Remote monitoring method and system for data transmission state
The method enhances data transmission reliability by encoding data with hash codes and time sequences, using multiple routes, and employing a twin support vector machine algorithm for real-time fault diagnosis, addressing inefficiencies in existing technologies.
Patent Information
- Application Number
- CN202411274213.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-12
- Publication Date
- 2024-12-27
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing data transmission methods struggle to maintain accuracy and continuity in complex network environments, leading to data loss, errors, and interruptions, while current solutions like data redundancy, error detection, and transmission path optimization are inefficient or complex.
A method involving data encoding with hash codes and time sequence numbers, multiple transmission routes, real-time link monitoring, and a partial binary tree twin support vector machine algorithm for fault diagnosis, with adaptive parameter adjustment and backup routes, ensures data integrity and reliability.
The method effectively prevents data loss and errors by diagnosing faults promptly and adapting transmission parameters, ensuring reliable and accurate data transmission in complex environments.
Smart Images
Figure CN119211064A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data transmission, and more specifically to a method and system for remotely monitoring data transmission status. Background Art
[0002] In the data transmission, data transmission at the data transmission end is an important link to ensure the stable operation of the data source. However, due to the complex network environment, unstable transmission link and other reasons, data may be lost, wrong or disordered during the transmission process. These problems will not only affect the accuracy of the data, but also may cause data transmission interruption and equipment failure.
[0003] In the prior art, common methods for analyzing data loss during data transmission include data redundancy, error detection and correction, and transmission path optimization, but these methods still have certain limitations when dealing with data transmission problems in complex industrial environments. For example, although data redundancy can improve the reliability of data transmission, it will increase the consumption of transmission bandwidth and storage space; error detection and correction technology requires efficient algorithm support and has limited effect in high packet loss environments; transmission path optimization requires real-time monitoring and adjustment, which increases the complexity of the system. Summary of the invention
[0004] In view of the shortcomings of the above-mentioned technology, the present invention provides a method and system for remote monitoring of data transmission status, and the technical solution is as follows:
[0005] A method for remotely monitoring data transmission status comprises the following steps:
[0006] Step 1: Data encoding is performed on the transmission data before data transmission at the data transmission end. The data encoding uses a hash code to perform redundant encoding on the data packet for detecting transmission data errors, and uses a time sequence code to attach a sequence number to each data packet for detecting loss and disorder of the transmission data;
[0007] Step 2: Establish multiple independent transmission routes for data transmission. When a transmission route fails, data on another route is transmitted as backup data.
[0008] Step 3: During the data transmission process, the state of the transmission link is monitored in real time by a network sensor, wherein the state of the transmission link includes signal strength, packet loss rate and delay;
[0009] Step 4: According to the historical data of the transmission link, a data transmission fault diagnosis algorithm model of the data transmission end is constructed through a partial binary tree twin support vector machine algorithm model, wherein the partial binary tree twin support vector machine algorithm model includes a data receiving module, a data processing module, a feature extraction module, a partial binary tree construction module, a parameter setting module, a twin support vector machine training module, a classification decision module, an evaluation optimization module, and a model output module;
[0010] Step 5: deploy the data transmission fault diagnosis algorithm model of the industrial data transmission end in the transmission link to perform fault diagnosis on the current transmission link, and set a checkpoint in the transmission link to verify the correctness and integrity of the transmission data in real time by checking the hash value and sequence code;
[0011] Step 6: Adaptively adjust transmission parameters and start backup transmission routes based on fault diagnosis results and real-time data analysis results;
[0012] Step 7: Record the data transmission log through the system log, which includes successfully transmitted data packets, lost data packets, transmission link failures and the number of times the backup transmission route is started, and provide a user interface to view the data transmission log, transmission link status and data analysis results.
[0013] As a further embodiment of the present invention, in step 1, the working method of the data encoding is:
[0014] Step (1), arranging and grouping the data to be transmitted into a plurality of data packets according to a predetermined size;
[0015] Step (2), generating an MD5 hash code for each data packet using the WinMD5 tool, and appending the generated MD5 hash code to the end of the data packet;
[0016] Step (3), generating a time serial number by using a Unix timestamp and a random number, and assigning the generated time serial number to each data packet in ascending order, and appending it to the beginning of the data packet;
[0017] Step (4), data transmission is performed using a data packet with a time sequence number and a hash code attached.
[0018] As a further embodiment of the present invention, in step three, the network sensor at least includes an optical fiber sensor, a radio frequency sensor and a vibration sensor.
[0019] As a further embodiment of the present invention, in step 4, the data receiving module is used to obtain historical data of the transmission link from the database, and the historical data includes status data and fault information of the transmission link; the data processing module is used to remove noise and outliers through Kalman filtering, and standardize and normalize the data; the feature extraction module is used to extract key features related to the fault from the processed data through principal component analysis, and divide the extracted feature data set into a training set and a test set; the partial binary tree construction module is used to initialize the partial binary tree structure according to the extracted features, and each node of the partial binary tree contains a twin support vector machine; the parameters The setting module is used to set the initial parameters of the twin support vector machine, which include penalty parameters, kernel function type and parameters, and learning rate; the twin support vector machine training module is used to train the twin support vector machine of each node through a training set; the classification decision module is used to determine the global classification decision rule of data transmission failure at the data transmission end according to the decision function of the twin support vector machine of each node; the evaluation and optimization module is used to evaluate the model performance through cross-validation of the test set, and adjust the model parameters to optimize the model according to the evaluation results; the model output module is used to output the final partial binary tree twin support vector machine algorithm model through an integrated compatible output interface;
[0020] The output end of the data receiving module is connected to the input end of the data processing module, the output end of the data processing module is connected to the input end of the feature extraction module, the output end of the feature extraction module is connected to the input end of the partial binary tree construction module, the output end of the partial binary tree construction module is connected to the input end of the parameter setting module, the output end of the parameter setting module is connected to the input end of the twin support vector machine training module, the output end of the twin support vector machine training module is connected to the input end of the classification decision module, the output end of the classification decision module is connected to the input end of the evaluation and optimization module and the model output module, and the output end of the evaluation and optimization module is connected to the input end of the partial binary tree construction module.
