Safety operation domain monitoring method for virtual marshalling train and related device

Abstract

The invention discloses a safety operation domain monitoring method for a virtual marshalling train and a related device, and relates to the technical field of rail transit train operation control. Comprising vehicle system related data, signal system data, comprehensive monitoring system data, intelligent operation and maintenance system data and environment sensing data; inputting the safe operation domain state data into the trained safe operation domain situation prediction model to obtain a safe operation domain change trend; inputting the change trend of the safe operation domain into a driving risk propagation model to obtain the probability of occurrence of a driving accident; the driving risk propagation model represents an association relationship between a monitoring index of a safe operation domain and a driving accident; determining a driving risk level according to the occurrence probability of the driving accident and the driving accident consequence severity level; and generating a risk relief strategy according to the driving risk level. Real-time decision support can be provided for the virtual marshalling train, and safe and reliable operation of the virtual marshalling train is ensured.

Description

Technical Field

[0001] The present application relates to the field of rail transit train operation control technology, and in particular to a safe operation domain monitoring method and related devices for virtual train formations. Background Art

[0002] In recent years, the development of technologies such as train-to-train communication, high-precision positioning, and coordinated control has led to the emergence of virtual train formations. This technology, through the dynamic coordinated operation of multiple trains, breaks free from the limitations of physical couplers. It can flexibly adjust the number of train formations and running intervals based on actual passenger demand, thereby precisely matching traffic volume needs. This effectively addresses the capacity waste of traditional fixed train formations and significantly improves operational economy and resource utilization.

[0003] During the operation of virtual trains, to ensure simultaneous departures and arrivals at stations and facilitate passenger boarding and alighting, as well as operational management, the train control system employs a "soft wall" protection method. This involves the trailing train dynamically adjusting its protection curve based on the leading train's estimated stopping position, rather than the traditional "hard wall" protection method, where the trailing train calculates its protection curve using the leading train's current position as a fixed stopping point. While this improvement increases operational flexibility, it also further compresses the driving safety margin and imposes stricter requirements on safe operating assumptions (referred to as the "safe operating domain"), including the operating environment, equipment performance, and clearance requirements.

[0004] However, due to factors such as braking technology and sensor accuracy, as well as external interference such as weather changes and emergencies, there is significant uncertainty in the safe operation domain, making it difficult for virtual trains to operate stably in all weather conditions, on all lines, and in all scenarios. Therefore, an online, accurate operational risk assessment method is urgently needed to provide real-time decision support for virtual train control systems and ensure their safe and reliable operation. Summary of the Invention

[0005] The purpose of this application is to provide a safe operation domain monitoring method and related devices for virtual train formations, which can realize online and accurate operation risk assessment, provide real-time decision support for the virtual train formation control system, and ensure its safe and reliable operation.

[0006] To achieve the above objectives, this application provides the following solutions:

[0007] In a first aspect, the present application provides a method for monitoring a safe operation domain of a virtual train formation, comprising:

[0008] Acquire safe operation domain status data; the safe operation domain status data includes vehicle system related data, signal system data, integrated monitoring system data, intelligent operation and maintenance system data, and environmental perception data;

[0009] Inputting the safe operation domain status data into the trained safe operation domain situation prediction model to obtain a safe operation domain change trend;

[0010] Inputting the change trend of the safe operation domain into a traffic risk propagation model to obtain the probability of traffic accidents; the traffic risk propagation model is expressed as the correlation between the monitoring indicators of the safe operation domain and traffic accidents;

[0011] Determining a driving risk level based on the probability of the driving accident and a preset severity level of the consequences of the driving accident;

[0012] A risk mitigation strategy is generated according to the driving risk level.

