A method and related device for online diagnosis of a drive-by-wire chassis based on causal reasoning and twin validation

Abstract

The application discloses a drive-by-wire chassis online diagnosis method based on causal reasoning and twin verification and related equipment, which can be applied to the technical field of intelligent networked vehicles. According to the application, after an observation vector sequence is constructed according to multi-source heterogeneous signals and corresponding instantaneous reconstruction errors are calculated, a window abnormal score is calculated and it is determined that an observation vector sequence exists in an abnormal state variable, a target causal relationship matrix is constructed based on historical normal operation data and a first causal parent set is extracted, and then counterfactual reasoning fault root cause positioning is performed to obtain a target fault root cause, so that the accuracy of fault diagnosis can be effectively improved. Then, the digital twin verification is performed on the correction strategy corresponding to the target fault root cause, and the causal weights in the target causal relationship matrix in the fault diagnosis model are updated and optimized in combination with the actual execution result, so that the running stability of the vehicle after control correction is improved, and the long-term adaptability is maintained by continuous optimization during vehicle operation.

Description

Technical Field

[0001] This application relates to the field of intelligent connected vehicle technology, and in particular to an online diagnostic method and related equipment for drive-by-wire chassis based on causal reasoning and twin verification. Background Technology

[0002] With the development of new energy vehicles, drive-by-wire chassis technologies such as EHB (Electro-hydraulic Braking) and SBW (Slewing Braking System) have been widely applied. These systems have high integration and numerous control units, resulting in strong internal coupling. In actual operation, faults often exhibit propagation and diversity, posing challenges to fault diagnosis and handling.

[0003] Currently, fault diagnosis for drive-by-wire chassis systems primarily relies on rule-based or fault code-based diagnostic methods, as well as data-driven model-based methods. These methods suffer from low accuracy and stability when dealing with complex operating conditions, compound faults, or simultaneous occurrences of multiple anomalies. Furthermore, existing fault handling procedures often execute corrective or control measures directly after the diagnostic results are generated, lacking prior assessment of the effectiveness and safety of the corrective strategies, which may affect the stability of system operation. In addition, onboard diagnostic models are typically difficult to dynamically adjust based on actual operating conditions after vehicle delivery, and their adaptability and diagnostic performance may gradually decline over long-term use.

[0004] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0005] The main objective of this application is to propose an online diagnostic method and related equipment for drive-by-wire chassis based on causal reasoning and twin verification. This method can effectively improve the accuracy and stability of fault diagnosis results, enhance the operational stability of the vehicle after control correction, and continuously optimize the system during vehicle operation to maintain long-term adaptability.

[0006] To achieve the above objectives, one aspect of this application proposes an online diagnostic method for drive-by-wire chassis based on causal reasoning and twin verification. The online diagnostic method is executed based on a fault diagnosis model and includes the following steps: Acquire multi-source heterogeneous signals from the target vehicle's drive-by-wire chassis, wherein the multi-source heterogeneous signals originate from different core components on the target vehicle; Construct an observation vector sequence based on the multi-source heterogeneous signals; The instantaneous reconstruction error corresponding to the observation vector sequence is calculated based on a preset encoder; Calculate the window anomaly score based on the instantaneous reconstruction error; When it is determined that there are abnormal state variables in the observation vector sequence based on the window anomaly score, a target causal relationship matrix is ​​constructed based on the historical normal operation data of the target vehicle, and the first causal parent set of all variable observation vectors in the observation vector sequence is extracted based on the target causal relationship matrix. The elements in the target causal relationship matrix are used to characterize the causal influence strength between the signals corresponding to different core components in the target vehicle, and the causal parent set is used to characterize the local causal dependency of the target variable observation vector. Based on the target causal relationship matrix, the abnormal state variables and the first causal parent set, the root cause of the failure is located by counterfactual reasoning of the abnormal state variables to obtain the target root cause of the failure. Digital twin verification is performed on the correction strategy corresponding to the target fault root cause; The causal weights within the target causal relationship matrix of the fault diagnosis model are updated and optimized based on the actual execution results of the correction strategy after verification.

[0007] In some embodiments, constructing the observation vector sequence based on the multi-source heterogeneous signals includes: Time alignment is performed on the multi-source heterogeneous signals; The time-aligned multi-source heterogeneous signals are processed using a sliding time window method to obtain the observation vector sequence.