[0021] As a further embodiment of the present invention, the structure of the skewed binary tree includes a root node, an internal node, a leaf node, a skewed structure, a twin support vector machine module and a structure adjustment module; the root node includes a twin support vector machine module for preliminary classification and decision-making of the input feature data; each of the internal nodes includes a twin support vector machine module for transferring data to the child node through a predetermined feature selection mechanism; each of the leaf nodes includes a twin support vector machine module for processing the final data classification decision and outputting the identified data transmission fault type; the left subtree and the right subtree of the skewed structured skewed binary tree have different branch depths and numbers, the left subtree has fewer branches and is used to process joint data, and the right subtree has more branches and is used to process general data; the twin support vector machine module is used to train and learn according to the set parameters and output the decision function of the node; the structure adjustment module is used to add, delete nodes and adjust the parameters of the twin support vector machine module according to the fault detection situation, and optimize the data classification and fault detection performance in real time;
[0022] The output end of the twin support vector machine module of the root node is connected to the input end of the twin support vector machine module of the internal nodes of the left subtree and the right subtree of the skew structure, the output end of the twin support vector machine module of the internal node is connected to the input end of the twin support vector machine module of the leaf node, the output end of the twin support vector machine module of the leaf node is connected to the input end of the structure adjustment module, and the output end of the structure adjustment module is connected to the input end of the root node.
[0023] As a further embodiment of the present invention, the working method of the twin support vector machine module is:
[0024] Step (1), construct two support vector machine models at each node of the partial binary tree, one for positive samples and one for negative samples;
[0025] The model of the constructed positive class support vector machine is:
[0026]
[0027] In formula (1), w + is the weight vector of the positive class support vector machine model, b + is the bias term of the positive class support vector machine model, C is the penalty parameter, n is the number of training samples, x i is the eigenvector of the i-th sample, max(0,·) controls the return value to be non-negative, max(0,1-(w + ·x i +b + ) is the slack variable of the positive class support vector machine model;
[0028] The constraints are:
[0029]
[0030] In formula (2), y i + is the label of the ith positive sample, x is the feature vector of the input sample, x i is the feature vector of the i-th input sample, exp is the exponential function, ||xx i || 2 x i is the Euclidean distance between x and , σ is the bandwidth parameter, is the Gaussian kernel function;
[0031] The model of the constructed negative class support vector machine is:
[0032]
[0033] In formula (3), w - is the weight vector of the negative class support vector machine model, b - is the bias term of the negative class support vector machine model, C is the penalty parameter, n is the number of training samples, x i is the eigenvector of the i-th sample, max(0,·) controls the return value to be non-negative, max(0,1+(w - ·x i +b - ) is the slack variable of the negative class support vector machine model;
[0034] The constraints are:
[0035]
[0036] In formula (2), y i - is the label of the i-th negative sample, x is the feature vector of the input sample, x i is the feature vector of the i-th input sample, exp is the exponential function, ||xx i || 2 x i is the Euclidean distance between x and , σ is the bandwidth parameter, is the Gaussian kernel function;
[0037] Step (2), respectively calculating the decision function values of the positive support vector machine and the negative support vector machine, wherein the positive support vector machine calculates the decision function value:
[0038]
[0039] In formula (5), f +(x) is the decision function value of the positive class support vector machine model for the input sample, w + is the weight vector of the positive class support vector machine model, b + is the bias term of the positive class support vector machine model, x is the feature vector of the input sample, is the Gaussian kernel function;
[0040] The negative class support vector machine calculates the decision function value:
[0041]
[0042] In formula (6), f - (x) is the decision function value of the negative class support vector machine model for the input sample, w - is the weight vector of the negative class support vector machine model, b - is the bias term of the negative class support vector machine model, x is the feature vector of the input sample, is the Gaussian kernel function;
[0043] Step (3), determining the classification result of the data transmission failure at the data transmission end according to the decision function values of the input features in the two support vector machine models;
[0044] Step (4), optimizing the parameters of the positive support vector machine and the negative support vector machine by grid search;
[0045] Step (5), at each node of the partial binary tree, a twin support vector machine is trained based on the local feature data to make local classification decisions, and the local classification results are passed to the next layer of nodes;
[0046] Step (6): Repeat the above steps (1) to (5) at each node until the leaf node of the partial binary tree outputs the final data transmission fault classification result of the data transmission end.
[0047] As a further embodiment of the present invention, in step five, the checkpoint is set at the key node of the transmission link, and the key node includes the data sending end, the data receiving end and the intermediate relay node. The checkpoint periodically verifies the value of the MD5 hash code and the time series code of the transmitted data through the network protocol analysis tool Wireshark, thereby verifying the correctness and integrity of the transmitted data in real time.
