Basic human error probability assessment method, device, equipment, medium and product

Abstract

The invention relates to the technical field of data analysis, in particular to a basic human error probability assessment method, device and equipment, a medium and a product, and the method comprises the steps: obtaining target case information needing to be analyzed; inputting the target case information into a pre-constructed multi-agent analysis framework, and outputting a multi-agent analysis result of the target case information by the multi-agent analysis framework; and according to a pre-constructed knowledge graph database and the multi-agent analysis result, generating values of a plurality of index parameters of the basic human error probability, and according to the knowledge graph database and the values of the plurality of index parameters, evaluating the basic human error probability of the target case information. Therefore, the problems of high dependence on expert knowledge, long evaluation time, low evaluation efficiency and the like in the prior art are solved.

Description

Technical Field

[0001] This application relates to the field of data analysis technology, and particularly relates to a basic human error probability assessment method, device, equipment, medium and product. Background Art

[0002] In the process of safety management and risk assessment in many industries, the occurrence of accidents is often closely related to human factors. Especially in high-risk fields such as nuclear power, aviation, and transportation, the vast majority of accidents can be traced back to the root cause of human error. Human error not only causes direct property losses and casualties, but also has a long-term negative impact on society and the environment. Therefore, how to evaluate and reduce the occurrence probability of human error has become a key issue in improving safety.

[0003] Existing basic human error probability assessment methods generally rely on expert knowledge and use empirical judgment to determine the probability of human error. The main problem with such methods is that they highly rely on the experience of experts, and the assessment process is often cumbersome and time-consuming. Especially in the case of complex systems and multiple factors, the assessment difficulty increases, resulting in low assessment efficiency. Summary of the Invention

[0004] This application provides a basic human error probability assessment method, device, equipment, medium and product to solve the problems in the prior art such as highly relying on expert knowledge, long assessment time, and low assessment efficiency.

[0005] The first aspect embodiment of this application provides a basic human error probability assessment method, including the following steps: obtaining target case information to be analyzed; inputting the target case information into a pre-constructed multi-agent analysis framework, and the multi-agent analysis framework outputs the multi-agent analysis result of the target case information; generating the values of multiple index parameters of the basic human error probability according to the pre-constructed knowledge graph database and the multi-agent analysis result, and evaluating the basic human error probability of the target case information according to the knowledge graph database and the values of the multiple index parameters.

[0006] Optionally, before generating the values of multiple index parameters of the basic human error probability according to the pre-constructed knowledge graph database and the multi-agent analysis result, it further includes: identifying the analysis requirements of the user; determining the type of basic human error probability analysis based on the analysis requirements; determining the target knowledge graph of the knowledge graph database based on the type; generating the values of multiple index parameters of the basic human error probability based on the target knowledge graph and the multi-agent analysis result.

[0007] Optionally, the multi-agent analysis framework includes a first to a fourth agent. Among them, the first agent is used to perform task analysis on the target case information; the second agent is used to perform context analysis on the target case information; the third agent is used to perform cognitive activity analysis on the target case information; the fourth agent is used to perform time constraint analysis on the target case information.

[0008] Optionally, before generating the values of multiple metric parameters of the basic human error probability according to the pre-constructed knowledge graph database and the multi-agent analysis results, it further includes: obtaining data related to the basic human error probability in the target database; using the target graph database management system and the relevant data to construct the knowledge graph database.

[0009] Optionally, the multi-agent analysis results include task analysis results, context analysis results, cognitive activity analysis results, and time constraint analysis results. The metric parameters include human factor performance influencing factors, cognitive error patterns, tasks, performance influencing factor variables, and other performance influencing factors. The knowledge graph database includes scene familiarity knowledge graphs, information availability and reliability knowledge graphs, and task complexity knowledge graphs.

[0010] Optionally, the process of task analysis includes at least one of: task overview, task classification, analyzing the goals of the task, checking error types and impacts, and determining the complexity of the task; the process of context analysis includes at least one of: identifying the background conditions where the task occurs, analyzing the support required for task execution, clarifying the initial conditions and requirements for task initiation, and checking error metric data; the process of cognitive activity analysis includes at least one of: checking the cognitive requirements for task execution, and understanding the psychological processes behind task performance; the process of time constraint analysis includes at least one of: analyzing the data sources and analyzing the time constraints.

[0011] An embodiment of the second aspect of this application provides a basic human error probability assessment device, including: an acquisition module, configured to acquire target case information that needs to be analyzed; an input module, configured to input the target case information into a pre-constructed multi-agent analysis framework, and the multi-agent analysis framework outputs the multi-agent analysis results of the target case information; an evaluation module, configured to generate the values of multiple metric parameters of the basic human error probability according to the pre-constructed knowledge graph database and the multi-agent analysis results, and evaluate the basic human error probability of the target case information according to the knowledge graph database and the values of the multiple metric parameters.

[0012] Optionally, it further includes: an identification module, configured to identify the analysis requirements of the user before generating the values of multiple metric parameters of the basic human error probability based on the pre-constructed knowledge graph database and the multi-agent analysis results; determine the type of basic human error probability analysis based on the analysis requirements; determine the target knowledge graph of the knowledge graph database based on the type; and generate the values of multiple metric parameters of the basic human error probability based on the target knowledge graph and the multi-agent analysis results.

[0013] Optionally, the multi-agent analysis framework includes the first to fourth agents. Among them, the first agent is configured to perform task analysis on the target case information; the second agent is configured to perform context analysis on the target case information; the third agent is configured to perform cognitive activity analysis on the target case information; and the fourth agent is configured to perform time constraint analysis on the target case information.

[0014] Optionally, it further includes: a construction module, configured to obtain the data related to the basic human error probability in the target database before generating the values of multiple metric parameters of the basic human error probability based on the pre-constructed knowledge graph database and the multi-agent analysis results; and construct a knowledge graph database by using the target graph database management system and the related data.

[0015] Optionally, the multi-agent analysis results include task analysis results, context analysis results, cognitive activity analysis results, and time constraint analysis results. The metric parameters include human performance influencing factors, cognitive error modes, tasks, performance influencing factor variables, and other performance influencing factors. The knowledge graph database includes a scenario familiarity knowledge graph, an information availability and reliability knowledge graph, and a task complexity knowledge graph.

