second fault detection module performs fault detection on various components of the second target base station based on the third fault parameters, and obtains the fault probability and environmental variables:

[0058] The fifth acquisition module is used to acquire fault probability time-series data and environmental variable time-series data of each component of the second target base station within a certain time period based on fault probability and environmental variables.

[0059] The prediction module is used to preprocess the time series data of fault probability and environmental variables, and then input them into the pre-trained prediction model to obtain the predicted fault probability of each component of the second target base station in the future within a set time period.

[0060] The prediction model is obtained by training a neural network based on high- and low-frequency cycles and environmental variables using historical fault probability time-series data samples of various components of the second target base station and historical environmental variable time-series data samples.

[0061] Compared to existing technologies, the artificial intelligence-based communication base station monitoring method and system provided in this application acquires operation and maintenance data of various components related to a first target base station; identifies corresponding feature factor sets from the operation and maintenance data, the feature factor sets including fault feature factor sets and stability feature factor sets; obtains the fault types of various components related to the first target base station based on the fault feature factor sets; calculates the target fault degree based on the fault feature factor sets and stability feature factor sets; when the target fault degree reaches a fault degree threshold, acquires a historical fault type set related to the first target base station, the historical fault type set including a first training set and a second training set; the first training set includes a plurality of first historical fault types, the second... The training set includes a plurality of second historical fault types; first fault parameters corresponding to each first historical fault type are obtained from the first training set, with each first fault parameter corresponding to one first historical fault type; second fault parameters corresponding to each second historical fault type are obtained from the second training set, with each second fault parameter corresponding to one second historical fault type; when the matching degree between a fault type and a first historical fault type reaches a first matching degree threshold, fault detection is performed on each component of the first target base station based on the first fault parameters; when the matching degree between a fault type and a second historical fault type reaches a second matching degree threshold, second fault parameters within the same preset space as the first target base station are obtained. The target base station includes at least one second target base station within a preset space. A third fault parameter is obtained by fusing the first and second fault parameters. Fault detection is performed on each component of the second target base station based on the third fault parameter to obtain fault probabilities and environmental variables. Based on the fault probabilities and environmental variables, time-series data of fault probabilities and environmental variables for each component of the second target base station are obtained over a certain time period. After preprocessing the time-series data, these data are input into a pre-trained prediction model to obtain the predicted fault probabilities of each component of the second target base station within a future set time period. The prediction model is based on the historical data of each component of the second target base station. The data is obtained by training a neural network based on high- and low-frequency cycles and environmental variable synergy using historical fault probability time-series data samples and historical environmental variable time-series data samples. This not only improves the efficiency of data processing during monitoring, ensuring that maintenance personnel can perform timely maintenance on communication base stations based on the first, second, and third fault parameters, thus guaranteeing the effective operation of communication base stations, but also generates accurate fault probability prediction results by combining dynamic changes in environmental variables. This can effectively cope with the impact of sudden and periodic environmental changes, achieve accurate prediction of communication base station faults, make the final prediction results more interpretable, and help improve the level of operation and maintenance. Attached Figure Description

[0062] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0063] Figure 1 A flowchart of the AI-based communication base station monitoring method provided in this application is shown;

[0064] Figure 2 A schematic diagram of the artificial intelligence-based communication base station monitoring system provided in this application is shown;

[0065] Figure 3 A schematic diagram of an electronic device provided in this application is shown. Detailed Implementation

[0066] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0067] It should be noted that, unless otherwise stated, the technical or scientific terms used in this application shall have the ordinary meaning as understood by one of ordinary skill in the art to which this application pertains.

[0068] Furthermore, the terms "first" and "second," etc., are used to distinguish different objects, not to describe a specific order. Additionally, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to those processes, methods, products, or apparatuses.

[0069] Please refer to Figure 1 , Figure 1 The communication base station monitoring method based on artificial intelligence provided in this application includes the following steps:

[0070] S101. Obtain the operation and maintenance data of each component related to the first target base station. The operation and maintenance data may include historical data and real-time data.

[0071] S102. Identify the corresponding feature factor set from the operation and maintenance data. The feature factor set includes the fault feature factor set and the stability feature factor set. Specifically, it can be automatically identified through feature extraction methods such as principal component analysis (PCA) and linear discriminant analysis (LDA).

[0072] S103. Obtain the fault type of each component related to the first target base station based on the fault feature factor set, such as by training a fault type identification model (support vector machine, random forest, gradient boosting decision tree, neural network).

[0073] S104. Calculate the target failure degree based on the fault characteristic factor set and the stability characteristic factor set, such as by calculating the target failure degree through feature weighting.

