Fault diagnosis method and device and computer readable storage medium

Abstract

The invention discloses a fault diagnosis method and device and a computer readable storage medium. The fault diagnosis method comprises the steps that hardware layer data, system layer data and application layer data are acquired; based on the hardware layer data, the system layer data and the application layer data, constructing fusion feature information; performing fault diagnosis based on the fused feature information to obtain fault diagnosis information; and generating a fault processing strategy based on the fault diagnosis information. According to the method, the multi-dimensional data of the hardware layer, the system layer and the application layer are acquired, the fusion feature information is constructed to perform fault diagnosis and generate the processing strategy, so that the performance of the fault at different levels can be comprehensively captured, the accuracy of positioning the fault reason is improved, missed diagnosis and misdiagnosis are reduced, the comprehensiveness and accuracy of fault diagnosis are improved, and the fault diagnosis efficiency is improved. Meanwhile, a user can conveniently and autonomously solve faults.

Description

Technical Field

[0001] This application relates to the field of fault detection technology, specifically to a fault diagnosis method, apparatus, and computer-readable storage medium. Background Technology

[0002] Current smart terminals utilize voice technology to achieve human-computer interaction via voice. This involves the user issuing voice commands to the terminal, which then uses speech recognition technology to convert the user's voice data captured by the microphone into text. The text is then semantically analyzed and processed, and the results are finally displayed on the screen or read aloud to the user. However, current smart terminal voice fault diagnosis relies on manual log analysis, and the accuracy and efficiency of fault diagnosis need improvement. Summary of the Invention

[0003] This application provides a fault diagnosis method, apparatus, and computer-readable storage medium, which improves the comprehensiveness and accuracy of fault diagnosis, and facilitates users to resolve faults independently.

[0004] The technical solution adopted by this invention to solve the problem is as follows:

[0005] In a first aspect, embodiments of this application provide a fault diagnosis method, comprising: acquiring hardware layer data, system layer data, and application layer data; constructing fused feature information based on the hardware layer data, the system layer data, and the application layer data; performing fault diagnosis based on the fused feature information to obtain fault diagnosis information; and generating a fault handling strategy based on the fault diagnosis information.

[0006] In some embodiments, constructing fused feature information based on the hardware layer data, the system layer data, and the application layer data includes: aligning the hardware layer data, the system layer data, and the application layer data to obtain aligned data information; extracting features from the aligned data information to obtain initial feature information and derived features corresponding to each original feature in the initial feature information; and fusing each original feature in the initial feature information and the derived features corresponding to that original feature to obtain fused feature information.

[0007] In some embodiments, the step of fusing each original feature and its corresponding derived feature in the initial feature information to obtain fused feature information includes: performing context encoding on the initial feature information to obtain a context feature corresponding to each original feature; performing calculation on each original feature and its corresponding context feature to obtain a first utility value corresponding to each original feature; performing calculation on each derived feature and its corresponding context feature to obtain a second utility value corresponding to each original feature; performing calculation on the first utility value and the second utility value to obtain first weight information and second weight information of each original feature and its corresponding derived feature; and performing weighted calculation on each original feature and its corresponding derived feature based on the first weight information and the second weight information to obtain the fused feature information.

[0008] In some embodiments, the step of performing fault diagnosis based on the fused feature information to obtain fault diagnosis information includes: acquiring problem description data; inputting the fused feature information and the problem description data into a target processing model, and outputting fault description information through the target processing model; matching the hardware layer data, the system layer data, and the application layer data with data in a pre-stored fault tree to obtain the cause of the fault; performing similarity matching on the fused feature information and fault case library data to obtain fault case information; and generating the fault diagnosis information based on the fault description information, the cause of the fault, and the fault case information.

[0009] In some embodiments, matching the hardware layer data, the system layer data, and the application layer data with data in a pre-stored fault tree to obtain the cause of the fault includes: matching node indicators in the pre-stored fault tree based on the hardware layer data, the system layer data, and the application layer data to obtain a target node indicator; and outputting the cause of the fault based on the target node indicator.

[0010] In some embodiments, the fused feature information includes a time-series feature to be diagnosed and a feature vector to be diagnosed, and the fault case database data includes historical time-series features and historical feature vectors; the step of performing similarity matching on the fused feature information and the fault case database data to obtain fault case information includes: calculating the similarity between the time-series feature to be diagnosed and the historical time-series feature to obtain first similarity information; calculating the similarity between the feature vector to be diagnosed and the historical feature vector to obtain second similarity information; performing calculation processing on the first similarity information and the second similarity information to obtain target similarity information; and determining the fault case information based on the target similarity information.

[0011] In some embodiments, generating a fault handling strategy based on the fault diagnosis information includes: confirming the fault type based on the fault diagnosis information; and generating a fault handling strategy based on the fault type.

[0012] In some embodiments, the fault diagnosis method further includes: when a user performs an operation based on the fault handling strategy and the operation fails, acquiring operation video data corresponding to the user; performing operation recognition based on the operation video data to obtain operation recognition result information of the user; if the operation recognition result information is first result information, generating prompt information; the first result information indicates that the user has an operation abnormality; if the operation recognition result information is second result information, generating an abnormality log based on the fault diagnosis information, and generating maintenance tool information and a spare parts list based on the abnormality log, so that the user can perform fault handling based on the maintenance tool information and the spare parts list.

