Signaling monitoring method, system and device integrated with AI technology and storage medium

Abstract

The invention belongs to the technical field of signaling monitoring, and relates to a signaling monitoring method, system and device integrated with an AI technology and a storage medium, and the method comprises the steps: 1) receiving original signaling data, obtaining an actual statistical value, calculating a theoretical statistical value through employing a signaling theoretical value calculation AI model, and comparing the actual statistical value with the theoretical statistical value, thereby obtaining a signaling monitoring result; judging the acquisition condition of the original signaling data; 2) carrying out preliminary analysis on the original signaling data after the source code self-inspection, detecting the analysis residual amount of the original signaling data after the source code self-inspection by utilizing an analysis self-inspection AI model so as to judge whether the decoding process is complete or not, and if the analysis is not complete, carrying out supplementary analysis; and 3) obtaining the value of the key index based on the analyzed file, calculating a dynamic threshold by using a statistical verification AI model, and determining whether statistics is abnormal based on the value of the key index and the dynamic threshold. The accuracy of signaling data is guaranteed, and a data abnormity alarm service can be provided in time.

Description

Technical Field

[0001] The present invention belongs to the technical field of signaling monitoring, and relates to a signaling monitoring method, system, device and storage medium, in particular to a signaling monitoring method, system, device and storage medium incorporating AI technology. Background Art

[0002] A signaling monitoring system is a system used to monitor and manage signaling messages in a communication network, and it plays an important role in ensuring the stable operation of the communication network, improving network performance and optimizing network resource allocation. The signaling monitoring system is divided into three layers, namely the signaling acquisition layer, the signaling sharing layer and the signaling application layer.

[0003] The signaling acquisition layer, whose main function is to obtain original signaling from the communication network. The acquisition layer supports various hardware device interfaces such as electrical ports and optical ports, and is responsible for copying various types of original signaling and protocol data from the communication network. The acquisition layer performs real-time acquisition of the original signaling data through the signaling interface. Among them, some interfaces adopt a soft acquisition method based on the signaling data mirror interface, and the acquired data needs to be connected to the signaling processing card board of the monitoring system. Other control plane interface data is collected distributively and processed centrally, and after data encapsulation, it is transmitted to the central station for storage and processing.

[0004] The signaling sharing layer is the core part of the signaling monitoring system, including two major modules: protocol processing and data processing. The protocol processing module is used for signaling acquisition, filtering, copying, and decoding data event synthesis, etc.; the data processing module is used for data warehousing, real-time statistical analysis of performance indicators, etc. The sharing layer receives the full amount of signaling data sent by the acquisition layer, and analyzes and processes the data to generate service signaling data, signaling data or statistical data required by the application system.

[0005] The signaling application layer is the end-user display layer of the signaling monitoring system, and is responsible for displaying various service functions such as real-time management, service analysis, user analysis, and network analysis to users. The signaling monitoring applications in the application layer will subscribe to signaling data from the sharing layer and receive the required XDR signaling data from the sharing layer. Other business systems can also obtain the required service signaling data from the sharing center through the subscription notification method.

[0006] The signaling monitoring system needs to process a large amount of original signaling data every day. How to effectively analyze and process the data and provide high-quality data services is the key factor affecting the application effect of the system. The existing signaling monitoring systems have the following three problems:

[0007] 1. During the collection process, when the resource collection in a certain area is insufficient or the network element data volume is adjusted, the processing capacity of the collection device may be insufficient, resulting in a situation where the collected data volume does not match the actual situation. Since there are many signaling data processing processes, the packet loss problem of the underlying collected data is not easily detected. Without a self-check and verification mechanism, these abnormal situations cannot be processed in a timely manner, thus affecting the integrity of the data and the accuracy of the application data.

[0008] 2. During the parsing process, when the original signaling data is not fully parsed and XDR is generated and sent to the statistical server, the statistical results will be inaccurate. This situation of only parsing without detecting the parsing completeness will also affect the accuracy of the application data.

