Fault analysis method for automatic testing equipment of vehicle machine

Abstract

The invention discloses a fault analysis method for vehicle-mounted automatic test equipment, and relates to the technical field of vehicle-mounted electronic equipment testing, and the method comprises the steps: collecting vehicle-mounted test data of the vehicle-mounted automatic test equipment, and storing the vehicle-mounted test data in a test database according to a time sequence; performing feature extraction on the test database through a deep learning model to generate a fault feature vector; inputting the fault feature vector into a fault diagnosis model, and generating a fault type judgment result and a fault probability value; and according to the fault type determination result and the fault probability values, screening corresponding fault recovery schemes from a preset fault recovery strategy library, and sorting the fault recovery schemes according to the fault probability values from high to low to form a fault processing report. According to the method, the fault processing report is generated based on the layered fault recovery strategy library and the priority grading mechanism, the technical problems that a traditional fault diagnosis method is insufficient in feature extraction and low in diagnosis precision are solved, and the accuracy and recovery efficiency of fault diagnosis are improved.

Description

Technical Field

[0001] The invention relates to the technical field of vehicle-mounted electronic equipment testing, and in particular to a fault analysis method for vehicle-mounted automatic testing equipment. Background Art

[0002] As a key link in ensuring the quality of vehicle-mounted systems, the reliability and fault diagnosis capabilities of vehicle-mounted computer automated testing equipment directly affect the test efficiency and result accuracy. At present, vehicle-mounted computer automated testing equipment generally adopts rule-based fault diagnosis methods, which detect and classify faults through preset thresholds and logical judgments. However, this traditional fault diagnosis method often exhibits problems such as low diagnostic accuracy and insufficient generalization capabilities when faced with complex test environments and diverse fault modes. In addition, existing fault diagnosis systems are mostly based on single fault type judgments, lack quantitative assessments of the probability of fault occurrence, and are difficult to provide testers with clear priority fault handling suggestions.

[0003] In recent years, with the development of artificial intelligence technology, fault diagnosis methods based on deep learning have gradually attracted the attention of academia and industry. However, the existing deep learning fault diagnosis methods still face many challenges when applied to vehicle automation test equipment: first, the timing characteristics of the test data are not fully utilized, resulting in unsatisfactory feature extraction results; second, the fault diagnosis model is not interpretable enough, making it difficult to establish a mapping relationship between fault types and specific repair strategies; third, the lack of a systematic fault repair solution management mechanism makes the fault handling process inefficient and the maintenance cost high. The existence of these problems seriously restricts the test efficiency and reliability of vehicle automation test equipment, and there is an urgent need for an intelligent fault diagnosis method that can fully utilize deep learning technology, has the ability to extract timing features, and can provide reliable repair strategies. Summary of the invention

[0004] In view of the problems that the existing vehicle computer automated test equipment has low fault diagnosis accuracy, insufficient feature extraction, and lack of a systematic fault handling mechanism, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is to utilize the time series characteristics of the test data to perform feature extraction, realize high-precision fault type identification, and automatically generate a repair plan with clear priority based on the fault probability.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, an embodiment of the present invention provides a fault analysis method for vehicle-computer automation test equipment, which includes collecting vehicle-computer test data of the vehicle-computer automation test equipment and storing it in a test database according to a time series; extracting features from the test database through a constructed deep learning model to generate a fault feature vector; inputting the fault feature vector into a constructed fault diagnosis model to generate a fault type determination result and a fault probability value; according to the fault type determination result and the fault probability value, screening a corresponding fault repair plan from a preset fault repair strategy library, and sorting the fault repair plan from high to low according to the fault probability value to form a fault handling report.

[0008] As a preferred solution of the fault analysis method for vehicle-mounted automated testing equipment described in the present invention, wherein: the preset fault repair strategy library includes a hardware repair strategy sub-library, a software repair strategy sub-library and an operation guidance strategy sub-library; the method for forming the fault handling report is to retrieve a repair instruction set in the corresponding strategy sub-library based on the fault type determination result, wherein the repair instruction set includes fault handling steps, repair parameter configuration and repair tool selection; mark the execution priority of the processing steps in the repair instruction set according to the fault probability value; sort the repair instruction sets with marked priorities from high to low according to the fault probability value, and generate a fault handling report, wherein the fault handling report includes a description of the fault phenomenon, an analysis of the cause of the fault, a list of repair steps, a list of required tools and an estimated repair time.