[0074] S105. When the target fault degree reaches the fault degree threshold, obtain the historical fault type set related to the first target base station. The historical fault type set includes a first training set and a second training set. The first training set includes a plurality of first historical fault types, and the second training set includes a plurality of second historical fault types.

[0075] S106. Obtain the first fault parameters corresponding to each first historical fault type from the first training set, where each first fault parameter corresponds to a first historical fault type.

[0076] S107. Obtain the second fault parameters corresponding to each second historical fault type from the second training set. Each second fault parameter corresponds to a second historical fault type.

[0077] S108. When the matching degree between the fault type and a first historical fault type reaches the first matching degree threshold, fault detection is performed on each component of the first target base station according to the first fault parameters.

[0078] S109. When the matching degree between the fault type and a second historical fault type reaches the second matching degree threshold, a second target base station in the same preset space as the first target base station is obtained. There is at least one second target base station in the preset space. The preset space includes the first target base station and the second target base station. The preset space is a circle with a circumference radius R, and the center of the preset space is the first target base station.

[0079] S110. By fusing the first fault parameter and the second fault parameter (such as through a weighted fusion calculation method), the third fault parameter is obtained.

[0080] S111. Based on the third fault parameters, perform fault detection on each component of the second target base station to obtain the fault probability and environmental variables:

[0081] S112. Based on the fault probability and environmental variables, obtain the fault probability time series data and environmental variable time series data of each component of the second target base station within a certain time period.

[0082] S113. After preprocessing the fault probability time series data and environmental variable time series data, input them into the pre-trained prediction model to obtain the predicted fault probability of each component of the second target base station in the future set time period.

[0083] The prediction model is obtained by training a neural network based on high- and low-frequency cycles and environmental variables using historical fault probability time-series data samples of various components of the second target base station and historical environmental variable time-series data samples.

[0084] In step S102, the identification of the corresponding feature factor set from the operation and maintenance data, the feature factor set including the fault feature factor set and the stability feature factor set, specifically includes:

[0085] Time-domain processing is performed on each piece of operational data to obtain a sequence set;

[0086] Obtain the fault feature factor set and the stability feature factor set from the sequence set. The first fault feature factor includes the following information:

[0087] Abnormal changes prior to a failure, such as sudden increases, sudden decreases, or changes in pattern;

[0088] Characteristics of the fault occurrence, such as the duration and magnitude of the fault state;

[0089] Recovery modes after a failure, such as recovery speed and recovery extent.

[0090] The first stable characteristic factor includes the following information:

[0091] The levels of indicators when the system is running normally, such as steady-state values ​​and fluctuation range;

[0092] Indicators of system stability, such as coefficient of variation and standard deviation.

[0093] The above methods can extract a set of feature factors that are helpful for fault prediction from operational data. These feature factor sets will serve as input to the neural network, helping the model learn the complex relationship between operational data and system faults, thereby improving the accuracy and efficiency of fault prediction.

[0094] In step S112, the fault detection of each component of the second target base station is performed according to the third fault parameter to obtain the fault probability and environmental variables. Specifically, based on the fault detection results, a statistical model (such as logistic regression) is used to calculate the fault probability of each component within a certain time period, and environmental variable data related to fault detection are collected within the same time period. Environmental variables may include physical environment (temperature, humidity, pressure, vibration, etc.) and operating environment (system load, operating mode, maintenance history, etc.).

[0095] The aforementioned fault probability time series data consists of fault probability data over a certain period of time, arranged according to a set resolution. The resolution can be set to specific time intervals such as hours, days, weeks, or months. For example, fault probability data can be statistically analyzed into time series data with a daily resolution; similarly, environmental variable time series data can be obtained.

[0096] In some specific embodiments, the method further includes: performing fault detection on each component of the second target base station based on the third fault parameters, and obtaining a fault detection report;

[0097] The fault detection report is transmitted to the operation and maintenance center for policy analysis and optimization to obtain regional joint operation and maintenance policy parameters. Based on the regional joint operation and maintenance policy parameters, fault operation and maintenance management is carried out in the regions of the first target base station and the second target base station.

[0098] The process of obtaining regional joint operation and maintenance strategy parameters includes: performing feature matching between fault detection reports and the fault operation and maintenance strategy library to obtain fault operation and maintenance strategies.

[0099] Based on the fault operation and maintenance strategy, a fault operation and maintenance strategy space is constructed, and the fault detection report is parsed according to the fault operation and maintenance strategy space to obtain the operation and maintenance strategy parameter thresholds.

[0100] The operation and maintenance strategy parameters are obtained by performing global optimization of the parameters within the preset search range and within the threshold of the operation and maintenance strategy parameters.

