Vehicle driving condition construction method and device and readable storage medium

By screening and sorting vehicle driving data, the vehicle driving conditions that reflect actual road conditions changes are constructed, which solves the problem of inaccurate working conditions in the existing technology and improves the accuracy of energy consumption and emission monitoring.

CN120183060APending Publication Date: 2025-06-20GREAT WALL NEW ENERGY COMMERCIAL VEHICLE CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311768299.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-12-20
Publication Date
2025-06-20

AI Technical Summary

Technical Problem

It is difficult for the prior art to accurately construct vehicle driving conditions that reflect changes in actual road conditions, resulting in a decrease in the accuracy of vehicle performance indicator measurement.

Method used

By collecting vehicle driving data, determining driving status, dividing kinematic segments, and screening and sorting the segments, the vehicle driving conditions are constructed. This method considers speed, slope and time sequence, improving the accuracy of the working conditions.

Benefits of technology

It improves the accuracy of monitoring indicators such as energy consumption or emissions in driving conditions, and can more accurately reflect changes in actual road conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120183060A_ABST
    Figure CN120183060A_ABST
Patent Text Reader

Abstract

The embodiment of the invention provides a vehicle driving condition construction method and device and a readable storage medium. The method comprises the following steps: collecting vehicle driving data and determining the driving state of a vehicle according to the vehicle driving data; determining kinematics fragments according to the driving state of the vehicle, and marking a sequence number for each kinematics fragment according to a time sequence; then, the processed kinematics fragments are screened, the screened kinematics fragments are arranged according to the serial numbers, and therefore the vehicle driving working condition is constructed according to the arranged kinematics fragments; the accuracy of monitoring indexes such as energy consumption or emission under the driving condition is improved, and the change of the actual road condition can be reflected more accurately.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of driving condition construction, and particularly to a method, device and readable storage medium for constructing a vehicle driving condition. Background Art

[0002] The driving cycle of an automobile, also known as the Driving Cycle, reflects the variation of vehicle speed over time during driving and can reflect the motion characteristics during vehicle driving. The driving cycle of an automobile is an important and common basic technology in the automobile industry, the basis for vehicle energy consumption / emission test methods and limit standards, and also the main benchmark for optimizing various performance indicators of automobiles. Currently, relatively mature driving cycles include NEDC (New European Driving Cycle) and WLTC (Worldwide Harmonized Light Vehicles Test Cycle). However, due to regional differences and traffic condition changes, foreign standard driving cycles are gradually not applicable to the actual road conditions in China.

[0003] The short-trip analysis method is a relatively commonly used driving condition construction method. It divides the entire trip into small kinematic segments, obtains the characteristics of the entire trip through the analysis of the kinematic segments, and mainly targets urban complex road conditions when constructing a driving condition cycle through the short-trip analysis method. However, currently, most scholars often cannot restore the time sequence of actual data for the vehicle driving condition segments obtained through kinematic segment analysis, and cannot reflect the changes in actual road conditions, thereby reducing the accuracy of the vehicle driving condition for measuring various performance indicators. Summary of the Invention

[0004] In view of the above problems, a method, device and readable storage medium for constructing a vehicle driving condition are provided to overcome or at least partially solve the above problems, including:

[0005] A method for constructing a vehicle driving condition, the method includes:

[0006] Collect vehicle driving data and determine the driving state of the vehicle according to the vehicle driving data;

[0007] Determine kinematic segments according to the driving state of the vehicle and number each kinematic segment in chronological order;

[0008] Screen the kinematic segments and arrange the screened kinematic segments in the order of the numbers, and construct a vehicle driving condition according to the arranged kinematic segments.

[0009] Optionally, at least speed data and gradient data are included in the vehicle driving data. Before determining the driving state of the vehicle based on the vehicle driving data, it further includes:

[0010] Preprocessing the vehicle driving data; wherein, the preprocessing at least includes smoothing the speed data and the gradient data, and when smoothing the speed data, the smoothing window used is larger than the smoothing window used when smoothing the gradient data.

[0011] Optionally, before screening the kinematic segments and arranging the screened kinematic segments in the order, it further includes:

[0012] Determining characteristic parameters based on the vehicle driving data, and processing the kinematic segments according to the characteristic parameters.

[0013] Optionally, the driving state at least includes an idle state. After determining the kinematic segments based on the driving state of the vehicle, it further includes:

[0014] In all the kinematic segments, detecting whether there is a kinematic segment in which the duration of the idle state is greater than a first threshold. If so, deleting a partial segment content in the first kinematic segment that is in the idle state according to a first preset rule, where the first kinematic segment is the kinematic segment in which the duration of the idle state is greater than the first threshold.

[0015] Optionally, determining the characteristic parameters based on the vehicle driving data includes:

[0016] Calculating acceleration data based on the vehicle driving data, and determining the characteristic parameters based on the acceleration data and the vehicle driving data;

[0017] Calculating a correlation matrix based on the characteristic parameters, and obtaining a correlation coefficient based on the correlation matrix;

[0018] Determining a plurality of first characteristic parameters from all the characteristic parameters based on the correlation coefficient, where the first characteristic parameters are the characteristic parameters with larger correlation coefficients.

[0019] Optionally, processing the kinematic segments according to the characteristic parameters includes:

[0020] Performing dimensionality reduction processing on the first characteristic parameters to obtain a dimensionality reduction processing result;

[0021] Performing clustering processing on all the kinematic segments according to the dimensionality reduction processing result to obtain multiple types of kinematic segments.

[0022] Optionally, screening the kinematic segments and arranging the screened kinematic segments according to the serial numbers includes:

[0023] Calculating the proportion of the duration of each kinematic segment in various types of kinematic segments in the set duration of the vehicle driving condition;

[0024] Performing a primary screening on each kinematic segment in various types of kinematic segments according to the proportion of the duration of each kinematic segment in the set duration to obtain a primary screening result;

[0025] For the primary screening result, performing a secondary screening on the primary screening result according to the distance between each kinematic segment in various types of kinematic segments and the clustering center to obtain a secondary screening result;

[0026] For the secondary screening result, performing a tertiary screening on the secondary screening result according to the magnitude of the serial numbers of each kinematic segment in various types of kinematic segments to obtain a tertiary screening result;

[0027] Arranging the kinematic segments in the tertiary screening result according to the magnitude of the serial numbers.

[0028] A vehicle driving condition construction device, the device includes:

[0029] A driving state determination module, configured to collect vehicle driving data and determine the driving state of the vehicle according to the vehicle driving data;

[0030] A kinematic segment determination module, configured to determine kinematic segments according to the driving state of the vehicle and label serial numbers for each kinematic segment in chronological order;

[0031] A vehicle driving condition construction module, configured to screen the processed kinematic segments, arrange the screened kinematic segments according to the serial numbers, and construct a vehicle driving condition according to the arranged kinematic segments.