[0009] As a preferred solution of the fault analysis method for vehicle-mounted automated testing equipment described in the present invention, wherein: the execution priority of the processing steps in the repair instruction set is marked according to the fault probability value, including: when the fault probability value is greater than a first threshold, the processing steps in the repair instruction set are marked as an emergency processing level; when the fault probability value is less than or equal to the first threshold and greater than a second threshold, the processing steps in the repair instruction set are marked as a priority processing level; when the fault probability value is less than or equal to the second threshold, the processing steps in the repair instruction set are marked as a regular processing level.

[0010] As a preferred solution of the fault analysis method for vehicle-mounted automated test equipment described in the present invention, wherein: the fault diagnosis model is supervised by an improved gradient boosting decision tree structure; the improved gradient boosting decision tree structure includes a backbone decision network and an auxiliary verification network; the backbone decision network is composed of a number of decision trees, and the depth of each decision tree is set to 8 layers; the auxiliary verification network adopts a three-layer fully connected structure; the method for generating the fault type determination result is to divide the fault feature vector into a hardware fault feature subset, a software fault feature subset and an operation fault feature subset through the backbone decision network, and calculate the fault type score of each subset respectively, and the specific formula is as follows:

[0011]

[0012] Among them, Score k is the score of the k-th fault feature subset, is the predicted weight value of the jth leaf node in the kth fault feature subset, T is the total number of decision trees, is the weight coefficient of the tth tree in the kth type of fault, X k is the k-th fault feature subset input vector, f t () is the prediction function of the t-th decision tree, γ k is the feature importance weight of the k-th fault feature subset, and Corr() is the feature correlation adjustment coefficient within the subset.

[0013] Based on a preset historical fault case library, a feature similarity matrix is ​​calculated through the auxiliary verification network; the fault type score and the feature similarity matrix are weightedly fused to generate a fault type determination result, wherein the fault type determination result includes a hardware fault code, a software fault code and an operation fault code; the fault type determination result is probability normalized to output various types of fault probability values, wherein the fault probability value reflects the confidence of the occurrence of the fault.

[0014] As a preferred solution of the fault analysis method for vehicle-mounted automated testing equipment described in the present invention, wherein: the method for generating the fault feature vector is that the deep learning model includes a multi-layer convolutional neural network and a long short-term memory network, wherein the multi-layer convolutional neural network contains three convolutional layers and two pooling layers; the long short-term memory network is provided with a memory unit; the spatial feature extraction of the hardware operating parameters in the test database is performed through the multi-layer convolutional neural network to generate a hardware feature tensor; at the same time, the temporal feature extraction of the software exception log and the manual operation record is performed through the long short-term memory network to generate a software operation feature tensor; the hardware feature tensor and the software operation feature tensor are input into the attention mechanism layer to calculate the key feature score, wherein the key feature score indicates the degree of correlation between hardware exceptions, software failures and operational errors; based on the key feature score, the hardware feature tensor and the software operation feature tensor are feature fused to form a fault feature vector, wherein each dimension in the fault feature vector corresponds to a potential fault feature.

[0015] As a preferred solution of the fault analysis method for vehicle-computer automated testing equipment described in the present invention, wherein: the test database includes vehicle-computer test data; the vehicle-computer test data includes hardware operating parameters, software exception logs and manual operation records; the hardware operating parameters include central processing unit utilization, memory occupancy, hard disk read and write rates, system bus bandwidth and power management status; the software exception logs include application crash records, system service exception information and resource scheduling conflict information, and the manual operation records include touch command sequences, key operation timings and voice control commands.

[0016] As a preferred solution of the fault analysis method for vehicle-computer automated testing equipment described in the present invention, the method for acquiring the test database is to convert the hardware operating parameters into digital signals through a data acquisition card, and to normalize the hardware operating parameters, the software exception log and the manual operation record through a data acquisition module to generate vehicle-computer operating status information; and to write the vehicle-computer operating status information into a timing data table in the test database in a unified timestamp format.

