Data processing method and system in rail transportation and vehicle operation system

Abstract

The embodiments of the present application provide a data processing method and system for rail transportation transportation and vehicle operation systems. The data processing method includes the following steps: extracting feature information from the data generated by the rail transportation transportation system; determining a feature sequence, wherein the feature sequence contains feature information extracted from the data generated at different times and arranged in chronological order; and performing fault diagnosis on the rail transportation transportation system based on the feature sequence. The technical solution provided by the embodiments of the present application improves the degree of automation of system fault diagnosis, and the diagnostic results can guide targeted maintenance, improve the efficiency of fault diagnosis, reduce delays caused by rail transportation parking, and so on.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a data processing method, system, device, computer storage medium and program product in a rail transportation vehicle operation system. Background Art

[0002] SACEM (Automatic Train Control and Automatic Train Protection) is a critical signaling system for the subway and must ensure high reliability. Therefore, SACEM-based equipment maintenance is one of the key guarantees for stable and reliable operation of the subway.

[0003] Common equipment maintenance methods include preventive maintenance and post-event maintenance. Preventive maintenance involves regular inspections and equipment replacement before an incident occurs, which is labor-intensive. Frequent equipment inspections and replacements lead to high labor costs and significant waste of spare components. Post-event maintenance involves quickly restoring subway operations after an incident, detecting the cause of the fault, and repairing and replacing components. Currently, most maintenance methods are manual, resulting in low efficiency. Summary of the Invention

[0004] In view of the problems existing in the prior art, the embodiments of the present application provide a data processing method, system, device, computer storage medium and program product in a rail transportation transportation and vehicle operation system.

[0005] In one embodiment of the present application, a data processing method in a rail transportation system is provided. The method includes:

[0006] Extracting feature information from data generated by the rail transportation system;

[0007] Determining a feature sequence, wherein the feature sequence includes feature information extracted from the data generated at different times and is arranged in chronological order;

[0008] Based on the characteristic sequence, fault diagnosis is performed on the rail transportation system.

[0009] In another embodiment of the present application, a data processing method in a vehicle operation system is provided. The vehicle operation system includes at least one vehicle and at least one working device for assisting the operation of the at least one vehicle. The data processing method in the vehicle operation system includes:

[0010] Acquire data generated by at least one working device in the vehicle operation system;

[0011] Extracting feature information from the data;

[0012] Determining a feature sequence, wherein the feature sequence includes feature information extracted from the data generated at different times and is arranged in chronological order;

[0013] Based on the feature sequence, fault diagnosis is performed on the vehicle operation system.

[0014] In another embodiment of the present application, a data processing method is provided. The method includes:

[0015] Extract feature information from system log information;

[0016] Determining a feature sequence, wherein the feature sequence includes feature information extracted from the log information generated at different times and is arranged in chronological order;

[0017] determining an event feature of at least one event based on the feature sequence;

[0018] A fault diagnosis is performed on the system according to an event feature of the at least one event.

[0019] In another embodiment of the present application, a data processing method is provided. The method includes:

[0020] Obtain historical data and historical operation status information generated by the rail transportation system during historical periods;

[0021] extracting feature information from the historical data;

[0022] Determining an association relationship between the characteristic information of the historical data and the historical operating status information;

[0023] Determining training samples based on the characteristic information of the historical data, the historical operating status information, and the correlation between the two;

[0024] Using the training samples, the model to be trained is trained to obtain a fault diagnosis model;

[0025] The fault diagnosis model is used to perform fault diagnosis on the rail transit transportation system according to the event characteristics of at least one event; the event characteristics of the at least one event are determined based on a feature sequence corresponding to data generated by the rail transit transportation system.

[0026] In another embodiment of the present application, a data processing method is provided. The method includes:

[0027] Obtain historical log information and historical operation status information generated by the system during historical periods;

[0028] Extracting feature information from the historical log information;

[0029] Determining the association relationship between the characteristic information of the historical log information and the historical operating status information;

[0030] Determining a training sample based on the characteristic information of the historical log information, the historical operating status information, and the correlation between the two;

[0031] Using the training samples, the model to be trained is trained to obtain a fault diagnosis model;

[0032] The fault diagnosis model is used to perform fault diagnosis on the system according to the event characteristics of at least one event; the event characteristics of the at least one event are determined based on a feature sequence corresponding to the log information generated by the system.

[0033] In another embodiment of the present application, a rail transportation system is provided. The rail transportation system includes:

[0034] at least one rail vehicle;

[0035] at least one working device for assisting the at least one rail vehicle in traveling;

[0036] A fault diagnosis device is configured to obtain data generated by the rail transit system, extract characteristic information from the data, determine a characteristic sequence, wherein the characteristic sequence includes characteristic information extracted from the data generated at different times and is arranged in chronological order; determine event characteristics belonging to an event based on the characteristic sequence; and perform fault diagnosis on the rail transit system based on the event characteristics;

[0037] The output device is used to output the diagnosis result of the fault diagnosis of the rail transportation system.

[0038] In another embodiment of the present application, a vehicle operation system is provided. The vehicle operation system includes:

[0039] At least one vehicle;

[0040] at least one working device for assisting the operation of the at least one vehicle;

[0041] a fault diagnosis device configured to obtain data generated by at least one working device in the vehicle operation system; extract characteristic information from the data; determine a characteristic sequence, wherein the characteristic sequence includes characteristic information extracted from the data generated at different times and arranged in chronological order; and perform fault diagnosis on the vehicle operation system based on the characteristic sequence;

[0042] An output device is used to output a diagnosis result of a fault diagnosis of the vehicle operation system.

[0043] In another embodiment of the present application, a data processing system is provided. The data processing system includes:

[0044] At least one working device, configured to generate log information;

[0045] a fault diagnosis device configured to obtain log information generated by the at least one working device; extract feature information from the log information; determine a feature sequence, wherein the feature sequence includes feature information extracted from the log information generated at different times and is arranged in chronological order; determine an event feature of at least one event based on the feature sequence; and perform fault diagnosis on the system based on the event feature of the at least one event;

[0046] An output device is used to output a diagnosis result of a fault diagnosis performed on the system.

[0047] In another embodiment of the present application, an electronic device is provided. The electronic device includes a memory and a processor, wherein:

[0048] The memory is used to store programs;

[0049] The processor is coupled to the memory and is configured to execute the program stored in the memory to implement the steps of the methods provided in the above embodiments, which will be further described below.

[0050] In another embodiment of the present application, a computer-readable storage medium is provided. The computer-readable storage medium is used to store a computer-readable program or instruction. When the program or instruction is executed by a processor, the steps of the method provided in each of the above embodiments are implemented. This will also be further described below.

[0051] In another embodiment of the present application, a computer program product is provided. The computer program product includes a computer program or instructions. When the computer program or instructions are executed by a processor, the steps of the methods provided in the above embodiments are implemented. This will also be further described below.

[0052] Some embodiments of the present application provide technical solutions that extract feature information from data generated by a rail transit transportation system or a vehicle operation system; the extracted feature information is used to determine a feature sequence; the feature sequence contains feature information extracted from the data generated at different times and is arranged in chronological order; then, fault diagnosis of the rail transit transportation system or the vehicle operation system is performed based on the feature sequence, thereby improving the degree of automation of system fault diagnosis, and the diagnosis results can guide targeted maintenance, improve the efficiency of fault diagnosis, reduce delays caused by rail transit stops, and so on.

[0053] Other embodiments of the present application provide technical solutions that extract feature information from log information generated by a system (such as a rail transit transportation system, a vehicle operation system, a network system, an intelligent manufacturing system, etc.) to determine a feature sequence based on the feature information extracted from the log information generated at different times, and the feature information in the feature sequence is arranged in chronological order; then, the event features of at least one event are determined based on the feature sequence, and the system is fault-diagnosed using the event features of at least one event. Because the log information contains evidence of normal and unusual activities occurring in the system, the technical solutions provided by the embodiments of the present application, by mining the features in the system log information and considering the overall perspective of the event (i.e., determining the event features based on the feature sequence), can provide more accurate diagnostic results, achieve early warning of faults, guide targeted maintenance, thereby reducing the failure rate, and reduce delays caused by rail transit parking, etc.

[0054] Some other embodiments of the present application also provide a solution for training a fault diagnosis model, that is, through the system's historical data (or historical log information) and historical operating status, training samples are determined, and the training model is trained using the training samples to obtain a fault diagnosis model; the fault diagnosis model can use the event features determined in the above embodiments as input and output as a diagnosis result. It can be seen that the technical solution provided by this embodiment makes the entire system more intelligent and automated; the fault diagnosis model can achieve early fault warning, guide targeted maintenance, reduce system failure rate, and reduce losses caused by the system not being able to work properly. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0056] Figure 1 A flow chart of a data processing method in a rail transportation transportation system provided by one embodiment of the present application;

[0057] Figure 2 A theoretical logic diagram of a data processing method in a rail transportation transportation system provided by an embodiment of the present application;

[0058] Figure 3 A flowchart of a data processing method in a vehicle operation system provided by another embodiment of the present application;

[0059] Figure 4A flowchart of a data processing method provided in one embodiment of the present application;

[0060] Figure 5 A flowchart of a data processing method provided by another embodiment of the present application;

[0061] Figure 6 A flowchart of a data processing method provided in another embodiment of the present application;

[0062] Figure 7 A schematic structural diagram of a rail transportation system provided in one embodiment of the present application;

[0063] Figure 8 A schematic diagram of the structure of a data processing device in a rail transportation transportation system provided by an embodiment of the present application;

[0064] Figure 9 A schematic structural diagram of a data processing device in a vehicle operation system provided in another embodiment of the present application;

[0065] Figure 10 A schematic structural diagram of a data processing device provided in another embodiment of the present application;

[0066] Figure 11 A schematic diagram of the structure of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0067] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0068] In some processes described in the specification, claims and the above-mentioned figures of this application, multiple operations that appear in a specific order are included, and these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. In addition, the embodiments described below are only some of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0069] With the development of artificial intelligence (AI), the concept of intelligent operations (AIOps) has emerged. This approach uses AI to analyze large amounts of data from a variety of operations tools and devices, automatically identifying and responding to system issues in real time, thereby improving system operations capabilities and automation.

