A functional safety analysis method and electronic device
By obtaining functional execution information and failure modes of the target vehicle from the functional safety knowledge base, the failure behavior and its hazardous events are determined, which solves the problem of the difficulty in comprehensively analyzing potential risks in the existing technology, and realizes full-dimensional functional safety analysis and precise safety design.
Patent Information
- Application Number
- CN202611528515.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-29
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies cannot fully reflect the potential risks of target functions during actual operation in functional safety analysis of automotive electronic systems, thus affecting the safety and reliability of the entire vehicle system.
By acquiring the target functions and failure modes of the target vehicle, functional execution information is obtained from the functional safety knowledge base to determine the failure behavior and its hazard events and fault propagation logic in the preset traffic scenario, thereby generating functional safety design constraints.
It enables a comprehensive analysis of target function failures, improves the consistency of functional safety analysis results and the completeness of risk identification, provides accurate safety design basis, and enhances the safety consistency of vehicle function development.
Smart Images

Figure CN123116455A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of automotive electronic functional safety analysis technology, and more specifically, to a functional safety analysis method and electronic device. Background Technology
[0002] As the complexity of automotive electronic systems continues to increase, the data interaction relationships between multiple functional domains such as body control, powertrain control, chassis control, cockpit control, and intelligent driving are becoming increasingly complex. Functional safety analysis has become a crucial step in the development process of automotive electronic systems. By analyzing the potential risks of target functions under different operating scenarios and generating corresponding safety analysis results, a safety design basis can be provided for subsequent functional development, thereby improving the safety and reliability of the entire vehicle system.
[0003] To meet the aforementioned functional safety analysis requirements, large language models can be used to analyze hazardous scenarios. For example, by identifying hazardous scenarios corresponding to autonomous driving functions and combining them with external road images, video data, or simulation scenarios, the risks in intelligent driving scenarios can be analyzed to assist in generating corresponding safety analysis results.
[0004] However, the above-mentioned safety analysis methods have limitations and cannot fully reflect the potential risks of the target function in actual operation, thereby affecting the safety and reliability of subsequent function development, system design, and vehicle operation. Summary of the Invention
[0005] In view of the above problems, this application is proposed to provide a functional safety analysis method and electronic device to achieve vehicle functional safety analysis tasks under different functional domains. The specific solution is as follows:
[0006] Firstly, this application provides a functional safety analysis method, including:
[0007] Obtain information about the object to be analyzed, including: the target function and the failure mode of the target function of the target vehicle;
[0008] Obtain the function execution information corresponding to the target function from the vehicle function specification file stored in the pre-configured functional safety knowledge base;
[0009] Based on the function execution information and the target function failure mode, determine the failure behavior of the target vehicle when the target function is in the target function failure mode;
[0010] The hazard event corresponding to the failure behavior in a preset traffic scenario is determined, and the fault propagation logic relationship corresponding to the hazard event is analyzed, including multiple logic nodes. The logic nodes are used to characterize at least one fault information of the vehicle control component, and the fault information has a fault causal relationship with the hazard event.
[0011] Based on the fault propagation logic relationship, functional safety design constraints are generated for the target function. These functional safety design constraints are used to indicate that the design of the vehicle control component corresponding to the logic node meets preset functional safety conditions.
[0012] In one possible implementation, in another implementation of the first aspect of the embodiments of this application,
[0013] The process of determining the failure behavior of the target vehicle when the target function is in the target function failure mode, based on the function execution information and the target function failure mode, includes:
[0014] A pre-configured large model is invoked to instruct the large model to predict the failure behavior of the target vehicle when the target function is in the target function failure mode, based on the function execution information.
[0015] In one possible implementation, in another implementation of the first aspect of the embodiments of this application,
[0016] The process of determining the failure behavior of the target vehicle when the target function is in the target function failure mode, based on the function execution information and the target function failure mode, includes:
[0017] Based on the function execution information, determine the target controller type associated with the target function;
[0018] Retrieve the set of candidate failure behaviors corresponding to the target controller type from the set of failure behaviors corresponding to different controller types stored in the functional safety knowledge base;
[0019] A pre-configured large model is invoked to instruct the large model to perform failure behavior analysis on the target vehicle based on the candidate failure behavior set and the target function failure mode, so as to obtain the failure behavior of the target vehicle when the target function is in the target function failure mode.
[0020] In one possible implementation, in another implementation of the first aspect of the embodiments of this application, the method further includes:
[0021] Calculate the semantic similarity between the failure behavior of the target vehicle and each candidate failure behavior in the candidate failure behavior set;
[0022] If the semantic similarity is lower than a preset threshold, the corresponding failure behavior of the target vehicle will be marked as a behavior pending review.
[0023] In one possible implementation, in another implementation of the first aspect of the embodiments of this application, the process of determining the hazard event corresponding to the failure behavior in a preset traffic scenario includes:
[0024] Based on the hazardous characteristics of the failure behavior and prior knowledge of functional safety, basic traffic scenarios that are causally related to the failure behavior are selected from preset traffic scenarios.
[0025] The factors influencing the harmful consequences of the failure behavior are combined to generate at least one scenario variable combination;
[0026] A pre-configured large model is invoked to instruct the large model to deduce the interaction risk between the target vehicle and external traffic participants under the failure behavior based on the basic traffic scenario and the combination of various scenario variables, and to obtain the corresponding hazard event.
[0027] In one possible implementation, in another implementation of the first aspect of the embodiments of this application, the process of analyzing the fault propagation logic relationship corresponding to the hazardous event includes:
[0028] The hazardous event is taken as the top event, and the primary fault cause that triggered the hazardous event is analyzed based on the functional execution information.
[0029] For each of the first-level fault causes, the corresponding lower-level fault causes are broken down level by level until the underlying fault causes that cannot be further decomposed or do not need to be further decomposed are obtained.
[0030] Establish the logical relationship between the hazardous events and the causes of failures at each level to obtain the fault propagation logical relationship corresponding to the hazardous events.
[0031] In one possible implementation, in another implementation of the first aspect of this application, the process of establishing the logical relationship between the hazardous event and the causes of failures at each level, and obtaining the fault propagation logical relationship corresponding to the hazardous event, includes:
[0032] Based on the logical relationship between the hazardous events and the causes of failures at each level, corresponding logical expressions are generated;
[0033] Based on the logical expression, the minimum cut set that triggers the occurrence of the hazardous event is determined, and the minimum cut set is used as the fault propagation logical relationship corresponding to the hazardous event.