[0016] Optionally, the process of task analysis includes at least one of: task overview, task classification, analyzing the goals of the task, checking error types and impacts, and determining the complexity of the task; the process of context analysis includes at least one of: identifying the background conditions where the task occurs, analyzing the support required for task execution, clarifying the initial conditions and requirements for task initiation, and checking error metric data; the process of cognitive activity analysis includes at least one of: checking the cognitive requirements for task execution and understanding the psychological processes behind task performance; the process of time constraint analysis includes at least one of: analyzing the data source and analyzing time constraints.

[0017] An embodiment of the third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor. The processor executes the program to perform the basic human error probability assessment method as described in the above embodiments.

[0018] The fourth aspect of the embodiments of the present application provides a computer-readable storage medium, on which a computer program or instruction is stored. The computer program or instruction is executed by a processor to execute the basic human error probability assessment method as described in the above embodiments.

[0019] The fifth aspect of the embodiments of the present application provides a computer program product, including a computer program or instruction. When the computer program or instruction is executed, it realizes the basic human error probability assessment method as described in the above embodiments.

[0020] Therefore, the present application has at least the following beneficial effects:

[0021] In the embodiments of the present application, the target case information to be analyzed can be input into a pre-constructed multi-agent analysis framework. The multi-agent analysis framework inputs the multi-agent analysis results of the target case information, combines it with the pre-constructed knowledge graph database to generate the values of multiple index parameters of the basic human error probability, and then evaluates the basic human error probability of the target case information according to the knowledge graph database and the values of the multiple index parameters, so as to help industry personnel more efficiently quantify and evaluate the probability of human error, thereby improving the work efficiency of safety risk assessment, and further providing strong decision-making support for improving human reliability. Moreover, it does not rely too much on expert knowledge, can quickly generate the basic human error probability through simple text description, reduces the evaluation time-consuming, and improves the evaluation efficiency. Thus, the technical problems of the prior art, such as high dependence on expert knowledge, long evaluation time, and low evaluation efficiency, are solved.

[0022] The additional aspects and advantages of the present application will be partially given in the following description, partially become obvious from the following description, or be understood through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The above and / or additional aspects and advantages of the present application will become obvious and easy to understand from the following description of the embodiments in conjunction with the drawings, where:

[0024] Figure 1 is a flowchart of the basic human error probability assessment method according to the embodiments of the present application;

[0025] Figure 2 is an example diagram of the basic human error probability assessment device according to the embodiments of the present application;

[0026] Figure 3 is a schematic structural diagram of an electronic device according to the embodiments of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] Embodiments of the present application will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present application, but should not be construed as limiting the present application.

[0028] The following describes a basic human error probability assessment method, device, equipment, medium, and product according to embodiments of the present application with reference to the accompanying drawings. In view of the problem in the above-mentioned background technology that existing basic human error probability assessment methods generally rely on expert knowledge and use empirical judgments to determine the probability of human error. The main problems of such methods are as follows: on the one hand, it highly depends on the experience of experts and lacks sufficient objectivity and comprehensiveness; on the other hand, the assessment process is often cumbersome and time-consuming, especially in the case of complex systems and multiple factors, the assessment difficulty increases, and it is difficult to achieve accuracy and efficiency. The present application provides a basic human error probability assessment method. In this method, the target case information to be analyzed can be input into a pre-constructed multi-agent analysis framework. The multi-agent analysis framework inputs the multi-agent analysis results of the target case information and combines them with a pre-constructed knowledge graph database to generate the values of multiple index parameters of the basic human error probability. Then, based on the knowledge graph database and the values of the multiple index parameters, the basic human error probability of the target case information is evaluated, so as to help industry personnel more efficiently quantify and evaluate the probability of human error, thereby improving the work efficiency of safety risk assessment, and further providing strong decision-making support for improving human reliability, and without relying too much on expert knowledge, quickly generating the basic human error probability through simple text descriptions, reducing the assessment time duration, and improving the assessment efficiency. Thus, the problems of the prior art such as highly relying on expert knowledge, long assessment time, and low assessment efficiency are solved.

[0029] Specifically, Figure 1 is a schematic flowchart of a basic human error probability assessment method provided by an embodiment of the present application.

[0030] As Figure 1 shown, the basic human error probability assessment method includes the following steps:

[0031] In step S101, obtain the target case information to be analyzed.

[0032] In step S102, input the target case information into a pre-constructed multi-agent analysis framework, and the multi-agent analysis framework outputs the multi-agent analysis results of the target case information.

[0033] Among them, the multi-agent analysis results include task analysis results, context analysis results, cognitive activity analysis results, and time constraint analysis results.

[0034] It can be understood that in the embodiments of the present application, the target case information can be input into a pre-constructed multi-agent analysis framework, and the multi-agent analysis framework outputs the multi-agent analysis results of the target case information.

[0035] In the embodiments of the present application, the multi-agent analysis framework includes the first to fourth agents. Among them, the first agent is used to perform task analysis on the target case information; the second agent is used to perform context analysis on the target case information; the third agent is used to perform cognitive activity analysis on the target case information; the fourth agent is used to perform time constraint analysis on the target case information.

[0036] It can be understood that the multi-agent analysis framework of the embodiments of the present application includes the first to fourth agents, that is, the multi-agent analysis framework deploys four large language model-driven agents, including the first agent, the second agent, the third agent, and the fourth agent. These agents are designed to further refine and analyze the target case information provided by the user, and more comprehensively reflect the human error process in the case. Among them,

[0037] The first agent is used to perform task analysis on the target case information; the second agent is used to perform context analysis on the target case information; the third agent is used to perform cognitive activity analysis on the target case information; the fourth agent is used to perform time constraint analysis on the target case information.

[0038] Specifically:

[0039] First, the main task of the first agent is to analyze the data source and identify the tasks related to the reported human error information. This task analysis includes several key components. First, conduct a task overview to outline the general process of the task. Second, perform task classification to classify the tasks according to their functions or types, so as to better understand the nature and background of the tasks. Third, analyze the goals of the tasks to clarify the specific goals of each task. Fourth, check the typical error types and their impacts, identify the common errors in the tasks and evaluate their potential consequences. Finally, determine the complexity of the tasks to provide in-depth insights into the task difficulty, covering tasks from simple to highly complex. The specific prompt engineering content is as follows:

[0040] Analyze the data source and identify the tasks reporting human error information. The following are specific guiding suggestions:

[0041] 1. Outline the task process.

