User long and short distance and daily journey identification system and method based on K-Means clustering

Abstract

The invention discloses a user long and short distance and daily travel identification system based on K-Means clustering, and the system comprises a data preprocessing module which obtains the mileage data of a single travel of a vehicle, filters the long-distance travel data to obtain a mileage data set of a short distance travel and a daily travel, and carries out the preprocessing to obtain a preprocessed mileage data set; a clustering analysis module randomly selects mileage data from the preprocessed mileage data set to serve as an initial clustering cluster center, clustering analysis is carried out, continuous iteration is carried out on the clustering cluster center, and a short-distance travel cluster center mileage value and a standard deviation are obtained; and the travel identification module obtains a short-distance travel mileage range value and a daily travel mileage range value according to the short-distance travel cluster center mileage value, the standard deviation and the long-distance travel threshold value, and identifies the long-distance travel, the short-distance travel and the daily travel of the user through the short-distance travel mileage range value, the daily travel mileage range value and the long-distance travel threshold value. The endurance matching degree and the user experience of the new energy electric vehicle are improved.

Description

Technical Field

[0001] This invention relates to the field of new energy electric vehicle user trip data analysis technology, specifically to a user long-distance and short-distance and daily trip identification system and method based on K-Means clustering. Background Technology

[0002] In recent years, driven by the deep integration of big data, cloud computing, and vehicle-to-everything (V2X) technologies, the automotive industry has accelerated its transformation towards intelligence and connectivity, gradually building a new industrial ecosystem centered on data. Vehicles generate real-time driving data through multi-dimensional sensors, providing crucial support for deeply understanding user travel characteristics and optimizing vehicle range performance. Among these, refined feature analysis of user travel mileage has become a core direction for technological breakthroughs in the industry. Precise segmentation of long and short distances, daily trips, high-frequency travel, and multi-scenario travel patterns can effectively deconstruct the complex relationship between user driving habits and energy consumption patterns. However, current research on segmenting user travel distances and daily trips relies on human experience, neglecting the differences in travel habits among different owners of new energy electric vehicles, leading to distorted individual travel pattern portrayals; it fails to exclude the interference of extreme values ​​from long-distance trips on clustering accuracy; and it lacks a data-driven dynamic calibration mechanism based on statistical mean or manual observation of preset boundary points, making it unable to adapt to changes in vehicle ownership or urban migration scenarios. Summary of the Invention

[0003] The purpose of this invention is to provide a user long-distance and short-distance and daily trip identification system and method based on K-Means clustering. This invention improves the range matching degree and user experience of new energy electric vehicles.

[0004] To achieve this objective, the present invention provides a user long / short distance and daily trip identification system based on K-Means clustering, comprising: The data preprocessing module is used to acquire the mileage data of all single trips of vehicles within a set time period. By using a preset long-distance travel threshold, long-distance travel data is filtered out from the mileage data of all single trips of vehicles within the set time period to obtain the mileage dataset of short-distance trips and daily trips. The mileage dataset of short-distance trips and daily trips is preprocessed to obtain the preprocessed mileage dataset. The cluster analysis module is used to randomly select a set number of mileage data from the preprocessed mileage dataset as the initial cluster centers. The preprocessed mileage dataset is then subjected to cluster analysis using the initial cluster centers. The cluster centers are iterated continuously during the cluster analysis to obtain the mileage value of the short-distance travel cluster center and the standard deviation of the short-distance travel cluster. The trip identification module is used to obtain the short-distance trip mileage range value and the daily trip mileage range value based on the center mileage value of the short-distance trip cluster, the standard deviation of the short-distance trip cluster, and the preset long-distance trip threshold. The module then identifies the user's long-distance, short-distance, and daily trips using the short-distance trip mileage range value, the daily trip mileage range value, and the preset long-distance trip threshold.