[0032] An electronic device, including a processor, a memory, and a computer program stored on the memory and capable of running on the processor, where when the computer program is executed by the processor, the vehicle driving condition construction method described above is implemented.

[0033] A computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the vehicle driving condition construction method described above is implemented.

[0034] The embodiments of the present invention have the following advantages:

[0035] In an embodiment of the present invention, vehicle driving data is collected and the driving state of the vehicle is determined based on the vehicle driving data; then kinematic segments are determined according to the driving state of the vehicle and serial numbers are marked for each of the kinematic segments in chronological order; subsequently, the processed kinematic segments are screened and the screened kinematic segments are arranged according to the serial numbers, so as to construct a vehicle driving condition based on the arranged kinematic segments; the accuracy of monitoring indexes such as energy consumption or emissions by the driving condition is improved; the sorting method of the driving condition segments restores the chronological order of the real data as much as possible, and can more accurately reflect the changes in the actual road conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for the description of the present invention will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0037] Figure 1 is a flowchart of the steps of a method for constructing a vehicle driving condition provided by an embodiment of the present invention;

[0038] Figure 2 is an implementation step diagram of the method for constructing a vehicle driving condition provided by the present invention;

[0039] Figure 3 is a flowchart of the steps of another method for constructing a vehicle driving condition provided by an embodiment of the present invention;

[0040] Figure 4 is a structural block diagram of a device for a method for constructing a vehicle driving condition provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] In order to make the above objects, features and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the drawings and specific embodiments. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0042] The driving cycle of a vehicle, also known as the driving cycle of an automobile, reflects how the speed of the vehicle changes over time during driving and can reflect the motion characteristics during vehicle driving. The driving cycle of a vehicle is an important and common basic technology in the automotive industry, the basis for vehicle energy consumption / emission test methods and limit standards, and also the main benchmark when optimizing various performance indicators of vehicles. Currently, the relatively mature driving cycles are NEDC and WLTC. However, due to regional differences and changes in traffic conditions, foreign standard driving cycles are gradually not applicable to the actual road conditions in China.

[0043] The short-trip analysis method is a commonly used method for constructing a driving cycle. It divides the entire trip into small kinematic segments and obtains the characteristics of the entire trip through the analysis of the kinematic segments. However, currently, most scholars only consider the characteristics related to speed and acceleration in the analysis of kinematic segments, with fewer characteristics considered. This will reduce the accuracy of the vehicle driving cycle in measuring various performance indicators, and the obtained vehicle driving cycle segments often cannot restore the chronological order of the actual data and cannot reflect the changes in the actual road conditions.

[0044] To solve the above problems, the present invention improves the vehicle driving cycle construction method in the related technology based on the short-trip analysis method. The following will describe the present invention in detail with reference to the accompanying drawings:

[0045] Refer to Figure 1 , which shows the step flow chart of a vehicle driving cycle construction method provided by an embodiment of the present invention, and specifically may include the following steps:

[0046] Step 101, collect vehicle driving data and determine the driving state of the vehicle according to the vehicle driving data;

[0047] In a specific implementation, vehicle driving data including speed data and slope data can be collected, and based on the collected data, the acceleration per second can be calculated according to the acceleration calculation formula in the national standard document "China Automotive Driving Cycle", and the driving state per second can be divided according to the division standards for acceleration, deceleration, constant speed, and idling in the national standard document.

[0048] In an embodiment of the present invention, the vehicle driving data at least includes speed data and slope data. Before determining the driving state of the vehicle according to the vehicle driving data, it further includes:

[0049] Preprocess the vehicle driving data; wherein, the preprocessing at least includes smoothing the speed data and the slope data, and when smoothing the speed data, the smoothing window used is larger than the smoothing window used when smoothing the slope data.

[0050] For the sake of easy understanding, some terms involved are explained as follows:

[0051] Pulse noise filter: The pulse noise filter is mainly used to identify and remove pulse noise in vehicle driving record data. Pulse noise is usually a sudden, short-term, and abnormal signal, which may be caused by improper driving or external interference, etc. Such noise is non-continuous, and the filter can identify data points (singular points) that do not conform to the normal driving pattern and remove or correct these points to ensure the coherence and accuracy of the data.

[0052] The pulse noise filter can perform corresponding processing according to the following formula:

[0053]

[0054] where v t is the vehicle driving speed at time t.

[0055] High-frequency noise filter: The high-frequency noise filter is mainly used to reduce or eliminate the high-frequency part in driving data. High-frequency noise usually shows rapid and short-term fluctuations in the signal, which may be caused by uneven road surfaces, vehicle vibrations, etc. This filter reduces the impact of high-frequency noise through smoothing processing, aiming to eliminate extreme changes in the short term and obtain a smoother signal curve to reflect a more real driving condition.

[0056] Currently, most scholars consider only the features related to speed and acceleration when analyzing kinematic segments, often ignoring the impact of slope on energy consumption, which will reduce the accuracy of the constructed driving conditions for measuring various performance indicators; therefore, vehicle driving data including speed data and slope data can be collected;

[0057] In practical applications, there may be cases of missing or duplicate data in the original driving data, and there may also be outliers in the collected data. Therefore, it is necessary to preprocess the data. When preprocessing, duplicate data can be directly deleted, and for missing data, when there are many missing data, the kinematic segment where the missing value is located can be selected for deletion, and when there are few missing data, interpolation processing can be performed. Specifically, since the collected data is often a time series, it is not reasonable and accurate enough to fill it with other indicators such as the average value or median, so linear interpolation can be used. Obvious abnormal data (such as a very large speed, a large change in slope in the previous and next two seconds, etc.) are deleted, and then linear interpolation is performed to ensure the continuity of time.

[0058] Further, noise reduction and smoothing processing can be performed on the speed data and slope data. Since the speed change within a short period of time may not be significant, but due to different road conditions, the slope data is very likely to have large fluctuations within a short period of time. Therefore, the smoothing coefficients used for smoothing processing should not be the same. The smoothing window used for speed data can be larger than the smoothing window used for slope data. When performing smoothing processing with a high-frequency noise filter, it can be carried out according to the following formula:

[0059]

[0060] In the above formula, K(x) is the weight value of the vehicle speed before and after the t moment. In this application, when performing smoothing processing on the speed, the value of h can be 6 seconds, and when performing smoothing processing on the slope, the value of h can be 2 seconds. The definition of K(x) is:

[0061]

[0062] It can be seen from the above smoothing formula that the smoothed speed or slope at the current time is calculated based on the speeds or slopes in the first few seconds and the last few seconds of the window size. For example, when the current time is the t-th second and the window size (i.e., the value of h) is 6, the value range of the window size is [-6, 6]. Calculate the coefficient of each value in the window within this interval, that is, K(x). Multiply K(x) by the speed at the (t + window value) second, and then sum and average the products of each K(x) and the speed at the (t + window value) second between [-6, 6], which is the smoothed speed at the current t-th second. That is to say, the smoothed speed at the t-th second is the weighted average of the speeds in the first 6 seconds and the last 6 seconds of the t-th second.

