Vehicle software flashing effect evaluation method and device based on association attribution analysis

Abstract

The invention relates to a vehicle software flashing effect evaluation method and device based on association attribution analysis. The method comprises the following steps: acquiring multidimensional data flashed by vehicle software in real time; through a Pearson correlation coefficient and a Granger causal test, an incidence relation matrix is generated; clustering the multi-dimensional data through a K-means clustering algorithm to obtain key parameters of the multi-dimensional data; on the basis of the incidence relation matrix, the key parameters and the flashing operation, through a Binseg algorithm and geometric topology characteristics, constructing a space-time evidence chain of vehicle software flashing; based on the space-time evidence chain, through Shapley value calculation, determining a plurality of flash factors; eliminating interference factors from the plurality of flashing factors through matching contrast and logistic regression of multi-dimensional data; and on the basis of the plurality of flashing factors after the interference factors are eliminated, calculating a software flashing score through an analytic hierarchy process. According to the method, through multi-dimensional data association analysis, evidence weight optimization and contribution degree quantification, the evaluation accuracy, the operation and maintenance efficiency and the flashing success rate are improved.

Description

Technical Field

[0001] This invention belongs to the field of vehicle software maintenance technology, specifically relating to a method and apparatus for evaluating the effect of vehicle software flashing based on association attribution analysis. Background Technology

[0002] Current vehicle software flashing effectiveness evaluation primarily employs traditional verification methods, including the following technical solutions: 1. **Indicator Comparison Analysis:** This method judges the flashing effect by comparing differences in system performance indicators before and after the flash. While simple, it cannot distinguish whether the indicator changes are caused by the flashing operation or other factors. 2. **Rule Matching Verification:** This method checks whether the system state after flashing meets expectations based on a predefined rule base. It relies on human experience and struggles to handle complex and unforeseen scenarios. 3. **Statistical Analysis Detection:** This method uses statistical methods to detect significant differences in system behavior before and after the flash. While it can identify changes, it cannot explain the causes and mechanisms of these changes. The above techniques have inherent problems. The above technologies have the following technical defects: 1. Insufficient attribution ability: It is unable to accurately determine the true relationship between system state changes and write operations, and is easily affected by other factors.

[0003] 2. Incomplete chain of evidence: There is a lack of systematic methods to construct and verify the chain of evidence for the brushing effect, resulting in limited credibility of the assessment conclusions.

[0004] 3. Difficulty in multi-factor analysis: It is difficult to distinguish the combined effects of multiple factors on the system state in a complex vehicle operating environment.

[0005] 4. Weak predictive guidance: It is unable to predict the possible effects of new flashing operations based on historical data, and lacks forward-looking guidance capabilities.

[0006] 5. Insufficient decision support: Evaluation results often remain at the level of describing phenomena, making it difficult to provide in-depth decision-making basis for subsequent optimization. Summary of the Invention

[0007] To address the problems raised in the background art, in one aspect of the present invention, a method for evaluating the effectiveness of vehicle software rewriting based on association attribution analysis is provided, comprising: real-time acquisition of multidimensional data on vehicle software rewriting, wherein the multidimensional data includes: rewriting operations, CAN bus data, software data, sensor data, and logs; generating an association matrix using Pearson correlation coefficient and Granger causality test; clustering the multidimensional data using K-means clustering algorithm to obtain key parameters of the multidimensional data; constructing a spatiotemporal evidence chain of vehicle software rewriting based on the association matrix, key parameters, and rewriting operations using Binseg algorithm and geometric topological features; determining multiple rewriting factors based on the spatiotemporal evidence chain using Shapley value calculation; eliminating interfering factors from the multiple rewriting factors through multidimensional data matching and logistic regression; and calculating the software rewriting score based on the multiple rewriting factors after eliminating interfering factors using analytic hierarchy process (AHP).

[0008] In some embodiments of the present invention, generating an association matrix by means of Pearson correlation coefficient and Granger causality test includes: determining the association relationship of the multidimensional data by calculating the Pearson correlation coefficient and mutual information value of the multidimensional data; testing the association relationship by means of Granger causality test; and generating an association matrix based on the test results.