[0005] Preferably, the range demand analysis module is used to calculate the individualized daily range demand of the corresponding new energy electric vehicle based on the center mileage value of the short-distance travel cluster, the center mileage value of the daily travel cluster, the preset long-distance travel threshold and the corresponding travel frequency, so as to formulate the corresponding battery specifications for new energy electric vehicle users. The daily travel cluster center mileage value is obtained by performing cluster analysis on the preprocessed mileage dataset using the initial cluster center, and by continuously iterating the cluster center during the cluster analysis.

[0006] Preferably, the process involves acquiring mileage data for all single trips of vehicles within a set time period, filtering out long-distance trip data from this data using a preset long-distance travel threshold to obtain mileage datasets for short-distance and daily trips, and then preprocessing these datasets to obtain the preprocessed mileage datasets. Monthly mileage datasets for individual vehicles are obtained through a vehicle monitoring system. ,in x i This represents the mileage of a single trip by a single vehicle. like x i Greater than the preset long-distance travel threshold m ,Will x i Labeling trips as long-distance journeys and removing the monthly trip mileage dataset D for single vehicles, we obtain mileage datasets for short-distance and daily trips. ; Mileage datasets for short-distance and daily travel Perform Z-Score standardization.

[0007] Preferred mileage datasets for short-distance and daily travel. The specific process of Z-Score standardization is as follows: Mileage datasets for short-distance and daily travel The formula for calculating Z-Score standardization is as follows: Where, x j This represents the mileage value of a single vehicle's short-distance trip or daily travel journey. The mean of the mileage dataset D1, representing short-distance and daily trips; The standard deviation of the mileage dataset D1 represents the distances of short-distance and daily trips; This represents the standardized mileage value for a single short-distance trip or daily commut by a single vehicle. Obtain the preprocessed mileage dataset .

[0008] Preferably, a set number of mileage data points are randomly selected from the preprocessed mileage dataset as initial cluster centers. K-Means clustering analysis is then performed on the preprocessed mileage dataset using these initial cluster centers. The K-Means clustering analysis iterates through the cluster centers to obtain the mileage values ​​for short-distance trip cluster centers, the standard deviation of short-distance trip cluster centers, and the mileage values ​​for daily trip cluster centers. The specific process is as follows: Preprocessed mileage dataset Divided into K clusters, due to the preprocessed mileage dataset Including mileage data from short-distance trips and daily trips, we get K=2, which means we divide the data into cluster 1 and cluster 2. From the preprocessed mileage dataset Two initial cluster centers were randomly selected from the data. , ; Calculate the preprocessed mileage dataset Each mileage data point Respectively with the current cluster center and Euclidean distance ( k ∈{1,2}), this mileage data point Assign to the cluster corresponding to the current cluster center with the closest Euclidean distance. ( k ∈{1,2}), minimizing the sum of squared Euclidean distances J for all mileage data points within the same cluster: Recalculate the current cluster centers corresponding clusters The mean of all mileage data points within the cluster is used as the new cluster center. The calculation formula is as follows: Recalculate the current cluster centers corresponding clusters The mean of all mileage data points within the cluster is used as the new cluster center. The calculation formula is as follows: If satisfied If the condition is met, the iteration stops, where ε represents the convergence threshold; otherwise, the iteration continues until the condition is met. ; Finally, the cluster center values ​​of cluster 1 were obtained. Cluster center value of cluster 2 ,set up < ,but This is the center mileage value for short-distance travel clusters. This refers to the mileage value at the center of the daily travel cluster. At this time, the standard deviation of the short-distance travel cluster The calculation formula is as follows: Where n is the number of samples of mileage data points in cluster 1. This represents the value of the mileage data point in cluster 1. This represents the sample average value of the mileage data points in cluster 1.

[0009] Preferably, the short-distance travel mileage range value x 短 For: 0 < x 短 ≤ The daily travel mileage range value x 日 for: <x 日 ≤m; then the range of long-distance travel distances is x. 长 For: x 长 >m; range of short-distance travel distances x 短 Daily travel mileage range value x 日 and the range value of long-distance travel mileage x 长 The system identifies users' long-distance, short-distance, and daily trips, and obtains the number of trips for each type of user in the monthly mileage of a single vehicle.