[0013] Secondly, embodiments of this application provide a fault diagnosis device, comprising: a first acquisition module for acquiring hardware layer data, system layer data, and application layer data; an information construction module for constructing fused feature information based on the hardware layer data, the system layer data, and the application layer data; a fault diagnosis module for performing fault diagnosis based on the fused feature information to obtain fault diagnosis information; and a strategy generation module for generating a fault handling strategy based on the fault diagnosis information.

[0014] Thirdly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, the computer program being loaded by a processor to perform the steps in the fault diagnosis method described above.

[0015] The beneficial effects of this application are as follows: This application provides a fault diagnosis method. It acquires hardware-layer data, system-layer data, and application-layer data; then constructs fused feature information based on these data; finally, it performs fault diagnosis based on the fused feature information to obtain fault diagnosis information; and finally, it generates a fault handling strategy based on the fault diagnosis information. This application, by acquiring multi-dimensional data from the hardware, system, and application layers and constructing fused feature information for fault diagnosis and strategy generation, can comprehensively capture the manifestation of faults at different levels, improve the accuracy of fault location, reduce missed diagnoses and misdiagnoses, and enhance the comprehensiveness and accuracy of fault diagnosis. It also facilitates user self-troubleshooting. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A schematic flowchart of an embodiment of the fault diagnosis method provided in this application;

[0018] Figure 2 A schematic flowchart illustrating a specific embodiment of step S2 provided in this application;

[0019] Figure 3 A schematic flowchart illustrating a specific embodiment of step S3 provided in this application;

[0020] Figure 4 A schematic flowchart of another embodiment of the fault diagnosis method provided in this application;

[0021] Figure 5 This is a schematic block diagram of one embodiment of the fault diagnosis device provided in this application.

[0022] Figure 6 This is a schematic diagram of an embodiment of the computer device provided in this application. Detailed Implementation

[0023] 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.

[0024] In the description of this application, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more features.

[0025] In this application, the term "exemplary" is used to mean "used as an example, illustration, or description." Any embodiment described as "exemplary" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use this application. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that this application can be made without using these specific details. In other instances, well-known structures and processes are not described in detail to avoid obscuring the description of this application with unnecessary detail. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0026] It should be noted that since the method in this application embodiment is executed in a computer device, the processing objects of each computer device exist in the form of data or information, such as time, which is essentially time information. It is understood that if size, quantity, position, etc. are mentioned in subsequent embodiments, they are all corresponding data that exist so that the computer device can process them. Specific details will not be elaborated here.

[0027] This application provides a fault diagnosis method, apparatus, and computer-readable storage medium. The fault diagnosis method includes: acquiring hardware layer data, system layer data, and application layer data; constructing fused feature information based on the hardware layer data, system layer data, and application layer data; performing fault diagnosis based on the fused feature information to obtain fault diagnosis information; and generating a fault handling strategy based on the fault diagnosis information to improve the comprehensiveness and accuracy of fault diagnosis, while facilitating users to resolve faults independently.

[0028] like Figure 1 The diagram shown is a flowchart of an embodiment of the fault diagnosis method in this application. The fault diagnosis method may include the following steps S1 to S4:

[0029] Step S1: Obtain hardware layer data, system layer data, and application layer data.

[0030] In this embodiment, hardware layer data is used to characterize the operating status of the physical device, including microphone array waveform data, etc. Microphone array waveform data is used to characterize acoustic signal quality, environmental noise characteristics, and hardware health status, etc. Microphone waveforms are used to determine if the quality of the user's input audio is abnormal; to identify environmental noise interference (such as acoustic feedback from the TV's own speakers), and abnormal signal-to-noise ratio caused by microphone hardware damage, etc.

[0031] System-level data is used to characterize the health status of the software runtime environment. System-level data includes process resource usage heatmaps, kernel exception events, and TV operating system version upgrade records. The process resource usage heatmap characterizes the resource usage of the voice processing process. Kernel exception events refer to (TV) operating system exception events, such as operating system-level crashes (e.g., blue screens, system freezes). TV operating system version upgrade records are a systematic record of relevant information during the process of updating the smart TV operating system from an older version to a newer version.

[0032] Application layer data is used to characterize the execution efficiency of application processes. Application layer data includes ASR (Automatic Speech Recognition) service response sequence diagrams, etc. An ASR service response sequence diagram can include request time, request body content, response time, and response body content.

[0033] Optionally, hardware layer data, system layer data, and application layer data can be obtained in real time through the device SDK (Software Development Kit).

[0034] Step S2: Construct fused feature information based on hardware layer data, system layer data, and application layer data.

[0035] In this embodiment, the fused feature information is feature information determined based on hardware layer data, system layer data, and application layer data. This embodiment determines the fused feature information based on hardware layer data, system layer data, and application layer data, ensuring that the obtained fused feature information includes information from these three layers, thus improving the comprehensiveness and accuracy of fault detection.