[0070] Under the AIOps trend, intelligent fault discovery and root cause diagnosis technologies driven by multi-source operation and maintenance data and centered on machine learning and other algorithms have attracted widespread attention. Multi-source operation and maintenance data include system runtime data and historical record data (such as forms, system update documents, system operation status, etc.). Compared with historical record data, system runtime data can reflect the dynamic characteristics of the system and the contextual information when the system fails, and has better detection and expression capabilities for unknown faults. The following embodiments of this application use the data generated by the system in real time to diagnose system faults.

[0071] Figure 1 The following is a flow chart of a data processing method provided by an embodiment of the present application. As shown in the figure, the method includes:

[0072] 101. Extract feature information from the data generated by the rail transit transportation system.

[0073] 102. Determine a feature sequence, wherein the feature sequence includes feature information extracted from the data generated at different times and is arranged in chronological order.

[0074] 103. Perform fault diagnosis on the rail transit transportation system based on the feature sequence.

[0075] In the above 101, the data generated during the operation of the system may include log information and monitoring data. Log information is text data generated by the printout code embedded in the program by the program developer to assist in debugging. It is used to record variable information, program execution status, etc. during system operation. Monitoring data refers to the specific operating conditions of the system under the operating state; such as the operating speed and suspension of rail vehicles in the rail transportation system. Log information and monitoring data are at different levels. Log information focuses on fine-grained application status and cross-component program execution logic, while monitoring data focuses on system status and coarse-grained application status.

[0076] Extracting feature information corresponding to data can be accomplished using a feature dictionary. This feature dictionary can be constructed based on historical data generated by the rail transit system over a period of time. Simply put, the feature dictionary provides mapping templates corresponding to data of different formats or types. When extracting feature information from data, the mapping template adapted for the data can be retrieved from the feature dictionary. The feature information of the data can then be determined using this mapping template.

[0077] When the rail transportation system is updated, such as software upgrade or hardware upgrade, the feature dictionary may be updated in a timely manner due to the emergence of new formats or types of data.

[0078] For example, the amount of log information generated by the rail transit system is large, and the structure, form, and type are diverse. Log information generally includes two categories, namely formatted logs and unformatted logs. The format of the formatted log is determined. The unformatted log basically includes: timestamp, device ID, information type, and detailed information. Among them, the timestamp indicates the specific time when the device generates the log message; the device ID indicates the identification of the device that generates the log message; the message type describes the general characteristics of the log message; and the detailed information describes the specific event described by the log message. The formats of unformatted logs are diverse, so it is necessary to map the information of the unformatted logs to the corresponding mapping templates, and then extract the required (i.e. valuable) feature information from the data based on the mapping templates.

[0079] In the above 102, the data generated by the rail transit system is continuously generated over time. This data generated by the rail transit system over time can be referred to as a data stream. Feature extraction is performed on each data item in the data stream, and the extracted feature information is then added to a feature sequence in chronological order. In specific implementations, the maximum number of feature information items that can be included in the feature sequence may be pre-defined, but this maximum number is not specifically limited in this embodiment.

[0080] In the above 103, the fault diagnosis of the rail transportation system based on the feature sequence can be implemented in the following two ways, but not limited to.

[0081] Method 1: Time slice-based method

[0082] That is, feature sequence segments are intercepted from the feature sequence according to the set time slice. For example, the time slice is set to the time period of 1:00 to 1:30; the corresponding interception process is: intercepting the feature sequence segments corresponding to the time period of 1:00 to 1:30 in the feature sequence. Then, based on the intercepted feature sequence segments, the rail transportation transportation system is diagnosed for faults. For example, in one solution, the intercepted feature sequence segments are processed using a fault diagnosis model, an expert system, etc. to obtain a diagnosis result. The fault diagnosis model can be a neural network model, etc. This embodiment does not specifically limit the specific implementation of the fault diagnosis model and the expert system.

[0083] Method 1: Event-based method

[0084] Before introducing the event-based approach, let's first briefly explain the concept of "event." An event is a single occurrence that occurs within a specific timeframe and environment, involves one or more actors, and consists of one or more actions. The event-based approach, taking an event-based approach, extracts a feature sequence segment corresponding to an event from a feature sequence as the event signature. This feature signature is then used to diagnose rail transportation system faults.

[0085] Specifically, the above step 103 of “performing fault diagnosis on the rail transportation system based on the feature sequence” may include the following steps:

[0086] 1031. Determine an event feature of at least one event based on the feature sequence;

[0087] 1032. Perform fault diagnosis on the rail transit transportation system based on event characteristics of the at least one event.

[0088] In one feasible solution, the above step 1031 of “determining event features of at least one event based on the feature sequence” can be implemented by the following steps:

[0089] S11, obtaining the travel information of rail vehicles in the rail transportation system;

[0090] S12. In combination with the travel information of the rail vehicle, perform at least one extraction operation on the feature sequence to extract at least one feature information belonging to the same event after the extraction operation to obtain an event feature of an event.

[0091] Unlike intercepting feature sequence segments from a feature sequence according to a set time slice, the second method is based on events. Taking events as the basis for interception, it is necessary to combine the real-time driving information of the rail vehicle (such as but not limited to: speed, acceleration, stop duration, etc.). In the above step S12, the extraction operation can be understood as the following process: from the feature sequence, identify the starting feature information and ending feature information that meet the requirements of an event; then, extract the starting feature information, ending feature information, and the feature sequence segment between the starting feature information and the ending feature information from the feature sequence as event features.

[0092] In specific implementations, the aforementioned process of identifying the starting and ending feature information, and extracting event features from the feature sequence, can be implemented by an event extraction model. Specifically, the feature sequence is input into the event extraction model, which processes the feature sequence to output event features for at least one event. In other words, the function of the event extraction model in this embodiment is to extract from the feature sequence a feature sequence segment belonging to an event of interest in this embodiment as an event feature.

[0093] That is, in another feasible technical solution, the above step 1031 of "determining event features of at least one event based on the feature sequence" can be implemented by the following steps:

[0094] S21, obtaining an event extraction model;

[0095] S22, taking the feature sequence and the travel information of the rail vehicle as inputs of the event extraction model, executing the event extraction model to output event features of at least one event;

[0096] The event extraction model is obtained by training the model to be trained through training samples.

[0097] The event extraction model can be obtained by training a machine learning model (such as a neural network model). For example, a supervised learning method is used to train the machine learning model. Among them, the selection of training samples for supervised learning and the training process of the event extraction model can be found in the relevant content of the prior art, and this embodiment does not specifically limit this. In another feasible technical solution, a clustering (unsupervised) learning method can also be used to directly extract events from raw corpus, reducing the dependence on the corpus. The K-means algorithm is a typical clustering algorithm based on distance, which uses distance as an indicator of similarity, that is, the similarity between two objects depends on the distance between them. What we want to discuss in this article is: event extraction requires clustering different elements (i.e., feature information) in the same event, and which cluster the different elements (i.e., feature information) belong to needs to be determined by calculating their similarity.

[0098] Furthermore, the solution provided in this embodiment can also provide the client with an interface for adjusting the event extraction model, so that staff (or domain experts) can review the event features output by the event extraction model and propose adjustment plans for the event extraction model to further optimize the event extraction model. That is, the method provided in this embodiment can also include the following steps:

[0099] 104. Send the output event feature of the at least one event to the client;

[0100] 105. Receive adjustment information for the event extraction model sent by the client;

[0101] 106. Adjust parameters in the event extraction model according to the adjustment information.

[0102] In a specific implementation, step 1032 may specifically include processing the event characteristics of the at least one event using a fault diagnosis model to obtain a fault diagnosis result for the rail transit system. The fault diagnosis model may be used to detect whether a system anomaly has occurred, analyze the root cause of the anomaly, the probability that the anomaly will cause a system failure in a future period, and analyze the root cause of the system failure if a system failure occurs.

[0103] In a specific embodiment, the step 1032 of performing fault diagnosis on the rail transportation system according to the event characteristics of the at least one event may include the following steps:

[0104] S31, obtaining a fault diagnosis model;

[0105] S32: Use the event feature of the at least one event as the input of the fault diagnosis model, execute the fault diagnosis model and output a diagnosis result.

[0106] The above-mentioned fault diagnosis model is obtained by training the model to be trained. That is, the method provided in this embodiment may also include the following steps:

[0107] 107. Obtain a feature dictionary and operating status information associated with dictionary items in the feature dictionary;

[0108] 108. Determine a training sample based on the feature dictionary and the operating status information associated with the dictionary items in the feature dictionary;

[0109] 109. Use the training samples to train the model to be trained to obtain the fault diagnosis model.

[0110] In another feasible technical solution, the method provided in this embodiment may further include the following steps:

[0111] 110. Obtain a fault feature dictionary and fault information associated with dictionary entries in the fault feature dictionary;

[0112] 111. Perform network structure training and parameter training on a Bayesian network model based on the fault feature dictionary and the fault information associated with the dictionary items in the fault feature dictionary to obtain the fault diagnosis model.

[0113] Among them, the purpose of training the network structure and parameters of the Bayesian network model is to train a model that can better fit the data; the data is the dictionary items (nodes) contained in the fault feature dictionary mentioned above and the fault information (nodes) associated with the dictionary items in the fault feature dictionary; the association between the dictionary items and the fault information (i.e., the directed edges between the nodes).

[0114] Specifically, Bayesian network model structure training methods may include: constraint-based methods (dependency sharing / conditional independence testing), scoring-based search methods (primarily considering the model's fit and complexity), methods based on a combination of the above, and methods based on random sampling. For details, please refer to the corresponding content in the prior art and will not be elaborated here.

