Intelligent driving vehicle fault accurate evaluation method and system, and medium

Abstract

The invention relates to an intelligent driving vehicle fault accurate evaluation method and system and a medium, and relates to the technical field of intelligent driving, and the method comprises the steps: obtaining fault related data of an intelligent driving vehicle; calculating a fault exposure degree score based on the fault related data, wherein the fault exposure degree score is used for representing the possibility that a fault is found by a user; calculating a user experience score based on the takeover mileage index and the tolerance of the user to the fault degradation frequency; according to the fault exposure degree score and the user experience score, determining a fault exact grade through an exposure degree-user experience two-dimensional score matrix; and outputting a corresponding out-of-order strategy according to the fault out-of-order level. According to the invention, the technical indexes and the user experience can be effectively combined, so that the rationality and scientificity of the fault standard of the intelligent driving vehicle are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving technology, and in particular to a method, system and medium for assessing fault exit of intelligent driving vehicles. Background Technology

[0002] With the rapid development of intelligent connected vehicle technology, the functions involved in intelligent driving vehicles are becoming increasingly complex, and the number of related faults is often several times that of non-intelligent driving vehicles, potentially reaching hundreds or even thousands. For fault exit design of intelligent driving vehicles, current methods used by companies to design exit standards include: referencing engineering experience from other projects, lessons learned by the company, and the company's accumulated fault database. Regarding the evaluation of intelligent driving functions, existing technologies collect users' natural driving data to evaluate the functions of the vehicle's advanced driver assistance systems from safety and comfort dimensions. In terms of evaluating the effectiveness of intelligent driving systems, existing technologies use multi-source data mining to establish vehicle models, random traffic scenario models, and occupant injury models to evaluate the effectiveness of intelligent driving systems in reducing occupant injury risks. Regarding the safety assessment of expected autonomous driving functions, existing technologies construct a mapping relationship between vehicle-level hazards, performance defects, and triggering conditions to assess the tolerance of triggering conditions.

[0003] However, the aforementioned existing technologies have certain limitations in the fault exit design of intelligent driving vehicles. Existing fault exit design methods are primarily based on experience, lacking systematic theoretical guidance and feedback from user experience, and neglecting the user's true tolerance for faults. Existing functional evaluation methods mainly focus on safety and comfort dimensions, failing to incorporate user tolerance for fault degradation frequency as an evaluation parameter for fault exit. Existing system effectiveness evaluation methods emphasize the assessment of occupant injury risk, failing to establish quantitative standards for fault exit from a user experience perspective. Existing expected functional safety assessment methods mainly evaluate triggering conditions, failing to comprehensively consider the correlation between fault exposure and user experience. Therefore, a fault exit assessment method for intelligent driving vehicles is needed that can balance technical thinking and user experience, incorporating the user group's tolerance for fault phenomena into the evaluation system. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a method, system and medium for evaluating fault exit criteria for intelligent driving vehicles, which can effectively combine technical indicators and user experience, thereby improving the rationality and scientificity of fault exit criteria for intelligent driving vehicles.

[0005] To achieve the above objectives, the present invention adopts the following technical solution.

[0006] Firstly, the present invention provides a fault exit assessment method for intelligent driving vehicles, which adopts the following technical solution: A fault exit assessment method for intelligent driving vehicles includes: Acquire fault-related data from intelligent driving vehicles; A fault exposure score is calculated based on the fault-related data, and the fault exposure score is used to characterize the likelihood that the fault will be discovered by the user. User experience scores are calculated based on takeover mileage metrics and user tolerance for the frequency of failure degradation. Based on the fault exposure score and the user experience score, a fault clearance level is determined using a two-dimensional scoring matrix of exposure-user experience; and Output the corresponding admission strategy based on the fault admission level.

[0007] Furthermore, in the aforementioned intelligent driving vehicle fault exit assessment method, the fault-related data is collected from at least one of the following channels: vehicle logs, cloud diagnostics, and after-sales or customer service feedback.

[0008] Furthermore, in the aforementioned intelligent driving vehicle fault exit assessment method, the fault exposure score is obtained by combining multiple sub-indicators, which include: Vehicle coverage rate is the ratio of the number of vehicles that experienced the fault to the total number of vehicles in the statistical period. Intelligent driving state trigger rate, representing the ratio of the number of triggers while intelligent driving is enabled to the total number of times intelligent driving is enabled or the duration of intelligent driving activation; and Exposure intensity is expressed as the number of triggers per thousand kilometers or per hour.

[0009] Furthermore, in the above-mentioned intelligent driving vehicle fault exit assessment method, the method also includes scene correction of the vehicle coverage rate. When the fault mainly occurs in the intelligent driving activation scenario, the weights of the intelligent driving state trigger rate and the exposure intensity are increased.

[0010] Furthermore, in the above-mentioned intelligent driving vehicle fault exit assessment method, the method also includes introducing a sample confidence factor, which is used to characterize the impact of statistical sample size, statistical period length, and version coverage on exposure stability. When the sample confidence factor is lower than a preset threshold, a conditional conclusion is output.

[0011] Furthermore, in the above-mentioned intelligent driving vehicle fault exit assessment method, the user experience score is jointly determined by the frequency dimension experience score and the impact dimension experience score, and the impact dimension experience score is used to characterize the perceived severity and recoverability of the fault to the user.

[0012] Furthermore, in the aforementioned intelligent driving vehicle fault exit assessment method, the impact dimension experience score is determined based on at least one of the following factors: Whether it triggers takeover, emergency braking, or steering intervention interruption; Does it cause core intelligent driving functions to become unavailable or significantly degraded? Whether obvious human-computer interaction alarms are generated and the duration of the alarms; and Whether it can be automatically restored or requires manual intervention.

[0013] Furthermore, in the aforementioned intelligent driving vehicle fault exit assessment method, the user experience score calculated based on the takeover mileage index and the user's tolerance for fault degradation frequency includes: The average user tolerance standard for takeover mileage is calculated, which is equal to the product of the average monthly intelligent driving mileage per vehicle, the user's average tolerance for fault degradation frequency, and the number of faults associated with the intelligent driving function; and The user experience score is determined by the ratio of the actual fault takeover mileage to the average user tolerance standard for the takeover mileage.

