Aggressive driving evaluation method, equipment, medium and product

Abstract

The invention discloses an aggressive driving evaluation method and device, a medium and a product, and relates to the technical field of aggressive driving evaluation.The method comprises the steps that feature extraction is conducted on multiple driving data, and multiple feature data points are determined; quantizing aggressive level distribution of the plurality of driving data by using a preset priori knowledge and experience cumulative distribution function, and determining an aggressive level quantization result of each feature data point; determining a plurality of aggressive driving data based on the aggressive level quantization result of each feature data point and a preset threshold; a fuzzy C-means clustering algorithm and semantic analysis are adopted to process the multiple aggressive driving data, and the driving operation composition of each aggressive driving data is determined; and according to the aggressive level quantification result corresponding to each aggressive driving data and the driving operation composition, evaluating each aggressive driving data, generating a potential aggressive driving evaluation result, and providing accurate, credible and targeted potential aggressive driving evaluation feedback for a driver.

Description

Technical Field

[0001] The present application relates to the technical field of aggressive driving evaluation, and in particular to an aggressive driving evaluation method, device, medium, and product. Background Art

[0002] Aggressive driving, including sudden acceleration, deceleration, and sharp turns, significantly impacts road safety. Studies show that over 90% of traffic accidents are caused by human error, with aggressive driving being a major factor. Evaluating drivers' aggressive driving behavior and providing feedback, encouragement, or guidance on adopting safer and more economical driving styles is an effective approach.

[0003] However, most current studies only conduct qualitative assessments of drivers' driving aggressiveness levels and lack the quantification and interpretation of given aggressiveness levels, resulting in inaccurate assessment results of aggressive driving data, which in turn reduces the credibility of aggressive driving data assessments. Summary of the Invention

[0004] The purpose of this application is to provide an aggressive driving evaluation method, device, medium and product, which can solve the problems of low accuracy of aggressive driving data evaluation and low credibility of aggressive driving data evaluation results in the prior art.

[0005] To achieve the above objectives, this application provides the following solutions.

[0006] In a first aspect, the present application provides an aggressive driving assessment method, comprising: extracting features from multiple driving data to determine multiple feature data points; the feature data points include longitudinal acceleration, lateral acceleration, and speed; quantifying the aggressiveness distribution of the multiple driving data using preset prior knowledge and an empirical cumulative distribution function to determine an aggressiveness quantification result for each feature data point; determining multiple aggressive driving data based on the aggressiveness quantification result for each feature data point and a preset threshold; processing the multiple aggressive driving data using a fuzzy C-means clustering algorithm and semantic analysis to determine a driving operation composition for each aggressive driving data; the driving operation composition includes a driving category and a degree of membership of each aggressive driving data to the driving category; the driving categories include sudden deceleration, sudden left turn, sudden acceleration, sudden right turn, or speeding; and each aggressive driving data corresponds to at least one driving category; and evaluating each aggressive driving data based on the aggressiveness quantification result and the driving operation composition corresponding to each aggressive driving data to generate a potential aggressive driving assessment result.

[0007] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the aggressive driving evaluation method described above.

[0008] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the aggressive driving evaluation method described above when executed by a processor.

[0009] In a fourth aspect, the present application provides a computer program product, comprising a computer program, which, when executed by a processor, implements any of the above-described aggressive driving evaluation methods.

[0010] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0011] This application provides an aggressive driving assessment method, device, medium, and product. First, the application extracts features from multiple driving data points to determine multiple feature data points. The system then quantifies the aggressiveness level of each feature data point using pre-set prior knowledge and an empirical cumulative distribution function. This allows for rapid and accurate estimation of the aggressiveness distribution of the driving data. The pre-set prior knowledge effectively quantifies the aggressiveness level of each feature data point. Furthermore, based on the quantified aggressiveness level of each feature data point and a pre-set threshold, multiple aggressive driving data points are determined. Simultaneously, a fuzzy C-means clustering algorithm and semantic analysis are employed to determine the driving operation composition of each aggressive driving data point. The driving operation composition of each aggressive driving data point provides an intuitive explanation of the driving operation composition of the aggressive driving data, enhancing the credibility of the data. Finally, based on the corresponding aggressiveness level quantification results and the driving operation composition of each aggressive driving data point, each aggressive driving data point is evaluated to accurately generate a potential aggressive driving assessment result. Ultimately, the system provides drivers with accurate, reliable, and targeted feedback on their potential aggressive driving behavior. This helps drivers better understand their driving habits and potential aggressive driving behaviors, encouraging them to proactively improve their driving behavior and fundamentally reduce the road safety hazards associated with aggressive driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0013] Figure 1 This is a flow chart of an aggressive driving evaluation method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0014] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0015] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0016] like Figure 1 As shown, the present application provides an aggressive driving evaluation method, including steps 101 to 105.