[0036] In some embodiments, such as Figure 2 As shown, the construction of fused feature information based on hardware layer data, system layer data, and application layer data in step S2 above may include the following steps S21 to S23:

[0037] Step S21: Align the hardware layer data, system layer data, and application layer data to obtain aligned data information.

[0038] In this embodiment, the alignment data information is data obtained by aligning hardware layer data, system layer data, and application layer data. This embodiment aligns hardware layer data, system layer data, and application layer data to obtain alignment data information, ensuring that the obtained alignment data information includes information from hardware layer data, system layer data, and application layer data, and that all information is synchronized, facilitating subsequent analysis and processing.

[0039] In some embodiments, hardware layer data, system layer data, and application layer data are aligned to obtain aligned data information. Specifically, this may include using a sliding window mechanism to align hardware layer data, system layer data, and application layer data in time to obtain aligned data information.

[0040] In this embodiment, hardware layer data, system layer data, and application layer data may be data in different formats, with different sampling frequencies, and from different sources. For example, microphone array waveform data may be high-frequency sampled time-series data, kernel exception events may be event-driven text logs, and process resource usage heatmaps may be periodically sampled numerical data. Because these data have different generation mechanisms and recording methods, they may be out of sync in terms of timestamps.

[0041] Furthermore, the sliding window mechanism is a time-series data processing method used to dynamically adjust the time window in a continuous data stream to reconcile the time differences between different data sources. For example, microphone array waveform data may be acquired at millisecond-level frequencies, while kernel exception events may be recorded at second-level frequencies. The sliding window mechanism ensures that different data sources are aligned in the time dimension by dynamically adjusting the window size and overlap rate.

[0042] Step S22: Extract features from the aligned data information to obtain initial feature information and derived features corresponding to each original feature in the initial feature information.

[0043] In this embodiment, the initial feature information is characteristic information obtained by feature extraction from the aligned data information. The initial feature information includes multiple raw features, which are basic data features extracted directly from the aligned data information without processing. They are direct reflections of the attributes of the data itself, such as the microphone real-time signal-to-noise ratio (SNR) and the instantaneous CPU (Central Processing Unit) occupancy rate.

[0044] Each original feature corresponds to one or more derived features. Derived features are new features obtained based on the original features through mathematical operations, statistical analysis, or logical transformations. They are used to reflect the patterns or states behind the data in greater depth. For example, the "average CPU utilization rate within 5 minutes" is calculated based on the "instantaneous CPU utilization rate" (original feature), and the "signal-to-noise ratio fluctuation standard deviation" is obtained based on the "microphone signal-to-noise ratio" (original feature).

[0045] Step S23: Perform fusion processing on each original feature and the corresponding derived feature in the initial feature information to obtain fused feature information.

[0046] In this embodiment, the fusion process specifically involves dynamically weighting and fusing the original features and derived features using a context-aware attention mechanism. The fused feature information in this embodiment is obtained by fusing each original feature and its corresponding derived feature from the initial feature information. This allows the fused feature information to retain the physical meaning of the original features while adaptively highlighting fault-related derived features, thereby improving the accuracy of fault detection.

[0047] In some embodiments, the method of fusing each original feature and the corresponding derived feature in the initial feature information to obtain fused feature information specifically includes the following steps S231 to S235:

[0048] Step S231: Context encoding is performed on the initial feature information to obtain the context features corresponding to each original feature.

[0049] In the embodiments of this application, contextual features are the semantic representation of each original feature in the context of global information, as well as the temporal dependencies and cross-layer associations between features.

[0050] In some embodiments, the method for performing context encoding on initial feature information to obtain context features corresponding to each original feature specifically includes: obtaining a preset neural network model and a time window length; performing calculations based on the preset neural network model, the time window length, and the initial feature information to obtain context features corresponding to each original feature.

[0051] Optionally, the process of performing context encoding on the initial feature information to obtain the context features corresponding to each original feature can be represented as: C(t,m,k)=LSTM(M[t-τ:t,m,k]), C(t,m,k)∈R d Where C(t,m,k) represents the context features; M[t,m,k] represents the initial feature information; τ = 100ms, where τ is the time window length; LSTM (Long Short-Term Memory) represents the preset neural network model, which is a special type of recurrent neural network designed specifically to solve long-term dependency problems.

[0052] Step S232: Calculate and process each original feature and its corresponding context feature to obtain the first utility value for each original feature.

[0053] In the embodiments of this application, the first utility value is used to characterize the degree of matching between the original feature and its context, and to quantify the contribution of the original feature to fault diagnosis.

[0054] In some embodiments, the step of calculating and processing each original feature and the context feature corresponding to the original feature to obtain a first utility value corresponding to each original feature may specifically include: obtaining a first parameter matrix and a first bias term; calculating and processing the original feature, the context feature corresponding to the original feature, the first parameter matrix and the first bias term to obtain a first utility value corresponding to each original feature.

