Urban distribution vehicle temporary stop multi-stage site selection method and device and storage medium

Abstract

The invention relates to a city distribution vehicle temporary stop multi-stage site selection method and device and a storage medium. The method comprises the following steps: S1, acquiring a planar map of a target area and meshing the planar map; s2, generating the weight of each grid according to the number of potential temporary stop points in each grid, clustering all grids to obtain a plurality of grid clusters, and taking each grid cluster as a research area; s3, in a single research area, pre-site selection is carried out on all potential temporary stop points, and a part of potential temporary stop points are obtained through screening to serve as core points; and S4, extracting all interest points related to the potential temporary stop points in the research area, respectively calculating weighted distance sums between each core point and all the interest points in combination with the weight of each interest point, and selecting the core point with the maximum weighted distance sum as the final temporary stop point. Compared with the prior art, the method meets the actual parking requirements of urban vehicles, and can reduce the influence on the normal life of a city.

Description

Technical Field

[0001] The present invention relates to the field of traffic information, and in particular to a method, device and storage medium for selecting a multi-stage site for a temporary stop of a city distribution vehicle. Background Art

[0002] Identifying and classifying the stops of freight vehicles is a basic task in freight data analysis. Its goal is to segment the continuous vehicle trajectory data into a series of meaningful trip segments and activity areas by identifying the location and purpose of the vehicle stop. This process is crucial for further freight characteristics analysis and travel pattern research. The types of stops are usually divided based on the activities related to the vehicle stop, which include the start and end points of cargo loading and unloading, stops due to rest, dining, refueling, accidents or severe traffic jams, and short stops due to traffic light control, minor congestion or tolls. In the process of identifying the stop points, researchers usually first determine a low-speed threshold to identify potential stop points by comparing the instantaneous speed of the truck or the average speed within a specific time window, and then use appropriate techniques to determine the specific type of the stop point.

[0003] In the existing literature, the methods for identifying the types of freight vehicle stop points can be mainly divided into the following three types:

[0004] (1) Auxiliary information method

[0005] This method mainly relies on auxiliary information, such as driver survey data and land use information, to identify the stop points. Alho et al. successfully identified the types of truck stop points by combining truck trajectory data and feedback from driver surveys. Although this method performs well when processing small sample data, its application is limited by the availability of auxiliary information and the size of the data, and its generalization ability is low.

[0006] (2) Rule-based discrimination

[0007] This method mainly uses the dwell time as the judgment criterion, and identifies the type of parking point by setting a specific dwell time threshold. Depending on the research background, the dwell time threshold will be different. For example, Richardson and Sharman set thresholds of 3 minutes and 5 minutes respectively in the urban freight study, while Gingerich set the threshold between 15 and 45 minutes in the cross-border freight study. In addition, Li Jiangtao identified parking points by setting the parking speed threshold of freight trucks, and further processed and analyzed the data through map matching and SCAN clustering methods. This method is simple to operate, but its classification results are relatively rough and are usually only used for preliminary screening. Yang et al. combined rule discrimination with auxiliary information methods, defined the speed threshold and multi-level time threshold required to identify truck stops, and considered the complexity of the urban freight environment, combined urban road network and POI data to improve the recognition accuracy, and effectively reflected the spatial distribution and operation mode of heavy truck freight in the city.

[0008] (3) Machine Learning Algorithms

[0009] These include spatial density-based algorithms such as K-means and DBSCAN, as well as support vector machines (SVM), random forests (RF), maximum entropy models, etc. These methods usually use features such as residence time and distance from a specific location for classification, and can achieve high recognition accuracy. Summary of the invention

[0010] The purpose of the present invention is to provide a multi-stage site selection method, device and storage medium for temporary parking points of urban distribution vehicles.

[0011] The purpose of the present invention can be achieved by the following technical solutions:

[0012] A multi-stage site selection method for temporary parking points of urban distribution vehicles, comprising:

[0013] Step S1: Obtain a plane map of the target area and grid it;

[0014] Step S2: Generate the weight of each grid according to the number of potential temporary stops in each grid, cluster all grids to obtain multiple grid clusters, and take each grid cluster as a research area;

[0015] Step S3: In a single study area, pre-select all potential temporary stop points, and select some potential temporary stop points as core points;

[0016] Step S4: In the study area, extract all points of interest related to potential temporary stopping points, calculate the weighted sum of distances between each core point and all points of interest based on the weights of each point of interest, and select the core point with the largest sum of weighted distances as the final temporary stopping point.