[0063] Step 102: Determine the kinematic segments according to the driving state of the vehicle and label each kinematic segment with a serial number in chronological order;

[0064] In specific implementation, the kinematic segment mainly refers to the driving process between two adjacent parking points. Therefore, the entire journey can be divided into multiple kinematic segments by taking the movement process of the vehicle from one idle state to the start of the next idle state as a kinematic segment, and each kinematic segment can be numbered in chronological order when dividing the kinematic segments, so as to facilitate the subsequent construction of the working conditions.

[0065] In an embodiment of the present invention, the driving state at least includes the idle state. After determining the kinematic segments according to the driving state of the vehicle, the following further includes:

[0066] In all kinematic segments, detect whether there is a kinematic segment in which the duration of the idle state is greater than a first threshold. If so, delete some segment contents in the first kinematic segment that are in the idle state according to a first preset rule, where the first kinematic segment is the kinematic segment in which the duration of the idle state is greater than the first threshold.

[0067] According to the regulations of the national standard document, the idle state refers to the state where the absolute value of the acceleration is less than 0.15 m / s 2 , and the speed is less than 0.5 km / h. In actual situations, the speed in most idle states is basically 0. Therefore, in practical applications, it is very likely that there is a situation of long-term parking in a certain kinematic segment, that is, the duration of the idle state is too long. The energy consumption in the idle state is generally very small, and a too long duration of the idle state will increase the duration of the kinematic segment. The total duration of the standard working condition is generally set to 1800 seconds, which means that too long a kinematic segment is not conducive to subsequent analysis. Therefore, for segments with too long a duration of the idle state, only part of the idle state can be retained. For example, only part of the idle state before the start of the kinematic segment can be retained. A dynamically adjustable threshold for the duration of the idle state can be set, and then detect whether there is a kinematic segment in which the duration of the idle state is greater than the threshold. If so, it is considered that the duration of the idle state is too long. For example, it can be stipulated that if a certain segment exceeds two-thirds of 1800 seconds (i.e., 1200 seconds), it can be judged that the duration of the idle state is too long. Then, part of the content in the kinematic segment in which the duration of the idle state is greater than the threshold can be deleted according to the rules preset by the user. Different road conditions have different retention quantities for the idle state. How many idle states need to be retained can be adjusted according to the results of multiple working condition cycles, but priority is given to making the proportion of the idle state in the constructed working condition cycle closer to the proportion in the original data.

[0068] Further, the method may further include:

[0069] Filter the kinematic segment according to a second preset rule.

[0070] For different road conditions, different analysis methods should be adopted. For example, in urban road conditions, due to the large number of traffic lights within the urban area, congestion during the morning and evening rush hours, or parking caused by other emergencies, the number of samples in the short-trip analysis method (i.e., the number of kinematic segments) will increase. The screening and filtering of samples should also vary according to different road conditions. Therefore, in practical applications, after dividing the kinematic segments, some kinematic segments have a short duration and cannot reflect the speed change characteristics. For example, the duration of some kinematic segments is less than 10 seconds or even shorter, some kinematic segments are only in an idling state, and some kinematic segments have an extremely long duration, even exceeding the time of 1800 seconds for the driving conditions of a general standard vehicle. For the above-mentioned several types of kinematic segments, they can be deleted or processed in other forms through filtering according to the rules preset by the user.

[0071] Step 103: Screen the processed kinematic segments, arrange the screened kinematic segments according to the serial numbers, and construct the vehicle driving conditions based on the arranged kinematic segments.

[0072] In practical applications, after completing the clustering, since the more similar the kinematic segments are, the more likely they are to be classified into the same category, the proportion of the total duration of the kinematic segments in each category to the total duration of all kinematic segments can be calculated. Then, multiply this proportion by 1800 seconds (i.e., the set duration of the vehicle driving conditions, which can also be called the total duration of the vehicle driving conditions) to obtain the total duration of each category in the vehicle driving conditions, and this can be used as one of the limiting conditions for selecting kinematic segments in the category. For example, when selecting kinematic segments in a category, while ensuring that the sum of the durations of the selected kinematic segments meets the proportion in 1800 seconds, try to select the kinematic segments closer to the clustering center. K-means clustering usually uses the Euclidean distance to calculate the distance between samples. The smaller the distance from the clustering center, the more typical the feature. Because a small distance from the clustering center indicates that the sample should be classified into the current category more, and the sample can better represent the characteristics of the category.

[0073] After the screening is completed, since one or more segments will be selected from each category after selecting the kinematic segments of the vehicle driving conditions, first, the kinematic segments within the category can be arranged in ascending order of the kinematic segment serial numbers. Then, compare the smallest kinematic segment serial numbers between different categories, and arrange the categories in ascending order of the smallest kinematic segment serial numbers. Construct the vehicle driving conditions based on the arranged kinematic segments. Through such a sorting method, the time sequence of the original data can be restored as much as possible, and putting the segments of each category together can better reflect the differences in inter-category characteristics.

[0074] Further, after the vehicle driving condition is constructed, the evaluation and verification of the vehicle driving condition can be carried out. The average vehicle speed, average smooth vehicle speed, average acceleration, average deceleration, acceleration ratio, deceleration ratio, constant speed ratio, and idle state ratio between the preliminarily filtered data and the vehicle driving condition data can be calculated. The calculation formula can be:

[0075] Relative error = |Original data feature - Constructed vehicle driving condition cycle feature| / |Original data feature| * 100%

[0076] Subsequently, the average value of the relative errors of these eigenvalue is calculated. After calculating multiple data, the average relative error of the technical solution provided by the present application can be minimized to within 5%, and the average relative error of the results of each vehicle driving condition can be basically controlled within 15%.

[0077] After verification by the chassis dynamometer test, the error of the fuel consumption per 100 kilometers of the vehicle driving condition constructed in the present application is approximately 0.97%; the error is effectively reduced, and the accuracy of the driving condition for monitoring indicators such as fuel consumption or emissions is improved; where the chassis dynamometer is a road simulator with a slope simulation function, which can not only perform power performance detection on the vehicle, but also measure multi-condition emission indicators and fuel consumption, etc., providing a solid experimental basis for studying the performance of the vehicle such as power performance, economy, comfort, and handling stability; therefore, the chassis dynamometer test can be used for relevant error verification.

[0078] In an embodiment of the present invention, before screening the kinematic segments and arranging the selected kinematic segments in the order, it further includes:

[0079] Determine characteristic parameters according to the vehicle driving data, and process the kinematic segments according to the characteristic parameters.

