Fault detection method and system based on mobile communication network

Abstract

The invention discloses a fault detection method and system based on a mobile communication network, and relates to the technical field of communication. Comprising the following steps: S1, detecting a mobile communication network signal and a base station transmission state; s2, establishing an SDN controller; s3, distributed storage and correctness verification are carried out on the data through the block chain node, and automatic fault detection is carried out based on a smart contract; s4, performing fault intelligent detection based on the improved convolutional neural network, and analyzing a fault occurrence position and reason in combination with the SDN controller and the block chain data; and S5, giving a corresponding alarm. According to the invention, the SDN controller, the block chain technology and the improved convolutional neural network are combined to realize automatic fault intelligent detection, the fault position is quickly positioned, and the troubleshooting cost is reduced; the fault is remotely analyzed and diagnosed based on the remote communication capability; and through real-time monitoring and analysis of the data, potential fault hidden dangers are found in advance, corresponding prevention measures are taken, the user power is fully exerted, the cost is reduced, and the community participation sense is enhanced.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a fault detection method and system based on a mobile communication network. Background Art

[0002] With the rapid development of information technology, mobile communication networks have become an indispensable part of people's lives and work. It not only provides people with convenient voice communication services, but also supports various data services, such as mobile Internet, Internet of Things, and intelligent transportation. The reliability and stability of mobile communication networks are crucial to ensuring people's daily lives and the normal operation of the economy and society.

[0003] In recent years, mobile communication network technology has continued to evolve, from 2G, 3G to 4G, and 5G, with network speeds increasing and service types becoming increasingly diverse. At the same time, with the rise of the Internet of Things, the number of devices connected to mobile communication networks has exploded, placing higher demands on network capacity and performance. In this context, how to ensure the stable operation of mobile communication networks and timely detect and troubleshoot faults has become an important technical challenge.

[0004] Traditional mobile communication network fault detection methods mainly rely on the network equipment's own alarm mechanism and manual inspection. When a fault occurs, the network equipment will send out an alarm signal to notify the operator to handle it. Manual inspection is to find potential fault hazards through regular inspection and maintenance of network equipment by technicians. However, these traditional methods have certain limitations. First, the alarm mechanism of network equipment can often only detect more serious faults, and it is difficult to detect some potential performance problems and minor faults. Second, manual inspection requires a lot of manpower and time, is inefficient, and it is difficult to cover all areas of the network. In addition, traditional methods are also difficult to locate and diagnose faults, and often require technicians to perform complex analysis and testing to determine the cause and location of the fault.

[0005] With the popularization of mobile communication networks and the increasing requirements of users for network quality, users' tolerance for network failures is also decreasing. Users hope to get timely feedback and understand the status of the network, quickly resolve network failures, and ensure their communication and business needs. Therefore, operators need to continuously improve the efficiency and accuracy of fault detection to meet user needs.

[0006] In order to solve the above problems, the present invention proposes a fault detection method and system based on a mobile communication network. Summary of the invention

[0007] The main purpose of the present invention is to provide a fault detection method and system based on a mobile communication network, which can effectively solve the problems in the background technology.

[0008] In order to achieve the above object, the present invention adopts the following technical solutions:

[0009] A fault detection method based on a mobile communication network comprises the following steps:

[0010] S1: Detect and provide real-time feedback on the mobile communication network signal status and base station transmission status, and self-adjust the fault information detected later based on the detection results;

[0011] S2: Build an SDN controller in the mobile communication network to detect network traffic, device status and connection status in real time, and dynamically adjust the mobile communication network configuration;

[0012] S3: Based on the SDN controller and network equipment, the signal data transmitted in the network is collected, and distributed backup storage of the collected signal data is performed through the blockchain node. The correctness of the collected signal data is verified by combining blockchain technology, and automatic fault detection and response is performed based on smart contracts;

[0013] S4: Intelligent fault detection and analysis based on improved convolutional neural network, combined with centralized control of SDN controller and distributed records of blockchain, to analyze the location and cause of fault;

[0014] S5: issuing a fault alarm based on the location and cause of the fault determined in S4.