[0017] In the second aspect, an embodiment of the present invention provides a fault analysis system for vehicle-computer automated testing equipment, comprising: an acquisition module, used to acquire vehicle-computer test data of the vehicle-computer automated testing equipment, and store it in a test database in time series; a feature extraction module, used to extract features from the test database through a constructed deep learning model to generate a fault feature vector; a generation module, used to input the fault feature vector into a constructed fault diagnosis model to generate a fault type determination result and a fault probability value; an analysis and processing module, used to screen corresponding fault repair plans from a preset fault repair strategy library according to the fault type determination result and the fault probability value, and sort the fault repair plans from high to low according to the fault probability value to form a fault handling report.

[0018] In a third aspect, an embodiment of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the fault analysis method for vehicle-mounted automated testing equipment as described in the first aspect of the present invention are implemented.

[0019] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the fault analysis method for vehicle-mounted automated testing equipment as described in the first aspect of the present invention are implemented.

[0020] The beneficial effects of the present invention are as follows: by constructing a test data collection and storage mechanism based on timing characteristics, unified management of hardware, software and operation data is achieved; by combining the multi-layer convolutional neural network and long short-term memory network in the deep learning model, multi-dimensional feature extraction and fusion of test data are performed, effectively improving the feature expression ability; by adopting an improved gradient boosting decision tree structure, high-precision fault type identification and probability evaluation are achieved through the synergy of the backbone decision network and the auxiliary verification network; based on a hierarchical fault repair strategy library and a priority grading mechanism, a fault handling report containing detailed repair instructions is automatically generated, overcoming the technical problems of insufficient feature extraction, low diagnostic accuracy and lack of systematic repair solutions in traditional fault diagnosis methods, thereby improving the accuracy of fault diagnosis and repair efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0022] Figure 1 This is a flow chart of a fault analysis method for vehicle computer automation testing equipment according to Example 1. DETAILED DESCRIPTION

[0023] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0024] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0025] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0026] Example 1

[0027] Reference Figure 1 , which is the first embodiment of the present invention, provides a fault analysis method for vehicle computer automatic test equipment, comprising:

[0028] S1: Collect vehicle test data from the vehicle automation test equipment and store it in the test database in time series.

[0029] Specifically, the vehicle computer test data includes hardware operating parameters, software exception logs and manual operation records; hardware operating parameters include CPU utilization, memory usage, hard disk read and write rates, system bus bandwidth and power management status; software exception logs include application crash records, system service exception information and resource scheduling conflict information, and manual operation records include touch command sequences, key operation timings and voice control commands.

[0030] Furthermore, the hardware operating parameters are converted into digital signals through the data acquisition card, and the hardware operating parameters, software abnormality logs and manual operation records are normalized through the data acquisition module to generate the vehicle computer operating status information; the adaptive data normalization method is used for normalization processing, and the specific formula is as follows:

[0031]

[0032] Among them, X norm is the normalized data value, indicating the output result after standardization processing, μ dynamic is the average value calculated in the current data window, σ dynamic is the dynamic standard deviation calculated in the current data window, A is the original input data value, and ∈ is the smoothing factor.

[0033] Furthermore, the vehicle computer operating status information is written into a time series data table in the test database in a unified timestamp format, wherein the time series data table records the change trend of the vehicle computer operating status information with a basic sampling interval of 10 milliseconds.

[0034] S2: Extract features from the test database using the constructed deep learning model to generate a fault feature vector.

[0035] Specifically, the deep learning model includes a multi-layer convolutional neural network and a long short-term memory network, wherein the multi-layer convolutional neural network contains three convolutional layers and two pooling layers; the long short-term memory network sets memory units.

[0036] Furthermore, the method for generating the fault feature vector is to extract the spatial features of the hardware operating parameters in the test database through a multi-layer convolutional neural network to generate a hardware feature tensor; at the same time, the time series features of the software exception log and the manual operation record are extracted through a long short-term memory network to generate a software operation feature tensor; the hardware feature tensor and the software operation feature tensor are input into the attention mechanism layer to calculate the key feature score, where the key feature score indicates the correlation between hardware anomalies, software failures and operational errors.

[0037] Furthermore, the specific formula for the key feature score is as follows:

[0038]

[0039] Among them, Score b is the key feature score, N is the number of hardware feature dimensions, M is the number of software feature dimensions, α i is the hardware feature weight coefficient of the i-th dimension, H i (t) is the time series value of the hardware feature of the i-th dimension, β i is the time decay factor, Z is the time window length, τ j is the software feature weight coefficient, S j (t) is the time series value of the j-th dimension software feature, E is the deviation between the current system state and the normal state, and θ is the normalization coefficient.