[0048] A remote monitoring system for data transmission status, applying the method according to any one of claims 1 to 7, the system comprising:
[0049] A data encoding module is used to encode data before data transmission and generate a data packet with a hash code and a time sequence number;
[0050] Multiple independent transmission routes are used for data transmission, and backup transmission routes are used when problems occur in the transmission routes;
[0051] A network sensor module is used to monitor the status of the transmission link in real time through sensors;
[0052] The data transmission fault diagnosis module uses the partial binary tree twin support vector machine algorithm model to diagnose transmission link faults;
[0053] The data verification module is used to set checkpoints at key nodes of the transmission link and perform data verification periodically to verify the correctness and integrity of the data in real time;
[0054] An adaptive adjustment module, used to adjust transmission parameters and start backup transmission routes according to fault diagnosis results and real-time data analysis results;
[0055] The logging module is used to record data transmission logs and provide a user interface to view data transmission logs, transmission link status, and data analysis results;
[0056] The output end of the data encoding module is connected to the input ends of the multiple independent transmission routes, the output ends of the multiple independent transmission routes are connected to the input ends of the network sensor module and the data verification module, the output end of the network sensor module is connected to the input end of the data transmission fault diagnosis module, the output ends of the data transmission fault diagnosis module and the data verification module are connected to the input end of the adaptive adjustment module, the output end of the adaptive adjustment module is connected to the input ends of the multiple independent transmission routes, and the output ends of the network sensor module, the data transmission fault diagnosis module and the data verification module are connected to the input end of the log recording module.
[0057] Positive beneficial effects:
[0058] Different from conventional technologies, the present invention discloses a remote monitoring method for data transmission status, which encodes data before data transmission, performs redundant encoding and serial number addition through hash codes and time series codes; establishes multiple independent transmission routes, enables backup routes when problems occur in the transmission routes, and promptly enables backup routes when problems occur in the transmission links, so as to ensure the continuity and reliability of data transmission; monitors the transmission link status in real time through network sensors; diagnoses data transmission faults using a partial binary tree twin support vector machine algorithm model, and can accurately identify data transmission faults through a machine learning algorithm; sets checkpoints in the transmission link, verifies the correctness and integrity of data by checking hash values and serial codes, and prevents data loss, errors and disorder; adaptively adjusts transmission parameters and starts backup transmission routes according to fault diagnosis results and real-time data analysis results, so as to ensure efficient and stable data transmission; records data transmission logs, and provides a user interface to view data transmission logs, transmission link status and data analysis results. The present invention realizes the reliability and accuracy of data transmission at the data transmission end. BRIEF DESCRIPTION OF THE DRAWINGS
[0059] Figure 1 A schematic diagram of the working process of a method for remote monitoring of data transmission status according to the present invention;
[0060] Figure 2 A schematic diagram of a data encoding workflow of a method for remotely monitoring data transmission status according to the present invention;
[0061] Figure 3 A schematic diagram of the structure of a partial binary tree twin support vector machine algorithm model of a method for remote monitoring of data transmission status according to the present invention;
[0062] Figure 4 A schematic diagram of a partial binary tree structure of a method for remote monitoring of data transmission status according to the present invention;
[0063] Figure 5 A schematic diagram of the working process of a twin support vector machine module of a method for remote monitoring of data transmission status of the present invention;
[0064] Figure 6 This is a schematic diagram of the overall structure of a remote monitoring system for data transmission status according to the present invention; DETAILED DESCRIPTION
[0065] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.
[0066] like Figure 1 As shown, a method for remotely monitoring data transmission status of the present invention comprises the following steps:
[0067] Step 1: Data encoding is performed on the transmission data before data transmission at the data transmission end. The data encoding uses a hash code to perform redundant encoding on the data packet for detecting transmission data errors, and uses a time sequence code to attach a sequence number to each data packet for detecting loss and disorder of the transmission data;
[0068] Step 2: Establish multiple independent transmission routes for data transmission. When a transmission route fails, data on another route is transmitted as backup data.
[0069] Step 3: During the data transmission process, the state of the transmission link is monitored in real time by a network sensor, wherein the state of the transmission link includes signal strength, packet loss rate and delay;
[0070] Step 4: According to the historical data of the transmission link, a data transmission fault diagnosis algorithm model of the data transmission end is constructed through a partial binary tree twin support vector machine algorithm model, wherein the partial binary tree twin support vector machine algorithm model includes a data receiving module, a data processing module, a feature extraction module, a partial binary tree construction module, a parameter setting module, a twin support vector machine training module, a classification decision module, an evaluation optimization module, and a model output module;
[0071] Step 5: deploy the data transmission fault diagnosis algorithm model of the industrial data transmission end in the transmission link to perform fault diagnosis on the current transmission link, and set a checkpoint in the transmission link to verify the correctness and integrity of the transmission data in real time by checking the hash value and sequence code;
[0072] Step 6: Adaptively adjust transmission parameters and start backup transmission routes based on fault diagnosis results and real-time data analysis results;
[0073] Step 7: Record the data transmission log through the system log, which includes successfully transmitted data packets, lost data packets, transmission link failures and the number of times the backup transmission route is started, and provide a user interface to view the data transmission log, transmission link status and data analysis results.