[0009] 3. During the statistical process, when the live network configuration is adjusted or the server resources are insufficient, errors may occur in the pre-statistics due to association problems, resulting in incorrect query results. Without a corresponding self-check and verification mechanism, such abnormal situations cannot be processed in a timely manner, and the accuracy of the statistical data cannot be guaranteed.

[0010] Therefore, in view of the defects existing in the above-mentioned prior art, it is necessary to develop a new signaling monitoring method, system, device and storage medium. Summary of the Invention

[0011] In order to overcome the defects of the prior art, the present invention proposes a signaling monitoring method, system, device and storage medium incorporating AI technology, which incorporates a self-check and verification mechanism based on an AI model in three parts: signaling collection, signaling parsing, and signaling statistics, thereby ensuring the accuracy of the signaling data and providing a data anomaly warning service in a timely manner.

[0012] In order to achieve the above object, the present invention provides the following technical solutions:

[0013] A signaling monitoring method incorporating AI technology, characterized by including the following steps:

[0014] 1) Source code self-check based on an AI model: Receive the original signaling data from the collection layer, perform statistics on the original signaling data to obtain the actual statistical value, and calculate the theoretical statistical value using the signaling theoretical value through the AI model. By comparing the actual statistical value with the theoretical statistical value, judge the collection situation of the original signaling data;

[0015] 2) Decoding detection based on an AI model: Receive the original signaling data after source code self-check, perform preliminary parsing on the original signaling data after source code self-check, and use the parsing self-check AI model to detect the remaining amount of parsing of the original signaling data after source code self-check to judge whether the decoding process is complete. If it is not fully parsed, perform supplementary parsing until all the original signaling data is completely parsed;

[0016] 3) Statistical verification based on the AI model: Receive the parsed file, obtain the values of the key metrics based on the parsed file, and use the statistical verification AI model to calculate the dynamic threshold of the key metrics of the parsed file. Determine whether the statistics are abnormal based on the values of the key metrics and the dynamic threshold.

[0017] Preferably, the signaling theoretical value calculation AI model in step 1) is a trained artificial neural network model, whose input is signaling data and output is the theoretical statistical value of the signaling data.

[0018] Preferably, step 1) specifically includes:

[0019] 11) When an original signaling data is actually received on the signaling acquisition device, the value of the signaling count counter is incremented by 1, and the actual statistical value of the original signaling data is obtained through such counting;

[0020] 12) Input the original signaling data into the signaling theoretical value calculation AI model, and output the theoretical statistical value of the original signaling data;

[0021] 13) Compare the actual statistical value and the theoretical statistical value in units of cells to check whether the actual statistical value and the theoretical statistical value of the original signaling data in a cell are consistent;

[0022] 14) If the actual statistical value and the theoretical statistical value of the original signaling data in each cell are consistent, it means that the original signaling data has been completely collected. If the actual statistical value and the theoretical statistical value of the original signaling data in a certain cell are inconsistent, output the inconsistent cell, and issue a separate alarm for this cell and perform supplementary collection.

[0023] Preferably, the parsing self-check AI model in step 2) is a trained artificial neural network model, whose input is signaling data and output is the XDR field of the signaling data.

[0024] Preferably, step 2) specifically includes:

[0025] 21) Read the original signaling data after source code self-check received within a unit time and perform parsing to obtain the parsed XDR field;

[0026] 22) Input the original signaling data after source code self-check received within a unit time into the parsing self-check AI model, and output the XDR field of the original signaling data;

[0027] 23) Compare whether there is a deviation between the parsed XDR field in step 21) and the XDR field output in step 22). If there is a deviation, generate an abnormal alarm and check and supplement the parsing of the original signaling data after source code self-check received within a unit time.

[0028] Preferably, the statistical verification AI model in step 3) is a trained artificial neural network model, whose input is the XDR file and output is the dynamic threshold of the key indicators.

[0029] Preferably, step 3) specifically includes:

[0030] 31) Store the parsed XDR file in the database, read the key indicators of the XDR file through the interface of the database, and calculate the ratio of the value of the key indicator to the value of the key indicator in the previous cycle;

[0031] 32) Input the parsed XDR file into the statistical verification AI model to output the dynamic threshold of the key indicators;

[0032] 33) Determine whether the ratio is within the dynamic threshold. If it is not within the dynamic threshold, output a statistical anomaly warning. If it is within the dynamic threshold, generate a report based on the value of the key indicator and store the statistical data in the database.