[0013] Optionally, the process of determining the trained safe operation domain situation prediction model specifically includes:

[0014] Constructing a safe operation domain model based on a preset safe operation assumption of a virtual marshaled train; the safe operation domain model includes element categories, monitoring indicators, and thresholds of the safe operation domain;

[0015] Acquire an initial safe operation domain historical state data set and a corresponding label data set; the initial safe operation domain historical state data set and the corresponding label data set are determined based on the safe operation domain model;

[0016] Performing multi-source heterogeneous data fusion on the initial safe operation domain historical state data set to obtain a feature vector set;

[0017] Based on the feature vector set and the label data set, a preset safe operation domain situation prediction model is trained to obtain a trained safe operation domain situation prediction model.

[0018] Optionally, multi-source heterogeneous data fusion is performed on the initial safe operation domain historical status data set to obtain a feature vector set, including:

[0019] Performing data cleaning on the initial safe operation domain historical state data set to obtain a first safe operation domain historical state data set; the data cleaning includes removing abnormal values ​​and redundant data;

[0020] Performing data conversion on the first safe operation domain historical status data set to obtain a second safe operation domain historical status data set; the data is converted into a unified timestamp and a standardized data format;

[0021] Performing spatiotemporal correlation and feature extraction on the second safe operation domain historical status data set to obtain a feature vector set.

[0022] Optionally, the safe operation domain change trend is input into a driving risk propagation model to obtain the probability of a driving accident, specifically including:

[0023] Inputting the change trend of the safe operation domain into the driving risk propagation model to obtain the driving risk propagation path;

[0024] Calculate the probability of traffic accidents based on the driving risk transmission path.

[0025] Optionally, the driving risk propagation path is composed of a monitoring indicator node of a safe operation domain, a driving accident node, and multiple risk propagation nodes between the monitoring indicator node and the corresponding driving accident node; the probability of a driving accident is calculated according to the driving risk propagation path, specifically as follows:

[0026] According to the formula P(traffic accident|current safe operation domain status)=∑ 所有行车风险传播路径 (∏ 行车风险传播路径上每个节点 P (current node status | parent node status) safe operation domain calculates the probability of traffic accidents;

[0027] Among them, P(traffic accident|current safe operation domain status) represents the probability of a traffic accident under the condition of the current safe operation domain status; ∑[all traffic risk propagation paths] represents the sum of the probability results of all traffic risk propagation paths; ∏[each node on the traffic risk propagation path] represents the multiplication of the conditional probability of each node on the traffic risk propagation path; P(current node status|parent node status) safe operation domain represents the conditional probability of the current node status under the parent node status under the condition of the current safe operation domain status. The current node is any node on the traffic risk propagation path, and the parent node status is the previous node.

[0028] Optionally, the trained safe operation domain situation prediction model includes an input layer, a feature encoding layer, a multi-scale time series extraction layer, an attention fusion layer, a recursive prediction layer, and an output layer connected in sequence; inputting the safe operation domain state data into the trained safe operation domain situation prediction model to obtain a safe operation domain change trend specifically includes:

[0029] Input the multi-modal safe operation domain state data into the input layer to obtain the pre-processed safe operation domain state data;

[0030] Inputting the pre-processed safe operation domain state data into the feature coding layer to obtain feature-coded time series data;

[0031] Inputting the feature-encoded time series data into a multi-scale time series extraction layer to obtain a multi-scale feature map; the multi-scale feature map includes local mutation features of the safe operation domain, state evolution features of the safe operation domain, and global trend features of the safe operation domain;

[0032] Inputting the multi-scale feature map into the attention fusion layer to obtain a weighted fused comprehensive feature vector;

[0033] Inputting the weighted fused comprehensive feature vector into the recursive prediction layer to obtain a hidden state sequence;

[0034] The hidden state sequence is input into the output layer to obtain the change trend of the safe operation domain.

[0035] Optionally, the risk mitigation strategy includes an on-board control strategy, a driving dispatch control strategy, and a maintenance strategy; and generating a risk mitigation strategy based on the driving risk level specifically includes:

[0036] When the driving risk level is high, a vehicle control strategy is generated; the vehicle control strategy is to issue an alarm to the vehicle or the driver and give a braking or speed reduction prompt;

[0037] When the driving risk level is medium risk, a driving dispatching control strategy is generated; the driving dispatching control strategy is to adjust the train operation plan or separation according to the specific type of driving risk.