[0101] The operation and maintenance strategy parameters are linked and integrated according to the inspection sequence of the base stations to obtain the regional linkage operation and maintenance strategy parameters.

[0102] In this embodiment, the method involves pre-training the prediction model in the following manner:

[0103] Acquire historical fault probability time-series data samples and historical environmental variable time-series data samples of each component of the second target base station; among them, historical fault probability time-series data samples: collect fault probability data of each component within a certain time range. These data usually exist in the form of time series, recording the probability of component failure at each time point or time period; historical environmental variable time-series data samples: collect environmental variable data related to fault probability, such as temperature, humidity, pressure, etc., which are also recorded in the form of time series.

[0104] After preprocessing (such as data cleaning and normalization) the historical failure probability time series data samples and the historical environmental variable time series data samples, the historical failure probability time series data samples and the historical environmental variable time series data samples are transformed into supervised failure probability data and supervised environmental variable data, respectively.

[0105] Supervised data on failure probabilities and supervised data on environmental variables are input into a neural network for training. A prediction model is obtained after the preset training cutoff condition is met.

[0106] In this embodiment, the neural network based on high- and low-frequency recurrent neural networks and environmental variable synergy includes a sequentially connected data input module, a data preprocessing module, a high-frequency feature extraction module, a low-frequency feature extraction module, an environmental variable processing module, a feature fusion module, a recurrent neural network module, and an output module, as detailed below:

[0107] Data input module: responsible for receiving input data, including various data sources such as high and low frequency signals and environmental variables;

[0108] Data preprocessing module: used for data cleaning and standardization;

[0109] High-frequency feature extraction module: used to capture short-term and local patterns in time series and convert the time series into the frequency domain to extract high-frequency components;

[0110] Low-frequency feature extraction module: used to capture long-term dependencies in time series and extract low-frequency trends and seasonal components of time series;

[0111] The environmental variable processing module includes an embedding layer and a fully connected layer. The embedding layer is used to convert categorical environmental variables into vectors of fixed dimensions, and the fully connected layer is used to perform non-linear transformations on numerical environmental variables.

[0112] Feature fusion module: used to concatenate and fuse high- and low-frequency feature vectors with environmental variable feature vectors in terms of dimensions;

[0113] Recurrent Neural Network Module: Further processes the fused feature sequences to capture complex temporal relationships.

[0114] Output module: Based on the specific task requirements, the fused information is transformed into the required output results.

[0115] In step S108, when the matching degree between the fault type and the first historical fault type reaches a first matching degree threshold, fault detection is performed on each component of the first target base station according to the first fault parameters, specifically including:

[0116] The matching degree is calculated based on the matching degree measurement algorithm, and the matching degree between the fault type and the first historical fault type is calculated. The specific measurement method can be cosine similarity, Euclidean distance, Manhattan distance, Jaccard similarity coefficient, etc.

[0117] When the matching degree reaches the first matching degree threshold, fault detection is performed on each target component based on the first fault parameter.

[0118] The above method can effectively identify current faults similar to the first historical fault type, and detect components based on specific fault parameters, thereby improving the accuracy and efficiency of fault detection.

[0119] In some implementations, calculating the matching degree between the fault type and the first historical fault type according to the matching degree measurement algorithm specifically includes:

[0120] Calculate the first state matrix of the fault type; the first state matrix is ​​a matrix representing the current state of the fault type, which usually contains the state of multiple features or parameters; these features may include the time of fault occurrence, frequency, scope of impact, relevant sensor readings, etc.

[0121] Calculate the second state matrix for the first historical fault type; the second state matrix is ​​similar to the first state matrix, and can also represent the state of the first historical fault type.

[0122] Calculating the matching degree between the first state matrix and the second state matrix can specifically include:

[0123] Align the two matrices to ensure that the same features are being compared;

[0124] A matching score is obtained by applying a metric algorithm.

[0125] In some embodiments, the method further includes: calculating a first fault-related factor set associated with the first target base station based on a set of historical fault types, wherein the first fault-related factor set includes a plurality of first critical fault-related factors and a plurality of second critical fault-related factors;

[0126] Obtain the first weight parameters corresponding to each first critical fault-related factor from the first training set, wherein each first weight parameter corresponds to one first critical fault-related factor;

[0127] Obtain the second weight parameters corresponding to each second critical fault-related factor from the second training set, wherein each second weight parameter corresponds to a second critical fault-related factor;

[0128] Calculate the critical parameters of the first critical failure corresponding to the factors related to the first critical failure;

[0129] Calculate the critical parameters of the second critical failure corresponding to the factors related to the second critical failure;

[0130] Calculate the first degree of integration for each critical parameter of the first fault and the critical parameter of the second fault; the first degree of integration is a comprehensive index used to evaluate the overall impact of each critical parameter of the fault.