[0033] In addition, the present invention also provides a signaling monitoring system incorporating AI technology, which is characterized by including:

[0034] A source code self-checking module based on an AI model, which is used to receive the original signaling data from the acquisition layer, perform statistics on the original signaling data to obtain the actual statistical value, calculate the theoretical statistical value by using the signaling theoretical value with the AI model, and judge the acquisition situation of the original signaling data by comparing the actual statistical value with the theoretical statistical value;

[0035] A decoding detection module based on an AI model, which is used to receive the original signaling data after source code self-checking, perform preliminary parsing on the original signaling data after source code self-checking, and use the parsing self-checking AI model to detect the remaining amount of parsing of the original signaling data after source code self-checking to judge whether the decoding process is complete. If it is not completely parsed, supplementary parsing is performed until all the original signaling data is completely parsed;

[0036] A statistical verification module based on an AI model, which is used to receive the parsed file, obtain the value of the key indicator based on the parsed file, calculate the dynamic threshold of the key indicator of the parsed file by using the statistical verification AI model, and determine whether the statistics are abnormal based on the value of the key indicator and the dynamic threshold.

[0037] Moreover, the present invention also provides a signaling monitoring device incorporating AI technology, which is characterized by including:

[0038] One or more processors;

[0039] A memory for storing one or more programs;

[0040] When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the signaling monitoring method incorporating AI technology as described above.

[0041] Finally, the present invention also provides a computer-readable storage medium, on which a computer program is stored, characterized in that when the program is executed by a processor, the steps of the signaling monitoring method incorporating AI technology as described above are implemented.

[0042] Compared with the prior art, the signaling monitoring method, system, device and storage medium incorporating AI technology of the present invention have one or more of the following beneficial technical effects:

[0043] 1. In the signaling monitoring method incorporating AI technology proposed by the present invention, source code self-checking based on AI, decoding detection based on AI and statistical verification based on AI are added, jointly constituting a powerful data processing and verification system, which cooperate with each other to jointly ensure the integrity and accuracy of the original signaling data.

[0044] 2. The present invention can timely detect and correct signaling data errors and anomalies, providing a strong guarantee for the stable operation of the signaling monitoring system.

[0045] 3. The present invention provides reliable data support for subsequent communication signaling analysis and optimization, laying a solid foundation for improving communication quality and user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a flowchart of the existing signaling monitoring method.

[0047] Figure 2 is a flowchart of the signaling monitoring method incorporating AI technology of the present invention.

[0048] Figure 3 is a flowchart of the source code self-checking based on the AI model of the present invention.

[0049] Figure 4 is a flowchart of the decoding detection based on the AI model of the present invention.

[0050] Figure 5 is a flowchart of the statistical verification based on the AI model of the present invention.

[0051] Figure 6 is a schematic diagram of the composition of the signaling monitoring system incorporating AI technology of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0052] Before describing any embodiments of the present invention in detail, it should be understood that the present invention is not limited in its application to the details of the construction and arrangement of components set forth in the following description or illustrated in the following drawings. The present invention is capable of other embodiments and of being practiced or carried out in various ways. Additionally, it should be understood that the language and terminology used herein are for the purpose of description and should not be regarded as limiting. As used herein, the terms "comprising", "having", and their variants are intended to cover the listed items and their equivalents as well as additional items. Unless otherwise specified or limited, the terms "mounted", "connected", "supported", and "coupled" and their variants are used broadly and cover both direct and indirect mounting, connection, support, and coupling. Further, "connected" and "coupled" are not limited to physical or mechanical connection or coupling.

[0053] And, in the first aspect, in the disclosure of the present invention, the orientation or positional relationship indicated by terms such as "longitudinal", "transverse", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. This is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, the above terms should not be construed as limiting the present invention; in the second aspect, the term "a" should be understood as "at least one" or "one or more". That is, in one embodiment, the number of an element can be one, while in other embodiments, the number of this element can be multiple. The term "a" should not be construed as limiting the quantity.