[0038] When the driving risk level is low risk, a maintenance strategy is generated; the maintenance strategy is to take necessary preventive measures based on the specific conditions of the driving risk.

[0039] In a second aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any one of the above-mentioned methods for monitoring the safe operation domain of a virtual train assembly.

[0040] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned methods for monitoring the safe operation domain of a virtual train formation.

[0041] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned methods for monitoring the safe operation domain of a virtual train formation.

[0042] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0043] The present application provides a method and related devices for monitoring the safe operation domain of a virtual train formation. By acquiring and analyzing vehicle systems, signal systems, integrated monitoring systems, intelligent operation and maintenance systems, and environmental perception data, various types of information that affect driving safety can be fully acquired. The combination of these data provides a more accurate basis for subsequent risk level prediction. Moreover, by processing these data using a safe operation domain situation prediction model, it is possible to capture the changing trend of the safe operation domain in a timely manner, and combine the changing trend of the safe operation domain with the driving risk propagation model, which can more accurately predict the probability of driving accidents. The combination of the probability of accidents and the severity level of the consequences of accidents helps to determine the driving risk level in different scenarios and achieve more accurate risk assessment and decision-making. In addition, based on the driving risk level, the system can generate corresponding risk mitigation strategies, thereby providing real-time decision support for the potential risks of virtual train formations and ensuring their safe and reliable operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0045] Figure 1 This is an application environment diagram of a safe operation domain monitoring method for virtual marshaled trains provided in Example 1 of the present application;

[0046] Figure 2 A flowchart of a method for monitoring a safe operation domain of a virtual train set provided in Example 1 of the present application;

[0047] Figure 3 Schematic diagram of the execution of the safe operation domain monitoring method for virtual marshaling trains provided in Example 2 of the present application;

[0048] Figure 4 Schematic diagram of the situation prediction model for the safe operation domain of a virtual marshaled train provided in Example 2 of this application;

[0049] Figure 5 A schematic diagram of the structure of the driving risk propagation model provided in Example 2 of the present application;

[0050] Figure 6 A schematic diagram of the structure of a computer device provided in Example 4 of the present application. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0052] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0053] Example 1

[0054] The safe operation domain monitoring method for virtual train formation provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can send the safe operation domain status data to the server 104, and the server 104 inputs the safe operation domain status data into the trained safe operation domain situation prediction model to obtain the safe operation domain change trend; inputs the safe operation domain change trend into the driving risk propagation model to obtain the probability of occurrence of driving accidents; determines the driving risk level according to the probability of occurrence of driving accidents and the preset driving accident consequence severity level; and generates a risk mitigation strategy according to the driving risk level. The server 104 can feedback the obtained risk mitigation strategy to the terminal 102.

[0055] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, and IoT devices. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or a cloud server.

[0056] In an exemplary embodiment, Figure 2 As shown, a method for monitoring the safe operation domain of a virtual train is provided. The method is executed by a computer device, specifically a computer device such as a terminal or a server, or a terminal and a server. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in the example is used as an example to illustrate the process, including the following steps 201 to 205.

[0057] Step 201, obtain safe operation domain status data; the safe operation domain status data includes vehicle system related data, signal system data, integrated monitoring system data, intelligent operation and maintenance system data and environmental perception data.

[0058] Step 202: Input the safe operation domain status data into the trained safe operation domain situation prediction model to obtain the safe operation domain change trend.

[0059] In step 203, the changing trend of the safe operation domain is input into the driving risk propagation model to obtain the probability of a driving accident; the driving risk propagation model is represented as the correlation between the monitoring indicators of the safe operation domain and driving accidents.

[0060] Step 204 : Determine the driving risk level based on the probability of the driving accident and the preset driving accident consequence severity level.