[0015] Preferably, in S1, the signal strength, transmission rate and network delay status of the mobile communication network are analyzed and fed back, and the connectivity of the base station and the operating status of the equipment are detected and fed back in real time.

[0016] Preferably, the SDN controller in S2 is connected to each network device of the base station through a southbound interface, and is connected to the mobile communication network control center through a northbound interface.

[0017] Preferably, the process of the SDN controller detecting a fault is specifically as follows:

[0018] Collect signal data, pre-process the collected signal data, make fault judgments based on the signal data characteristics obtained by pre-processing, further combine blockchain to locate and analyze the causes of signal data judged to have faults, perform alarm processing, and continuously monitor signals judged not to have faults.

[0019] Preferably, in S3, a smart contract is set according to preset fault detection rules and conditions, the network status is automatically detected, and corresponding response measures are triggered when a fault is detected; a fault-tolerant mechanism is set in the smart contract, and the fault-tolerant mechanism includes a validity check mechanism, a permission management mechanism, a multiple condition judgment mechanism, a backup mechanism, a timeout mechanism and a retry mechanism.

[0020] Preferably, the improved convolutional neural network in S4 performs division and collection of signal data based on the time dimension and the space dimension, performs feature extraction processing on the collected signal data of different dimensions, and performs feature extraction processing on the time dimension matrix X obtained by feature extraction. m and the spatial dimension matrix X n Fusion based on probability matrix;

[0021] The time dimension matrix X m and the spatial dimension matrix X n The probability matrix p m and p n The calculation is as follows:

[0022]

[0023] Among them, σ is the activation function; ω m and ω n is the learning weight; b m and b n is bias;

[0024] Based on the improved butterfly optimization algorithm, the matrices of different dimensions are weighted to obtain the fusion signal data feature X; the details are as follows:

[0025] X=λ((p m ) T +p n )X m +(1-λ)(p m +(p n ) T )X n ;

[0026] Among them, λ is the weight value; T is the transpose of the vector;

[0027] The improved butterfly optimization algorithm is specifically as follows:

[0028] The butterfly scent is calculated based on the detection accuracy, the inverse of the error function is used as the fitness function, and the updated butterfly position is used as the weighted value. The butterfly scent is calculated as follows:

[0029] s=fδ α ;

[0030] Among them, f is the adjustment factor; δ is the detection accuracy; α is the power exponent;

[0031] At the beginning of the iteration, each butterfly is randomly arranged in the search area, and the fitness function of the butterfly at this time is calculated; the size of the fragrance produced by each butterfly and its position at this time are calculated; the global search is as follows:

[0032]

[0033] in, is the position vector of the i-th butterfly obtained at the t-th iteration; is the position vector of the ith butterfly obtained in the t+1th global search iteration; s i is the fragrance produced by the i-th butterfly; r1, r2 and r3 are all random numbers between (0,1); u g is the global optimal position vector; β(t+1) is the Gaussian random walk function; Guss(·) is the Gaussian function; is the best butterfly position vector obtained from the tth global search iteration; μ is the standard deviation;

[0034] The local search is as follows:

[0035]

[0036] in, is the position vector of the ith butterfly obtained by the t+1th local search iteration; r4 is a random number between (0,1); is the position vector of the ath butterfly obtained at the tth iteration; is the position vector of the bth butterfly obtained at the tth iteration;

[0037] The improved convolutional network includes 7 convolutional layers, 2 pooling layers and 2 fully connected layers. A residual structure is added after the third, fourth and seventh layers of the convolutional pooling respectively. The weight of the residual structure is obtained based on the improved butterfly optimization algorithm. The Dropout technology is also used for optimization between the two fully connected layers. Finally, the center loss function and the Softmax loss function are used as objective functions to perform fault detection on the fused signal data features.

[0038] Preferably, the SDN controller determines the fault area based on the network topology and traffic information, obtains historical fault information and related diagnostic clues in combination with the fault records on the blockchain, and determines the type and severity of the fault by analyzing the fault characteristics and related signal data.

