Data-driven urban green freight new energy vehicle parking purpose identification method

Abstract

The invention relates to a data-driven urban green freight new energy vehicle parking purpose identification method. The method comprises the following steps: S1, obtaining an original vehicle positioning data sequence; s2, preprocessing and repairing the obtained original positioning data to obtain one or more target vehicle positioning data sequences; s3, on the basis of the target vehicle positioning data sequence, comparing the distance and the average speed of the adjacent track points with a pre-configured distance threshold value and an average speed threshold value respectively, screening out all vehicle stop points, and constructing a vehicle stop point cluster by the continuous vehicle stop points; s4, judging whether each stop point set is an effective stop point set or an undetermined stop point set based on the time difference between the last stop point and the first stop point in the constructed stop point set; s5, for each to-be-determined stop point set, determining whether each stop point set is an effective stop point set or not in combination with turning angle and stroke chain analysis; and S6, determining the parking purpose of all the effective stop point sets. Compared with the prior art, the method has the advantages of high anti-interference capability and the like.

Description

Technical Field

[0001] The present invention relates to the field of vehicle transportation, and in particular to a data-driven method for identifying parking purposes of urban green freight new energy vehicles. Background Art

[0002] The acquisition of vehicle stopping points is an important basis for municipal planning. Different from the traditional manual identification and survey method, some existing technologies have proposed a stopping point identification method based on vehicle location data. For example, Chinese patent CN118469429A discloses a method for identifying truck stopping points based on the spatiotemporal characteristics of GPS data, including: arranging each trajectory point in order from morning to night according to the timestamp of each group of GPS trajectory data, calculating the number of neighborhood points of each trajectory point, cutting the trajectory density distribution curve according to the dense interval division threshold, and determining the dense interval of trajectory points; using cluster analysis to identify spatial clusters in multiple groups of original GPS trajectory data of the same truck; obtaining multiple dense intervals of trajectory points included or intersected by each spatial cluster, judging whether the multiple dense intervals of trajectory points included or intersected by the spatial cluster are continuous in time, so as to determine the candidate stopping point set; if the vehicle displacement, vehicle residence time, vehicle moving speed and vehicle travel mode in the candidate stopping point set all meet the preset conditions, then the candidate stopping point set is the final stopping point.

[0003] Through the above application, the stopping point can be identified by relying on GPS data without the need for survey, thus breaking through the spatial constraints and being identified by specialized software tools.

[0004] However, on the one hand, faced with the growing proportion of new energy trucks, the above method cannot distinguish between traditional stopping points and charging demand stopping points, resulting in the obtained data having no application value. On the other hand, when faced with situations such as overpass areas and poor network of vehicle-mounted positioning devices that may cause excessive data errors, the above method will be interfered by some data with smaller errors, resulting in a low accuracy rate in the final recognition result. Summary of the invention

[0005] The purpose of the present invention is to provide a data-driven method for identifying the parking purpose of urban green freight new energy vehicles.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] A data-driven urban green freight new energy vehicle parking purpose identification method, comprising:

[0008] Step S1: obtaining an original vehicle positioning data sequence, wherein each track point in the vehicle positioning data sequence includes latitude and longitude, timestamp, remaining power percentage, charging state and instantaneous speed;

[0009] Step S2: preprocessing and repairing the acquired original positioning data to obtain one or more target vehicle positioning data sequences;

[0010] Step S3: Based on the target vehicle positioning data sequence, all vehicle stop points are screened out by comparing the distance and average speed of adjacent trajectory points with the pre-configured distance threshold and average speed threshold, and the continuous vehicle stop points are constructed into a vehicle stop point cluster;

[0011] Step S4: judging whether each stop point set is a valid stop point set or a pending stop point set based on the time difference between the last stop point and the first stop point in the constructed stop point set;

[0012] Step S5: for each undetermined stop point set, combining the turning angle and the trip chain analysis, determine whether each stop point set is a valid stop point set;

[0013] Step S6: Determine the parking purpose of all valid stop point sets.