[0061] Step 205: Generate a risk mitigation strategy based on the driving risk level; the risk mitigation strategy is used to send warning information to the dispatch center and the vehicle control unit, and provide countermeasure suggestions or auxiliary decision-making.

[0062] By implementing steps 201 to 205 above, and acquiring and analyzing data from vehicle systems, signal systems, integrated monitoring systems, intelligent operation and maintenance systems, and environmental perception, a comprehensive range of information that affects driving safety can be obtained. The combination of this data provides a more accurate basis for subsequent risk level predictions. Furthermore, by processing this data using a safe operation domain situation prediction model, it is possible to timely capture changing trends in the safe operation domain. Combining these changing trends with the driving risk propagation model allows for a more accurate prediction of the probability of driving accidents. Combining the probability of accidents with the severity of their consequences helps determine driving risk levels in different scenarios, enabling more accurate risk assessment and decision-making. Furthermore, based on the driving risk level, the system can generate corresponding risk mitigation strategies, thereby providing real-time decision support for potential risks in virtual marshaling trains and ensuring their safe and reliable operation.

[0063] Example 2

[0064] This application builds a safe operation domain model and monitors the driving status in real time, which can timely predict and evaluate the driving risk level and generate risk mitigation strategies online, thereby effectively avoiding the occurrence of driving accidents. In order to achieve the above purpose, the specific execution steps are as follows: Figure 3 shown.

[0065] S1: Driving safety operation domain monitoring

[0066] S1.1 Define safe operation domain

[0067] The present application provides a method for defining a safe operation domain, and obtains a safe operation domain model based on the safe operation assumptions of virtual marshaling trains. Specifically, the safe operation assumptions stipulate the basic conditions for the safe operation of the virtual marshaling train control system, such as line conditions, equipment working environment conditions, vehicle braking performance requirements, and operator operation requirements. The text-based and decentralized safe operation assumptions designed during the development phase of the virtual marshaling train control system are not conducive to supervision and management. Based on the safe operation assumptions of the virtual marshaling train control system, the safe operation domain is defined according to element categories, monitoring indicators, and thresholds to facilitate online dynamic supervision.

[0068] Specifically, based on the safe operation assumptions, the performance parameter requirements of the virtual marshaling train control system related to speed measurement, positioning, and vehicle emergency braking are taken as an example to define the safe operation domain, as shown in Table 1 below:

[0069] Table 1 Safe operation domain table

[0070]

[0071] S1.2 Data Collection

[0072] The virtual marshaling train control system in this application collects operational data through multiple interfaces, covering multiple systems such as the rolling stock system, signaling system, integrated monitoring system, intelligent operation and maintenance system, meteorological station system, construction management and control system, and clearance obstacle monitoring system. The data set consists of data in various formats, including text information such as operation logs and dispatch timetables, image information such as track conditions and obstacles, and numerical information such as train performance parameters.

[0073] Specifically, the vehicle system provides key operating parameters such as the train's real-time position, speed, acceleration and braking status; the signal system provides signal data such as the status of the on-board ATP / ATO and track occupancy; the integrated monitoring system integrates data from multiple systems such as communications, fire and power to provide comprehensive monitoring of the train operating environment; the intelligent operation and maintenance system is responsible for status monitoring and fault warning of train equipment; the meteorological station system provides real-time weather information, such as wind speed, rainfall and visibility; the construction management and control system records information such as construction plans and construction areas to ensure the safe operation of trains in construction areas; the clearance obstacle monitoring system monitors the obstacles above the track in real time to prevent safety accidents caused by the intrusion of foreign objects; the data collected by the meteorological station system, construction management and control system and clearance obstacle monitoring system are used as environmental perception data.

[0074] S1.3 Multi-source heterogeneous data fusion

[0075] This application integrates data collected from vehicle systems, signal systems, integrated monitoring systems, intelligent operation and maintenance systems, and meteorological observatory systems (i.e., the initial safe operation domain historical status data set) into multi-modal high-dimensional data through spatiotemporal correlation, comprehensively and in real time reflecting the safe operation domain status of virtual train formations. The application includes the following specific steps:

[0076] The first step is data cleaning, which removes outliers and redundant data that may mislead the analysis. These data include duplicate records and irrelevant fields.