[0034] In one possible implementation, in another implementation of the first aspect of the embodiments of this application, the method further includes:
[0035] Obtain a set of functional safety experience parameters and perform context correction on the currently generated analysis results. The analysis results include: the failure behavior and / or the hazard event and / or the fault propagation logic relationship, as well as the functional safety design constraints.
[0036] In one possible implementation, in another implementation of the first aspect of the embodiments of this application, the method further includes:
[0037] Based on preset audit rules, the consistency audit of the currently generated analysis results is performed. The analysis results include: the failure behavior and / or the hazard event and / or the fault propagation logic relationship, as well as the functional safety design constraints.
[0038] If a logical conflict or missing item is detected in the analysis results, the current step is re-executed or the previous one or more steps are returned to be executed.
[0039] Secondly, this application provides an electronic device, including: a memory and a processor;
[0040] The memory is used to store programs;
[0041] The processor is configured to execute the program to implement the functional safety analysis method described in any of the first aspects of this application.
[0042] Thirdly, this application provides a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the functional safety analysis method described in any of the first aspects of this application.
[0043] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the functional safety analysis method described in any of the first aspects of this application.
[0044] By means of the above technical solution, the functional safety analysis method provided in this application obtains the functional execution information corresponding to the target function of the target vehicle from the vehicle functional information stored in the pre-configured functional safety knowledge base. This enables the subsequent functional safety analysis to be carried out around the execution logic, control relationship and other execution information of the target function itself. This breaks through the analysis boundary of related technical solutions that only focus on the external driving environment, and can conduct targeted analysis on the impact of target function failure based on the vehicle's functional operation mechanism.
[0045] Furthermore, based on functional execution information and the target function failure mode, the failure behavior exhibited by the vehicle when the target function is in a failed state is determined. This transforms the abstract functional failure mode into a perceptible actual behavioral manifestation during vehicle operation, fully deriving the actual impact of the target function failure. On this basis, the potential hazardous events caused by the vehicle's failure behavior in preset traffic scenarios are identified, establishing a correlation between functional failure and the external traffic environment. This allows functional safety analysis to not only consider the operational logic of the target function itself but also to conduct joint analysis in conjunction with external environmental factors, thereby improving the consistency of functional safety analysis results and the completeness of risk identification across the entire functional domain.
[0046] Based on this, the fault propagation logic of hazardous events is analyzed to locate vehicle control components with causal relationships to the hazardous events and the fault propagation paths, thereby achieving comprehensive identification of target function failure tracing and external scenario risks. Finally, functional safety design constraints are generated based on the fault propagation logic, providing precise safety basis for the development and design of vehicle control components, thereby improving the safety consistency and engineering application capabilities in the whole vehicle function development process, and realizing a more comprehensive whole vehicle-level functional safety analysis for any target function across all functional domains. Attached Figure Description
[0047] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:
[0048] Figure 1 A schematic diagram of an implementation system architecture for the functional safety analysis method provided in this application embodiment;
[0049] Figure 2 A multi-agent workflow flowchart is provided for embodiments of this application;
[0050] Figure 3 A flowchart illustrating a functional safety analysis method provided in an embodiment of this application;
[0051] Figure 4 A schematic diagram of a data structure provided in an embodiment of this application;
[0052] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0053] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0054] This application relates to technical fields such as artificial intelligence, automotive safety design and development, and automotive electronic functional safety analysis, and can be applied to safety analysis scenarios of vehicle functions such as steering control, braking control, power control, intelligent driving, chassis control, and multi-domain fusion control.
[0055] Understandably, as vehicle electronic systems evolve towards centralized and software-defined vehicle architectures, the functional coupling, signal interaction, and fault propagation relationships between different controllers become increasingly complex. A single functional failure can affect multiple functional domains through communication or control links. Therefore, during vehicle development, it is necessary to conduct system-level functional safety analysis of the failure modes of target functions to identify potential hazards in advance and generate corresponding safety design justifications.
[0056] Currently, most functional safety analysis schemes focus on risk analysis based on external traffic scenarios, emphasizing the impact of changes in road environment, traffic participants, and environmental conditions on vehicle behavior. Analysis of the vehicle's internal functional execution logic, controller coordination relationships, and fault propagation mechanisms is relatively insufficient. However, dangerous vehicle behavior can be triggered not only by the external environment but also directly by internal vehicle anomalies. For example, abnormal controller logic, sensor signals, communication links, or actuators can all lead to unexpected steering, abnormal acceleration, abnormal braking force, or loss of vehicle stability.
[0057] Based on the actual needs in the above application scenarios and the shortcomings of existing technologies, this application provides a functional safety analysis method. The functional safety analysis method of this application embodiment will be described in detail below with reference to the accompanying drawings.
[0058] See Figure 1 , Figure 1 A schematic diagram of a system architecture is shown. The system may include a terminal 100 and a server 200, wherein the server 200 may include one or more servers (…). Figure 1 (This example uses a server as an illustration).
[0059] Either terminal 100 or server 200 can be used independently to execute the functional safety analysis method provided in the embodiments of this application. Alternatively, terminal 100 and server 200 can also be used collaboratively to execute the functional safety analysis method provided in the embodiments of this application.
[0060] In this application embodiment, the terminal 100 can be an in-vehicle domain controller, an electronic control unit (ECU), a functional safety analysis workstation, an automotive electronics development platform, a cloud-based functional safety analysis server, an in-vehicle testing and diagnostic equipment, a vehicle architecture development terminal, an automotive-grade in-vehicle computing platform, an in-vehicle information processing terminal, an automotive R&D simulation terminal, etc. This application embodiment does not impose any restrictions on this.
[0061] In some possible implementations, multiple agents can be deployed within the system comprised of terminal 100 and server 200 to perform various analysis steps in the functional safety analysis method provided in this application embodiment. (Refer to...) Figure 2 Multiple collaborative intelligent agents can include: a function extraction agent, a function analysis agent, a function failure analysis agent, a failure scenario review agent, a hazard analysis agent, a security incident analysis agent, and a security requirement analysis agent. An additional functional safety knowledge base is introduced as the core knowledge data source and context support for each agent, providing data and experience assurance for functional safety analysis and improving the reliability of the analysis results.