[0055] Optionally, the process of calculating and processing each original feature and its corresponding context features to obtain the first utility value for each original feature can be represented as follows: Among them, S raw Indicates the first utility value; C represents the first parameter matrix; t,m,k Represents contextual features; f raw (t,m,k) represents the original features; b1 represents the first bias term; This indicates a vector concatenation operation.

[0056] Step S233: Calculate and process the derived features corresponding to each original feature and the context features corresponding to the original feature to obtain the second utility value corresponding to each original feature.

[0057] In the embodiments of this application, the second utility value is used to characterize the degree of matching between the derived feature and its context, and to quantify the contribution of the derived feature to fault diagnosis.

[0058] In some embodiments, the step of calculating and processing the derived features corresponding to each original feature and the context features corresponding to the original feature to obtain the second utility value corresponding to each original feature may specifically include: obtaining the second parameter matrix and the second bias term; calculating and processing the derived features corresponding to each original feature, the context features corresponding to the original feature, the second parameter matrix and the second bias term to obtain the second utility value corresponding to each original feature.

[0059] Optionally, the process of calculating the derived features corresponding to each original feature and the context features corresponding to that original feature to obtain the second utility value corresponding to each original feature can be represented as follows: Among them, S derived Indicates the second utility value; C represents the second parameter matrix; t,m,k Represents contextual features; f derived (t,m,k) represents the derived feature; b2 represents the second bias term; This indicates a vector concatenation operation.

[0060] Step S234: Calculate and process the first utility value and the second utility value to obtain the first weight information corresponding to each original feature and the second weight information of the derived feature corresponding to the original feature.

[0061] In this embodiment, the first weight information is the importance weight of the original feature in the fused feature information, which is dynamically converted from the first utility value. The second weight information is the importance weight of the derived feature in the fused feature information, which is dynamically converted from the second utility value.

[0062] In some embodiments, the step of calculating and processing the first utility value and the second utility value to obtain the first weight information corresponding to each original feature and the second weight information of the derived feature corresponding to the original feature may specifically include: obtaining a first function; and calculating and processing the first utility value and the second utility value based on the first function to obtain the first weight information corresponding to each original feature and the second weight information of the derived feature corresponding to the original feature.

[0063] Optionally, the process of calculating and processing the first utility value and the second utility value to obtain the first weight information corresponding to each original feature and the second weight information of the derived feature corresponding to the original feature can be represented as follows: Where α represents the first weight information; β represents the second weight information; S raw S represents the first utility value; derived The second utility value is represented by GatedSoftmax; the first function is represented by GatedSoftmax, which is defined as follows: Where T is the temperature coefficient, T = 0.5.

[0064] Step S235: Based on the first weight information and the second weight information, perform weighted calculation on each original feature and the derived feature corresponding to the original feature to obtain fused feature information.

[0065] In some embodiments, the step of performing weighted calculations on each original feature and the derived features corresponding to the original feature based on first weight information and second weight information to obtain fused feature information may specifically include: performing weighted calculations on each original feature based on first weight information and performing weighted calculations on the derived features corresponding to the original feature based on second weight information to obtain fused feature information.

[0066] Optionally, the process of weighting each original feature and its corresponding derived feature based on the first weight information and the second weight information to obtain the fused feature information can be expressed as: M(t,m,k)=α*f_raw(t,m,k)+β*f_derived(t,m,k); where M(t,m,k) represents the fused feature information, α represents the first weight information, β represents the second weight information, f_raw(t,m,k) represents the original feature, and f_derived(t,m,k) represents the derived feature.

[0067] Step S3: Perform fault diagnosis based on fused feature information to obtain fault diagnosis information.

[0068] In this embodiment of the application, the fault diagnosis information is the fusion feature information based on the fusion of data from the hardware layer, system layer, and application layer, and is obtained through diagnostic logic or model analysis.

[0069] In some embodiments, such as Figure 3 As shown, the method for fault diagnosis based on fused feature information to obtain fault diagnosis information includes the following steps S31 to S35:

[0070] Step S31: Obtain problem description data.

[0071] In this embodiment, the problem description data is a description of a voice malfunction. This problem description data can be text, voice, or a combination of text and voice. For example, user-reported problem text information can be used as problem description data, such as "voice commands cannot wake up the device," "the output function does not match the input voice," or "poor voice recognition performance." During the user's use of the device, the raw audio data collected by the microphone during the period when the problem occurs can also be used as problem description data.

[0072] Step S32: Input the fused feature information and problem description data into the target processing model, and output the fault description information through the target processing model.

[0073] In this embodiment, the fault description information is the descriptive information output by the target processing model after inputting fused feature information and problem description data. This fault description information can be text, such as "Microphone array signal loss caused voice wake-up function failure," etc.