[0077] The second step is data conversion, which ensures that data from different systems can be compared on a unified time basis. This includes synchronization of timestamps and standardization of data formats to ensure data integrity and accuracy.

[0078] The third step involves spatial and temporal data correlation. This involves using timestamps to identify similar attributes and data items across different data sources, matching and correlating data. This allows the extraction of key feature vectors from complex data sets, such as train speed change rate, position deviation, and delay error, that are crucial for predicting train safety. These feature vectors not only reflect the train's immediate status during operation but also contain information about potential safety risks.

[0079] Through data fusion technology, the virtual marshaling train control system integrates multi-source data, allowing data from different data sources and in different formats to complement and verify each other, ultimately obtaining more accurate, complete and consistent data, and constructing a safe operation domain status data set (including feature vector sets and label data sets) to realize the monitoring of the safe operation domain of driving.

[0080] S2: Security Situation Assessment

[0081] S2.1 Predicting driving safety trends

[0082] This application uses fused multi-modal high-dimensional data to predict driving safety trends. Specifically, it includes the following steps:

[0083] The safe operation domain status data set is used to train the safe operation domain situation prediction model, and is divided into two parts: training set and test set.

[0084] Among them, the training set plays a core role in the model training stage. Its main purpose is to improve the accuracy and generalization ability of the safe operation domain situation prediction model through repeated iteration and optimization; while the test set is used to comprehensively evaluate the trained safe operation domain situation prediction model to verify its effectiveness and performance in actual applications.

[0085] The safe operation domain situation prediction model can adapt to cross-scale prediction needs. For example, for slow-evolving features such as pantograph wear and track degradation, the model can capture long-term temporal dependencies; and for fast-evolving features such as communication quality fluctuations and icing and disconnection from the grid, a short-term time series identification strategy is constructed to deal with them.

[0086] Real-time data collected about the safe operating domain status is fed into a trained safe operating domain situation prediction model. Based on this input data, the model applies internal algorithms and logic to efficiently calculate future trends in the safe operating domain. This allows for accurate and reliable predictions of trends in the safe operating domain of virtual trains, providing crucial support for train safety warnings and risk mitigation decisions.

[0087] The trained safe operation domain situation prediction model includes an input layer, a feature coding layer, a multi-scale time series extraction layer, an attention fusion layer, a recursive prediction layer and an output layer connected in sequence; specifically, the input layer is responsible for receiving multimodal safe operation domain status data. This layer designs a unified data interface and preprocessing module to ensure that data of different formats and systems can be effectively input into subsequent network layers; the encoder of the feature coding layer adopts a fully connected neural network structure to process feature representation data such as vehicle systems, signal systems, integrated monitoring systems, intelligent operation and maintenance systems and meteorological station systems, and establishes a unified feature coding mechanism for numerical operating parameters, text log information and image monitoring data to obtain feature-encoded time series data; the multi-scale time series extraction layer includes short-term extraction branches, medium-term extraction branches and long-term extraction branches set in parallel: the short-term extraction branch adopts a convolutional neural network to extract sudden equipment failures and communication errors through multiple convolution kernels of different sizes. The mid-term extraction branch adopts a bidirectional long short-term memory network to extract the state evolution characteristics of the safe operation domain with a medium time span; the long-term extraction branch adopts a multi-head self-attention mechanism to establish long-term temporal dependencies and effectively identify long-term safety operation domain change trends such as equipment aging and seasonal environmental changes (i.e., global trend characteristics of the safe operation domain); the attention fusion layer adopts an adaptive weight mechanism to perform weighted fusion on the output features of the three branches to obtain a comprehensive feature vector after weighted fusion; the recursive prediction layer adopts a multi-layer gated recurrent unit network, which includes a reset gate, an update gate and a candidate state module, and realizes the selective memory and forgetting of historical information through the gating mechanism, and models the temporal dependency of the safe operation domain state; the output layer includes multiple prediction heads, which output the prediction results of the safe operation domain change trend of different time spans. The safe operation domain situation prediction model is as follows Figure 4 shown.