[0115] A Bayesian network model, also known as a belief network or directed acyclic graph model, is a probabilistic graphical model (a type of statistical model) that represents a set of variables and their conditional dependencies via a directed acyclic graph (DAG). Bayesian networks are well-suited for taking an event that has already occurred and predicting the likelihood that any one of several possible known causes was a contributing factor. Formally, a Bayesian network is a directed acyclic graph (DAG) whose nodes represent variables in the Bayesian sense. These can be observables, latent variables, unknown parameters, or hypotheses. Edges represent conditional dependencies; unconnected nodes (no path connecting one node to another) represent variables that are conditionally independent of each other. Each node is associated with a probability function that takes as input a set of feature values ​​of the node's parent variables and gives (as output) the probability of the variable represented by the node.

[0116] Therefore, Bayesian network model learning can be broken down into two sub-stages: 1. Learning the network topology, or directed acyclic graph, referred to as structure learning; 2. Learning the local conditional probability distribution of each variable in the network, referred to as parameter learning. Similarly, the parameter training process for Bayesian network models can be found in related literature and is not limited here.

[0117] In addition to the Bayesian network model, this embodiment may also use models such as support vector machines and neural networks for training to obtain a fault diagnosis model.

[0118] Similarly, the solution provided in this embodiment may also provide an interface for manually labeling the diagnosis results, so that the diagnosis results and the event features based on the diagnosis results can be used as training samples to further train the fault diagnosis model. That is, the solution provided in this embodiment also includes the following steps:

[0119] 112. Sending a diagnosis result obtained by performing fault diagnosis on the rail transportation system to a client;

[0120] 113. After receiving the correct mark or error mark for the diagnosis result fed back by the user through the client, the correct mark or error mark corresponding to the diagnosis result and the event feature of the at least one event are used as sample data to train the fault diagnosis model.

[0121] The system failure process includes three stages: the occurrence of the root cause of the failure, abnormal system behavior, and system operation failure. The root cause of the failure causes abnormal system behavior, which in turn causes the system failure. The root cause of the failure, the abnormality, and the failure appear in sequence during the software system failure process, with a temporal and causal relationship. The purpose of fault diagnosis is to detect abnormal information represented by the system during the abnormal system behavior stage; predict possible future failures; and diagnose the root cause of the failure. Therefore, fault diagnosis includes three tasks: anomaly detection, root cause diagnosis, and fault prediction. That is, the diagnostic results output by the fault diagnosis model in step S32 of the method provided in this embodiment may include, but are not limited to, at least part of the following: the root cause of the abnormality when the rail transit system has an abnormality, the probability of a failure in the future due to the abnormality of the rail transit system, and the root cause of the failure when the rail transit system fails.

[0122] In the technical solution provided in this embodiment, feature information is extracted from the data generated by the rail transit transportation system to determine a feature sequence based on the feature information extracted from the data generated at different times; then, the event features of at least one event are determined according to the feature sequence, and the event features of at least one event are used to diagnose faults in the rail transit transportation system. This is in line with the thinking logic of people in fault diagnosis, and considering the overall perspective of the event, the diagnostic results are more accurate, and early warning of faults can be achieved, guiding targeted maintenance, thereby reducing the failure rate and delays caused by rail transit stops, etc.

[0123] The rail transit system in this embodiment can be a SACEM system (rail transit signaling system), such as a subway SACEM system. To summarize, the method provided in this embodiment first extracts event features corresponding to at least one event from a feature sequence based on an event extraction model. Then, based on the event features of the at least one event, a diagnosis is performed on the fault probability, fault type, and root cause. The feature sequence is determined based on real-time data (e.g., log information) generated by the rail transit system.

[0124] In the prior art, after a train failure occurs, experts manually review the SACEM system's log information and determine the fault type based on their experience. Due to the large volume of SACEM system logs, manual fault diagnosis is time-consuming and cannot quickly locate the cause. Furthermore, experts' varying experience can lead to misjudgments, resulting in recurring faults. Furthermore, prior art diagnoses faults post-facto and cannot provide real-time fault warnings. Compared to prior art, the technical solution provided in this embodiment can automatically process large amounts of log information, eliminating the need for extensive human resources and unaffected by expert expertise. Furthermore, it allows for timely fault diagnosis, enabling timely or proactive preventive measures to reduce the failure rate.

[0125] As described above, in step 101 of this embodiment, feature information extraction can be implemented using a feature dictionary. Specifically, in one feasible technical solution, the step of "extracting feature information from data generated by the rail transit system" may include:

[0126] 1011. Obtain feature dictionary;

[0127] 1012. Using a feature matching algorithm, match the data with dictionary items in the feature dictionary;

[0128] 1013. Determine the feature information according to the dictionary item matched by the data.

[0129] The above-mentioned dictionary items may include the mapping templates mentioned above.

[0130] It was mentioned above that the data generated by the system includes log information and monitoring data. In fault diagnosis, log information has more advantages than monitoring data. Fault diagnosis based on monitoring data can only locate abnormal fluctuations in a specific monitoring indicator, while fault diagnosis technology based on log information can locate specific error logs and events. In addition, log information can reflect the operation trajectory of the entire system in a more fine-grained manner, and the fault location can be accurately located by processing the log information. Therefore, in this embodiment, the data for extracting feature information in the above step 101 is log information. That is, the data generated by the rail transit transportation system includes log information; the feature dictionary includes a log template set (the same as the mapping template mentioned above), and the feature matching algorithm is a template matching algorithm. Accordingly, the above step 1012 "using a feature matching algorithm to match the data with the dictionary items in the feature dictionary, and determining the feature information according to the dictionary items matched to the data" includes:

[0131] S41, using the template matching algorithm, matching the log information with the log template in the log template set;

[0132] S42. Obtain a template identifier corresponding to a target log template, wherein the target log template matches a constant in the log information;

[0133] S43. Determine variables in the log information based on the target log template;

[0134] S44. Extracting log features from the variables of the log information;

[0135] S45. Obtain the feature information according to the template identifier corresponding to the target log template and the log feature.

[0136] The feature dictionary can be constructed based on historical data. That is, the feature dictionary is constructed by mining and selecting feature information of historical data generated during the system's historical period. That is, the solution provided in this embodiment also includes:

[0137] 114. Obtain historical data generated during the historical period of the rail transit system;

[0138] 115. Construct the feature dictionary based on the historical data.

[0139] Specifically, the historical data mentioned above can be historical log information of the system. Generally, log information includes constants and variables. Constants (which can be understood as information frameworks) summarize the events corresponding to the log information and events corresponding to similar log information. Variables are data within the information framework that changes as the system's real-time operation changes.

[0140] Therefore, when constructing a feature dictionary based on historical log information, constants of the historical log information may be identified first, and then the feature dictionary may be constructed based on the identified constant information.

[0141] Of course, before constructing a feature dictionary based on the historical log information, the historical log information can also be pre-processed, such as filtering out duplicate log information. For example, if the information described in one historical log is completely contained in another historical log, the two historical logs can be determined to be identical, and one of them can be deleted. On the other hand, if the information described in the two historical logs is largely the same, with only negligible minor differences, the two historical logs are determined to be similar and can be merged into a single record.

[0142] Furthermore, in specific implementation, the solution provided by this embodiment may also add operating status information associated with the dictionary item in the feature dictionary. That is, the method provided by this embodiment may also include the following steps:

[0143] 116. Obtain historical operating status information of the rail transit transportation system within a historical period;

[0144] 117. Determine the association relationship between the historical data and the historical operating status information;

[0145] 118. Determine the operating status information associated with the dictionary item in the feature dictionary based on the historical data, the historical operating status information, and the association relationship therebetween.

[0146] The historical operating status information may include normal status and fault status; in the case of a fault status, it may also include fault symptoms, root causes of the fault, etc.

[0147] Of course, in a specific embodiment, step 116 may only acquire historical fault information of the rail transit system. This historical fault information may include, but is not limited to, fault symptoms and root causes. That is, the acquired historical operating status information includes historical fault information. Accordingly, step 118, "determining the operating status information associated with the dictionary item in the feature dictionary based on the historical data, the historical operating status information, and the association between the two," may include:

[0148] 1181. Determine data related to the historical fault information in the historical data based on the historical data, the historical fault information, and a correlation between the historical data and the historical fault information;

[0149] 1182. Determine a fault feature dictionary based on data related to historical fault information in the historical data;

[0150] 1183. Associating dictionary items in the fault feature dictionary with corresponding historical fault information;

[0151] The fault information includes: fault symptoms and fault root causes.

[0152] In order to improve the accuracy of the feature dictionary, this embodiment may also provide a client interface to facilitate users (or domain experts) to adjust dictionary items in the feature dictionary according to the client. Specifically, the method provided in this embodiment may also include the following steps:

[0153] 119. Send the fault feature dictionary and historical fault information associated with dictionary items in the fault feature dictionary to the client, so that the user can review, confirm or modify the fault feature dictionary and / or historical fault information associated with dictionary items in the fault feature dictionary through the client.

[0154] See also Figure 2 The technical solution provided by this embodiment can be briefly described as follows:

[0155] Part I: Training

[0156] Collect historical data (i.e., data accumulation). For example, process this data using feature construction and selection algorithms to generate a feature dictionary. This feature dictionary contains dictionary entries and the system operating status information (e.g., fault information) associated with each entry. Based on the feature dictionary, training samples are generated. The training samples are used to train the model to be trained, resulting in a fault diagnosis model.

[0157] Figure 2 An example of selecting a Bayesian network model is shown in FIG. Based on training samples and using a Bayesian network structure learning algorithm, the Bayesian network structure is first trained. Then, a Bayesian network parameter learning algorithm is used to train the parameters in the Bayesian network structure, thereby obtaining a fault diagnosis model.

[0158] In addition, the solution provided by this embodiment provides interfaces for user participation in three stages, such as Figure 2 As shown, during the feature dictionary construction phase using the feature construction and selection algorithm, users can view the feature dictionary through the corresponding interface and, based on experience, adjust the feature dictionary when errors exist, or optimize the feature construction and selection algorithm to improve the accuracy of dictionary item generation in subsequent feature dictionaries. During the Bayesian network structure learning phase, users can participate in confirming and adjusting the Bayesian network structure or adjusting the Bayesian network structure learning algorithm through the corresponding interface. During the Bayesian network parameter learning phase, users can also participate in confirming and adjusting the Bayesian network parameters or adjusting the Bayesian network parameter learning algorithm through the corresponding interface.