[0074] In some embodiments, the target processing model can be built based on a large model. A large model refers to an artificial neural network model with a very large number of parameters. In the field of artificial intelligence, a large model typically refers to a model with hundreds of millions to trillions of parameters. The model usually needs to be trained on large-scale datasets and requires significant computational resources for optimization and tuning. Large models are commonly used to solve complex tasks such as natural language processing, computer vision, and speech recognition. In this embodiment, the large model can be a large-scale pre-trained model such as the ChatGPT series, BERT, XLNet, Zhipu model, Claude, Moonshot AI model, ChatGLM model, Tongwen Qianyi model, MiniMax model, Xinghuo model, Llama model, 360GPT model, Qwen model, Baichuan model, Yunque model, vivoLM model, deepseek, Tencent Yuanbao, and Wenxin Yiyan, etc., and this embodiment does not limit the scope of the large model.

[0075] Step S33: Match the hardware layer data, system layer data, and application layer data with the data in the pre-stored fault tree to obtain the cause of the fault.

[0076] In this embodiment, the pre-stored fault tree is a pre-stored fault causal relationship tree, which includes multiple nodes and multiple edges. The nodes are connected by edges, and each node represents a fault event, while the edges represent causal relationships. For example, in the pre-stored fault tree, root node 1 represents voice function failure; child node A (connected to root node 1) represents hardware layer failure, child node B (connected to root node 1) represents system layer failure, child node C (connected to root node 1) represents application layer failure, and child node D (connected to root node 1) represents user audio failure; child node A1 (connected to child node A) represents microphone configuration error (e.g., 4 microphones configured as 2 microphones), child node A2 (connected to child node A) represents microphone damage, and so on. The nodes and edges in the pre-stored fault tree can be set according to actual needs, and are not listed here.

[0077] In this embodiment, the cause of the fault is obtained through fault tree matching, such as the microphone driver version being incompatible with the current system kernel, or the sampling rate parameter being set incorrectly in the ASR service configuration file.

[0078] In some embodiments, the method of matching hardware layer data, system layer data, and application layer data with data in a pre-stored fault tree to obtain the cause of the fault specifically includes: matching node indicators in the pre-stored fault tree based on hardware layer data, system layer data, and application layer data to obtain target node indicators; and outputting the cause of the fault based on the target node indicators.

[0079] In this embodiment of the application, the target node index refers to the feature parameters or judgment criteria that are identified as the nodes most relevant to the current fault state in the pre-stored fault tree after being matched with hardware layer data, system layer data and application layer data.

[0080] Step S34: Perform similarity matching on the fused feature information and the fault case database data to obtain fault case information.

[0081] In this embodiment, the fault case database is a collection of historical fault records. The fault case database may include user requests (e.g., far-field voice - power-on command not working), fault causes (e.g., microphone configuration error), and data metrics (e.g., microphone signal-to-noise ratio). The fault case information is obtained by similarity matching between the fused feature information and the fault case database.

[0082] In some embodiments, the fused feature information includes the time-series features to be diagnosed and the feature vector to be diagnosed, and the fault case database data includes historical time-series features and historical feature vectors. The method for performing similarity matching between the fused feature information and the fault case database data to obtain fault case information specifically includes: calculating the similarity between the time-series features to be diagnosed and the historical time-series features to obtain first similarity information; calculating the similarity between the feature vector to be diagnosed and the historical feature vector to obtain second similarity information; processing the first similarity information and the second similarity information to obtain target similarity information; and determining fault case information based on the target similarity information.

[0083] Optionally, the process of calculating the similarity between the temporal features to be diagnosed and the historical temporal features to obtain the first similarity information can be represented as follows:

[0084] Among them, Q 时序 Indicates the timing characteristics to be diagnosed (such as microphone array waveforms, etc.), C 时序 It indicates the characteristics of historical time sequence.

[0085] Optionally, the process of calculating the similarity between the feature vector to be diagnosed and the historical feature vectors to obtain the second similarity information can be represented as follows:

[0086] Among them, Q 向量 C represents the feature vector to be diagnosed (such as the HTTP status code vector of an application-layer ASR service). 向量 This represents the historical feature vector.

[0087] Step S35: Generate fault diagnosis information based on fault description information, fault cause and fault case information.

[0088] In this embodiment, the fault diagnosis information is a comprehensive and conclusive description of the fault, which is a complete diagnostic result formed by integrating fault description information, fault cause, and fault case information. The fault diagnosis information may include fault phenomenon definition, root cause of the fault, historical solutions, etc.

[0089] Optionally, the fault phenomenon is defined based on the fault description information, such as "the device's voice wake-up function is malfunctioning and the wake-up success rate is less than 10%"; the root cause of the fault is determined based on the fault cause, such as "poor contact of the microphone hardware circuit, resulting in the signal-to-noise ratio of the acquired signal being lower than the threshold"; and the historical solution is determined based on fault case information, such as "in similar cases, it can be repaired by re-plugging and unplugging the microphone interface, with a repair success rate of 92%".

[0090] Step S4: Generate a fault handling strategy based on the fault diagnosis information.

[0091] In this embodiment, the fault handling strategy is a specific action plan generated based on fault diagnosis information to resolve or mitigate the fault.

[0092] In some embodiments, the method for generating a fault handling strategy based on fault diagnosis information specifically includes: confirming the fault type based on the fault diagnosis information; and generating a fault handling strategy based on the fault type.