[0074] In the above embodiment, if Figure 2 As shown, in step 1, the working method of the data encoding is:
[0075] Step (1), arranging and grouping the data to be transmitted into a plurality of data packets according to a predetermined size;
[0076] Step (2), generating an MD5 hash code for each data packet using the WinMD5 tool, and appending the generated MD5 hash code to the end of the data packet;
[0077] Step (3), generating a time serial number by using a Unix timestamp and a random number, and assigning the generated time serial number to each data packet in ascending order, and appending it to the beginning of the data packet;
[0078] Step (4), data transmission is performed using a data packet with a time sequence number and a hash code attached.
[0079] In a specific embodiment, the data packet for which an MD5 hash code needs to be generated is selected through the WinMD5 tool. After the data packet is selected, the WinMD5 tool will automatically calculate and display the MD5 hash value of the data packet, copy the MD5 hash value generated by the WinMD5 tool and add it to the end of the data packet, and ensure that each data packet corresponds to a unique MD5 hash value.
[0080] In a specific embodiment, the time serial number is generated by combining the Unix timestamp and a random number through the script tool Bash and is added to the beginning of the data packet in increasing order.
[0081] In the above embodiment, in step three, the network sensor at least includes an optical fiber sensor, a radio frequency sensor and a vibration sensor.
[0082] In a specific embodiment, the status of the transmission link is collected through various network sensors deployed around the transmission line. Fiber optic sensors are arranged along the transmission link to monitor the line status and environmental changes in real time. Radio frequency sensors can be installed around industrial communication equipment to monitor the strength and quality of wireless signals. Vibration sensors can be fixed on key equipment or structures to collect vibration data in real time.
[0083] In a specific embodiment, the temperature sensor adopts OptaSense ODH4, with a maximum transmission distance of 50 kilometers, a sampling frequency of 1000 Hz, and a sensitivity of 1με; the radio frequency sensor adopts PE-W15M001, with a frequency range of 1 MHz to 1 GHz and a sensitivity of -110 dBm; the vibration sensor adopts WilcoxonResearch 786A, with a range of ±500g, a sensitivity of 10mV / g, and a frequency range of 1 Hz to 5000 Hz.
[0084] In the above embodiment, if Figure 3As shown, in the step 4, the data receiving module is used to obtain the historical data of the transmission link from the database, and the historical data includes the status data and fault information of the transmission link; the data processing module is used to remove noise and outliers through Kalman filtering, and standardize and normalize the data; the feature extraction module is used to extract the key features related to the fault from the processed data through principal component analysis, and divide the extracted feature data set into a training set and a test set; the partial binary tree construction module is used to initialize the partial binary tree structure according to the extracted features, and each node of the partial binary tree contains a twin support vector machine; the parameter setting module is used to The method is used to set the initial parameters of the twin support vector machine, which include penalty parameters, kernel function type and parameters, and learning rate; the twin support vector machine training module is used to train the twin support vector machine of each node through a training set; the classification decision module is used to determine the global classification decision rule of data transmission failure at the data transmission end according to the decision function of the twin support vector machine of each node; the evaluation and optimization module is used to evaluate the model performance through cross-validation of the test set, and adjust the model parameters according to the evaluation results to optimize the model; the model output module is used to output the final partial binary tree twin support vector machine algorithm model through an integrated compatible output interface;
[0085] The output end of the data receiving module is connected to the input end of the data processing module, the output end of the data processing module is connected to the input end of the feature extraction module, the output end of the feature extraction module is connected to the input end of the partial binary tree construction module, the output end of the partial binary tree construction module is connected to the input end of the parameter setting module, the output end of the parameter setting module is connected to the input end of the twin support vector machine training module, the output end of the twin support vector machine training module is connected to the input end of the classification decision module, the output end of the classification decision module is connected to the input end of the evaluation and optimization module and the model output module, and the output end of the evaluation and optimization module is connected to the input end of the partial binary tree construction module.
[0086] In a specific embodiment, the historical data of the transmission link is stored in a MySQL database, and the data receiving module connects to the MySQL database and reads the signal strength, packet loss rate and delay data of the transmission link.
[0087] In a specific embodiment, the working method of the principal component analysis is:
[0088] S1. Construct the signal strength, packet loss rate and delay data of the preprocessed transmission link into a data matrix, where each row represents a data sample and each column represents a transmission fault feature;
[0089] S2. Describe the linear relationship between the feature and the transmission failure by calculating the covariance matrix of the data matrix, and perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues and eigenvectors. The eigenvalues represent the importance of each principal component, and the eigenvectors represent the direction of the principal component.
[0090] S4. Sort the principal components according to the size of the eigenvalues, and select the first few principal components with large eigenvalues, and the principal components contain the most transmission fault information in the data;
[0091] S5. Use the selected principal components to transform the original data to obtain a new data set, and extract key features related to the transmission failure from the transformed data set to obtain a data transmission failure feature data set at the data transmission end.