[0054] To overcome the deficiencies of the prior art, the present invention provides a signaling monitoring method, system, device, and storage medium incorporating AI technology. It incorporates a self-check and verification mechanism based on an AI model in three aspects: signaling acquisition, signaling parsing, and signaling statistics, thereby ensuring the accuracy of signaling data and providing timely data anomaly warning services.

[0055] Figure 2 The flowchart of the signaling monitoring method incorporating AI technology of the present invention is shown. As Figure 2 shown, the signaling monitoring method incorporating AI technology of the present invention includes the following steps:

[0056] I. Source code self-check based on an AI model.

[0057] And Figure 1The signaling collection in the prior art shown is different. In the present invention, an AI model is incorporated into the signaling collection process to form a source code self-check based on the AI model. That is, the original signaling data from the collection layer is received, the original signaling data is statistically analyzed to obtain the actual statistical value, and the theoretical statistical value is calculated by the AI model using the signaling theoretical value. By comparing the actual statistical value with the theoretical statistical value, the collection situation of the original signaling data is judged.

[0058] Among them, the signaling theoretical value calculation AI model is a trained artificial neural network model (ANN model), whose input is signaling data and output is the theoretical statistical value of the signaling data.

[0059] Before use, the existing artificial neural network model needs to be trained to form the signaling theoretical value calculation AI model, and its training process specifically includes:

[0060] 1. Collect a large number of open-source existing signaling data.

[0061] 2. Preprocess the signaling data.

[0062] Preprocess the large number of open-source existing signaling data collected, including removing invalid data, filling in missing values, standardizing the data format, etc. These are all prior arts and will not be described in detail for the sake of simplicity.

[0063] 3. Label the preprocessed signaling data.

[0064] Manually label the preprocessed signaling data to mark which signaling data is complete, which signaling data is missing or incorrect, and respectively mark the theoretical statistical value corresponding to the complete signaling data and the theoretical statistical value corresponding to the missing or incorrect signaling data, thus forming a training data set.

[0065] 4. Use the training data set to train the ANN model.

[0066] Use the labeled complete signaling data and missing or incorrect signaling data as input, and their corresponding theoretical statistical values as output to train the ANN model. During the training process, the ANN model will learn the characteristics and patterns of the complete signaling data, as well as the characteristics of the missing or incorrect signaling data.

[0067] After training the signaling theoretical value calculation AI model, the source code self-check based on the AI model can be carried out, as Figure 3 shown, and it specifically includes:

[0068] 1. Statistically analyze the actual statistical value of the original signaling data.

[0069] Similar to the prior art, when an original signaling data is actually received by the signaling acquisition device, the value of the signaling count counter is incremented by 1, and the actual statistical value of the original signaling data is obtained through such counting.

[0070] 2. Input the original signaling data into the AI model for calculating the theoretical value of the signaling, and the AI model for calculating the theoretical value of the signaling outputs the theoretical statistical value of the original signaling data.

[0071] 3. Data comparison.

[0072] Since the collected original signaling data is massive, when performing data comparison, the present invention automatically compares in units of cells. The actual statistical value of the original signaling data of all call records occurring in a specific cell is counted and compared with the theoretical statistical value of the original signaling data output by the AI model for calculating the theoretical value of the signaling. Through such data comparison processing, it can be verified whether the actual statistical value and the theoretical statistical value of the original signaling data are consistent.

[0073] 4. If the actual statistical value and the theoretical statistical value of the original signaling data in each cell are consistent, it indicates that the original signaling data has been completely collected and the collection process is accurate; if the two are inconsistent, it means that there is data loss or collection error, then the inconsistent cell is output, and the cell is separately alarmed and re-collected.

[0074] Through the source code self-check based on the AI model, the present invention can check whether the original signaling data has been completely collected, thereby ensuring the reliability of subsequent processing.

[0075] II. Decoding detection based on the AI model.