[0008] In some embodiments, calculating the instantaneous reconstruction error corresponding to the observation vector sequence based on a preset encoder includes: The observation vector sequence is input into the preset encoder to map and reconstruct each original observation vector in the observation vector sequence to obtain a reconstructed output vector; The instantaneous reconstruction error corresponding to each of the original observation vectors is calculated based on the reconstructed output vector and the original observation vector.

[0009] In some embodiments, constructing a target causal relationship matrix based on the historical normal operation data of the target vehicle includes: Analyze the historical normal operation data of the target vehicle to construct an initial causal relationship matrix between variables corresponding to different core components; The initial causal relationship matrix is ​​sparsified to obtain the target causal relationship matrix.

[0010] In some embodiments, the step of locating the root cause of a fault by performing counterfactual reasoning on the abnormal state variables based on the target causal relationship matrix, the abnormal state variables, and the first causal parent set, to obtain the target root cause of the fault, includes: Using the abnormal state variable as the first abnormal node, extract the second abnormal node from the observation vector sequence according to the target causal relationship matrix and extract the second causal parent set corresponding to all target abnormal nodes from the first causal parent set to form a candidate fault root cause node set. The target abnormal node includes the first abnormal node or the second abnormal node. Construct the counterfactual states corresponding to the candidate root cause nodes within the set of candidate root cause nodes; The counterfactual state is compared with the current abnormal state corresponding to the candidate root cause node to determine the degree of influence of the candidate root cause node. The target root cause is determined from the set of candidate root cause nodes based on the degree of impact.

[0011] In some embodiments, determining the target root cause from the set of candidate root cause nodes based on the degree of influence includes: All candidate root cause nodes in the candidate root cause node set are sorted according to the degree of influence. The target root cause is determined from the set of candidate root cause nodes based on the sorting results.

[0012] In some embodiments, the digital twin verification of the correction strategy corresponding to the target fault root cause includes: A correction action is generated based on the target fault root cause and amplitude constraints. The correction strategy is determined based on the correction action and the reference control input; Construct a digital twin model of the target vehicle's drive-by-wire chassis; The correction strategy is input into the digital twin model for simulation to obtain the simulation response. All the simulated response quantities are subjected to safety boundary verification in the prediction time domain.

[0013] To achieve the above objectives, another aspect of this application proposes an online diagnostic device for a drive-by-wire chassis based on causal reasoning and twin verification. The data processing within the device is executed based on a fault diagnosis model. The device includes: The first module is used to acquire multi-source heterogeneous signals from the target vehicle's wire-controlled chassis, wherein the multi-source heterogeneous signals come from different core components on the target vehicle. The second module is used to construct an observation vector sequence based on the multi-source heterogeneous signals; The third module is used to calculate the instantaneous reconstruction error corresponding to the observation vector sequence based on the preset encoder; The fourth module is used to calculate the window anomaly score based on the instantaneous reconstruction error; The fifth module is used to construct a target causal relationship matrix based on the historical normal operation data of the target vehicle when it is determined that there are abnormal state variables in the observation vector sequence according to the window anomaly score, and to extract the first causal parent set of all variable observation vectors in the observation vector sequence according to the target causal relationship matrix. The elements in the target causal relationship matrix are used to characterize the causal influence strength between the corresponding signals of different core components in the target vehicle, and the causal parent set is used to characterize the local causal dependency of the target variable observation vector. The sixth module is used to locate the root cause of the fault by counterfactual reasoning of the abnormal state variables based on the target causal relationship matrix, the abnormal state variables and the first causal parent set, so as to obtain the target root cause of the fault. The seventh module is used to perform digital twin verification of the correction strategy corresponding to the target fault root cause; The eighth module is used to update and optimize the causal weights in the target causal relationship matrix of the fault diagnosis model based on the actual execution results of the correction strategy after verification.

[0014] To achieve the above objectives, another aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described above.

[0015] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods described above.

[0016] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer program product, including a computer program that, when executed by a processor, implements the aforementioned method.