[0080] In specific implementation, since the finally constructed driving condition is a curve of speed and time, when selecting characteristic parameters, those related to speed and acceleration should be selected as much as possible; the following characteristic parameters can be selected: namely, the highest vehicle speed, average vehicle speed, highest acceleration, average acceleration, highest deceleration, average deceleration, acceleration ratio, deceleration ratio, constant speed ratio, idle state ratio, maximum slope, average slope, acceleration standard deviation, speed standard deviation, slope standard deviation;

[0081] After selecting the characteristic parameters, since the dimensions of the selected characteristic parameters are different, the data needs to be standardized before the next operation to eliminate the influence of dimensions. The Z-Score (Z-Score Standardization) standardization method can be used. The third-party library pandas in Python comes with a standardization function, and the specific formula is as follows:

[0082]

[0083] In the above formula, j refers to a certain characteristic parameter, and μ j is the expectation of all values of the characteristic parameter j, and σ j is the standard deviation of the eigenvalue j.

[0084] Furthermore, for the convenience of understanding the subsequent content, the following terms and operation processes that may be involved are now explained:

[0085] Correlation coefficient:

[0086] The correlation coefficient can represent the degree of correlation between different variables. Correlation coefficient analysis represents the degree of correlation between different variables through calculation, and its calculation formula is as follows:

[0087]

[0088] where cov(x,y) is the covariance between variable x and variable y, and σ x represents the standard deviation of variable x, and σ y represents the standard deviation of variable y. The calculation formula of the standard deviation is as follows:

[0089]

[0090] The calculation result ρ xy of the correlation coefficient ranges from [-1,1]. When the value is 1, it means that there is a perfect positive correlation between these two variables. When the value is -1, it means that there is a perfect negative correlation between these two variables. When the value is 0, it means that there is no correlation between these two variables. In addition, the closer the calculation result is to 0, the weaker the degree of correlation between the variables.

[0091] Principal component analysis:

[0092] Principal component analysis uses orthogonal transformation to convert the observed data represented by linearly correlated variables into data represented by a few linearly independent variables. The linearly independent variables are called principal components, thereby achieving the purpose of dimensionality reduction. Therefore, the number of principal components is less than the number of original variables.

[0093] In principal component analysis, first, the given data is normalized (or standardized) so that the mean of each variable in the data is 0 and the variance is 1. Then, an orthogonal transformation is performed on the data. The data originally represented by linearly correlated variables is transformed through the orthogonal transformation into data represented by a number of linearly independent new variables. The new variables are those for which the sum of the variances of the variables in the possible orthogonal transformations is the largest (the larger the variance, the more information is retained). The variance represents the amount of information on the new variables. The new variables are successively called the first principal component, the second principal component, etc. This is the basic idea of principal component analysis; it also includes the following theorem: Theorem 1: For any positive integer q, 1 ≤ q ≤ m, consider the orthogonal linear transformation

[0094] y = B T x

[0095] where y is a q-dimensional vector and B T is a q × m matrix. Let the covariance matrix of y be

[0096] ∑ y = B T ∑B

[0097] Then the set tr(∑ y ) reaches its maximum value when B = A y , where the matrix A q is composed of the first q columns of the orthogonal matrix A. q

[0098] Theorem 2: Consider the orthogonal transformation y = B T x, where B T is a p × m matrix. The definitions of A and ∑ y are the same as those in Theorem 1. Then tr(∑ y ) reaches its minimum value when B = A p , where the matrix A p is composed of the last p columns of the orthogonal matrix A.

[0099] Theorem 2 can be understood as follows: when discarding the last p columns of A, that is, discarding the last p principal components of the variable x, the information loss of the variance of the original variables is the least. The above two theorems can be used as the theoretical basis for selecting k principal components. The specific method for selecting k usually uses the variance contribution rate.

[0100] Theorem 3: The variance contribution rate of the k-th principal component y k is defined as the ratio of the variance of y k to the sum of all variances, denoted as η k

[0101]

[0102] The k principal components y1, y2,..., y k ​The cumulative variance contribution rate is defined as the ratio of the sum of k variances to the sum of all variances.

[0103]

[0104] Usually, k is taken such that the cumulative variance contribution rate reaches a specified percentage or more, usually 85%. The cumulative variance contribution rate reflects the proportion of information retained by the principal components, but it cannot reflect the proportion of information retained for a certain original variable x. i To reflect the proportion of information retained for a certain original variable x, the contribution rate of the k principal components y1, y2,..., yk to the original variable x is usually used. k For the original variable x i The contribution rate.

[0105] Silhouette coefficient:

[0106] The silhouette coefficient is an index for evaluating the clustering effect, with a value range of [-1, 1]. The closer it is to 1, the better the clustering effect. It can be used to select the number of clusters during clustering. By traversing the number of clusters, calculating the corresponding silhouette coefficients, and selecting the clustering model corresponding to the maximum silhouette coefficient. K-means clustering traverses the initial number of centers to find the clustering model that maximizes the silhouette coefficient.

[0107] The formula for calculating the silhouette coefficient of a single sample i:

[0108]

[0109] where a(i) is the average distance between sample i and other samples in the cluster to which it belongs. If there is only sample i in the cluster, then let s(i) = 0; b(i) is the minimum of the average distances between sample i and samples in other clusters.

[0110] The silhouette coefficient of the overall clustering is the average of the silhouette coefficients of all samples. Given that the total number of samples is N, the specific formula is as follows:

[0111]

[0112] K-means clustering:

[0113] Clustering is a data analysis method that divides a given set of samples into several "classes" or "clusters" based on the similarity or distance of the sample features. A class is a subset of the given sample set, and samples with similar features are grouped into the same class, while dissimilar samples are assigned to different classes; clustering belongs to unsupervised learning because it only classifies samples based on their similarity or distance, and the classes or clusters are not known in advance; common clustering methods include K-means clustering, hierarchical clustering, density clustering, etc.

[0114] A class is a subset of the given sample set, and samples with similar features are grouped into the same class, while dissimilar samples are assigned to different classes; clustering belongs to unsupervised learning because it only classifies samples based on their similarity or distance, and the classes or clusters are not known in advance; common clustering methods include K-means clustering, hierarchical clustering, density clustering, etc.