[0010] Preferably, the specific process of calculating the individualized daily range requirement of the corresponding new energy electric vehicle based on the center mileage value of the short-distance travel cluster, the center mileage value of the daily travel cluster, the preset long-distance travel threshold, and the corresponding travel frequency, and thus formulating the corresponding battery specifications for new energy electric vehicle users, is as follows: Based on the short-distance travel cluster center mileage value, the daily travel cluster center mileage value, the preset long-distance travel threshold, and the number of long-distance, short-distance, and daily trips for each user in the corresponding monthly mileage of a single vehicle, the calculation formula for the individualized daily average range requirement of new energy electric vehicles is as follows: in, f 1 For the average daily frequency of short-distance trips, f2 For daily travel frequency, f 3 For the average daily frequency of long-distance travel, This is the center mileage value for short-distance travel clusters. This refers to the mileage value at the center of the daily travel cluster. The battery capacity requirement is obtained by multiplying the daily average range requirement by the number of days in a week by the weekly range redundancy coefficient, thereby enabling the formulation of corresponding battery specifications for new energy electric vehicle users.

[0011] A method for identifying user long and short trips based on K-Means clustering includes the following steps: Acquire all vehicle single trip mileage data within a set time period. Filter out long-distance trip data from all vehicle single trip mileage data within the set time period using a preset long-distance travel threshold to obtain short-distance and daily trip mileage datasets. Preprocess the short-distance and daily trip mileage datasets to obtain preprocessed mileage datasets. A set number of mileage data points are randomly selected from the preprocessed mileage dataset as initial cluster centers. Cluster analysis is performed on the preprocessed mileage dataset using the initial cluster centers, and the cluster centers are iterated continuously during the cluster analysis to obtain the mileage value of the short-distance travel cluster center and the standard deviation of the short-distance travel cluster. The short-distance travel mileage range and daily travel mileage range are obtained based on the center mileage value of the short-distance travel cluster, the standard deviation of the short-distance travel cluster, and the preset long-distance travel threshold. The user's long-distance, short-distance, and daily trips are identified by the short-distance travel mileage range, daily travel mileage range, and the preset long-distance travel threshold.

[0012] A computer program product includes a computer program that, when executed by a processor, implements the steps of the above-described method.

[0013] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method.

[0014] The beneficial effects of this invention are: This invention improves recognition accuracy through precise travel pattern analysis, avoiding subjective errors caused by manually set thresholds. The recognition results are more in line with the actual driving habits of new energy electric vehicle users, providing data-driven decision-making basis for the range design of new energy electric vehicles. Based on the cluster analysis results of each new energy electric vehicle, the core distribution range of individual travel distances of new energy electric vehicle users can be extracted, directly quantifying the actual range needs of new energy electric vehicle users. It can aggregate the cluster results of new energy electric vehicle users in the same city to generate city-level range design benchmarks. Automakers can formulate differentiated battery specifications according to regional differences, realizing personalized customization of battery specifications. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the structure of the present invention; Figure 2 This is a flowchart of the present invention. Detailed Implementation