[0017] The embodiments of this application include at least the following beneficial effects: This application provides an online diagnostic method and related equipment for a drive-by-wire chassis based on causal reasoning and twin verification. This scheme is based on a fault diagnosis model. During execution, after acquiring multi-source heterogeneous signals from the drive-by-wire chassis of the target vehicle, an observation vector sequence is constructed based on the multi-source heterogeneous signals. The instantaneous reconstruction error corresponding to the observation vector sequence is calculated based on a preset encoder. Then, a window anomaly score is calculated based on the instantaneous reconstruction error. When it is determined that there are abnormal state variables in the observation vector sequence based on the window anomaly score, a target causal relationship matrix is ​​constructed based on the historical normal operation data of the target vehicle, and the target causal relationship matrix is ​​used to extract... The first causal parent set of all variable observation vectors in the observation vector sequence is used. Then, based on the target causal relationship matrix, abnormal state variables, and the first causal parent set, counterfactual reasoning is performed on the abnormal state variables to locate the root cause of the fault, thereby obtaining the target fault root cause. This can effectively improve the accuracy of fault diagnosis. Next, the correction strategy corresponding to the target fault root cause is verified by digital twin. Based on the actual execution result of the correction strategy after verification, the causal weights in the target causal relationship matrix in the fault diagnosis model are updated and optimized, thereby improving the operational stability of the vehicle after control correction. At the same time, continuous optimization is carried out during vehicle operation to maintain long-term adaptability, which can further improve the accuracy of the fault diagnosis process. Attached Figure Description

[0018] Figure 1 This is a flowchart of the online diagnostic method for drive-by-wire chassis based on causal reasoning and twin verification provided in the embodiments of this application; Figure 2 This is a schematic diagram of the structure of the online diagnostic device for drive-by-wire chassis based on causal reasoning and twin verification provided in the embodiments of this application; Figure 3 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0020] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various concepts, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if,” “when,” or “in response to a determination” as used herein may be interpreted as “when…” or “when…” or “in response to a determination.”

[0021] As used in this application, the terms "at least one", "multiple", "each", "any", etc., "at least one" includes one, two or more, "multiple" includes two or more, "each" refers to each of the corresponding multiples, and "any" refers to any one of the multiples.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0023] In related technologies, existing fault diagnosis for drive-by-wire chassis systems primarily relies on rule-based or fault code-based diagnostic methods, as well as data-driven model-based diagnostic methods. These methods exhibit low accuracy and stability in complex operating conditions, compound faults, or simultaneous occurrences of multiple anomalies. Furthermore, existing fault handling procedures often execute corrective or control measures directly after the diagnostic results are generated, lacking prior assessment of the effectiveness and safety of the corrective strategies, which may affect the stability of system operation. In addition, on-board diagnostic models are typically difficult to dynamically adjust based on actual operating conditions after vehicle delivery, and their adaptability and diagnostic performance may gradually decline over long-term use.

[0024] In view of this, this application provides a method and related equipment for online diagnosis of drive-by-wire chassis based on causal reasoning and twin verification, which can effectively improve the accuracy and stability of fault diagnosis results and improve the operational stability of the vehicle after control correction, while continuously optimizing during vehicle operation to maintain long-term adaptability.

[0025] The online diagnostic method for drive-by-wire chassis based on causal reasoning and twin verification provided in this application relates to the field of intelligent connected vehicle technology. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, or in-vehicle terminal, but is not limited to these. The server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network. The software can be an application implementing the online diagnostic method for drive-by-wire chassis based on causal reasoning and twin verification, but is not limited to the above forms.

[0026] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0027] The embodiments of this application will be described in detail below with reference to the accompanying drawings: Figure 1 This is an optional flowchart of the online diagnostic method for drive-by-wire chassis based on causal reasoning and twin verification provided in the embodiments of this application. Figure 1 The method may include, but is not limited to, steps S110 to S180: Step S110: Obtain multi-source heterogeneous signals from the target vehicle's drive-by-wire chassis. The multi-source heterogeneous signals come from different core components on the target vehicle. Step S120: Construct an observation vector sequence based on multi-source heterogeneous signals; Step S130: Calculate the instantaneous reconstruction error corresponding to the observation vector sequence based on the preset encoder; Step S140: Calculate the window anomaly score based on the instantaneous reconstruction error; Step S150: When it is determined that there are abnormal state variables in the observation vector sequence based on the window anomaly score, construct the target causal relationship matrix based on the historical normal operation data of the target vehicle, and extract the first causal parent set of all variable observation vectors in the observation vector sequence based on the target causal relationship matrix. The elements in the target causal relationship matrix are used to characterize the causal influence strength between the signals corresponding to different core components in the target vehicle, and the causal parent set is used to characterize the local causal dependency of the target variable observation vector. Step S160: Based on the target causal relationship matrix, abnormal state variables and the first causal parent set, perform counterfactual reasoning to locate the root cause of the fault and obtain the target root cause of the fault. Step S170: Perform digital twin verification of the correction strategy corresponding to the root cause of the target fault; Step S180: Update and optimize the causal weights in the target causal relationship matrix of the fault diagnosis model based on the actual execution results of the correction strategy after verification.