[0042] 2. Classify the tasks according to their functions or types, which helps to understand the nature and background of the tasks. For example:

[0043] Operation tasks: such as "transport fuel components using a fuel machine".

[0044] Monitoring tasks: e.g., "indicator check".

[0045] Diagnostic tasks: e.g., "diagnosis that requires interpreting a large number of indications and alarms".

[0046] Decision-making tasks: e.g., "choosing the wrong strategy".

[0047] Communication tasks: e.g., "pilot repeating communication" and "communication in nuclear facility operation".

[0048] Training tasks: e.g., "students performing tests on relational working memory".

[0049] 3. Specific objectives of the analysis tasks, such as:

[0050] Operation tasks: The objective is to complete a certain physical or logical action (such as transportation, executing a program).

[0051] Monitoring tasks: The objective is to correctly identify signals or states (such as alarm detection, indicator check).

[0052] Diagnostic tasks: The objective is to correctly understand complex or ambiguous information (such as interpreting alarms and indications).

[0053] Decision-making tasks: The objective is to select the correct strategy or solution.

[0054] Communication tasks: The objective is to accurately transmit or receive information.

[0055] Training tasks: The objective is to improve capabilities or evaluate performance.

[0056] 4. Typical error types and their impacts in analysis tasks, such as:

[0057] Omission errors: e.g., "nuclear power plant maintenance (omitting one instruction)".

[0058] Recognition errors: e.g., "highly experienced driver simulating driving (missed detection rate of peripheral detection tasks)".

[0059] Execution errors: e.g., "delayed execution (wrong assessment of margin)".

[0060] Communication errors: e.g., "pilot repeating communication (error rate of misreporting messages)".

[0061] 5. Determine the complexity level of the tasks, including:

[0062] Simple tasks: e.g., "reading a meter" and "signal detection".

[0063] Complex tasks: e.g., "diagnosis that requires interpreting a large number of indications and alarms" and "controlled operations that require monitoring the results of actions and adjusting actions accordingly".

[0064] Answer my question in Chinese. The content to be analyzed is: {data_scource}, where data_scource is the case information to be analyzed.

[0065] Second, the core task of the second agent is to analyze the data source and the context related to task execution. Specifically, this includes four key tasks: First, identify the background conditions under which the task occurs; second, analyze the support required for task execution; third, clarify the initial conditions and requirements for task initiation; and finally, examine the error metric data to evaluate the task-related performance and results. The specific prompt engineering content is as follows:

[0066] Analyze the data source and the context. The following are specific guiding suggestions:

[0067] 1. Identify the background conditions under which the task occurs, such as:

[0068] Whether it is completed in a simulation environment (e.g., "Nuclear power plant operators perform emergency operating procedures (EOPs) on a simulator").

[0069] Whether it involves dynamic changes or complex situations (e.g., "Operating a controller while monitoring a dynamic display").

[0070] Whether it is an emergency task in a special situation (e.g., "Nuclear power plant staff perform a series of tasks for a steam generator tube rupture (SGTR) event").

[0071] 2. Analyze the support required for task execution:

[0072] Tools and equipment: Such as whether instruments, control panels, or simulators are required.

[0073] Procedures and guidance: Whether there are clear operating procedures or scripts.

[0074] Team collaboration: Such as whether communication or coordination with others is required for the task.

[0075] 3. Clarify the situation and requirements for task initiation:

[0076] Regular trigger: For example, periodic tasks or routine maintenance tasks.

[0077] Abnormal trigger: For example, emergency tasks when a fault occurs (e.g., "Nuclear power plant staff perform a series of tasks for a steam generator tube rupture (SGTR) event").

[0078] Situation dependence: For example, tasks in a dynamic display or multi-task environment.

[0079] 4. Analyze the error metric data, such as:

[0080] Error rate range: The error rate is specified for certain tasks (e.g., 5% to 50% in "Pilot repeats communication").

[0081] Severity of errors: Some errors may lead to system failures or catastrophic consequences.

[0082] Answer my question in Chinese. The content to be analyzed is: {data_scource}.

[0083] III. The main task of the third agent is to analyze the data source, describe the task, and identify the specific cognitive activities involved. This includes examining the cognitive requirements for task execution and understanding the mental processes behind task performance. The specific prompt engineering content is as follows:

[0084] Analyze the data source, describe the task to identify the cognitive activities involved in the task. For example:

[0085] 1. Experience: Does the task require a skilled operator (e.g., "Military operator reads a meter").

[0086] 2. Cognitive abilities: Such as memory ability (e.g., "Student performs a test on relational working memory") or perceptual ability (e.g., "Detecting signals in nuclear facility operations").

[0087] 3. Decision-making ability: Such as the need to evaluate ambiguous information (e.g., "Diagnosis that requires interpreting a large number of instructions and alarms").

[0088] The content to be analyzed is: {data_scource}.

[0089] IV. The main task of the fourth agent is to analyze the data source and evaluate the time constraints related to task execution. This involves identifying time limits, deadlines, or time-sensitive conditions that affect task execution and outcomes. The specific prompt engineering content is as follows:

[0090] Analyze the data source, analyze the time constraints. The content to be analyzed is: {data_scource}.

[0091] In the embodiments of the present application, the process of task analysis includes at least one of: task overview, task classification, analyzing the goal of the task, checking the error type and its impact, determining the complexity of the task; the process of context analysis includes at least one of: identifying the background conditions where the task occurs, analyzing the support required for task execution, clarifying the initial conditions and requirements for task initiation, checking the error metric data; the process of cognitive activity analysis includes at least one of: examining the cognitive requirements for task execution, understanding the mental processes behind task performance; the process of time constraint analysis includes at least one of: analyzing the data source, analyzing the time constraints.