[0062] Furthermore, a multi-agent collaborative analysis architecture is formed based on a preset workflow. The controller coordinates and schedules each agent or the agents automatically flow to implement the solution of this application. Each agent can be responsible for task processing at different analysis stages, and collaborative reasoning is achieved through context passing, task scheduling, and result correction, forming a multi-stage joint analysis process. It should be noted that the specific analysis steps, input-output relationships, reasoning logic, and collaboration methods implemented by each agent can be understood by referring to the detailed description of the method embodiments of this application below, and will not be repeated here to avoid repetition.
[0063] This application provides a functional safety analysis method, which is illustrated by applying the method to a computer device. Specifically, the computer device may be... Figure 1 The system consists of terminal 100 or a combination of terminal 100 and server 200. (Refer to...) Figure 3 The present application provides a flowchart of a functional safety analysis method, which may include steps S110 to S150, and these steps are described in detail below.
[0064] Step S110: Obtain information about the object to be analyzed. The information about the object to be analyzed may include the target function and the failure mode of the target function of the target vehicle.
[0065] The target vehicle can be any vehicle that requires functional safety analysis, such as new energy vehicles, intelligent driving vehicles, and vehicles with multi-domain electronic control architectures. The target function can be the vehicle control function to be analyzed in the target vehicle, such as steering control, braking control, power output, vehicle stability control, intelligent driving control, and automatic parking.
[0066] The target function failure mode can be understood as the abnormal operating mode when the target function deviates from the expected working state. Refer to Table 1 below for a list of function failure modes and explanations of each mode. Target function failure modes may include, but are not limited to, false triggering, non-triggered operation, stuck mode, excessive mode, insufficient mode, premature mode, delayed mode, oscillation mode, partial failure mode, and reverse mode. However, in practical applications, other failure mode types can be extended according to industry standards or enterprise-defined standards, as long as they can characterize the abnormal functional state. This application does not impose a single limitation on this.
[0067] Table 1. List of functional failure modes and explanations of each mode.
[0068]
[0069] Based on the above understanding, step S110 can be triggered by the front end uploading vehicle development files or specification files, etc. The system responds to the files uploaded by the front end, extracts the target function name by automatically parsing the file content, and obtains the list of preset failure modes associated with the target function name as the target function failure mode, thus advancing the execution of step S120.
[0070] Alternatively, step S110 can be triggered by the vehicle's fault diagnosis results. In response to these results, the system automatically identifies potential risky target functions based on fault codes, anomaly logs, controller alarm information, or communication anomaly records collected during vehicle testing. Furthermore, when the system detects an anomaly in the vehicle control component corresponding to the target function, it can automatically select the appropriate target function failure mode based on the anomaly type.
[0071] Step S120: Obtain the function execution information corresponding to the target function from the vehicle function specification file stored in the pre-configured functional safety knowledge base.
[0072] A functional safety knowledge base is a collection of functional safety data pre-built by developers using technologies such as Retrieval-Augmented Generation (RAG), knowledge management, and vector databases. It can store a large amount of vehicle functional specification documents, failure behavior sets, sets of potential hazardous events for different controllers, and HARA (Hazard Analysis and Risk Assessment) scenario libraries. Vehicle functional specification documents can be understood as document data describing the vehicle's functional operating logic and control relationships, and may include, for example, functional requirements specifications, functional definition documents, control logic specifications, system design documents, controller interface documents, or functional flow description documents.
[0073] Optionally, each document can first be processed by text parsing, chapter splitting, and semantic slicing to be broken down into several text blocks with independent semantics; then, semantic vectorization processing is performed on each text block to generate a corresponding text block index, and the index and text blocks are stored synchronously in the functional safety knowledge base so that semantic retrieval can be performed more accurately in the future, providing knowledge support for functional safety analysis and significantly improving the accuracy and standardization of the analysis results.
[0074] In this embodiment, the target function can be filled into a preset function extraction prompt template to generate standardized retrieval and extraction instructions. Then, a pre-configured function extraction agent is called to instruct the function extraction agent to generate a corresponding retrieval vector based on the retrieval and extraction instructions, the function name, function description or function keywords corresponding to the target function, and the most semantically relevant text block to the target function is retrieved from the functional safety knowledge base based on the vector similarity, as the context input of the large model.
[0075] Subsequently, the function extraction agent invokes the large model, inputting information such as "Based on the following functional specifications, analyze the functional logic, input-output relationships, and other functional execution information of the target function" to instruct the large model to analyze and extract the corresponding functional execution information from the context. This functional execution information can be understood as representing the actual execution flow and associated parameters of the target function within the vehicle, specifically including information such as function triggering conditions, participating controllers, control logic relationships, execution links, input-output interfaces, and input-output signals.
[0076] This application embodiment, through this retrieval and generation mechanism based on RAG technology, can accurately locate functional logic details, avoid attention distraction caused by context window limitations when processing full files in a large model, thereby improving the accuracy and relevance of functional execution information extraction.
[0077] Step S130: Based on the function execution information and the target function failure mode, determine the failure behavior of the target vehicle when the target function is in the target function failure mode.
[0078] Understandably, failure behavior can refer to the abnormal behavior exhibited by the target vehicle during actual operation when the target function experiences an abnormal failure mode. For example, when the power steering function is in a false trigger failure mode, the corresponding failure behavior might be the vehicle unexpectedly veering to the left; when the power output function is in an excessive failure mode, the corresponding failure behavior might be excessive power output, causing the vehicle to suddenly lurch forward or accelerate abnormally.
[0079] Therefore, determining the failure behavior involves mapping abstract functional failure modes to specific vehicle physical behaviors. Specifically, based on the acquired functional execution information, the control logic under normal conditions can be understood. Then, combined with the determined target function failure mode, the actual vehicle performance under abnormal conditions can be deduced. For example, if the target function is power steering, the functional execution information shows its function is to provide steering torque assistance. When the target function failure mode is exceeded, the determined failure behavior is unexpected excessive steering or over-steering. Optionally, a set of failure behaviors corresponding to the target function failure mode can be retrieved from the failure behavior set stored in the functional safety knowledge base. Optionally, a functional failure analysis agent can also be invoked, using functional execution information, the target function, and the target function failure mode as inference input. Based on preset failure analysis rules and the semantic reasoning capabilities of a large model, the agent can accurately deduce the failure behavior and ultimately output the failure behavior corresponding to the target function.
[0080] Based on this, by transforming logical anomalies at the functional level into physical failure behaviors at the vehicle level, a mapping relationship from software logic to the physical performance of the whole vehicle is established, providing a direct basis for subsequent assessment of its degree of harm in traffic scenarios.