[0093] In the embodiments of this application, the fault types include a first fault type, a second fault type, and a third fault type. The first fault type is used to characterize fault types that can be repaired by simpler operations (compared to the second fault type), such as adjusting certain parameters, restarting the device, etc. The second fault type is used to characterize fault types that can be repaired by more complex operations (compared to the first fault type), and the third fault type is used to characterize fault types that have not been saved.

[0094] In some embodiments, the step of generating a fault handling strategy based on the fault type may include: when the fault type is a first fault type, determining the fault handling strategy as a first handling strategy; when the fault type is a second fault type, determining the fault handling strategy as a second handling strategy; and when the fault type is a third fault type, determining the fault handling strategy as a third handling strategy.

[0095] In this embodiment, the first processing strategy is to convert technical terms into visual guidance (such as demonstrating the working status of a microphone array through 3D animation) and generate a QR code so that the user can scan the QR code to determine and execute debugging commands (such as ADB (Android Debug Bridge) debugging command sequences) to repair the fault. The second processing strategy is to convert technical terms into visual guidance and generate an operation video so that the user can operate based on the operation video to repair the fault. The third processing strategy is to convert technical terms into visual guidance and save the fault diagnosis information to a fault case database.

[0096] In some embodiments, such as Figure 4 As shown, the fault diagnosis method further includes the following steps S5 to S8:

[0097] Step S5: When a user performs an operation based on the fault handling strategy and the operation fails, obtain the corresponding operation video data of the user.

[0098] In this embodiment, the operation video data is video data of user operations based on fault handling strategies, which is collected and saved in real time.

[0099] Step S6: Perform operation recognition based on the operation video data to obtain the user's operation recognition result information.

[0100] In this embodiment, the operation identification result information is used to characterize whether the user's operation is correct. The operation identification result information includes first result information and second result information. The first result information is used to characterize that the user's operation is abnormal, and the second result information is used to characterize that the user's operation is correct.

[0101] Step S7: If the operation recognition result information is the first result information, then generate a prompt message. The first result information indicates that the user has an operation error.

[0102] In this embodiment, the prompt message is a notification to the user indicating an operational error. The prompt message may be in the form of text, voice, light, or a combination of multiple forms. For example, it may output the text "Operation error, please try again," the voice message "Operation error, please try again," or a flashing red light.

[0103] Step S8: If the operation identification result information is the second result information, an anomaly log is generated based on the fault diagnosis information, and maintenance tool information and spare parts list are generated based on the anomaly log so that the user can handle the fault based on the maintenance tool information and spare parts list.

[0104] In this embodiment of the application, the exception log is used to characterize failure cases where the user's operation was correct but the failure still could not be fixed.

[0105] The repair tool information is a dynamically recommended set of hardware or software tools based on anomaly logs. This information is used to perform self-tests and repair operations. Optionally, the repair tool information may include physical toolkits and virtual toolkits. For example, a physical toolkit may include a precision multimeter and its model, a tweezers set and its specifications, etc., while a virtual toolkit may include circuit diagram design documents (marked with the coordinates of the faulty module), voltage fluctuation analysis scripts, etc.

[0106] The spare parts list is a list of hardware modules or software components that need to be replaced based on anomaly logs. For example, the spare parts list may include surface-mount capacitors and their specifications, power supply regulator chips and their models, etc.

[0107] This application provides a fault diagnosis method. It acquires hardware-layer, system-layer, and application-layer data; then constructs fused feature information based on these data; finally, it performs fault diagnosis based on this fused feature information to obtain fault diagnosis information; and finally, it generates a fault handling strategy based on this fault diagnosis information. This application, by acquiring multi-dimensional data from the hardware, system, and application layers and constructing fused feature information for fault diagnosis and strategy generation, can comprehensively capture the manifestation of faults at different levels, improve the accuracy of fault location, reduce missed diagnoses and misdiagnoses, and enhance the comprehensiveness and accuracy of fault diagnosis. It also facilitates user self-troubleshooting.

[0108] To better implement the fault diagnosis method in the embodiments of this application, a fault diagnosis device is also provided in the embodiments of this application, based on the fault diagnosis method, such as... Figure 5 As shown, the fault diagnosis device 200 includes:

[0109] The first acquisition module 201 is used to acquire hardware layer data, system layer data, and application layer data.

[0110] Information construction module 202 is used to construct fused feature information based on hardware layer data, system layer data and application layer data.

[0111] The fault diagnosis module 203 is used to perform fault diagnosis based on fused feature information to obtain fault diagnosis information.

[0112] The strategy generation module 204 is used to generate fault handling strategies based on fault diagnosis information.

[0113] In some embodiments, the information construction module 202 is specifically used to: perform alignment processing on hardware layer data, system layer data and application layer data to obtain aligned data information; perform feature extraction on the aligned data information to obtain initial feature information and derived features corresponding to each original feature in the initial feature information; and perform fusion processing on each original feature in the initial feature information and the derived features corresponding to the original feature to obtain fused feature information.