[0115] K-means clustering divides the sample set into K subsets, and each sample can only belong to one class. The specific algorithm process of K-means clustering is as follows; where D represents the sample set, C represents the cluster partition, and μ i is the mean vector of cluster C i and λ j represents the cluster label:

[0116]

[0117]

[0118] Furthermore, in the specific implementation, the seaborn library in Python can be used to draw the correlation matrix between the selected feature parameters. The calculation of the correlation matrix can be implemented using the built-in function corr() in the third-party library pandas in Python. Combining with the heatmap in the seaborn library can more intuitively observe the correlation between the feature parameters. For example, the correlation coefficients in the correlation matrix can be observed. After observation, it can be found that there is a large correlation between some feature parameters. Generally, the correlation coefficient |r| < 0.4 can indicate a weak correlation, 0.4 <= |r| <= 0.7 can indicate a medium correlation, and |r| > 0.7 can indicate a large correlation coefficient. It can be understood that if the absolute value of the correlation coefficient calculated by some feature parameters is greater than 0.7 or there are many between 0.4 and 0.7, then it means that there is a large correlation between these feature parameters and dimensionality reduction is required. Therefore, the principal component analysis method can be used to perform dimensionality reduction processing on the feature parameters with a large correlation coefficient. It is also possible to select feature parameters in any correlation degree interval for dimensionality reduction processing according to the corresponding requirements, and the specific division of the relevant interval can also be set by oneself. Principal Component Analysis (PCA) is a technique for analyzing and simplifying data sets. PCA is often used to reduce the dimensionality of a data set while retaining the features that contribute the most to the variance in the data set. The final number of principal components is determined based on the cumulative variance contribution rate, and the cumulative variance contribution rate is used to measure the explanatory ability of the principal components for the variability of the original variables. Usually, if the cumulative variance contribution rate reaches more than 85%, it is considered that selecting the first k principal components can well retain the information of the original samples.

[0119] Further, after dimensionality reduction by principal component analysis, the K-means algorithm can be used to cluster the PCA results of each kinematic segment, and segments with similar features are classified into one category; the value of K in K-means clustering is the number of clusters and needs to be determined in advance; in this application, the value of K can be determined by the silhouette coefficient, and the range of the silhouette coefficient is between [-1, 1]. The better the clustering effect, the closer the silhouette coefficient is to 1, and the worse the clustering effect, the closer the silhouette coefficient is to -1; the goal of clustering is to minimize the within-class difference and maximize the between-class difference. The closer the silhouette coefficient is to 1, the more obvious the differences between categories, and the more it can reflect the dissimilarity between each category.

[0120] In an embodiment of the present invention, the determining the characteristic parameters according to the vehicle driving data includes:

[0121] Calculating acceleration data based on the vehicle driving data, and determining the characteristic parameters according to the acceleration data and the vehicle driving data;

[0122] Calculating a correlation matrix according to the characteristic parameters, and obtaining a correlation coefficient according to the correlation matrix;

[0123] Determining a plurality of first characteristic parameters from all the characteristic parameters according to the correlation coefficient, where the first characteristic parameters are the characteristic parameters with larger correlation coefficients.

[0124] In practical applications, the seaborn library in Python can be used to draw the correlation matrix between the selected characteristic parameters. The calculation of the correlation matrix can be implemented using the built-in function corr() in the third-party library pandas in Python. Coupled with the heatmap in the seaborn library, the correlation between the characteristic parameters can be more intuitively observed. For example, the correlation coefficients in the correlation matrix can be observed. After observation, it can be found that there is a large correlation between some characteristic parameters. Generally, the correlation coefficient |r| < 0.4 can indicate a weak degree of correlation, 0.4 <= |r| <= 0.7 can indicate a medium degree of correlation, and |r| > 0.7 can indicate a large correlation coefficient. It can be understood that if the absolute value of the correlation coefficient calculated for some characteristic parameters is greater than 0.7 or there are more between 0.4 and 0.7, then it means that the correlation between these characteristic parameters is large and dimensionality reduction is required. Therefore, the principal component analysis method can be used to perform dimensionality reduction processing on the characteristic parameters with a large correlation coefficient, or characteristic parameters in any correlation degree interval can be selected for dimensionality reduction processing according to corresponding requirements. The specific division of the relevant interval can also be set by oneself. Principal component analysis is a technique for analyzing and simplifying data sets. PCA is often used to reduce the dimensionality of a data set while retaining the features that contribute the most to the variance in the data set. The final number of principal components is determined based on the cumulative variance contribution rate, and the cumulative variance contribution rate is used to measure the explanatory ability of the principal components for the variability of the original variables. Usually, if the cumulative variance contribution rate reaches more than 85%, it is considered that selecting the first k principal components can well retain the information of the original sample.

[0125] In an embodiment of the present invention, processing the kinematic segment according to the characteristic parameter includes:

[0126] Performing dimensionality reduction processing on the first characteristic parameter to obtain a dimensionality reduction processing result;

[0127] Performing clustering processing on all the kinematic segments according to the dimensionality reduction processing result to obtain multiple categories of kinematic segments.

[0128] In specific implementation, the seaborn library in Python can be used to draw the correlation matrix between the selected characteristic parameters. The calculation of the correlation matrix can be implemented using the built-in function corr() in the third-party library pandas in Python. Combining with the heatmap in the seaborn library can more intuitively observe the correlation between the characteristic parameters. For example, the correlation coefficients in the correlation matrix can be observed. After observation, it can be found that there is a large correlation between some characteristic parameters. Generally, when the correlation coefficient |r| < 0.4, it can indicate a weak correlation degree; when 0.4 <= |r| <= 0.7, it can indicate a medium correlation degree; when |r| > 0.7, it can indicate a strong correlation degree. It can be understood that if the absolute value of the correlation coefficient calculated for some characteristic parameters is greater than 0.7 or there are more between 0.4 and 0.7, then it means that the correlation between these characteristic parameters is large and dimensionality reduction is required. Therefore, the principal component analysis method can be used to perform dimensionality reduction processing on the characteristic parameters with a strong correlation degree, or the characteristic parameters in any correlation degree interval can be selected for dimensionality reduction processing according to corresponding requirements. The specific division of the relevant interval can also be set by oneself. Principal component analysis is a technique for analyzing and simplifying data sets. PCA is often used to reduce the dimensionality of a data set while retaining the features that contribute the most to the variance in the data set. The final number of principal components is determined based on the cumulative variance contribution rate, and the cumulative variance contribution rate is used to measure the explanatory ability of the principal components for the variability of the original variables. Usually, if the cumulative variance contribution rate reaches more than 85%, it is considered that selecting the first k principal components can well retain the information of the original sample.

[0129] Furthermore, after dimensionality reduction by principal component analysis, the K-means algorithm can be used to cluster the PCA results of each kinematic segment to divide the segments with similar features into one category to determine the type of kinematic segment. The K value in K-means clustering is the number of clusters and needs to be determined in advance. In this application, the K value can be determined by the silhouette coefficient. The range of the silhouette coefficient is between [-1, 1]. The better the clustering effect, the closer the silhouette coefficient is to 1; the worse the clustering effect, the closer the silhouette coefficient is to -1. The goal of clustering is to make the within-class difference small and the between-class difference as large as possible. The closer the silhouette coefficient is to 1, the more obvious the differences between different categories are, and the more it can reflect the dissimilarity between each category.