[0014] Furthermore, in the aforementioned intelligent driving vehicle fault exit assessment method, the user's tolerance for the frequency of fault degradation is obtained in layers according to user groups and usage scenarios to obtain different tolerance parameters. The user groups include high-frequency intelligent driving users and low-frequency users, and the usage scenarios include urban commuting and long-distance highway travel.

[0015] Furthermore, in the aforementioned intelligent driving vehicle fault exit assessment method, the exit strategy includes a structured exit action package, which includes: The result of the determination of whether or not to grant permission or conditional permission; Fault location directions that need to be optimized first; User suggestions, explanations, or downgrade strategies; and Version release strategy, which includes grayscale ratio, target audience, monitoring metrics and rollback conditions.

[0016] Secondly, the present invention provides an intelligent driving vehicle fault exit assessment system, which adopts the following technical solution: A fault exit assessment system for intelligent driving vehicles includes: The data acquisition module is configured to acquire fault-related data of intelligent driving vehicles; The exposure calculation module is configured to calculate a fault exposure score based on the fault-related data, wherein the fault exposure score is used to characterize the likelihood that the fault will be discovered by the user. The user experience calculation module is configured to calculate the user experience score based on the takeover mileage metric and the user's tolerance for the frequency of failure degradation. The fault clearance module is configured to determine the fault clearance level based on the fault exposure score and the user experience score using a two-dimensional scoring matrix of exposure-user experience; and The strategy output module is configured to output the corresponding admission strategy according to the fault admission level.

[0017] Thirdly, the present invention provides a readable storage medium, which adopts the following technical solution: A readable storage medium storing computer instructions that, when executed by a processor, implement the intelligent driving vehicle fault clearance assessment method as described in any of the first aspects above.

[0018] In summary, compared with the prior art, the present invention has at least one of the following beneficial technical effects: The fault exit assessment method for intelligent driving vehicles provided by this invention, by acquiring fault-related data and calculating fault exposure scores and user experience scores, can quantitatively correlate technical fault performance with user tolerance levels, thereby establishing a more scientific and reasonable fault exit assessment system. Determining the fault exit level through a two-dimensional scoring matrix of exposure and user experience can, to some extent, avoid the subjectivity and uncertainty inherent in relying solely on engineering experience for exit judgments. Calculating the user experience score based on takeover mileage indicators and user tolerance for fault degradation frequency makes the exit standards closer to actual user experience, helping to improve user satisfaction and trust in intelligent driving functions. Outputting corresponding exit strategies based on the fault exit level provides structured decision support for product development and version releases, facilitating continuous iterative optimization of intelligent driving functions while ensuring user experience. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A flowchart of an embodiment of the intelligent driving vehicle fault clearance assessment method of the present invention is shown.

[0021] Figure 2 A flowchart of an embodiment of the fault exposure score calculation method of the present invention is shown.

[0022] Figure 3 This diagram illustrates an embodiment of the exposure grading definition in the exposure-user experience two-dimensional rating matrix of the present invention.

[0023] Figure 4 A flowchart of an embodiment of the user experience score calculation method of the present invention is shown.

[0024] Figure 5 A data table illustrating an embodiment of the user tolerance survey data of the present invention is shown.

[0025] Figure 6 This diagram illustrates an embodiment of the correspondence between the user experience level and the actual MPD range of the fault according to the present invention.

[0026] Figure 7 A schematic diagram of an embodiment of the exposure-user experience two-dimensional scoring matrix of the present invention is shown.

[0027] Figure 8 The flowchart illustrates an embodiment of the sample confidence factor judgment and admission strategy generation process of the present invention.

[0028] Figure 9 A data table is shown as an embodiment of the sensor signal timeout fault statistics of the present invention.

[0029] Figure 10 This diagram illustrates another embodiment of the correspondence between the user experience level and the actual MPD range of the fault according to the present invention.

[0030] Figure 11 The diagram shows an architecture block diagram of an embodiment of the intelligent driving vehicle fault exit assessment system of the present invention. Detailed Implementation

[0031] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Furthermore, it should be understood that the specific embodiments described herein are only for illustration and explanation of this application and are not intended to limit this application.

[0032] It should be noted that the order of description of the following embodiments is not intended to limit the preferred order of the embodiments of this application. Furthermore, the descriptions of each embodiment in the following embodiments have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0033] The method steps described in this embodiment of the invention can be executed in the order described in the specific implementation, or the execution order of each step can be adjusted according to actual needs, provided that the technical problem can be solved. These are not listed one by one here.

[0034] The present invention will be further described in detail below with reference to the accompanying drawings.

[0035] Reference Figure 1 , Figure 1 The flowchart illustrates an embodiment of a fault exit assessment method for intelligent driving vehicles according to the present invention. The fault exit assessment method for intelligent driving vehicles includes acquiring fault-related data of the intelligent driving vehicle, calculating a fault exposure score based on the fault-related data, calculating a user experience score based on the takeover mileage index and the user's tolerance for fault degradation frequency, determining the fault exit level through a two-dimensional scoring matrix of exposure-user experience, and outputting a corresponding exit strategy according to the fault exit level.

[0036] like Figure 1 As shown, the intelligent driving vehicle fault exit assessment method begins with step 100. In step 100, fault-related data of the intelligent driving vehicle is acquired. This fault-related data is collected from at least one of the following channels: vehicle logs, cloud-based diagnostics, and after-sales or customer service feedback. In some implementations, after acquiring the fault-related data, the definitions of faults, statistical periods, vehicle base, and triggering scenarios are standardized to ensure consistency in the data foundation for subsequent calculations.

[0037] Continue to refer to Figure 1 In step 102, a fault exposure score is calculated based on fault-related data. The fault exposure score characterizes the likelihood that a fault will be discovered by a user. The higher the proportion of faulty vehicles, the higher the risk exposed to the user, and the lower the corresponding fault exposure score.