[0092] In the above embodiment, if Figure 4 As shown, the structure of the skewed binary tree includes a root node, an internal node, a leaf node, a skewed structure, a twin support vector machine module and a structure adjustment module; the root node includes a twin support vector machine module for preliminary classification and decision-making of the input feature data; each of the internal nodes includes a twin support vector machine module for transferring data to the child node through a predetermined feature selection mechanism; each of the leaf nodes includes a twin support vector machine module for processing the final data classification decision and outputting the identified data transmission fault type; the left subtree and the right subtree of the skewed structured skewed binary tree have different branch depths and numbers, the left subtree has fewer branches and is used to process joint data, and the right subtree has more branches and is used to process general data; the twin support vector machine module is used to train and learn according to the set parameters and output the decision function of the node; the structure adjustment module is used to add, delete nodes and adjust the parameters of the twin support vector machine module according to the fault detection situation, and optimize the data classification and fault detection performance in real time;
[0093] The output end of the twin support vector machine module of the root node is connected to the input end of the twin support vector machine module of the internal nodes of the left subtree and the right subtree of the skew structure, the output end of the twin support vector machine module of the internal node is connected to the input end of the twin support vector machine module of the leaf node, the output end of the twin support vector machine module of the leaf node is connected to the input end of the structure adjustment module, and the output end of the structure adjustment module is connected to the input end of the root node.
[0094] In a specific embodiment, Figure 5 As shown, the working method of the twin support vector machine module is:
[0095] Step (1), construct two support vector machine models at each node of the partial binary tree, one for positive samples and one for negative samples;
[0096] The model of the constructed positive class support vector machine is:
[0097]
[0098] In formula (1), w + is the weight vector of the positive class support vector machine model, which is used to determine the direction and position of the classification hyperplane;
[0099] b + It is the bias term of the positive support vector machine model, which is used to adjust the position of the hyperplane so that the hyperplane can better separate positive samples from negative samples;
[0100] C is a penalty parameter, which is used to control the degree of penalty imposed by the model on misclassified samples. A larger C value will make the model more focused on reducing misclassification errors, while a smaller C value will allow more misclassification errors.
[0101] n is the number of training samples, which is used to represent the sum of penalty items for all training samples;
[0102] x i is the feature vector of the i-th sample, which contains the feature values of the sample and is used for model training and classification decisions;
[0103] max(0,·) controls the return value to be non-negative. When the expression in the brackets is less than 0, the max(0,·) function returns 0; when the expression in the brackets is greater than 0, the max(0,·) function returns the value in the brackets.
[0104] max(0,1-(w + ·x i +b + ) is the slack variable of the positive class support vector machine model, representing the sample x i The distance from the classification boundary minus 1, when this value is greater than 0, it means that the sample x i within the classification interval or misclassified;
[0105] The constraints are:
[0106]
[0107] In formula (2), y i + is the label of the i-th positive sample, which is used for category identification in supervised learning;
[0108] x is the feature vector of the input sample, which indicates the sample that needs to be classified or processed at present;
[0109] x i is the feature vector of the i-th input sample, representing the i-th sample in the training set;
[0110] exp is an exponential function, which is used to calculate the value of the Gaussian kernel function;
[0111] ||xx i || 2 x i is the Euclidean distance between x and , which is used to measure the similarity between samples;
[0112] σ is the bandwidth parameter, which is used to control the width of the Gaussian kernel function;
[0113] It is a Gaussian kernel function, which is used to measure the similarity between samples and map the data into a high-dimensional feature space;
[0114] The model of the constructed negative class support vector machine is:
[0115]
[0116] In formula (3), w - is the weight vector of the negative class support vector machine model, which is used to determine the direction and position of the classification hyperplane;
[0117] b - It is the bias term of the negative support vector machine model, which is used to adjust the position of the hyperplane so that the hyperplane can better separate positive samples from negative samples;
[0118] C is a penalty parameter, which is used to control the degree of penalty imposed by the model on misclassified samples. A larger C value will make the model more focused on reducing misclassification errors, while a smaller C value will allow more misclassification errors.
[0119] n is the number of training samples, which is used to represent the sum of penalty items for all training samples;
[0120] x i is the feature vector of the i-th sample, which contains the feature values of the sample and is used for model training and classification decisions;
[0121] max(0,·) controls the return value to be non-negative. When the expression in the brackets is less than 0, the max(0,·) function returns 0; when the expression in the brackets is greater than 0, the max(0,·) function returns the value in the brackets.
[0122] max(0,1+(w - ·x i +b - ) is the slack variable of the negative class support vector machine model, which represents the distance of sample i from the classification boundary plus 1. When this value is greater than 0, it means that sample i is within the classification interval or is misclassified;
[0123] The constraints are:
[0124]
[0125] In formula (2), y i - is the label of the i-th negative sample, used for category identification in supervised learning;
[0126] x is the feature vector of the input sample, which indicates the sample that needs to be classified or processed at present;
[0127] x i is the feature vector of the i-th input sample, representing the i-th sample in the training set;
[0128] exp is an exponential function, which is used to calculate the value of the Gaussian kernel function;
[0129] ||xx i || 2 x i is the Euclidean distance between x and , which is used to measure the similarity between samples;
[0130] σ is the bandwidth parameter, which is used to control the width of the Gaussian kernel function;
[0131] It is a Gaussian kernel function, which is used to measure the similarity between samples and map the data into a high-dimensional feature space;
[0132] Step (2), respectively calculating the decision function values of the positive support vector machine and the negative support vector machine, wherein the positive support vector machine calculates the decision function value:
[0133]
[0134] In formula (5), f + (x) is the decision function value of the positive class support vector machine model for the input sample, which is used to output the classification decision value of the input sample;
[0135] w + It is the weight vector of the positive class support vector machine model, which determines the direction of the classification hyperplane and the weight of each feature;
[0136] b + It is the bias term of the positive class support vector machine model, which is used to adjust the offset of the classification hyperplane so that the classification hyperplane does not pass through the origin;
[0137] x is the feature vector of the input sample, which indicates the sample that needs to be classified and contains all the feature information of the sample;
[0138] It is a Gaussian kernel function, which maps samples to a high-dimensional feature space through the kernel function, making the originally linearly inseparable data become linearly separable in the high-dimensional space;
[0139] The negative class support vector machine calculates the decision function value:
[0140]
[0141] In formula (6), f - (x) is the decision function value of the negative class support vector machine model for the input sample, which is used to output the classification decision value of the input sample;
[0142] w - is the weight vector of the negative class support vector machine model, which determines the direction of the classification hyperplane and the weight of each feature;
[0143] b - It is the bias term of the negative class support vector machine model, which is used to adjust the offset of the classification hyperplane so that the classification hyperplane does not pass through the origin;
[0144] x is the feature vector of the input sample, which indicates the sample that needs to be classified and contains all the feature information of the sample;
[0145] The Gaussian kernel function maps samples to a high-dimensional feature space through the kernel function, making the originally linearly inseparable data become linearly separable in the high-dimensional space;
[0146] Step (3), determining the classification result of the data transmission failure at the data transmission end according to the decision function values of the input features in the two support vector machine models;
[0147] Step (4), optimizing the parameters of the positive support vector machine and the negative support vector machine by grid search;
[0148] Step (5), at each node of the partial binary tree, a twin support vector machine is trained based on the local feature data to make local classification decisions, and the local classification results are passed to the next layer of nodes;
[0149] Step (6): Repeat the above steps (1) to (5) at each node until the leaf node of the partial binary tree outputs the final data transmission fault classification result of the data transmission end.