[0081] Step S140: Determine the hazard events corresponding to the failure behavior in the preset traffic scenario, and analyze the fault propagation logic relationship corresponding to the hazard events.
[0082] In this context, a hazardous event refers to a potential safety accident caused by a failed behavior in a preset traffic scenario, such as a vehicle collision, a rear-end collision, or pedestrian injury. This application embodiment, after determining the hazardous event that exists after the interaction between the external traffic scenario and the failed behavior, further traces inwards, using methods such as Fault Tree Analysis (FTA), Event Tree Analysis (ETA), and dependency analysis to analyze the internal fault propagation path leading to the hazardous event, obtaining the corresponding fault propagation logical relationship. Specifically, this fault propagation logical relationship can include multiple logical nodes, where each logical node can be used to represent at least one fault information of the vehicle control component. This fault information has a causal relationship with the hazardous event. For example, the fault information represented by the logical node can include sensor signal drift, controller algorithm error, actuator jamming, etc. Optionally, the logical node can also be used to represent other fault causes besides the fault information of the vehicle control component, such as steering column mechanical jamming, power supply abnormality, communication link abnormality, or driver misoperation. It is understandable that each logical node can be regarded as a fault factor that leads to the occurrence of a hazardous event. The logical nodes are interconnected based on the corresponding fault causal relationship or logical relationship, thereby representing the fault propagation path of the vehicle's internal faults being transmitted step by step and ultimately leading to the hazardous event.
[0083] Specifically, failure behaviors can be input into vehicle dynamics simulation models or traffic flow simulation environments. Through simulation analysis, changes in vehicles under different speeds, road surface adhesion conditions, and traffic densities can be analyzed. Based on the simulation results, the interaction changes between vehicles and traffic participants can be analyzed to generate corresponding hazard events.
[0084] Optionally, a hazard analysis agent can be invoked, taking the failure behavior, target function, function execution information, and a pre-defined database of typical traffic scenarios as input. Relying on hazard analysis rules, exposure rate / severity / controllability assessment parameters, etc., from the functional safety knowledge base, it directly outputs the hazard event corresponding to the failure behavior through large-scale model semantic reasoning and scenario generalization analysis. Then, a hazard event analysis agent is invoked to decompose the fault cause leading to the hazard event step-by-step from the hazard event and the function execution information corresponding to the target function. For example, the fault propagation path can be analyzed from dimensions such as input anomalies, communication anomalies, control logic anomalies, execution anomalies, and power supply anomalies. Subsequently, logical relationships between different logical nodes are established to identify the fault propagation logic.
[0085] Understandably, hazard events can accurately correlate vehicle internal function failures with external traffic scenario risks, comprehensively identifying the real driving safety risks caused by function failures. Meanwhile, the fault propagation logic can accurately describe the hierarchical transmission chain, logical connections, and evolution process of vehicle internal component failures, clarifying how underlying hardware failures, software anomalies, and signal deviations are transmitted step-by-step through control logic, execution links, and communication interactions, ultimately triggering the corresponding hazard event.
[0086] Therefore, the analysis method adopted in step S140 breaks through the limitations of existing technologies that only focus on external perception. It realizes a full-dimensional integrated analysis of external traffic interaction risks and internal component fault tracing. It can not only transform internal hidden faults into explicit driving risks, but also reverse locate the internal fault propagation chain from the hazard event, thus comprehensively ensuring the integrity, accuracy and depth of functional safety risk assessment.
[0087] Step S150: Generate functional safety design constraints corresponding to the target function based on the fault propagation logic relationship.
[0088] Functional safety design constraints are used to indicate that the design of the vehicle control components corresponding to the logic nodes meets preset functional safety conditions. These preset functional safety conditions can refer to industry standards or internally defined safety indicators or design specifications. The purpose of generating functional safety design constraints in this step is to provide specific design guidance for the subsequent development and design of vehicle control components.
[0089] Specifically, functional safety requirement templates can be pre-built for different fault types. Communication faults can be associated with functional safety requirements such as communication timeout detection, communication redundancy verification, and abnormal communication isolation. Combining these with logical nodes in the fault propagation logic, the system automatically matches and adapts the functional safety requirement templates, generating functional safety design constraints corresponding to the target function. Optionally, a safety requirement analysis agent can be invoked, using the fault propagation logic, fault types corresponding to logical nodes, at-risk components, and hazardous events as inference inputs. Relying on the functional safety knowledge base's writing standards and engineering experience, the agent autonomously derives functional safety requirements, defines attributes, and allocates components through a large model, intelligently generating customized functional safety design constraints adapted to the target failure scenario.
[0090] Understandably, this step transforms the fault propagation logic relationship analyzed in step S140 into an executable and verifiable design language, which can guide the implementation of hardware selection, software architecture design, and safety mechanisms, thereby blocking the fault propagation path during the development phase and ensuring that the design of the target vehicle meets the preset functional safety conditions.
[0091] In summary, the functional safety analysis method provided in this application obtains the functional execution information corresponding to the target function of the target vehicle from the vehicle functional information stored in the pre-configured functional safety knowledge base. This enables subsequent functional safety analysis to focus on the execution logic, control relationships, and other execution information of the target function itself. This breaks through the analysis boundary of related technical solutions that only focus on the external driving environment, and enables targeted analysis of the impact of target function failure based on the vehicle's functional operation mechanism.
[0092] Furthermore, based on functional execution information and the target function failure mode, the failure behavior exhibited by the vehicle when the target function is in a failed state is determined. This transforms the abstract functional failure mode into a perceptible actual behavioral manifestation during vehicle operation, fully deriving the actual impact of the target function failure. On this basis, the potential hazardous events caused by the vehicle's failure behavior in preset traffic scenarios are identified, establishing a correlation between functional failure and the external traffic environment. This allows functional safety analysis to not only consider the operational logic of the target function itself but also to conduct joint analysis in conjunction with external environmental factors, thereby improving the consistency of functional safety analysis results and the completeness of risk identification across the entire functional domain.
[0093] Based on this, the fault propagation logic of hazardous events is analyzed to locate vehicle control components with causal relationships to the hazardous events and the fault propagation paths, thereby achieving comprehensive identification of target function failure tracing and external scenario risks. Finally, functional safety design constraints are generated based on the fault propagation logic, providing precise safety basis for the development and design of vehicle control components, thereby improving the safety consistency and engineering application capabilities in the whole vehicle function development process, and realizing a more comprehensive whole vehicle-level functional safety analysis for any target function across all functional domains.
