Road surface grading method and device based on vehicle driving characteristics, equipment and storage medium
By calculating vehicle driving feature data and iterative training verification models, screening and clustering analysis, and dynamically dividing pavement levels, the poor adaptability problem caused by fixed pavement levels in the existing technology is solved, and the pavement level definition is more suitable for the vehicle chassis system.
Patent Information
- Application Number
- CN202411013430.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-26
- Publication Date
- 2024-11-15
AI Technical Summary
Due to the fixed grade of the pavement, the existing pavement level classification method cannot adapt to the needs of different application scenarios, resulting in poor adaptability to the vehicle chassis system.
By calculating the preset parameter values in the vehicle driving feature data, iteratively train and verify the model, filtering target parameters with high correlation, performing cluster analysis, dynamically divide road surface levels, and breaking away from the traditional road unevenness division method.
It has achieved dynamic and adaptive improvements in pavement level division, which is more suitable for the needs of vehicle chassis systems, and provides a more popular and applicable pavement level definition.
Smart Images

Figure CN118968449A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of road surface recognition, and in particular to a method, device, computer equipment and storage medium for classifying road surfaces based on vehicle driving characteristics. Background Art
[0002] At present, the official standard road surface grade is fixedly divided into 8 grades from A to H according to the unevenness of the road surface. However, when the road surface grade is actually used to divide and apply the road surface, it often does not match the road surface grade division actually required by the application scenario due to the fixed road surface grade division, resulting in poor fit of the popular road grade definition for chassis system comfort research.
[0003] Therefore, it is necessary to provide a road surface grade division method to solve the above-mentioned problem that the road surface grade division is not matched with the actual needs of the application scenario due to the fixed existing road surface grade division. Summary of the invention
[0004] In view of the above problems, the present application provides a method, device, equipment and storage medium for road surface grade division based on vehicle driving characteristics, which is used to solve the problem in the prior art that the existing road surface grade division is fixed and does not match the road surface grade division actually required by the application scenario. By calculating the parameter value corresponding to the parameter name carried by the preset feature, the target parameter is obtained, and the preset feature is expanded; at the same time, the target parameter is calculated for correlation through the target model after iterative training verification, and the target parameter that meets the preset threshold is selected as the initial feature, and the target parameter is screened based on the correlation; the screened target parameter is used as the initial feature, and the target feature is obtained after processing, and the target feature is clustered and analyzed to obtain the road surface grade. Finally, cluster analysis is performed according to the target feature to obtain the road surface grade division. Different from the conventional road surface division according to road roughness, the clustered road division grade is more suitable for the road surface due to feature expansion, correlation calculation and screening, and is more suitable for the adaptability research of the road surface in the vehicle chassis system.
[0005] According to one aspect of an embodiment of the present application, a method for road surface grade division based on vehicle driving characteristics is provided, the method comprising: calculating a parameter value corresponding to a parameter name carried by a preset feature based on sample data, and obtaining a target parameter according to the parameter name and the parameter value; wherein the sample data is driving characteristic data collected in real time during vehicle driving; performing iterative training and verification processing on a preset model according to the sample data to obtain a target model for calculating an associated value; wherein the associated value represents the degree of association between the target parameter and the road surface grade division; screening the target parameter carrying the associated value output by the target model, and taking the target parameter whose associated value meets a preset threshold as the initial feature; preprocessing the initial feature to obtain the target feature, and performing cluster analysis based on the target feature to obtain the road surface grade division.
[0006] In an optional manner, before the step of calculating the parameter value corresponding to the parameter name carried by the preset feature based on the sample data, it also includes: dividing the sample data according to the preset cycle length and the preset cycle interval; constructing the parameter name of the preset feature based on the divided sample data to obtain the preset feature carrying the parameter name.
[0007] In an optional manner, the parameter name includes an average value parameter, a maximum value parameter, a minimum value parameter, a standard deviation parameter and a zero-crossing frequency ratio parameter; the step of calculating the parameter value corresponding to the parameter name carried by the preset feature based on the sample data further includes: comparing the numerical values corresponding to the preset feature in the sample data to obtain the maximum and minimum values of the preset feature; summing the numerical values corresponding to the preset feature in the sample data and dividing them by the number of the numerical values to calculate the average value of the preset feature; obtaining the standard deviation of the preset feature based on the average value and the numerical value corresponding to the preset feature in the sample data; sorting the numerical values corresponding to the preset feature in the sample data according to the collection time, and counting the number of times that two adjacent numerical values have different positive and negative signs, proportionally calculating the number of the statistics and the number of the numerical values, and calculating the zero-crossing frequency ratio.
[0008] In one optional manner, the sample data includes multiple copies; the step of iteratively training and verifying the preset model according to the sample data to obtain a target model for calculating the associated value further includes: according to the sampling period corresponding to each sub-sample in the sample data, the sub-sample whose sampling period matches the preset sampling period is used as a verification set, and the sub-samples other than the verification set are used as a training set; the preset model is trained with the training set to obtain the trained preset model, and the verification set is used to verify whether the trained preset model meets the preset standard; the trained preset model is used as a new preset model, and the steps of iteratively repeating the steps of training the preset model with the training set to obtain the trained preset model, and the verification set is used to verify whether the trained preset model meets the preset standard; if each of the training sets has trained the preset model or the trained preset model meets the preset standard, the iterative repetition step is stopped, and the trained preset model when the iterative repetition step is stopped is used as the target model for calculating the associated value.
[0009] In an optional manner, the step of preprocessing the initial features to obtain target features further includes: performing dimensionality reduction processing on the initial features having a dimension higher than two dimensions to obtain initial features having a dimension equal to two dimensions after dimensionality reduction; and using the initial features having a dimension equal to or lower than two dimensions and the initial features having a dimension equal to two dimensions after dimensionality reduction as target features.