[0038] In step 104, a user experience score is calculated based on the Disengagement Mileage (MPD) metric and the user's tolerance for the frequency of fault degradation. The MPD metric measures the frequency of disengagement due to faults during the use of intelligent driving functions. The user's tolerance for the frequency of fault degradation is obtained through user surveys and reflects the user group's acceptance of fault phenomena.

[0039] In step 106, the fault admission level is determined based on the fault exposure score and user experience score using a two-dimensional exposure-user experience scoring matrix. This matrix uses fault exposure and user experience as two scoring dimensions, and the corresponding fault admission level is obtained by mapping the scores of these two dimensions together.

[0040] In step 108, a corresponding exit strategy is output based on the fault exit level. The exit strategy provides corresponding design optimization directions, prompts or degrading strategies, and release strategies according to different fault exit levels, thereby guiding the fault design and exit decisions of intelligent driving vehicles.

[0041] Furthermore, refer to Figure 2 , Figure 2 The flowchart for calculating the fault exposure score is shown. The fault exposure score is obtained by combining multiple sub-indicators, including vehicle coverage, intelligent driving trigger rate, and exposure intensity.

[0042] like Figure 2 As shown, the fault exposure score calculation process begins with step 200. In step 200, the vehicle coverage rate E1 is calculated. The vehicle coverage rate E1 represents the ratio of the number of vehicles experiencing the fault to the total number of vehicles in the statistical period. The vehicle coverage rate E1 reflects the distribution range of the fault within the vehicle group; the higher the vehicle coverage rate E1, the wider the range of vehicles affected by the fault.

[0043] Continue to refer to Figure 2 In step 202, the intelligent driving state trigger rate E2 is calculated. The intelligent driving state trigger rate E2 represents the ratio of the number of triggers in the intelligent driving activated state to the number of times intelligent driving is activated or the duration of intelligent driving activation. The intelligent driving state trigger rate E2 is used to measure the frequency of fault occurrence during the activation of the intelligent driving function, thereby reflecting the degree of correlation between the fault and the intelligent driving function.

[0044] In step 204, the exposure intensity E3 is calculated. Exposure intensity E3 represents the number of triggers per thousand kilometers or per hour. Exposure intensity E3 is used to quantify the density of fault occurrence per unit of mileage or unit of time, thereby assessing the frequency of faults encountered by the user during actual use.

[0045] like Figure 2 As shown, after calculating the three sub-indicators E1 (vehicle coverage rate), E2 (intelligent driving trigger rate), and E3 (exposure intensity), the process proceeds to step 206. In step 206, it is determined whether the fault occurs in the intelligent driving activation scenario. The result of step 206 determines the subsequent processing method for the weights of each sub-indicator.

[0046] When the judgment result of step 206 is yes, meaning the fault occurs in the intelligent driving scenario, the process proceeds to step 208. In step 208, the weights of E2 and E3 are increased for scenario correction. By increasing the weights of the intelligent driving state trigger rate E2 and exposure intensity E3, the exposure score more accurately reflects the risk of the fault being discovered by the user in the intelligent driving usage scenario. After the weight adjustment is completed in step 208, the process proceeds to step 212, where multiple sub-indicators are combined to obtain the fault exposure score.

[0047] If the judgment result of step 206 is negative, meaning the fault occurs in a non-intelligent driving scenario or is difficult for the user to perceive, the process proceeds to step 210. In step 210, the original weights are maintained. In some implementations, when the fault occurs in a non-intelligent driving scenario or is difficult for the user to perceive, the contribution of the fault to the exposure score is reduced, thereby avoiding misjudgments caused by relying solely on the vehicle percentage. After step 210 is completed, the process proceeds to step 214, where multiple sub-indicators are combined to obtain the fault exposure score.

[0048] Through the above process, the fault exposure score E is obtained by combining three sub-indicators: vehicle coverage rate E1, intelligent driving trigger rate E2, and exposure intensity E3. Scenario correction is applied to the vehicle coverage rate E1. When a fault occurs in an intelligent driving scenario, the weights of the intelligent driving trigger rate E2 and exposure intensity E3 are increased, allowing the fault exposure score to more accurately reflect the likelihood of the fault being discovered by the user.

[0049] For example, refer to Figure 3 , Figure 3 This paper illustrates one implementation of the exposure grading definition in a two-dimensional exposure-user experience rating matrix. Exposure is defined into five levels using the normal distribution 3σ principle, where 1σ = 68.2%, 2σ = 95.4%, and 3σ = 99.7%. Based on the normal distribution principle, a correspondence is established between the exposure level and the percentage of faulty vehicles to quantify the probability that a fault is detected by the user.

[0050] like Figure 3 As shown, exposure level 1 corresponds to a faulty vehicle percentage of 31.8% to 100%, indicating that the probability of the fault being discovered by the user is extremely high. When the percentage of faulty vehicles is within this range, the fault is widely distributed among the vehicle population, and the likelihood of the user encountering the fault is at its highest level.

[0051] Continue to refer to Figure 3 Exposure level 2 corresponds to a faulty vehicle percentage of 18.2% to 31.8%, indicating a relatively high probability that the fault will be discovered by users. Exposure level 3 corresponds to a faulty vehicle percentage of 4.6% to 18.2%, indicating a moderate probability that the fault will be discovered by users. Exposure level 4 corresponds to a faulty vehicle percentage of 0.3% to 4.6%, indicating a relatively low probability that the fault will be discovered by users. Exposure level 5 corresponds to a faulty vehicle percentage of 0% to 0.3%, indicating an extremely low probability that the fault will be discovered by users.

[0052] By adopting the normal distribution 3σ principle to define the exposure level, the exposure-user experience two-dimensional scoring matrix can divide the probability of a fault being discovered by the user into 5 levels based on the statistical data of the proportion of faulty vehicles, thus providing a quantitative basis for the subsequent determination of the fault access level.

[0053] Reference Figure 4 , Figure 4 A flowchart illustrating the user experience score calculation method is provided. The user experience score is calculated based on the takeover mileage metric and the user's tolerance for the frequency of failure degradation. The user experience score is jointly determined by the frequency dimension experience score and the impact dimension experience score.