[0159] Part II Diagnostic Process

[0160] The system acquires the real-time data stream generated by the system—a set of data arranged in chronological order. Then, based on a feature dictionary, a feature matching algorithm is used to extract feature information from the data and sort it chronologically to obtain a feature sequence. Subsequently, an event extraction model is used to extract feature sequence segments from the feature sequence as event features for a single event. Finally, the event features are used as input to a fault diagnosis model, which is then executed to obtain a fault diagnosis result.

[0161] See also Figure 2 As shown in , the solution provided in this embodiment also provides a user interface for adjusting the event extraction model. Users can use this interface to view the extracted feature sequence segments and then adjust the parameters in the event extraction model based on experience to improve the calculation accuracy of the event extraction model.

[0162] The fault diagnosis results can be sent to the user's client, where they can confirm the results, such as labeling them as correct or incorrect. The labeled and confirmed fault diagnosis results and event features can be used as new training data for regular or irregular training and optimization of the fault diagnosis model.

[0163] The above embodiment provides a data processing method in a rail transportation system. In essence, the inventive concept of the technical solution provided by the above embodiment is to obtain the data generated by the system in real time (such as log information), extract feature information; determine the feature sequence; and then perform fault diagnosis on the system based on the feature information with time continuity contained in the feature sequence; it can be applied to the following scenarios: road transportation system, air transportation system, maritime transportation system, etc. Of course, the inventive concept of the technical solution provided by the above embodiment can also be applied to intelligent processing systems. Accordingly, Figure 3 A flow chart of a data processing method in a vehicle operation system provided by an embodiment of the present application is shown. The vehicle operation system includes at least one vehicle and at least one working device for assisting the operation of the at least one vehicle. In different application scenarios, the specific implementation of the vehicle is different. When the vehicle operation system is a transportation system, the vehicle in the corresponding system is a vehicle; wherein the vehicle is any one of the following: a motor vehicle, an electric vehicle, a logistics distribution vehicle, an aircraft, a ship, and the like. When the vehicle operation system is a material sorting system, the vehicle in the corresponding system is a sorting robot or a sorting vehicle, and the like. When the vehicle operation system is an intelligent processing system, the vehicle in the corresponding system is processing equipment, a robot, a robotic arm, and the like.

[0164] The at least one working device may be mounted on the vehicle, independently mounted and communicatively connected to the vehicle, or partially mounted on the vehicle and partially external to the vehicle and communicatively connected to the vehicle. The working device may include, but is not limited to, various sensors for monitoring vehicle operating parameters (e.g., lidar, cameras, gyroscopes, etc.), storage media storing corresponding control programs or instructions, microcontrollers or processors for executing the control programs or instructions, and other actuators (e.g., relays, modems, etc.).

[0165] See also Figure 3 As shown, the data processing method in the vehicle operation system includes:

[0166] 201. Acquire data generated by at least one working device in the vehicle operation system;

[0167] 202. Extract feature information from the data;

[0168] 203. Determine a feature sequence, wherein the feature sequence includes feature information extracted from the data generated at different times and is arranged in chronological order;

[0169] 204. Perform fault diagnosis on the vehicle operation system based on the feature sequence.

[0170] In the above 201, the data generated by at least one working device may include log information.

[0171] The above step 204 of “performing fault diagnosis on the vehicle operation system based on the feature sequence” may specifically include:

[0172] 2041. Determine an event feature of at least one event based on the feature sequence;

[0173] 2042. Perform fault diagnosis on the vehicle operation system based on the event characteristics of the at least one event.

[0174] For the specific implementation of the above steps 202 to 204 and 2041 to 2042, please refer to the corresponding content in the above embodiment and will not be described in detail here.

[0175] Figure 4 The present application also provides a flow chart of a data processing method according to an embodiment. The type of system is not limited in this embodiment, and it can be the system mentioned in the above embodiment, or a CDN network (content distribution network) system, a switch system, an intelligent manufacturing system, etc. Figure 4 As shown, the data processing method includes:

[0176] 301. Extract characteristic information from the system log information;

[0177] 302. Determine a feature sequence, wherein the feature sequence includes feature information extracted from the log information generated at different times and arranged in chronological order;

[0178] 303. Determine an event feature of at least one event based on the feature sequence;

[0179] 304. Perform fault diagnosis on the system according to event characteristics of the at least one event.

[0180] The specific implementation of the above steps 301 to 304 can refer to the corresponding content in the above embodiments.

[0181] From another perspective, the above step 301 of "extracting feature information from system log information" may include:

[0182] 3021. Map the log information to a target log template in the log template set;

[0183] 3022. Determine variables in the log information based on the target log template;

[0184] 3023. Extract log features from the variables in the log information;

[0185] 3024. Obtain the feature information according to the template identifier corresponding to the target log template and the log feature.

[0186] Only by properly processing system log information can key features be effectively extracted. Typically, system log information is processed by extracting mapping templates from historical system log information and then mapping the log information to the corresponding mapping template. When software or hardware upgrades occur, new types of log information are generated, necessitating the creation of new mapping templates.

[0187] That is, the method provided in this embodiment may further include the following steps:

[0188] 305. If the log information cannot be mapped to any log template in the log template set, determine a corresponding log template based on the log information;

[0189] 306. Add the log template determined based on the log information to the log template set to update the log template set.

[0190] In step 305, a system log learning method or algorithm is used to automatically learn mapping templates from system log information. These mapping templates are then used to extract features from the log information. Furthermore, because the number of different mapping templates is far smaller than the number of different log information, the overhead of subsequent feature extraction is low after the log information is mapped to the corresponding mapping template.

[0191] It should be noted here that the log template set in this embodiment can be included in the feature dictionary in the above embodiment.

[0192] The solution provided in this embodiment may also provide a client interface, through which users (or domain experts) can improve or modify the log template set. Specifically, the method provided in this embodiment may also include the following steps:

[0193] 307. Sending a log template determined based on the log information to the client;

[0194] 308. When the client feeds back confirmation information regarding the log template, triggering a step of adding the log template determined based on the log information to the log template set;

[0195] 309. When the client feeds back modification information for the log template, modify the log template according to the modification information, so as to add the modified log template to the log template set.

[0196] Furthermore, in the method of this embodiment, step 303 of “determining event features of at least one event based on the feature sequence” may specifically include:

[0197] Using an event extraction model, determining starting feature information and ending feature information belonging to the same event in the feature sequence, and extracting the starting feature information, the ending feature information, and a feature sequence segment between the starting feature information and the ending feature information from the feature sequence;

[0198] The extraction result is used as the event feature of the event.

[0199] Of course, the event extraction model can also extract event features of more than one event from the feature sequence at one time, which is not limited in this embodiment.

[0200] Similarly, the solution provided in this embodiment also provides a client interface, through which users (or domain experts) can adjust the event extraction model. That is, the method provided in this embodiment can also include the following steps:

[0201] 310 : After receiving adjustment information for the event extraction model fed back by the client based on the extraction result, adjust parameters in the event extraction model according to the adjustment information.

[0202] Furthermore, in the method provided in this embodiment, step 304 of “performing fault diagnosis on the system according to the event characteristics of the at least one event” includes:

[0203] 3041. Obtain a fault diagnosis model;

[0204] 3042. Use the event feature of the at least one event as the input of the fault diagnosis model, execute the fault diagnosis model and output a diagnosis result.

[0205] Specifically, the fault diagnosis model is a Bayesian network model that has undergone structure training and parameter training.

[0206] In addition to the Bayesian network model, this embodiment may also use models such as support vector machines and neural networks for training to obtain a fault diagnosis model.

[0207] In the above embodiments, it is mentioned that the fault diagnosis model is used to diagnose the system fault. The fault diagnosis model can be obtained by training (or self-learning). A fault diagnosis model determination scheme will be provided below. Specifically, Figure 5 The flowchart of the data processing method provided by an embodiment of the present application is shown in FIG. Figure 5 As shown, the method includes:

[0208] 401. Obtain historical data and historical operation status information generated by the rail transportation system during historical periods;

[0209] 402. Extract feature information from the historical data;

[0210] 403. Determine the association between the characteristic information of the historical data and the historical operating status information;

[0211] 404. Determine a training sample based on the characteristic information of the historical data, the historical operating status information, and the correlation between the two.

[0212] 405. Using the training samples, train the model to be trained to obtain a fault diagnosis model;

[0213] The fault diagnosis model is used to perform fault diagnosis on the rail transit transportation system according to the event characteristics of at least one event; the event characteristics of the at least one event are determined based on a feature sequence corresponding to data generated by the rail transit transportation system.

[0214] The solution provided in this embodiment may further include the following steps:

[0215] 406. Construct a feature dictionary based on multiple historical data;

[0216] The feature dictionary provides support for extracting feature information from data, that is, using a feature matching algorithm to match the data with dictionary items in the feature dictionary to determine the feature information of the data according to the dictionary items matched to the data.

[0217] Furthermore, the historical operating status information includes historical fault information. Accordingly, the above step 404 "determining training samples based on the characteristic information of the historical data, the historical operating status information and the correlation between the two" includes:

[0218] 4041. Determine characteristic information related to the historical fault information based on the characteristic information of the historical data, the historical fault information, and the correlation between the characteristic information and the historical fault information.

[0219] 4042. Using feature information related to the historical fault information and the historical fault information as training samples;

[0220] The historical fault information includes fault symptoms and fault root causes.

[0221] In a specific implementation, step 405 of the method provided in this embodiment, "using the training samples to train the model to be trained to obtain a fault diagnosis model," can be specifically as follows:

[0222] The Bayesian network model is trained on network structure and parameters using a plurality of the training samples to obtain the fault diagnosis model.

[0223] The training of the fault diagnosis model can be triggered irregularly or periodically, so as to train the fault diagnosis model with continuously updated training samples and continuously optimize the fault diagnosis model.