[0028] It is understood that the method in this embodiment is based on a fault diagnosis model. This fault diagnosis model can be an intelligent diagnostic system that links a pre-trained large language model with vehicle data. This intelligent diagnostic system can be continuously updated and optimized based on real-time multi-source heterogeneous signals during vehicle application, thereby achieving online evolution throughout the entire life cycle.

[0029] It is understood that the core components of this embodiment may include, but are not limited to, EHB (Electro-Hydraulic Braking System), SBW (Smart Brake Assist System), MCU (Microcontroller Unit), and in-vehicle CANFD (Controller Area Network Fast Transfer) network. Specifically, after obtaining the multi-source heterogeneous signals, this embodiment performs time alignment on the multi-source heterogeneous signals, and then uses a sliding time window method to process the time-aligned multi-source heterogeneous signals to obtain the observation vector sequence. The observation vector sequence is given by the following formula: ; In the formula, This indicates the length of the sliding window, corresponding to the length of historical data within a preset time range; This represents the total number of observable nodes, which include both physical and logical state variables in the drive-by-wire chassis system. Indicates at time Multi-source synchronous observation vector.

[0030] It is understood that in this embodiment, after the observation vector sequence, the observation vector sequence can be input into a preset encoder to map and reconstruct each original observation vector in the observation vector sequence to obtain a reconstructed output vector. Then, the instantaneous reconstruction error corresponding to each original observation vector is calculated based on the reconstructed output vector and the original observation vector. Specifically, in this embodiment, the original observation vector refers to the element in the observation vector sequence. The preset encoder in this embodiment can be a shallow autoencoder calibrated on the production line, and its forward reconstruction process can be expanded into two layers of nonlinear transformation: ; ; In the formula, This is the input observation vector at the current time t; Hidden layer features; , It is a fixed weight matrix; , It is the bias vector; It is the ReLU activation function, i.e. .

[0031] The reconstructed output vector obtained by the above mapping reconstruction With the original observation vector The difference between them is used to characterize the degree of abnormality in the vehicle system's operating state, and its instantaneous reconstruction error is as follows: ; In the formula, Operators that represent Euclidean norms.

[0032] By extracting abnormal features and calculating reconstruction errors as described above, the degree of deviation of the multi-source state of the drive-by-wire chassis in normal mode can be quantified, providing a basic indicator for subsequent abnormal scoring and diagnostic triggering.

[0033] Understandably, to avoid transient noise interfering with the anomaly assessment results, this embodiment uses a sliding window method to statistically generate a window anomaly score for the transient reconstruction error. Specifically, this embodiment uses a sliding window method with a length of... Within a sliding window, the mean of the instantaneous reconstruction error is calculated to obtain a window anomaly score: ; In the formula, Indicates time The corresponding average window reconstruction error; This represents the i-th instantaneous reconstruction error among all instantaneous reconstruction errors; This indicates the length of the sliding window, which should be consistent with the observation window mentioned above.

[0034] This embodiment models the statistical distribution of window anomaly scores based on historical normal operation data, and the anomaly triggering judgment condition is defined as follows: ; In the formula, This represents the expected value of the average window reconstruction error under normal operating conditions. This represents the corresponding standard deviation; This is a configurable sensitivity coefficient used to balance the false alarm rate and the false negative rate.

[0035] When the window anomaly score meets the above judgment conditions, it indicates that there is an abnormal state vector in the current observation vector sequence. Therefore, this embodiment constructs a target causal relationship matrix based on the historical normal operation data of the target vehicle, and extracts the first causal parent set of all variable observation vectors in the observation vector sequence according to the target causal relationship matrix. Specifically, the construction process of the target causal relationship matrix in this embodiment can be achieved by analyzing the historical normal operation data of the target vehicle to construct an initial causal relationship matrix between the corresponding variables of different core components, and then performing sparsification processing on the initial causal relationship matrix to obtain the target causal relationship matrix. The initial causal relationship matrix is ​​as follows: ; In the formula, Representing variables right The strength of causal influence; when When the time is right, it means there is no direct causal relationship between the two.