[0014] The step S2 comprises:

[0015] Step S2-1: Delete all data points with longitude or latitude of 0 in the vehicle positioning data sequence, and data points that exceed the longitude threshold range or latitude threshold range;

[0016] Step S2-2: Delete all data points whose timestamps do not change continuously in the vehicle positioning data sequence;

[0017] Step S2-3: Calculate the average speed between adjacent data points respectively. If the average speed of adjacent data points is greater than the pre-configured maximum speed, delete the corresponding data point.

[0018] Step S2-4: When the time stamp difference between adjacent data points exceeds a preconfigured time threshold, the vehicle positioning data sequence is segmented to obtain multiple target vehicle positioning data sequences.

[0019] The step S3 comprises:

[0020] Step S3-1: configure the distance threshold and the average speed threshold;

[0021] Step S3-2: respectively calculating the distance and average speed between each adjacent data point in the target vehicle positioning data sequence; if the distance between adjacent data points is less than the distance threshold, and the average speed is less than the average speed threshold, then determining the previous data point among the adjacent data points as the vehicle stop point;

[0022] Step S3-3: Construct all consecutive vehicle stop points into a vehicle stop point cluster.

[0023] The distance threshold is determined by: constructing a distance logarithmic distribution curve of two adjacent trajectory data points based on the target vehicle positioning data sequence, the logarithmic distribution curve presents a bimodal mixed distribution, and taking the target point of the mixed distribution as the distance threshold;

[0024] The average speed threshold is determined in the following manner: based on the target vehicle positioning data sequence, a logarithmic distribution curve of the average speed of two adjacent trajectory data points is constructed, the logarithmic distribution curve presents a bimodal mixed distribution, and the target point of the mixed distribution is used as the average speed threshold.

[0025] The step S4 comprises:

[0026] Step S4-1: Determine a first parking time threshold and a second parking time threshold by fitting a power law distribution on a double logarithmic axis, wherein the first parking time threshold is greater than the second parking time threshold;

[0027] Step S4-2: Determine whether the time difference between the last stop point and the first stop point in the vehicle stop point cluster exceeds the first stop time threshold. If yes, the vehicle stop point cluster is a valid stop point set. Otherwise, execute step S4-3:

[0028] Step S4-3: Determine whether the time difference between the last stop point and the first stop point in the vehicle stop point cluster exceeds a second stop time threshold. If yes, the vehicle stop point cluster is a pending stop point set.

[0029] In the step S4-1, in the process of fitting the power law distribution on the double logarithmic axis, it is grouped into four sections.

[0030] The step S5 comprises:

[0031] Step S5-1: for each pending stop point set, determine whether the charging status of each stop point is charging or charging completed. If yes, calculate whether the remaining power percentage of the last stop point is greater than the remaining power percentage of the first stop point. If yes, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-2.

[0032] Step S5-2: Calculate the turning angle of each pending stop point set respectively. If the turning angle is greater than a pre-configured angle threshold, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-3.

[0033] Step S5-3: Based on the trip chain analysis, determine whether the pending stop point set is the starting point or the end point of the trip chain. If yes, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-4.

[0034] Step S5-4: Determine whether the pending stop point set is located after the valid stop point set for charging. If yes, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-5:

[0035] Step S5-5: Determine whether the location of the pending stop point set is identified as a valid stop point set in historical determinations. If yes, mark the pending stop point set as a valid stop point set.

[0036] The calculation process of the turning angle of the pending stop point set includes:

[0037] Step S5-2-1: Determine the actual stop position of the pending stop point set;

[0038] Step S5-2-2: Based on the position of the pending stop point set, extracting from the target vehicle positioning data sequence a front moving point cluster located before and adjacent to the pending stop point set, and a rear moving point cluster located after and adjacent to the pending stop point set;

[0039] Step S5-2-3: Based on the front moving point cluster and the rear moving point cluster, combined with the actual stop position of the to-be-determined stop point set, determine the turning angle of the to-be-determined stop point set:

[0040]

[0041] Where: α(s c ) is the turning angle of the set of pending stop points, K is the number of single moving points selected for calculation, and w k is the weight of the moving point, Forward moving point m i-k , actual stop position, rear moving point m I+k The turning angle, dis(m i-k ,m I+k ) is the forward moving point m i-k and then move point m I+k , I1 is the serial number of the last rear moving point in the rear moving point cluster, I is the serial number of the last stopping point in the pending stopping point set, and I0 is the serial number of the first rear moving point in the front moving point cluster.