[0017] Step 101: extract features from a plurality of driving data to determine a plurality of feature data points; the feature data points include longitudinal acceleration, lateral acceleration and speed.

[0018] In practical applications, the acquired driving data is extracted including but not limited to lateral and longitudinal acceleration and speed as features, and then preprocessed, including but not limited to filtering, denoising, downsampling, etc. Suppose the set of all n preprocessed driving data is {X1,X2,...,X n}, where the i-th data is and Represents the longitudinal and lateral acceleration of the i-th data (a x ,a y ) and speed (v) characteristics.

[0019] Step 102: quantify the aggressiveness distribution of the plurality of driving data using preset prior knowledge and an empirical cumulative distribution function, and determine an aggressiveness quantification result for each characteristic data point.

[0020] In some embodiments, step 102 specifically includes steps 201 to 203 .

[0021] Step 201: Define a "left tail" empirical cumulative distribution function and a "right tail" empirical cumulative distribution function, and use the "left tail" empirical cumulative distribution function and the "right tail" empirical cumulative distribution function as the empirical cumulative distribution function; wherein, the "left tail" represents the extent to which the driving data belongs to the driving category of sudden deceleration or sudden left turn; the "right tail" represents the extent to which the driving data belongs to the driving category of sudden acceleration, sudden right turn or speeding.

[0022] Step 202: According to

[0023] and Calculate the i-th feature data point The lateral acceleration a x Radical horizontal distribution, longitudinal acceleration a y The radical level distribution of and the radical level distribution of speed v; where, and Represents the longitudinal acceleration a of the i-th characteristic data point y , lateral acceleration a x and velocity v; and Denote the lateral acceleration a x and longitudinal acceleration a y The radical level distribution, and Respectively represent a x The “left-tail” empirical cumulative distribution function and the “right-tail” empirical cumulative distribution function of ; and A y The "left-tail" empirical cumulative distribution function and the "right-tail" empirical cumulative distribution function of represents the lateral acceleration of the i-th characteristic data point; Γ is the indicator function; 1 means when When the indicator function has a value of 1, 0 means when When , the value of the indicator function is 0; x represents the independent variable of the empirical cumulative distribution function; R represents the set of driving data; n represents the total number of driving data; represents the radical level distribution of velocity v, represents the “right-tail” empirical cumulative distribution function of v;

[0024] Step 203: According to Determine the radical level quantization result of each characteristic data point; wherein S(X i ) represents the feature data point X i The radical level of quantitative results; and Represents X i In a x 、a y and the radical level quantization results on v, Indicates that Substitution The result after Indicates that Substitution The result after Indicates that Substitution The result after.

[0025] In practical applications, it is necessary to obtain prior knowledge of each feature, for example, x and a y Smaller and larger values ​​(near the left and right tails of the distribution) have higher aggressiveness levels, corresponding to emergency (slowing down / turning left) and urgent (accelerating / turning right), respectively. For v, larger values ​​(near the right tail of the distribution) have higher aggressiveness levels, corresponding to speeding. If other features are available, their prior knowledge can be obtained in a similar way.

[0026] Furthermore, based on prior knowledge, the aggressiveness distribution of a feature can be estimated by its tail characteristics, because tail data may mean a larger acceleration or speed, and a larger acceleration or speed is directly positively correlated with the aggressiveness level. Considering that the tail includes two types, the left tail and the right tail, two empirical cumulative distribution functions, "left tail" and "right tail", are defined. This is to quantify the degree to which driving data belongs to the left tail or right tail of a feature. Based on this, the aggressiveness distribution of the feature can be estimated, with a x For example, it is represented as and The calculation is as follows (for a y Same as v).