[0131] Components whose first degree of integration reaches the integration threshold are identified as the first critical components of the first target base station. This method effectively identifies the most critical components of the base station, enabling targeted maintenance and monitoring to ensure stable operation.

[0132] In some implementations, the method further includes: when the first fault critical parameter is equal to 0 and the second fault critical parameter is equal to 0, performing fault detection on the first target base station. Through this detailed fault detection process, the risk of the base station can be more comprehensively assessed, further ensuring the safety and reliability of the base station.

[0133] Furthermore, the step of determining the component whose first fusion degree reaches the fusion degree threshold as the first key component of the first target base station specifically includes:

[0134] Calculate the first degree of fusion for each critical parameter of the first fault and the critical parameter of the second fault;

[0135] By combining (e.g., weighted fusion) the two first fusion degrees above, a second fusion degree is obtained;

[0136] The component whose second degree of fusion reaches the fusion threshold is identified as the first key component of the first target base station.

[0137] This process identifies components critical to base station operation, whose failures can significantly impact the overall performance of the base station. Identifying these critical components helps the operations and maintenance team prioritize and ensure their reliability, thereby improving the stability and service quality of the entire base station system.

[0138] Furthermore, the step of identifying the component whose second fusion degree reaches the fusion degree threshold as the first key component of the first target base station specifically includes:

[0139] Calculate the first degree of fusion for each critical parameter of the first fault and the critical parameter of the second fault;

[0140] By combining (e.g., weighted fusion) the two first fusion degrees above, a second fusion degree is obtained;

[0141] The component with a second fusion degree greater than or equal to 0 is identified as the first key component of the first target base station.

[0142] This process ensures that all components that have a significant impact on base station operation are identified, allowing for targeted maintenance and monitoring.

[0143] Compared to existing technologies, the artificial intelligence-based communication base station monitoring method and system provided in this application acquires operation and maintenance data of various components related to a first target base station; identifies corresponding feature factor sets from the operation and maintenance data, the feature factor sets including fault feature factor sets and stability feature factor sets; obtains the fault types of various components related to the first target base station based on the fault feature factor sets; calculates the target fault degree based on the fault feature factor sets and stability feature factor sets; when the target fault degree reaches a fault degree threshold, acquires a historical fault type set related to the first target base station, the historical fault type set including a first training set and a second training set; the first training set includes a plurality of first historical fault types, the second... The training set includes a plurality of second historical fault types; first fault parameters corresponding to each first historical fault type are obtained from the first training set, with each first fault parameter corresponding to one first historical fault type; second fault parameters corresponding to each second historical fault type are obtained from the second training set, with each second fault parameter corresponding to one second historical fault type; when the matching degree between a fault type and a first historical fault type reaches a first matching degree threshold, fault detection is performed on each component of the first target base station based on the first fault parameters; when the matching degree between a fault type and a second historical fault type reaches a second matching degree threshold, second fault parameters within the same preset space as the first target base station are obtained. The target base station includes at least one second target base station within a preset space. A third fault parameter is obtained by fusing the first and second fault parameters. Fault detection is performed on each component of the second target base station based on the third fault parameter to obtain fault probabilities and environmental variables. Based on the fault probabilities and environmental variables, time-series data of fault probabilities and environmental variables for each component of the second target base station are obtained over a certain time period. After preprocessing the time-series data, these data are input into a pre-trained prediction model to obtain the predicted fault probabilities of each component of the second target base station within a future set time period. The prediction model is based on the historical data of each component of the second target base station. The data is obtained by training a neural network based on high- and low-frequency cycles and environmental variable synergy using historical fault probability time-series data samples and historical environmental variable time-series data samples. This not only improves the efficiency of data processing during monitoring, ensuring that maintenance personnel can perform timely maintenance on communication base stations based on the first, second, and third fault parameters, thus guaranteeing the effective operation of communication base stations, but also generates accurate fault probability prediction results by combining dynamic changes in environmental variables. This can effectively cope with the impact of sudden and periodic environmental changes, achieve accurate prediction of communication base station faults, make the final prediction results more interpretable, and help improve the level of operation and maintenance.

[0144] In the above embodiments, an artificial intelligence-based communication base station monitoring method was provided. Correspondingly, this application also provides an artificial intelligence-based communication base station monitoring system. Figure 2 As shown, the AI-based communication base station monitoring system provided in this application embodiment can implement the above-described AI-based communication base station monitoring method. This AI-based communication base station monitoring system can be implemented through software, hardware, or a combination of both. For example, the AI-based communication base station monitoring system may include integrated or separate functional modules or units to perform the corresponding steps in the above methods, including:

[0145] The first acquisition module 101 is used to acquire operation and maintenance data of various components related to the first target base station;

[0146] The identification module 102 is used to identify the corresponding feature factor set from the operation and maintenance data. The feature factor set includes the fault feature factor set and the stability feature factor set.