[0010] In an optional manner, the target features include body features and suspension features, wherein the body features include multiple types of sub-features; and the step of performing cluster analysis based on the target features to obtain a road surface division level further includes: taking the sub-features of the type that meet the preset type conditions as the sub-features to be processed, traversing each of the sub-features to be processed, and taking the traversed sub-features to be processed as the target sub-features; performing cluster analysis on the target sub-features and the suspension features to obtain an initial clustering result; performing cluster analysis on the initial clustering result to obtain a sub-road surface division level corresponding to the target sub-feature; based on the sub-road surface division level, obtaining the road surface division level corresponding to the target feature; wherein the road surface division level includes multiple sub-road surface division levels.
[0011] In an optional manner, the initial clustering result includes multiple types of clustering data; the step of performing cluster analysis on the initial clustering result to obtain the sub-road surface division level corresponding to the target sub-feature further includes: traversing each type of clustering data in the initial clustering result, and using the traversed clustering data as the target clustering data; performing cluster analysis on the target clustering data to obtain the single-class road surface division level corresponding to the target clustering data, and then obtaining the single-class road surface division level corresponding to each type of clustering data; obtaining the sub-road surface division level corresponding to the target sub-feature based on each type of clustering data and the single-class road surface division level corresponding to each type of clustering data.
[0012] According to another aspect of an embodiment of the present application, a device for road surface grade division based on vehicle driving characteristics is provided, the device comprising: a parameter calculation module, for calculating a parameter value corresponding to a parameter name carried by a preset feature based on sample data, and obtaining a target parameter according to the parameter name and the parameter value; wherein the sample data is driving feature data collected in real time during vehicle driving; a model training module, for iteratively training and verifying a preset model according to the sample data, so as to obtain a target model for calculating an associated value; wherein the associated value represents the degree of association between the target parameter and the road surface grade division; a feature screening module, for screening the target parameter carrying the associated value output by the target model, and taking the target parameter whose associated value meets the preset threshold as the initial feature; a road surface recognition module, for preprocessing the initial feature to obtain the target feature, and performing cluster analysis based on the target feature to obtain the road surface grade division.
[0013] According to another aspect of an embodiment of the present application, a device is provided, comprising: a controller; and a memory for storing one or more programs, wherein when the one or more programs are executed by the controller, the controller implements the method for road grade classification based on vehicle driving characteristics as described in any one of the above claims.
[0014] According to another aspect of an embodiment of the present application, a storage medium is provided, in which a computer program is stored. The computer program includes at least one executable instruction, and the executable instruction enables the device / equipment to perform the operations in the method for road surface grade classification based on vehicle driving characteristics as described in any one of the above claims.
[0015] The embodiment of the present application obtains the target parameter by calculating the parameter value corresponding to the parameter name carried by the preset feature, thereby expanding the preset feature; at the same time, the target parameter is calculated for correlation through the target model after iterative training verification, and the target parameter that meets the preset threshold is selected as the initial feature, thereby screening the target parameter based on correlation; the screened target parameter is then used as the initial feature, and the target feature is obtained after processing, and the target feature is clustered and analyzed to obtain the road surface classification level. Finally, cluster analysis is performed according to the target feature to obtain the road surface classification level. Different from the conventional road surface classification based on road roughness, the clustered road classification level can be more adapted to the road surface due to feature expansion, correlation calculation and screening, and is therefore more suitable for road surface adaptability research in the vehicle chassis system.
[0016] The above description is only an overview of the technical solution of the embodiment of the present application. In order to more clearly understand the technical means of the embodiment of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the embodiment of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] The accompanying drawings are only used to illustrate the embodiments and are not to be considered as limiting the present application. In addition, the same reference symbols are used to represent the same components throughout the accompanying drawings. In the accompanying drawings:
[0018] Figure 1 A schematic diagram showing a flow chart of an embodiment of a method for road surface grade classification based on vehicle driving characteristics provided by the present application;
[0019] Figure 2 A schematic flow chart of another embodiment of a road surface grade classification method based on vehicle driving characteristics provided by the present application is shown;
[0020] Figure 3 A schematic diagram showing a division process of sample data in one embodiment is shown;
[0021] Figure 4 A schematic flow chart of another embodiment of a road surface grade classification method based on vehicle driving characteristics provided by the present application is shown;
[0022] Figure 5 A schematic flow chart of another embodiment of a road surface grade classification method based on vehicle driving characteristics provided by the present application is shown;
[0023] Figure 6 A schematic flow chart of another embodiment of a road surface grade classification method based on vehicle driving characteristics provided by the present application is shown;
[0024] Figure 7A diagram showing initial clustering of feature data after dimensionality reduction in one embodiment;
[0025] Figure 8 A schematic flow chart of another embodiment of a road surface grade classification method based on vehicle driving characteristics provided by the present application is shown;
[0026] Fig. 9 A schematic flow chart of another embodiment of a road surface grade classification method based on vehicle driving characteristics provided by the present application is shown;
[0027] Fig.10 A data analysis schematic diagram showing an embodiment in which the slope amplitude is selected as a fixed value of 0.03;
[0028] Fig.11 Shows Fig.10 Schematic diagram of the data after cluster analysis;
[0029] Fig.12 A schematic diagram showing the weight feature importance of features related to vehicle body pitch angle features, left front suspension speed features, and left rear suspension speed features under different road surface grades in one embodiment;
[0030] Fig.13 A schematic diagram showing the weight feature importance of features related to the left front suspension speed feature and the left rear suspension speed feature under different road surface grades in one embodiment;
[0031] Fig.14 A schematic diagram showing the weight feature importance of vehicle body pitch angle-related features at different road surface levels in one embodiment;
[0032] Fig.15 A schematic diagram showing the result of cluster analysis using vehicle body vertical velocity characteristics as sub-features in one embodiment is shown;
[0033] Fig.16 A schematic diagram showing the structure of an embodiment of a road surface grade classification device based on vehicle driving characteristics provided by the present application;
[0034] Fig.17 A schematic structural diagram of an embodiment of the device provided by the present application is shown.