[0009] In some embodiments of the present invention, the step of constructing a spatiotemporal evidence chain for vehicle software flashing based on the association matrix, key parameters, and flashing operation, using the Binseg algorithm and geometric topological features, includes: detecting abrupt changes in each key parameter before and after the flashing operation using the Binseg algorithm; constructing a temporal evidence chain based on the abrupt changes; and constructing a spatial evidence chain based on the consistency of the connection relationships and change trends of the key parameters in the vector space of multidimensional data.

[0010] Furthermore, the construction of the spatiotemporal evidence chain of vehicle software flashing based on the correlation matrix, key parameters, and flashing operation, through the Binseg algorithm and geometric topological features, also includes: calculating the reliability score of the spatiotemporal evidence chain based on the consistency between the temporal evidence chain and the spatial evidence chain.

[0011] refer to Figure 2In some embodiments of the present invention, the step of determining multiple writing factors based on the spatiotemporal evidence chain and through Shapley value calculation includes: constructing multiple candidate feature combinations based on the spatiotemporal evidence chain and environmental data; determining weights based on the number of features in each candidate feature combination; calculating the Shapley value of each candidate feature combination based on the weights; calculating the contribution of each candidate feature combination based on the ratio of the Shapley value of each feature combination to the sum of the Shapley values ​​of all candidate feature combinations; and determining multiple writing factors based on the contribution of each candidate feature combination.

[0012] In some embodiments of the present invention, the step of removing interfering factors from the plurality of brushing factors through multidimensional data matching and logistic regression includes: selecting multiple covariates from the multidimensional data; determining the brushing vehicle group and the corresponding control vehicle group based on the multiple covariates through logistic regression and nearest neighbor matching algorithms; comparing the change values ​​of the brushing vehicle group and the control vehicle group on each brushing factor, and removing factors whose change values ​​are lower than a preset threshold from the plurality of brushing factors.

[0013] A second aspect of the present invention provides a vehicle software rewriting effect evaluation device based on association attribution analysis, comprising: an acquisition module for acquiring multidimensional data of vehicle software rewriting in real time, the multidimensional data including: rewriting operations, CAN bus data, software data, sensor data, and logs; a clustering module for generating an association matrix using Pearson correlation coefficient and Granger causality test; clustering the multidimensional data using K-means clustering algorithm to obtain key parameters of the multidimensional data; a determination module for constructing a spatiotemporal evidence chain of vehicle software rewriting based on the association matrix, key parameters, and rewriting operations using Binseg algorithm and geometric topological features; determining multiple rewriting factors based on the spatiotemporal evidence chain using Shapley value calculation; a removal module for removing interfering factors from the multiple rewriting factors through multidimensional data matching and logistic regression; and a calculation module for calculating the software rewriting score based on the multiple rewriting factors after removing interfering factors using analytic hierarchy process (AHP).

[0014] A third aspect of the present invention provides an electronic device comprising: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the vehicle software flashing effect evaluation method based on association attribution analysis provided in the first aspect of the present invention.

[0015] In a fourth aspect, the present invention provides a computer-readable medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the vehicle software rewriting effect evaluation method based on association attribution analysis provided in the first aspect of the present invention.

[0016] The beneficial effects of this invention are: This invention improves assessment quality and operational efficiency by employing a multi-dimensional evidence association analysis architecture, a dynamic association network construction mechanism, an evidence weight optimization algorithm, and a contribution quantification analysis engine. It reduces the false positive rate to below 8%, achieves 95% consistency in assessment results, shortens problem analysis time by 60%, improves decision-making efficiency by 45%, and reduces the incidence of recurring problems by 50%. It also reduces invalid writes and repetitive work, lowers costs by 35%, increases write success rate, reduces after-sales service expenses, and avoids potential major losses through preventative analysis. Attached Figure Description