[0036] In this embodiment, the causal relationship modeling constrains the causal structure to prevent circular dependencies during the construction process, ensuring that the obtained causal relationships meet the requirements of a directed acyclic structure. Among these, variables... It represents an observable state quantity in the drive-by-wire chassis system, corresponding to physical or logical quantities from various subsystems of the chassis.

[0037] To reduce the structural complexity of the causal relationship matrix while meeting the requirements of real-time vehicle computing, this embodiment performs sparsification on the initial causal relationship matrix to retain causal connections that have a major impact on system state changes. Specifically, the sparsification process in this embodiment involves performing a threshold filtering operation on each element in the causal weight matrix: ; In the formula, A preset threshold is used to filter out causal connections with a relatively small impact.

[0038] This embodiment uses sparsification to retain only a limited number of effective connection weights in the target causal relationship matrix, thereby reducing the scale and complexity of subsequent inference calculations.

[0039] It is understood that, after processing the target causal relationship matrix, this embodiment uses the set of upstream causal variables directly associated with each observable variable extractor in the observation vector sequence based on the target causal relationship matrix to characterize the local causal dependency of the observable variables.

[0040] For any observable variable Its corresponding first causal parent set is defined as: ; In the formula, Represents the observable variable The set of variables that have a direct causal effect.

[0041] Understandably, this embodiment performs root cause analysis on the abnormal state variables after obtaining the target causal relationship matrix and the first causal parent set corresponding to all observable variables. This embodiment can construct counterfactual states to assess the influence of different candidate variables on system anomalies, thereby achieving quantitative localization of the root cause of the failure. Specifically, this embodiment can use the abnormal state variables as the first abnormal node, extract the second abnormal node from the observation vector sequence based on the target causal relationship matrix, and extract the second causal parent set corresponding to all target abnormal nodes from the first causal parent set to form candidate root cause nodes. A set. The target abnormal node includes either the first abnormal node or the second abnormal node. Candidate root cause nodes. The physical or logical state quantities that may cause abnormal states in the corresponding drive-by-wire chassis system are used to apply counterfactual interventions in subsequent steps and assess their impact on system abnormalities.

[0042] For any candidate root cause node This embodiment constructs counterfactual states corresponding to candidate root cause nodes within a set of candidate root cause nodes to simulate the system's possible operating state when the variable does not exhibit anomalies. Then, the counterfactual states are compared with the current anomalous states corresponding to the candidate root cause nodes to determine the degree of influence of the candidate root cause nodes. Specifically, this embodiment uses candidate root cause nodes... Replace the current abnormal value with its reference value within the normal operating range. And keeping the values ​​of the other state variables unchanged, thus obtaining the corresponding counterfactual state vector: ; In the formula, Determined based on historical normal operation data or system calibration information.

[0043] Then, the counterfactual state is compared with the current abnormal state corresponding to the candidate root cause node to determine the influence degree of the candidate root cause node. Based on the influence degree, the target root cause is determined from the set of candidate root cause nodes. Specifically, this embodiment compares the counterfactual state with the current abnormal state to evaluate the degree of improvement of key system performance indicators, thereby quantifying the influence degree of the candidate root cause node on system anomalies. Here, candidate root cause nodes are defined. degree of influence for: ; In the formula, Key performance indicators under abnormal conditions; This indicates the performance metrics corresponding to a counterfactual state.

[0044] After sorting all candidate root cause nodes in the candidate root cause node set according to their degree of influence, the target root cause is determined from the candidate root cause node set based on the sorting results. For example, the candidate root cause node with the highest degree of influence can be selected as the high-confidence target root cause or a high-confidence root cause set can be formed.