[0039] Preferably, the SDN controller is combined with a blockchain node to collect fault information through a crowdsourcing mode; in the crowdsourcing mode, users provide feedback on signal data and fault information through mobile applications, SMS feedback, customer service hotlines or online platforms, and an incentive mechanism is established to reward users with accurate feedback.

[0040] Preferably, the incentive mechanism includes point rewards, honorary titles and lucky draws.

[0041] A fault detection system based on a mobile communication network, comprising:

[0042] Network self-check module: used to detect and provide real-time feedback on the mobile communication network signal status and base station transmission status, and to self-adjust the fault information detected later based on the detection results;

[0043] Control building module: build an SDN controller in the mobile communication network to detect network traffic, device status and connection status in real time, and dynamically adjust the network configuration;

[0044] Intelligent detection module; used to realize intelligent fault detection of signal data;

[0045] Fault analysis module: used to perform intelligent fault detection and analysis based on improved convolutional neural networks, and analyze the location and cause of faults by combining the centralized control of the SDN controller and the distributed records of the blockchain;

[0046] Fault alarm module: used to issue fault alarm based on the fault location and cause determined in the fault analysis module;

[0047] The network self-check module includes a signal detection unit and a base station detection unit; respectively used to detect the mobile communication network signal status and the base station transmission status;

[0048] The intelligent detection module includes a signal data acquisition unit, a signal data backup unit, a signal data verification unit and an intelligent response unit;

[0049] The signal data acquisition unit is used to collect signal data transmitted in the network based on the SDN controller and the network equipment;

[0050] The signal data backup unit is used to perform distributed backup storage of the collected signal data through blockchain nodes;

[0051] The signal data verification unit is used to verify the correctness of the signal data collected by the SDN controller in combination with the blockchain technology;

[0052] The intelligent response unit is used to perform automated fault detection and response based on smart contracts.

[0053] Compared with the prior art, the present invention provides a fault detection method and system based on a mobile communication network, which has the following beneficial effects:

[0054] The present invention can more comprehensively understand the operation of the network by integrating the signal data in the mobile communication network, and realize automatic intelligent fault detection by combining SDN controller, blockchain technology and improved convolutional neural network technology, which can automatically identify the fault mode, quickly locate the fault location, and reduce the manpower and time costs of troubleshooting; and based on the remote communication capability of the mobile communication network, it can realize remote analysis and diagnosis of faults, further improving the efficiency of troubleshooting; and through real-time monitoring and analysis of data, it can discover potential fault hazards in advance, take corresponding preventive measures, and avoid the occurrence of faults. It also improves the comprehensiveness, accuracy and real-time nature of fault detection by giving full play to the power of users, while reducing costs and enhancing community participation. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a flow chart of the method mentioned in Example 1 of the present invention;

[0056] Figure 2 This is a control framework diagram of the SDN controller mentioned in Example 1 of the present invention;

[0057] Figure 3 This is a schematic diagram of the improved convolutional network structure mentioned in Example 1 of the present invention;

[0058] Figure 4 This is a fault detection flow chart of the SDN controller mentioned in Embodiment 1 of the present invention;

[0059] Figure 5 This is a system block diagram mentioned in Example 2 of the present invention.

[0060] In the figure:

[0061] 100. Network self-check module; 110. Signal detection unit; 120. Base station detection unit; 200. Control construction module; 300. Intelligent detection module; 310. Signal data acquisition unit; 320. Signal data backup unit; 330. Signal data verification unit; 340. Intelligent response unit; 400. Fault analysis module; 500. Fault alarm module. DETAILED DESCRIPTION

[0062] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0063] Embodiment 1:

[0064] See also Figure 1-Figure 4 The present invention provides a fault detection method based on a mobile communication network, comprising:

[0065] S1: Detect and provide real-time feedback on the mobile communication network signal status and base station transmission status, and self-adjust the fault information detected later based on the detection results; the details are as follows:

[0066] Use the built-in signal strength detection function of the mobile device to obtain the signal strength index of the currently connected mobile communication network in real time. You can develop a special application to regularly read and record the signal strength value, while recording the location information and timestamp. For example, the application can run continuously in the background, perform a signal strength detection at regular intervals, and upload the signal data to the server for analysis. In this way, a signal strength distribution map of different locations can be established to help operators understand the signal coverage situation.