[0017] Figure 1 This is a basic flowchart illustrating the vehicle software flashing effect evaluation method based on association attribution analysis in some embodiments of the present invention. Figure 2 This is a schematic diagram illustrating the specific process of constructing a spatiotemporal evidence chain in some embodiments of the present invention; Figure 3 This is a schematic diagram of the vehicle software flashing effect evaluation device based on association attribution analysis in some embodiments of the present invention. Figure 4 This is a schematic diagram of the structure of an electronic device in some embodiments of the present invention. Detailed Implementation

[0018] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0019] Example 1 refer to Figure 1 and Figure 2In a first aspect, the present invention provides a method for evaluating the effect of vehicle software rewriting based on association attribution analysis, comprising: S100. acquiring multidimensional data of vehicle software rewriting in real time, the multidimensional data including: rewriting operations, CAN bus data, software data, sensor data, and logs; S200. generating an association matrix through Pearson correlation coefficient and Granger causality test; clustering the multidimensional data using K-means clustering algorithm to obtain key parameters of the multidimensional data; S300. constructing a spatiotemporal evidence chain of vehicle software rewriting based on the association matrix, key parameters, and rewriting operations using Binseg algorithm and geometric topological features; determining multiple rewriting factors based on the spatiotemporal evidence chain by Shapley value calculation; S400. removing interfering factors from the multiple rewriting factors through multidimensional data matching and logistic regression; S500. calculating the software rewriting score based on the multiple rewriting factors after removing interfering factors using analytic hierarchy process.

[0020] In step S100 of some embodiments of the present invention, multi-dimensional data of vehicle software flashing is acquired in real time. The multi-dimensional data includes: flashing operation, CAN bus data, software data, sensor data and logs. Specifically, multi-dimensional data acquisition and preprocessing are performed. Inputs include real-time vehicle CAN bus streaming data (such as ECU ID, CPU load, memory usage, and message frequency), software repository metadata (version number, checksum, and flash timestamp), environmental sensor data (GPS positioning, network signal strength, and ambient temperature), and diagnostic system logs (fault codes and system status change records). The processing includes data access and parsing: parsing the CAN bus DBC file using the CANoe configuration template to extract key ECU parameters; deploying a version sniffing agent to periodically capture software version information and calculate hash values; and establishing an environmental data acquisition channel to sample sensor data once per second. Data cleaning and repair includes: detecting missing data periods (marking data segments with more than 5 consecutive missing sampling points), handling missing values ​​(using linear interpolation for random missing points and forward padding for consecutive missing segments), and outlier removal (calculating the 3σ range of the 30-day historical distribution of each parameter and removing values ​​outside the range). Data standardization and alignment include: numerical standardization (scaling each parameter to the [0,1] interval using MinMaxScaler), time axis alignment (establishing an analysis window of [-30min, +60min] based on the completion time T0), and time synchronization (using the NTP protocol to ensure time errors of all data sources are <100ms). The output is a time-aligned standardized data table (containing vehicle ID, timestamp, parameter values, etc.) and a data quality report (completeness score, proportion of outlier data).

[0021] In step S200 of some embodiments of the present invention, generating an association matrix by means of Pearson correlation coefficient and Granger causality test includes: determining the association relationship of the multidimensional data by calculating the Pearson correlation coefficient and mutual information value of the multidimensional data; testing the association relationship by means of Granger causality test; and generating an association matrix based on the test results.

[0022] Specifically, the input is a preprocessed, standardized data table. Processing includes correlation calculation: calculating the Pearson correlation coefficient for all parameter pairs and retaining linear associations with |r|>0.3; calculating the mutual information value of parameter pairs and retaining non-linear associations with MI>0.2; and performing a Granger causality test (lag=5, p<0.05) on software version and performance metrics. Pattern recognition and dimensionality reduction: using K-means (k=3) to cluster parameter change patterns within one hour after the refresh; performing PCA on 20 core parameters and selecting principal components with a cumulative contribution rate >95%; and calculating the importance score of each parameter to the refresh effect based on random forest. Output includes a correlation matrix (parameter pairs, correlation types, and strength values), parameter cluster labels (the refresh effect category to which each vehicle belongs), and a list of core parameters (the top 10 key parameters sorted by importance).