[0017] The grid is a square grid.

[0018] The step S2 comprises:

[0019] Step S2-1: Generate the weight of each grid according to the number of potential temporary stop points in each grid, cluster the grids by K-Means weighted clustering method, and obtain preliminary clustering results, wherein the weight of the grid is positively correlated with the number of potential temporary stop points in the grid;

[0020] Step S2-2: Modify the preliminary clustering results to obtain multiple grid clusters, and take each grid cluster as a research area.

[0021] The correction process in step S2-2 includes:

[0022] (1) Density-centered principle: Each grid cluster contains at least one super-high-density grid or high-density grid as the core of the division, where the super-high-density grid is a grid with a number of potential temporary stops greater than a first set threshold, and the high-density grid is a grid with a number of potential temporary stops greater than a second set threshold;

[0023] (2) Exclusivity principle: Each grid can only belong to one grid cluster;

[0024] (3) Expansion principle: With a high-density grid or an ultra-high-density grid as the center, expand to the surrounding areas. If there are medium-density grids or above among the four adjacent grids, the medium-density grids or above are added to the grid cluster where the high-density grid or the ultra-high-density grid is located;

[0025] (4) Edge extension principle: If a high-density grid or ultra-high-density grid has clustered potential temporary stops that extend to adjacent grids, and the adjacent grids are medium-density or higher-density grids, then the medium-density or higher-density grids are added to the grid cluster where the high-density grid or ultra-high-density grid is located.

[0026] Step S3 divides the potential temporary stop points in the study area into core points, edge points and noise points.

[0027] The core point is: the number of potential temporary stopping points within a set radius around the core point is greater than the minimum number of neighbors;

[0028] The edge point is: the distance from any core point is less than the set radius, and the number of potential temporary stop points within the set radius around the edge point is less than the minimum number of neighbors;

[0029] The noise points are potential temporary stopping points other than core points and edge points.

[0030] The step S4 comprises:

[0031] Step S4-1: extracting all points of interest related to potential temporary stops in the study area;

[0032] Step S4-2: screening the points of interest obtained in step S4-1;

[0033] Step S4-3: Obtain the weights of all the filtered points of interest;

[0034] Step S4-4: Calculate the weighted distance sum between each core point and all interest points respectively, and select the core point with the largest weighted distance sum as the final temporary stop point.

[0035] The screening method in step S4-2 includes:

[0036] Excluding service industry points of interest with low relevance to freight delivery;

[0037] Exclude medical and health-related points of interest;

[0038] Exclude shopping and consumption points of interest.

[0039] The weighted distance sum D is:

[0040]

[0041] Where: n is the number of points of interest, w i is the weight of the i-th interest point, d i is the distance between the core point and the i-th interest point.

[0042] A multi-stage site selection device for temporary parking points of urban distribution vehicles includes a memory, a processor, and a program stored in the memory. When the processor executes the program, the method as described above is implemented.

[0043] A storage medium stores a program, which implements the above method when executed.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] 1. By creating grid units, potential temporary stops are gridded, and the grids are initially divided by weighted clustering method, and on this basis, the study area is further divided based on empirical rules; then, potential temporary stops are pre-selected; and the temporary stops are optimized based on points of interest that are significantly related to urban distribution in the study area to ensure that the selected locations meet the needs of urban freight to the greatest extent; this method can provide a scientific basis for the site selection of temporary stops for urban distribution vehicles based on the data information of temporary stops for urban distribution vehicles, provide technical support for improving the efficiency of freight distribution of urban distribution vehicles and alleviating traffic congestion, and will help promote the sustainable development of urban transportation.

[0046] 2. Use weighted clustering to divide data points into K clusters, where each cluster consists of data points closest to its centroid (center point). Unlike the traditional K-Means method, weighted K-Means assigns a weight value to each data point during the clustering process, so that different data points have different influences in determining the cluster center. In the scenario of urban freight, the number of stops in each grid can be regarded as a weight value, which means that grids with more stops have higher priority in the clustering process.