[0067] Use the network speed test tool on your mobile device to perform transmission rate tests regularly. You can test indicators such as download speed, upload speed, and latency, and upload the test results to the server. The application can automatically test the transmission rate when the user is using a specific application to understand the network performance in different application scenarios. For example, test while watching a video, downloading a file, or playing an online game.

[0068] Applications on mobile devices can use the Ping command or other network latency testing tools to periodically send test requests to the server and measure the response time. These tests can be performed automatically in the background or when the user initiates a specific action.

[0069] Applications can record network delays for different operations, such as opening a web page, sending a message, or playing an online game. These signal data can help operators understand the network's latency performance in different application scenarios.

[0070] The base station device can periodically send connection status reports to the server, including indicators such as the number of connections with mobile devices, connection quality, and signal strength. The server can detect the connectivity of the base station based on these signal data and discover connection problems in a timely manner.

[0071] Operators can deploy detection devices near base stations, which can monitor the signal transmission and reception of base stations and detect the connection quality between base stations and mobile devices. These devices can communicate with servers to provide more detailed connectivity signal data.

[0072] Base station equipment can have a built-in self-detection function to regularly check the operating status of hardware equipment, such as processor temperature, memory usage, power supply, etc. These signal data can be transmitted to the server via the network for remote detection by operators.

[0073] Operators can use remote management tools to monitor and manage base station equipment in real time. These tools can remotely view equipment status information, configuration parameters and log records, and promptly detect equipment failures and abnormal conditions.

[0074] Based on the detected mobile communication network signal status and base station transmission status, corresponding adaptive adjustments are made to the fault information detected later, such as location, time, etc., to ensure the accuracy of the fault information detected later.

[0075] S2: Build an SDN controller in the mobile communication network to detect network traffic, device status and connection status in real time, and dynamically adjust the mobile communication network configuration; the details are as follows:

[0076] The framework of the SDN controller can be found in Figure 2 ,The SDN controller separates the control plane of the mobile communication network from the signal data plane, connects the various devices of the base station through the southbound interface, and connects to the mobile communication network control center through the northbound interface, realizing centralized network control and management. Through the SDN controller, network traffic, device status and connection status can be detected in real time, providing a comprehensive network view for fault detection.

[0077] The SDN controller allows administrators to dynamically adjust network configurations as needed, such as adjusting routing policies, allocating bandwidth resources, etc. During fault detection, the flexibility of SDN can be used to quickly isolate the fault area and reduce the impact of the fault on the entire network.

[0078] SDN controllers can perform in-depth analysis of network traffic to identify abnormal traffic patterns and potential signs of failure. For example, sudden traffic spikes, changes in traffic distribution, or traffic anomalies for specific applications may be precursors to network failures.

[0079] S3: Based on the SDN controller and network equipment, the signal data transmitted in the network is collected, and distributed backup storage of the collected signal data is performed through the blockchain node. The correctness of the collected signal data is verified by combining blockchain technology, and automatic fault detection and response is performed based on smart contracts; the details are as follows:

[0080] By using the detection function of SDN controllers and network devices, various signal data in the mobile communication network are collected in real time, including signal strength, connection status, traffic signal data, equipment performance indicators, etc. from mobile devices, base stations and detection equipment, and stored in the blockchain. Blockchain can use distributed storage technology to ensure the security and reliability of signal data. Signal data can be classified and stored according to dimensions such as time, location, and device type for subsequent analysis and query. At the same time, blockchain nodes can also collect and verify fault information to ensure the integrity and accuracy of signal data. In the mobile communication network, according to the preset fault detection rules and conditions, smart contracts are set to automatically detect the network status and trigger corresponding response measures when a fault is detected; for example, the FaultRecord structure is used in the smart contract to store detailed information of the fault, including timestamp, fault description, location and severity level. The functions of the smart contract can be called through software or interfaces that interact with the blockchain node to realize an automated fault detection and response mechanism. At the same time, the functions of the smart contract can be further expanded and customized according to actual needs to adapt to different fault scenarios and response strategies.