[0076] Different from the signaling parsing in the prior art shown in Figure 1 In the present invention, an AI model is incorporated into the signaling parsing process to form a decoding detection based on the AI model, that is, the original signaling data after source code self-check is received, the original signaling data after source code self-check is preliminarily parsed, and the parsing self-check AI model is used to detect the remaining amount of parsing of the original signaling data after source code self-check to determine whether the decoding process is complete. If the parsing is not complete, supplementary parsing is performed until all the original signaling data is completely parsed.

[0077] Among them, the parsing self-check AI model is a trained artificial neural network model (ANN model), whose input is signaling data and the output is the XDR field of the signaling data.

[0078] Before use, the existing artificial neural network model needs to be trained to form the parsing self-check AI model, and its training process specifically includes:

[0079] 1. Collect a large amount of existing open-source signaling data.

[0080] 2. Parse the large amount of existing open-source signaling data collected and obtain the logs, metrics, XDR fields, and other data of the parsed files.

[0081] In the present invention, the logs, metrics, XDR fields, and other data of the parsed files obtained mainly include:

[0082] 1). The number of single-interface fields: Determine the range of the single-interface field data trend.

[0083] 2). The correlation of different interface fields: Which fields appear in different data interfaces, such as the "TAC" and "ECI" fields in both the S1-MME and S11 interfaces.

[0084] 3). Special fields carried in abnormal processes, such as the cause field, which is only filled in when the process fails.

[0085] 3. Clean and format the obtained data to ensure the accuracy and consistency of the data.

[0086] 4. Split the cleaned and formatted data into words and symbols and convert them into numbers or vectors representing their meanings so that the ANN model can process them.

[0087] 5. Use the large amount of existing open-source signaling data collected as input and the corresponding converted numbers or vectors as output to train the ANN model.

[0088] After training the parsing self-checking AI model, decoding detection based on the AI model can be performed, as Figure 4 shown, which specifically includes:

[0089] 1. Similar to the prior art, first read the original signaling data after source code self-checking received per unit time and parse it to obtain the parsed XDR fields.

[0090] 2. Different from the prior art, input the original signaling data after source code self-checking received per unit time into the parsing self-checking AI model, and the parsing self-checking AI model outputs the XDR fields of the original signaling data.

[0091] 3. Compare whether there is a deviation between the parsed XDR fields in step 1 and the XDR fields output in step 2. If there is a deviation, it means that the original signaling data has not been completely parsed. At this time, an abnormal alarm needs to be generated and the original signaling data after source code self-checking received per unit time needs to be checked and supplemented with parsing until the parsing is complete.

[0092] 4. Finally, based on the fully parsed file, synthesize the XDR file to obtain the XDR file.

[0093] Through the decoding detection based on the AI model, the present invention can check whether the decoding process is sufficiently parsed to ensure that the decoded data can truly reflect the content of the original signaling data, so as to ensure the integrity and accuracy of the decoded data and provide a reliable data basis for subsequent processing.

[0094] III. Statistical verification based on the AI model.

[0095] Unlike Figure 1 the signaling statistics in the prior art shown, in the present invention, an AI model is incorporated into the signaling statistics process to form a statistical verification based on the AI model, that is, receive the parsed file, obtain the values of the key indicators based on the parsed file, and use the statistical verification AI model to calculate the dynamic threshold of the key indicators of the parsed file, and determine whether the statistics are abnormal based on the values of the key indicators and the dynamic threshold.

[0096] Among them, the statistical verification AI model is a trained artificial neural network model (ANN model), whose input is the XDR file and the output is the dynamic threshold of the key indicators.

[0097] Before use, it is necessary to train the existing artificial neural network model to form the statistical verification AI model, and its training process specifically includes:

[0098] 1. Collect a large number of open-source existing XDR files.

[0099] 2. Filter the large number of open-source existing XDR files collected, and select the XDR files with complete filled values for storage.

[0100] 3. Manually annotate the stored XDR files, annotate the complete XDR files and their dynamic thresholds of key indicators and the missing XDR files and their dynamic thresholds of key indicators to form a training data set.