[0040] Specifically, the key feature score establishes a quantitative indicator system for fault diagnosis through a comprehensive evaluation of hardware operating parameters and software operation data, enabling the system to objectively measure the severity of various abnormal conditions; the key feature score mechanism enables dynamic tracking of fault development trends by introducing weight distribution of timing features, and can timely capture subtle changes in the operating status of equipment; in addition, the key feature score also establishes a correlation mapping between hardware anomalies, software failures and human operational errors, and provides a basis for locating the source of the fault through the fusion analysis of multi-dimensional features.

[0041] Furthermore, based on the key feature scores, the hardware feature tensor and the software operation feature tensor are fused to form a fault feature vector, where each dimension in the fault feature vector corresponds to a potential fault feature.

[0042] S3: Input the fault feature vector into the constructed fault diagnosis model to generate a fault type determination result and a fault probability value.

[0043] Specifically, the fault diagnosis model adopts an improved gradient boosting decision tree structure for supervision; the improved gradient boosting decision tree structure includes a backbone decision network and an auxiliary verification network; the backbone decision network is composed of several decision trees, and the depth of each decision tree is set to 8 layers; the auxiliary verification network adopts a three-layer fully connected structure.

[0044] Furthermore, the method for generating the fault type determination result is to divide the fault feature vector into hardware fault feature subsets, software fault feature subsets and operation fault feature subsets through the backbone decision network, and calculate the fault type score of each subset respectively. The specific formula is as follows:

[0045]

[0046] Among them, Scorek is the score of the k-th fault feature subset, is the predicted weight value of the jth leaf node in the kth fault feature subset, T is the total number of decision trees, is the weight coefficient of the tth tree in the kth type of fault, X k is the k-th fault feature subset input vector, f t () is the prediction function of the t-th decision tree, γ k is the feature importance weight of the k-th fault feature subset, and Corr() is the feature correlation adjustment coefficient within the subset.

[0047] Furthermore, based on the preset historical fault case library, the feature similarity matrix is ​​calculated through the auxiliary verification network. The specific formula is as follows:

[0048]

[0049] Among them, S ab is the feature similarity between fault case a and the fault case, x a (z) is the time series feature sequence of the ath fault case, x b (z) is the time series feature sequence of the bth fault case, Z is the time window length, is the reference time point, is the static eigenvalue of the g-th dimension of the a-th fault case, is the g-th dimension static eigenvalue of the b-th fault case, and n is the number of static feature dimensions.

[0050] Specifically, the fault type score and the feature similarity matrix are weightedly fused to generate a fault type determination result, wherein the fault type determination result includes a hardware fault code, a software fault code, and an operation fault code.

[0051] Furthermore, the fault type determination results are subjected to probability normalization processing, and various fault probability values ​​are output, wherein the fault probability value reflects the confidence level of the fault occurrence.

[0052] S4: According to the fault type determination result and the fault probability value, a corresponding fault repair solution is selected from a preset fault repair strategy library, and the fault repair solutions are sorted from high to low according to the fault probability value to form a fault handling report.

[0053] Specifically, the preset fault repair strategy library includes a hardware repair strategy sub-library, a software repair strategy sub-library and an operation guidance strategy sub-library, wherein each sub-library sets a repair instruction template and a repair parameter threshold; the method for forming a fault handling report is to retrieve a repair instruction set in a corresponding strategy sub-library based on a fault type determination result, wherein the repair instruction set includes fault handling steps, repair parameter configuration and repair tool selection; and the processing steps in the repair instruction set are marked with execution priorities according to the fault probability value.

[0054] Furthermore, the calculation method of the fault probability value integrates two key dimensions: fault type score and historical case similarity, and achieves a dynamic balance between the two through weighted coefficients. The fault type score reflects the characteristic performance of the current fault, while the historical case similarity aggregates the reference value of historical fault cases through exponential decay, so that cases with closer time intervals have higher reference weights. In order to ensure the stability and comparability of the calculation results, the system reference benchmark variance is introduced as a normalization factor, and the mapping of probability values ​​in the [0,1] interval is achieved through the fractional structure.

[0055] Furthermore, when the fault probability value is greater than a first threshold, the processing steps in the repair instruction set are marked as an emergency processing level; when the fault probability value is less than or equal to the first threshold and greater than a second threshold, the processing steps in the repair instruction set are marked as a priority processing level; when the fault probability value is less than or equal to the second threshold, the processing steps in the repair instruction set are marked as a regular processing level.