[0027]

[0028] in, and A x The "left tail" and "right tail" empirical cumulative distribution functions, the value of the "left tail" empirical cumulative distribution function gradually increases as the data goes from left to right, and the value of the "right tail" empirical cumulative distribution function gradually increases as the data goes from right to left.

[0029] Specifically, the radical level distribution of each feature is estimated by combining prior knowledge and empirical cumulative distribution function. x and a y The left and right tails of these two feature distributions have radical meanings (such as sudden acceleration / deceleration, sharp turns, etc.), which effectively characterizes a x and a y This characteristic of decreasing radicalization levels from the two tails of its distribution toward the center is consistent with prior knowledge.

[0030] Furthermore, for feature v, the calculation formula of its radical level distribution is: That is, the “right tail” empirical cumulative distribution function is taken because the right tail of the characteristic distribution of v has a radical meaning (such as speeding). This treatment effectively represents the characteristic that the radical level of v decreases from the right tail to the left tail of its distribution, which is consistent with prior knowledge.

[0031] The aggressiveness of the driving data is quantified based on the aggressiveness distribution of each feature of each feature data point. The negative logarithm is taken to map the value range originally belonging to [0, 1] to [0, +∞], making the result more intuitive in representing the aggressiveness level. For example, a higher value indicates a higher aggressiveness level.

[0032] Step 103 : Determine a plurality of aggressive driving data based on the quantified result of the aggressiveness level of each characteristic data point and a preset threshold.

[0033] In some embodiments, step 103 specifically includes: sorting the quantization results of the aggressiveness level of each of the feature data points in descending order to obtain a data set; selecting data from the data set that is less than or equal to a preset threshold; and using the feature data points corresponding to the data as the multiple aggressive driving data.

[0034] In practical applications, according to the above S(X i ) calculation formula, we can get the radical level quantization results of all feature data points, sort these results in descending order, and the results are recorded as a set As shown below:

[0035]

[0036] in, This value represents the characteristic data point with the maximum quantified aggressiveness level. A certain percentage of data is selected from high to low as aggressive driving data. Considering that aggressive driving behavior accounts for a relatively small proportion of driving data, a smaller value, such as 5%, is recommended.

[0037] Step 104: Using a fuzzy C-means clustering algorithm and semantic analysis, the plurality of aggressive driving data are processed to determine a driving operation component for each aggressive driving data; the driving operation component includes a driving category and a degree of membership of each aggressive driving data to the driving category; the driving categories include sudden deceleration, sudden left turn, sudden acceleration, sudden right turn, or speeding; and each aggressive driving data corresponds to at least one driving category.

[0038] In some embodiments, step 104 specifically includes steps 301 and 302.

[0039] Step 301: Clustering the multiple aggressive driving data using a fuzzy C-means clustering algorithm to determine the degree of membership of each aggressive driving data item to a driving category and the centroid of each driving category; wherein the centroid of each driving category represents the average position of multiple feature data points corresponding to each driving category.

[0040] In some embodiments, step 301 specifically includes: for any aggressive driving data and the membership of the aggressive driving data to any driving category, initializing a membership set Among them, U represents the membership set; μ fj represents χ f with c j The membership degree between f represents the fth aggressive driving data, l represents the total number of aggressive driving data; c j represents the centroid of the j-th driving category, μ' fj represents μ fj The initial value of τ represents the fuzzification parameter; m represents the number of initialized driving categories; according to the centroid of the j-th driving category, update μ fj , determining a target value of the membership; using the target value as the initial value to update the centroid of the j-th driving category; determining an updated value of the membership based on the updated centroid of the j-th driving category; stopping the updating when the difference between the updated value and the target value is less than a target threshold; determining the updated value as the membership degree of the aggressive driving data belonging to the j-th driving category, and using the updated centroid of the j-th driving category as the centroid of the j-th driving category.

[0041] Step 302: Perform semantic analysis on the centroid of each driving category to determine the driving category corresponding to each aggressive driving data.

[0042] In practical applications, the Fuzzy C-means (FCM) clustering algorithm and semantic analysis are used to identify specific driving operations on aggressive driving data, and the driving operation composition of each feature data point is quantified.

[0043] The principle of the fuzzy C-means algorithm is as follows.

[0044] Step 1: Initialize the number of categories m, the membership set U and the fuzzification parameter τ.