[0023] In step S300 of some embodiments of the present invention, the construction of a spatiotemporal evidence chain of vehicle software flashing based on the association matrix, key parameters, and flashing operation using the Binseg algorithm and geometric topological features includes: detecting abrupt change points of each key parameter before and after the flashing operation using the Binseg algorithm; constructing a temporal evidence chain based on the abrupt change points; and constructing a spatial evidence chain based on the consistency of the connection relationship and change trend of the key parameters in the vector space of multidimensional data.

[0024] Furthermore, the construction of the spatiotemporal evidence chain of vehicle software flashing based on the correlation matrix, key parameters, and flashing operation, through the Binseg algorithm and geometric topological features, also includes: calculating the reliability score of the spatiotemporal evidence chain based on the consistency between the temporal evidence chain and the spatial evidence chain.

[0025] Specifically, the inputs include correlation analysis results, time-series data, and write operation records. The processing involves: constructing a temporal evidence chain: using the Binseg algorithm to detect abrupt changes before and after the write operation for each key parameter, and checking whether these abrupt changes are concentrated within the [T0-2min, T0+10min] window for temporal consistency verification; constructing a spatial evidence chain: identifying closely connected parameter groups in the correlation network, and checking whether the direction of change of the correlation parameters after the write operation is consistent (both rising / falling) for direction verification; quantifying evidence quality: calculating the completeness score (0-1) based on the proportion of complete data points in the evidence chain, calculating a temporal consistency score (0-1) based on the proximity of the abrupt change point to the write operation time, calculating a logical consistency score (0-1) based on the consistency of the direction of change of the correlation parameters, and then calculating a comprehensive quality score of 0.3 × completeness + 0.4 × temporal consistency + 0.3 × logical consistency. The outputs include a set of evidence chains (each containing the parameters involved, the location of the abrupt change point, and the direction of change) and an evidence quality scoring table (evidence ID, scores for each item, and comprehensive quality score).

[0026] In step S300 of some embodiments of the present invention, determining multiple writing factors based on the spatiotemporal evidence chain and through Shapley value calculation includes: constructing multiple candidate feature combinations based on the spatiotemporal evidence chain and environmental data; determining weights based on the number of features in each candidate feature combination; calculating the Shapley value of each candidate feature combination based on the weights; calculating the contribution of each candidate feature combination based on the ratio of the Shapley value of each feature combination to the sum of the Shapley values ​​of all candidate feature combinations; and determining multiple writing factors based on the contribution of each candidate feature combination.

[0027] Specifically, the input includes a set of evidence chains, parameter change data, and environmental factor data. Processing includes feature subset generation: constructing all possible feature combinations (e.g., "software version only," "software version + temperature," "temperature only," etc.) and training a prediction model (e.g., gradient boosting tree) for each subset to predict system state changes; Shapley value calculation: for each feature subset, calculating the model performance improvement after adding the "software brushing" feature as the marginal contribution, assigning weights according to subset size (smaller subsets have larger weights), and then summing the marginal contributions of all subsets according to their weights to obtain the Shapley value for software brushing; Contribution normalization: normalizing the Shapley values ​​of all factors to a percentage form, with the software brushing contribution equal to the software Shapley value divided by the sum of all factor Shapley values. Output includes a contribution distribution table (factor name, contribution value, confidence interval) and a software brushing-specific contribution (e.g., brushing operation contributed 42% to the system performance change).

[0028] In step S400 of some embodiments of the present invention, the step of removing interfering factors from the plurality of brushing factors through multidimensional data matching and logistic regression includes: selecting a plurality of covariates from the multidimensional data; determining the brushing vehicle group and the corresponding control vehicle group based on the plurality of covariates through logistic regression and nearest neighbor matching algorithms; comparing the change values ​​of the brushing vehicle group and the control vehicle group on each brushing factor, and removing factors whose change values ​​are lower than a preset threshold from the plurality of brushing factors.