[0056] It should be noted that the first threshold is determined based on the safety risk level of system operation, the possibility of damage to key equipment and the severity of the impact on production continuity; the second threshold is determined based on the level of equipment performance degradation, the degree of reduction in system operation stability and the reasonable allocation needs of maintenance resources.

[0057] Specifically, the priority-marked repair instruction set is sorted from high to low according to the fault probability value, and a fault handling report is generated, wherein the fault handling report includes a description of the fault phenomenon, an analysis of the cause of the fault, a list of repair steps, a list of required tools, and an estimated repair time.

[0058] Furthermore, the present embodiment also provides a fault analysis system for vehicle-computer automated testing equipment, comprising: an acquisition module, used to acquire vehicle-computer test data of the vehicle-computer automated testing equipment, and store it in a test database in time series; a feature extraction module, used to extract features from the test database through a constructed deep learning model to generate a fault feature vector; a generation module, used to input the fault feature vector into a constructed fault diagnosis model to generate a fault type determination result and a fault probability value; an analysis and processing module, used to screen corresponding fault repair plans from a preset fault repair strategy library according to the fault type determination result and the fault probability value, and sort the fault repair plans from high to low according to the fault probability value to form a fault handling report.

[0059] This embodiment also provides a computer device, which is suitable for the fault analysis method of vehicle-mounted automated testing equipment, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the fault analysis method for vehicle-mounted automated testing equipment proposed in the above embodiment.

[0060] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0061] The present embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps: collecting vehicle computer test data of a vehicle computer automated test device, and storing the data in a test database in a time series; performing feature extraction on the test database through a constructed deep learning model to generate a fault feature vector; inputting the fault feature vector into a constructed fault diagnosis model to generate a fault type determination result and a fault probability value; based on the fault type determination result and the fault probability value, screening corresponding fault repair plans from a preset fault repair strategy library, and sorting the fault repair plans from high to low according to the fault probability value to form a fault handling report.

[0062] In summary, the present invention realizes the unified management of hardware, software and operation data by constructing a test data collection and storage mechanism based on timing characteristics; combines the multi-layer convolutional neural network and long short-term memory network in the deep learning model to extract and fuse multi-dimensional features of the test data, effectively improving the feature expression ability; adopts an improved gradient boosting decision tree structure, through the synergy of the backbone decision network and the auxiliary verification network, to achieve high-precision fault type identification and probability evaluation; based on a hierarchical fault repair strategy library and a priority grading mechanism, a fault handling report containing detailed repair instructions is automatically generated, overcoming the technical problems of insufficient feature extraction, low diagnostic accuracy, and lack of systematic repair solutions in traditional fault diagnosis methods, thereby improving the accuracy of fault diagnosis and repair efficiency.

[0063] Example 2

[0064] Referring to Table 1, which is a second embodiment of the present invention, this embodiment provides a fault analysis method for vehicle computer automated testing equipment. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0065] Specifically, a 2024 smart in-vehicle entertainment system of a certain brand was selected as the test object and tested in a standard laboratory environment (temperature 25±2℃, relative humidity 45±5%). The test equipment includes a high-precision data acquisition card (sampling rate 100kHz, resolution 16 bits), a multi-function signal analyzer and a special vehicle computer test tooling.

[0066] Furthermore, the test object is installed on the test fixture and connected to the data acquisition equipment and signal analyzer. Through the self-developed data acquisition software, the operation data is continuously collected for 72 hours with a basic sampling interval of 10ms. The collection content includes: CPU usage, memory usage (real-time sampling), hard disk read and write rate, system bus bandwidth utilization, and power management status data. At the same time, a special logging module is deployed to capture abnormal information at the software level, including application crash records, system service exceptions, and resource scheduling conflict information.

[0067] Furthermore, in order to simulate real-world usage scenarios, a standardized manual operation test process was designed, including 50 different touch operation sequences, 30 key combination operations, and 20 voice control commands. During the test, an adaptive data normalization algorithm was used to process the collected data in real time, and data of different dimensions were uniformly converted to the [0,1] interval. The data was feature extracted through the configured deep learning model (containing 3 convolutional layers, the pooling layer adopts the maximum pooling strategy, the convolution kernel size is 3×3, and the number of channels in each layer is 32, 64, and 128 respectively). The long short-term memory network sets 128 memory units to capture timing features.