[0016] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments: Example 1 A user long / short distance and daily itinerary identification system based on K-Means clustering, such as Figure 1 As shown, it includes: The data preprocessing module is used to acquire all vehicle single trip mileage data within a set time period (set time is one month). By using a preset long-distance travel threshold (preset long-distance travel threshold is 100km), long-distance travel data is filtered out from all vehicle single trip mileage data within the set time period to obtain short-distance and daily travel mileage datasets. The short-distance and daily travel mileage datasets are preprocessed (Z-Score normalization preprocessing) to obtain the preprocessed mileage dataset. This design avoids interference from extreme long-distance trips to subsequent clustering by filtering out extreme long-distance trips. The clustering analysis module is used to randomly select a set number (2 in this embodiment) of mileage data from the preprocessed mileage dataset as the initial cluster centers. The preprocessed mileage dataset is then subjected to clustering analysis using the initial cluster centers. The cluster centers are iterated continuously during the clustering analysis to obtain the mileage value of the short-distance travel cluster center and the standard deviation of the short-distance travel cluster. This design outputs two cluster centers through K-Means clustering analysis. The output results are clear and easy for new energy electric vehicle business personnel to understand and use. The obtained results of the mileage value of the short-distance travel cluster center and the standard deviation of the short-distance travel cluster are stable and less affected by outliers. The trip identification module is used to obtain the short-distance trip mileage range value and the daily trip mileage range value based on the center mileage value of the short-distance trip cluster, the standard deviation of the short-distance trip cluster, and the preset long-distance trip threshold. By using the short-distance trip mileage range value, the daily trip mileage range value, and the preset long-distance trip threshold, the module identifies the user's long-distance, short-distance, and daily trips. This design divides the trips of new energy electric vehicle users into three non-overlapping and continuous mileage intervals: short-distance, daily, and long-distance, with clear classification boundaries. It can more accurately reflect the trip distribution of individual new energy electric vehicle users and meet personalized service needs.

[0017] In the above technical solution, the range demand analysis module is used to calculate the individualized daily range demand of the corresponding new energy electric vehicle based on the center mileage value of the short-distance travel cluster, the center mileage value of the daily travel cluster, the preset long-distance travel threshold and the corresponding travel frequency, so as to formulate the corresponding battery specifications for new energy electric vehicle users. The daily travel cluster center mileage value is obtained by clustering the preprocessed mileage dataset with the initial cluster center, and the cluster center is continuously iterated during the cluster analysis; the above design can realize the calculation of individual range demand of new energy electric vehicle users and support the precise customization of battery specifications.

[0018] In the above technical solution, the mileage data of all single trips of vehicles within a set time period is obtained. Long-distance trip data is then filtered out from the mileage data of all single trips of vehicles within the set time period using a preset long-distance travel threshold, resulting in a mileage dataset for short-distance and daily trips. The specific process of preprocessing the mileage datasets of short-distance and daily trips to obtain the preprocessed mileage dataset is as follows: Monthly mileage datasets for individual vehicles are obtained through a vehicle monitoring system. (i.e., the mileage data of all single trips of vehicles within a set time period), where x i This represents the mileage of a single trip by a single vehicle. like x i Greater than the preset long-distance travel threshold m ( m (for 100km), x i Labeling trips as long-distance journeys and removing the monthly trip mileage dataset D for single vehicles, we obtain mileage datasets for short-distance and daily trips. ; Mileage datasets for short-distance and daily travel Z-Score standardization is performed; the above design directly filters out long-distance travel trips by setting a preset long-distance travel threshold, which can avoid the impact of extreme data values ​​on cluster centers and ensure the accuracy of the division between short-distance travel trips and daily travel trips.

[0019] The above technical solution involves mileage datasets for short-distance and daily travel. The specific process of Z-Score standardization is as follows: Mileage datasets for short-distance and daily travel The formula for calculating Z-Score standardization is as follows: Where, x j This represents the mileage value of a single vehicle's short-distance trip or daily travel journey. The mean of the mileage dataset D1, representing short-distance and daily trips; The standard deviation of the mileage dataset D1 represents the distances of short-distance and daily trips; This represents the standardized mileage value for a single short-distance trip or daily commut by a single vehicle. Obtain the preprocessed mileage dataset The above design avoids clustering bias caused by differences in metric by standardizing the mileage dataset D1 for short-distance and daily travel using Z-Score.