[0092] It can be understood that the process of the four agents in the embodiments of the present application analyzing the target case information is as follows: The process of the first agent performing task analysis includes at least one of: task overview, task classification, analyzing the goal of the task, checking the error type and impact, and determining the complexity of the task; The process of the second agent performing context analysis includes at least one of: identifying the background conditions under which the task occurs, analyzing the support required for task execution, clarifying the initial conditions and requirements for task initiation, and checking the error metric data; The process of the third agent performing cognitive activity analysis includes at least one of: checking the cognitive requirements for task execution, understanding the mental processes behind task performance; The process of the fourth agent performing time constraint analysis includes at least one of: analyzing the data source, analyzing the time constraint.

[0093] The following uses an embodiment to describe the process and results of the multi-agent analysis framework for analysis.

[0094] The case information is as follows: The task requires the pilot to listen to and repeat the instructions of the air traffic controller. The pilot stores the information in memory and repeats it at the end of the transmission, and the above content is input into the pre-constructed multi-agent analysis framework.

[0095] I. For the first agent, the output results of the task analysis are as follows:

[0096] 1. Task overview

[0097] The task requires the pilot to listen to the instructions of the air traffic controller and accurately repeat these instructions after the information transmission is completed. This involves storing the information in memory and ensuring the accurate reproduction of key information.

[0098] 2. Task classification

[0099] The task can be divided into two categories: a communication task, which involves repeating the instructions of the controller; a memory task, which requires short-term memory of the transmitted information. The main goals of the task are to ensure communication accuracy - transmitting key information without omission or misunderstanding; and to ensure memory reliability - ensuring that no information is lost before repetition.

[0100] 3. Common errors and their impacts

[0101] Two common error types have been identified: communication errors, i.e., key information is repeated incorrectly or omitted; memory errors, i.e., the pilot fails to recall all key information. Both of these errors can affect flight safety, resulting in incorrect operation instructions or incomplete actions.

[0102] 4. Task complexity

[0103] The task is classified as moderately complex.

[0104] 5. Difficulty of the task

[0105] It is mainly affected by the key information volume and the transmission methods (listening, memorizing, and repeating). As the complexity of the information increases, the difficulty of the task also increases.

[0106] II. For the second agent, the output results of the context analysis are as follows:

[0107] 1. Task background

[0108] This task is executed in a real operating environment rather than a simulated environment. The task occurs in a static background, that is, there is no dynamic change involved. However, the complexity and quantity of the information may vary. There is no time limit for this task, allowing the pilot to have sufficient time to accurately repeat the information.

[0109] 2. Support conditions required for task execution

[0110] The equipment required for this task is extremely simple, only relying on wireless communication devices. No other additional instruments or tools are mentioned. Since the repetition of instructions mainly depends on the pilot's experience, there are no clear operating procedures or guidelines. The task is executed independently without the need for teamwork or external verification.

[0111] 3. Task triggering conditions

[0112] This task is usually triggered by a communication request from an air traffic controller and is part of routine flight operations. The execution of the task is affected by the complexity of the communication instructions and the information transmission method.

[0113] 4. Error metric data

[0114] The error rate is not clearly reported, but repetition errors may occur under different information complexities. Although the severity of the errors is not quantified, such errors may affect the accuracy of flight instructions and pose potential risks to operational safety.

[0115] III. For the third agent, the output results of the cognitive activity analysis are as follows:

[0116] 1. Perceptual ability

[0117] It is required to have the ability to listen to and identify key information. One must understand the logical relationships in the information and the intentions behind them.

[0118] 2. Memory capacity

[0119] It is crucial to have the ability to hold multiple instructions in short-term memory. The information must be repeated without omission or distortion.

[0120] 3. Communication skills

[0121] The ability to accurately repeat information is crucial. The integrity and accuracy of key information must be maintained.

[0122] 4. Cognitive Load

[0123] When the information complexity is high, the cognitive load may increase. This phenomenon is particularly obvious, especially in the case of multitasking.

[0124] IV. For the fourth agent, the output results of the time constraint analysis are as follows:

[0125] 1. Time Constraint

[0126] This task has no time limit, allowing the pilot to have enough time to complete the repetition task. The time factor will not directly affect the error rate.

[0127] 2. Potential Impact

[0128] Although there is plenty of time, the high information complexity may indirectly increase the difficulty of the task.

[0129] In the embodiment of the present application, the multi-agent analysis results include that the index parameters include human factor performance influencing factors, cognitive failure modes, tasks, performance influencing factor variables, and other performance influencing factors, and the knowledge graph database includes scene familiarity knowledge graphs, information availability and reliability knowledge graphs, and task complexity knowledge graphs.

[0130] In step S103, based on the pre-constructed knowledge graph database and the multi-agent analysis results, the values of multiple index parameters of the basic human factor error probability are generated, and the basic human factor error probability of the target case information is evaluated according to the knowledge graph database and the values of the multiple index parameters.

[0131] Among them, the index parameters include PIF (Performance Influencing Factors, human factor performance influencing factors), CFM (Cognitive Failure Mode, cognitive failure mode), Task (and error measure), PIF Measure (performance influencing factor variable), and Other PIFs (and Uncertainty) (other performance influencing factors (and uncertainty)).

[0132] It can be understood that the embodiments of the present application can generate the values of multiple index parameters of the basic human error probability according to the pre-constructed knowledge graph database and the multi-agent analysis results, and evaluate the basic human error probability of the target case information according to the knowledge graph database and the values of the multiple index parameters, so as to help industry personnel more efficiently quantify and evaluate the probability of human error, thereby improving the work efficiency of safety risk assessment, and further providing strong decision-making support for improving human reliability.

[0133] In the embodiments of the present application, before generating the values of multiple index parameters of the basic human error probability according to the pre-constructed knowledge graph database and the multi-agent analysis results, it further includes: obtaining the data related to the basic human error probability in the target database; constructing a knowledge graph database by using the target graph database management system and the related data.

[0134] Among them, the target database is some publicly available human reliability databases at present, such as the human reliability database (Integrated Human Event Analysis System for Human ReliabilityData, IDHEAS-DATA) publicly available in the United States in 2021; the target graph data management system can be selected according to specific circumstances, such as Neo4j.

[0135] It can be understood that the embodiments of the present application can obtain the data related to the basic human error probability in the target database, and then construct a knowledge graph database by using the target graph database management system and the related data.