[0029] Specifically, the input includes all vehicle data, flashing tags, and vehicle attribute data. Processing involves: 10 covariates are selected, including vehicle model, mileage, and usage region. A logistic regression model is trained using flashed / unflashed as labels to obtain a propensity score, and the probability of each vehicle being selected for the flashing group is calculated. Matching execution: Nearest neighbor matching is used to find the unflashed vehicle with the closest propensity score for each flashed vehicle group (caliper = 0.02), and the distribution of the two groups on the covariates is checked for balance after matching (p > 0.05). Net effect evaluation: The difference in average parameter changes between the flashed group and the matched control group is calculated; the net effect equals the change in the flashed group minus the change in the control group. Output includes a matching table (flashed vehicle IDs, matched unflashed vehicle IDs) and a net effect evaluation report (the true flashing effect value after excluding interference).

[0030] In step S500 of some embodiments of the present invention, the score of software brushing is calculated by the hierarchical analysis method based on multiple brushing factors after removing interference factors.

[0031] Specifically, the input includes net effect data and multi-dimensional indicator definitions. Processing includes indicator-level calculations: system stability (failure rate variation, number of crashes, number of abnormal restarts), functional integrity (functional availability, response time, processing accuracy), and user experience (operational smoothness, interface responsiveness, user satisfaction, if applicable); dimension-level aggregation: system stability score = 0.4 × failure rate + 0.3 × performance fluctuation + 0.3 × resource efficiency, functional integrity score = 0.5 × functional availability + 0.3 × response time + 0.2 × processing accuracy, user experience score = 0.6 × operational smoothness + 0.4 × interface responsiveness; comprehensive score calculation: comprehensive effect score = 0.4 × system stability + 0.3 × functional integrity + 0.3 × user experience, and the results are standardized and mapped to the [0, 100] range. Outputs include a hierarchical scoring table (indicator-level score, dimension-level score, comprehensive score) and effect level labels (excellent ≥90, good ≥75, average ≥60, poor <60).

[0032] Specifically, it also includes risk assessment and early warning. Inputs include effect evaluation results and historical failure data. Processing involves quantifying risk indicators: system stability risk is calculated based on the increase and severity of failure rate; core function risk is based on the assessment of the possibility of key function failure; performance degradation risk is calculated based on the decline in performance indicators; and user complaint risk is based on the predicted complaint rate of similar historical writes. Risk score calculation: Risk score = 0.4 × stability risk + 0.3 × functional risk + 0.2 × performance risk + 0.1 × complaint risk, and normalized to the [0,1] interval. Risk level classification: high risk (≥0.7) requires immediate remedial measures; medium risk (≥0.3) requires monitoring and preparation of optimization plans; and low risk (<0.3) can be observed normally. Outputs include a risk scoring card (various risk values, comprehensive risk score, risk level) and early warning suggestions (immediate action, key attention, normal observation).

[0033] And, optimization suggestion generation. Inputs include contribution analysis, risk assessment, and a historical optimization library. Processing includes suggestion template matching: identifying key issues (selecting factors with a contribution > 0.1 and positively correlated with negative effects), and matching solution templates for similar issues in the historical optimization library; personalized suggestion generation: replacing general parameters in the templates with current specific values, and adjusting the implementation intensity of the suggestions according to the severity of the issues; priority ranking: calculating the expected impact based on contribution and risk value (0.7 weight), estimating the implementation cost based on required resources and time (0.3 weight), and the priority score = 0.7 × expected impact + 0.3 × (1 - implementation cost). Outputs include a list of optimization suggestions (suggestion content, expected effect, implementation cost, priority) and an implementation roadmap (high-priority suggestions to be implemented immediately, medium-priority suggestions to be included in the plan).

[0034] Continuous learning and optimization. Inputs include historical evaluation data, new write cases, and model performance metrics. Processing includes performance monitoring: collecting accuracy, recall, and F1 scores from the last 5 evaluations, and calculating the percentage difference between the current average performance and the historical average; triggering mechanism: if performance drops by more than 5% or the performance is poor in 3 consecutive evaluations, a model update is triggered; model iteration includes data augmentation (adding newly validated write cases to the training set), parameter retuning (retraining core models such as association analysis and contribution calculation), A / B testing (comparing the performance of the old and new models on the test set), and model deployment (gradually replacing the online model after confirming performance improvement). Outputs include a model performance report (current performance, historical comparison, and update suggestions), updated evaluation model files, and version update logs. The final output summary includes a structured evaluation report (including effect score, risk level, and contribution analysis), actionable optimization suggestions (specific improvement measures ranked by priority), a model update package (the continuously optimized algorithm model), and a decision support dashboard (a visual interface for monitoring write performance).