[0042] The step S5-2-1 specifically includes:

[0043] Step S5-2-1-1: Determine whether the number of stop points in the pending stop point set is 1 or 2. If yes, calculate the average longitude and latitude based on all the stop points based on time weighting as the actual stop position of the pending stop point set. Otherwise, execute step S5-2-1-2;

[0044] Step S5-2-1-2: determine whether the number of the stop points in the pending stop point set is 3. If yes, take the longitude and latitude of the second stop point as the actual stop position of the pending stop point set. Otherwise, execute step S5-2-1-3.

[0045] Step S5-2-1-3: Based on the second to the second to last stop point in the set of pending stop points, calculate the average longitude and latitude based on time weighting as the actual stop position of the set of pending stop points;

[0046] In the steps S5-2-1-1 and S5-2-1-3, the weight of the kth stopping point in the set of pending stopping points participating in the calculation is:

[0047]

[0048] Where: t k+1 is the timestamp of the k+1th stop point, t k is the timestamp of the kth stop point, The timestamp of the last stop point in the stop point set. The timestamp of the first stop point in the stop point set.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] 1. The charging demand is taken into account, so that the effective stopping point clusters of charging demand can be identified, which improves the accuracy and information content of the identification.

[0051] 2. Based on the above specially designed turning angles and other criteria, the effective stop point clusters are identified. In this way, when facing situations such as overpass areas and poor network of vehicle positioning devices that may cause excessive data errors, interference from local smaller errors can be avoided, thereby improving the identification accuracy of the effective stop point clusters. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a schematic flow chart of the main steps of the method of the present invention;

[0053] Figure 2 A curve diagram of a method for identifying parking purposes of urban green freight new energy vehicles driven by data in this application;

[0054] Figure 3 This is a schematic diagram of the distance between two adjacent trajectory points in this application;

[0055] Figure 4 Schematic diagram of logarithmic distribution, where (a) is the schematic diagram of logarithmic distribution of distances between adjacent tracks, and (b) is the schematic diagram of logarithmic distribution of average speeds of adjacent tracks;

[0056] Figure 5 This is a schematic diagram of the cumulative stop time distribution of the stop point cluster for this application;

[0057] Figure 6 This is a process tree diagram for identifying effective stopping points for freight new energy vehicles based on the trip chain for this application;

[0058] Figure 7 Schematic diagram of actual stop position recognition, wherein (a) is a schematic diagram of stop point recognition based on no turning angle; (b) is a schematic diagram of stop point recognition based on a turning angle; (c) is a schematic diagram of distance-weighted turning angle calculation;

[0059] Figure 8 It is a schematic diagram of the turning angle distribution of the stop point cluster of this application. DETAILED DESCRIPTION

[0060] 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.

[0061] Explanation of professional terms in this application:

[0062] (1) Vehicle trajectory: The vehicle’s positioning trajectory on day d (generally using GPS data, but also using data from Beidou or other positioning systems) is represented by G d {(x1,y1,t1),(x2,y2,t2),...,(x i ,y i ,t i ),...,(x I ,y I ,t I )}, t1<t2,...,<t I , t I -t1=T d . Where (x i ,y i ,t i ) is t i The GPS track point at time, denoted as p i ;x i ,y i t i The latitude and longitude at the time; I represents the number of trajectory points; T d represents the time span of the dth day, d = 1, 2, ..., D, where D is the number of days of trajectory data;

[0063] (2) Moving distance: trajectory point p i and p i+1The geographical distance between i+1 It is calculated by formula (1), where R is the average radius of the earth, which is about 6371 km;