[0047] 3. The DBSCAN clustering method is used to distinguish the potential temporary stops in the study area into core points, edge points, and noise points, so that irregular or disorganized data sets can be effectively processed, the true clustering structure can be identified, and isolated data points or outliers can be ignored.

[0048] 4. It can meet the actual parking needs of urban distribution vehicles and reduce the impact on normal urban life. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic flow chart of the main steps of the method of the present invention;

[0050] Figure 2 This is a schematic diagram of the POI correlation coefficient of urban distribution vehicles in this application. DETAILED DESCRIPTION

[0051] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0052] A multi-stage site selection method for temporary parking points of urban distribution vehicles, such as Figure 1 As shown, including:

[0053] Step S1: Obtain a plane map of the target area and grid it;

[0054] Generally, the grid is a square grid. In this embodiment, a square grid with a side length of 1 kilometer is used.

[0055] Step S2: Generate the weight of each grid based on the number of potential temporary stops in each grid, cluster all grids to obtain multiple grid clusters, and take each grid cluster as a research area, including:

[0056] Step S2-1: Generate the weight of each grid according to the number of potential temporary stop points in each grid, cluster the grids by K-Means weighted clustering method, and obtain preliminary clustering results, wherein the weight of the grid is positively correlated with the number of potential temporary stop points in the grid;

[0057] Due to the weighted clustering method, each grid is assigned a weight, so that different grids have different influences in determining the cluster center. K-Means weighted clustering is a partition-based clustering method that aims to divide the grids into k sets, each set consisting of the grids closest to its centroid (center point).

[0058] The K value is determined by the elbow method or the silhouette coefficient method:

[0059] (1) The elbow method determines the optimal k value by observing the change in clustering cost under different k values. As the k value increases, the clustering cost decreases. The clustering cost calculation formula is:

[0060]

[0061] Where: n1 is the number of grids, Distortion is the clustering cost, μ j is the center of set j, x i is the i-th grid,

[0062] (2) The silhouette coefficient method is to calculate the average distance a between a grid and other grids in its set and the average distance b between all grids in the nearest set. The calculation formula of the silhouette coefficient s is:

[0063]

[0064] Among them: s is between -1 and 1, and the higher the s value, the better the clustering effect.

[0065] Step S2-2: Modify the preliminary clustering results to obtain multiple grid clusters, and use each grid cluster as a research area. The modification process includes:

[0066] (1) Density-centered principle: Each grid cluster contains at least one super-high-density grid or high-density grid as the core of the division, where super-high-density refers to a grid with a number of potential temporary stops greater than a first set threshold, and high-density refers to a grid with a number of potential temporary stops greater than a second set threshold;

[0067] (2) Exclusivity principle: Each grid can only belong to one grid cluster;

[0068] (3) Expansion principle: With a high-density grid or ultra-high-density grid as the center, expand to the surrounding areas. If there are medium-density grids or higher among the four adjacent grids, then the medium-density grids or higher are added to the grid cluster where the high-density grid or ultra-high-density grid is located;

[0069] (4) Edge extension principle: If a high-density grid or ultra-high-density grid has clustered potential temporary stops that extend to adjacent grids, and the adjacent grids are medium-density or above grids, then the medium-density or above grids will be added to the grid cluster where the high-density grid or ultra-high-density grid is located.

[0070] In addition, in this embodiment, the present invention also includes:

[0071] (5) Morphological optimization principle: When forming a grid cluster, if a rectangular area can be formed by completing the medium-density grid, it can be considered to be included in the grid cluster that may become a rectangular area, but this is not a mandatory requirement.

[0072] (6) Principles for handling special cases: For areas containing special landmarks such as large commercial centers, they should be completely classified into the same grid cluster based on actual conditions.

[0073] (7) Partition area principle: Low-density areas are generally regarded as partition areas and used as an important reference for grid cluster division.

[0074] (8) Principle of cutting large areas: For grid clusters that are too large due to dense stopover points, they should be further cut and subdivided based on characteristics such as street groups to obtain smaller and more practical grid clusters.