[0081] A validity check mechanism for input parameters is also added to the smart contract. Before recording a new fault, verify whether the fault description, location and other parameters meet reasonable format requirements. For example, check whether the fault description is an empty string and whether the location information contains valid geographic coordinates or address descriptions. If the input parameters are illegal, return an error message instead of blindly recording possibly erroneous fault information.

[0082] A specific permission management mechanism is also set up so that only specific administrator roles or verified maintenance personnel can call this function. This can prevent unauthorized users from misoperating or maliciously tampering with the fault status.

[0083] Set up multiple conditional judgments and backup measures. For example, when an automatic response action is required based on severity level and time, you can first check whether the backup network configuration is available. If the primary network fails and the backup configuration is unavailable, an alarm mechanism can be triggered to notify relevant personnel for emergency processing, rather than blindly trying to switch to a backup network that may also have problems.

[0084] Set up timeout and retry mechanisms to deal with possible blockchain network congestion, node failures, etc. When calling a smart contract function, if there is no response within a certain period of time, it can automatically retry a certain number of times. If multiple retries still fail, the error log can be recorded and the administrator can be notified for manual intervention.

[0085] By adding these fault-tolerant mechanisms, the reliability and robustness of smart contracts in mobile communication network fault detection can be greatly improved, ensuring normal operation and effective fault handling in various complex situations.

[0086] S4: Intelligent fault detection and analysis based on improved convolutional neural network, combined with centralized control of SDN controller and distributed records of blockchain, to analyze the location and cause of the fault; details are as follows:

[0087] The collected signal data is analyzed and processed using an improved convolutional neural network, which can analyze the changing trends and distribution of indicators such as signal strength, transmission rate, network delay, base station connectivity, and equipment operating status.

[0088] First, the signal data is divided and collected according to the time dimension and space dimension, and the signal data of different dimensions are extracted and processed. The time dimension matrix X obtained by feature extraction is m and the spatial dimension matrix X n Fusion based on probability matrix;

[0089] The time dimension matrix X m and the spatial dimension matrix X n The probability matrix p m and p n The calculation is as follows:

[0090]

[0091] Among them, σ is the activation function; ω m and ω n is the learning weight; b m and b n is bias;

[0092] There are certain limitations for signal data in a single time dimension or a single space dimension. The probability fusion weighted approach can more accurately capture the fault characteristics in the signal data and improve the accuracy of fault detection.

[0093] Based on the improved butterfly optimization algorithm, the matrices of different dimensions are weighted to obtain the fusion signal data feature X; the details are as follows:

[0094] X=λ((p m ) T +p n )X m +(1-λ)(p m +(p n ) T )X n ;

[0095] Among them, λ is the weight value; T is the transpose of the vector;

[0096] The improved butterfly optimization algorithm is as follows:

[0097] The butterfly scent is calculated based on the detection accuracy, the inverse of the error function is used as the fitness function, and the updated butterfly position is used as the weighted value. The butterfly scent is calculated as follows:

[0098] s=fδ α ;

[0099] Among them, f is the adjustment factor; δ is the detection accuracy; α is the power exponent;

[0100] At the beginning of the iteration, each butterfly is randomly arranged in the search area, and the fitness function of the butterfly at this time is calculated; the size of the fragrance produced by each butterfly and its position at this time are calculated; the global search is as follows:

[0101]

[0102] in, is the position vector of the i-th butterfly obtained at the t-th iteration; is the position vector of the ith butterfly obtained in the t+1th global search iteration; s i is the fragrance produced by the i-th butterfly; r1, r2 and r3 are all random numbers between (0,1); u g is the global optimal position vector; β(t+1) is the Gaussian random walk function; Guss(·) is the Gaussian function; is the best butterfly position vector obtained from the tth global search iteration; μ is the standard deviation;