[0101] 4. Train the ANN model with the training data set. Use the complete XDR files and the missing XDR files as inputs, and their corresponding dynamic thresholds of key indicators as outputs to train the ANN model. During the training process, the ANN model will learn the reasonable threshold situation that the complete XDR file should have.

[0102] After training the statistical verification AI model, the statistical verification based on the AI model can be carried out, as Figure 5 shown, which specifically includes:

[0103] 1. Store the parsed XDR file in the database. Through the interface of the database, read the key metrics of the XDR file; and calculate the ratio of the value of the key metric to the value of the key metric in the previous period.

[0104] 2. Input the parsed XDR file into the statistical verification AI model, and the statistical verification AI model outputs the dynamic threshold of the key metric.

[0105] 3. Determine whether the ratio is within the dynamic threshold. If it is not within the dynamic threshold, it indicates that the statistical data is incorrect, and a statistical anomaly alarm is output. If it is within the dynamic threshold, it indicates that the statistics are correct. Generate a report based on the value of the key metric obtained in step 1 and store the statistical data in the database.

[0106] 4. Then, the statistical data can be called from the database for the presentation of the statistical data.

[0107] In the present invention, through the statistical verification based on the AI model, it is possible to check whether the preliminary statistical result is correct.

[0108] Figure 6 The schematic diagram of the composition of the signaling monitoring system integrating AI technology of the present invention is shown. As Figure 6 shown, the signaling monitoring system integrating AI technology of the present invention includes:

[0109] 1. The source code self-checking module based on the AI model.

[0110] The source code self-checking module based on the AI model is used to receive the original signaling data from the acquisition layer, perform statistics on the original signaling data to obtain the actual statistical value, and calculate the theoretical statistical value by using the signaling theoretical value through the AI model. By comparing the actual statistical value with the theoretical statistical value, the acquisition situation of the original signaling data is judged.

[0111] 2. The decoding detection module based on the AI model.

[0112] The decoding detection module based on the AI model is used to receive the original signaling data after source code self-checking, perform preliminary parsing on the original signaling data after source code self-checking, and use the parsing self-checking AI model to detect the remaining amount of parsing of the original signaling data after source code self-checking to judge whether the decoding process is complete. If it is not completely parsed, supplementary parsing is performed until all the original signaling data is completely parsed.

[0113] 3. The statistical verification module based on the AI model.

[0114] The statistical verification module based on the AI model is used to receive the parsed file, obtain the values of the key metrics based on the parsed file, and calculate the dynamic threshold of the key metrics of the parsed file using the statistical verification AI model, and determine whether the statistics are abnormal based on the values of the key metrics and the dynamic threshold.

[0115] In addition, the present invention also provides a signaling monitoring device incorporating AI technology, which includes: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the signaling monitoring method incorporating AI technology as described above.

[0116] Finally, the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the signaling monitoring method incorporating AI technology as described above are implemented.

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting the protection scope of the present invention. Those skilled in the art can modify or equivalently replace the technical solutions of the present invention according to the idea of the present invention, without departing from the essence and scope of the technical solutions of the present invention.

Claims

1. A signaling monitoring method incorporating AI technology, characterized in that: The following steps are involved: 1) Source code self-check based on AI model: Receive the original signaling data from the collection layer, perform statistics on the original signaling data to obtain the actual statistical value, and use the signaling theoretical value calculation AI model to calculate the theoretical statistical value. By comparing the actual statistical value with the theoretical statistical value, the collection status of the original signaling data can be judged; 2) Decoding detection based on AI model: Receive the original signaling data after source code self-check, perform preliminary analysis on the original signaling data after source code self-check, and use the analysis self-check AI model to detect the analysis residual amount of the original signaling data after source code self-check to determine whether the decoding process is complete. If it is not completely analyzed, perform supplementary analysis until all the original signaling data are completely analyzed; 3) Statistical verification based on AI model: Receive the parsed file, obtain the value of key indicators based on the parsed file and use the statistical verification AI model to calculate the dynamic threshold values ​​of the key indicators of the parsed file, and determine whether the statistics are abnormal based on the value of the key indicators and the dynamic threshold values.