[0035] A1-body pitch angle standard deviation feature; A2-body pitch angle zero crossing times ratio feature; D1-left front suspension speed standard deviation feature; D2-left front suspension speed zero crossing times ratio feature; E1-left rear suspension speed standard deviation feature; E2-left rear suspension speed zero crossing times ratio feature. DETAILED DESCRIPTION
[0036] Here, exemplary embodiments will be described in detail, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are only examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims.
[0037] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.
[0038] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.
[0039] The term "multiple" as used in this application refers to two or more than two. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the related objects are in an "or" relationship.
[0040] At present, some road surface grade identification generally belongs to one-stop modeling classification based only on supervised learning algorithms. The road surface grade is generally divided into 8 grades from A to H according to the road roughness. Then the road surface model and the quarter vehicle suspension model can be established to obtain the vehicle unsprung mass acceleration output under different road surface grades to further construct the model features. This method requires the establishment of a professional vehicle simulation model and data collection under different real road conditions, and then further integration into sequence data for end-to-end LSTM neural network modeling. Although a good fitting effect can be obtained, the model feature data construction is highly professional, the overall cost is high, and the divided grades cannot be directly given to the vehicle chassis system for road adaptability research, and further business conversion of the results is required.
[0041] Some other non-machine learning methods, such as road grade recognition based on vehicle speed and suspension travel, have difficulty in all-round collection and professional processing of acquired signals. The core is the module establishment of the feature matrix, the precise calculation and judgment of the deviation matrix, and finally the identification of road grades from A to H. It belongs to the precise calculation and division of mathematical models, and cannot perform extended calculations for new or special road grades, that is, it has poor generalization and is difficult to adapt to the road grade requirements for the vehicle chassis system.
[0042] Based on the above, the present invention proposes a new road surface grade classification method based on vehicle driving characteristics to solve the above problems, which is described in detail with the following embodiments.
[0043] Figure 1 The flowchart of an embodiment of the method for road surface classification based on vehicle driving characteristics of the present application is shown, and the method is executed by a computer device. Figure 1 As shown, the method comprises the following steps:
[0044] Step S110: Calculate the parameter value corresponding to the parameter name carried by the preset feature based on the sample data, and obtain the target parameter according to the parameter name and the parameter value.
[0045] Among them, the sample data is the driving characteristic data collected in real time during the vehicle driving process.
[0046] Specifically, the speed level distribution takes the range of 20-120km / h, and 6 speed points are taken at intervals of 20km / h for sample data collection. The preset features include the vertical speed of the vehicle body, the pitch angle of the vehicle body, the roll angle of the vehicle body, the left front suspension speed, the left rear suspension speed, the right front suspension speed, and the right rear suspension speed. The parameter names include the average value, the maximum value, the minimum value, the standard deviation, and the proportion of zero crossing times. Each preset feature carries multiple parameter names, corresponding to multiple target parameters of multiple preset features.
[0047] Step S120: Perform iterative training and verification processing on the preset model according to the sample data to obtain a target model for calculating the correlation value.
[0048] Among them, the correlation value represents the correlation degree between the target parameter and the road surface grade classification.
[0049] Specifically, the preset model can adopt a random forest model, the input of the target model is sample data and target parameters, and the output is the target parameter with associated values. Here, a supervised fit is adopted from the preset model to the target model. The numerical value of the associated value is positively correlated with the degree of association. The larger the associated value, the higher the degree of association between the target parameter and the road surface grade classification.
[0050] Step S130: Filter the target parameters with associated values output by the target model, and use the target parameters whose associated values meet the preset threshold as initial features.
[0051] Specifically, the target parameters are screened according to the correlation value, and the target parameters with high correlation value are screened as the initial features; the preset threshold is determined according to the correlation value of the actual feature, and the smallest correlation value among the top 40% of the target parameters is preferably used as the preset threshold. In this way, irrelevant target parameters can be effectively removed, avoiding the situation where there are too many data or irrelevant data that lead to deviation in clustering results when the features are used for clustering analysis later.
[0052] Step S140: pre-processing the initial features to obtain target features, and performing cluster analysis based on the target features to obtain road surface classification levels.
[0053] Among them, the preprocessing is dimensionality reduction processing.
[0054] Specifically, the initial features after dimensionality reduction are used as target features, and cluster analysis is performed to obtain the road surface classification level.
[0055] Beneficial effects: The embodiment of the present application obtains the target parameter by calculating the parameter value corresponding to the parameter name carried by the preset feature, thereby expanding the preset feature; at the same time, the target parameter is calculated for correlation through the target model after iterative training verification, and the target parameter that meets the preset threshold is selected as the initial feature, thereby screening the target parameter based on correlation; the screened target parameter is then used as the initial feature, and the target feature is obtained after processing, and the target feature is clustered and analyzed to obtain the road surface classification level. Finally, cluster analysis is performed according to the target feature to obtain the road surface classification level. Different from the conventional road surface classification based on road roughness, the clustered road classification level can be more adapted to the road surface due to feature expansion, correlation calculation and screening, and is therefore more suitable for road surface adaptability research in the vehicle chassis system.
[0056] This application uses the target model to screen out features with high correlation, and then obtains the conventional grade classification of the road surface through preprocessing and cluster analysis. It is separated from the official grade classification of road roughness A to H. The clustered road grade is more popular and more suitable for the adaptability research of the vehicle chassis system to the road surface.
[0057] In some embodiments, Figure 2 As shown, before the step of calculating the parameter value corresponding to the parameter name carried by the preset feature based on the sample data in the above step S110, it also includes:
[0058] Step S210: dividing the sample data according to the preset cycle length and the preset cycle interval.
[0059] Specifically, based on similar "sliding window" technology (such as Figure 3 As shown in the figure, the window size window (i.e., the preset period length) is the number of data records used to calculate the features, and the sliding step stride (i.e., the preset period interval) is the number of interval records between two adjacent windows, which is used to construct some statistical features.