[0103] The local search is as follows:

[0104]

[0105] in, is the position vector of the ith butterfly obtained by the t+1th local search iteration; r4 is a random number between (0,1); is the position vector of the ath butterfly obtained at the tth iteration; is the position vector of the bth butterfly obtained at the tth iteration;

[0106] Optimization based on the improved butterfly optimization algorithm with detection accuracy as the goal can further improve the detection accuracy of network faults and provide guarantee for network security operation.

[0107] Reference Figure 3The improved convolutional network includes 7 convolutional layers, 2 pooling layers and 2 fully connected layers. A residual structure is added after the third, fourth and seventh layers of the convolutional pooling respectively. The weight of the weight layer in the residual structure is obtained based on the improved butterfly optimization algorithm. The Dropout technology is also used for optimization between the two fully connected layers. Finally, the center loss function and the Softmax loss function are used as the objective function to perform fault detection on the fused signal data features.

[0108] Reference Figure 4 , detect faults based on SDN controllers: collect signal data, preprocess the collected signal data, make fault judgments based on the signal data characteristics obtained by preprocessing, further combine blockchain to locate and analyze the causes of signal data that are judged to have faults, and quickly locate the location and cause of the fault by combining SDN centralized control and blockchain distributed records. The SDN controller can determine the fault area based on network topology and traffic information, and the fault records on the blockchain can provide historical fault information and related diagnostic clues. By analyzing fault characteristics and related signal data, the type and severity of the fault can be determined, and alarm processing can be performed, and the signal that is judged to have no fault can be continuously monitored.

[0109] It is also possible to analyze and predict potential faults based on the improved convolutional neural network, so that corresponding preventive measures can be taken in time to avoid the occurrence of faults.

[0110] S5: Fault alarm is issued based on the location and cause of the fault determined in S4. The details are as follows:

[0111] Once a fault is identified, the smart contract can automatically trigger the corresponding response measures. This may include notifying the administrator, starting backup equipment, adjusting network configuration, or executing fault repair procedures. At the same time, the fault record on the blockchain can provide a reference for subsequent fault analysis and optimization.

[0112] It is also possible to combine the crowdsourcing model to widely collect usage data and fault information of mobile communication networks. Users can report the problems they encounter in the use of mobile communication networks, such as signal interruption, slow network speed, poor call quality, etc., through special mobile applications, SMS feedback or online platforms. These feedback data from a large number of users can provide more comprehensive and realistic network fault conditions, helping operators to locate and solve problems more quickly. Based on the results of signal data analysis, feedback and suggestions are provided to operators. Operators can take corresponding measures to optimize the network and troubleshoot problems based on these feedbacks. After the fault is repaired, the mobile communication network is continuously monitored to ensure that the fault does not recur. At the same time, the signal data in the process of fault detection and repair is analyzed, lessons learned are summarized, and fault detection methods and network management strategies are continuously improved.

[0113] It can also provide network performance reports and recommendations to mobile device users to help them understand network conditions and optimize device settings. For example, users can choose appropriate applications or adjust the device's network settings based on network latency.

[0114] Embodiment 2:

[0115] See also Figure 5 The present invention provides a fault detection system based on a mobile communication network, comprising:

[0116] The network self-check module 100 detects and provides real-time feedback on the mobile communication network signal status and base station transmission status. Based on the detection results of the network self-check module 100, the fault information detected later, such as location, time, etc., can be self-adjusted to improve the accuracy of fault detection;

[0117] The control building module 200 builds an SDN controller in the mobile communication network. The SDN controller connects to the network devices of the base station through the southbound interface and connects to the mobile communication network control center through the northbound interface. It detects the network traffic, device status and connection status in real time and makes dynamic adjustments to the network configuration.