[0064] d i+1 =R·arc cos(sin(y i )·sin(y i+1 )+cos(y i )·cos(y i+1 )·cos(x i+1 -x i )) (1)

[0065] (3) Average velocity: p of two adjacent trajectory points i and p i+1 The average speed between i+1 =d i+1 / (t i+1 -t i );

[0066] (4) Stopping point and stopping cluster: If the vehicle is at t i The moment is stopped, then the stopping point s i Defined as Stop cluster is defined as S i ={s i ,s i+1 ,...,s I}, The number of trajectory points included is

[0067] (5) Stop point set: the set of all stop point clusters;

[0068] (6) Mobile points and mobile clusters: If the vehicle is at t i The moment is moving, then the moving point m i Defined as The moving point cluster is defined as M i ={m i ,m i+1 ,...,m I}, The number of trajectory points included is

[0069] (7) Cumulative moving distance: the sum of moving distances of trajectory points in the moving point cluster;

[0070] (8) Stop point cluster range: the maximum value of the pairwise stop point distances in the stop point cluster;

[0071] (9) Actual stop position: the representative point in the stop point cluster, which may be the first or last stop point in the stop point cluster;

[0072] (10) Stop accumulating time: Stop cluster S i The cumulative stop time is expressed as

[0073] (11) Turning angle: p i The turning angle at is defined as in Calculates the angle of trajectory points.

[0074] A data-driven method for identifying parking purposes of urban green freight new energy vehicles, such as Figure 1 and Figure 2 As shown, including:

[0075] Step S1: obtaining an original vehicle positioning data sequence, wherein each track point in the vehicle positioning data sequence includes latitude and longitude, timestamp, remaining power percentage, charging state and instantaneous speed;

[0076] The specific data format is shown in Table 1:

[0077] Table 1

[0078]

[0079]

[0080] Among them, the charging status Charge Status is 1, 3, and 4, respectively indicating parking charging, not charging, and charging completion, and PDATime indicates a timestamp.

[0081] Step S2: preprocessing and repairing the acquired original positioning data to obtain one or more target vehicle positioning data sequences, including:

[0082] Step S2-1: Delete all data points with longitude or latitude of 0 in the vehicle positioning data sequence, and data points that exceed the longitude threshold range or latitude threshold range;

[0083] Step S2-2: Delete all data points whose timestamps do not change continuously in the vehicle positioning data sequence;

[0084] Step S2-3: Calculate the average speed between adjacent data points respectively. If the average speed of adjacent data points is greater than the pre-configured maximum speed, delete the corresponding data point.

[0085] Step S2-4: When the time stamp difference between adjacent data points exceeds a preconfigured time threshold, the vehicle positioning data sequence is segmented to obtain multiple target vehicle positioning data sequences.

[0086] Step S3: Based on the target vehicle positioning data sequence, all vehicle stop points are screened out by comparing the distance and average speed of adjacent trajectory points with the pre-configured distance threshold and average speed threshold, and the continuous vehicle stop points are constructed into a vehicle stop point cluster, including:

[0087] Step S3-1: configure the distance threshold and the average speed threshold;

[0088] Among them, Figure 4 As shown in part (a) of FIG. 1 , the distance threshold is determined by: based on the target vehicle positioning data sequence, a distance logarithmic distribution curve of two adjacent trajectory data points is constructed, and the logarithmic distribution curve presents a bimodal mixed distribution. The target point of the mixed distribution is used as the distance threshold. In this embodiment, the distance threshold L0=14.7m,

[0089] like Figure 4 As shown in part (b), the average speed threshold is determined by the following method: based on the target vehicle positioning data sequence, a logarithmic distribution curve of the average speed of two adjacent trajectory data points is constructed, and the logarithmic distribution curve presents a bimodal mixed distribution. The target point of the mixed distribution is used as the average speed threshold. In this embodiment, the average speed threshold V0=1.33km / h.