[0224] Figure 6 A method for determining a fault diagnosis model provided by another embodiment of the present application is shown. Specifically, this embodiment provides a data processing method, including:

[0225] 501. Obtain historical log information and historical operation status information generated by the system during historical periods;

[0226] 502. Extract feature information from the historical log information;

[0227] 503. Determine the association between the characteristic information of the historical log information and the historical operating status information;

[0228] 504. Determine a training sample based on the feature information of the historical log information, the historical operating status information, and the association between the feature information and the historical operating status information;

[0229] 505. Using the training samples, train the model to be trained to obtain a fault diagnosis model;

[0230] The fault diagnosis model is used to perform fault diagnosis on the system according to the event characteristics of at least one event; the event characteristics of the at least one event are determined based on a feature sequence corresponding to the log information generated by the system.

[0231] Furthermore, the method provided in this embodiment may further include the following steps:

[0232] 506. Construct a log template set based on characteristic information of multiple historical log messages;

[0233] Among them, the log template set provides support for extracting feature information from log information, that is, by mapping the log information to a target log template in the log template set, using the target log template to extract log features, and determining the feature information of the log information based on the template identifier corresponding to the target log template and the log features.

[0234] In a specific embodiment, the historical operating status information includes historical fault information. Accordingly, step 504 of this embodiment, "determining training samples based on the characteristic information of the historical log information, the historical operating status information, and the relationship between the two," may include:

[0235] 5041. Determine characteristic information related to the historical fault information based on the characteristic information of the historical log information, the historical fault information, and the correlation between the characteristic information and the historical fault information;

[0236] 5042. Using feature information related to the historical fault information and the historical fault information as training samples;

[0237] The historical fault information includes fault symptoms and fault root causes.

[0238] Furthermore, in the method of this embodiment, step 505 of “using the training samples to train the model to be trained to obtain a fault diagnosis model” may include:

[0239] The Bayesian network model is trained on network structure and parameters using a plurality of the training samples to obtain the fault diagnosis model.

[0240] Similarly, the training of the fault diagnosis model can be triggered irregularly or periodically, so as to train the fault diagnosis model with continuously updated training samples and continuously optimize the fault diagnosis model.

[0241] The execution subject of the methods provided in the above embodiments can be a fault diagnosis device in the system. The fault diagnosis device can be an independent device deployed in the system, or it can be deployed in a working device in the system, or it can be deployed on a server device on the network side. This embodiment does not specifically limit this. Accordingly, this application also provides a system that applies the methods of the above embodiments.

[0242] Figure 7 Figure 1 shows a schematic diagram of the structure of a rail transportation system. Figure 7 As shown, the system includes: at least one rail vehicle 14, at least one working device 11, a fault diagnosis device 12 and an output device 13.

[0243] at least one working device 11 for assisting the at least one rail vehicle 14 in traveling;

[0244] The fault diagnosis device 12 is configured to obtain data generated by the rail transit system, extract characteristic information from the data, determine a characteristic sequence, wherein the characteristic sequence includes characteristic information extracted from the data generated at different times and arranged in chronological order, and perform fault diagnosis on the rail transit system based on the characteristic sequence;

[0245] The output device 13 is used to output the diagnosis result of the fault diagnosis of the rail transportation system.

[0246] The at least one working device may include, but is not limited to, devices deployed on rail vehicles, or devices deployed at stations or tracks. Devices deployed on rail vehicles may include, but are not limited to, accelerometers, communication devices, speed controllers, and vehicle door control devices (such as door sensors and door opening and closing actuators). Devices deployed at stations or tracks may include, but are not limited to, access points (APs), communication devices, and cables.

[0247] The output device 13 may be a display device connected to the fault diagnosis device, for example, a large screen deployed in a subway or high-speed rail control room.

[0248] The vehicle operation system can also adopt the technical solution provided by this application. That is, this application also provides an embodiment of a vehicle operation system, the corresponding diagram of which is not shown in the accompanying drawings of the specification. Figure 7 Similar, Figure 7The icon pointed to by the middle number 14 can be replaced with the corresponding image of a vehicle, AGV, drone, airplane, ship, etc. Specifically, the vehicle operation system includes: at least one vehicle, at least one working device, fault diagnosis equipment and output equipment.

[0249] At least one vehicle;

[0250] at least one working device for assisting the operation of the at least one vehicle;

[0251] a fault diagnosis device configured to obtain data generated by at least one working device in the vehicle operation system; extract characteristic information from the data; determine a characteristic sequence, wherein the characteristic sequence includes characteristic information extracted from the data generated at different times and arranged in chronological order; and perform fault diagnosis on the vehicle operation system based on the characteristic sequence;

[0252] An output device is used to output a diagnosis result of a fault diagnosis of the vehicle operation system.

[0253] In addition to the rail transportation system, other systems, such as communication systems composed of network equipment, intelligent manufacturing systems (such as manufacturing systems composed of assembly line equipment), etc., can also adopt the technical solutions provided in the above embodiments. That is, another embodiment of the present application provides a data processing system. The data processing system includes: at least one working device, a fault diagnosis device, and an output device. Among them,

[0254] At least one working device, configured to generate log information;

[0255] a fault diagnosis device configured to obtain log information generated by the at least one working device; extract feature information from the log information; determine a feature sequence, wherein the feature sequence includes feature information extracted from the log information generated at different times and arranged in chronological order; determine an event feature of at least one event based on the feature sequence; and perform fault diagnosis on the system based on the event feature of the at least one event;

[0256] An output device is used to output a diagnosis result of a fault diagnosis performed on the system.

[0257] The specific implementation of these working devices will vary across different systems. For example, in an intelligent manufacturing system, the working devices may be manufacturing equipment (such as CNC machining equipment, robotic arms, etc.). In a communication system composed of network devices, the working devices may be switches, routers, servers, and so on.

[0258] It should be added here that: in addition to implementing the functional steps described in the respective embodiments, the fault diagnosis equipment in the above-mentioned system embodiments can also implement other functional steps in the above-mentioned method embodiments. The specific content can be found in the description above and will not be repeated here.

[0259] Figure 8 FIG. 1 shows a schematic diagram of the structure of a data processing device in a rail transportation system provided by an embodiment of the present application. Figure 8 As shown, the device includes: an extraction module 21, a determination module 22, and a diagnosis module 23. The extraction module 21 is configured to extract feature information from data generated by the rail transit system. The determination module 22 is configured to determine a feature sequence, wherein the feature sequence includes feature information extracted from the data generated at different times and arranged in chronological order. The diagnosis module 23 is configured to perform fault diagnosis on the rail transit system based on the feature sequence.

[0260] Furthermore, when extracting feature information from the data generated by the rail transportation system, the extraction module 21 is specifically used to:

[0261] Acquire a feature dictionary; use a feature matching algorithm to match the data with dictionary items in the feature dictionary; and determine the feature information according to the dictionary items matched by the data.

[0262] Furthermore, the data generated by the rail transit system includes log information; the feature dictionary includes a set of log templates, and the feature matching algorithm is a template matching algorithm. Accordingly, when the extraction module 21 uses the feature matching algorithm to match the data with dictionary items in the feature dictionary and determines the feature information according to the dictionary items matched by the data, it is specifically used to:

[0263] Using the template matching algorithm, matching the log information with the log templates in the log template set;

[0264] Obtaining a template identifier corresponding to a target log template, wherein the target log template matches a constant in the log information;

[0265] Determining variables in the log information based on the target log template;

[0266] Extracting log features from the variables of the log information;

[0267] The characteristic information is obtained according to the template identifier corresponding to the target log template and the log characteristic.

[0268] Furthermore, the apparatus provided in this embodiment may further include an acquisition module and a construction module. The acquisition module is used to acquire historical data generated by the rail transit system during a historical period. The construction module is used to construct the feature dictionary based on the historical data.

[0269] Furthermore, the acquisition module is further configured to acquire historical operating status information of the rail transit system within a historical period. The determination module is further configured to determine an association between the historical data and the historical operating status information; and to determine the operating status information associated with a dictionary item in the feature dictionary based on the historical data, the historical operating status information, and the association between the two.

[0270] Furthermore, the acquired historical operating status information includes historical fault information. Accordingly, when the determination module 22 determines the operating status information associated with the dictionary item in the feature dictionary based on the historical data, the historical operating status information, and the association between the two, it is specifically used to:

[0271] Determining data related to the historical fault information in the historical data based on the historical data, the historical fault information, and a correlation between the historical data and the historical fault information;

[0272] determining a fault feature dictionary based on data related to historical fault information in the historical data;

[0273] Associating dictionary items in the fault feature dictionary with corresponding historical fault information;

[0274] The fault information includes: fault symptoms and fault root causes.

[0275] Furthermore, the apparatus provided in this embodiment further includes a sending module. The sending module is configured to send the fault feature dictionary and historical fault information associated with dictionary entries in the fault feature dictionary to a client, so that a user can review, confirm, or modify the fault feature dictionary and / or historical fault information associated with dictionary entries in the fault feature dictionary through the client.

[0276] Furthermore, when the diagnostic module 23 performs fault diagnosis on the rail transportation system based on the feature sequence, it is specifically configured to:

[0277] determining an event feature of at least one event based on the feature sequence;

[0278] Performing fault diagnosis on the rail transit transportation system according to the event characteristics of the at least one event.

[0279] Furthermore, when the diagnostic module 23 determines the event characteristics of at least one event based on the feature sequence, it is specifically used to: obtain the driving information of the rail vehicle in the rail transportation transportation system; and perform at least one extraction operation on the feature sequence in combination with the driving information of the rail vehicle, so as to extract at least one feature information belonging to the same event after the extraction operation to obtain the event characteristics of an event.

[0280] Alternatively, when determining the event feature of at least one event based on the feature sequence, the diagnosis module 23 is specifically configured to: obtain an event extraction model; obtain travel information of a rail vehicle in the rail transportation system; use the feature sequence and the travel information of the rail vehicle as inputs to the event extraction model, execute the event extraction model, and output the event feature of at least one event;

[0281] The event extraction model is obtained by training the model to be trained through training samples.