[0060] Step S220: constructing parameter names of preset features according to the divided sample data, and obtaining preset features carrying parameter names.
[0061] Specifically, the following statistical features are mainly constructed: (1) mean value, (2) maximum value, (3) minimum value, (4) standard deviation, and (5) percentage of zero-crossing times. The percentage of zero-crossing times is obtained from a relatively common experience in the vehicle field. After obtaining a collected road condition, after a period of collection, the data of the six preset features mentioned in the above embodiment are collected every second according to time, and sorted according to time to obtain sample data.
[0062] Beneficial effect: In this embodiment, after the sample data is divided and processed according to the preset cycle length and the preset cycle interval, the parameter names of the established preset features are more comprehensive, and the sliding window refinement method is used to make the calculated data more representative when the parameter values are subsequently calculated.
[0063] In some embodiments, the parameter names include a mean value parameter, a maximum value parameter, a minimum value parameter, a standard deviation parameter, and a zero-crossing frequency ratio parameter; Figure 4 As shown, the step of calculating the parameter value corresponding to the parameter name carried by the preset feature based on the sample data in step S110 further includes:
[0064] Step S111: compare the values corresponding to the preset features in the sample data to obtain the maximum and minimum values of the preset features.
[0065] Specifically, according to the comparison of the values from large to small, the maximum value and the minimum value of each preset feature are determined, and then the maximum value parameter and the minimum value parameter of each preset feature are determined.
[0066] Step S112: sum the values corresponding to the preset features in the sample data and divide the sum by the number of values to calculate the average value of the preset features.
[0067] Among them, in the sample data, one preset feature corresponds to multiple values.
[0068] Specifically, multiple values of the same preset feature are summed up, and then the sum is divided by the number of values to obtain the average value corresponding to the preset feature, that is, the average value of the preset feature, and then the average value parameter of each preset feature is determined.
[0069] Step S113: Obtain the standard deviation of the preset feature according to the average value and the value corresponding to the preset feature in the sample data.
[0070] Specifically, the standard deviation is calculated based on the average value and each numerical value corresponding to the preset feature to obtain the standard deviation value of the preset feature, and then to obtain the standard deviation parameter of the preset feature.
[0071] Step S114: sort the values corresponding to the preset features in the sample data according to the collection time, count the number of times two adjacent values have different positive and negative signs, calculate the ratio of the counted number of times and the number of values, and calculate the proportion of zero-crossing times.
[0072] Among them, the proportion of zero-crossing times is a common experience in the vehicle field. For the vehicle field, it is necessary to count the proportion of zero-crossing times.
[0073] Specifically, the sample data is collected according to time, so it is necessary to sort the multiple values corresponding to the preset features according to the order of collection time, and then count the zero-crossing points of the sorted values. The zero-crossing point is the statistics of crossing the zero point, and the proportion is calculated based on the number of times obtained by statistics and the number of values, so as to obtain the proportion of the number of zero-crossing points of the preset feature in the sample data. Then, the zero-crossing point proportion parameter of each preset feature is determined.
[0074] Beneficial effect: In this embodiment, the minimum and maximum values of the preset features are determined by judging the maximum and minimum values according to multiple numerical values corresponding to the preset features in the sample data; at the same time, the average value and standard deviation of the preset features are calculated according to the numerical values corresponding to the preset features; and then the proportion of the number of zero crossings is counted according to the numerical values corresponding to the preset features, so as to calculate the parameter value corresponding to the parameter name carried by the preset feature, and the calculation of the parameter value corresponding to the parameter name carried by the preset feature is realized through simple mathematical calculation, thereby improving the overall processing rate.
[0075] In some embodiments, Figure 5 As shown, the sample data includes multiple copies; the above step S120 further includes:
[0076] Step S121: according to the sampling period corresponding to each sub-sample in the sample data, the sub-sample whose sampling period matches the preset sampling period is used as a validation set, and the sub-samples other than the validation set are used as a training set.
[0077] Specifically, for the N data samples collected, and including the 35 features obtained in the previous embodiment, considering the high similarity of adjacent data, there may be data leakage problems in randomly extracting and dividing the data set, so 10% of the continuous data is extracted from each data sample as the validation set, and the other data is used as the training set (i.e., 91 training validation ratio) to build a random forest model for supervised data model fitting. It should be noted that the validation set is continuous data, so the sampling period is used as a unit to distinguish the validation set from the training set.
[0078] Step S122: Use the training set to train the preset model to obtain the trained preset model, and use the verification set to verify whether the trained preset model meets the preset standard.
[0079] Specifically, the trained preset model is verified using a validation set, and the preset standard is that the accuracy reaches a preset value.
[0080] Step S123: taking the trained preset model as a new preset model, iteratively and repeatedly using the training set to train the preset model to obtain the trained preset model, and using the verification set to verify whether the trained preset model meets the preset standard.
[0081] Specifically, adopting an iterative and repeated training method can reduce errors and make the accuracy reach the preset value. Moreover, the training set used in each iteration comes from different sample data, which can make the model training more accurate.
[0082] Step S124: If each training set has been trained with a preset model or the trained preset model meets the preset standard, the iterative repetition step is stopped, and the trained preset model when the iterative repetition step is stopped is used as the target model for calculating the associated value.
[0083] Specifically, when the preset model meets the preset standard or the collected sample data is trained, the iterative repetition step is stopped, and the preset model trained when the iterative repetition step is stopped is used as the target model for calculating the associated value. In this way, a model with a higher accuracy can be obtained. Finally, the importance of each feature can be obtained.
[0084] Among the features output by the target model, the first 40% important features are mainly the standard deviation and maximum value of the right front suspension acceleration, the standard deviation and minimum value of the left front suspension acceleration, the maximum value, minimum value and standard deviation of the vehicle body pitch angle, the minimum value and maximum value of the left front suspension acceleration, the minimum value and standard deviation of the vehicle body vertical velocity and the standard deviation of the right front suspension velocity, etc. Based on the above, it can be concluded that the constructed original and aggregated features are effective for grade fitting of road conditions, so the next step of dimensionality reduction and clustering can be done based on the obtained important features.