[0147] The fault type acquisition module 103 is used to obtain the fault type of each component related to the first target base station based on the fault feature factor set;

[0148] Calculation module 104 is used to calculate the target fault degree based on the fault characteristic factor set and the stability characteristic factor set;

[0149] The second acquisition module 105 is used to acquire a set of historical fault types related to the first target base station when the target fault degree reaches the fault degree threshold. The set of historical fault types includes a first training set and a second training set. The first training set includes a plurality of first historical fault types, and the second training set includes a plurality of second historical fault types.

[0150] The second acquisition module 106 is used to acquire first fault parameters corresponding to each first historical fault type from the first training set, wherein each first fault parameter corresponds to a first historical fault type.

[0151] The third acquisition module 107 is used to acquire second fault parameters corresponding to each second historical fault type from the second training set, and each second fault parameter corresponds to a second historical fault type.

[0152] The first fault detection module 108 is used to perform fault detection on each component of the first target base station according to the first fault parameters when the matching degree between the fault type and a first historical fault type reaches a first matching degree threshold.

[0153] The fourth acquisition module 109 is used to acquire a second target base station in the same preset space as the first target base station when the matching degree between the fault type and a second historical fault type reaches a second matching degree threshold. There is at least one second target base station in the preset space.

[0154] The fusion module 110 is used to fuse the first fault parameters and the second fault parameters to obtain the third fault parameters;

[0155] The second fault detection module 111 performs fault detection on each component of the second target base station based on the third fault parameters, and obtains the fault probability and environmental variables:

[0156] The fifth acquisition module 112 is used to acquire fault probability time-series data and environmental variable time-series data of each component of the second target base station within a certain time period based on fault probability and environmental variables.

[0157] The prediction module 113 is used to preprocess the fault probability time series data and environmental variable time series data, and then input them into the pre-trained prediction model to obtain the predicted fault probability of each component of the second target base station in the future set time period.

[0158] The prediction model is obtained by training a neural network based on high- and low-frequency cycles and environmental variables using historical fault probability time-series data samples of various components of the second target base station and historical environmental variable time-series data samples.

[0159] The system and method provided in this application are based on the same inventive concept and have the same beneficial effects as the methods they employ, operate, or implement.

[0160] This application also provides an electronic device corresponding to the method provided in the foregoing embodiments. The electronic device may be an electronic device for a client, such as a mobile phone, laptop computer, tablet computer, desktop computer, etc., to execute the above method.

[0161] Please refer to Figure 3 This illustrates a schematic diagram of an electronic device provided by some embodiments of this application. For example... Figure 3 As shown, the electronic device 20 includes: a processor 200, a memory 201, a bus 202, and a communication interface 203. The processor 200, the communication interface 203, and the memory 201 are connected via the bus 202. The memory 201 stores a computer program that can run on the processor 200. When the processor 200 runs the computer program, it executes the method provided in any of the foregoing embodiments of this application.

[0162] The memory 201 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 203 (which can be wired or wireless), such as the Internet, wide area network, local area network, or metropolitan area network.

[0163] Bus 202 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. The memory 201 is used to store programs. After receiving an execution instruction, the processor 200 executes the program. The methods disclosed in any of the foregoing embodiments of this application can be applied to the processor 200, or implemented by the processor 200.

[0164] The processor 200 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of the processor 200 or by instructions in software form. The processor 200 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules may reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 201. The processor 200 reads the information in memory 201 and, in conjunction with its hardware, completes the steps of the above method.

[0165] The electronic devices and methods provided in the embodiments of this application are based on the same inventive concept and have the same beneficial effects as the methods they employ, operate, or implement.

[0166] This application also provides a computer-readable storage medium corresponding to the method provided in the foregoing embodiments, which stores a computer program (i.e., a program product) thereon. When the computer program is run by a processor, it executes the method provided in any of the foregoing embodiments.

[0167] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical and magnetic storage media, which will not be elaborated here.

[0168] The computer-readable storage medium provided in the above embodiments of this application and the method provided in the embodiments of this application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the applications stored therein.

[0169] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application, and they should all be covered within the scope of the claims and specification of this application.