[0088] S2.2 Comprehensive assessment of operational risks

[0089] This application proposes a driving risk propagation model based on historical accident data and expert experience.

[0090] Specifically, the driving risk propagation model integrates and analyzes historical accident data, identifies the frequency, type and potential causes of accidents, and combines expert experience to conduct in-depth analysis of the accident propagation path and impact range to form a systematic risk propagation logic. On this basis, the correlation between the monitoring indicators of the safe operation domain and driving accidents is established. According to the probability reasoning algorithm to represent the dependency relationship of all relevant factors of the driving risk propagation model, the driving risk propagation model includes multiple monitoring indicator nodes of the safe operation domain, risk propagation nodes and driving accident nodes. The driving risk propagation model is as follows: Figure 5 shown.

[0091] To ensure the accuracy and reliability of the traffic risk propagation model, a combination of expert experience and training and learning was used to determine model parameters. On the one hand, the model parameters were initially set by drawing on the expertise and practical experience of experts. On the other hand, the model was trained and learned using historical accident data, and the accuracy of the model was improved through continuous parameter optimization.

[0092] After the model is built, the real-time monitored trend of changes in the safe operation domain is input into the driving risk propagation model. Based on this data, the model will infer possible driving risk propagation paths and calculate the probability of driving accidents under the current safe operation domain status conditions. By referring to the empirical values ​​of the severity level of the consequences of driving accidents, the current driving risk level assessment is obtained through comprehensive analysis of these data. The driving risk propagation path consists of a monitoring indicator node of the safe operation domain, a driving accident node, and multiple risk propagation nodes between the monitoring indicator node and the corresponding driving accident node. The node connection relationship is: the monitoring indicator node of the safe operation domain is connected to the corresponding risk propagation node through a directed edge, and there is a mutually influential connection relationship between the risk propagation nodes. The risk propagation node is connected to the corresponding driving accident node through a directed edge; according to the driving risk propagation path, the probability of driving accidents is calculated, specifically:

[0093] According to the formula P(traffic accident|current safe operation domain status)=∑ 所有行车风险传播路径 (∏ 行车风险传播路径上每个节点 P (current node status | parent node status) safe operation domain calculates the probability of traffic accidents;

[0094] Among them, P(traffic accident|current safe operation domain status) represents the probability of a traffic accident under the condition of the current safe operation domain status; ∑[all traffic risk propagation paths] represents the sum of the probability results of all traffic risk propagation paths; ∏[each node on the traffic risk propagation path] represents the multiplication of the conditional probability of each node on the traffic risk propagation path; P(current node status|parent node status) safe operation domain represents the conditional probability of the current node status under the parent node status under the condition of the current safe operation domain status. The current node is any node on the traffic risk propagation path, and the parent node status is the previous node.

[0095] S3: Risk Mitigation Strategy Generation

[0096] The risk mitigation strategy for this application was formulated taking into account factors such as risk type, severity, scope of impact, and available resources.

[0097] Among them, for rapidly evolving hazards (i.e. high risks), such risks change in seconds or minutes and require immediate action, which is usually responded to by on-board control strategies; the on-board control strategy is to issue an alarm to the vehicle or driver and give braking or speed reduction prompts.

[0098] Among them, for hazardous sources that evolve at a medium speed (i.e., medium risk), this type of risk evolves in hours or days. It can be responded to in an orderly manner according to the established plan, and a train scheduling control strategy is usually adopted; the train scheduling control strategy is to adjust the train operation plan or detrain according to the specific type of train risk.

[0099] Among them, for slowly evolving hazards (i.e. low risks), this type of risk gradually changes in units of days, months or even longer periods of time, and can be handled during non-operating hours, usually adopting a maintenance strategy; the maintenance strategy is to take necessary preventive measures based on the specific conditions of the driving risk.