[0045] Understandably, this embodiment verifies the safety and effectiveness of the correction strategy generated based on the target fault root cause before performing correction actions on the vehicle. This embodiment can use a digital twin model to perform forward simulation of the correction strategy, predicting its impact on key system state variables and performance indicators, thereby avoiding new safety risks caused by blindly executing correction actions. Specifically, this embodiment can generate correction actions based on the target fault root cause and amplitude constraints, determine the correction strategy based on the correction actions and baseline control inputs, and simultaneously construct a digital twin model corresponding to the target vehicle's drive-by-wire chassis. The correction strategy is input into the digital twin model for simulation to obtain the simulation response, and then all simulation response quantities are checked for safety boundaries in the prediction time domain.

[0046] In this embodiment, the correction strategy within the preset time domain is represented as a correction amount to the current control input: ; In the formula, This serves as the baseline control input for the current operating condition. Corrective actions are generated in response to the root cause of the fault.

[0047] To avoid generating unexecutable correction strategies, this embodiment imposes amplitude constraints on the correction actions: ; This embodiment adjusts the method based on the type of root cause of the fault. The range or variation of values ​​is used to form multiple candidate correction strategies, and these multiple candidate correction strategies are used as inputs for subsequent digital twin verification steps.

[0048] This embodiment constructs a lightweight digital twin model corresponding to the dynamic characteristics of the key subsystems of the drive-by-wire chassis to simulate the evolution of the system state under corrective actions. The lightweight digital twin model adopts a discrete state update format: ; In the formula, Indicates the simulation time The system status, This is the corrected control input.

[0049] In this embodiment, all candidate correction strategies are input into the digital twin model, and forward simulation calculations are performed in the prediction time domain at discrete time steps: ; In the formula, Indicates the first The system state at each simulation moment This is the control input after applying the correction action.

[0050] This embodiment calculates key response quantities for evaluating the correction effect and safety based on the state trajectory obtained from simulation, which are obtained by mapping the system state: ; In the formula, Indicates the first Key performance indicators at each simulation moment.

[0051] Subsequently, safety boundary checks are performed on the critical response quantities throughout the entire prediction time domain. When the critical response quantities meet the preset safety constraints at all simulation times, the corresponding correction strategy is determined to be an executable strategy; if the safety constraints are violated at any time, the correction strategy is determined to be unexecutable and is discarded.

[0052] It is understood that after the verification and correction strategy is passed, this embodiment obtains the corresponding actual execution results, and corrects the causal weights in the target causal relationship matrix in the fault diagnosis model online based on the actual execution results, forming a closed-loop mechanism of diagnosis, correction and update, and adjusts the local causal connection strength through correction feedback to maintain the long-term accuracy and adaptability of the diagnosis model.

[0053] Specifically, in this embodiment, after obtaining the actual execution results corresponding to the verified correction strategy, key performance indicators are obtained from the actual execution results. When key performance indicators If the system recovers to the preset normal range, the correction is considered successful; otherwise, the correction is considered unsuccessful. Supervision labels are constructed based on the actual execution results. ; In the formula, Used to characterize the effectiveness of the current "root cause-correction strategy" combination in real vehicles.

[0054] This embodiment is based on the supervision label. The weights of causal connections corresponding to the current root cause node and its related variables are locally updated. When the correction is successful, the weights of the relevant causal connections are increased. When the correction fails, reduce the weight of the relevant causal connections. The update magnitude is determined by a preset update step size. To ensure the stability of the model update process, this embodiment performs an update after each update. Perform amplitude limiting to keep it within the preset range. This avoids excessive increases or decreases in weights, which could lead to model divergence.

[0055] It is understood that after completing the local weight update, this embodiment can use the updated fault diagnosis model for subsequent anomaly detection and root cause localization, and repeatedly execute the method of this application during vehicle operation, thereby realizing the continuous online self-evolution of the fault diagnosis model throughout its entire life cycle.

[0056] As described above, the method in this embodiment improves the accuracy and interpretability of fault root cause localization by introducing an online diagnostic process based on causal reasoning; it performs pre-simulation verification of correction actions before actual execution by combining a correction strategy verification process with a digital twin model; and it establishes an online update mechanism for the diagnostic model based on correction feedback, enabling the fault diagnosis model to continuously optimize and maintain long-term adaptability during vehicle operation. Therefore, the method in this embodiment effectively improves the accuracy and stability of fault diagnosis results, enhances the operational stability of the vehicle after control correction, and continuously optimizes during vehicle operation to maintain long-term adaptability.