[0045] Step 2: According to Update the centroids of all categories.

[0046] in, represents the fth aggressive driving data, l is the total number of aggressive driving data; represents the centroid of the jth class, μ fj is χ f with c j The membership value between f Belong to c j degree of membership set τ is a fuzzification parameter used to control the degree of membership between categories. A larger τ value increases fuzziness, while a smaller τ makes the categories clearer. It is usually set in the range of 1<τ<3, with τ=2 being the most common choice.

[0047] Step 3: According to Update the membership set, where ||·|| represents the Euclidean distance or the two-norm.

[0048] Step 4: When the change in membership is lower than the target threshold (usually a small positive number, such as 10 -4 ), stop the iteration; otherwise return to step 2).

[0049] In practical applications, it is necessary to consider the optimal number of categories. Here, the "elbow rule" is used to determine the optimal number. Specifically, a series of different numbers of categories are preset, and the maximum preset number of categories is generally 20-30. Fuzzy C-means clustering is performed under each number of categories, and the corresponding sum of squared errors (SSE) is calculated. For example, when the number of categories m is 5, the calculation formula is as follows:

[0050]

[0051] Among them, Φ j is the set of aggressive driving data in class j.

[0052] Finally, plot all SSE values ​​and identify the "elbow point," the point where the rate of decline slows significantly. This point represents a balance between reducing SSE and avoiding overfitting. Smaller SSE values ​​indicate more compact clustering of aggressive driving data within the same category, indicating higher clustering quality. However, this value should not be too small, as this can lead to overfitting and hinder the generation of meaningful categories. The optimal number of categories corresponding to this "elbow point" is therefore denoted by m.

[0053] The centroid coordinates of m categories are semantically analyzed to obtain aggressive driving operations, such as sudden acceleration, sharp steering, and speeding. For example, if the centroid coordinates correspond to the longitudinal acceleration a y The value of is large, the lateral acceleration a x The value of is very small, which can be semantically interpreted as a sudden acceleration driving operation.

[0054] The membership vector of each feature data point is used to quantify the composition of the driving operation, such as μ f =(μ f1 ,μ f2 ,...,μ fm ) represents the driving operation composition in the f-th radical feature data point. f1Indicates the proportion of the first type of driving operation in the feature data point. Regarding the composition of driving operations, for example, 20% belongs to sudden acceleration, 40% belongs to sudden steering, and 40% belongs to speeding; among them, the i-th feature data point is 20% and 40% are both membership degrees.

[0055] Step 105 : Evaluate each aggressive driving data according to the aggressiveness level quantification result and the driving operation composition corresponding to each aggressive driving data, and generate a potential aggressive driving evaluation result.

[0056] In some embodiments, step 105 specifically includes steps 401 and 402:

[0057] Step 401: According to R(χ f )=μ f S(χ f ), evaluate the f-th aggressive driving data χ f , generate the f-th aggressive driving data χ f The evaluation results of f ) represents the f-th aggressive driving data χ f The evaluation results of μ f represents χ f degree of membership in the driving category; S(χ f ) represents χ f The radical level of quantitative results; and Represents χ f In a x 、a y and v on the aggressiveness level quantization results.

[0058] Step 402: According to the f-th aggressive driving data χ f The evaluation results of the potential aggressive driving are generated.

[0059] In some embodiments, step 402 specifically includes: based on the f-th aggressive driving data x f The evaluation results of the aggressive driving data are used to construct an evaluation result set of the aggressive driving data; and the evaluation results of each aggressive driving data in the evaluation result set are added together to generate a potential aggressive driving evaluation result.

[0060] In practical applications, each aggressive driving data is evaluated and calculated as follows:

[0061]

[0062] Among them, R(χ f ) represents the f-th aggressive driving data χ fThe evaluation results of μ f represents the membership of the feature data point to all driving operations, i.e., its specific driving operation composition; S(χ f ) represents the aggressiveness level of the feature data point. This allows for a refined evaluation of each aggressive driving feature data point.

[0063] In practical applications, drivers are evaluated based on all aggressive driving data. The calculation method is illustrated using the t-th driver as an example. Assuming that the set of all aggressive driving data for this driver is {R(χ1), R(χ3), R(χ5), R(χ8)}, the evaluation calculation is as follows.