[0136] Taking the knowledge graph database built based on Neo4j and IDHEAS-DATA as an example below.

[0137] The data related to the basic human reliability is divided into three situations: scene familiarity (the specific content is Table 1 and Table 2), information availability and reliability (the specific content is Table 3 and Table 4), task complexity (the specific content is Table 5 and Table 6). The type of specific analysis of the basic human error probability is specified by the user. Among them, Table 1 is the attribute identifier and description of the performance impact factor - scene familiarity, Table 2 is the basic human error probability of the performance impact factor scene familiarity, Table 3 is the attribute identifier and description of the performance impact factor - information availability and reliability, Table 4 is the basic human error probability of information availability and reliability, Table 5 is the attribute identifier and description of the performance impact factor - task complexity, and Table 6 is the basic human error probability of task complexity.

[0138] Table 1

[0139] ID PIF Attribute SF0 No-impact·frequently performed tasks in well-trained scenarios,·routine tasks SF1 Unpredictable dynamics in known scenarios SF1.1 Shifting objectives ... ... SF4.3 Preference for wrong strategies in decision making

[0140] Table 2

[0141]

[0142]

[0143] Table 3

[0144]

[0145] Table 4

[0146]

[0147]

[0148] Table 5

[0149]

[0150] Table 6

[0151]

[0152]

[0153] In the embodiment of the present application, before generating the values of multiple index parameters of the basic human error probability according to the pre-constructed knowledge graph database and the multi-agent analysis results, the following steps are further included: identifying the analysis requirements of the user; determining the type of basic human error probability analysis based on the analysis requirements; determining the target knowledge graph of the knowledge graph database based on the type; and generating the values of multiple index parameters of the basic human error probability based on the target knowledge graph and the multi-agent analysis results.

[0154] Among them, the analysis types include scene familiarity, information availability and reliability, task complexity, etc., and the knowledge graph database includes a scene familiarity knowledge graph, an information availability and reliability knowledge graph, and a task complexity knowledge graph.

[0155] It can be understood that the embodiment of the present application can identify the analysis requirements of the user, determine the type of basic human error probability analysis based on the analysis requirements, and then determine the target knowledge graph of the knowledge graph database based on the type, and generate the values of multiple index parameters of the basic human error probability based on the target knowledge graph and the multi-agent analysis results.

[0156] For example, if the type of basic human error probability analysis selected by the user is "scene database", the knowledge graph database will automatically switch to the nodes and information corresponding to scene familiarity, that is, the scene familiarity knowledge graph.

[0157] Specifically, based on the analysis results output by the above knowledge graph data path and multi-agent analysis framework, the embodiments of the present application generate the values of each parameter of the basic human error probability based on the prompt engineering of the large language model. Specifically, the user needs to select the type of basic human error probability analysis. Suppose the user selects "task complexity" among the three options: scenario familiarity, information availability and reliability, and task complexity. Thereby, the knowledge graph automatically switches to the nodes and information corresponding to "task complexity". Then, combining the knowledge of the knowledge graph and the results output of the above multi-agent analysis framework, the values of each parameter of the basic human error probability are generated using the prompt engineering of the large language model. These parameters include Performance Influencing Factors (PIF), Cognitive Failure Mode (CFM), Task (and error measure), PIF Measure, and Other PIFs (and Uncertainty).

[0158] I. The prompt engineering content for Task (and error measure) is as follows:

[0159] Now, please analyze that the human error rates of these tasks have been reported in the data source, along with the definitions of the human errors measured in these tasks. The following is the reference information for analysis:

[0160] Knowledge base:

[0161] Task information: scen_task;

[0162] Task context information: scen_context;

[0163] Task-related cognitive activity information: scen_cog;

[0164] Task time limit information: scen_time;

[0165] Please write the analysis process of selecting the most suitable output Task (and error measure) from {task_lib} according to the following information requirements, and the final result is output between <>. Answer my question in English and strictly follow the required format for output.

[0166] Among them, scen_task, scen_context, scen_cog, and scen_time are the output results of the above task analysis, context analysis, cognitive activity analysis, and time constraint analysis agents respectively, and task_lib is the Task(and error measure) attribute information of the corresponding knowledge graph in the "task complexity". After extraction, the Task(and error measure) is "Pilots listen to and read back key messages".

[0167] II. The prompt engineering content for the Cognitive Failure Mode (CFM) is as follows:

[0168] Now, please analyze the CFMs (Cognitive Failure Modes). The CFMs are labeled as D, U, DM, E, and T, representing Detection Failure, Understanding Failure, Decision-making Failure, Action Execution Failure, and Interteam Coordination Failure respectively. Note that a task may involve multiple applicable CFMs.

[0169] D (Detection): Failure to correctly detect or identify the required information (such as alarms, instrument readings).

[0170] U (Understanding): Failure to correctly understand or interpret information (such as misunderstanding procedures, misinterpreting system status)

[0171] DM (Decision-making): Failure to make the correct decision (such as choosing an inappropriate strategy).

[0172] E (Action Execution): Failure to correctly execute the required action (such as pressing the wrong button, operation delay).

[0173] T (Interteam Coordination): Failure in communication or collaboration between teams (such as incorrect information transmission or poor coordination).

[0174] If indistinguishable CFMs are reported in the task completion time, the output is Unsp, indicating unspecified CFMs. If the process involves multiple CFMs, output according to logic, with "&" representing logical AND and " / " representing logical OR. For example, if both D and U exist in the process, the output is "<D&U>".

[0175] The following is the reference information for analysis,

[0176] Knowledge base:

[0177] Task information: scen_task;

[0178] Task context information: scen_context;

[0179] Cognitive activity information involved in the task: scen_cog;

[0180] Task time limit information: scen_time;

[0181] Task(and error measure): Pilots listen to and read back key messages.

[0182] Please select the most suitable output CFMs from {CFMs_lib}, output the analysis process, and the final result is output between <>. Answer my question in English and strictly follow the required format for output.