[0068] Furthermore, as shown in Table 1, in terms of fault diagnosis accuracy, the method of the present invention reaches 95.8%, which is 13.3 percentage points higher than the traditional method. This improvement is mainly due to the innovative deep learning model architecture and feature fusion mechanism. It is particularly noteworthy that the method of the present invention has made a breakthrough in the average response time, and only 150ms is required to complete the fault diagnosis, which is 82.4% shorter than the traditional method, which is of great significance for the real-time response requirements of the vehicle system.

[0069] Table 1 Comparison between the present invention and traditional diagnostic methods

[0070] Test indicators Traditional diagnostic methods Method of the present invention Fault diagnosis accuracy (%) 82.5 95.8 Average response time (ms) 850 150 False alarm rate (%) 8.6 2.1 Missing rate (%) 7.2 1.8 Fault classification accuracy (%) 84.3 96.2 Hardware fault identification rate (%) 83.7 95.5 Software fault recognition rate (%) 81.9 94.8 Operation error recognition rate (%) 85.2 96.7 Resource utilization (%) 45.6 25.4 Diagnosis real-time (ms) 920 180

[0071] Furthermore, in terms of reliability indicators, the false alarm rate and missed alarm rate of the method of the present invention are reduced to 2.1% and 1.8% respectively, which are 75.6% and 75% lower than those of the traditional method. This shows that the method of the present invention has made significant breakthroughs in the accuracy and reliability of fault judgment. In particular, the fault classification accuracy has reached a high level of 96.2%, thanks to the improved gradient boosting decision tree structure and the innovative feature similarity calculation method.

[0072] Specifically, the method of the present invention shows comprehensive advantages in terms of the ability to identify different types of faults: the hardware fault recognition rate is 95.5%, the software fault recognition rate is 94.8%, and the operation error recognition rate is 96.7%, which are 11.8, 12.9, and 11.5 percentage points higher than the traditional method, respectively. This comprehensive performance improvement proves the effectiveness of the multi-dimensional feature fusion and dynamic weight allocation mechanism proposed in the present invention. At the same time, the resource occupancy rate of the method of the present invention is only 25.4%, which is 44.3% lower than that of the traditional method, which shows that the present invention not only improves the diagnostic performance, but also optimizes the computing efficiency. In terms of diagnostic real-time, the method of the present invention only takes 180ms, which is 80.4% higher than the traditional method. This efficient processing capability has important practical value for real-time fault diagnosis of vehicle-mounted systems.

[0073] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A fault analysis method for vehicle computer automated testing equipment, characterized in that: include, Collect vehicle test data from vehicle automation test equipment and store it in the test database in time series; Extracting features from the test database using the constructed deep learning model to generate a fault feature vector; Inputting the fault feature vector into the constructed fault diagnosis model to generate a fault type determination result and a fault probability value; According to the fault type determination result and the fault probability value, a corresponding fault repair plan is screened from a preset fault repair strategy library, and the fault repair plans are sorted from high to low according to the fault probability value to form a fault handling report.

2. The fault analysis method for vehicle computer automated testing equipment according to claim 1, characterized in that: The preset fault repair strategy library includes a hardware repair strategy sub-library, a software repair strategy sub-library and an operation guidance strategy sub-library; the method for forming the fault handling report is: Retrieving a repair instruction set in a corresponding strategy sub-library based on the fault type determination result, wherein the repair instruction set includes fault handling steps, repair parameter configuration, and repair tool selection; Marking execution priorities of processing steps in the repair instruction set according to the fault probability value; The priority-marked repair instruction set is sorted from high to low according to the fault probability value to generate a fault handling report, wherein the fault handling report includes a description of the fault phenomenon, an analysis of the cause of the fault, a list of repair steps, a list of required tools, and an estimated repair time.

3. The fault analysis method for vehicle computer automated testing equipment according to claim 2, characterized in that: Marking execution priorities of processing steps in the repair instruction set according to the fault probability value includes: When the fault probability value is greater than a first threshold, the processing steps in the repair instruction set are marked as an emergency processing level; When the fault probability value is less than or equal to the first threshold and greater than the second threshold, the processing steps in the repair instruction set are marked as a priority processing level; When the fault probability value is less than or equal to the second threshold, the processing steps in the repair instruction set are marked as a conventional processing level.