[0054] like Figure 4 As shown, the user experience score calculation process begins at step 300. In step 300, the average monthly intelligent driving mileage per vehicle is calculated. The formula for calculating the average monthly intelligent driving mileage per vehicle is: (Intelligent driving mileage / Total number of vehicles / Number of statistical days) × 30. This formula converts the intelligent driving mileage data within the statistical period into the average monthly mileage per vehicle, thus providing basic data for subsequently calculating the average user tolerance standard for takeover mileage.

[0055] Continue to refer to Figure 4 In step 302, the average tolerance of users for the frequency of fault degradation is obtained. This average tolerance is obtained through user surveys and reflects the user group's acceptance of fault degradation. In some implementations, user tolerance for fault degradation frequency is obtained in stratified manner according to user groups and usage scenarios to obtain different tolerance parameters. User groups include high-frequency intelligent driving users and low-frequency users, and usage scenarios include urban commuting and long-distance highway travel.

[0056] In step 304, the average user tolerance standard T for takeover mileage is calculated. The average user tolerance standard T for takeover mileage is equal to the product of the average monthly intelligent driving mileage per vehicle, the user's average tolerance for fault degradation frequency, and the number of faults associated with the intelligent driving function. The number of faults associated with the intelligent driving function refers to the total number of faults that would cause the function to degrade. By multiplying the above three parameters, the average user tolerance standard for fault takeover mileage is obtained, which serves as a benchmark value for evaluating user experience.

[0057] like Figure 4 As shown, in step 306, the frequency dimension experience score is determined based on the ratio of the actual fault takeover mileage to the user's average tolerance standard T for takeover mileage. The actual fault takeover mileage, also known as the actual fault MPD value, represents the average mileage traveled between each fault occurrence. A higher ratio of the actual fault MPD to the user's average tolerance standard T for takeover mileage indicates a lower frequency of fault occurrences, a better user experience, and a higher frequency dimension experience score.

[0058] Continue to refer to Figure 4In step 308, the impact dimension experience score is determined based on the fault influencing factors. The impact dimension experience score is used to characterize the perceived severity and recoverability of the fault to the user. The impact dimension experience score is determined based on at least one of the following factors: whether it triggers takeover, emergency braking, or steering intervention interruption; whether it causes core intelligent driving functions to become unavailable or significantly degraded; whether it generates obvious human-machine interaction alarms and the duration of the alarms; and whether it can be automatically recovered or requires manual intervention. By evaluating the above fault influencing factors, the impact dimension experience score can reflect the actual degree of impact of the fault on the user's driving experience.

[0059] In step 310, the frequency dimension experience score and the impact dimension experience score are combined to determine the final user experience score. In some implementations, a conservative strategy is used to determine the user experience score, that is, the smaller value between the frequency dimension experience score and the impact dimension experience score is taken as the final user experience score, prioritizing safety and a minimum acceptable level of user experience. In some implementations, a weighted strategy is used to determine the user experience score, that is, the frequency dimension experience score and the impact dimension experience score are weighted and summed, where the weights are preset values. By combining the frequency and impact dimensions, the user experience score can more comprehensively reflect the true impact of the fault on the user.

[0060] For example, Figure 5 The table shows user tolerance survey data. Users' average tolerance for failure degradation frequency was obtained through a user survey, and the survey results were used to determine the average tolerance ratio coefficient corresponding to different failure degradation frequencies. Figure 5 As shown, when the failure degradation frequency is once every 3 months, the average tolerance proportionality coefficient is 0.5; when the failure degradation frequency is once every 6 months, the average tolerance proportionality coefficient is 1; when the failure degradation frequency is once every 9 months, the average tolerance proportionality coefficient is 1.5; and when the failure degradation frequency is once every 12 months, the average tolerance proportionality coefficient is 2. The average tolerance proportionality coefficient reflects the proportional relationship between different failure degradation frequencies and the baseline tolerance, where once every 6 months is used as the baseline tolerance, corresponding to a proportionality coefficient of 1.

[0061] Continue to refer to Figure 5 Users' tolerance for the frequency of fault degradation is obtained by stratifying users based on user groups and usage scenarios to obtain different tolerance parameters. User groups include high-frequency intelligent driving users and low-frequency users. High-frequency intelligent driving users refer to users who frequently use intelligent driving functions, while low-frequency users refer to users who occasionally use intelligent driving functions. In some implementations, user group stratification also includes a distinction between novice users and experienced users. Novice users refer to users who are just starting to use intelligent driving functions, while experienced users refer to users who have extensive experience using intelligent driving functions.

[0062] Use cases include urban commuting and long-distance highway driving. Urban commuting refers to users using intelligent driving functions for daily commutes in urban road environments, while long-distance highway driving refers to users using intelligent driving functions for long-distance travel in highway environments. Users' tolerance for fault degradation varies across different use cases. By obtaining tolerance parameters stratified by use case, we can more accurately reflect users' actual experience needs in different scenarios.

[0063] When determining exit criteria, the average user tolerance standard T for takeover mileage at each level is calculated separately. group For different user groups such as high-frequency intelligent driving users, low-frequency users, novice users, and experienced users, as well as different usage scenarios such as urban commuting and long-distance highway driving, corresponding tolerance parameters are obtained. Based on the tolerance parameters of each layer, the average user tolerance standard T for takeover mileage is calculated. group The average user tolerance standard T is calculated through hierarchical calculation of takeover mileage. group It can output cluster exit conclusions, thereby supporting release strategies such as whitelisting or gray-scale release.

[0064] Furthermore, Figure 6 A table showing the correspondence between User Experience Level 4 and Actual MPD Range 5 is presented. User Experience Level 4 is defined with 5 levels based on the takeover mileage metric, and the Actual MPD Range 5 corresponds to User Experience Level 4, where T is the average user tolerance standard for takeover mileage.