[0183] Among them, scen_task, scen_context, scen_cog, and scen_time are the output results of the above task analysis, context analysis, cognitive activity analysis, and time constraint analysis agents respectively. The part of Task(and error measure): Pilots listen to and read back key messages is the Task(and error measure) information determined in the previous step. CFMs_lib is the CFMs attribute information of the knowledge graph corresponding to "task complexity" where Task(and error measure) is "Pilots listen to and read back key messages". After extraction, the result of CFMs is "U".

[0184] III. The prompt engineering content for human factor performance influencing factors is as follows:

[0185] Now please analyze the basic PIF attributes. The following is the reference information for analysis,

[0186] Task information: scen_task;

[0187] Task context information: scen_context;

[0188] Cognitive activity information involved in the task: scen_cog;

[0189] Task time limit information: scen_time;

[0190] Task (and error measure): Pilots listen to and read back key messages;

[0191] CFMs: U (Understanding).

[0192] Please select the most suitable output PIF attribute from PIF_Attribute_lib, output the analysis process, and the final result is output between <>. Answer my question in English and strictly follow the required format for output.

[0193] Among them, scen_task, scen_context, scen_cog, and scen_time are the output results of the above task analysis, context analysis, cognitive activity analysis, and time constraint analysis agents respectively. The parts of Task (and error measure): Pilots listen to and read back key messages and CFMs: U (Understanding) are the already determined Task (and error measure) and CFMs information. PIF_Attribute_lib is the PIF attribute information of the knowledge graph corresponding to "task complexity" where Task (and error measure) is "Pilots listen to and read back key messages" and CFMs is "U (Understanding)". After extraction, the output result gives the PIF as "C11".

[0194] IV. The prompt engineering content for Performance Impact Factor Measure (PIF Measure) is as follows:

[0195] Now please analyze the PIF attribute measure - in the data source, it is used to describe a task-specific factor or variable during task execution, and this factor or variable is related to the measurement of the human error rate. The following is the reference information for analysis,

[0196] Task information: scen_task;

[0197] Task context information: scen_context;

[0198] Cognitive activity information involved in the task: scen_cog;

[0199] Task time limit information: scen_time;

[0200] Task (and error measure): Pilots listen to and read back key messages;

[0201] CFMs: U (Understanding);

[0202] PIF: C11.

[0203] Please select the most suitable output PIF attribute measure from {PIF_Measure_lib}, output the analysis process, and the final result is output between <>. Answer my question in English and strictly follow the required format for output.

[0204] Among them, the parts of Task (and error measure): Pilots listen to and read back key messages, CFMs: U (Understanding), and PIF: C11 are the determined Task (and error measure) and CFMs information. PIF_Attribute_lib is the PIF Measure attribute information of the knowledge graph corresponding to "task complexity" where Task (and error measure) is "Pilots listen to and read back key messages", CFMs is "U (Understanding)", and PIF is "C11". After extraction, the output result shows that the PlF measure is "Message complexity of key messages in one transmission"

[0205] V. Tips for engineering on other performance impact factors (and uncertainties) Other PIFs (and Uncertainty) are as follows:

[0206] Now, please analyze other PIFs (and uncertainties). In addition to the PIF attributes in the study, the task background in the data source may also contain other PIF attributes that may exist during task execution, and thus they can affect the reported human error rate. This section records the other PIF attributes that exist, particularly recording whether the task is executed under time constraints. Information about time availability is very important for inferring the underlying HEP (human error probability) from the reported human error data. If the available time is insufficient, the reported human error rate corresponds to the sum of the probabilities of the underlying HEP and the error probability (P,) due to insufficient time. It also records the uncertainties in the data source and the uncertainties when mapping to CFM and PIF attributes. These uncertainties will affect how the reported error rates are integrated to inform the underlying HEP.

[0207] There are uncertainties in the data source and when mapping to IDHEAS-G CFM. In particular, if the number of task executions is not large enough, the reported error rate may not represent the lowest HEP.

[0208] Task information: scen_task;

[0209] Task context information: scen_context;

[0210] Information on cognitive activities involved in the task: scen_cog;

[0211] Task time limit information: scen_time;

[0212] Task (and error measure): Pilots listen to and read back key messages;

[0213] CFMs: U (Understanding);

[0214] PIF: C11;

[0215] PIF Measure: Message complexity of key messages in one transmission.

[0216] Please write according to the following information requirements: Please select the most suitable output OtherPIFs (and Uncertainty) from {Other_pif_lib}, where the content in () is the uncertainty content, output the analysis process, and finally output the result between <>. Answer my question in English and strictly follow the required format.

[0217] Among them, for Task (and error measure): Pilots listen to and read back key messages, CFMs: U (Understanding), PIF: C11, and PIF Measure: Message complexity of key messages in one transmission, the parts of Task (and error measure), CFMs information PIF_Attribute_lib are the determined Task (and error measure) and CFMs information. The PIF Measure attribute information of the knowledge graph corresponding to "task complexity" with Task (and error measure) being "Pilots listen to and read back key messages", CFMs being "U (Understanding)", PIF being "C11", and PIF Measure being "Message complexity of key messages in one transmission". After extraction of the output result, Other PlFs (and Uncertainty) are obtained as (Mixture of normal and emergent operations other PlF attributes may exist).

[0218] After obtaining the values of each metric parameter, they are input into the knowledge graph database for search. After the search, it is found that there is only one node in the knowledge graph. After querying with Neo4j, among the attributes of this node, the Error Rate is "M5 = 0.036, M8 = 0.05, M11 = 0.11, M15 = 0.23, M17 = 0.32, M>20 = 0.5", and thus the basic human error probability of this case is obtained.

[0219] The following describes the basic human error probability assessment method of the present application through an embodiment, including:

[0220] 1. Construct a multi-agent analysis framework.

[0221] Constructing a multi-agent analysis framework to disassemble the case information and more comprehensively reflect the human error process in the case. This method deploys four large language model-driven agents to perform different tasks: task analysis, context analysis, cognitive activity analysis, and time constraint analysis. These agents are designed to further refine and analyze the case information provided by the user.

[0222] 2. Construct a knowledge graph database for basic human error probability analysis.

[0223] Collect data on basic human error probability from the Human Reliability Database publicly available in the United States in 2021 (Integrated Human Event Analysis System for Human Reliability Data, IDHEAS-DATA), and combine with Neo4j to build a knowledge graph database for basic human error probability.