4. The fault analysis method for vehicle computer automated testing equipment according to claim 2, characterized in that: The fault diagnosis model adopts an improved gradient boosting decision tree structure for supervision; the improved gradient boosting decision tree structure includes a backbone decision network and an auxiliary verification network; the backbone decision network is composed of a number of decision trees, and the depth of each decision tree is set to 8 layers; the auxiliary verification network adopts a three-layer fully connected structure; the method for generating the fault type determination result is, The fault feature vector is divided into a hardware fault feature subset, a software fault feature subset and an operation fault feature subset through the backbone decision network, and the fault type score of each subset is calculated respectively. The specific formula is as follows: Among them, Score k is the score of the k-th fault feature subset, is the predicted weight value of the jth leaf node in the kth fault feature subset, T is the total number of decision trees, is the weight coefficient of the tth tree in the kth type of fault, X k is the input vector of the k-th fault feature subset, f t () is the prediction function of the t-th decision tree, γ k is the feature importance weight of the k-th fault feature subset, and Corr() is the feature correlation adjustment coefficient within the subset; Based on a preset historical fault case library, a feature similarity matrix is ​​calculated through the auxiliary verification network; Performing weighted fusion on the fault type score and the feature similarity matrix to generate a fault type determination result, wherein the fault type determination result includes a hardware fault code, a software fault code, and an operation fault code; The fault type determination result is subjected to probability normalization processing, and various fault probability values ​​are output, wherein the fault probability value reflects the confidence level of the fault occurrence.

5. The fault analysis method for vehicle computer automated testing equipment according to claim 4, characterized in that: The method for generating the fault feature vector is: The deep learning model includes a multi-layer convolutional neural network and a long short-term memory network, wherein the multi-layer convolutional neural network contains three convolutional layers and two pooling layers; the long short-term memory network is provided with a memory unit; Extracting spatial features of hardware operating parameters in the test database through the multi-layer convolutional neural network to generate hardware feature tensors; At the same time, the long short-term memory network is used to extract time series features from software exception logs and manual operation records to generate software operation feature tensors; Inputting the hardware feature tensor and the software operation feature tensor into an attention mechanism layer to calculate a key feature score, wherein the key feature score indicates the degree of correlation between hardware anomalies, software failures, and operation errors; The hardware feature tensor and the software operation feature tensor are feature fused based on the key feature score to form a fault feature vector, wherein each dimension in the fault feature vector corresponds to a potential fault feature.

6. The fault analysis method for vehicle computer automated testing equipment according to claim 5, characterized in that: The test database includes vehicle computer test data; the vehicle computer test data includes hardware operation parameters, software abnormality logs and manual operation records; The hardware operating parameters include CPU usage, memory occupancy, hard disk read and write rates, system bus bandwidth, and power management status; the software exception log includes application crash records, system service exception information, and resource scheduling conflict information; the manual operation records include touch command sequences, key operation timings, and voice control commands.

7. The fault analysis method for vehicle computer automated testing equipment according to claim 6, characterized in that: The method for obtaining the test database is: The hardware operation parameters are converted into digital signals through a data acquisition card, and the hardware operation parameters, the software abnormality log and the manual operation record are standardized through a data acquisition module to generate vehicle computer operation status information; The vehicle computer operating status information is written into the time series data table in the test database in a unified timestamp format.

8. A fault analysis system for vehicle-mounted automated test equipment, based on the fault analysis method for vehicle-mounted automated test equipment according to any one of claims 1 to 7, characterized in that: include, The acquisition module is used to collect the vehicle computer test data of the vehicle computer automation test equipment and store it in the test database according to the time series; A feature extraction module, used to extract features from the test database through a constructed deep learning model to generate a fault feature vector; A generation module, used for inputting the fault feature vector into the constructed fault diagnosis model to generate a fault type determination result and a fault probability value; The analysis and processing module is used to select corresponding fault repair solutions from a preset fault repair strategy library according to the fault type determination result and the fault probability value, and sort the fault repair solutions from high to low according to the fault probability value to form a fault handling report.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the fault analysis method for vehicle-mounted automated testing equipment described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the fault analysis method for vehicle-mounted automated testing equipment described in any one of claims 1 to 7 are implemented.

Images