[0075] Step S3: In a single study area, pre-select all potential temporary stop points, and select some potential temporary stop points as core points;

[0076] Step S3 divides the potential temporary stop points in the study area into core points, edge points and noise points.

[0077] Among them, a core point is: the number of potential temporary stops within a set radius around the core point is greater than the minimum number of neighbors;

[0078] An edge point is one where the distance to any core point is less than the set radius, and the number of potential temporary stopover points within the set radius around the edge point is less than the minimum number of neighbors;

[0079] Noise points are potential temporary stopping points other than core points and edge points.

[0080] In this embodiment, the DBSCAN density clustering method is specifically used to pre-select the potential temporary stop points; the DBSCAN density clustering method is a density-based clustering algorithm.

[0081] Step S4: In the study area, extract all points of interest related to potential temporary stop points, calculate the weighted distance sum between each core point and all points of interest based on the weight of each point of interest, and select the core point with the largest weighted distance sum as the final temporary stop point, including:

[0082] Step S4-1: extracting all points of interest related to potential temporary stops in the study area;

[0083] The introduction of point of interest data related to urban distribution, namely POI data, can make up for the limitations of site selection based only on the number of temporary stops and regional density. Taking into account the differences in freight volume, cargo type and importance of freight services involved in different stops, the site selection method based solely on the absolute number of stops may not accurately reflect these subtle differences. In addition, the actual stop location may not always correspond exactly to the specific destination of freight demand. For example, vehicles may choose to stop at non-target locations such as residential areas around demand points and near schools, which may not only affect the normal travel of local residents and students, but also interfere with the accuracy of site selection.

[0084] Therefore, this application will use the distribution and weight of POI points with high relevance to urban freight delivery in the study area to make site selection decisions. This method can ensure that the final selected temporary stop can not only meet the actual parking needs of urban distribution vehicles, but also reduce the impact on normal urban life.

[0085] Step S4-2: screening the points of interest obtained in step S4-1;

[0086] like Figure 2 As shown in Figure 1, it is a schematic diagram of the correlation coefficient of urban delivery vehicles POI. In order to screen the POI categories that are closely related to urban freight delivery, Figure 2 The following three types of POIs are excluded:

[0087] (1) Excluding service industry POIs with low relevance to freight delivery;

[0088] (2) Excluding medical and health POIs;

[0089] (3) Eliminate shopping consumption POIs that may lead to duplicate counting;

[0090] Step S4-3: Obtain the weights of all the filtered points of interest;

[0091] After screening, the POI types with high relevance to the temporary parking demand of urban distribution vehicles are obtained as shown in Table 1. Since the similarity between the sub-category POIs in the catering and hotel accommodation categories is high, they are merged. The merged results in this embodiment are shown in Table 1:

[0092] Table 1

[0093]

[0094] Step S4-4: Calculate the weighted distance sum between each core point and all interest points respectively, and select the core point with the largest weighted distance sum as the final temporary stop point.

[0095] After screening out the types of POIs related to urban distribution, the expert judgment method is used to assign weights to each type of POI. After assigning weights to each type of POI related to urban distribution, the POI weighted distance model is then constructed. In the POI weighted distance model, each geographical location is assigned a weight, and the distance is adjusted by adjusting the weight. The weighted short distance uses the inverse of the distance as the weight. The total weighted distance calculation formula for all urban distribution vehicle-related POIs to the pre-selected temporary stop is:

[0096]

[0097] Where: n is the number of points of interest, w i is the weight of the i-th interest point, d i is the distance between the core point and the i-th interest point.

[0098] Finally, in this embodiment, the site selection results are evaluated and analyzed through the Voronoi diagram, and the analysis process is as follows:

[0099] Step S5-1: Determine seed points, determine a series of seed points, and these points may be geographical location points;

[0100] Step S5-2: Construct the Voronoi boundary. For each pair of seed points in the seed point set, find the set of all points that are equidistant from the two seed points. This set of points forms a line, called the Voronoi edge.

[0101] Step S5-3: Generate Voronoi cells, repeat step S5-2 until the Voronoi edges between each pair of seed points are determined, and these edges together define the Voronoi cells around each seed point, forming a complete Voronoi diagram;

[0102] Step S5-4: Analysis and Application: By evaluating the size and shape of these Voronoi cells, the spatial coverage of temporary stops can be evaluated, thereby optimizing their location and distribution.