[0150] In a specific embodiment, the types of the transmission link fault diagnosis results include:
[0151] Serious failure: signal strength is extremely low or no signal, packet loss rate exceeds 50%, delay time exceeds 500 milliseconds, and jitter value is extremely large;
[0152] Moderate failure: signal strength is low, packet loss rate is between 10% and 50%, delay time is between 200 and 500 milliseconds, and jitter value is large and regular;
[0153] Minor failure: signal strength is slightly lower than normal, packet loss rate is less than 10%, delay time is between 100 and 200 milliseconds, and jitter value is slightly higher but within acceptable range;
[0154] Normal: The signal strength is within the normal range, the packet loss rate is close to 0%, the delay time is less than 100 milliseconds, the jitter value is low, and the transmission frequency is stable.
[0155] In a specific embodiment, the data transmission fault diagnosis algorithm model of the data transmission end output by the partial binary tree twin support vector machine algorithm model training is used to perform fault diagnosis on the five transmission links of the data transmission of the data transmission end. The diagnosis results are as follows:
[0156] Table 1 Data transmission link fault diagnosis record table at the data transmission end
[0157] Transmission link number Signal Strength Packet loss rate Delay time Jitter value Fault type 1 normal 0.2％ 58ms normal normal 2 Low 7％ 156ms normal Minor fault 3 normal 0％ 34ms normal normal 4 normal 1.5％ 63ms normal normal 5 Very low 70％ 870ms great Serious failure
[0158] In the above embodiment, the checkpoint is set at the key node of the transmission link, and the key node includes the data sending end, the data receiving end and the intermediate relay node. The checkpoint periodically verifies the value of the MD5 hash code and the time series code of the transmitted data through the network protocol analysis tool Wireshark, and verifies the correctness and integrity of the transmitted data in real time.
[0159] In a specific embodiment, the distributed log storage system Splunk is used to store log data, and the log data is indexed for fast query and analysis, and the original log data is parsed into a structured format through Logstash.
[0160] like Figure 6 As shown, in a specific embodiment, a remote monitoring system for data transmission status includes:
[0161] A data encoding module is used to encode data before data transmission and generate a data packet with a hash code and a time sequence number;
[0162] Multiple independent transmission routes are used for data transmission, and backup transmission routes are used when problems occur in the transmission routes;
[0163] A network sensor module is used to monitor the status of the transmission link in real time through sensors;
[0164] The data transmission fault diagnosis module uses the partial binary tree twin support vector machine algorithm model to diagnose transmission link faults;
[0165] The data verification module is used to set checkpoints at key nodes of the transmission link and perform data verification periodically to verify the correctness and integrity of the data in real time;
[0166] An adaptive adjustment module, used to adjust transmission parameters and start backup transmission routes according to fault diagnosis results and real-time data analysis results;
[0167] The logging module is used to record data transmission logs and provide a user interface to view data transmission logs, transmission link status, and data analysis results;
[0168] The output end of the data encoding module is connected to the input end of the multiple independent transmission routes, the output end of the multiple independent transmission routes is connected to the input end of the network sensor module and the data verification module, the output end of the network sensor module is connected to the input end of the data transmission fault diagnosis module, the output end of the data transmission fault diagnosis module and the data verification module are connected to the input end of the adaptive adjustment module, the output end of the adaptive adjustment module is connected to the input end of the multiple independent transmission routes, and the output end of the network sensor module, the data transmission fault diagnosis module and the data verification module is connected to the input end of the log recording module. In a specific embodiment, the standard liquid flow calibration device can accurately control and provide standard liquid with a known flow rate to ensure stable flow rate, can adjust and record environmental parameters including temperature, humidity, pressure and electromagnetic interference to simulate the actual use environment, has data recording and analysis functions, can record the measurement results of the flow meter and compare them with the standard flow rate, and automatically calculate the measurement error and accuracy.