[0114] In some embodiments, the information construction module 202 is further configured to: perform context encoding on the initial feature information to obtain the context feature corresponding to each original feature; perform calculation processing on each original feature and the context feature corresponding to the original feature to obtain a first utility value corresponding to each original feature; perform calculation processing on the derived feature corresponding to each original feature and the context feature corresponding to the original feature to obtain a second utility value corresponding to each original feature; perform calculation processing on the first utility value and the second utility value to obtain a first weight information corresponding to each original feature and a second weight information of the derived feature corresponding to the original feature; and perform weighted calculation on each original feature and the derived feature corresponding to the original feature based on the first weight information and the second weight information to obtain fused feature information.

[0115] In some embodiments, the fault diagnosis module 203 is specifically used for: acquiring problem description data; inputting the fused feature information and problem description data into the target processing model, and outputting fault description information through the target processing model; matching hardware layer data, system layer data, and application layer data with data in a pre-stored fault tree to obtain the cause of the fault; performing similarity matching on the fused feature information and fault case library data to obtain fault case information; and generating fault diagnosis information based on the fault description information, the cause of the fault, and the fault case information.

[0116] In some embodiments, the fault diagnosis module 203 is further configured to include: matching node indicators in a pre-stored fault tree based on hardware layer data, system layer data, and application layer data to obtain target node indicators; and outputting fault causes based on target node indicators.

[0117] In some embodiments, the fused feature information includes the time-series features to be diagnosed and the feature vector to be diagnosed, and the fault case database data includes historical time-series features and historical feature vectors. The fault diagnosis module 203 is further configured to: calculate the similarity between the time-series features to be diagnosed and the historical time-series features to obtain first similarity information; calculate the similarity between the feature vector to be diagnosed and the historical feature vector to obtain second similarity information; process the first similarity information and the second similarity information to obtain target similarity information; and determine fault case information based on the target similarity information.

[0118] In some embodiments, the strategy generation module 204 is specifically used to: confirm the fault type based on fault diagnosis information; and generate a fault handling strategy based on the fault type.

[0119] In some embodiments, the fault diagnosis device further includes:

[0120] The second acquisition module 205 is used to acquire the corresponding operation video data of the user when the user performs an operation based on the fault handling strategy and the operation fails.

[0121] The operation recognition module 206 is used to perform operation recognition based on operation video data to obtain the user's operation recognition result information.

[0122] The first prompt module 207 is used to generate a prompt message if the operation recognition result information is the first result information. The first result information indicates that the user has performed an abnormal operation.

[0123] The second prompt module 208 is used to generate an anomaly log based on the fault diagnosis information if the operation identification result information is the second result information, and to generate maintenance tool information and a spare parts list based on the anomaly log, so that the user can handle the fault based on the maintenance tool information and spare parts list.

[0124] This application also provides a computer device that integrates any of the fault diagnosis devices provided in this application. The computer device includes:

[0125] One or more processors;

[0126] Memory; and

[0127] One or more applications, wherein the applications are stored in memory and configured to be executed by a processor of any of the steps in the above-described fault diagnosis method embodiments.

[0128] This application also provides a computer device that integrates any of the fault diagnosis devices provided in this application. For example... Figure 6 As shown, it illustrates a structural schematic diagram of the computer device involved in the embodiments of this application, specifically:

[0129] The computer device may include components such as a processor 801 with one or more processing cores, a memory 802 with one or more computer-readable storage media, a power supply 803, and an input unit 804. Those skilled in the art will understand that... Figure 6 The computer device structure shown does not constitute a limitation on the computer device and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein:

[0130] The processor 801 is the control center of the computer device. It connects various parts of the computer device via various interfaces and lines. By running or executing software programs and / or modules stored in the memory 802, and by calling data stored in the memory 802, it performs various functions of the computer device and processes data, thereby providing overall monitoring of the computer device. Optionally, the processor 801 may include one or more processing cores; preferably, the processor 801 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 801.

[0131] The memory 802 can be used to store software programs and modules. The processor 801 executes various functional applications and data processing by running the software programs and modules stored in the memory 802. The memory 802 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 802 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 802 may also include a memory controller to provide the processor 801 with access to the memory 802.

[0132] The computer device also includes a power supply 803 that supplies power to the various components. Preferably, the power supply 803 can be logically connected to the processor 801 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 803 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0133] The computer device may also include an input unit 804, which can be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.

[0134] Although not shown, the computer device may also include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 801 in the computer device loads the executable files corresponding to the processes of one or more application programs into the memory 802 according to the following instructions, and the processor 801 runs the application programs stored in the memory 802 to realize various functions, as follows:

[0135] Acquire hardware layer data, system layer data, and application layer data;

[0136] Based on hardware layer data, system layer data, and application layer data, we construct fused feature information.

[0137] Fault diagnosis is performed based on fused feature information to obtain fault diagnosis information;

[0138] Based on fault diagnosis information, a fault handling strategy is generated.

[0139] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0140] Therefore, embodiments of this application provide a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), a magnetic disk, or an optical disk, etc. A computer program is stored thereon, and the computer program is loaded by a processor to execute the steps in any of the fault diagnosis methods provided in embodiments of this application. For example, the computer program loaded by the processor can execute the following steps:

[0141] Acquire hardware layer data, system layer data, and application layer data;

[0142] Based on hardware layer data, system layer data, and application layer data, we construct fused feature information.