[0020] In the above technical solution, a set number of mileage data points are randomly selected from the preprocessed mileage dataset as initial cluster centers. K-Means clustering analysis is then performed on the preprocessed mileage dataset using these initial cluster centers. The cluster centers are iterated continuously during the K-Means clustering analysis to obtain the mileage values ​​for short-distance travel cluster centers, the standard deviation of short-distance travel clusters, and the mileage values ​​for daily travel cluster centers. The specific process is as follows: Preprocessed mileage dataset Divided into K clusters, due to the preprocessed mileage dataset Including mileage data from short-distance trips and daily trips, we get K=2, which means we divide the data into cluster 1 and cluster 2. From the preprocessed mileage dataset Two initial cluster centers were randomly selected from the data. , ; Calculate the preprocessed mileage dataset Each mileage data point Respectively with the initial cluster centers and Euclidean distance ( k ∈{1,2}), this mileage data point Assign to the cluster corresponding to the initial cluster center with the closest Euclidean distance. ( k∈{1,2}), minimizing the sum of squared Euclidean distances J for all mileage data points within the same cluster: Recalculate the initial cluster centers corresponding clusters The mean of all mileage data points within the cluster is used as the new cluster center. The calculation formula is as follows: Recalculate the initial cluster centers corresponding clusters The mean of all mileage data points within the cluster is used as the new cluster center. The calculation formula is as follows: Calculate the preprocessed mileage dataset Each mileage data point Each with the new cluster center and Euclidean distance ( k (∈{1,2}), repeat the above process, iterate t times to obtain the current cluster center. and ; Calculate the preprocessed mileage dataset Each mileage data point Respectively with the current cluster center and Euclidean distance ( k ∈{1,2}), this mileage data point Assign to the cluster corresponding to the current cluster center with the closest Euclidean distance. ( k ∈{1,2}), minimizing the sum of squared Euclidean distances J for all mileage data points within the same cluster: Recalculate the current cluster centers corresponding clusters The mean of all mileage data points within the cluster is used as the new cluster center. The calculation formula is as follows: Recalculate the current cluster centers corresponding clusters The mean of all mileage data points within the cluster is used as the new cluster center. The calculation formula is as follows: If satisfied If the condition is met, the iteration stops, where ε represents the convergence threshold (the convergence threshold ε can be set to 0.0001; mileage data is usually rounded to one decimal place, ε=0.0001 is equivalent to the value change of the cluster center being less than 0.01%, which is higher than the accuracy requirement of mileage data, so the convergence threshold ε is set to 0.0001); otherwise, the iteration continues until the condition is met. ; Finally, the cluster center values ​​of cluster 1 were obtained. Cluster center value of cluster 2 ,set up < ,but This is the center mileage value for short-distance travel clusters. This refers to the mileage value at the center of the daily travel cluster. At this time, the standard deviation of the short-distance travel cluster The calculation formula is as follows: Where n is the number of samples of mileage data points in cluster 1. This represents the value of the mileage data point in cluster 1. This represents the sample mean of the mileage data points in cluster 1; the above design is iteratively optimized using K-Means clustering analysis (K=2) to dynamically calculate the mileage value of the short-distance travel cluster center. and the mileage value of the daily travel cluster center It breaks through the limitations of fixed thresholds (such as the traditional method of preset 15km / 30km) and accurately adapts to different user travel patterns.

[0021] In the above technical solution, the short-distance travel mileage range value x 短 For: 0 < x 短 ≤ The daily travel mileage range value x 日 for: <x 日 ≤m; then the range of long-distance travel distances is x. 长 For: x 长 >m; range of short-distance travel distances x 短 Daily travel mileage range value x 日 and the range value of long-distance travel mileage x 长 The system identifies users' long-distance, short-distance, and daily trips to obtain the number of trips for each type of trip within the monthly mileage of a single vehicle; the above design is based on the mileage value of the center of the short-distance trip cluster. Standard deviation of short-distance travel cluster It generates short-distance travel mileage range values, breaking through fixed thresholds for short-distance travel (such as 30km), and accurately matches the fluctuations of users' actual short-distance travel trips; by complementing non-overlapping ranges of users' long-distance, short-distance, and daily trips, it can reasonably resolve the boundary disputes between short-distance and daily travel trips in traditional methods; it accurately counts the number of trips of the three types, providing real behavioral data for the calculation of range requirements and avoiding errors in empirical estimation.