[0094] Next, other possible implementations of the functional safety analysis method proposed in this application will be described through the following embodiments.
[0095] In one possible implementation, the implementation process of step S130 may include: determining the target controller type associated with the target function based on the function execution information; retrieving the candidate failure behavior set corresponding to the target controller type from the failure behavior set corresponding to different controller types stored in the functional safety knowledge base; and calling a pre-configured large model to instruct the large model to perform failure behavior analysis on the target vehicle based on the candidate failure behavior set and the target function failure mode, so as to obtain the failure behavior of the target vehicle when the target function is in the target function failure mode.
[0096] First, the system parses the function execution information obtained in step S120 and extracts the control unit identifier that executes the target function to determine the target controller type. For example, if the target function is power steering, the associated target controller type is identified as an EPS (Electric Power Steering) controller; if the target function is engine torque management, the target controller type is identified as an EMS (Engine Management System) controller.
[0097] After determining the type of the target controller, the set of candidate failure behaviors corresponding to that type is precisely retrieved from the functional safety knowledge base. For example, for an EPS controller, the set of candidate failure behaviors may include physically observable abnormal behaviors such as unexpected steering, oversteering, understeering, steering jamming, and steering vibration.
[0098] The candidate failure behavior set obtained above is used as the inference context or constraint input of the large model, defining an inference boundary for the large model that conforms to vehicle physics, industry engineering common sense, and functional safety experience. Within this inference boundary, the large model, in conjunction with the failure modes of the target function, matches, filters, and combines the candidate failure behaviors from the set, outputting the most realistic failure behavior. Alternatively, the aforementioned candidate failure behavior set can be used as the basic prior knowledge of the large model, enabling it to learn the failure behavior deduction logic that conforms to physical rules and engineering practices under corresponding failure modes. Furthermore, by combining multi-dimensional information such as vehicle operating status, function execution information, and target function failure modes, the large model can accurately deduce the failure behavior of the target vehicle.
[0099] It is understandable that having a set of candidate failure behaviors to provide prior knowledge for the large model can prevent the large model from developing illusions due to a lack of vertical domain knowledge, thereby improving the professionalism and credibility of the analysis results. Furthermore, the specific form in which the aforementioned set of candidate failure behaviors affects the reasoning of the large model can be flexibly configured by adjusting the prompt word template input to the large model; this application does not impose a unique limitation on this.
[0100] Based on this, in order to balance the accuracy of functional safety analysis with the ability to explore new risks, it may also include: calculating the semantic similarity between the failure behavior of the target vehicle and each candidate failure behavior in the candidate failure behavior set; when the semantic similarity is lower than a preset threshold, marking the corresponding failure behavior of the target vehicle as a behavior to be reviewed.
[0101] Specifically, after the large model outputs the failure behaviors of the target vehicle, the semantic similarity between each failure behavior and each candidate failure behavior in the candidate failure behavior set is calculated using methods such as cosine similarity, Jaccard coefficient, or distance metric based on embedding vectors. Furthermore, the semantic similarity is compared with a preset threshold, which can be configured according to the inference bias tolerance of the actual application scenario.
[0102] If the semantic similarity between a candidate failure behavior and the failure behavior output by the model is not lower than the preset threshold, it indicates that the failure behavior output by the large model belongs to the known experience range, and it can be confirmed as a valid analysis result and automatically transferred to the next stage.
[0103] However, if the semantic similarity between a candidate failure behavior and the failure behavior output by the model is lower than a preset threshold, it means that the large model has made an erroneous inference that deviates from reality, or that the large model has inferred a novel failure behavior that has not yet been included in the knowledge base based on prior knowledge. To avoid missing any potential innovative risks and ensure output quality, this embodiment of the application marks such failure behaviors with semantic similarity lower than the preset threshold as "behaviors pending review" and pushes them to the front end for verification. If the front end confirms that it is a real novel risk, it is recognized as a valid analysis result and automatically transferred to the next stage, and added to the knowledge base to achieve dynamic evolution of the knowledge base. Otherwise, it is considered a hallucination of the large model and is removed from the output results.
[0104] It is understood that the embodiments of this application, through this collaborative mechanism of automated initial screening based on semantic similarity and front-end fallback review, can ensure the efficiency and accuracy of routine analysis while retaining a window for discovering unknown risks, thus balancing analytical efficiency and innovation.
[0105] Furthermore, based on the failure behavior exhibited by the target vehicle when the target function is in failure mode, as output by the above large model, step S140 is executed to determine the hazard event corresponding to the failure behavior in the preset traffic scenario.
[0106] In one possible implementation, the process of determining the hazard event corresponding to the failure behavior in a preset traffic scenario in step S140 may include: selecting basic traffic scenarios that are causally related to the failure behavior from the preset traffic scenarios based on the hazard characteristics of the failure behavior and prior knowledge of functional safety; combining variables of factors affecting the hazard outcome of the failure behavior to generate at least one scenario variable combination; and calling a pre-configured large model to instruct the large model to deduce the interaction risk between the target vehicle and external traffic participants under the failure behavior based on the basic traffic scenarios and each set of scenario variable combinations, thereby obtaining the corresponding hazard event.
[0107] Specifically, real-world traffic scenarios have almost infinite combinations of possibilities. Directly allowing a large model to traverse all possible scenarios for risk analysis would not only consume enormous computational resources but also easily lead to divergent analysis, omission of key risks, or the generation of a large number of invalid results due to the excessively large search space. To address this issue, embodiments of this application use filtering to converge the infinite scenarios into a finite set of computable scenarios.
[0108] Specifically, firstly, the hazard characteristics of the failure behavior are extracted, namely the physical properties and potential destructive power of the failure behavior itself. Secondly, preset traffic scenarios are obtained from the HARA scenario library of the functional safety knowledge base. These preset traffic scenarios almost approximate all traffic scenarios, resulting in a massive dataset. Based on prior functional safety knowledge, basic traffic scenarios with causal relationships to the hazard characteristics are selected from the preset traffic scenarios. Irrelevant and low-risk redundant scenarios are eliminated, significantly narrowing the analysis scope of the large model. This avoids the inefficiency and analytical bias caused by quantifying all scenarios, improving the accuracy of hazard events.