[0035] Example 2 refer to Figure 3 In a second aspect, the present invention provides a vehicle software rewriting effect evaluation device 1 based on association attribution analysis, comprising: an acquisition module 11, used to acquire multidimensional data of vehicle software rewriting in real time, the multidimensional data including: rewriting operations, CAN bus data, software data, sensor data, and logs; a clustering module 12, used to generate an association matrix through Pearson correlation coefficient and Granger causality test; and to cluster the multidimensional data using K-means clustering algorithm to obtain key parameters of the multidimensional data; a determination module 13, used to construct a spatiotemporal evidence chain of vehicle software rewriting based on the association matrix, key parameters, and rewriting operations, using the Binseg algorithm and geometric topological features; and to determine multiple rewriting factors based on the spatiotemporal evidence chain by Shapley value calculation; a removal module 14, used to remove interfering factors from the multiple rewriting factors through multidimensional data matching and logistic regression; and a calculation module 15, used to calculate the software rewriting score based on the multiple rewriting factors after removing interfering factors by using the analytic hierarchy process.

[0036] Furthermore, the clustering module 12 includes: a determination unit, used to determine the association relationship of the multidimensional data by calculating the Pearson correlation coefficient and mutual information value of the multidimensional data; and a generation unit, used to test the association relationship by Granger causality test and generate an association relationship matrix based on the test results.

[0037] Example 3 refer to Figure 4 A third aspect of the present invention provides an electronic device comprising: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method of the first aspect of the present invention.

[0038] Electronic device 500 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 501, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 502 or a program loaded from storage device 508 into random access memory (RAM) 503. The RAM 503 also stores various programs and data required for the operation of electronic device 500. The processing unit 501, ROM 502, and RAM 503 are interconnected via bus 504. An input / output (I / O) interface 505 is also connected to bus 504.

[0039] Typically, the following devices can be connected to I / O interface 505: input devices 506 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 507 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 508 including, for example, hard disks; and communication devices 509. Communication device 509 allows electronic device 500 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 4 An electronic device 500 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 4 Each box shown can represent a device or multiple devices as needed.

[0040] Specifically, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 509, or installed from a storage device 508, or installed from a ROM 502. When the computer program is executed by a processing device 501, it performs the functions defined in the methods of embodiments of this disclosure. It should be noted that the computer-readable medium described in embodiments of this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0041] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more computer programs, which, when executed by the electronic device, cause the electronic device to: Computer program code for performing the operations of embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, C++, and Python—and conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0042] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0043] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for evaluating the effectiveness of vehicle software flashing based on association attribution analysis, characterized in that, include: Real-time acquisition of multi-dimensional data from vehicle software flashing, including flashing operations, CAN bus data, software data, sensor data, and logs; A correlation matrix is ​​generated using Pearson correlation coefficient and Granger causality test; the multidimensional data is then clustered using K-means clustering algorithm to obtain the key parameters of the multidimensional data. Based on the aforementioned correlation matrix, key parameters, and flashing operations, a spatiotemporal evidence chain of vehicle software flashing is constructed using the Binseg algorithm and geometric topological features; based on the spatiotemporal evidence chain, multiple flashing factors are determined through Shapley value calculation. By matching and comparing multidimensional data and performing logistic regression, interfering factors are eliminated from the multiple brushing factors. Based on multiple brushing factors after eliminating interfering factors, the score of software brushing is calculated using the analytic hierarchy process (AHP).