[0143] Fault diagnosis is performed based on fused feature information to obtain fault diagnosis information;

[0144] Based on fault diagnosis information, a fault handling strategy is generated.

[0145] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the detailed descriptions of other embodiments above, which will not be repeated here.

[0146] In practice, each of the above units or structures can be implemented as an independent entity or can be arbitrarily combined to be implemented as the same or several entities. For the specific implementation of each of the above units or structures, please refer to the previous method embodiments, which will not be repeated here.

[0147] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0148] The technical solutions provided by the embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A fault diagnosis method, characterized in that, include: Acquire hardware layer data, system layer data, and application layer data; Based on the hardware layer data, the system layer data, and the application layer data, a fused feature information is constructed. Fault diagnosis is performed based on the fused feature information to obtain fault diagnosis information; Based on the fault diagnosis information, a fault handling strategy is generated.

2. The fault diagnosis method according to claim 1, characterized in that, The construction of fused feature information based on the hardware layer data, the system layer data, and the application layer data includes: The hardware layer data, the system layer data, and the application layer data are aligned to obtain aligned data information; Feature extraction is performed on the alignment data information to obtain initial feature information and derived features corresponding to each original feature in the initial feature information; Each original feature and its corresponding derived feature in the initial feature information are fused to obtain fused feature information.

3. The fault diagnosis method according to claim 2, characterized in that, The process of fusing each original feature and its corresponding derived feature in the initial feature information to obtain fused feature information includes: The initial feature information is context-encoded to obtain the context feature corresponding to each original feature; Each original feature and its corresponding context feature are processed to obtain a first utility value for each original feature. The derived features and context features corresponding to each original feature are calculated and processed to obtain the second utility value corresponding to each original feature. The first utility value and the second utility value are calculated and processed to obtain the first weight information corresponding to each original feature and the second weight information of the derived feature corresponding to the original feature; The fused feature information is obtained by weighting each original feature and its corresponding derived feature based on the first weight information and the second weight information.

4. The fault diagnosis method according to claim 1, characterized in that, The fault diagnosis based on the fused feature information to obtain fault diagnosis information includes: Obtain problem description data; The fused feature information and the problem description data are input into the target processing model, and the target processing model outputs the fault description information. The hardware layer data, the system layer data, and the application layer data are matched with the data in the pre-stored fault tree to obtain the cause of the fault; The fused feature information and the fault case database data are matched for similarity to obtain fault case information; The fault diagnosis information is generated based on the fault description information, the fault cause, and the fault case information.

5. The fault diagnosis method according to claim 4, characterized in that, The step of matching the hardware layer data, the system layer data, and the application layer data with data in a pre-stored fault tree to obtain the cause of the fault includes: Based on the hardware layer data, the system layer data, and the application layer data, node indicators are matched in the pre-stored fault tree to obtain the target node indicators. Based on the target node indicators, the cause of the failure is output.

6. The fault diagnosis method according to claim 4, characterized in that, The fused feature information includes the time-series features to be diagnosed and the feature vector to be diagnosed, and the fault case database data includes historical time-series features and historical feature vectors; The step of performing similarity matching between the fused feature information and the fault case database data to obtain fault case information includes: The similarity between the time-series features to be diagnosed and the historical time-series features is calculated to obtain the first similarity information; The similarity between the feature vector to be diagnosed and the historical feature vector is calculated to obtain second similarity information; The first similarity information and the second similarity information are calculated and processed to obtain the target similarity information; The fault case information is determined based on the target similarity information.

7. The fault diagnosis method according to claim 1, characterized in that, The generation of a fault handling strategy based on the fault diagnosis information includes: Based on the fault diagnosis information, the fault type is confirmed; Based on the fault type, a fault handling strategy is generated.

8. The fault diagnosis method according to claim 1, characterized in that, The fault diagnosis method further includes: When a user performs an operation based on the fault handling strategy and the operation fails, the corresponding operation video data of the user is obtained. Based on the operation video data, operation recognition is performed to obtain the user's operation recognition result information; If the operation identification result information is the first result information, then a prompt message is generated; the first result information indicates that the user has an operation abnormality. If the operation identification result information is the second result information, then an anomaly log is generated based on the fault diagnosis information, and maintenance tool information and spare parts list are generated based on the anomaly log, so that the user can perform fault handling based on the maintenance tool information and the spare parts list.

9. A fault diagnosis device, characterized in that, include: The first acquisition module is used to acquire hardware layer data, system layer data, and application layer data. The information construction module is used to construct fused feature information based on the hardware layer data, the system layer data, and the application layer data; The fault diagnosis module is used to perform fault diagnosis based on the fused feature information to obtain fault diagnosis information; The strategy generation module is used to generate a fault handling strategy based on the fault diagnosis information.

10. A computer-readable storage medium, characterized in that, It stores a computer program, which is loaded by a processor to perform the steps in the fault diagnosis method according to any one of claims 1 to 8.

Images