Claims

1. A communication base station monitoring method based on artificial intelligence, characterized in that, Includes the following steps: Obtain operation and maintenance data for each component related to the first target base station; Identify the corresponding feature factor set from the operation and maintenance data. The feature factor set includes the fault feature factor set and the stability feature factor set. The fault feature factor set is used to capture abnormal changes before the fault, the characteristics when the fault occurs, and the recovery mode after the fault. The stability feature factor set is used to capture the indicators when the system is running normally and the indicators of system stability. The fault types of each component related to the first target base station are obtained based on the fault feature factor set; Calculate the target fault degree based on the fault characteristic factor set and the stability characteristic factor set; When the target fault degree reaches the fault degree threshold, the historical fault type set related to the first target base station is obtained. The historical fault type set includes a first training set and a second training set. The first training set includes a plurality of first historical fault types, and the second training set includes a plurality of second historical fault types. Obtain first fault parameters corresponding to each first historical fault type from the first training set, with each first fault parameter corresponding to a first historical fault type; Obtain the second fault parameters corresponding to each second historical fault type from the second training set, with each second fault parameter corresponding to a second historical fault type. When the matching degree between the fault type and a first historical fault type reaches a first matching degree threshold, fault detection is performed on each component of the first target base station according to the first fault parameters. When the matching degree between the fault type and a second historical fault type reaches the second matching degree threshold, a second target base station in the same preset space as the first target base station is obtained, and there is at least one second target base station in the preset space. By combining the first fault parameters and the second fault parameters, the third fault parameters are obtained; Based on the third fault parameter, fault detection is performed on each component of the second target base station to obtain the fault probability and environmental variables: Based on the failure probability and environmental variables, obtain the time-series data of failure probability and environmental variable of each component of the second target base station within a certain time period; After preprocessing the time series data of failure probability and environmental variables, the data is input into a pre-trained prediction model to obtain the predicted failure probability of each component of the second target base station within a set time period in the future. The prediction model is obtained by training a neural network based on high- and low-frequency cycles and environmental variables through historical fault probability time-series data samples of various components of the second target base station and historical environmental variable time-series data samples. The method further includes: Based on the third fault parameters, fault detection is performed on each component of the second target base station to obtain a fault detection report; The fault detection report is transmitted to the operation and maintenance center for strategy parsing and optimization to obtain regional linkage operation and maintenance strategy parameters. Based on the regional linkage operation and maintenance strategy parameters, fault operation and maintenance management is carried out in the regions of the first target base station and the second target base station. The process of obtaining regional linkage operation and maintenance strategy parameters includes: performing feature matching between the fault detection report and the fault operation and maintenance strategy library to obtain the fault operation and maintenance strategy. Based on the fault operation and maintenance strategy, a fault operation and maintenance strategy space is constructed, and the fault detection report is parsed according to the fault operation and maintenance strategy space to obtain the operation and maintenance strategy parameter thresholds. The operation and maintenance strategy parameters are obtained by performing global optimization of the parameters within the preset search range and within the threshold of the operation and maintenance strategy parameters. The operation and maintenance strategy parameters are linked and integrated according to the inspection sequence of the base station to obtain the regional linkage operation and maintenance strategy parameters. The method is described in which the prediction model is pre-trained in the following manner: Obtain historical fault probability time-series data samples and historical environmental variable time-series data samples for each component of the second target base station; After preprocessing the historical failure probability time series data samples and the historical environmental variable time series data samples, the historical failure probability time series data samples and the historical environmental variable time series data samples are transformed into supervised failure probability data and supervised environmental variable data, respectively. Supervised data on failure probability and supervised data on environmental variables are input into the neural network for training. A prediction model is obtained after the preset training cutoff condition is met. The neural network based on high- and low-frequency recurrent neural networks and environmental variables includes a sequentially connected data input module, a data preprocessing module, a high-frequency feature extraction module, a low-frequency feature extraction module, an environmental variable processing module, a feature fusion module, a recurrent neural network module, and an output module. The data input module receives input data from various data sources, including high- and low-frequency signals and environmental variables. The data preprocessing module cleans and standardizes the data. The high-frequency feature extraction module captures short-term and local patterns in the time series and converts the time series to the frequency domain to extract high-frequency components. The low-frequency feature extraction module captures long-term dependencies in the time series and extracts low-frequency trends and seasonal components. The environmental variable processing module includes an embedding layer and a fully connected layer. The embedding layer converts categorical environmental variables into fixed-dimensional vectors, while the fully connected layer performs nonlinear transformations on numerical environmental variables. The feature fusion module concatenates and fuses high- and low-frequency feature vectors with environmental variable feature vectors in terms of dimension. The recurrent neural network module further processes the fused feature sequence to capture complex temporal relationships. The output module transforms the fused information into the desired output results according to the specific task requirements. The method further includes: calculating a first fault-related factor set related to the first target base station based on a historical fault type set, wherein the first fault-related factor set includes a plurality of first key fault-related factors and a plurality of second key fault-related factors; Obtain the first weight parameters corresponding to each first critical fault-related factor from the first training set, wherein each first weight parameter corresponds to one first critical fault-related factor; Obtain the second weight parameters corresponding to each second critical fault-related factor from the second training set, wherein each second weight parameter corresponds to a second critical fault-related factor; Calculate the critical parameters of the first critical failure corresponding to the factors related to the first critical failure; Calculate the critical parameters of the second critical failure corresponding to the factors related to the second critical failure; Calculate the first degree of integration for each critical parameter of the first fault and the critical parameter of the second fault; the first degree of integration is a comprehensive index used to evaluate the overall impact of each critical parameter of the fault. Furthermore, the component whose first degree of fusion reaches the fusion threshold is identified as the first key component of the first target base station, specifically including: Calculate the first degree of fusion for each critical parameter of the first fault and the critical parameter of the second fault; By combining the two first degrees of fusion above, we obtain the second degree of fusion; The component whose second degree of fusion reaches the fusion degree threshold is identified as the first key component of the first target base station; Furthermore, the step of identifying the component whose second fusion degree reaches the fusion degree threshold as the first key component of the first target base station specifically includes: Calculate the first degree of fusion for each critical parameter of the first fault and the critical parameter of the second fault; By combining the two first degrees of fusion above, we obtain the second degree of fusion; The component with a second fusion degree greater than or equal to 0 is identified as the first key component of the first target base station.