2. The method for evaluating the effect of vehicle software flashing based on association attribution analysis according to claim 1, characterized in that, The process of generating the association matrix using Pearson correlation coefficient and Granger causality test includes: The association relationships of the multidimensional data are determined by calculating the Pearson correlation coefficient and mutual information value of the multidimensional data; The association was tested using the Granger causality test, and an association matrix was generated based on the test results.

3. The method for evaluating the effect of vehicle software flashing based on association attribution analysis according to claim 1, characterized in that, The construction of the spatiotemporal evidence chain for vehicle software flashing based on the aforementioned correlation matrix, key parameters, and flashing operations, using the Binseg algorithm and geometric topological features, includes: The Binseg algorithm is used to detect abrupt changes in each key parameter before and after the write operation; a time-based evidence chain is constructed based on these abrupt changes. Based on the consistency of the connection relationships and changing trends of key parameters in the vector space of multidimensional data, a spatial evidence chain is constructed.

4. The method for evaluating the effect of vehicle software flashing based on association attribution analysis according to claim 3, characterized in that, The construction of the spatiotemporal evidence chain for vehicle software flashing based on the aforementioned correlation matrix, key parameters, and flashing operations, using the Binseg algorithm and geometric topological features, further includes: Based on the consistency between the temporal and spatial evidence chains, the reliability score of the spatiotemporal evidence chain is calculated.

5. The method for evaluating the effect of vehicle software flashing based on association attribution analysis according to claim 1, characterized in that, Based on the spatiotemporal evidence chain, the multiple write factors determined through Shapley value calculation include: Based on the spatiotemporal evidence chain and environmental data, multiple candidate feature combinations are constructed; The weights are determined based on the number of features in each candidate feature combination. Based on the weights, calculate the Shapley value for each candidate feature combination; based on the ratio of the Shapley value of each feature combination to the sum of the Shapley values ​​of all candidate feature combinations, calculate the contribution of each candidate feature combination. Based on the contribution of each candidate feature combination, multiple brushing factors are determined.

6. The method for evaluating the effect of vehicle software flashing based on association attribution analysis according to claim 1, characterized in that, The process of eliminating interfering factors from the multiple brushing factors through multidimensional data matching and logistic regression includes: Multiple covariates are selected from the multidimensional data; based on the multiple covariates, the vehicle groups to be smeared and the corresponding control vehicle groups are determined by logistic regression and nearest neighbor matching algorithms. The changes in each writing factor are compared between the written vehicle group and the control vehicle group, and factors with changes below a preset threshold are removed from the multiple writing factors.

7. A device for evaluating the effect of vehicle software flashing based on association attribution analysis, characterized in that, include: The acquisition module is used to acquire multi-dimensional data of vehicle software flashing in real time. The multi-dimensional data includes: flashing operation, CAN bus data, software data, sensor data and logs. The clustering module is used to generate an association matrix through Pearson correlation coefficient and Granger causality test; and to cluster the multidimensional data using the K-means clustering algorithm to obtain the key parameters of the multidimensional data. The determination module is used to construct a spatiotemporal evidence chain of vehicle software flashing based on the correlation matrix, key parameters, and flashing operations, using the Binseg algorithm and geometric topological features; and to determine multiple flashing factors based on the spatiotemporal evidence chain by calculating Shapley values. The elimination module is used to eliminate interfering factors from the multiple brushing factors by matching and comparing multidimensional data and logistic regression. The calculation module is used to calculate the score of the software's brushing based on multiple brushing factors after removing interference factors, using the analytic hierarchy process.

8. The vehicle software flashing effect evaluation device based on association attribution analysis according to claim 7, characterized in that, The clustering module includes: The determining unit is used to determine the association relationship of the multidimensional data by calculating the Pearson correlation coefficient and mutual information value of the multidimensional data; The generation unit is used to test the association relationship through Granger causality test and generate an association relationship matrix based on the test results.

9. An electronic device, comprising: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the vehicle software flashing effect evaluation method based on association attribution analysis as described in any one of claims 1 to 6.

10. A computer-readable medium having a computer program stored thereon, wherein, When the computer program is executed by the processor, it implements the vehicle software rewriting effect evaluation method based on association attribution analysis as described in any one of claims 1 to 6.

Images