[0169] In a specific embodiment, the accuracy of data loss during data transmission at the data transmission end is compared between a method and system for remotely monitoring data transmission status of the present invention and a conventional method and system as follows:
[0170] Table 2 Comparison of accuracy of remote monitoring of data transmission status
[0171]
[0172] The above experiments show that the method of this study has positive and beneficial technical effects.
[0173] Although the specific embodiments of the present invention are described above, it should be understood by those skilled in the art that these specific embodiments are only illustrative, and those skilled in the art may omit, replace, and change the details of the above methods and systems in various ways without departing from the principles and essence of the present invention. For example, merging the above method steps so as to perform substantially the same functions in substantially the same manner to achieve substantially the same results is within the scope of the present invention. Therefore, the scope of the present invention is limited only by the appended claims.
Claims
1. A method for remote monitoring of data transmission status, characterized in that: The following steps are involved: Step 1: Data encoding is performed on the transmission data before data transmission at the data transmission end. The data encoding uses a hash code to perform redundant encoding on the data packet for detecting transmission data errors, and uses a time sequence code to attach a sequence number to each data packet for detecting loss and disorder of the transmission data; Step 2: Establish multiple independent transmission routes for data transmission. When a transmission route fails, data on another route is transmitted as backup data. Step 3: During the data transmission process, the state of the transmission link is monitored in real time by a network sensor, wherein the state of the transmission link includes signal strength, packet loss rate and delay; Step 4: According to the historical data of the transmission link, a data transmission fault diagnosis algorithm model of the data transmission end is constructed through a partial binary tree twin support vector machine algorithm model, wherein the partial binary tree twin support vector machine algorithm model includes a data receiving module, a data processing module, a feature extraction module, a partial binary tree construction module, a parameter setting module, a twin support vector machine training module, a classification decision module, an evaluation optimization module, and a model output module; Step 5: deploy the data transmission fault diagnosis algorithm model of the industrial data transmission end in the transmission link to perform fault diagnosis on the current transmission link, and set a checkpoint in the transmission link to verify the correctness and integrity of the transmission data in real time by checking the hash value and sequence code; Step 6: Adaptively adjust transmission parameters and start backup transmission routes based on fault diagnosis results and real-time data analysis results; Step 7: Record the data transmission log through the system log, which includes successfully transmitted data packets, lost data packets, transmission link failures and the number of times the backup transmission route is started, and provide a user interface to view the data transmission log, transmission link status and data analysis results.
2. According to claim 1, a method for remotely monitoring data transmission status, characterized in that: In the step 1, the working method of the data encoding is: Step (1), arranging and grouping the data to be transmitted into a plurality of data packets according to a predetermined size; Step (2), generating an MD5 hash code for each data packet using the WinMD5 tool, and appending the generated MD5 hash code to the end of the data packet; Step (3), generating a time serial number by using a Unix timestamp and a random number, and assigning the generated time serial number to each data packet in ascending order, and appending it to the beginning of the data packet; Step (4), data transmission is performed using a data packet with a time sequence number and a hash code attached.
3. A method for remotely monitoring data transmission status according to claim 2, characterized in that: In the step three, the network sensor includes at least an optical fiber sensor, a radio frequency sensor and a vibration sensor.
4. A method for remotely monitoring data transmission status according to claim 1, characterized in that: In the step 4, the data receiving module is used to obtain historical data of the transmission link from the database, and the historical data includes status data and fault information of the transmission link; the data processing module is used to remove noise and outliers through Kalman filtering, and standardize and normalize the data; The feature extraction module is used to extract key fault-related features from the processed data through principal component analysis, and divide the extracted feature data set into a training set and a test set; the partial binary tree construction module is used to initialize the partial binary tree structure according to the extracted features, and each node of the partial binary tree contains a twin support vector machine; the parameter setting module is used to set the initial parameters of the twin support vector machine, and the initial parameters include penalty parameters, kernel function type and parameters and learning rate; the twin support vector machine training module is used to train the twin support vector machine of each node through the training set; the classification decision module is used to determine the global classification decision rules for data transmission failures at the data transmission end according to the decision function of the twin support vector machine of each node; the evaluation and optimization module is used to evaluate the model performance through cross-validation of the test set, and adjust the model parameters to optimize the model according to the evaluation results; the model output module is used to output the final partial binary tree twin support vector machine algorithm model through an integrated compatible output interface; The output end of the data receiving module is connected to the input end of the data processing module, the output end of the data processing module is connected to the input end of the feature extraction module, the output end of the feature extraction module is connected to the input end of the partial binary tree construction module, the output end of the partial binary tree construction module is connected to the input end of the parameter setting module, the output end of the parameter setting module is connected to the input end of the twin support vector machine training module, the output end of the twin support vector machine training module is connected to the input end of the classification decision module, the output end of the classification decision module is connected to the input end of the evaluation and optimization module and the model output module, and the output end of the evaluation and optimization module is connected to the input end of the partial binary tree construction module.