2. The method according to claim 1, characterized in that, The step of identifying the corresponding feature factor set from the operation and maintenance data, including the fault feature factor set and the stability feature factor set, specifically includes: Time-domain processing is performed on each piece of operational data to obtain a sequence set; Obtain the fault characteristic factor set and the stability characteristic factor set from the sequence set.

3. The method according to claim 1, characterized in that, When the matching degree between the fault type and the first historical fault type reaches a first matching degree threshold, fault detection is performed on each component of the first target base station according to the first fault parameters, specifically including: Calculate the matching degree between the fault type and the first historical fault type based on the matching degree measurement algorithm; When the matching degree reaches the first matching degree threshold, fault detection is performed on each component of the first target base station according to the first fault parameters.

4. The method according to claim 3, characterized in that, The step of calculating the matching degree between the fault type and the first historical fault type according to the matching degree measurement algorithm specifically includes: Calculate the first state matrix for the fault type; Calculate the second state matrix for the first historical fault type; Calculate the matching degree between the first state matrix and the second state matrix.

5. The method according to claim 1, characterized in that, Also includes: When the first critical fault parameter is equal to 0 and the second critical fault parameter is equal to 0, fault detection is performed on the first target base station.

6. An artificial intelligence-based communication base station monitoring system, characterized in that, include: The first acquisition module is used to acquire operation and maintenance data of various components related to the first target base station; The identification module is used to identify the corresponding feature factor set from the operation and maintenance data. The feature factor set includes the fault feature factor set and the stability feature factor set. The fault feature factor set is used to capture abnormal changes before the fault, the characteristics when the fault occurs, and the recovery mode after the fault. The stability feature factor set is used to capture the indicators when the system is running normally and the indicators of system stability. The fault type acquisition module is used to obtain the fault type of each component related to the first target base station based on the fault feature factor set; The calculation module is used to calculate the target fault degree based on the fault characteristic factor set and the stability characteristic factor set; The second acquisition module is used to acquire a set of historical fault types related to the first target base station when the target fault degree reaches the fault degree threshold. The set of historical fault types includes a first training set and a second training set. The first training set includes a plurality of first historical fault types, and the second training set includes a plurality of second historical fault types. The second acquisition module is used to acquire first fault parameters corresponding to each first historical fault type from the first training set, and each first fault parameter corresponds to a first historical fault type. The third acquisition module is used to acquire second fault parameters corresponding to each second historical fault type from the second training set, and each second fault parameter corresponds to a second historical fault type. The first fault detection module is used to perform fault detection on each component of the first target base station according to the first fault parameters when the matching degree between the fault type and a first historical fault type reaches a first matching degree threshold. The fourth acquisition module is used to acquire a second target base station in the same preset space as the first target base station when the matching degree between the fault type and a second historical fault type reaches a second matching degree threshold. There is at least one second target base station in the preset space. The fusion module is used to fuse the first fault parameters and the second fault parameters to obtain the third fault parameters. The second fault detection module performs fault detection on various components of the second target base station based on the third fault parameters, and obtains the fault probability and environmental variables: The fifth acquisition module is used to acquire fault probability time-series data and environmental variable time-series data of each component of the second target base station within a certain time period based on fault probability and environmental variables. The prediction module is used to preprocess the time series data of fault probability and environmental variables, and then input them into the pre-trained prediction model to obtain the predicted fault probability of each component of the second target base station in the future within a set time period. The prediction model is obtained by training a neural network based on high- and low-frequency cycles and environmental variables through historical fault probability time-series data samples of various components of the second target base station and historical environmental variable time-series data samples. The system also includes: Based on the third fault parameters, fault detection is performed on each component of the second target base station to obtain a fault detection report; The fault detection report is transmitted to the operation and maintenance center for strategy parsing and optimization to obtain regional linkage operation and maintenance strategy parameters. Based on the regional linkage operation and maintenance