[0022] In the above technical solution, the specific process of calculating the individualized daily range requirement of the corresponding new energy electric vehicle based on the short-distance travel cluster center mileage value, the daily travel cluster center mileage value, the preset long-distance travel threshold, and the corresponding travel frequency, and thus formulating the corresponding battery specifications for new energy electric vehicle users, is as follows: Based on the short-distance travel cluster center mileage value, the daily travel cluster center mileage value, the preset long-distance travel threshold, and the number of long-distance, short-distance, and daily trips for each user in the corresponding monthly mileage of a single vehicle, the calculation formula for the individualized daily average range requirement of new energy electric vehicles is as follows: in, f 1 For the average daily frequency of short-distance trips, f 2 For daily travel frequency, f 3 The average daily frequency of long-distance travel (because long-distance travel accounts for a small proportion of the monthly mileage of a single vehicle and is prone to extreme values, the preset long-distance travel threshold m is used as the mileage value when calculating the average daily range requirement). This is the center mileage value for short-distance travel clusters. This refers to the mileage value at the center of the daily travel cluster. The battery capacity requirement is calculated by multiplying the daily average range requirement by the number of days in a week and the weekly range redundancy coefficient (based on historical failure rates, the weekly range redundancy coefficient can be set to 1.2; specifically, it is determined by analyzing cases of insufficient range failures in new energy electric vehicles and adjusting the redundancy coefficient accordingly, ensuring that the covered failure scenarios meet the requirements while avoiding over-design that would increase battery costs). This allows for the development of appropriate battery specifications for new energy electric vehicle users. The above design, through the calculation formula for individualized daily average range requirements of new energy electric vehicles, can scientifically and rationally obtain battery capacity requirements, avoiding over- or under-capacity battery configurations. Calculating the weekly range requirement based on the daily average range requirement provides a data-driven basis for battery configuration, reducing resource waste.

[0023] Example 2 A method for identifying user long and short distances and daily itineraries based on K-Means clustering, such as Figure 2As shown, the process involves acquiring single-trip mileage data for vehicles, filtering out long-distance travel data to obtain mileage datasets for short-distance and daily travel, and preprocessing this data to obtain a preprocessed mileage dataset. Mileage data is randomly selected from the preprocessed dataset as initial cluster centers, and cluster analysis is performed, iterating through the cluster centers to obtain the mileage values ​​and standard deviations of the short-distance travel cluster centers. Based on the short-distance travel cluster center mileage values ​​and standard deviations, as well as the long-distance travel threshold, the mileage range values ​​for short-distance and daily travel are obtained. These mileage range values, along with the long-distance travel threshold, are used to identify the user's long-distance, short-distance, and daily travel.

[0024] The specific methods for identifying user long / short distance and daily itinerary include the following steps: Acquire all vehicle single trip mileage data within a set time period. Filter out long-distance trip data from all vehicle single trip mileage data within the set time period using a preset long-distance travel threshold to obtain short-distance and daily trip mileage datasets. Preprocess the short-distance and daily trip mileage datasets to obtain preprocessed mileage datasets. A set number of mileage data points are randomly selected from the preprocessed mileage dataset as initial cluster centers. Cluster analysis is performed on the preprocessed mileage dataset using the initial cluster centers, and the cluster centers are iterated continuously during the cluster analysis to obtain the mileage value of the short-distance travel cluster center and the standard deviation of the short-distance travel cluster. The short-distance travel mileage range and daily travel mileage range are obtained based on the center mileage value of the short-distance travel cluster, the standard deviation of the short-distance travel cluster, and the preset long-distance travel threshold. The user's long-distance, short-distance, and daily trips are identified by the short-distance travel mileage range, daily travel mileage range, and the preset long-distance travel threshold.

[0025] Example 3 A computer program product includes a computer program that, when executed by a processor, implements the steps of the method described in Embodiment 2.

[0026] Example 4 A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described in Embodiment 2.