[0130] In an embodiment of the present invention, the screening of the kinematic segments and arranging the screened kinematic segments in the order includes:

[0131] Calculate the proportion of the duration of each kinematic segment in each type of kinematic segment in the set duration of the vehicle driving condition.

[0132] According to the ratio of the duration of each kinematic segment to the set duration, each kinematic segment in various types of kinematic segments is initially screened to obtain an initial screening result;

[0133] For the initial screening result, according to the distance between each kinematic segment and the cluster center in various types of kinematic segments, the initial screening result is secondarily screened to obtain a secondary screening result;

[0134] For the secondary screening result, according to the magnitude of the serial numbers of each kinematic segment in various types of kinematic segments, the secondary screening result is tertially screened to obtain a tertiary screening result;

[0135] Arrange each kinematic segment in the tertiary screening result in ascending order of the serial number.

[0136] In practical applications, after clustering is completed, since the more similar the kinematic segments are, the more likely they are to be classified into the same category. Therefore, the ratio of the total duration of the kinematic segments in each category to the total duration of all kinematic segments can be calculated, and then this ratio is multiplied by 1800 seconds (i.e., the set duration of the vehicle driving condition, which can also be called the total duration of the vehicle driving condition) to obtain the total duration of each category in the vehicle driving condition, and this can be used as one of the limiting conditions for selecting kinematic segments in the category; for example, when selecting kinematic segments in the category, while ensuring that the sum of the durations of the selected kinematic segments meets the proportion in 1800 seconds, try to select kinematic segments closer to the cluster center. K-means clustering usually uses the Euclidean distance to calculate the distance between samples. The smaller the distance from the cluster center, the more typical the feature. Because the small distance from the cluster center indicates that the sample should be classified into the current category more, and the sample can better represent the features of the category. Therefore, when actually screening, the initial screening can be first performed according to the ratio of the duration of each kinematic segment to the set duration to obtain the initial screening result. Then, based on the initial screening result, the second screening can be performed according to the distance between each kinematic segment and the cluster center in various types of kinematic segments. Specifically, kinematic segments with the smallest distance from the cluster center can be preferentially selected, or kinematic segments with the largest or other distance magnitudes can be selected according to the corresponding actual requirements;

[0137] After the second screening is completed, since one or more segments will be screened out in each category after selecting the kinematic segments of the vehicle driving conditions, the third screening can first be carried out within the category in ascending order of the kinematic segment numbers. Then, according to the results of the third screening, compare the smallest kinematic segment numbers among different categories, arrange the categories in ascending order of the smallest kinematic segment numbers, and construct the vehicle driving conditions based on the arranged kinematic segments. Through such a sorting method, the time sequence of the original data generation can be restored as much as possible, and putting the segments of each category together can better reflect the differences in inter-class characteristics.

[0138] Of course, in addition to selecting the smallest kinematic segment number, the largest or medium-sized kinematic segments can also be selected according to actual needs.

[0139] On the basis of the above content, for the convenience of understanding and more detailed description of the technical solution of the present invention, refer to Figure 2 , which shows the implementation step diagram of the vehicle driving condition construction method provided by the present invention. Its working steps can be summarized as: (1) data cleaning, (2) denoising and smoothing processing of speed and slope, (3) calculating acceleration and dividing driving states, (4) kinematic segment division, (5) kinematic segment filtering, (6) feature parameter selection and calculation, (7) feature parameter standardization, (8) dimensionality reduction by principal component analysis, (9) clustering by K-Means algorithm, (10) selection of kinematic segments of the synthetic condition (the synthetic condition is the vehicle driving condition constructed in the above content of this application), (11) sorting of kinematic segments of the synthetic condition, (12) evaluation and verification of the synthetic condition; it can be seen that the condition construction method provided by this application has good applicability to urban complex road conditions and innovatively considers the influence of slope. The construction of vehicle driving conditions can, on the one hand, reduce the labor cost and time cost invested in chassis dynamometer tests in some departments of the automotive manufacturing industry. On the other hand, vehicle driving conditions provide a basis for optimizing various vehicle performance indicators, such as reducing energy consumption and exhaust emissions in autonomous driving.

[0140] Refer to Figure 3 , which shows the step flow chart of another vehicle driving condition construction method provided by an embodiment of the present invention, and specifically may include the following steps:

[0141] Step 301, collect vehicle driving data, preprocess the vehicle driving data and determine the driving state of the vehicle according to the vehicle driving data; wherein, the vehicle driving data at least includes speed data and slope data;

[0142] In a specific implementation, vehicle driving data including speed data and slope data can be collected, and based on the above collected data, the acceleration of each second can be calculated according to the acceleration calculation formula in the national standard document "China's Automobile Driving Conditions", and the driving state of each second can be divided according to the division standards for acceleration, deceleration, constant speed and idle states in the national standard document.

[0143] There may be missing or duplicate data in the original driving data, and there may also be abnormal values ​​in the collected data. Therefore, the data needs to be preprocessed. During preprocessing, the duplicate data can be directly deleted. For missing data, when there are many missing data, the kinematic segments where the missing values ​​are located can be deleted, and when there are fewer missing data, interpolation can be performed. In specific processing, because the collected data is often a time series, it is not reasonable and accurate to use other indicators such as the average or median to fill in, so linear interpolation can be used. For obviously abnormal data (such as high speed, large changes in slope before and after two seconds, etc.), delete them and then perform linear interpolation to ensure time continuity.

[0144] Step 302, determining kinematic segments according to the driving state of the vehicle and marking a sequence number for each kinematic segment in chronological order;

[0145] In practical applications, a kinematic segment mainly refers to the driving process between two adjacent parking points. Therefore, the entire journey can be divided into multiple kinematic segments by taking the vehicle's movement process from one idle state to the next idle state as a kinematic segment. When dividing the kinematic segments, each kinematic segment can be labeled with a serial number in chronological order to facilitate the subsequent construction of working conditions.

[0146] Step 303, determining characteristic parameters according to the vehicle driving data, standardizing the characteristic parameters and processing the kinematic segment according to the characteristic parameters;

[0147] In the specific implementation, since the working condition finally constructed is a curve of speed and time, when selecting characteristic parameters, try to choose those related to speed and acceleration; the following characteristic parameters can be selected: maximum vehicle speed, average vehicle speed, maximum acceleration, average acceleration, maximum deceleration, average deceleration, acceleration ratio, deceleration ratio, uniform speed ratio, idling ratio, maximum slope, average slope, acceleration standard deviation, speed standard deviation, slope standard deviation;

[0148] After selecting the characteristic parameters, since the dimensions of the selected characteristic parameters are different, it is necessary to standardize the data before proceeding to the next step to eliminate the influence of dimensions. The Z-Score standardization method can be adopted. The third-party library pandas in Python comes with a standardization function, and the specific formula is as follows:

[0149]

[0150] In the above formula, j refers to a certain characteristic parameter, and μ j is the expectation of all values of the characteristic parameter j, and σ j is the standard deviation of the eigenvalue j.