[0065] like Figure 6 As shown, when the user experience level 4 is 1, the actual fault MPD range 5 is 0 to 0.5T, indicating that the actual fault takeover mileage is within 0 to 50% of the average user tolerance standard T for takeover mileage, and the user experience is at its lowest level. When the user experience level 4 is 2, the actual fault MPD range 5 is 0.5T to 1T, indicating that the actual fault takeover mileage is within 50% to 100% of the average user tolerance standard T for takeover mileage.

[0066] Continue to refer to Figure 6 When the user experience level 4 is 3, the actual fault MPD range is 5, which is 1T to 1.5T. This indicates that the actual fault takeover mileage reaches or exceeds the user's average tolerance standard T for takeover mileage, and the user experience is at a medium level. When the user experience level 4 is 4, the actual fault MPD range is 5, which is 1.5T to 2T. This indicates that the actual fault takeover mileage is within 150% to 200% of the user's average tolerance standard T for takeover mileage, and the user experience is at a high level. When the user experience level 4 is 5, the actual fault MPD range is 5, which is greater than or equal to 2T. This indicates that the actual fault takeover mileage reaches or exceeds 200% of the user's average tolerance standard T for takeover mileage, and the user experience is at the highest level.

[0067] By comparing the actual MPD range (5) of a fault with the average user tolerance standard (T) for takeover mileage, user experience level 4 can quantify the impact of fault frequency on user experience. A higher actual MPD value indicates a longer average mileage between fault occurrences, a lower fault frequency, and a higher user experience level 4.

[0068] When determining the overall user experience score U, the overall user experience score U is composed of the frequency dimension experience score U. freq And the impact on the dimensional experience rating U sev Jointly determined. In some implementations, a conservative strategy is used to determine the overall user experience score U, i.e. Frequency-based experience rating U freq And the impact on the dimensional experience rating U sev The smaller value among these is used as the overall user experience score U, prioritizing safety and ensuring a basic level of user experience. In some implementations, a weighted strategy is used to determine the overall user experience score U, i.e. Where α is a preset weight, and the experience score U is determined by frequency dimension. freq And the impact on the dimensional experience rating U sev The weighted sum is used to obtain the overall user experience score U.

[0069] Furthermore, refer to Figure 7 , Figure 7 This diagram illustrates an embodiment of an exposure-user experience two-dimensional rating matrix. The exposure-user experience two-dimensional rating matrix uses exposure and user experience as two rating dimensions. Exposure is divided into five levels from 1 to 5, and user experience is also divided into five levels from 1 to 5. Depending on the different combinations of exposure and user experience, each cell in the exposure-user experience two-dimensional rating matrix corresponds to a different clearance level, including four levels: S, A, B, and C.

[0070] like Figure 7 As shown, when the exposure is 1 and the user experience is 1 to 4, the exit level is C; when the exposure is 1 and the user experience is 5, the exit level is B. When the exposure is 2 and the user experience is 1 to 3, the exit level is C; when the exposure is 2 and the user experience is 4 to 5, the exit level is B. When the exposure is 3 and the user experience is 1 to 2, the exit level is C; when the exposure is 3 and the user experience is 3 to 4, the exit level is B; when the exposure is 3 and the user experience is 5, the exit level is A.

[0071] Continue to refer to Figure 7When the exposure is 4 and the user experience is 1, the clearance level is C; when the exposure is 4 and the user experience is 2 to 3, the clearance level is B; when the exposure is 4 and the user experience is 4 to 5, the clearance level is A. When the exposure is 5 and the user experience is 1 to 2, the clearance level is B; when the exposure is 5 and the user experience is 3 to 4, the clearance level is A; when the exposure is 5 and the user experience is 5, the clearance level is S.

[0072] Level S indicates that the fault is allowed to be released. When both the exposure score and user experience score of the fault reach the highest level, the fault meets the release criteria and can be released. Level A indicates that the fault is recommended to be released. When the exposure score and user experience score of the fault are at a high level, the decision on whether to release the fault needs to be made based on the specific exposure and user experience of the actual vehicle. Levels B and C indicate that the fault is rejected from release. When the exposure score or user experience score of the fault is at a low level, the fault design needs to be optimized based on the actual vehicle data, and the release evaluation should be carried out again after optimization.

[0073] In some implementations, the exposure-user experience two-dimensional scoring matrix can establish corresponding thresholds and mapping relationships for different vehicle platforms. Different vehicle platforms have different intelligent driving function configurations and user habits; by establishing corresponding thresholds and mapping relationships for different vehicle platforms, fault detection assessments can be made more consistent with the actual situation of each vehicle platform.

[0074] In some implementations, the exposure-user experience two-dimensional scoring matrix can establish corresponding thresholds and mapping relationships for different software versions. As software versions iterate and update, the performance and stability of intelligent driving functions will change. By establishing corresponding thresholds and mapping relationships for different software versions, it is possible to adapt to the fault clearance requirements during the software version evolution process.

[0075] In some implementations, the exposure-user experience two-dimensional scoring matrix can establish corresponding thresholds and mapping relationships for different regional regulations. Different regions have different regulations regarding the safety and reliability requirements of intelligent driving vehicles. By establishing corresponding thresholds and mapping relationships for different regional regulations, it can be ensured that fault exit assessments comply with the regulatory requirements of each region.

[0076] In some implementations, the exposure-user experience two-dimensional scoring matrix can establish corresponding thresholds and mapping relationships for different user groups. Different user groups have different tolerances for faults and different experience needs. By establishing corresponding thresholds and mapping relationships for different user groups, fault exit assessment can be made more in line with the actual experience needs of each user group.

[0077] Furthermore, refer to Figure 8 , Figure 8This paper illustrates the sample confidence factor determination and admission strategy generation process in the fault admission assessment method for intelligent driving vehicles. The method also includes the introduction of a sample confidence factor, which characterizes the impact of statistical sample size, statistical period length, and version coverage on exposure stability.

[0078] like Figure 8 As shown, the process for determining the sample confidence factor and generating the admission strategy begins at step 400. In step 400, the sample confidence factor C is calculated. conf Sample confidence factor C conf The sample confidence factor C is used to assess the stability of exposure calculation results by comprehensively considering the sufficiency of the statistical sample size, the reasonableness of the statistical period length, and the completeness of version coverage. When the statistical sample size is small, the statistical period is short, or the version coverage is insufficient, the sample confidence factor C... conf The low value indicates that the reliability of the current statistical data needs to be improved.