[0103] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

Claims

1. A multi-stage site selection method for temporary parking points of urban distribution vehicles, characterized in that: include: Step S1: Obtain a plane map of the target area and grid it; Step S2: Generate the weight of each grid according to the number of potential temporary stops in each grid, cluster all grids to obtain multiple grid clusters, and take each grid cluster as a research area; Step S3: In a single study area, pre-select all potential temporary stop points, and select some potential temporary stop points as core points; Step S4: In the study area, extract all points of interest related to potential temporary stopping points, calculate the weighted sum of distances between each core point and all points of interest based on the weights of each point of interest, and select the core point with the largest sum of weighted distances as the final temporary stopping point.

2. A multi-stage site selection method for temporary parking points of urban distribution vehicles according to claim 1, characterized in that: The grid is a square grid.

3. A multi-stage site selection method for temporary parking points of urban distribution vehicles according to claim 1, characterized in that: The step S2 comprises: Step S2-1: Generate the weight of each grid according to the number of potential temporary stop points in each grid, cluster the grids by K-Means weighted clustering method, and obtain preliminary clustering results, wherein the weight of the grid is positively correlated with the number of potential temporary stop points in the grid; Step S2-2: Modify the preliminary clustering results to obtain multiple grid clusters, and take each grid cluster as a research area.

4. A multi-stage site selection method for temporary parking points of urban distribution vehicles according to claim 3, characterized in that: The correction process in step S2-2 includes: (1) Density-centered principle: Each grid cluster contains at least one super-high-density grid or high-density grid as the core of the division, where the super-high-density grid is a grid with a number of potential temporary stops greater than a first set threshold, and the high-density grid is a grid with a number of potential temporary stops greater than a second set threshold; (2) Exclusivity principle: Each grid can only belong to one grid cluster; (3) Expansion principle: With a high-density grid or an ultra-high-density grid as the center, expand to the surrounding areas. If there are medium-density grids or above among the four adjacent grids, the medium-density grids or above are added to the grid cluster where the high-density grid or the ultra-high-density grid is located; (4) Edge extension principle: If a high-density grid or ultra-high-density grid has clustered potential temporary stops that extend to adjacent grids, and the adjacent grids are medium-density or higher-density grids, then the medium-density or higher-density grids are added to the grid cluster where the high-density grid or ultra-high-density grid is located.

5. A multi-stage site selection method for temporary parking points of urban distribution vehicles according to claim 1, characterized in that: Step S3 divides the potential temporary stop points in the study area into core points, edge points and noise points. The core point is: the number of potential temporary stopping points within a set radius around the core point is greater than the minimum number of neighbors; The edge point is: the distance from any core point is less than the set radius, and the number of potential temporary stop points within the set radius around the edge point is less than the minimum number of neighbors; The noise points are potential temporary stopping points other than core points and edge points.

6. A multi-stage site selection method for temporary parking points of urban distribution vehicles according to claim 1, characterized in that: The step S4 comprises: Step S4-1: extracting all points of interest related to potential temporary stops in the study area; Step S4-2: screening the points of interest obtained in step S4-1; Step S4-3: Obtain the weights of all the filtered points of interest; Step S4-4: Calculate the weighted distance sum between each core point and all interest points respectively, and select the core point with the largest weighted distance sum as the final temporary stop point.

7. A multi-stage site selection method for temporary parking points of urban distribution vehicles according to claim 6, characterized in that: The screening method in step S4-2 includes: Excluding service industry points of interest with low relevance to freight delivery; Exclude medical and health-related points of interest; Exclude shopping and consumption points of interest.

8. A multi-stage site selection method for temporary parking points of urban distribution vehicles according to claim 6, characterized in that: The weighted distance sum D is: Where: n is the number of points of interest, w i is the weight of the i-th interest point, d i is the distance between the core point and the i-th interest point.

9. A multi-stage site selection device for temporary parking points of urban distribution vehicles, comprising a memory, a processor, and a program stored in the memory, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 8 is implemented.

10. A storage medium having a program stored thereon, characterized in that: When the program is executed, the method according to any one of claims 1 to 8 is implemented.

Images