[0118] The intelligent detection module 300 realizes intelligent fault detection of signal data based on the SDN controller combined with blockchain technology;

[0119] The fault analysis module 400 performs intelligent fault detection and analysis based on the improved convolutional neural network, and analyzes the location and cause of the fault by combining the centralized control of the SDN controller and the distributed records of the blockchain;

[0120] The fault alarm module 500 generates a fault alarm based on the fault location and cause determined in the fault analysis module 400;

[0121] The network self-check module 100 includes a signal detection unit 110 and a base station detection unit 120, which are used to detect the mobile communication network signal status and the base station transmission status respectively;

[0122] The intelligent detection module 300 includes a signal data acquisition unit 310, a signal data backup unit 320, a signal data verification unit 330 and an intelligent response unit 340;

[0123] The signal data collection unit 310 is used to collect signal data transmitted in the network based on the SDN controller and the network equipment;

[0124] The signal data backup unit 320 is used to perform distributed backup storage of the collected signal data through blockchain nodes;

[0125] The signal data verification unit 330 is used to verify the correctness of the signal data collected by the SDN controller in combination with the blockchain technology;

[0126] The intelligent response unit 340 is used to perform automated fault detection and response based on smart contracts.

[0127] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A fault detection method based on a mobile communication network, characterized in that: The steps include: S1: Detect and provide real-time feedback on the mobile communication network signal status and base station transmission status, and self-adjust the fault information detected later based on the detection results; S2: Build an SDN controller in the mobile communication network to detect network traffic, device status and connection status in real time, and dynamically adjust the mobile communication network configuration; S3: Based on the SDN controller and network equipment, the signal data transmitted in the network is collected, and distributed backup storage of the collected signal data is performed through the blockchain node. The correctness of the collected signal data is verified by combining blockchain technology, and automatic fault detection and response is performed based on smart contracts; S4: Intelligent fault detection and analysis based on improved convolutional neural network, combined with centralized control of SDN controller and distributed records of blockchain, to analyze the location and cause of fault; S5: issuing a fault alarm based on the location and cause of the fault determined in S4.

2. A fault detection method based on a mobile communication network according to claim 1, characterized in that: In S1, the signal strength, transmission rate and network delay of the mobile communication network are analyzed and fed back, and the connectivity of the base station and the operating status of the equipment are detected and fed back in real time.

3. A fault detection method based on a mobile communication network according to claim 1, characterized in that: The SDN controller in S2 is connected to each network device of the base station through a southbound interface and is connected to the mobile communication network control center through a northbound interface.

4. A fault detection method based on a mobile communication network according to claim 3, characterized in that: The process of the SDN controller detecting a fault is as follows: Collect signal data, pre-process the collected signal data, make fault judgments based on the signal data characteristics obtained by pre-processing, further combine blockchain to locate and analyze the causes of signal data judged to have faults, perform alarm processing, and continuously monitor signals judged not to have faults.

5. A fault detection method based on a mobile communication network according to claim 1, characterized in that: In the S3, a smart contract is set according to preset fault detection rules and conditions to automatically detect the network status and trigger corresponding response measures when a fault is detected; a fault-tolerant mechanism is set in the smart contract, and the fault-tolerant mechanism includes a validity check mechanism, a permission management mechanism, a multiple condition judgment mechanism, a backup mechanism, a timeout mechanism and a retry mechanism.