[0282] Furthermore, the apparatus provided in this embodiment includes a sending module, a receiving module, and an adjustment module. The sending module is configured to send the output event features of at least one event to a client. The receiving module is configured to receive adjustment information for the event extraction model sent by the client. The adjustment module is configured to adjust parameters in the event extraction model according to the adjustment information.

[0283] Furthermore, when the diagnostic module 23 performs fault diagnosis on the rail transit transportation system based on the event characteristics of the at least one event, it is specifically used to: obtain a fault diagnosis model; use the event characteristics as input of the fault diagnosis model, execute the fault diagnosis model and output a diagnosis result.

[0284] Furthermore, the apparatus provided in this embodiment also includes an acquisition module and a training module. The acquisition module is configured to acquire a feature dictionary and operating status information associated with dictionary entries in the feature dictionary; and to determine training samples based on the feature dictionary and the operating status information associated with dictionary entries in the feature dictionary. The training module is configured to train a model to be trained using the training samples to obtain the fault diagnosis model.

[0285] Alternatively, the apparatus provided in this embodiment includes an acquisition module configured to acquire a fault feature dictionary and fault information associated with dictionary entries in the fault feature dictionary. The training module is configured to perform network structure training and parameter training on a Bayesian network model based on the fault feature dictionary and the fault information associated with dictionary entries in the fault feature dictionary to obtain the fault diagnosis model.

[0286] Furthermore, the apparatus provided in this embodiment includes a sending module and a sample data determination module. The sending module is configured to send a diagnostic result obtained by performing fault diagnosis on the rail transit system to a client. The sample data determination module is configured to, upon receiving a correct or incorrect labeling of the diagnostic result from a user via the client, use the correct or incorrect labeling corresponding to the diagnostic result and the event features as sample data for training the fault diagnosis model.

[0287] Furthermore, the diagnosis results in this embodiment include at least part of the following:

[0288] The root cause of the abnormality when the rail transit transportation system has an abnormality, the probability of a failure in the future due to the abnormality of the rail transit transportation system, and the root cause of the failure when the rail transit transportation system fails.

[0289] It should be noted here that the data processing device provided in the above embodiment can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can be found in the corresponding contents in the above method embodiments, which will not be repeated here.

[0290] Figure 9 FIG. 1 shows a schematic diagram of the structure of a data processing device in a vehicle operation system provided by another embodiment of the present application. Figure 9 As shown, the data processing device in the vehicle operation system includes: an acquisition module 31, an extraction module 32, a determination module 33 and a diagnosis module 34. Among them, the acquisition module 31 is used to obtain data generated by at least one working device in the vehicle operation system. The extraction module 32 is used to extract feature information from the data. The determination module 33 is used to determine a feature sequence, wherein the feature sequence contains feature information extracted from the data generated at different times and arranged in chronological order. The diagnosis module 34 is used to perform fault diagnosis on the vehicle operation system based on the feature sequence.

[0291] Furthermore, when the diagnostic module 34 performs fault diagnosis on the vehicle operation system based on the feature sequence, it is specifically used to: determine the event characteristics of at least one event based on the feature sequence; and perform fault diagnosis on the vehicle operation system according to the event characteristics of the at least one event.

[0292] Furthermore, the vehicle operation system is a transportation system, and the vehicle in the corresponding system is a vehicle; wherein the vehicle is any one of the following: a motor vehicle, an electric vehicle, a logistics delivery vehicle, an aircraft, or a ship. Alternatively, the vehicle operation system is a material sorting system, and the vehicle in the corresponding system is a sorting robot or a sorting vehicle. Alternatively, the vehicle operation system is an intelligent processing system, and the vehicle in the corresponding system is processing equipment, a robot, or a robotic arm.

[0293] It should be noted here that the data processing device in the vehicle operation system provided by the above embodiment can implement the technical solution described in the above corresponding method embodiment. The specific implementation principles of the above modules or units can be found in the corresponding contents in the above method embodiment, which will not be repeated here.

[0294] Another embodiment of the present application provides a data processing device, the structural block diagram of which is the same as Figure 8 As shown. The data processing device includes: an extraction module, a determination module and a diagnosis module. The extraction module is used to extract feature information from the system log information. The determination module is used to determine a feature sequence, wherein the feature sequence contains feature information extracted from the log information generated at different times and is arranged in chronological order; based on the feature sequence, the event feature of at least one event is determined. The diagnosis module is used to perform fault diagnosis on the system based on the event feature of the at least one event.

[0295] Furthermore, when extracting feature information from the system log information, the extraction module is specifically used to:

[0296] The log information is mapped to a target log template in a log template set; based on the target log template, variables in the log information are determined; log features are extracted from the variables in the log information; and the feature information is obtained according to the template identifier corresponding to the target log template and the log features.

[0297] Furthermore, the determination module is further configured to determine a corresponding log template based on the log information when the log information cannot be mapped to any log template in the log template set. Accordingly, the apparatus provided in this embodiment further includes an adding module. The adding module is configured to add the log template determined based on the log information to the log template set to update the log template set.

[0298] Furthermore, the device provided in this embodiment also includes a sending module and a triggering module. The sending module is used to send the log template determined based on the log information to the client. The triggering module is used to trigger the adding module to add the log template determined based on the log information to the log template set when the client feeds back confirmation information for the log template; when the client feeds back modification information for the log template, the log template is modified according to the modification information, and the adding module is triggered to add the modified log template to the log template set.

[0299] Furthermore, when determining the event feature of at least one event based on the feature sequence, the determining module is specifically configured to:

[0300] Using an event extraction model, determining starting feature information and ending feature information belonging to the same event in the feature sequence, and extracting the starting feature information, the ending feature information, and a feature sequence segment between the starting feature information and the ending feature information from the feature sequence;

[0301] The extraction results are used as event features of an event.

[0302] Furthermore, the apparatus provided in this embodiment includes a sending module, a receiving module, and an adjustment module. The sending module is configured to send the extraction results to the client; the receiving module is configured to receive adjustment information for the event extraction model provided by the client based on the extraction results; and the adjustment module is configured to adjust parameters in the event extraction model based on the adjustment information received by the receiving module.

[0303] Furthermore, when the diagnosis module performs fault diagnosis on the system according to the event characteristics of the at least one event, it is specifically configured to:

[0304] Acquire a fault diagnosis model; use the event feature of the at least one event as input to the fault diagnosis model, execute the fault diagnosis model and output a diagnosis result.

[0305] Specifically, the fault diagnosis model is a Bayesian network model that has undergone structure training and parameter training.

[0306] It should be noted here that the data processing device provided in the above embodiment can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can be found in the corresponding contents in the above method embodiments, which will not be repeated here.

[0307] Figure 10The structural diagram of the data processing device provided by an embodiment of the present application is shown. The device includes: an acquisition module 41, an extraction module 42, a determination module 43 and a training module 44. Among them, the acquisition module 41 is used to obtain historical data and historical operating status information generated by the rail transit transportation system in the historical period. The extraction module 42 is used to extract feature information from the historical data. The determination module 43 is used to determine the correlation between the feature information of the historical data and the historical operating status information; and determine the training sample based on the feature information of the historical data, the historical operating status information and the correlation between the two. The training module 44 is used to use the training sample to train the model to be trained to obtain a fault diagnosis model. Among them, the fault diagnosis model is used to perform fault diagnosis on the rail transit transportation system based on the event characteristics of at least one event; the event characteristics of the at least one event are determined based on the feature sequence corresponding to the data generated by the rail transit transportation system.

[0308] Furthermore, the apparatus provided in this embodiment also includes a construction module. The construction module is configured to construct a feature dictionary based on multiple historical data. The feature dictionary supports extracting feature information from the data, i.e., using a feature matching algorithm to match the data with dictionary entries in the feature dictionary to determine the feature information of the data based on the dictionary entries matched to the data.

[0309] Furthermore, the historical operating status information includes historical fault information. Correspondingly, when determining the training sample based on the characteristic information of the historical data, the historical operating status information and the correlation between the two, the determination module is specifically used to:

[0310] Determining characteristic information related to the historical fault information based on the characteristic information of the historical data, the historical fault information, and a correlation between the two;

[0311] Using feature information related to the historical fault information and the historical fault information as training samples;

[0312] The historical fault information includes fault symptoms and fault root causes.

[0313] Furthermore, when the training module 44 uses the training samples to train the model to be trained and obtain the fault diagnosis model, it is specifically used to: use the training samples to perform network structure training and parameter training on the Bayesian network model to obtain the fault diagnosis model.

[0314] It should be noted here that the data processing device provided in the above embodiment can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can be found in the corresponding contents in the above method embodiments, which will not be repeated here.

[0315] Another embodiment of the present application provides a data processing device, which has the same structure as above. Figure 10 . Specifically, the data processing device includes: an acquisition module, an extraction module, a determination module and a training module. Among them, the acquisition module is used to obtain historical log information and historical operating status information generated by the system in a historical period. The extraction module is used to extract feature information from the historical log information. The determination module is used to determine the correlation between the feature information of the historical log information and the historical operating status information; and determine the training sample based on the feature information of the historical log information, the historical operating status information and the correlation between the two. The training module is used to use the training sample to train the model to be trained to obtain a fault diagnosis model. Among them, the fault diagnosis model is used to perform fault diagnosis on the system based on the event characteristics of at least one event; the event characteristics of the at least one event are determined based on the feature sequence corresponding to the log information generated by the system.

[0316] Furthermore, the data processing device provided in this embodiment also includes a construction module. The construction module is configured to construct a log template set based on the characteristic information of multiple historical log messages. The log template set supports extracting characteristic information from log messages by mapping the log messages to a target log template in the log template set, extracting log features using the target log template, and determining the characteristic information of the log messages based on the template identifier corresponding to the target log template and the log features.

[0317] Furthermore, the historical operating status information includes historical fault information. Correspondingly, when determining the training sample based on the characteristic information of the historical log information, the historical operating status information and the correlation between the two, the determination module is specifically used to:

[0318] Determining characteristic information related to the historical fault information based on the characteristic information of the historical log information, the historical fault information, and a correlation between the characteristic information and the historical fault information;

[0319] Using feature information related to the historical fault information and the historical fault information as training samples;

[0320] The historical fault information includes fault symptoms and fault root causes.