[0224] 3. Input information and generate the multi-agent analysis results of the described case.

[0225] The user inputs the case information to be analyzed into the multi-agent framework, and outputs the task analysis, context analysis, cognitive activity analysis, and time constraint analysis results related to the task.

[0226] 4. Combine the output of the knowledge graph and multi-agent results, and generate the values of each parameter of the basic human error probability based on the large language model.

[0227] Combine the knowledge graph database in step 2 and the output results of the multi-agent framework in step 3, and generate the values of each parameter of the basic human error probability based on the large language model. Specifically, these parameters include Performance Influencing Factors (PIF), Cognitive Failure Mode (CFM), Task (and error measure), PIF Measure, Other PIFs (and Uncertainty).

[0228] 5. Combine the knowledge graph search and output the final basic human error probability value.

[0229] Input the values of each parameter obtained in step 4 into the knowledge graph database for search to obtain the final basic human failure probability, and conduct a quantitative risk assessment of the described case.

[0230] In summary, a basic human error probability assessment method based on a multi-agent framework and a knowledge graph according to an embodiment of the present application combines large language model technology, knowledge graph technology, and a multi-agent framework. First, a multi-agent analysis framework is constructed. By disassembling case information through a multi-agent system, it can more comprehensively reflect the process of human error in the case. Next, a knowledge graph database for basic human error probability analysis is constructed. The knowledge graph can systematically store the experience of domain experts, relevant data, and standards and specifications, and provide structured knowledge support for analysis. Then, the user inputs the corresponding case information, and the multi-agent analysis result of the described case is generated through an automated process. Next, in combination with the knowledge graph and the large language model, the values of each parameter of the basic human error probability are generated. Finally, by combining the search function of the knowledge graph, the final value of the basic human error probability is output, so as to provide an efficient and accurate quantitative result for human reliability assessment.

[0231] According to the basic human error probability assessment method proposed in the embodiment of the present application, the target case information to be analyzed can be input into a pre-constructed multi-agent analysis framework. The multi-agent analysis framework inputs the multi-agent analysis result of the target case information, and combines it with a pre-constructed knowledge graph database to generate the values of multiple index parameters of the basic human error probability. Furthermore, based on the knowledge graph database and the values of multiple index parameters, the basic human error probability of the target case information is evaluated, so as to help industry personnel more efficiently quantify and evaluate the probability of human error, thereby improving the work efficiency of safety risk assessment, and further providing strong decision-making support for improving human reliability. Moreover, it does not rely too much on expert knowledge, and can quickly generate the basic human error probability through simple text description, reducing the evaluation time and improving the evaluation efficiency.

[0232] Secondly, a basic human error probability assessment device according to an embodiment of the present application is described with reference to the accompanying drawings.

[0233] Figure 2 It is a block diagram of a basic human error probability assessment device according to an embodiment of the present application.

[0234] As Figure 2 shown, the basic human error probability assessment device 10 includes: an acquisition module 100, an input module 200, and an evaluation module 300.

[0235] Among them, the acquisition module 100 is used to acquire the target case information to be analyzed; the input module 200 is used to input the target case information into a pre-constructed multi-agent analysis framework, and the multi-agent analysis framework outputs the multi-agent analysis results of the target case information; the evaluation module 300 is used to generate the values of multiple index parameters of the basic human error probability according to the pre-constructed knowledge graph database and the multi-agent analysis results, and evaluate the basic human error probability of the target case information according to the knowledge graph database and the values of the multiple index parameters.

[0236] In the embodiment of the present application, the device 10 of the embodiment of the present application further includes: an identification module.

[0237] Among them, the identification module is used to identify the analysis requirements of the user before generating the values of multiple index parameters of the basic human error probability according to the pre-constructed knowledge graph database and the multi-agent analysis results; determine the type of basic human error probability analysis based on the analysis requirements; determine the target knowledge graph of the knowledge graph database based on the type; generate the values of multiple index parameters of the basic human error probability based on the target knowledge graph and the multi-agent analysis results.

[0238] In the embodiment of the present application, the multi-agent analysis framework includes the first to fourth agents. Among them, the first agent is used to perform task analysis on the target case information; the second agent is used to perform context analysis on the target case information; the third agent is used to perform cognitive activity analysis on the target case information; the fourth agent is used to perform time constraint analysis on the target case information.

[0239] In the embodiment of the present application, the device 10 of the embodiment of the present application further includes: a construction module.

[0240] Among them, the construction module is used to acquire the data related to the basic human error probability in the target database before generating the values of multiple index parameters of the basic human error probability according to the pre-constructed knowledge graph database and the multi-agent analysis results; construct the knowledge graph database by using the target graph database management system and the related data.

[0241] In the embodiment of the present application, the multi-agent analysis results include task analysis results, context analysis results, cognitive activity analysis results and time constraint analysis results. The index parameters include human factor performance influencing factors, cognitive error modes, tasks, performance influencing factor variables and other performance influencing factors. The knowledge graph database includes scene familiarity knowledge graph, information availability and reliability knowledge graph and task complexity knowledge graph.

[0242] In the embodiments of the present application, the process of task analysis includes at least one of: task overview, task classification, analyzing the goals of the task, checking error types and impacts, and determining the complexity of the task; the process of context analysis includes at least one of: identifying the background conditions where the task occurs, analyzing the support required for task execution, clarifying the initial conditions and requirements for task initiation, and checking error metric data; the process of cognitive activity analysis includes at least one of: checking the cognitive requirements for task execution, and understanding the psychological processes behind task performance; the process of time constraint analysis includes at least one of: analyzing data sources, and analyzing time constraints.

[0243] It should be noted that the foregoing explanation of the embodiments of the basic human error probability assessment method also applies to the basic human error probability assessment device of this embodiment, and will not be elaborated here.