[0079] Continue to refer to Figure 8 In step 402, the sample confidence factor C is determined. conf Whether it is below a preset threshold. The judgment result of step 402 determines the direction of the subsequent process. When the sample confidence factor C conf When the value is below a preset threshold, it indicates that the current statistical data is insufficient to support a reliable exit decision, and the process proceeds to step 404. When the sample confidence factor C... conf If the value is not lower than the preset threshold, it indicates that the current statistical data has sufficient reliability, and the process proceeds to step 406.

[0080] like Figure 8 As shown, when the judgment result of step 402 is yes, the process proceeds to step 404. In step 404, a conditional conclusion is output. The conditional conclusion includes suggestions such as supplementing samples, extending the statistical period, and expanding the grayscale range before re-review. By outputting the conditional conclusion, subsequent data collection work is guided to obtain more sufficient statistical data to support the admission determination.

[0081] Continue to refer to Figure 8 If the judgment result of step 402 is negative, the process proceeds to step 406. In step 406, the fault clearance level is determined. The fault clearance level is determined based on the fault exposure score and user experience score through a two-dimensional scoring matrix of exposure-user experience. The clearance level includes four levels: S, A, B, and C.

[0082] like Figure 8As shown, after determining the fault clearance level in step 406, the process proceeds to step 408. In step 408, a judgment result is generated regarding whether to grant clearance or conditional clearance. The judgment result is determined based on the fault clearance level. When the clearance level is S, the judgment result is "allow clearance"; when the clearance level is A, the judgment result is "recommend clearance," requiring a decision based on the specific circumstances of the actual vehicle; when the clearance level is B or C, the judgment result is "deny clearance," requiring fault design optimization.

[0083] Continue to refer to Figure 8 After step 408 is completed, the process proceeds to step 410. In step 410, priority fault location directions for optimization and user-specific prompting strategies are generated. The fault location directions guide the development team in locating and optimizing faults, while the prompting strategies provide users with appropriate explanations or downgrade prompts when faults occur, thereby reducing the negative impact of faults on user experience.

[0084] like Figure 8 As shown, after step 410 is completed, the process proceeds to step 412. In step 412, the version release strategy is output. The version release strategy includes the canary scale, target audience, monitoring metrics, and rollback conditions. The canary scale is used to determine the coverage of the new version release, the target audience is used to determine the target user group for the new version release, the monitoring metrics are used to continuously monitor the failure performance after release, and the rollback conditions are used to determine the threshold for triggering a version rollback when the failure performance deteriorates.

[0085] The exit strategy includes a structured exit action package, which includes: the determination result of whether to allow or conditionally allow exit; the fault location direction that needs to be optimized first; prompts, explanations, or downgrade strategies for users; and the version release strategy, which includes the gray-scale ratio, applicable users, monitoring indicators, and rollback conditions. By generating a structured exit action package, the fault exit assessment method for intelligent driving vehicles can provide complete strategic guidance for fault exit decisions.

[0086] In some implementations, the exit strategy outputs segmented exit conclusions, thereby supporting release strategies such as whitelisting or gradual rollout. Different exit conclusions are output for different user groups, such as high-frequency intelligent driving users, low-frequency users, novice users, and experienced users, as well as different usage scenarios, such as urban commuting and long-distance highway travel. Through segmented exit conclusions, differentiated release strategies can be adopted for different user groups and usage scenarios, gradually expanding the release scope while ensuring user experience.

[0087] In some implementations, thresholds, weights, or matrix boundaries are updated based on new real-vehicle data and user feedback within each statistical period, thereby achieving dynamic calibration and closed-loop iteration of the fault exit criteria. With the continuous accumulation of real-vehicle data and the ongoing collection of user feedback, the thresholds in the exposure-user experience two-dimensional scoring matrix, the weights of each sub-indicator, and the boundary conditions of the exit level can be continuously optimized. Through dynamic calibration and closed-loop iteration mechanisms, the fault exit assessment method for intelligent driving vehicles can adapt to the evolution of intelligent driving functions and changes in user needs, continuously improving the accuracy and effectiveness of fault exit assessment.

[0088] For example, Figure 9 A statistical table of sensor signal timeout faults is presented as an application example of the fault exit assessment method for intelligent driving vehicles. The table displays the actual status of sensor signal timeout faults within a statistical period, including the number of faulty vehicles, the total number of vehicles, the percentage of faulty vehicles, the number of fault occurrences during intelligent driving, intelligent driving mileage, and MPD (Minimum Mileage Discount).

[0089] like Figure 9 As shown, 139 vehicles experienced sensor signal timeout faults, out of a total of 8635 vehicles. The percentage of faulty vehicles is the ratio of the number of faulty vehicles to the total number of vehicles, i.e., 139 / 8635 = 1.6%. The number of fault occurrences in intelligent driving mode was 52, indicating the number of times the fault was triggered when the intelligent driving function was enabled. The intelligent driving mileage was 2,520,476.7 kilometers, representing the total intelligent driving mileage of all vehicles within the statistical period. The MPD (Maximum Mileage Displacement) was 48,470.7 kilometers, representing the takeover mileage index for this fault, i.e., the ratio of intelligent driving mileage to the number of fault occurrences in intelligent driving mode.

[0090] Continue to refer to Figure 9 Exposure was calculated using the aforementioned statistical data. The proportion of faulty vehicles was 1.6%. According to the definition of exposure level, a proportion of 1.6% for faulty vehicles falls within the range of 0.3% to 4.6%, corresponding to an exposure level of 4, indicating that the probability of the fault being discovered by users is relatively low.