[0027] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

Claims

1. A user long / short distance and daily itinerary identification system based on K-Means clustering, characterized in that, It includes: The data preprocessing module is used to acquire the mileage data of all single trips of vehicles within a set time period. By using a preset long-distance travel threshold, long-distance travel data is filtered out from the mileage data of all single trips of vehicles within the set time period to obtain the mileage dataset of short-distance trips and daily trips. The mileage dataset of short-distance trips and daily trips is preprocessed to obtain the preprocessed mileage dataset. The cluster analysis module is used to randomly select a set number of mileage data from the preprocessed mileage dataset as the initial cluster centers. The preprocessed mileage dataset is then subjected to cluster analysis using the initial cluster centers. The cluster centers are iterated continuously during the cluster analysis to obtain the mileage value of the short-distance travel cluster center and the standard deviation of the short-distance travel cluster. The trip identification module is used to obtain the short-distance trip mileage range value and the daily trip mileage range value based on the center mileage value of the short-distance trip cluster, the standard deviation of the short-distance trip cluster, and the preset long-distance trip threshold. The module then identifies the user's long-distance, short-distance, and daily trips using the short-distance trip mileage range value, the daily trip mileage range value, and the preset long-distance trip threshold.

2. The user long / short distance and daily itinerary identification system based on K-Means clustering according to claim 1, characterized in that, It also includes: The range demand analysis module is used to calculate the individualized daily range demand of the corresponding new energy electric vehicle based on the center mileage value of the short-distance travel cluster, the center mileage value of the daily travel cluster, the preset long-distance travel threshold and the corresponding travel frequency, so as to formulate the corresponding battery specifications for new energy electric vehicle users. The daily travel cluster center mileage value is obtained by performing cluster analysis on the preprocessed mileage dataset using the initial cluster center, and by continuously iterating the cluster center during the cluster analysis.

3. The user long / short distance and daily itinerary identification system based on K-Means clustering according to claim 2, characterized in that: The process involves acquiring mileage data for all single trips of vehicles within a set time period. Long-distance trip data is then filtered out from this data using a preset long-distance travel threshold, resulting in mileage datasets for short-distance and daily trips. These short-distance and daily trip mileage datasets are then preprocessed to obtain the preprocessed mileage dataset. Monthly mileage datasets for individual vehicles are obtained through a vehicle monitoring system. ,in x i This represents the mileage of a single trip by a single vehicle. like x i Greater than the preset long-distance travel threshold m ,Will x i Labeling trips as long-distance journeys and removing the monthly trip mileage dataset D for single vehicles, we obtain mileage datasets for short-distance and daily trips. ; Mileage datasets for short-distance and daily travel Perform Z-Score standardization.

4. The user long / short distance and daily itinerary identification system based on K-Means clustering according to claim 3, characterized in that: Mileage datasets for short-distance and daily travel The specific process of Z-Score standardization is as follows: Mileage datasets for short-distance and daily travel The formula for calculating Z-Score standardization is as follows: Where, x j This represents the mileage value of a single vehicle's short-distance trip or daily travel journey. The mean of the mileage dataset D1, representing short-distance and daily trips; The standard deviation of the mileage dataset D1, representing short-distance and daily travel distances; This represents the standardized mileage value for a single short-distance trip or daily commut by a single vehicle. Obtain the preprocessed mileage dataset .