[0100] The virtual marshaling train control system includes multiple pre-set risk mitigation strategies, including onboard control strategies, train scheduling control strategies, and maintenance strategies. These strategies address varying levels of risk by adjusting train speeds, rerouting routes, and initiating emergency braking. Furthermore, the system supports flexible adjustment of strategy parameters based on actual conditions, ensuring the effectiveness and relevance of risk mitigation measures.

[0101] Example 3

[0102] The present application also provides an application scenario, which applies the above-mentioned safe operation domain monitoring method. Specifically: the safe operation domain monitoring method provided in this embodiment can be applied in the risk mitigation scenario of virtual train formation. The risk mitigation scenario of virtual train formation includes a risk mitigation strategy generation link and a risk mitigation strategy execution link; the risk mitigation strategy generation link is used to predict the risk level and generate a risk mitigation strategy based on real-time safe operation domain status data; the risk mitigation strategy execution link is used to send early warning information or assist in decision-making to the dispatching center and the on-board control unit. The safe operation domain monitoring method provided in this embodiment belongs to the risk mitigation strategy generation link.

[0103] Example 4

[0104] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O) and a communication interface. The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store processing data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a safe operation domain monitoring method for a virtual train is implemented.

[0105] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0106] In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0107] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0108] In an example embodiment, a computer program product is provided, including a computer program which, when executed by a processor, implements the steps of the above-mentioned method embodiments.

[0109] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0110] It can be understood by those skilled in the art that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. Any reference to memory, database or other medium used in the embodiments provided by the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0111] The database involved in the embodiments provided by the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on blockchain, etc., without being limited thereto. The processor involved in the embodiments provided by the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0112] Any technical features in the above embodiments can be combined, and for the sake of brevity, not all possible combinations are described above, however, it should be understood that the application encompasses all possible combinations of the technical features described above.

[0113] The principles and implementation manners of the present application are described herein by using specific examples, and the above embodiments are only used to help understand the method of the present application and its core idea; meanwhile, according to the idea of the present application, the specific implementation manners and application scopes will be changed by those skilled in the art. In conclusion, the content of the present specification should not be understood as a limitation of the present application.

Claims

1. A method for monitoring the safe operation domain of a virtual train, characterized in that: include: Obtain security operation domain status data; The safe operation domain status data includes vehicle system related data, signal system data, integrated monitoring system data, intelligent operation and maintenance system data and environmental perception data; Inputting the safe operation domain status data into the trained safe operation domain situation prediction model to obtain a safe operation domain change trend; Inputting the change trend of the safe operation domain into a traffic risk propagation model to obtain the probability of traffic accidents; the traffic risk propagation model is expressed as the correlation between the monitoring indicators of the safe operation domain and traffic accidents; Determining a driving risk level based on the probability of the driving accident and a preset severity level of the consequences of the driving accident; A risk mitigation strategy is generated according to the driving risk level.

2. The method for monitoring the safe operation domain of a virtual train set according to claim 1, characterized in that: The process of determining the trained safe operation domain situation prediction model specifically includes: Constructing a safe operation domain model based on a preset safe operation assumption of a virtual marshaled train; the safe operation domain model includes element categories, monitoring indicators, and thresholds of the safe operation domain; Acquire an initial safe operation domain historical state data set and a corresponding label data set; the initial safe operation domain historical state data set and the corresponding label data set are determined based on the safe operation domain model; Performing multi-source heterogeneous data fusion on the initial safe operation domain historical state data set to obtain a feature vector set; Based on the feature vector set and the label data set, a preset safe operation domain situation prediction model is trained to obtain a trained safe operation domain situation prediction model.