[0151] Furthermore, in the specific implementation, the seaborn library in Python can be used to draw the correlation matrix between the selected characteristic parameters. The calculation of the correlation matrix can be implemented using the built-in function corr() in the third-party library pandas of Python. Combining with the heatmap in the seaborn library can more intuitively observe the correlation between the characteristic parameters; it can be found that there is a large correlation between some characteristic parameters. Generally, the correlation coefficient |r| < 0.4 indicates a weak correlation, 0.4 <= |r| <= 0.7 indicates a medium correlation, and |r| > 0.7 indicates a large correlation coefficient. Therefore, the principal component analysis method can be used to reduce the dimension of the selected characteristic parameters; principal component analysis is a technique for analyzing and simplifying data sets; PCA is often used to reduce the dimension of the data set while retaining the characteristics that contribute the most to the variance in the data set; the final number of principal components is determined based on the cumulative variance contribution rate, and the cumulative variance contribution rate is used to measure the explanatory ability of the principal components for the variability of the original variables; usually, if the cumulative variance contribution rate reaches more than 85%, it is considered that selecting the first k principal components can well retain the information of the original sample.

[0152] Furthermore, after dimensionality reduction by principal component analysis, the K-means algorithm can be used to cluster the PCA results of each kinematic segment, and the segments with similar characteristics are divided into one category; the K value in K-means clustering is the number of clusters and needs to be determined in advance; in this application, the K value can be determined by the silhouette coefficient. The range of the silhouette coefficient is between [-1, 1]. The closer the silhouette coefficient is to 1, the better the clustering effect, and the closer the silhouette coefficient is to -1, the worse the clustering effect; the goal of clustering is to make the within-class difference small and the between-class difference as large as possible. The closer the silhouette coefficient is to 1, the more obvious the differences between categories are, and the more it can reflect the dissimilarity between each category.

[0153] Step 304: Screen the processed kinematic segments, arrange the screened kinematic segments according to the serial numbers, and construct a vehicle driving condition based on the arranged kinematic segments.

[0154] In practical applications, after clustering is completed, since the more similar kinematic segments are more likely to be classified into the same category, the proportion of the total duration of the kinematic segments in each category to the total duration of all kinematic segments can be calculated. Then, multiply this proportion by 1800 seconds (i.e., the set duration of the vehicle driving condition, which can also be called the total duration of the vehicle driving condition) to obtain the total duration of each category in the vehicle driving condition, and this can be used as one of the limiting conditions for selecting kinematic segments in the category. For example, when selecting kinematic segments in a category, while ensuring that the sum of the durations of the selected kinematic segments meets the proportion in 1800 seconds, try to select kinematic segments closer to the clustering center. K-means clustering usually uses the Euclidean distance to calculate the distance between samples. The smaller the distance from the clustering center, the more typical the feature, because a small distance from the clustering center indicates that the sample should be classified into the current category more, and the sample can better represent the features of the category.

[0155] After screening is completed, since one or more segments will be selected from each category after selecting the kinematic segments of the vehicle driving condition, first, the kinematic segments within the category can be arranged in ascending order of the kinematic segment serial numbers, and then the smallest kinematic segment serial numbers between categories can be compared. The categories are arranged in ascending order of the smallest kinematic segment serial numbers, and a vehicle driving condition is constructed based on the arranged kinematic segments. Through this sorting method, the time sequence generated by the original data can be restored as much as possible, and putting the segments of each category together can better reflect the differences in inter-category features.

[0156] Further, after the vehicle driving condition is constructed, the evaluation and verification of the vehicle driving condition can be carried out. The relative errors between the average vehicle speed, average smooth vehicle speed, average acceleration, average deceleration, acceleration ratio, deceleration ratio, uniform speed ratio, and idle state ratio of the preliminarily filtered data and the vehicle driving condition data can be calculated. The calculation formula can be:

[0157] Relative error = |Original data feature - Constructed vehicle driving condition cycle feature| / |Original data feature| * 100%

[0158] Then, the average value of the relative errors of these characteristic values is calculated. After calculating multiple data sets, the average relative error of the technical solution provided by this application can be minimized to within 5%, and the average relative error of the results of each vehicle driving condition can basically be controlled within 15%.

[0159] After verification by the chassis dynamometer test, the error of the energy consumption per 100 kilometers of the vehicle driving cycle constructed in this application is approximately 0.97%; the error is effectively reduced, and the accuracy of the driving cycle for monitoring indicators such as energy consumption or emissions is improved; among them, the chassis dynamometer is a road simulator with a slope simulation function, which can not only perform power performance tests on automobiles, but also measure multi-condition emission indicators and fuel consumption, etc., providing a solid experimental basis for studying the power performance, economy, comfort and handling stability of automobiles; therefore, the chassis dynamometer test can be used for relevant error verification.

[0160] It should be noted that for the method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the embodiments of the present invention are not limited by the described action sequence, because according to the embodiments of the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0161] Referring to Figure 4 , a structural block diagram of an apparatus for a vehicle driving cycle construction method provided by an embodiment of the present invention is shown, which may specifically include the following modules:

[0162] A driving state determination module 401, configured to collect vehicle driving data and determine the driving state of the vehicle according to the vehicle driving data;

[0163] A kinematic segment determination module 402, configured to determine kinematic segments according to the driving state of the vehicle and number each of the kinematic segments in chronological order;

[0164] A vehicle driving cycle construction module 403, configured to screen the processed kinematic segments and arrange the screened kinematic segments according to the serial numbers, and construct a vehicle driving cycle according to the arranged kinematic segments.

[0165] In another embodiment of the present invention, the driving state determination module 401 further includes:

[0166] A preprocessing sub-module, configured to preprocess the vehicle driving data; wherein, the preprocessing at least includes smoothing the speed data and the slope data, and when smoothing the speed data, the smoothing window used is larger than the smoothing window used when smoothing the slope data.

[0167] In another embodiment of the present invention, the apparatus further includes:

[0168] A kinematic segment processing module, configured to determine characteristic parameters according to the vehicle driving data and process the kinematic segments according to the characteristic parameters.

[0169] In another embodiment of the present invention, the driving state at least includes an idle state, and the driving state determination module 401 further includes:

[0170] An idle state detection sub-module, configured to detect, among all kinematic segments, whether there is a kinematic segment in which the duration of the idle state is greater than a first threshold. If so, delete part of the segment content in the first kinematic segment that is in the idle state according to a first preset rule, where the first kinematic segment is the kinematic segment in which the duration of the idle state is greater than the first threshold.