[0057] Please see Figure 2 This application also provides an online diagnostic device for a drive-by-wire chassis based on causal reasoning and twin verification. The data processing within the device is executed based on a fault diagnosis model. The device includes: The first module is used to acquire multi-source heterogeneous signals from the target vehicle's drive-by-wire chassis, wherein the multi-source heterogeneous signals come from different core components on the target vehicle. The second module is used to construct observation vector sequences based on multi-source heterogeneous signals; The third module is used to calculate the instantaneous reconstruction error corresponding to the observation vector sequence based on the preset encoder; The fourth module is used to calculate the window anomaly score based on the instantaneous reconstruction error; The fifth module is used to construct a target causal relationship matrix based on the historical normal operation data of the target vehicle when it is determined that there are abnormal state variables in the observation vector sequence according to the window anomaly score. The first causal parent set of all variable observation vectors in the observation vector sequence is extracted according to the target causal relationship matrix. The elements in the target causal relationship matrix are used to characterize the causal influence strength between the corresponding signals of different core components in the target vehicle, and the causal parent set is used to characterize the local causal dependency of the target variable observation vector. The sixth module is used to locate the root cause of the fault by counterfactual reasoning of the abnormal state variables based on the target causal relationship matrix, abnormal state variables and the first causal parent set, so as to obtain the target fault root cause. The seventh module is used to perform digital twin verification of the correction strategy corresponding to the root cause of the target failure; The eighth module is used to update and optimize the causal weights in the target causal relationship matrix in the fault diagnosis model based on the actual execution results of the correction strategy after verification.

[0058] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0059] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0060] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0061] Please see Figure 3 , Figure 3 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 310 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory 320 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 320 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 320 and is called and executed by the processor 310 using the methods described in the embodiments of this application. Input / output interface 330 is used to realize information input and output; The communication interface 340 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 350 transmits information between various components of the device (e.g., processor 310, memory 320, input / output interface 330, and communication interface 340); The processor 310, memory 320, input / output interface 330 and communication interface 340 are connected to each other within the device via bus 350.

[0062] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0063] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0064] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0065] It is understood that the content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0066] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0067] This application provides an online diagnostic method and related equipment for drive-by-wire chassis based on causal reasoning and twin verification. By modeling and analyzing the causal relationships between multi-source state variables in the drive-by-wire chassis system, it achieves accurate location of the root cause of faults in complex fault scenarios, avoiding misjudging propagation anomalies as fundamental faults. By introducing digital twin simulation verification before the actual execution of the correction strategy, the effectiveness and safety of the correction strategy are evaluated, reducing the adverse effects of improper correction actions on system stability and operational safety. Simultaneously, the actual execution results of the correction strategy are used to correct the diagnostic model online, enabling the diagnostic model to continuously adjust and optimize as the vehicle operates, thereby maintaining stable diagnostic performance and good adaptability throughout the vehicle's entire lifecycle.

[0068] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0069] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0070] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; 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.

[0071] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0072] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0073] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0074] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0075] The units described above as separate components may or may not be physically separate. 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 units can be selected to achieve the purpose of this embodiment according to actual needs.

[0076] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0077] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0078] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A method for online diagnosis of drive-by-wire chassis based on causal reasoning and twin verification, characterized in that, The online diagnostic method is based on a fault diagnosis model and includes the following steps: Acquire multi-source heterogeneous signals from the target vehicle's drive-by-wire chassis, wherein the multi-source heterogeneous signals originate from different core components on the target vehicle; Construct an observation vector sequence based on the multi-source heterogeneous signals; The instantaneous reconstruction error corresponding to the observation vector sequence is calculated based on a preset encoder; Calculate the window anomaly score based on the instantaneous reconstruction error; When it is determined that there are abnormal state variables in the observation vector sequence based on the window anomaly score, a target causal relationship matrix is ​​constructed based on the historical normal operation data of the target vehicle, and the first causal parent set of all variable observation vectors in the observation vector sequence is extracted based on the target causal relationship matrix. The elements in the target causal relationship matrix are used to characterize the causal influence strength between the signals corresponding to different core components in the target vehicle, and the causal parent set is used to characterize the local causal dependency of the target variable observation vector. Based on the target causal relationship matrix, the abnormal state variables and the first causal parent set, the root cause of the failure is located by counterfactual reasoning of the abnormal state variables to obtain the target root cause of the failure. The correction strategy corresponding to the target fault root cause is verified by digital twin; The causal weights within the target causal relationship matrix of the fault diagnosis model are updated and optimized based on the actual execution results of the correction strategy after verification.