[0321] Furthermore, when the training module uses the training samples to train the model to be trained to obtain the fault diagnosis model, it is specifically used to:

[0322] The Bayesian network model is trained on network structure and parameters using a plurality of the training samples to obtain the fault diagnosis model.

[0323] It should be noted here that the data processing device provided in the above embodiment can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can be found in the corresponding contents in the above method embodiments, which will not be repeated here.

[0324] Figure 11 FIG. 1 shows a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Figure 11 As shown, the electronic device includes a memory 51 and a processor 52. The memory 51 can be configured to store various other data to support operations on the electronic device. Examples of such data include instructions for any application or method operating on the electronic device. The memory 51 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0325] The memory 51 is used to store programs;

[0326] The processor 52 is coupled to the memory 51 and is configured to execute the program stored in the memory 51 to:

[0327] Extract feature information from data generated by rail transportation systems;

[0328] Determining a feature sequence, wherein the feature sequence includes feature information extracted from the data generated at different times and is arranged in chronological order;

[0329] Based on the characteristic sequence, fault diagnosis is performed on the rail transportation system.

[0330] In particular, when the processor 52 executes the program in the memory 51 , in addition to the above functions, it can also implement other functions, for details, please refer to the description of the corresponding method embodiment above.

[0331] Further, if Figure 11 As shown, the electronic device also includes: a communication component 53, a display 54, a power component 55, an audio component 56 and other components. Figure 11 Only some components are shown schematically, which does not mean that the electronic device only includes Figure 11 Components shown.

[0332] Another embodiment of the present application provides an electronic device with the same structure as above. Figure 11 Specifically, the electronic device includes a memory and a processor, wherein:

[0333] The memory is used to store programs;

[0334] The processor is coupled to the memory and is configured to execute the program stored in the memory to:

[0335] Acquire data generated by at least one working device in the vehicle operation system;

[0336] Extracting feature information from the data;

[0337] Determining a feature sequence, wherein the feature sequence includes feature information extracted from the data generated at different times and is arranged in chronological order;

[0338] Based on the feature sequence, fault diagnosis is performed on the vehicle operation system.

[0339] Similarly, when the processor executes the program in the memory, in addition to the above functions, it can also implement other functions. For details, please refer to the description of the corresponding method embodiment above.

[0340] Another embodiment of the present application provides an electronic device with the same structure as above. Figure 11 Specifically, the electronic device includes a memory and a processor, wherein:

[0341] The memory is used to store programs;

[0342] The processor is coupled to the memory and is configured to execute the program stored in the memory to:

[0343] Extract feature information from system log information;

[0344] Determining a feature sequence, wherein the feature sequence includes feature information extracted from the log information generated at different times and is arranged in chronological order;

[0345] determining an event feature of at least one event based on the feature sequence;

[0346] A fault diagnosis is performed on the system according to an event feature of the at least one event.

[0347] Similarly, when the processor executes the program in the memory, in addition to the above functions, it can also implement other functions. For details, please refer to the description of the corresponding method embodiment above.

[0348] Another embodiment of the present application provides an electronic device, the structure of which is the same as the above figure. Specifically, the electronic device includes a memory and a processor, wherein:

[0349] The memory is used to store programs;

[0350] The processor is coupled to the memory and is configured to execute the program stored in the memory to:

[0351] Obtain historical data and historical operation status information generated by the rail transportation system during historical periods;

[0352] extracting feature information from the historical data;

[0353] Determining an association relationship between the characteristic information of the historical data and the historical operating status information;

[0354] Determining training samples based on the characteristic information of the historical data, the historical operating status information, and the correlation between the two;

[0355] Using the training samples, the model to be trained is trained to obtain a fault diagnosis model;

[0356] The fault diagnosis model is used to perform fault diagnosis on the rail transit transportation system according to the event characteristics of at least one event; the event characteristics of the at least one event are determined based on a feature sequence corresponding to data generated by the rail transit transportation system.

[0357] Similarly, when the processor executes the program in the memory, in addition to the above functions, it can also implement other functions. For details, please refer to the description of the corresponding method embodiment above.

[0358] Another embodiment of the present application provides an electronic device, the structure of which is the same as the above figure. Specifically, the electronic device includes a memory and a processor, wherein:

[0359] The memory is used to store programs;

[0360] The processor is coupled to the memory and is configured to execute the program stored in the memory to:

[0361] Obtain historical log information and historical operation status information generated by the system during historical periods;

[0362] Extracting feature information from the historical log information;

[0363] Determining the association relationship between the characteristic information of the historical log information and the historical operating status information;

[0364] Determining training samples based on the characteristic information of the historical log information, the historical operating status information, and the correlation between the two;

[0365] Using the training samples, the model to be trained is trained to obtain a fault diagnosis model;

[0366] The fault diagnosis model is used to perform fault diagnosis on the system according to the event characteristics of at least one event; the event characteristics of the at least one event are determined based on a feature sequence corresponding to the log information generated by the system.

[0367] Similarly, when the processor executes the program in the memory, in addition to the above functions, it can also implement other functions. For details, please refer to the description of the corresponding method embodiment above.

[0368] Accordingly, an embodiment of the present application also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, the method steps or functions provided by the above-mentioned method embodiments can be implemented.

[0369] The present application also provides a computer program product. The computer program product includes a computer program or instructions. When the computer program or instructions are executed by a processor, the computer program or instructions can implement the method steps or functions provided in the above-mentioned method embodiments.

[0370] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0371] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0372] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A data processing method in a rail transportation system, characterized in that: include: Extracting feature information from data generated by the rail transportation system; Determining a feature sequence, wherein the feature sequence includes feature information extracted from the data generated at different times and is arranged in chronological order; Performing fault diagnosis on the rail transportation system based on the characteristic sequence; Performing fault diagnosis on the rail transportation system based on the characteristic sequence includes: determining an event feature of at least one event based on the feature sequence; Performing fault diagnosis on the rail transportation system according to the event characteristics of the at least one event; Determining an event feature of at least one event based on the feature sequence includes: Acquiring travel information of rail vehicles in the rail transportation system; In combination with the travel information of the rail vehicle, performing at least one extraction operation on the feature sequence to extract at least one feature information belonging to the same event after the extraction operation to obtain an event feature of an event; Performing fault diagnosis on the rail transportation system according to the event characteristics of the at least one event includes: Obtaining a fault diagnosis model; Taking the event feature of the at least one event as input of the fault diagnosis model, executing the fault diagnosis model and outputting a diagnosis result; Obtaining a fault feature dictionary and fault information associated with dictionary items in the fault feature dictionary; According to the fault feature dictionary and the fault information associated with the dictionary items in the fault feature dictionary, network structure training and parameter training are performed on the Bayesian network model to obtain the fault diagnosis model.

2. The method according to claim 1, characterized in that Extracting characteristic information from the data generated by the rail transportation system, including: Get the feature dictionary; Using a feature matching algorithm, matching the data with dictionary items in the feature dictionary; The feature information is determined according to the dictionary item matched to the data.

3. The method according to claim 2, characterized in that The data generated by the rail transportation system includes log information; the feature dictionary includes a log template set, and the feature matching algorithm is a template matching algorithm; and using a feature matching algorithm to match the data with dictionary items in the feature dictionary, and determining the feature information according to the dictionary items matched by the data, including: Using the template matching algorithm, matching the log information with the log templates in the log template set; Obtaining a template identifier corresponding to a target log template, wherein the target log template matches a constant in the log information; Determining variables in the log information based on the target log template; Extracting log features from the variables of the log information; The characteristic information is obtained according to the template identifier corresponding to the target log template and the log characteristic.

4. The method according to claim 2 or 3, characterized in that Also includes: Obtaining historical data generated during a historical period of the rail transit system; The feature dictionary is constructed based on the historical data.

5. The method according to claim 4, characterized in that Also includes: Obtaining historical operating status information of the rail transit transportation system within a historical period; Determining an association relationship between the historical data and the historical operating status information; The operating status information associated with the dictionary items in the feature dictionary is determined according to the historical data, the historical operating status information and the association relationship therebetween.

6. The method according to claim 5, characterized in that The acquired historical operating status information includes historical fault information; and Determining the operating status information associated with a dictionary item in the feature dictionary based on the historical data, the historical operating status information, and the association relationship therebetween includes: Determining data related to the historical fault information in the historical data based on the historical data, the historical fault information, and the correlation between the historical data and the historical fault information; determining a fault feature dictionary based on data related to historical fault information in the historical data; Associating dictionary items in the fault feature dictionary with corresponding historical fault information; The fault information includes: fault symptoms and fault root causes.

7. The method according to claim 6, characterized in that Also includes: The fault feature dictionary and historical fault information associated with dictionary items in the fault feature dictionary are sent to the client, so that the user can review, confirm or modify the fault feature dictionary and / or historical fault information associated with dictionary items in the fault feature dictionary through the client.

8. The method according to claim 1, characterized in that Determining an event feature of at least one event based on the feature sequence includes: Get the event extraction model; Taking the feature sequence and the travel information of the rail vehicle in the rail transportation system as inputs of the event extraction model, executing the event extraction model to output event features of at least one event; The event extraction model is obtained by training the model to be trained through training samples.

9. The method according to claim 8, characterized in that Also includes: sending the output event characteristics of at least one event to the client; Receiving adjustment information for the event extraction model sent by the client; According to the adjustment information, the parameters in the event extraction model are adjusted.

10. The method according to claim 1, characterized in that Also includes: Obtaining a feature dictionary and operating status information associated with dictionary items in the feature dictionary; determining a training sample based on the feature dictionary and the operating status information associated with the dictionary items in the feature dictionary; The training sample is used to train the model to be trained to obtain the fault diagnosis model.