[0064] Q t =R(χ1)+R(χ3)+R(χ5)+R(χ8).

[0065] Among them, Q t It represents the aggressive driving evaluation result of the t-th driver. The calculation for other drivers is similar.

[0066] Compared with the existing technology, this application has the following main advantages.

[0067] 1. Efficient and Accurate Quantification of Aggressiveness: The aggressiveness of driving data is quantified by combining prior knowledge and the empirical cumulative distribution function (ECDF). As a non-parametric statistical method, the ECDF quickly and accurately estimates the data distribution without complex hyperparameter tuning. It also effectively quantifies the aggressiveness of driving data by incorporating prior domain knowledge (for example, the effect of high lateral acceleration on emergency cornering).

[0068] 2. Interpretable and Detailed Aggressive Maneuver Identification: The Fuzzy C-means algorithm clusters aggressive driving data and uses semantic analysis to identify interpretable aggressive driving maneuvers. Furthermore, the Fuzzy C-means algorithm allows feature data points to belong to multiple clusters simultaneously. Therefore, the membership metric can be used to quantify the composition of different driving maneuvers within each feature data point, enabling detailed identification of aggressive driving maneuvers.

[0069] 3. An effective feedback mechanism to enhance driving safety: By quantifying the driver's aggressive driving level and identifying aggressive driving maneuvers, the system provides intuitive and targeted feedback to the driver. This feedback mechanism helps drivers better understand their driving habits and potential aggressive behavior, encouraging them to proactively improve their driving style and fundamentally reduce road safety hazards caused by aggressive driving.

[0070] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the above method when executing the computer program.

[0071] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the above method when executed by a processor.

[0072] In an exemplary embodiment, a computer program product is provided, including a computer program, which implements the above method when executed by a processor.

[0073] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0074] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. In particular, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (RdM), magnetic random access memory (MRdM), ferroelectric random access memory (FRdM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RdM) or external cache memory, etc. By way of illustration and not limitation, RdM may be in various forms, such as static random access memory (SRdM) or dynamic random access memory (DRdM).

[0075] The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may include, but are not limited to, general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic units, data processing logic units based on quantum computing, and the like.

[0076] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0077] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. An aggressive driving evaluation method, characterized in that: include: Perform feature extraction on multiple driving data to determine multiple feature data points; The characteristic data points include longitudinal acceleration, lateral acceleration and velocity; quantifying the aggressiveness distribution of the plurality of driving data using preset prior knowledge and an empirical cumulative distribution function, and determining an aggressiveness quantification result for each characteristic data point; determining a plurality of aggressive driving data based on the quantified result of the aggressiveness level of each characteristic data point and a preset threshold; Processing the plurality of aggressive driving data using a fuzzy C-means clustering algorithm and semantic analysis to determine a driving operation component for each aggressive driving data; the driving operation component including a driving category and a degree of membership of each aggressive driving data to the driving category; The driving categories include sudden deceleration, sudden left turn, sudden acceleration, sudden right turn or speeding; each aggressive driving data corresponds to at least one driving category; Each aggressive driving data is evaluated based on the aggressive level quantification result and driving operation composition corresponding to each aggressive driving data to generate a potential aggressive driving evaluation result.

2. The aggressive driving evaluation method according to claim 1, characterized in that: Quantifying the aggressiveness distribution of the plurality of driving data using preset prior knowledge and an empirical cumulative distribution function, and determining an aggressiveness quantification result for each characteristic data point, specifically comprising: A "left tail" empirical cumulative distribution function and a "right tail" empirical cumulative distribution function are defined, and the "left tail" empirical cumulative distribution function and the "right tail" empirical cumulative distribution function are used as the empirical cumulative distribution function; wherein the "left tail" represents the degree to which the driving data belongs to the driving category of sudden deceleration or sudden left turn; and the "right tail" represents the degree to which the driving data belongs to the driving category of sudden acceleration, sudden right turn, or speeding; according to and Calculate the i-th feature data point The lateral acceleration a x Radical horizontal distribution, longitudinal acceleration a y The radical level distribution of and the radical level distribution of speed v; where, and Represents the longitudinal acceleration a of the i-th characteristic data point y , lateral acceleration a x and velocity v; and Denote the lateral acceleration a x and longitudinal acceleration a y The radical level distribution, and Respectively represent a x The "left-tail" empirical cumulative distribution function and the "right-tail" empirical cumulative distribution function of ; and Respectively represent a y The "left-tail" empirical cumulative distribution function and the "right-tail" empirical cumulative distribution function of represents the lateral acceleration of the i-th characteristic data point; Γ is the indicator function; 1 means when When the indicator function has a value of 1, 0 means when When , the value of the indicator function is 0; x represents the independent variable of the empirical cumulative distribution function; R represents the set of driving data; n represents the total number of driving data; represents the radical level distribution of velocity v, represents the "right tail" empirical cumulative distribution function of v; according to Determine the radical level quantization result of each characteristic data point; wherein S(X i ) represents the feature data point X i The radical level of quantitative results; and Represents X i In a x 、a y and the radical level quantization results on v, Indicates that Substitution The result after Indicates that Substitution The result after Indicates that Substitution The result after.