5. A method for remotely monitoring data transmission status according to claim 4, characterized in that: The structure of the skewed binary tree includes a root node, an internal node, a leaf node, a skewed structure, a twin support vector machine module and a structure adjustment module; the root node includes a twin support vector machine module for preliminary classification and decision-making of input feature data; each of the internal nodes includes a twin support vector machine module for transmitting data to a child node through a predetermined feature selection mechanism; each of the leaf nodes includes a twin support vector machine module for processing the final data classification decision and outputting the identified data transmission fault type; the left subtree and the right subtree of the skewed structure of the skewed binary tree have different branch depths and numbers, the left subtree has fewer branches and is used to process joint data, and the right subtree has more branches and is used to process general data; The twin support vector machine module is used to train and learn according to the set parameters and output the decision function of the node; The structure adjustment module is used to add or delete nodes and adjust the parameters of the twin support vector machine module according to the fault detection situation, so as to optimize the data classification and fault detection performance in real time; The output end of the twin support vector machine module of the root node is connected to the input end of the twin support vector machine module of the internal nodes of the left subtree and the right subtree of the skew structure, the output end of the twin support vector machine module of the internal node is connected to the input end of the twin support vector machine module of the leaf node, the output end of the twin support vector machine module of the leaf node is connected to the input end of the structure adjustment module, and the output end of the structure adjustment module is connected to the input end of the root node.
6. A method for remotely monitoring data transmission status according to claim 5, characterized in that: The working method of the twin support vector machine module is: Step (1), construct two support vector machine models at each node of the partial binary tree, one for positive samples and one for negative samples; The model of the constructed positive class support vector machine is: In formula (1), w + is the weight vector of the positive class support vector machine model, b + is the bias term of the positive class support vector machine model, C is the penalty parameter, n is the number of training samples, x i is the eigenvector of the i-th sample, max(0,·) controls the return value to be non-negative, max(0,1-(w + ·x i +b + ) is the slack variable of the positive class support vector machine model; The constraints are: (2) In formula (2), y i + is the label of the ith positive sample, x is the feature vector of the input sample, x i is the feature vector of the i-th input sample, exp is the exponential function, ||xx i || 2 x i is the Euclidean distance between x and , σ is the bandwidth parameter, is the Gaussian kernel function; The model of the constructed negative class support vector machine is: In formula (3), w - is the weight vector of the negative class support vector machine model, b - is the bias term of the negative class support vector machine model, C is the penalty parameter, n is the number of training samples, x i is the eigenvector of the i-th sample, max(0,·) controls the return value to be non-negative, max(0,1+(w - ·x i +b - ) is the slack variable of the negative class support vector machine model; The constraints are: In formula (2), y i - is the label of the i-th negative sample, x is the feature vector of the input sample, x i is the feature vector of the i-th input sample, exp is the exponential function, ||xx i || 2 x i is the Euclidean distance between x and , σ is the bandwidth parameter, is the Gaussian kernel function; Step (2), respectively calculating the decision function values of the positive support vector machine and the negative support vector machine, wherein the positive support vector machine calculates the decision function value: In formula (5), f + (x) is the decision function value of the positive class support vector machine model for the input sample, w + is the weight vector of the positive class support vector machine model, b + is the bias term of the positive class support vector machine model, x is the feature vector of the input sample, is the Gaussian kernel function; The negative class support vector machine calculates the decision function value: In formula (6), f - (x) is the decision function value of the negative class support vector machine model for the input sample, w - is the weight vector of the negative class support vector machine model, b - is the bias term of the negative class support vector machine model, x is the feature vector of the input sample, is the Gaussian kernel function; Step (3), determining the classification result of the data transmission failure at the data transmission end according to the decision function values of the input features in the two support vector machine models; Step (4), optimizing the parameters of the positive support vector machine and the negative support vector machine by grid search; Step (5), at each node of the partial binary tree, a twin support vector machine is trained based on the local feature data to make local classification decisions, and the local classification results are passed to the next layer of nodes; Step (6): Repeat the above steps (1) to (5) at each node until the leaf node of the partial binary tree outputs the final data transmission fault classification result of the data transmission end.
7. A method for remotely monitoring data transmission status according to claim 1, characterized in that: In step five, the checkpoints are set at key nodes of the transmission link, including the data sending end, the data receiving end and the intermediate relay nodes. The checkpoints periodically verify the value of the MD5 hash code and the time series code of the transmitted data through the network protocol analysis tool Wireshark to verify the correctness and integrity of the transmitted data in real time.
8. A data transmission status remote monitoring system, characterized in that: Applying the method according to any one of claims 1 to 7, the system comprises: A data encoding module is used to encode data before data transmission and generate a data packet with a hash code and a time sequence number; Multiple independent transmission routes are used for data transmission, and backup transmission routes are used when problems occur in the transmission routes; A network sensor module is used to monitor the status of the transmission link in real time through sensors; The data transmission fault diagnosis module uses the partial binary tree twin support vector machine algorithm model to diagnose transmission link faults; The data verification module is used to set checkpoints at key nodes of the transmission link and perform data verification periodically to verify the correctness and integrity of the data in real time; An adaptive adjustment module, used to adjust transmission parameters and start backup transmission routes according to fault diagnosis results and real-time data analysis results; The logging module is used to record data transmission logs and provide a user interface to view data transmission logs, transmission link status, and data analysis results; The output end of the data encoding module is connected to the input ends of the multiple independent transmission routes, the output ends of the multiple independent transmission routes are connected to the input ends of the network sensor module and the data verification module, the output end of the network sensor module is connected to the input end of the data transmission fault diagnosis module, the output ends of the data transmission fault diagnosis module and the data verification module are connected to the input end of the adaptive adjustment module, the output end of the adaptive adjustment module is connected to the input ends of the multiple independent transmission routes, and the output ends of the network sensor module, the data transmission fault diagnosis module and the data verification module are connected to the input end of the log recording module.