11. The method according to claim 1, wherein The diagnosis result includes at least part of the following: The root cause of the abnormality when the rail transit transportation system has an abnormality, the probability of a failure in the future due to the abnormality of the rail transit transportation system, and the root cause of the failure when the rail transit transportation system fails.

12. The method according to claim 1, characterized in that Also includes: Sending the diagnosis result obtained by performing fault diagnosis on the rail transportation system to the client; After receiving the correct mark or error mark for the diagnosis result fed back by the user through the client, the correct mark or error mark corresponding to the diagnosis result and the event feature of the at least one event are used as sample data to train the fault diagnosis model.

13. A data processing method in a vehicle operation system, characterized in that: The method comprises: Acquire data generated by at least one working device in the vehicle operation system; Extracting feature information from the data; Determining a feature sequence, wherein the feature sequence includes feature information extracted from the data generated at different times and is arranged in chronological order; Performing fault diagnosis on the vehicle operation system based on the feature sequence; performing fault diagnosis on the vehicle operation system based on the feature sequence includes: determining an event feature of at least one event based on the feature sequence; Performing fault diagnosis on the vehicle operation system according to event characteristics of the at least one event; Determining an event feature of at least one event based on the feature sequence includes: Acquiring the driving information of the rail vehicle in the vehicle operation system; In combination with the travel information of the rail vehicle, performing at least one extraction operation on the feature sequence to extract at least one feature information belonging to the same event after the extraction operation to obtain an event feature of an event; Performing fault diagnosis on the vehicle operation system according to the event characteristics of the at least one event includes: Obtaining a fault diagnosis model; Taking the event feature of the at least one event as input of the fault diagnosis model, executing the fault diagnosis model and outputting a diagnosis result; Obtaining a fault feature dictionary and fault information associated with dictionary items in the fault feature dictionary; According to the fault feature dictionary and the fault information associated with the dictionary items in the fault feature dictionary, network structure training and parameter training are performed on the Bayesian network model to obtain the fault diagnosis model.

14. The method according to claim 13, characterized in that The vehicle operation system is a transportation system, and the vehicle in the corresponding system is a vehicle; wherein the vehicle is any one of the following: a motor vehicle, an electric vehicle, a logistics delivery vehicle, an aircraft, and a ship; or The carrier operation system is a material sorting system, and the carrier in the corresponding system is a sorting robot or a sorting vehicle; or The carrier operation system is an intelligent processing system, and the carriers in the corresponding system are processing equipment, robots, and robotic arms.

15. A data processing method, characterized in that: include: Extract feature information from the log information of the data processing system; Determining a feature sequence, wherein the feature sequence includes feature information extracted from the log information generated at different times and is arranged in chronological order; determining an event feature of at least one event based on the feature sequence; performing fault diagnosis on the data processing system according to event characteristics of the at least one event; Performing fault diagnosis on the data processing system according to the event characteristics of the at least one event includes: Obtaining a fault diagnosis model; Taking the event feature of the at least one event as input of the fault diagnosis model, executing the fault diagnosis model and outputting a diagnosis result; The fault diagnosis model is a Bayesian network model that has undergone structure training and parameter training; Determining an event feature of at least one event based on the feature sequence includes: Using an event extraction model, determining starting feature information and ending feature information belonging to the same event in the feature sequence, and extracting the starting feature information, the ending feature information, and a feature sequence segment between the starting feature information and the ending feature information from the feature sequence; The extraction results are used as event features of an event.

16. The method according to claim 15, characterized in that Extract characteristic information from the log information of the data processing system, including: Mapping the log information to a target log template in a log template set; Determining variables in the log information based on the target log template; Extracting log features from the variables of the log information; The characteristic information is obtained according to the template identifier corresponding to the target log template and the log characteristic.

17. The method according to claim 16, characterized in that Also includes: If the log information cannot be mapped to any log template in the log template set, determining a corresponding log template based on the log information; Adding the log template determined based on the log information to the log template set to update the log template set.

18. The method according to claim 17, characterized in that Also includes: Sending a log template determined based on the log information to the client; When the client feeds back confirmation information for the log template, triggering a step of adding the log template determined based on the log information to the log template set; In the case where the client feeds back modification information for the log template, the log template is modified according to the modification information, so as to add the modified log template to the log template set.

19. The method according to claim 18, characterized in that Also includes: Send the extraction results to the client; After receiving the adjustment information for the event extraction model fed back by the client based on the extraction result, the parameters in the event extraction model are adjusted according to the adjustment information.

20. A data processing method, characterized in that: The invention is characterized by comprising: Obtain historical data and historical operation status information generated by the rail transportation system during historical periods; extracting feature information from the historical data; Determining an association relationship between the characteristic information of the historical data and the historical operating status information; Determining training samples based on the characteristic information of the historical data, the historical operating status information, and the correlation between the two; Using the training samples, the model to be trained is trained to obtain a fault diagnosis model; The fault diagnosis model is used to perform fault diagnosis on the rail transit system based on an event feature of at least one event; the event feature of the at least one event is determined based on a feature sequence corresponding to data generated by the rail transit system; The training sample is used to train the model to be trained to obtain a fault diagnosis model, including: The training samples are used to perform network structure training and parameter training on the Bayesian network model to obtain the fault diagnosis model.

21. The method according to claim 20, characterized in that Also includes: Build a feature dictionary based on multiple historical data; The feature dictionary provides support for extracting feature information from data, that is, using a feature matching algorithm to match the data with dictionary items in the feature dictionary to determine the feature information of the data according to the dictionary items matched to the data.

22. The method according to claim 20 or 21, characterized in that The historical operating status information includes historical fault information; as well as Determining a training sample based on the characteristic information of the historical data, the historical operating status information, and the correlation between the two includes: Determining characteristic information related to the historical fault information based on the characteristic information of the historical data, the historical fault information, and a correlation between the two; Using feature information related to the historical fault information and the historical fault information as training samples; The historical fault information includes fault symptoms and fault root causes.

23. A data processing method, characterized in that: include: Obtain historical log information and historical operating status information generated by the data processing system during historical periods; Extracting feature information from the historical log information; Determining the association relationship between the characteristic information of the historical log information and the historical operating status information; Determining a training sample based on the characteristic information of the historical log information, the historical operating status information, and the correlation between the two; Using the training samples, the model to be trained is trained to obtain a fault diagnosis model; The fault diagnosis model is used to perform fault diagnosis on the data processing system based on an event feature of at least one event; the event feature of the at least one event is determined based on a feature sequence corresponding to log information generated by the data processing system; The training sample is used to train the model to be trained to obtain a fault diagnosis model, including: The training samples are used to perform network structure training and parameter training on the Bayesian network model to obtain the fault diagnosis model.

24. The method according to claim 23, wherein Also includes: Construct a log template set based on the characteristic information of multiple historical log messages; Among them, the log template set provides support for extracting feature information from log information, that is, by mapping the log information to a target log template in the log template set, using the target log template to extract log features, and determining the feature information of the log information based on the template identifier corresponding to the target log template and the log features.

25. The method according to claim 23 or 24, characterized in that The historical operating status information includes historical fault information; as well as Determining a training sample based on the feature information of the historical log information, the historical operating status information, and the correlation between the two includes: Determining characteristic information related to the historical fault information based on the characteristic information of the historical log information, the historical fault information, and a correlation between the two; Using feature information related to the historical fault information and the historical fault information as training samples; The historical fault information includes fault symptoms and fault root causes.

26. A rail transportation system, characterized in that: Applying the method of claims 1 to 12, the rail transportation system comprises: at least one rail vehicle; at least one working device for assisting the at least one rail vehicle in traveling; A fault diagnosis device is configured to obtain data generated by the rail transit system and extract characteristic information from the data; determine a characteristic sequence, wherein the characteristic sequence includes characteristic information extracted from the data generated at different times and is arranged in chronological order; and perform fault diagnosis on the rail transit system based on the characteristic sequence; The output device is used to output the diagnosis result of the fault diagnosis of the rail transportation system.

27. A vehicle operation system, characterized in that: Applying the method of claims 13 to 14, the vehicle operation system comprises: At least one vehicle; at least one working device for assisting the operation of the at least one vehicle; a fault diagnosis device configured to obtain data generated by at least one working device in the vehicle operation system; extract characteristic information from the data; determine a characteristic sequence, wherein the characteristic sequence includes characteristic information extracted from the data generated at different times and arranged in chronological order; and perform fault diagnosis on the vehicle operation system based on the characteristic sequence; An output device is used to output a diagnosis result of a fault diagnosis of the vehicle operation system.

28. A data processing system, characterized in that: Applying the method of claims 15 to 19, the data processing system comprises: At least one working device, configured to generate log information; a fault diagnosis device configured to obtain log information generated by the at least one working device; extract feature information from the log information; determine a feature sequence, wherein the feature sequence includes feature information extracted from the log information generated at different times and is arranged in chronological order; determine an event feature of at least one event based on the feature sequence; and perform fault diagnosis on the data processing system based on the event feature of the at least one event; An output device is used to output a diagnosis result of a fault diagnosis performed on the data processing system.

29. An electronic device, characterized in that: comprising a memory and a processor, wherein, The memory is used to store programs; The processor, coupled to the memory, is configured to execute the program stored in the memory to implement the steps of the method described in any one of claims 1 to 12; or implement the steps of the method described in any one of claims 13 to 14; or implement the steps of the method described in any one of claims 15 to 19; or implement the steps of the method described in any one of claims 20 to 22; or implement the steps of the method described in any one of claims 23 to 25.

30. A computer-readable storage medium, characterized in that Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps of the method described in any one of claims 1 to 12; or implement the steps of the method described in any one of claims 13 to 14; or implement the steps of the method described in any one of claims 15 to 19; or implement the steps of the method described in any one of claims 20 to 22; or implement the steps of the method described in any one of claims 23 to 25.

31. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instructions are executed by a processor, it can implement the steps of the method described in any one of claims 1 to 12; or implement the steps of the method described in any one of claims 13 to 14; or implement the steps of the method described in any one of claims 15 to 19; or implement the steps of the method described in any one of claims 20 to 22; or implement the steps of the method described in any one of claims 23 to 25.

Images