strategy parameters, fault operation and maintenance management is carried out in the regions of the first target base station and the second target base station. The process of obtaining regional linkage operation and maintenance strategy parameters includes: performing feature matching between the fault detection report and the fault operation and maintenance strategy library to obtain the fault operation and maintenance strategy. Based on the fault operation and maintenance strategy, a fault operation and maintenance strategy space is constructed, and the fault detection report is parsed according to the fault operation and maintenance strategy space to obtain the operation and maintenance strategy parameter thresholds. The operation and maintenance strategy parameters are obtained by performing global optimization of the parameters within the preset search range and within the threshold of the operation and maintenance strategy parameters. The operation and maintenance strategy parameters are linked and integrated according to the inspection sequence of the base station to obtain the regional linkage operation and maintenance strategy parameters. The system is pre-trained with the prediction model in the following manner: Obtain historical fault probability time-series data samples and historical environmental variable time-series data samples for each component of the second target base station; After preprocessing the historical failure probability time series data samples and the historical environmental variable time series data samples, the historical failure probability time series data samples and the historical environmental variable time series data samples are transformed into supervised failure probability data and supervised environmental variable data, respectively. Supervised data on failure probability and supervised data on environmental variables are input into the neural network for training. A prediction model is obtained after the preset training cutoff condition is met. The neural network based on high- and low-frequency recurrent neural networks and environmental variables includes a sequentially connected data input module, a data preprocessing module, a high-frequency feature extraction module, a low-frequency feature extraction module, an environmental variable processing module, a feature fusion module, a recurrent neural network module, and an output module. The data input module receives input data from various data sources, including high- and low-frequency signals and environmental variables. The data preprocessing module cleans and standardizes the data. The high-frequency feature extraction module captures short-term and local patterns in the time series and converts the time series to the frequency domain to extract high-frequency components. The low-frequency feature extraction module captures long-term dependencies in the time series and extracts low-frequency trends and seasonal components. The environmental variable processing module includes an embedding layer and a fully connected layer. The embedding layer converts categorical environmental variables into fixed-dimensional vectors, while the fully connected layer performs nonlinear transformations on numerical environmental variables. The feature fusion module concatenates and fuses high- and low-frequency feature vectors with environmental variable feature vectors in terms of dimension. The recurrent neural network module further processes the fused feature sequence to capture complex temporal relationships. The output module transforms the fused information into the desired output results according to the specific task requirements. The system further includes: calculating a first fault-related factor set related to the first target base station based on a historical fault type set, wherein the first fault-related factor set includes a plurality of first key fault-related factors and a plurality of second key fault-related factors; Obtain the first weight parameters corresponding to each first critical fault-related factor from the first training set, wherein each first weight parameter corresponds to one first critical fault-related factor; Obtain the second weight parameters corresponding to each second critical fault-related factor from the second training set, wherein each second weight parameter corresponds to a second critical fault-related factor; Calculate the critical parameters of the first critical failure corresponding to the factors related to the first critical failure; Calculate the critical parameters of the second critical failure corresponding to the factors related to the second critical failure; Calculate the first degree of integration for each critical parameter of the first fault and the critical parameter of the second fault; the first degree of integration is a comprehensive index used to evaluate the overall impact of each critical parameter of the fault. Furthermore, the step of determining the component whose first fusion degree reaches the fusion degree threshold as the first key component of the first target base station specifically includes: Calculate the first degree of fusion for each critical parameter of the first fault and the critical parameter of the second fault; By combining the two first degrees of fusion above, we obtain the second degree of fusion; The component whose second degree of fusion reaches the fusion degree threshold is identified as the first key component of the first target base station; Furthermore, the step of identifying the component whose second fusion degree reaches the fusion degree threshold as the first key component of the first target base station specifically includes: Calculate the first degree of fusion for each critical parameter of the first fault and the critical parameter of the second fault; By combining the two first degrees of fusion above, we obtain the second degree of fusion; The component with a second fusion degree greater than or equal to 0 is identified as the first key component of the first target base station.

Images