[0085] Beneficial effect: By dividing the sample data into a validation set and a training set, and using the training set to train and verify the preset model based on an iterative repeated training method, a target model with a higher accuracy is obtained, thereby ensuring the accuracy of subsequent predictions.
[0086] In some embodiments, Figure 6 As shown, in step S140, the initial features are preprocessed to obtain the target features, further comprising:
[0087] Step S141: Perform dimensionality reduction processing on the initial features with a dimension higher than two dimensions to obtain the initial features with a dimension equal to two dimensions after dimensionality reduction.
[0088] Specifically, the dimensions of the initial features of the standard deviation and the proportion of zero-crossing times are higher than two dimensions. Initial features higher than two dimensions are difficult to display on the interface and difficult to aggregate with each other. Therefore, it is necessary to reduce the dimensionality of initial features higher than two dimensions.
[0089] Step S142: taking the initial features whose dimension is equal to or less than two dimensions and the initial features whose dimension is equal to two dimensions after dimensionality reduction as target features.
[0090] Specifically, the initial features whose dimensions are equal to or less than two dimensions after dimensionality reduction and the initial features whose dimensions are equal to or less than two dimensions are both used as target features. Figure 7 As shown in, they are all initial features after dimensionality reduction, that is, target features. Figure 7 The left figure in the middle is clustered into the four types of data shown in the four small figures on the right.
[0091] Beneficial effect: By reducing the dimension of the initial features with a dimension higher than two dimensions, the reduced-dimensional initial features are obtained, making the initial features easier to display and conducive to mutual clustering analysis.
[0092] In some embodiments, the target features include body features and suspension features, wherein the body features include multiple types of sub-features; Figure 8 As shown, the step S140 of performing cluster analysis according to the target features to obtain a road surface classification step further includes:
[0093] Step S141: taking sub-features whose types meet preset type conditions as sub-features to be processed, traversing each sub-feature to be processed, and taking the traversed sub-features to be processed as target sub-features.
[0094] Among them, the body features include body pitch angle features, body roll angle features and body vertical speed features.
[0095] Specifically, because the body pitch angle features and body vertical velocity features have the feasibility of further feature dimension clustering, while the body roll angle features do not have this feasibility, the sub-features that meet the preset type conditions have the feasibility of further feature dimension clustering, so the sub-features to be processed include the body pitch angle features and body vertical velocity features.
[0096] Step S142: performing cluster analysis on the target sub-features and suspension-type features to obtain initial clustering results.
[0097] Specifically, the suspension features are clustered with the body pitch angle features and the body vertical speed features. The suspension features are the left front suspension speed features, the left rear suspension speed features, the right front suspension speed features, and the right rear suspension speed features.
[0098] Step S143: Perform cluster analysis on the initial clustering results to obtain the sub-road surface division level corresponding to the target sub-feature.
[0099] Specifically, cluster analysis is performed again on the initial clustering results to obtain the sub-road surface division level corresponding to the target sub-feature. Each initial clustering result corresponds to a sub-road surface division level.
[0100] Step S144: Obtain the road surface classification level corresponding to the target feature according to the sub-road surface classification level.
[0101] The road surface division level includes multiple sub-road surface division levels.
[0102] Specifically, multiple sub-road surface division levels constitute the road surface division level corresponding to the target feature. In the present application, there are sub-road surface division levels corresponding to the vehicle body pitch angle feature and the vehicle body vertical speed feature. The road surface division level includes the sub-road surface division level corresponding to the vehicle body pitch angle feature and the sub-road surface division level corresponding to the vehicle body vertical speed feature.
[0103] Beneficial effect: By dividing the target features into body features and suspension features, and then clustering the sub-features of the body features with the suspension features, the road surface classification level corresponding to each body sub-feature is obtained. In addition, the features that cannot be clustered are screened out, and then the two dimensions of the road surface level are effectively classified in a popular way based on the "divide and conquer" thinking.
[0104] In some embodiments, the initial clustering results include multiple types of clustering data, such as Fig. 9 As shown, step S143 further includes:
[0105] Step S1431: traverse each type of cluster data in the initial clustering result, and use the traversed cluster data as target cluster data.
[0106] Specifically, the initial clustering result corresponds to the sub-road surface division level, and each type of clustering data corresponds to a single-type road surface division level, and multiple single-type road surface division levels constitute the sub-road surface division level.
[0107] Step S1432: Perform cluster analysis on the target cluster data to obtain the single-type road surface classification level corresponding to the target cluster data, and then obtain the single-type road surface classification level corresponding to each type of cluster data.
[0108] Specifically, cluster analysis is performed on the target cluster data, and then single-class road surface classification levels can be obtained; if there are several types of cluster data in each initial clustering result, then there are correspondingly several single-class road surface classification levels.
[0109] Step S1433: Obtain the sub-road surface division level corresponding to the target sub-feature according to each type of cluster data and the single-type road surface division level corresponding to each type of cluster data.
[0110] Specifically, the single-type road surface division level corresponding to each type of clustering data and each type of clustering data included in the initial clustering result can obtain the sub-road surface division level corresponding to the target sub-feature.
[0111] Beneficial effect: By further refining the clustering analysis of the initial clustering results, the sub-road surface division level steps corresponding to the target sub-features are obtained, thereby enriching and improving the solution of the present application, thereby making the road surface division more detailed and accurate.