[0090] Step S3-2: Figure 3 As shown, the distance and average speed between each adjacent data point in the target vehicle positioning data sequence are calculated respectively. If the distance between adjacent data points is less than the distance threshold, and the average speed is less than the average speed threshold, the previous data point among the adjacent data points is determined as the vehicle stop point;

[0091] Step S3-3: Construct all consecutive vehicle stop points into a vehicle stop point cluster.

[0092] Step S4: judging whether each stop point set is a valid stop point set or a pending stop point set based on the time difference between the last stop point and the first stop point in the constructed stop point set, including:

[0093] Step S4-1: Determine a first parking time threshold and a second parking time threshold by fitting a power law distribution on a double logarithmic axis, wherein the first parking time threshold is greater than the second parking time threshold;

[0094] Step S4-2: Determine whether the time difference between the last stop point and the first stop point in the vehicle stop point cluster exceeds the first stop time threshold. If yes, the vehicle stop point cluster is a valid stop point set. Otherwise, execute step S4-3:

[0095] Step S4-3: Determine whether the time difference between the last stop point and the first stop point in the vehicle stop point cluster exceeds a second stop time threshold. If yes, the vehicle stop point cluster is a pending stop point set.

[0096] like Figure 5 As shown, in step S4-1, in the process of fitting the power law distribution on the double logarithmic axis, it is grouped into four segments.

[0097] The cumulative stop time of each section and the method for identifying the effective stop point are as follows:

[0098] In the first segment, the cumulative stop time is less than 2.16 minutes. It is determined that the broken trajectory data is a short stop point caused by waiting for traffic lights and traffic congestion, not a valid stop point, and is excluded;

[0099] In the second section, the cumulative stop time is greater than or equal to 2.16 minutes and less than 55.44 minutes. It is judged to contain invalid stop points and valid stop points and is recorded as a short stop;

[0100] The third segment, where the cumulative stop time was greater than 55.44 minutes and less than 7.35 hours, was judged to include invalid stops, temporary stops, and stops at the end of the trip, and was recorded as a mid-term stop;

[0101] In the fourth section, the cumulative stop time is greater than 7.35 hours, which is considered a valid stop point.

[0102] That is, the second parking time threshold is 55.44 minutes, and the first parking time threshold is 7.35 hours.

[0103] In addition, in this embodiment, in order to eliminate the stop point clusters with too large range, the maximum distance between each stop point in the stop point cluster is counted. In this application, the maximum distance of the stop points in 99% of the stop point clusters is less than 312.36 meters. Therefore, the stop point clusters with a range greater than 312.36 meters are marked as invalid stop points and are excluded.

[0104] Step S5: For each undetermined stop point set, combine the turning angle and the trip chain analysis to determine whether each stop point set is a valid stop point set. Figure 6 As shown, including:

[0105] Step S5-1: for each pending stop point set, determine whether the charging status of each stop point is charging or charging completed. If yes, calculate whether the remaining power percentage of the last stop point is greater than the remaining power percentage of the first stop point. If yes, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-2.

[0106] Step S5-2: Calculate the turning angle of each pending stop point set respectively. If the turning angle is greater than a pre-configured angle threshold, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-3.

[0107] like Figure 7 As shown, the calculation process of the turning angle of the pending stop point set includes:

[0108] Step S5-2-1: Determine the actual stop position of the pending stop point set, specifically including:

[0109] Step S5-2-1-1: Determine whether the number of stop points in the pending stop point set is 1 or 2. If yes, calculate the average longitude and latitude based on all the stop points based on time weighting as the actual stop position of the pending stop point set. Otherwise, execute step S5-2-1-2;

[0110] Step S5-2-1-2: determine whether the number of the stop points in the pending stop point set is 3. If yes, take the longitude and latitude of the second stop point as the actual stop position of the pending stop point set. Otherwise, execute step S5-2-1-3.