[0109] After completing the basic traffic scenario screening, the key influencing factors affecting the harmful consequences of the failure behavior are further identified, and these key influencing factors are combined to generate at least one scenario variable combination. These key influencing factors may include: vehicle speed range (low speed, medium speed, high speed), road surface conditions (e.g., dry, slippery, icy / snowy), weather conditions (e.g., sunny, rainy, foggy), lighting conditions (e.g., daytime, nighttime, backlighting), and the type of traffic participant (e.g., passenger cars, trucks, pedestrians, cyclists), etc., and the above factors are combined as variables.
[0110] Optionally, the aforementioned key influencing factors can be categorized into different levels and orthogonally combined to ensure diversity among the combinations. For example, for the basic traffic scenario of "this vehicle driving alongside another vehicle," multiple sets of scenario variable combinations can be automatically generated, such as "high speed + slippery road surface + rainy day," "medium speed + dry road surface + nighttime," and "low speed + icy / snowy road surface + foggy day." This structured variable combination ensures that, within a limited basic scenario framework, key environmental factors affecting the severity of hazards can be systematically covered, avoiding coverage blind spots caused by random sampling.
[0111] Subsequently, the selected basic traffic scenarios and combinations of scenario variables are used to construct structured prompts, which are then used to perform interactive risk simulations using a pre-configured large model. During this process, the large model, based on vehicle dynamics, traffic rule understanding, and physical world knowledge, simulates the dynamic interaction between the target vehicle and external traffic participants under specific failure behaviors, predicting possible collision patterns, damage levels, and accident severity, thereby identifying the corresponding hazardous events. For example, when the input is "unexpected steering failure behavior + basic scenario of this vehicle driving alongside an adjacent vehicle + high speed + slippery road surface variable combination," the large model can deduce the specific hazardous event: "The vehicle loses stability due to a sudden change in lateral force, intrudes into the adjacent lane, and collides sideways with the adjacent vehicle. Due to the high speed and low road adhesion, the driver has difficulty correcting, which may lead to serious injuries or fatalities." Furthermore, it can assess the risk level based on this.
[0112] In this process, this embodiment of the application uses the hazardous characteristics of failure behavior and prior knowledge of functional safety to screen basic traffic scenarios that are causally related to the current failure behavior from almost all traffic scenarios. It then further combines variables of key influencing factors affecting the hazardous outcome, thereby converging the originally infinitely divergent traffic risk analysis space into a finite, structured set of scenarios with risk relevance. This avoids the problems of excessive computational resource consumption, divergent risk analysis, and omission of key hazardous scenarios caused by directly traversing all traffic scenarios. Simultaneously, this embodiment of the application further combines the reasoning capabilities of large models on vehicle dynamics, traffic rules, and the interactive behavior of traffic participants to deduce the dynamic interaction process of the target vehicle under different combinations of scenario variables. This not only improves the scenario coverage and accuracy of the hazardous event analysis process but also enhances the realism and engineering interpretability of hazardous behavior deduction in complex traffic scenarios. This enables a more systematic, accurate, and efficient functional safety analysis of the hazardous outcomes of vehicle failure behavior in real traffic environments, providing comprehensive and reliable support for subsequent risk level assessment, fault propagation analysis, and safety design.
[0113] Based on the process described above, up to step S140, we can obtain the causal relationships between the target function, failure mode, failure behavior, traffic scenario, and hazardous event. (Refer to...) Figure 4 This application provides a schematic diagram of a data structure. The data structure is a hierarchical causal relationship graph, with the vehicle's target functions (e.g., functions A to Z) as the top-level nodes, and multi-dimensional causal relationship links constructed layer by layer downwards:
[0114] The first level is the association between target functions and failure modes. Each target function (such as function A) corresponds to multiple possible failure modes (such as failure mode α1 to failure mode αn), which characterize the abnormal types of the target function that deviate from the expected working state.
[0115] The second level is the association between failure modes and failure behaviors. Each failure mode is further mapped to a specific failure behavior in vehicle operation (such as failure mode α1 corresponding to failure behaviors β1 to failure behaviors βn), realizing the concrete expression of abstract failure modes to vehicle behavior at the physical level.
[0116] The third level is the association between failure behavior and traffic scenarios. Each failure behavior can act on different preset traffic scenarios (e.g., failure behavior β1 corresponds to traffic scenarios γ1 to γn), representing the external environmental conditions in which the failure behavior occurs.
[0117] The fourth level is the association between traffic scenarios and hazardous events. In each traffic scenario, the failure behavior will trigger a corresponding specific hazardous event (such as traffic scenario γ1 corresponding to hazardous events δ1 to δn), which represents the final safety risk result caused by the failure behavior in a specific environment.
[0118] This data structure, through a hierarchical and structured approach, fully preserves the causal logic of the entire chain from "target function" to "failure mode," "failure behavior," "traffic scenario" and "hazard event," providing a traceable and queryable standardized data foundation for subsequent fault propagation analysis, safety target formulation, and functional safety requirement derivation. It realizes the full-chain correlation and structured storage of vehicle internal failure mechanisms and external driving risks.
[0119] For each hazard event in this data structure, the corresponding target function can be traced back in reverse. Then, the causes leading to the hazard event can be deduced from the function execution information of the target function, thus tracing the complete fault chain. In one possible implementation, the process of analyzing the fault propagation logic relationship corresponding to the hazard event may include: performing fault tree analysis with the hazard event as the top event to obtain the fault propagation logic relationship corresponding to the hazard event.
[0120] Fault Tree Analysis (FTA) is a top-down deductive failure analysis method used to identify all possible combinations of causes leading to a hazardous event. In the embodiments of this application, this process can be automatically executed based on a large model and knowledge base. First, the hazardous event is directly defined as the top event of the fault tree. Then, a pre-built hazardous event analysis agent is invoked, and combined with the functional execution information of the target function, the fault causes at each level leading to the top event are automatically traced downwards. This transforms the implicit fault propagation process within the vehicle into a standardized explicit calculation process, significantly improving analysis efficiency and consistency.
[0121] Specifically, the hazard event is taken as the top event, and the primary fault cause that caused the hazard event is analyzed based on the functional execution information. For each primary fault cause, the corresponding lower-level fault causes are decomposed down level by level until the underlying fault causes that cannot be further decomposed or do not need to be decomposed are obtained. The logical relationship between the hazard event and the fault causes at each level is established to obtain the fault propagation logical relationship corresponding to the hazard event.