[0112] In some embodiments, Figure 7 , Fig.10 and Fig.11 As shown in the figure, the body feature includes multiple types of sub-features, taking the body pitch angle sub-feature as an example. First, the dimensional feature data such as the body pitch angle, the standard deviation of each suspension acceleration, and the number of zero crossings are reduced based on the PCA machine learning model. For example, the interpretability of one of the two-dimensional data is [0.7809, 0.1332], which can represent 91% of the data information. The effect is shown above. Figure 7 Then, based on the agglomerative hierarchical clustering algorithm, Figure 7 The dimensionality reduction data on the left side is clustered as follows Figure 7The 4 categories shown on the right, and the 4th category is flatter than the 3rd category, so it can be manually classified into one category for the convenience of analysis, so it is finally clustered into 3 categories (i.e. 3 categories of clustered data), and the results are well interpretable. Then the slope amplitude is selected as a fixed value of 0.03 (i.e. one category of clustered data is used as an example for explanation), and then data analysis is performed based on different pitch sizes (which can be converted into the wavelength size of the simulated road surface) and driving speeds (such as Fig.10 As shown in the figure, we can find that: ① the pitch size is positively correlated with the slope amplitude; ② the pitch size is related to the driving speed; ③ a specific pitch size value can be formed on the road surface of each wavelength; ④ a smaller wavelength means that the road surface is relatively steeper and it is almost impossible to produce a speed above 60. Fig.10 The data analysis results in the data are further clustered and modeled, and finally we can get Fig.11 The clustering effect is that it is clustered into three clusters (three types) of road surface classification levels, that is, the road surface classification level of cluster data with the slope amplitude selected as a fixed value of 0.03 is obtained. Fig.11 The road surface classification level obtained in is to divide the road surface into three categories. This is the single-class road surface classification level when the slope amplitude is selected as a fixed value of 0.03 (this type of clustering data). This single-class road surface classification level can divide the road surface into three categories: "relatively flat road", "relatively gentle" and "relatively steep slope".
[0113] The roads classified as "relatively flat roads" include: roads corresponding to various vehicle speeds under long wave 30, roads corresponding to the average vehicle speed in the range of 20-100km / h under long wave 25, roads corresponding to the average vehicle speed in the range of 20-80km / h under long wave 20, roads corresponding to the average vehicle speed in the range of 20-60km / h under long wave 15, roads corresponding to the average vehicle speed in the range of 20-40km / h under long wave 10, and roads corresponding to the average vehicle speed in the range of 20km / h under long wave 5.
[0114] "Relatively smooth" includes: roads with an average vehicle speed of 120km / h under long wave 25, roads with an average vehicle speed of 100-120km / h under long wave 20, roads with an average vehicle speed of 80-100km / h under long wave 15, and roads with an average vehicle speed of 60km / h under long wave 10.
[0115] The "relatively steep slopes" include: roads with an average vehicle speed of 120km / h under long wave 15, roads with an average vehicle speed of 80-120km / h under long wave 10, and roads with an average vehicle speed of 40-60km / h under long wave 5.
[0116] According to the situation of this embodiment, the single-type road surface division levels corresponding to the clustering data of the slope amplitude selected as other fixed value classes are clustered and analyzed, and the sub-road surface division levels corresponding to the vehicle body pitch angle characteristics are obtained.
[0117] In some embodiments, Figure 12-14 As shown, the cluster intersection obtained by using two cluster analyses in the above embodiment is verified, and then three features are used for classification. The feature weights are as follows: Figure 12-14 As shown, Fig.12 In order to verify the fitting of the classification model, the vehicle pitch angle features, the left front suspension speed features and the left rear suspension speed features were used, A1 is the standard deviation feature of the vehicle pitch angle, A2 is the proportion feature of the number of times the vehicle pitch angle crosses zero, D1 is the standard deviation feature of the left front suspension speed, D2 is the proportion feature of the number of times the left front suspension speed crosses zero, E1 is the standard deviation feature of the left rear suspension speed, and E2 is the proportion feature of the number of times the left rear suspension speed crosses zero. The accuracy rate can reach 100%. Fig.13 In order to verify the fitting of the classification model, the left front suspension speed features and the left rear suspension speed features were used. D1 is the standard deviation feature of the left front suspension speed, D2 is the proportion feature of the number of zero-crossing times of the left front suspension speed, E1 is the standard deviation feature of the left rear suspension speed, and E2 is the proportion feature of the number of zero-crossing times of the left rear suspension speed. The accuracy rate can reach 99.6%. Fig.14 Only the relevant features of the vehicle body pitch angle characteristics were used for classification model fitting verification. A1 is the standard deviation feature of the vehicle body pitch angle, and A2 is the proportion of the number of times the vehicle body pitch angle crosses zero. The accuracy rate is 96.9%. In summary, the clustering effect after integrating the overall relevant features of the vehicle body pitch angle characteristics, the left front suspension speed characteristics, and the left rear suspension speed characteristics can meet the standard.
[0118] Beneficial effect: Through the verification of this embodiment, it can be seen that the features selected for cluster analysis by the solution provided in this application are accurate, and the results of the subsequent cluster analysis are also accurate, which proves the feasibility of the solution of this application.
[0119] In some embodiments, Fig.15 As shown in the figure, when the vehicle body vertical speed feature is used as a sub-feature, a cluster analysis is performed, and the following is obtained: Fig.15 The cluster analysis results shown in the figure show that Fig.15It can be seen that in each ah graph, there are at most two categories, so the road surface corresponding to the body vertical dimension feature is divided into two categories, including "high vertical vibration" and "low vertical vibration". Since clustering analysis based on the body vertical speed feature as a sub-feature is not very relevant to the subsequent chassis application in practice, the example of the clustering scheme of this application is described in detail in the previous embodiment with the body pitch angle feature as a sub-feature, so here we only show that the clustering scheme based on this application is feasible, and no detailed clustering description is given.
[0120] Fig.16 The schematic diagram of the structure of the embodiment of the road surface grade classification device based on vehicle driving characteristics of the present application is shown. Fig.16 As shown, the device 300 includes: a parameter calculation module 310, a model training module 320, a feature screening module 330 and a road surface recognition module 340.