[0091] like Figure 9 As shown, the user experience score is calculated using the above statistical data. The formula for calculating the average monthly intelligent driving mileage per vehicle is: (Intelligent driving mileage / Total number of vehicles / Number of statistical days) × 30. Based on Figure 9The data shows that the average monthly mileage of intelligent driving per vehicle is 2,520,476.7 / 8,635 = 292 kilometers (assuming a statistical period of 30 days). The average user tolerance standard T for takeover mileage is equal to the product of the average monthly intelligent driving mileage per vehicle, the user's average tolerance for fault degradation frequency, and the number of faults associated with the intelligent driving function. Assuming the user's average tolerance for fault degradation frequency is once every 6 months, and the number of faults associated with the intelligent driving function is 100, then the average user tolerance standard T for takeover mileage is 292 × 6 × 100 = 175,200 kilometers / time.

[0092] Continue to refer to Figure 9 The actual MPD (Maximum Mileage Distance) for this sensor signal timeout fault is 48470.7 km. According to the correspondence between user experience level and the actual MPD range, when the actual MPD is within the range of 0 to 0.5T, the user experience level is 1. Since 0.5T = 0.5 × 175200 = 87600 km, and the actual MPD of this fault is 48470.7 km, falling within the range of 0 to 87600 km, the user experience score for this fault is 1, indicating that the user experience is at the lowest level.

[0093] Based on the above calculations, the exposure score for this sensor signal timeout fault is 4, and the user experience score is 1. According to the two-dimensional scoring matrix of exposure and user experience, when the exposure is 4 and the user experience is 1, the exit level is C. An exit level of C indicates that the fault is denied exit, requiring fault design optimization based on real vehicle data, and a re-evaluation of exit criteria after optimization.

[0094] Furthermore, refer to Figure 10 , Figure 10 A table showing the correspondence between user experience levels and the actual MPD range of a fault is presented to illustrate the process of determining the user experience score in an application example of a sensor signal timeout fault. Based on the aforementioned statistical data, the average monthly intelligent driving mileage per vehicle is 292 km. The average user tolerance standard T for takeover mileage is calculated based on the product of three parameters: the average monthly intelligent driving mileage per vehicle, the average user tolerance for fault degradation frequency, and the number of faults associated with the intelligent driving function, i.e., T = 292 × 6 × 100 = 175200 km / time.

[0095] like Figure 10As shown, a correspondence is established between user experience level and the actual MPD (Maximum Mileage Distance) range of the fault. This is used to determine the user experience score based on the ratio of the actual MPD to the user's average tolerance standard T for takeover mileage. When the user experience level is 1, the actual MPD range is 0 to 0.5T, corresponding to 0 to 87,600 km. When the user experience level is 2, the actual MPD range is 0.5T to 1T, corresponding to 87,600 to 175,200 km. When the user experience level is 3, the actual MPD range is 1T to 1.5T, corresponding to 175,200 to 262,800 km. When the user experience level is 4, the actual MPD range is 1.5T to 2T, corresponding to 262,800 to 350,400 km. When the user experience level is 5, the actual MPD range is greater than or equal to 2T, corresponding to greater than or equal to 350,400 km.

[0096] Continue to refer to Figure 10 The actual MPD of the sensor signal timeout fault was 48470.7 km. According to... Figure 10 The relationship between the user experience level and the actual MPD range of the fault is shown. 48470.7km falls within the range of 0 to 87600km, corresponding to a user experience level of 1, indicating the lowest user experience level. Combining the previously calculated exposure score of 4, a lookup table is performed using the exposure-user experience two-dimensional scoring matrix. When the exposure is 4 and the user experience is 1, the exit level is C. An exit level C indicates that the fault is denied exit; this sensor signal timeout fault does not meet the fault exit requirements and requires fault design optimization based on real-vehicle data.

[0097] This invention also discloses an intelligent driving vehicle fault exit assessment system.

[0098] Specifically, refer to Figure 11 The diagram illustrates the architecture of an intelligent driving vehicle fault exit assessment system 500. The system includes a data acquisition module 502, an exposure calculation module 504, a user experience calculation module 506, a level determination module 508, and a policy output module 510. The system is used to assess the exit status of intelligent driving vehicles based on faults. By integrating the calculation results of fault exposure and user experience, it achieves fault exit assessment and policy output.

[0099] like Figure 11As shown, the data acquisition module 502 is configured to acquire fault-related data of the intelligent driving vehicle. The data acquisition module 502 collects fault-related data from channels such as vehicle logs, cloud diagnostics, and after-sales or customer service feedback, and standardizes the definitions, statistical periods, vehicle base, and triggering scenarios. The data acquisition module 502 transmits the acquired fault-related data to the exposure calculation module 504 and the user experience calculation module 506, respectively, providing a data foundation for subsequent exposure calculations and user experience calculations.

[0100] Continue to refer to Figure 11 The exposure calculation module 504 is configured to calculate a fault exposure score based on fault-related data. The fault exposure score characterizes the likelihood that a fault will be detected by the user. The exposure calculation module 504 receives fault-related data transmitted by the data acquisition module 502 and calculates the fault exposure score based on sub-indicators such as vehicle coverage, intelligent driving trigger rate, and exposure intensity. The exposure calculation module 504 then transmits the calculated fault exposure score to the level determination module 508.

[0101] like Figure 11 As shown, the user experience calculation module 506 is configured to calculate a user experience score based on the takeover mileage metric and the user's tolerance for the frequency of fault degradation. The user experience calculation module 506 receives fault-related data transmitted by the data acquisition module 502, calculates the average user tolerance standard for takeover mileage, and determines the frequency dimension experience score based on the ratio of the actual fault takeover mileage to the average user tolerance standard for takeover mileage. In some embodiments, the user experience calculation module 506 also determines the impact dimension experience score based on fault influencing factors, and combines the frequency dimension experience score and the impact dimension experience score to determine the final user experience score. The user experience calculation module 506 transmits the calculated user experience score to the level determination module 508.

[0102] Continue to refer to Figure 11 The fault admission determination module 508 is configured to determine the fault admission level based on the fault exposure score and user experience score using a two-dimensional exposure-user experience scoring matrix. The module 508 receives the fault exposure score transmitted by the exposure calculation module 504 and the user experience score transmitted by the user experience calculation module 506, and maps the combination of the fault exposure score and user experience score to the corresponding admission level based on the exposure-user experience two-dimensional scoring matrix. The admission levels include four levels: S, A, B, and C, corresponding to different admission judgment results such as allowed admission, recommended admission, and denied admission, respectively. The module 508 transmits the determined fault admission level to the policy output module 510.