[0122] Specifically, firstly, after identifying the top event, a thorough analysis of key elements in the function execution information, such as "enabling conditions," "trigger conditions," and "execution outputs," is conducted to identify the direct causes that may have led to the top event, i.e., primary fault causes or primary intermediate events. These can include involved vehicle control components, abnormal behaviors, parameter records, etc. For example, if the top event is "unintended steering leading to a collision," then primary fault causes might include "the EPS controller issuing an incorrect torque command," "mechanical jamming of the steering column," or "driver misoperation," etc.
[0123] Subsequently, for each primary fault cause, a recursive breakdown is performed. For example, the intermediate event of "EPS controller issuing incorrect torque command" can be further broken down into next-level causes such as "torque sensor signal drift," "control algorithm parameter overflow," or "abnormal power supply voltage." This breakdown process continues until the lowest level is reached, revealing the underlying fault cause. The underlying fault cause or event typically refers to hardware failures, software defects, or human errors that cannot be further decomposed, or events that do not require further decomposition based on the required analytical granularity.
[0124] Finally, logic gates (such as AND gates, OR gates, and voting gates) can be used to clearly express the causal logic relationships between the causes of faults at each level. For example, "torque sensor signal drift" and "redundancy check failure" must occur simultaneously for "erroneous torque command not to be intercepted," and the relationship between the two is an AND gate relationship. Through this hierarchical structure, the embodiments of this application transform textual functional descriptions into computer-processable symbolic modeling language, laying the foundation for subsequent quantitative analysis.
[0125] Optionally, the smallest cut set that triggers the occurrence of a hazardous event in the logical expression generated by the logic gates can be used as the fault propagation logic relationship corresponding to the hazardous event.
[0126] The minimal cut set is the smallest set of basic events that leads to the top event. It can be understood that the minimal cut set represents the set of weakest links in a functional failure. Using this as the logical relationship for fault propagation, functional safety design constraints can be formulated to directly focus on the critical underlying fault paths that lead to hazardous events, avoiding redundant safety constraints on non-critical vehicle control components, and significantly improving the targeting and effectiveness of functional safety design constraints. It can also optimize the allocation of development resources, avoiding over-development and cost waste caused by indiscriminate safety design, concentrating limited resources on the weakest links in the system, and balancing the contradiction between safety goals and development costs and cycles. Furthermore, by covering all critical fault paths, it can systematically block the minimum fault occurrence link of hazardous events, avoiding the omission of critical risk points, significantly improving the robustness of system fault protection, and ensuring vehicle driving safety.
[0127] Based on the functional safety analysis method obtained by all the above possible implementation methods or combinations thereof, it may also include: obtaining a set of functional safety empirical parameters, performing context correction on the currently generated analysis results, and the analysis results include: failure behavior and / or hazard events and / or fault propagation logical relationships, as well as functional safety design constraints.
[0128] Specifically, the functional safety experience parameter set is a benchmark dataset pre-extracted and structured from massive amounts of historical engineering data, industry standard documents, and enterprise best practices. Its purpose is to provide deterministic anchors for the generated content of large models, preventing large models from creating the illusion of deviating from engineering reality due to probabilistic sampling. This parameter set typically includes, but is not limited to, a typical HARA (Hazard Analysis and Risk Assessment) scenario library, a whole vehicle hazardous behavior library, a risk level determination criterion table, and a standard safety mechanism library for each functional domain.
[0129] Context correction can be performed after any of steps S110-S150. The analysis results generated in the current step are used as the query object to retrieve corresponding standard entries or reference values from the functional safety experience parameter set. The retrieved parameters are then injected as new context into the prompts of the large model, instructing the model to compare its own output with the standard parameters, thereby correcting or confirming the analysis results.
[0130] For example, in step S140, after the large model generates a dangerous event of "avoiding unexpected acceleration" and the corresponding risk level assessment for "ESC vehicle stability control function", it automatically retrieves the HARA standard scenario library for ESC function in the parameter set to check whether the exposure rate, severity and controllability scores of the model assessment fall within the reasonable range recognized by the industry. If the risk level given by the model is B (lower level), while the parameter set shows that the historical level for similar scenarios is D (higher level), a correction process will be triggered to improve the risk level. The model is required to re-examine the assessment basis or directly adopt the recommended value in the parameter set. This effectively solves the shortcomings of the large model in terms of professional compliance and ensures that the analysis results can be directly connected to the subsequent engineering design process.
[0131] In addition, functional safety analysis methods may also include: performing consistency audits on the currently generated analysis results based on preset audit rules, the analysis results including: failure behaviors and / or hazard events and / or fault propagation logical relationships and functional safety design constraints; when logical conflicts or missing items are detected in the analysis results, re-executing the current step or returning to execute at least the previous step.
[0132] Pre-defined audit rules are real-time verification logic for each step's output stage. They are used to perform preliminary checks on the semantics, causal chains, and data structure completeness and consistency of the output content after each analysis result is generated. Audit rules can include: causal relationship verification rules, such as checking whether the bottom event in the fault tree can truly deduce the top event through logic gates, and whether there are broken causal chains; attribute integrity rules, such as checking whether each generated functional safety requirement includes mandatory fields such as the Automotive Safety Integrity Level (ASIL), allocated components, and time constraints; and cross-step consistency rules, such as checking whether the functional safety design constraints generated in step 150 cover all high-risk hazard events identified in step S140, and whether there are any omissions. Only after the output results of each step pass the corresponding rule verification can they proceed to the next stage of analysis, achieving phased quality control in the streaming output process.
[0133] After all analysis results are generated and verified according to the step-by-step rules, a reflection mechanism based on a large model agent is triggered. Specifically, if logical conflicts or missing items are detected in the analysis results during the review, such as a vehicle control component being assigned mutually exclusive safety mechanisms simultaneously, or a critical event having an undefined failure probability, the system will not directly terminate with an error report. Instead, an automated self-correcting feedback loop will be triggered. The specific step that caused the problem will be located based on the error type, and the error message will be used as a negative feedback prompt. The large model will then be called back to execute that step or revert to the preceding related steps for re-reasoning, until the entire process passes the reflection review, forming a closed-loop iterative optimization mechanism.
[0134] The aforementioned context correction and consistency audit reflection can be executed serially or in parallel, or even embedded into the internal processing loop of each agent to form a real-time quality protection barrier. This invention does not limit this.