2. The signaling monitoring method integrating AI technology according to claim 1, characterized in that: The signaling theoretical value calculation AI model in step 1) is a trained artificial neural network model, whose input is signaling data and whose output is the theoretical statistical value of the signaling data.

3. The signaling monitoring method integrating AI technology according to claim 2, characterized in that: The step 1) specifically includes: 11) When a piece of original signaling data is actually received on the signaling collection device, the value of the signaling number counter is increased by 1, and the actual statistical value of the original signaling data is obtained through such counting; 12) Inputting the original signaling data into the signaling theoretical value calculation AI model, and outputting the theoretical statistical value of the original signaling data; 13) Compare the actual statistical value with the theoretical statistical value in cells to check whether the actual statistical value of the original signaling data in a cell is consistent with the theoretical statistical value; 14) If the actual statistical value and theoretical statistical value of the original signaling data in each cell are consistent, it means that the original signaling data has been fully collected. If the actual statistical value and theoretical statistical value of the original signaling data in a cell are inconsistent, the inconsistent cell is output, and an alarm and supplementary sampling are performed for the cell separately.

4. The signaling monitoring method integrating AI technology according to claim 1, characterized in that: The analytical self-checking AI model in step 2) is a trained artificial neural network model, whose input is signaling data and output is the XDR field of the signaling data.

5. The signaling monitoring method incorporating AI technology according to claim 4 is characterized in that: The step 2) specifically includes: 21) Read the original signaling data after source code self-check received within a unit time and parse it to obtain the parsed XDR field; 22) Input the original signaling data after source code self-check received within a unit time into the parsing self-check AI model, and output the XDR field of the original signaling data; 23) Compare the XDR field parsed in step 21) with the XDR field output in step 22) to see if there is a deviation. If there is a deviation, an abnormal alarm is generated and the original signaling data received after source code self-check within a unit time is checked and supplemented with analysis.

6. The signaling monitoring method integrating AI technology according to claim 1, characterized in that: The statistical verification AI model in step 3) is a trained artificial neural network model, whose input is an XDR file and whose output is a dynamic threshold value of key indicators.

7. The signaling monitoring method incorporating AI technology according to claim 6, characterized in that: The step 3) specifically includes: 31) The parsed XDR file is stored in the database, and the key indicators of the XDR file are read through the database interface, and the ratio of the value of the key indicator to the value of the key indicator of the previous period is calculated; 32) Input the parsed XDR file into the statistical verification AI model and output the dynamic threshold of key indicators; 33) Determine whether the ratio is within the dynamic threshold value. If it is not within the dynamic threshold value, output a statistical abnormality alarm. If it is within the dynamic threshold value, generate a report based on the value of the key indicator and store the statistical data.

8. A signaling monitoring system incorporating AI technology, characterized in that: include: The source code self-check module based on the AI ​​model is used to receive the original signaling data from the collection layer, perform statistics on the original signaling data to obtain the actual statistical value, and use the signaling theoretical value calculation AI model to calculate the theoretical statistical value. By comparing the actual statistical value with the theoretical statistical value, the collection status of the original signaling data can be judged; The decoding detection module based on the AI ​​model is used to receive the original signaling data after the source code self-check, perform preliminary analysis on the original signaling data after the source code self-check, and use the analysis self-check AI model to detect the analysis residual amount of the original signaling data after the source code self-check to determine whether the decoding process is complete. If it is not completely analyzed, supplementary analysis is performed until all the original signaling data are completely analyzed; A statistical verification module based on an AI model is used to receive a parsed file, obtain the value of a key indicator based on the parsed file, and calculate the dynamic threshold values ​​of the key indicators of the parsed file using a statistical verification AI model, and determine whether the statistics are abnormal based on the value of the key indicator and the dynamic threshold values.

9. A signaling monitoring device incorporating AI technology, characterized in that: include: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the signaling monitoring method integrating AI technology as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the signaling monitoring method incorporating AI technology as described in any one of claims 1 to 7 are implemented.

Images