3. The method for monitoring the safe operation domain of a virtual train set according to claim 2, characterized in that: Perform multi-source heterogeneous data fusion on the initial safe operation domain historical status data set to obtain a feature vector set, including: Performing data cleaning on the initial safe operation domain historical state data set to obtain a first safe operation domain historical state data set; the data cleaning includes removing abnormal values ​​and redundant data; Performing data conversion on the first safe operation domain historical status data set to obtain a second safe operation domain historical status data set; the data is converted into a unified timestamp and a standardized data format; Performing spatiotemporal correlation and feature extraction on the second safe operation domain historical status data set to obtain a feature vector set.

4. The method for monitoring the safe operation domain of a virtual train set according to claim 1, characterized in that: The safe operation domain change trend is input into the driving risk propagation model to obtain the probability of driving accidents, which specifically includes: Inputting the change trend of the safe operation domain into the driving risk propagation model to obtain the driving risk propagation path; Calculate the probability of traffic accidents based on the driving risk transmission path.

5. The method for monitoring the safe operation domain of a virtual train set according to claim 4, characterized in that: The driving risk propagation path consists of a monitoring indicator node of a safe operation domain, a driving accident node, and multiple risk propagation nodes between the monitoring indicator node and the corresponding driving accident node. The probability of a driving accident is calculated based on the driving risk propagation path, specifically: According to the formula P(traffic accident|current safe operation domain status)=∑ 所有行车风险传播路径 (∏ 行车风险传播路径上每个节点 P (current node status | parent node status) safe operation domain calculates the probability of traffic accidents; Among them, P(traffic accident|current safe operation domain status) represents the probability of a traffic accident under the condition of the current safe operation domain status; ∑[all traffic risk propagation paths] represents the sum of the probability results of all traffic risk propagation paths; ∏[each node on the traffic risk propagation path] represents the multiplication of the conditional probability of each node on the traffic risk propagation path; P(current node status|parent node status) safe operation domain represents the conditional probability of the current node status under the parent node status under the condition of the current safe operation domain status. The current node is any node on the traffic risk propagation path, and the parent node status is the previous node.

6. The method for monitoring the safe operation domain of a virtual train set according to claim 1, characterized in that: The trained safe operation domain situation prediction model includes an input layer, a feature encoding layer, a multi-scale time series extraction layer, an attention fusion layer, a recursive prediction layer and an output layer connected in sequence; Inputting the safe operation domain status data into the trained safe operation domain situation prediction model to obtain the safe operation domain change trend, specifically including: Input the multi-modal safe operation domain state data into the input layer to obtain the pre-processed safe operation domain state data; Inputting the pre-processed safe operation domain state data into the feature coding layer to obtain feature-coded time series data; Inputting the feature-encoded time series data into a multi-scale time series extraction layer to obtain a multi-scale feature map; the multi-scale feature map includes local mutation features of the safe operation domain, state evolution features of the safe operation domain, and global trend features of the safe operation domain; Inputting the multi-scale feature map into the attention fusion layer to obtain a weighted fused comprehensive feature vector; Inputting the weighted fused comprehensive feature vector into the recursive prediction layer to obtain a hidden state sequence; The hidden state sequence is input into the output layer to obtain the change trend of the safe operation domain.

7. The method for monitoring the safe operation domain of a virtual train set according to claim 1, characterized in that: The risk mitigation strategy includes vehicle control strategy, vehicle dispatch control strategy and maintenance strategy; Based on the driving risk level, a risk mitigation strategy is generated, specifically including: When the driving risk level is high, a vehicle control strategy is generated; the vehicle control strategy is to issue an alarm to the vehicle or the driver and give a braking or speed reduction prompt; When the driving risk level is medium risk, a driving dispatching control strategy is generated; the driving dispatching control strategy is to adjust the train operation plan or separation according to the specific type of driving risk. When the driving risk level is low risk, a maintenance strategy is generated; the maintenance strategy is to take necessary preventive measures based on the specific conditions of the driving risk.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the safe operation domain monitoring method for a virtual train assembly as described in any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for monitoring the safe operation domain of a virtual train set according to any one of claims 1 to 7 is implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method for monitoring the safe operation domain of a virtual train set according to any one of claims 1 to 7 is implemented.

Images