[0135] This application also provides an electronic device in its embodiments. (See reference...) Figure 5 The diagram illustrates a structural schematic suitable for implementing the electronic device in the embodiments of this application. The electronic device in the embodiments of this application may include, but is not limited to, fixed terminals such as mobile phones, tablets, large-screen teaching displays, wearable devices, etc. Figure 5 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.
[0136] like Figure 5 As shown, the electronic device may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 1, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 2 or a program loaded from a storage device 8 into a random access memory (RAM) 3, to implement the functional safety analysis method of the foregoing embodiments of this application. When the electronic device is powered on, the RAM 3 also stores various programs and data required for the operation of the electronic device. The processing unit 1, ROM 2, and RAM 3 are interconnected via a bus 4. An input / output (I / O) interface 5 is also connected to the bus 4.
[0137] Typically, the following devices can be connected to I / O interface 5: input devices 6 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 7 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 8 including, for example, memory cards, hard drives, etc.; and communication devices 9. Communication device 9 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 5Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown. More or fewer devices may be implemented or have instead.
[0138] This application also provides a computer program product, including computer-readable instructions, which, when executed on an electronic device, cause the electronic device to implement any of the functional safety analysis methods provided in this application.
[0139] This application also provides a readable storage medium that carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any of the functional safety analysis methods provided in this application.
[0140] It should also be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. In addition, in the device embodiment drawings provided in this application, the connection relationship between modules indicates that they have a communication connection, which can be implemented as one or more communication buses or signal lines.
[0141] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware, or it can be implemented by special-purpose hardware including application-specific integrated circuits, special-purpose CPUs, special-purpose memory, special-purpose components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for this application, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk, or optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, training equipment, or network device, etc.) to execute the methods described in the various embodiments of this application.
[0142] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product.
[0143] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, training device, or data center to another website, computer, training device, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).
[0144] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.
Claims
1. A functional safety analysis method, characterized in that, include: Obtain information about the object to be analyzed, including: the target function and the failure mode of the target function of the target vehicle; Obtain the function execution information corresponding to the target function from the vehicle function specification file stored in the pre-configured functional safety knowledge base; Based on the function execution information and the target function failure mode, determine the failure behavior of the target vehicle when the target function is in the target function failure mode; The hazard event corresponding to the failure behavior in a preset traffic scenario is determined, and the fault propagation logic relationship corresponding to the hazard event is analyzed. The fault propagation logic relationship includes multiple logic nodes, and the logic nodes are used to characterize at least one fault information of the vehicle control component. The fault information has a fault causal relationship with the hazard event. Based on the fault propagation logic relationship, functional safety design constraints are generated for the target function. These functional safety design constraints are used to indicate that the design of the vehicle control component corresponding to the logic node meets preset functional safety conditions.
2. The functional safety analysis method according to claim 1, characterized in that, The process of determining the failure behavior of the target vehicle when the target function is in the target function failure mode, based on the function execution information and the target function failure mode, includes: A pre-configured large model is invoked to instruct the large model to predict the failure behavior of the target vehicle when the target function is in the target function failure mode, based on the function execution information.
3. The functional safety analysis method according to claim 1, characterized in that, The process of determining the failure behavior of the target vehicle when the target function is in the target function failure mode, based on the function execution information and the target function failure mode, includes: Based on the function execution information, determine the target controller type associated with the target function; Retrieve the set of candidate failure behaviors corresponding to the target controller type from the set of failure behaviors corresponding to different controller types stored in the functional safety knowledge base; A pre-configured large model is invoked to instruct the large model to perform failure behavior analysis on the target vehicle based on the candidate failure behavior set and the target function failure mode, so as to obtain the failure behavior of the target vehicle when the target function is in the target function failure mode.
4. The functional safety analysis method according to claim 3, characterized in that, The method further includes: Calculate the semantic similarity between the failure behavior of the target vehicle and each candidate failure behavior in the candidate failure behavior set; If the semantic similarity is lower than a preset threshold, the corresponding failure behavior of the target vehicle will be marked as a behavior pending review.
5. The functional safety analysis method according to claim 1, characterized in that, The process of determining the hazard event corresponding to the failure behavior in a preset traffic scenario includes: Based on the hazardous characteristics of the failure behavior and prior knowledge of functional safety, basic traffic scenarios that are causally related to the failure behavior are selected from preset traffic scenarios. The factors influencing the harmful consequences of the failure behavior are combined to generate at least one scenario variable combination; A pre-configured large model is invoked to instruct the large model to deduce the interaction risk between the target vehicle and external traffic participants under the failure behavior based on the basic traffic scenario and the combination of various scenario variables, and to obtain the corresponding hazard event.
6. The functional safety analysis method according to claim 1, characterized in that, The process of analyzing the fault propagation logic relationship corresponding to the aforementioned hazardous event includes: The hazardous event is taken as the top event, and the primary fault cause that triggered the hazardous event is analyzed based on the functional execution information. For each of the first-level fault causes, the corresponding lower-level fault causes are broken down level by level until the underlying fault causes that cannot be further decomposed or do not need to be further decomposed are obtained. Establish the logical relationship between the hazardous events and the causes of failures at each level to obtain the fault propagation logical relationship corresponding to the hazardous events.
7. The functional safety analysis method according to claim 6, characterized in that, The process of establishing the logical relationship between the hazardous event and the causes of failure at each level, and obtaining the fault propagation logical relationship corresponding to the hazardous event, includes: Based on the logical relationship between the hazardous events and the causes of failures at each level, corresponding logical expressions are generated; Based on the logical expression, the minimum cut set that triggers the occurrence of the hazardous event is determined, and the minimum cut set is used as the fault propagation logical relationship corresponding to the hazardous event.
8. The functional safety analysis method according to any one of claims 1-7, characterized in that, The method further includes: Obtain a set of functional safety experience parameters and perform context correction on the currently generated analysis results. The analysis results include: the failure behavior and / or the hazard event and / or the fault propagation logic relationship, as well as the functional safety design constraints.
9. The functional safety analysis method according to any one of claims 1-7, characterized in that, The method further includes: Based on preset audit rules, the consistency audit of the currently generated analysis results is performed. The analysis results include: the failure behavior and / or the hazard event and / or the fault propagation logic relationship, as well as the functional safety design constraints. If a logical conflict or missing item is detected in the analysis results, the current step is re-executed or the previous one or more steps are returned to be executed.
10. An electronic device, characterized in that, include: Memory and processor; The memory is used to store programs; The processor is configured to execute the program to implement the various steps of the functional safety analysis method as described in any one of claims 1 to 9.