[0171] In another embodiment of the present invention, the kinematic segment processing module includes:

[0172] A characteristic parameter determination sub-module, configured to calculate acceleration data according to the vehicle driving data and determine energy consumption according to the acceleration data and the vehicle driving data;

[0173] A correlation coefficient calculation sub-module, configured to calculate the correlation coefficient between each characteristic parameter and the energy consumption;

[0174] A first characteristic parameter determination sub-module, configured to determine a plurality of first characteristic parameters from all the characteristic parameters according to the correlation coefficient, where the first characteristic parameter is the characteristic parameter with a larger correlation coefficient.

[0175] In another embodiment of the present invention, the kinematic segment processing module further includes:

[0176] A dimensionality reduction processing result acquisition sub-module, configured to determine a correlation matrix according to a plurality of the first characteristic parameters and perform dimensionality reduction processing on the first characteristic parameters to obtain a dimensionality reduction processing result;

[0177] A multi-class kinematic segment acquisition sub-module, configured to perform clustering processing on all the kinematic segments according to the correlation matrix and the dimensionality reduction processing result to obtain multi-class kinematic segments.

[0178] In another embodiment of the present invention, the vehicle driving condition construction module 403 includes:

[0179] A kinematic segment duration ratio calculation sub-module, configured to calculate the ratio of the duration of each kinematic segment in each type of kinematic segment to the set duration of the vehicle driving condition;

[0180] The initial screening result acquisition sub-module is used to perform an initial screening on each kinematic segment in various types of kinematic segments according to the ratio of the duration of each kinematic segment to the set duration, and obtain the initial screening result;

[0181] The secondary screening result acquisition sub-module is used to perform a secondary screening on the initial screening result according to the distance between each kinematic segment and the clustering center in various types of kinematic segments for the initial screening result, and obtain the secondary screening result;

[0182] The tertiary screening result acquisition sub-module is used to perform a tertiary screening on the secondary screening result according to the magnitude of the serial numbers of each kinematic segment in various types of kinematic segments for the secondary screening result, and obtain the tertiary screening result;

[0183] The kinematic segment arrangement sub-module is used to arrange each kinematic segment in the tertiary screening result in ascending order of the serial number.

[0184] An embodiment of the present invention further provides an electronic device, which may include a processor, a memory, and a computer program stored on the memory and capable of running on the processor. When the computer program is executed by the processor, the above vehicle driving condition construction method is implemented.

[0185] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by the processor, the above vehicle driving condition construction method is implemented.

[0186] For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple. For the relevant parts, refer to the partial description of the method embodiment.

[0187] Each embodiment in this specification is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other.

[0188] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a device, or a computer program product. Therefore, the embodiments of the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0189] Embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal devices generate means for implementing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0190] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0191] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable terminal device provide steps for implementing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0192] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they know the basic creative concept. Therefore, the appended claims are intended to be construed as including the preferred embodiments and all changes and modifications falling within the scope of the present invention.

[0193] Finally, it should also be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or terminal device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or terminal device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or terminal device comprising the above elements.

[0194] The above provides a detailed introduction to a method, device and readable storage medium for constructing a vehicle driving condition. In this text, specific examples are used to illustrate the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for constructing a vehicle driving condition, characterized in that, The method includes: Collecting vehicle driving data and determining the driving state of the vehicle according to the vehicle driving data; Determining kinematic segments according to the driving state of the vehicle and numbering each kinematic segment in chronological order; Screening the kinematic segments and arranging the screened kinematic segments according to the serial numbers, and constructing a vehicle driving condition according to the arranged kinematic segments.

2. The method according to claim 1, characterized in that, At least speed data and slope data are included in the vehicle driving data. Before determining the driving state of the vehicle according to the vehicle driving data, it further includes: Preprocessing the vehicle driving data; wherein, the preprocessing at least includes smoothing the speed data and the slope data, and when smoothing the speed data, the smoothing window used is larger than the smoothing window used when smoothing the slope data.

3. The method according to claim 1, characterized in that, Before screening the kinematic segments and arranging the screened kinematic segments according to the serial numbers, it further includes: Determining characteristic parameters according to the vehicle driving data and processing the kinematic segments according to the characteristic parameters.

4. The method according to claim 1, characterized in that, The driving state at least includes an idle state. After determining the kinematic segments according to the driving state of the vehicle, it further includes: In all kinematic segments, detecting whether there is a kinematic segment in which the duration of the idle state is greater than a first threshold. If so, deleting part of the segment content in the first kinematic segment that is in the idle state according to a first preset rule, where the first kinematic segment is the kinematic segment in which the duration of the idle state is greater than the first threshold.

5. The method according to claim 3, characterized in that, Determining the characteristic parameters according to the vehicle driving data includes: Calculating acceleration data according to the vehicle driving data, and determining the characteristic parameters according to the acceleration data and the vehicle driving data; Calculating a correlation matrix according to the characteristic parameters and obtaining a correlation coefficient according to the correlation matrix; Determining a plurality of first characteristic parameters from all the characteristic parameters according to the correlation coefficient, where the first characteristic parameters are the characteristic parameters with larger correlation coefficients.

6. The method according to claim 5, characterized in that, Processing the kinematic segments according to the characteristic parameters includes: Performing dimensionality reduction processing on the first characteristic parameters to obtain a dimensionality reduction processing result; Performing clustering processing on all the kinematic segments according to the dimensionality reduction processing result to obtain multiple types of kinematic segments.

7. The method according to claim 6, characterized in that, Screening the kinematic segments and arranging the screened kinematic segments according to the serial numbers includes: Calculating the proportion of the duration of each kinematic segment in each type of kinematic segment in the set duration of the vehicle driving condition; Performing a primary screening on each kinematic segment in each type of kinematic segment according to the proportion of the duration of each kinematic segment in the set duration to obtain a primary screening result; For the primary screening result, performing a secondary screening on the primary screening result according to the distance between each kinematic segment in each type of kinematic segment and the clustering center to obtain a secondary screening result; For the secondary screening results, perform a tertiary screening on the secondary screening results according to the magnitudes of the sequence numbers of the kinematic segments in each type of kinematic segment to obtain the tertiary screening results; Arrange the kinematic segments in the tertiary screening results in ascending order of their sequence numbers.

8. A device for constructing a vehicle driving condition, characterized in that, The device includes: A driving state determination module, configured to collect vehicle driving data and determine the driving state of the vehicle according to the vehicle driving data; A kinematic segment determination module, configured to determine kinematic segments according to the driving state of the vehicle and label a sequence number for each of the kinematic segments in chronological order; A vehicle driving condition construction module, configured to screen the processed kinematic segments, arrange the screened kinematic segments in the order of the sequence numbers, and construct a vehicle driving condition according to the arranged kinematic segments.

9. An electronic device, characterized in that, Comprising a processor, a memory, and a computer program stored on the memory and capable of running on the processor, where when the computer program is executed by the processor, it implements the vehicle driving condition construction method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, it implements the vehicle driving condition construction method according to any one of claims 1 to 7.