2. The method according to claim 1, characterized in that, The step of constructing the observation vector sequence based on the multi-source heterogeneous signals includes: Time alignment is performed on the multi-source heterogeneous signals; The time-aligned multi-source heterogeneous signals are processed using a sliding time window method to obtain the observation vector sequence.

3. The method according to claim 1, characterized in that, The calculation of the instantaneous reconstruction error corresponding to the observation vector sequence based on the preset encoder includes: The observation vector sequence is input into the preset encoder to map and reconstruct each original observation vector in the observation vector sequence, thereby obtaining a reconstructed output vector; The instantaneous reconstruction error corresponding to each of the original observation vectors is calculated based on the reconstructed output vector and the original observation vector.

4. The method according to claim 1, characterized in that, The construction of the target causal relationship matrix based on the historical normal operation data of the target vehicle includes: Analyze the historical normal operation data of the target vehicle to construct an initial causal relationship matrix between variables corresponding to different core components; The initial causal relationship matrix is ​​sparsified to obtain the target causal relationship matrix.

5. The method according to claim 1, characterized in that, The step of locating the root cause of a fault by performing counterfactual reasoning on the abnormal state variables based on the target causal relationship matrix, the abnormal state variables, and the first causal parent set, to obtain the target root cause of the fault, includes: Using the abnormal state variable as the first abnormal node, extract the second abnormal node from the observation vector sequence according to the target causal relationship matrix and extract the second causal parent set corresponding to all target abnormal nodes from the first causal parent set to form a candidate fault root cause node set. The target abnormal node includes the first abnormal node or the second abnormal node. Construct the counterfactual states corresponding to the candidate root cause nodes within the set of candidate root cause nodes; The counterfactual state is compared with the current abnormal state corresponding to the candidate root cause node to determine the degree of influence of the candidate root cause node. The target root cause is determined from the set of candidate root cause nodes based on the degree of impact.

6. The method according to claim 5, characterized in that, Determining the target root cause from the set of candidate root cause nodes based on the degree of influence includes: All candidate root cause nodes in the candidate root cause node set are sorted according to the degree of influence. The target root cause is determined from the set of candidate root cause nodes based on the sorting results.

7. The method according to claim 1, characterized in that, The digital twin verification of the correction strategy corresponding to the target fault root cause includes: A correction action is generated based on the target fault root cause and amplitude constraints. The correction strategy is determined based on the correction action and the reference control input; Construct a digital twin model of the target vehicle's drive-by-wire chassis; The correction strategy is input into the digital twin model for simulation to obtain the simulation response. All the simulated response quantities are subjected to safety boundary verification in the prediction time domain.

8. An online diagnostic device for a drive-by-wire chassis based on causal reasoning and twin verification, characterized in that, The data processing within the device is executed based on a fault diagnosis model, and the device includes: The first module is used to acquire multi-source heterogeneous signals from the target vehicle's wire-controlled chassis, wherein the multi-source heterogeneous signals come from different core components on the target vehicle. The second module is used to construct an observation vector sequence based on the multi-source heterogeneous signals; The third module is used to calculate the instantaneous reconstruction error corresponding to the observation vector sequence based on the preset encoder; The fourth module is used to calculate the window anomaly score based on the instantaneous reconstruction error; The fifth module is used to construct a target causal relationship matrix based on the historical normal operation data of the target vehicle when it is determined that there are abnormal state variables in the observation vector sequence according to the window anomaly score, and to extract the first causal parent set of all variable observation vectors in the observation vector sequence according to the target causal relationship matrix. The elements in the target causal relationship matrix are used to characterize the causal influence strength between the corresponding signals of different core components in the target vehicle, and the causal parent set is used to characterize the local causal dependency of the target variable observation vector. The sixth module is used to locate the root cause of the fault by counterfactual reasoning of the abnormal state variables based on the target causal relationship matrix, the abnormal state variables and the first causal parent set, so as to obtain the target root cause of the fault. The seventh module is used to perform digital twin verification of the correction strategy corresponding to the target fault root cause; The eighth module is used to update and optimize the causal weights in the target causal relationship matrix of the fault diagnosis model based on the actual execution results of the correction strategy after verification.

9. An electronic device, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images