[0103] like Figure 11As shown, the strategy output module 510 is configured to output corresponding exit strategies based on the fault exit level. The strategy output module 510 receives the fault exit level transmitted by the level determination module 508 and generates corresponding exit strategies based on different exit levels. The exit strategy includes a structured exit action package, which includes the determination result of whether to allow or conditionally allow exit, the fault location direction to be prioritized for optimization, user-specific prompts or downgrade strategies, and a version release strategy. The version release strategy includes the grayscale ratio, applicable user groups, monitoring indicators, and rollback conditions. Through the exit strategies output by the strategy output module 510, the intelligent driving vehicle fault exit assessment system 500 can provide complete strategy guidance for fault exit decisions.

[0104] This invention also discloses a readable storage medium.

[0105] A computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the intelligent driving vehicle fault clearance assessment method described in any of the above embodiments. The computer-readable storage medium may include any entity or device capable of carrying a computer program, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), and a software distribution medium, etc. The computer program includes computer program code. The computer program code may be in the form of source code, object code, an executable file, or some intermediate form, etc. The computer-readable storage medium may include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), and a software distribution medium, etc.

[0106] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0107] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a system including a processing module or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0108] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A fault exit assessment method for intelligent driving vehicles, characterized in that, include: Acquire fault-related data from intelligent driving vehicles; A fault exposure score is calculated based on the fault-related data, and the fault exposure score is used to characterize the likelihood that the fault will be discovered by the user. User experience scores are calculated based on takeover mileage metrics and user tolerance for the frequency of failure degradation. Based on the fault exposure score and the user experience score, the fault clearance level is determined by a two-dimensional scoring matrix of exposure-user experience. as well as Output the corresponding admission strategy based on the fault admission level.

2. The intelligent driving vehicle fault exit assessment method according to claim 1, characterized in that, The fault-related data is collected from at least one of the following channels: vehicle logs, cloud diagnostics, and after-sales or customer service feedback.

3. The intelligent driving vehicle fault exit assessment method according to claim 1, characterized in that, The fault exposure score is obtained by combining multiple sub-indicators, which include: Vehicle coverage rate is the ratio of the number of vehicles that experienced the fault to the total number of vehicles in the statistical period. Intelligent driving state trigger rate, representing the ratio of the number of triggers while intelligent driving is enabled to the total number of times intelligent driving is enabled or the duration of intelligent driving activation; and Exposure intensity is expressed as the number of triggers per thousand kilometers or per hour.

4. The intelligent driving vehicle fault exit assessment method according to claim 3, characterized in that, The method also includes scene correction of the vehicle coverage, and when the fault mainly occurs in the intelligent driving activation scenario, the weights of the intelligent driving state trigger rate and the exposure intensity are increased.

5. The intelligent driving vehicle fault exit assessment method according to claim 1, characterized in that, The method further includes introducing a sample confidence factor, which is used to characterize the impact of statistical sample size, statistical period length, and version coverage on exposure stability. When the sample confidence factor is lower than a preset threshold, a conditional conclusion is output.

6. The intelligent driving vehicle fault exit assessment method according to claim 1, characterized in that, The user experience score is determined by a combination of frequency-based experience score and impact-based experience score. The impact-based experience score is used to characterize the perceived severity and recoverability of the fault to the user.

7. The intelligent driving vehicle fault exit assessment method according to claim 6, characterized in that, The influence dimension experience score is determined based on at least one of the following factors: Whether it triggers takeover, emergency braking, or steering intervention interruption; Does it cause core intelligent driving functions to become unavailable or significantly degraded? Whether a significant human-computer interaction alarm is generated and the duration of the alarm; as well as Whether it can be automatically restored or requires manual intervention.

8. The intelligent driving vehicle fault exit assessment method according to claim 1, characterized in that, User experience scores are calculated based on takeover mileage metrics and user tolerance for the frequency of failure degradation, including: The average user tolerance standard for takeover mileage is calculated, which is equal to the product of the average monthly intelligent driving mileage per vehicle, the user's average tolerance for fault degradation frequency, and the number of faults associated with the intelligent driving function; and The user experience score is determined by the ratio of the actual fault takeover mileage to the average user tolerance standard for the takeover mileage.

9. The intelligent driving vehicle fault exit assessment method according to claim 8, characterized in that, The user's tolerance for the frequency of fault degradation is obtained by stratifying the user group and usage scenario to obtain different tolerance parameters. The user group includes high-frequency intelligent driving users and low-frequency users, and the usage scenario includes urban commuting and long-distance highway travel.

10. The intelligent driving vehicle fault exit assessment method according to claim 1, characterized in that, The exit strategy includes a structured exit action package, which includes: The result of the determination of whether or not to grant permission or conditional permission; Fault location directions that need to be optimized first; User suggestions, explanations, or downgrade strategies; and Version release strategy, which includes grayscale ratio, target audience, monitoring metrics and rollback conditions.

11. A fault exit assessment system for intelligent driving vehicles, characterized in that, include: The data acquisition module is configured to acquire fault-related data of intelligent driving vehicles; The exposure calculation module is configured to calculate a fault exposure score based on the fault-related data, wherein the fault exposure score is used to characterize the likelihood that the fault will be discovered by the user. The user experience calculation module is configured to calculate the user experience score based on the takeover mileage metric and the user's tolerance for the frequency of failure degradation. The fault clearance module is configured to determine the fault clearance level based on the fault exposure score and the user experience score using a two-dimensional scoring matrix of exposure-user experience. as well as The strategy output module is configured to output the corresponding admission strategy according to the fault admission level.

12. A readable storage medium, characterized in that, The readable storage medium stores computer instructions that, when executed by a processor, implement the intelligent driving vehicle fault exit assessment method as described in any one of claims 1-10.

Images