[0244] The basic human error probability assessment device proposed according to the embodiments of the present application can input the target case information to be analyzed into a pre-constructed multi-agent analysis framework. The multi-agent analysis framework inputs the multi-agent analysis results of the target case information, and combines it with the pre-constructed knowledge graph database to generate the values of multiple index parameters of the basic human error probability. Furthermore, based on the knowledge graph database and the values of multiple index parameters, it evaluates the basic human error probability of the target case information, so as to help industry personnel more efficiently quantify and evaluate the probability of human error, thereby improving the work efficiency of safety risk assessment, and further providing strong decision-making support for improving human reliability. Moreover, it does not rely too much on expert knowledge, can quickly generate the basic human error probability through simple text descriptions, reduces the evaluation time duration, and improves the evaluation efficiency.

[0245] Figure 3 It is a schematic structural diagram of an electronic device provided by the embodiments of the present application. The electronic device may include:

[0246] A memory 301, a processor 302, and a computer program stored on the memory 301 and executable on the processor 302.

[0247] When the processor 302 executes the program, it implements the basic human error probability assessment method provided in the foregoing embodiments.

[0248] Furthermore, the electronic device further includes:

[0249] A communication interface 303 for communication between the memory 301 and the processor 302.

[0250] The memory 301 is used to store a computer program executable on the processor 302.

[0251] The memory 301 may include high-speed RAM memory and may also include non-volatile memory, such as at least one disk memory.

[0252] If the memory 301, the processor 302, and the communication interface 303 are implemented independently, the communication interface 303, the memory 301, and the processor 302 can be interconnected through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity of representation, Figure 3 only a thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0253] Optionally, in a specific implementation, if the memory 301, the processor 302, and the communication interface 303 are integrated on a chip, the memory 301, the processor 302, and the communication interface 303 can communicate with each other through an internal interface.

[0254] The processor 302 may be a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0255] The embodiments of the present application also provide a computer-readable storage medium, on which a computer program or instruction is stored. When the computer program or instruction is executed by a processor, the above-mentioned basic human error probability assessment method is implemented.

[0256] The embodiments of the present application also provide a computer program product, including a computer program or instruction. When the computer program or instruction is executed, the above-mentioned basic human error probability assessment method is implemented.

[0257] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of this application. In this specification, the schematic descriptions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0258] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In the description of this application, the meaning of "N" is at least two, such as two, three, etc., unless otherwise specifically defined.

[0259] Any process or method description in the flowchart or described in other ways herein can be understood as representing a module, segment, or part of code including one or N executable instructions for implementing a customized logical function or process, and the scope of the preferred embodiments of this application includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in a reverse order according to the involved functions, rather than in the order shown or discussed, which should be understood by those skilled in the technical field to which the embodiments of this application belong.

[0260] It should be understood that each part of this application can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware as in another embodiment, it can be implemented by any one or a combination of the following techniques well known in the art: discrete logic circuits with logic gate circuits for implementing logical functions on data signals, application-specific integrated circuits with suitable combinational logic gate circuits, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0261] Those of ordinary skill in the art of this technology can understand that all or part of the steps carried by the methods of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

Claims

1. A basic human error probability assessment method, characterized in that: The following steps are involved: Obtain target case information that needs to be analyzed; Inputting the target case information into a pre-built multi-agent analysis framework, and the multi-agent analysis framework outputs a multi-agent analysis result of the target case information; The values ​​of multiple indicator parameters of basic human error probability are generated according to the pre-constructed knowledge graph database and the multi-agent analysis results, and the basic human error probability of the target case information is evaluated according to the knowledge graph database and the values ​​of the multiple indicator parameters.

2. The basic human error probability assessment method according to claim 1, characterized in that: Before generating the values ​​of multiple index parameters of basic human error probability according to the pre-built knowledge graph database and the multi-agent analysis results, it also includes: Identify users’ analytical needs; Determining the type of the basic human error probability analysis based on the analysis requirements; Determine a target knowledge graph of the knowledge graph database based on the type; Based on the target knowledge graph and the multi-agent analysis results, the values ​​of multiple indicator parameters of basic human error probability are generated.

3. The basic human error probability assessment method according to claim 1, characterized in that: The multi-agent analysis framework includes first to fourth agents, wherein: The first intelligent agent is used to perform task analysis on the target case information; The second agent is used to perform context analysis on the target case information; The third intelligent agent is used to perform cognitive activity analysis on the target case information; The fourth agent is used to perform time constraint analysis on the target case information.

4. The basic human error probability assessment method according to claim 1, characterized in that: Before generating the values ​​of multiple index parameters of basic human error probability according to the pre-built knowledge graph database and the multi-agent analysis results, it also includes: Acquire data related to the basic human error probability in a target database; The knowledge graph database is constructed using the target graph database management system and the related data.

5. The basic human error probability assessment method according to claim 1, characterized in that: The multi-agent analysis results include task analysis results, context analysis results, cognitive activity analysis results and time constraint analysis results; the indicator parameters include human performance influencing factors, cognitive error patterns, tasks, performance influencing factor variables and other performance influencing factors; the knowledge graph database includes scene familiarity knowledge graph, information availability and reliability knowledge graph and task complexity knowledge graph.

6. The basic human error probability assessment method according to claim 3, characterized in that: The task analysis process includes at least one of: task overview, task classification, analyzing task objectives, checking error types and impacts, and determining task complexity; The process of the context analysis includes at least one of: identifying the background conditions for the occurrence of the task, analyzing the support required for executing the task, clarifying the initial conditions and requirements for starting the task, and checking error measurement data; The process of cognitive activity analysis includes: examining at least one of the cognitive demands required for task execution and understanding the psychological processes behind task performance; The process of the time constraint analysis includes: analyzing at least one of data sources and analyzing time constraints.

7. A basic human error probability assessment device, characterized in that: include: An acquisition module is used to obtain target case information that needs to be analyzed; An input module, used for inputting the target case information into a pre-built multi-agent analysis framework, and the multi-agent analysis framework outputs the multi-agent analysis result of the target case information; An evaluation module is used to generate values ​​of multiple indicator parameters of basic human error probability based on a pre-constructed knowledge graph database and the multi-agent analysis results, and to evaluate the basic human error probability of the target case information based on the knowledge graph database and the values ​​of the multiple indicator parameters.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the basic human error probability assessment method as claimed in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: The computer program or instruction is executed by a processor to implement the basic human error probability assessment method as described in any one of claims 1-6.

10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed, the basic human error probability assessment method as described in any one of claims 1 to 6 is implemented.

Images