[0111] Step S5-2-1-3: Based on the second to the second to last stop point in the set of pending stop points, calculate the average longitude and latitude based on time weighting as the actual stop position of the set of pending stop points;

[0112] In step S5-2-1-1 and step S5-2-1-3, the weight of the kth stop point in the set of pending stop points participating in the calculation is:

[0113]

[0114] Where: t k+1 is the timestamp of the k+1th stop point, t k is the timestamp of the kth stop point, The timestamp of the last stop point in the stop point set. The timestamp of the first stop point in the stop point set.

[0115] The specific position calculation formula is shown in formula (2):

[0116]

[0117] Where: x c and are the latitude and longitude of the actual docking location, is the number of stopping points.

[0118] Step S5-2-2: Based on the position of the pending stop point set, extracting from the target vehicle positioning data sequence a front moving point cluster located before and adjacent to the pending stop point set, and a rear moving point cluster located after and adjacent to the pending stop point set;

[0119] Step S5-2-3: Determine the turning angle of the pending stop point set based on the front moving point cluster and the rear moving point cluster and the actual stop position of the pending stop point set

[0120] The turning angle at the cluster of the stopping points is calculated based on the actual stopping position of the truck and the front and rear moving points. Due to the influence of data recording time and vehicle speed, the front and rear moving points may be far away from the actual parking position of the truck. To solve this problem, the distance weighted method is used to calculate the turning angle:

[0121]

[0122] Where: α(s c ) is the turning angle of the set of pending stop points, K is the number of single moving points selected for calculation, and w k is the weight of the moving point, Forward moving point m i-k , actual stop position, rear moving point m I+k The turning angle, dis(m i-k ,m I+k ) is the forward moving point m i-k and then move point m I+k , I1 is the serial number of the last rear moving point in the rear moving point cluster, I is the serial number of the last stopping point in the pending stopping point set, and I0 is the serial number of the first rear moving point in the front moving point cluster.

[0123] Among them, the larger the distance between the paired moving points, the smaller the weight

[0124] like Figure 7 As shown in (c), for each pair of turning angles calculated for the forward and backward moving points, the larger the distance between them, the less credible they are and the smaller their weights.

[0125] like Figure 8 As shown, the turning angle at each stop point cluster is modeled and its probability distribution is shown. The distribution is interrupted at 84.45°, each part represents a different turning field, and the angle threshold is determined to be 84.45°.

[0126] Step S5-3: Based on the trip chain analysis, determine whether the pending stop point set is the starting point or the end point of the trip chain. If yes, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-4.

[0127] Step S5-4: Determine whether the pending stop point set is located after the valid stop point set for charging. If yes, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-5:

[0128] Step S5-5: Determine whether the location of the pending stop point set is identified as a valid stop point set in historical determinations. If yes, mark the pending stop point set as a valid stop point set.

[0129] Step S6: Determine the parking purpose of all valid stop point sets.

[0130] Specifically, on the one hand, if the remaining power percentage of the last stopping point in the valid stopping point set is greater than the remaining power percentage of the first stopping point, and the charging status is charging or charging completed, then the charging demand is located for the parking purpose of the valid stopping point set, and other parking purposes can be determined in combination with existing methods.

[0131] 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 data-driven method for identifying parking purposes of urban green freight new energy vehicles, characterized in that: include: Step S1: obtaining an original vehicle positioning data sequence, wherein each track point in the vehicle positioning data sequence includes latitude and longitude, timestamp, remaining power percentage, charging state and instantaneous speed; Step S2: preprocessing and repairing the acquired original positioning data to obtain one or more target vehicle positioning data sequences; Step S3: Based on the target vehicle positioning data sequence, all vehicle stop points are screened out by comparing the distance and average speed of adjacent trajectory points with the pre-configured distance threshold and average speed threshold, and the continuous vehicle stop points are constructed into a vehicle stop point cluster; Step S4: judging whether each stop point set is a valid stop point set or a pending stop point set based on the time difference between the last stop point and the first stop point in the constructed stop point set; Step S5: for each undetermined stop point set, combining the turning angle and the trip chain analysis, determine whether each stop point set is a valid stop point set; Step S6: Determine the parking purpose of all valid stop point sets.