[0121] The parameter calculation module 310 is used to calculate the parameter value corresponding to the parameter name carried by the preset feature based on the sample data, and obtain the target parameter according to the parameter name and the parameter value; wherein the sample data is the driving feature data collected in real time during the vehicle driving process;
[0122] The model training module 320 is used to perform iterative training and verification processing on the preset model according to the sample data to obtain a target model for calculating the correlation value; wherein the correlation value represents the degree of correlation between the target parameter and the road surface grade classification;
[0123] A feature screening module 330 is used to screen target parameters with associated values output by the target model, and use the target parameters whose associated values meet a preset threshold as initial features;
[0124] The road surface recognition module 340 is used to pre-process the initial features to obtain target features, and perform cluster analysis based on the target features to obtain road surface classification levels.
[0125] Beneficial effects: The embodiment of the present application obtains the target parameter by calculating the parameter value corresponding to the parameter name carried by the preset feature, thereby expanding the preset feature; at the same time, the target parameter is calculated for correlation through the target model after iterative training verification, and the target parameter that meets the preset threshold is selected as the initial feature, thereby screening the target parameter based on correlation; the screened target parameter is then used as the initial feature, and the target feature is obtained after processing, and the target feature is clustered and analyzed to obtain the road surface classification level. Finally, cluster analysis is performed according to the target feature to obtain the road surface classification level. Different from the conventional road surface classification based on road roughness, the clustered road classification level can be more adapted to the road surface due to feature expansion, correlation calculation and screening, and is therefore more suitable for road surface adaptability research in the vehicle chassis system.
[0126] This application uses the target model to screen out features with high correlation, and then obtains the conventional grade classification of the road surface through preprocessing and cluster analysis. It is separated from the official grade classification of road roughness A to H. The clustered road grade is more popular and more suitable for the adaptability research of the vehicle chassis system to the road surface.
[0127] It should be noted that the road surface grade classification device based on vehicle driving characteristics provided in the above embodiment and the road surface grade classification method based on vehicle driving characteristics provided in the aforementioned embodiment belong to the same concept, and the specific manner in which each module and unit performs operations has been described in detail in the method embodiment and will not be repeated here.
[0128] Fig.17 A schematic diagram of the structure of an embodiment of the device of the present application is shown, which shows a schematic diagram of the structure of a computer system suitable for implementing the device of the embodiment of the present application. The specific embodiment of the present application does not limit the specific implementation of the device.
[0129] See also Fig.17 As shown, the device includes: a controller; a memory for storing one or more programs, and when the one or more programs are executed by the controller, the above-mentioned road surface grade classification method based on vehicle driving characteristics is executed.
[0130] Please continue reading Fig.17 As shown, the computer system 400 of the device includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage part 408 to the random access memory (RAM) 403, such as executing the method in the above embodiment. In RAM 403, various programs and data required for system operation are also stored. CPU 401, ROM 402 and RAM 403 are connected to each other through bus 404. Input / output (I / O) interface 405 is also connected to bus 404.
[0131] The following components are connected to the I / O interface 405: an input section 406 including a keyboard, a mouse, etc.; an output section 407 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as needed so that a computer program read therefrom is installed into the storage section 408 as needed.
[0132] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication section 409, and / or installed from a removable medium 411. When the computer program is executed by a central processing unit (CPU) 401, various functions defined in the system of the present application are executed.
[0133] Another aspect of the present application also provides a storage medium on which a computer program is stored, and when the computer program is executed by a processor, the road surface grade classification method based on vehicle driving characteristics as described above is implemented. The computer readable storage medium may be included in the device described in the above embodiment, or may exist independently without being assembled into the electronic device.
[0134] Another aspect of the present application also provides a computer program product or a computer program, which includes at least one executable instruction. When the executable instruction is executed on a device / equipment, the device / equipment executes the road surface grade classification method based on vehicle driving characteristics as described above.
[0135] Beneficial effects: The embodiment of the present application obtains the target parameter by calculating the parameter value corresponding to the parameter name carried by the preset feature, thereby expanding the preset feature; at the same time, the target parameter is calculated for correlation through the target model after iterative training verification, and the target parameter that meets the preset threshold is selected as the initial feature, thereby screening the target parameter based on correlation; the screened target parameter is then used as the initial feature, and the target feature is obtained after processing, and the target feature is clustered and analyzed to obtain the road surface classification level. Finally, cluster analysis is performed according to the target feature to obtain the road surface classification level. Different from the conventional road surface classification based on road roughness, the clustered road classification level can be more adapted to the road surface due to feature expansion, correlation calculation and screening, and is therefore more suitable for road surface adaptability research in the vehicle chassis system.
[0136] This application uses the target model to screen out features with high correlation, and then obtains the conventional grade classification of the road surface through preprocessing and cluster analysis. It is separated from the official grade classification of road roughness A to H. The clustered road grade is more popular and more suitable for the adaptability research of the vehicle chassis system to the road surface.
[0137] It should be noted that the computer-readable medium shown in the embodiment of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it. In the present application, a computer-readable signal medium can include a data signal propagated in a baseband or as a part of a carrier wave, wherein a computer-readable computer program is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.
[0138] The flowchart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program product according to various embodiments of the present application. Wherein, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
[0139] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. The names of these units do not, in some cases, constitute limitations on the units themselves.
[0140] According to one aspect of an embodiment of the present application, a computer system is also provided, including a central processing unit (CPU), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) or a program loaded from a storage portion into a random access memory (RAM), such as executing the method in the above embodiment. In RAM, various programs and data required for system operation are also stored. CPU, ROM and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.
[0141] The following components are connected to the I / O interface: an input part including a keyboard, a mouse, etc.; an output part including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage part including a hard disk, etc.; and a communication part including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication part performs communication processing via a network such as the Internet. A drive is also connected to the I / O interface as needed. Removable media, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., are installed on the drive as needed so that the computer program read therefrom is installed into the storage part as needed.