5. The user long / short distance and daily itinerary identification system based on K-Means clustering according to claim 4, characterized in that: The process involves randomly selecting a set number of mileage data points from the preprocessed mileage dataset as initial cluster centers. K-Means clustering analysis is then performed on the preprocessed mileage dataset using these initial cluster centers. The K-Means clustering analysis iterates through the cluster centers to obtain the mileage values ​​for short-distance trip cluster centers, the standard deviation of short-distance trip cluster centers, and the mileage values ​​for daily trip cluster centers. Preprocessed mileage dataset Divided into K clusters, due to the preprocessed mileage dataset Including mileage data from short-distance trips and daily trips, we get K=2, which means we divide the data into cluster 1 and cluster 2. From the preprocessed mileage dataset Two initial cluster centers were randomly selected from the data. , ; Calculate the preprocessed mileage dataset Each mileage data point Respectively with the current cluster center and Euclidean distance ( k ∈{1,2}), this mileage data point Assign to the cluster corresponding to the current cluster center with the closest Euclidean distance. ( k ∈{1,2}), minimizing the sum of squared Euclidean distances J for all mileage data points within the same cluster: Recalculate the current cluster centers corresponding clusters The mean of all mileage data points within the cluster is used as the new cluster center. The calculation formula is as follows: Recalculate the current cluster centers corresponding clusters The mean of all mileage data points within the cluster is used as the new cluster center. The calculation formula is as follows: If satisfied If ε = 0, then the iteration stops, where ε represents the convergence threshold. Otherwise, continue iterating until the condition is met. ; Finally, the cluster center values ​​of cluster 1 were obtained. Cluster center value of cluster 2 ,set up < ,but This is the center mileage value for short-distance travel clusters. This refers to the mileage value at the center of the daily travel cluster. At this time, the standard deviation of the short-distance travel cluster The calculation formula is as follows: Where n is the number of samples of mileage data points in cluster 1. This represents the value of the mileage data point in cluster 1. This represents the sample average value of the mileage data points in cluster 1.

6. The user long / short distance and daily itinerary identification system based on K-Means clustering according to claim 5, characterized in that: The short-distance travel mileage range value x 短 For: 0 < x 短 ≤ The daily travel mileage range value x 日 for: <x 日 ≤m; then the range of long-distance travel distances is x. 长 For: x 长 >m; range of short-distance travel distance values ​​x 短 Daily travel mileage range value x 日 and the range value of long-distance travel mileage x 长 The system identifies users' long-distance, short-distance, and daily trips, and obtains the number of trips for each type of user in the monthly mileage of a single vehicle.

7. The user long / short distance and daily itinerary identification system based on K-Means clustering according to claim 6, characterized in that: Based on the short-distance travel cluster center mileage value, the daily travel cluster center mileage value, the preset long-distance travel threshold, and the corresponding travel frequency, the individualized daily average range requirement of the corresponding new energy electric vehicle is calculated, and the specific process of formulating corresponding battery specifications for new energy electric vehicle users is as follows: Based on the short-distance travel cluster center mileage value, the daily travel cluster center mileage value, the preset long-distance travel threshold, and the number of long-distance, short-distance, and daily trips for each user in the corresponding monthly mileage of a single vehicle, the calculation formula for the individualized daily average range requirement of new energy electric vehicles is as follows: in, f 1 For the average daily frequency of short-distance trips, f 2 For daily travel frequency, f 3 For the average daily frequency of long-distance travel, This is the center mileage value for short-distance travel clusters. This refers to the mileage value at the center of the daily travel cluster. The battery capacity requirement is obtained by multiplying the daily average range requirement by the number of days in a week by the weekly range redundancy coefficient, thereby enabling the formulation of corresponding battery specifications for new energy electric vehicle users.

8. A method for identifying user long and short trips based on K-Means clustering, characterized in that, It includes the following steps: Acquire all vehicle single trip mileage data within a set time period. Filter out long-distance trip data from all vehicle single trip mileage data within the set time period using a preset long-distance travel threshold to obtain short-distance and daily trip mileage datasets. Preprocess the short-distance and daily trip mileage datasets to obtain preprocessed mileage datasets. A set number of mileage data points are randomly selected from the preprocessed mileage dataset as initial cluster centers. Cluster analysis is performed on the preprocessed mileage dataset using the initial cluster centers, and the cluster centers are iterated continuously during the cluster analysis to obtain the mileage value of the short-distance travel cluster center and the standard deviation of the short-distance travel cluster. The short-distance travel mileage range and daily travel mileage range are obtained based on the center mileage value of the short-distance travel cluster, the standard deviation of the short-distance travel cluster, and the preset long-distance travel threshold. The user's long-distance, short-distance, and daily trips are identified by the short-distance travel mileage range, daily travel mileage range, and the preset long-distance travel threshold.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method of claim 8.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it implements the steps of the method of claim 8.

Images