6. A fault detection method based on a mobile communication network according to claim 1, characterized in that: The improved convolutional neural network described in S4 divides and collects signal data based on the time dimension and the space dimension, and performs feature extraction processing on the collected signal data of different dimensions, and extracts the time dimension matrix X obtained by feature extraction. m and the spatial dimension matrix X n Fusion based on probability matrix; The time dimension matrix X m and the spatial dimension matrix X n The probability matrix p m and p n The calculation is as follows: Among them, σ is the activation function; ω m and ω n is the learning weight; b m and b n is bias; Based on the improved butterfly optimization algorithm, the matrices of different dimensions are weighted to obtain the fusion signal data feature X; the details are as follows: X=λ((p m ) T +p n )X m +(1-λ)(p m +(p n ) T )X n ; Among them, λ is the weight value; T is the transpose of the vector; The improved butterfly optimization algorithm is specifically as follows: The butterfly scent is calculated based on the detection accuracy, the inverse of the error function is used as the fitness function, and the updated butterfly position is used as the weighted value. The butterfly scent is calculated as follows: s=fδ α ; Among them, f is the adjustment factor; δ is the detection accuracy; α is the power exponent; At the beginning of the iteration, each butterfly is randomly arranged in the search area, and the fitness function of the butterfly at this time is calculated; the size of the fragrance produced by each butterfly and its position at this time are calculated; the global search is as follows: in, is the position vector of the i-th butterfly obtained at the t-th iteration; is the position vector of the ith butterfly obtained in the t+1th global search iteration; s i is the fragrance produced by the i-th butterfly; r1, r2 and r3 are all random numbers between (0,1); u g is the global optimal position vector; β(t+1) is the Gaussian random walk function; Guss(·) is the Gaussian function; is the best butterfly position vector obtained from the tth global search iteration; μ is the standard deviation; The local search is as follows: in, is the position vector of the ith butterfly obtained by the t+1th local search iteration; r4 is a random number between (0,1); is the position vector of the ath butterfly obtained at the tth iteration; is the position vector of the bth butterfly obtained at the tth iteration; The improved convolutional network includes 7 convolutional layers, 2 pooling layers and 2 fully connected layers. A residual structure is added after the third, fourth and seventh layers of the convolutional pooling respectively. The weight of the residual structure is obtained based on the improved butterfly optimization algorithm. The Dropout technology is also used for optimization between the two fully connected layers. Finally, the center loss function and the Softmax loss function are used as objective functions to perform fault detection on the fused signal data features.

7. A fault detection method based on a mobile communication network according to claim 6, characterized in that: The SDN controller determines the fault area based on the network topology and traffic information, obtains historical fault information and related diagnostic clues in combination with the fault records on the blockchain, and determines the type and severity of the fault by analyzing the fault characteristics and related signal data.

8. A fault detection method based on a mobile communication network according to claim 7, characterized in that: The SDN controller is combined with blockchain nodes to collect fault information through a crowdsourcing model; in the crowdsourcing model, users provide feedback on signal data and fault information through mobile applications, SMS feedback, customer service hotlines or online platforms, and an incentive mechanism is established to reward users who provide accurate feedback.

9. A fault detection method based on a mobile communication network according to claim 8, characterized in that: The incentive mechanism includes point rewards, honorary titles and lucky draws.

10. A fault detection system based on a mobile communication network applied to any one of claims 1-9, characterized in that: include: Network self-check module (100): used to detect and provide real-time feedback on the mobile communication network signal status and base station transmission status, and to self-adjust the fault information detected later based on the detection results; Control building module (200): building an SDN controller in the mobile communication network, detecting network traffic, device status and connection status in real time, and dynamically adjusting network configuration; An intelligent detection module (300); used to realize intelligent fault detection of signal data; Fault analysis module (400): used to perform intelligent fault detection and analysis based on an improved convolutional neural network, and analyze the location and cause of the fault by combining the centralized control of the SDN controller and the distributed records of the blockchain; A fault alarm module (500): used for generating a fault alarm based on the fault location and cause determined in the fault analysis module (400); The network self-check module (100) comprises a signal detection unit (110) and a base station detection unit (120); the two units are used to detect the signal status of the mobile communication network and the transmission status of the base station respectively; The intelligent detection module (300) comprises a signal data acquisition unit (310), a signal data backup unit (320), a signal data verification unit (330) and an intelligent response unit (340); The signal data collection unit (310) is used to collect signal data transmitted in the network based on the SDN controller and the network equipment; The signal data backup unit (320) is used to perform distributed backup storage of the collected signal data through blockchain nodes; The signal data verification unit (330) is used to verify the correctness of the signal data collected by the SDN controller in combination with blockchain technology; The intelligent response unit (340) is used to perform automated fault detection and response based on smart contracts.

Images