[0142] The above content is only a preferred exemplary embodiment of the present application and is not intended to limit the implementation scheme of the present application. A person skilled in the art can easily make corresponding changes or modifications based on the main concept and spirit of the present application. Therefore, the scope of protection of the present application shall be based on the scope of protection required by the claims.
Claims
1. A road surface grade classification method based on vehicle driving characteristics, characterized in that: The method comprises: Calculate the parameter value corresponding to the parameter name carried by the preset feature based on the sample data, and obtain the target parameter according to the parameter name and the parameter value; wherein the sample data is the driving feature data collected in real time during the vehicle driving process; Iteratively training and verifying the preset model according to the sample data to obtain a target model for calculating the correlation value; wherein the correlation value represents the degree of correlation between the target parameter and the road surface grade classification; Filtering target parameters with associated values output by the target model, and taking target parameters with associated values that meet a preset threshold as initial features; The initial features are preprocessed to obtain target features, and cluster analysis is performed based on the target features to obtain road surface classification levels.
2. The method according to claim 1, characterized in that Before the step of calculating the parameter value corresponding to the parameter name carried by the preset feature based on the sample data, the method further includes: Divide and process the sample data according to the preset cycle length and the preset cycle interval; According to the sample data after the division process, the parameter name of the preset feature is constructed to obtain the preset feature carrying the parameter name.
3. The method according to claim 1 or 2, characterized in that: The parameter names include average value parameter, maximum value parameter, minimum value parameter, standard deviation parameter and zero crossing times ratio parameter; The step of calculating the parameter value corresponding to the parameter name carried by the preset feature based on the sample data further includes: Comparing the values corresponding to the preset features in the sample data to obtain the maximum and minimum values of the preset features; The values corresponding to the preset features in the sample data are summed and then divided by the number of the values to calculate an average value of the preset features; Obtaining a standard deviation of the preset feature according to the average value and a value corresponding to the preset feature in the sample data; The values corresponding to the preset features in the sample data are sorted according to the collection time, and the number of times two adjacent values have different positive and negative signs is counted, and the ratio of the counted number to the number of values is calculated to obtain the proportion of zero-crossing times.
4. The method according to claim 1, characterized in that: The sample data includes multiple copies; The step of iteratively training and validating the preset model according to the sample data to obtain a target model for calculating the correlation value further includes: According to the sampling period corresponding to each sub-sample in the sample data, the sub-sample whose sampling period matches the preset sampling period is used as a validation set, and the sub-samples other than the validation set are used as a training set; Using the training set to train a preset model to obtain a trained preset model, and using the verification set to verify whether the trained preset model meets the preset standard; The trained preset model is used as a new preset model, and the steps of iteratively repeating the steps of using the training set to train the preset model to obtain the trained preset model, and using the verification set to verify whether the trained preset model meets the preset standard; If each of the training sets has been trained with a preset model or the trained preset model meets the preset standard, the iterative repetition step is stopped, and the trained preset model when the iterative repetition step is stopped is used as the target model for calculating the associated value.
5. The method according to claim 1, characterized in that The step of preprocessing the initial features to obtain target features further includes: Performing dimensionality reduction processing on the initial features having a dimension higher than two dimensions to obtain initial features having a dimension equal to two dimensions after dimensionality reduction; The initial features whose dimension is equal to or lower than two dimensions and the initial features whose dimension is equal to two dimensions after dimensionality reduction are used as target features.
6. The method according to claim 1, characterized in that The target features include body features and suspension features, wherein the body features include multiple types of sub-features; The step of performing cluster analysis according to the target features to obtain a road surface classification level further includes: Taking the sub-features of the type that meet the preset type conditions as the sub-features to be processed, traversing each of the sub-features to be processed, and taking the traversed sub-features to be processed as target sub-features; Performing cluster analysis on the target sub-features and the suspension-related features to obtain an initial clustering result; Performing cluster analysis on the initial clustering results to obtain the sub-road surface division level corresponding to the target sub-feature; According to the sub-road surface division level, the road surface division level corresponding to the target feature is obtained; wherein the road surface division level includes a plurality of sub-road surface division levels.
7. The method according to claim 6, characterized in that The initial clustering result includes multiple types of clustering data; The step of performing cluster analysis on the initial clustering results to obtain the sub-road surface division level corresponding to the target sub-feature further includes: Traversing each type of clustering data in the initial clustering result, and taking the traversed clustering data as target clustering data; Performing cluster analysis on the target cluster data to obtain the single-type road surface classification level corresponding to the target cluster data, and then obtaining the single-type road surface classification level corresponding to each type of cluster data; The sub-road surface division level corresponding to the target sub-feature is obtained according to each type of cluster data and the single-type road surface division level corresponding to each type of cluster data.
8. A road surface grade classification device based on vehicle driving characteristics, characterized in that: The device comprises: A parameter calculation module, used to calculate the parameter value corresponding to the parameter name carried by the preset feature based on the sample data, and obtain the target parameter according to the parameter name and the parameter value; wherein the sample data is the driving characteristic data collected in real time during the vehicle driving process; A model training module, used for iteratively training and verifying a preset model according to the sample data to obtain a target model for calculating a correlation value; wherein the correlation value represents the degree of correlation between the target parameter and the road surface grade classification; A feature screening module, used for screening target parameters with associated values output by the target model, and taking target parameters with associated values meeting a preset threshold as initial features; The road surface recognition module is used to pre-process the initial features to obtain target features, and perform cluster analysis based on the target features to obtain road surface classification levels.
9. A device, characterized in that: include: Controller; A memory for storing one or more programs, which, when executed by the controller, enables the controller to implement the road grade classification method based on vehicle driving characteristics as described in any one of claims 1 to 7.
10. A storage medium, characterized in that: The storage medium stores a computer program, which includes at least one executable instruction. When the executable instruction is executed on the device / equipment, the device / equipment performs the operation of the road surface recognition method based on vehicle driving characteristics as described in any one of claims 1 to 7.