3. The aggressive driving evaluation method according to claim 1, characterized in that: Based on the quantified result of the aggressiveness level of each characteristic data point and the preset threshold, multiple aggressive driving data are determined, including: sorting the radical level quantization results of each characteristic data point in descending order to obtain a data set; Selecting data in the data set that is less than or equal to a preset threshold; The characteristic data points corresponding to the data are used as the multiple aggressive driving data.

4. The aggressive driving evaluation method according to claim 2, characterized in that: The plurality of aggressive driving data are processed using a fuzzy C-means clustering algorithm and semantic analysis to determine the driving operation composition of each aggressive driving data, specifically including: Clustering the multiple aggressive driving data using a fuzzy C-means clustering algorithm to determine the degree of membership of each aggressive driving data to a driving category and the centroid of each driving category; wherein the centroid of each driving category represents the average position of multiple feature data points corresponding to each driving category; A semantic analysis is performed on the centroid of each driving category to determine the driving category corresponding to each aggressive driving data.

5. The aggressive driving evaluation method according to claim 4, characterized in that: Clustering the plurality of aggressive driving data according to a fuzzy C-means clustering algorithm to determine the membership degree of each aggressive driving data to a driving category and the centroid of each driving category, specifically including: For any aggressive driving data and the membership of the aggressive driving data to any driving category, initialize the membership set Among them, U represents the membership set; μ fj represents χ f with c j The membership degree between f represents the fth aggressive driving data, l represents the total number of aggressive driving data; c j represents the centroid of the j-th driving category, μ' fj represents μ fj The initial value of ; τ represents the fuzzification parameter; m represents the number of initialized driving categories; According to the centroid of the j-th driving category, update μ fj , determine the target value of membership; Using the target value as the initial value, updating the centroid of the j-th driving category; Determine the updated value of the membership degree according to the updated centroid of the j-th driving category; When the difference between the updated value and the target value is less than the target threshold, stopping the update; The updated value is determined as the membership degree of the aggressive driving data to the j-th driving category, and the updated centroid of the j-th driving category is used as the centroid of the j-th driving category.

6. The aggressive driving evaluation method according to claim 2, characterized in that: Based on the aggressiveness level quantification results and driving operation components corresponding to each aggressive driving data, each aggressive driving data is evaluated to generate a potential aggressive driving evaluation result, including: According to R(χ f )=μ f S(χ f ), evaluate the f-th aggressive driving data χ f , generate the f-th aggressive driving data χ f The evaluation results of f ) represents the f-th aggressive driving data χ f The evaluation results of μ f represents χ f degree of membership in the driving category; S(χ f ) represents χ f The radical level of quantitative results; and Represents χ f In a x 、a y and the radical level quantization results on v; According to the f-th aggressive driving data χ f The evaluation results of the potential aggressive driving are generated.

7. The aggressive driving evaluation method according to claim 6, characterized in that: According to the f-th aggressive driving data χ f The evaluation results of potential aggressive driving are generated based on the evaluation results, including: Based on the f-th aggressive driving data χ f The evaluation results of aggressive driving data are constructed; The evaluation results of each aggressive driving data in the evaluation result set are added together to generate a potential aggressive driving evaluation result.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the aggressive driving evaluation method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the aggressive driving evaluation method according to any one of claims 1 to 7 is implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the aggressive driving evaluation method according to any one of claims 1 to 7 is implemented.

Images