2. According to claim 1, a data-driven urban green freight new energy vehicle parking purpose identification method is characterized in that: The step S2 comprises: Step S2-1: Delete all data points with longitude or latitude of 0 in the vehicle positioning data sequence, and data points that exceed the longitude threshold range or latitude threshold range; Step S2-2: Delete all data points whose timestamps do not change continuously in the vehicle positioning data sequence; Step S2-3: Calculate the average speed between adjacent data points respectively. If the average speed of adjacent data points is greater than the pre-configured maximum speed, delete the corresponding data point. Step S2-4: When the time stamp difference between adjacent data points exceeds a preconfigured time threshold, the vehicle positioning data sequence is segmented to obtain multiple target vehicle positioning data sequences.

3. According to claim 1, a data-driven urban green freight new energy vehicle parking purpose identification method is characterized in that: The step S3 comprises: Step S3-1: configure the distance threshold and the average speed threshold; Step S3-2: respectively calculating the distance and average speed between each adjacent data point in the target vehicle positioning data sequence; if the distance between adjacent data points is less than the distance threshold, and the average speed is less than the average speed threshold, then determining the previous data point among the adjacent data points as the vehicle stop point; Step S3-3: Construct all consecutive vehicle stop points into a vehicle stop point cluster.

4. A data-driven urban green freight new energy vehicle parking purpose identification method according to claim 3, characterized in that: The distance threshold is determined by: constructing a distance logarithmic distribution curve of two adjacent trajectory data points based on the target vehicle positioning data sequence, the logarithmic distribution curve presents a bimodal mixed distribution, and taking the target point of the mixed distribution as the distance threshold; The average speed threshold is determined in the following manner: based on the target vehicle positioning data sequence, a logarithmic distribution curve of the average speed of two adjacent trajectory data points is constructed, the logarithmic distribution curve presents a bimodal mixed distribution, and the target point of the mixed distribution is used as the average speed threshold.

5. According to claim 1, a data-driven urban green freight new energy vehicle parking purpose identification method is characterized in that: The step S4 comprises: Step S4-1: Determine a first parking time threshold and a second parking time threshold by fitting a power law distribution on a double logarithmic axis, wherein the first parking time threshold is greater than the second parking time threshold; Step S4-2: Determine whether the time difference between the last stop point and the first stop point in the vehicle stop point cluster exceeds the first stop time threshold. If yes, the vehicle stop point cluster is a valid stop point set. Otherwise, execute step S4-3: Step S4-3: Determine whether the time difference between the last stop point and the first stop point in the vehicle stop point cluster exceeds a second stop time threshold. If yes, the vehicle stop point cluster is a pending stop point set.

6. A data-driven urban green freight new energy vehicle parking purpose identification method according to claim 5, characterized in that: In the step S4-1, in the process of fitting the power law distribution on the double logarithmic axis, it is grouped into four sections.

7. A data-driven urban green freight new energy vehicle parking purpose identification method according to claim 5, characterized in that: The step S5 comprises: Step S5-1: for each pending stop point set, determine whether the charging status of each stop point is charging or charging completed. If yes, calculate whether the remaining power percentage of the last stop point is greater than the remaining power percentage of the first stop point. If yes, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-2. Step S5-2: Calculate the turning angle of each pending stop point set respectively. If the turning angle is greater than a pre-configured angle threshold, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-3. Step S5-3: Based on the trip chain analysis, determine whether the pending stop point set is the starting point or the end point of the trip chain. If yes, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-4. Step S5-4: Determine whether the pending stop point set is located after the valid stop point set for charging. If yes, mark the pending stop point set as a valid stop point set. Otherwise, execute step S5-5: Step S5-5: Determine whether the location of the pending stop point set is identified as a valid stop point set in historical determinations. If yes, mark the pending stop point set as a valid stop point set.

8. A data-driven urban green freight new energy vehicle parking purpose identification method according to claim 7, characterized in that: The calculation process of the turning angle of the pending stop point set includes: Step S5-2-1: Determine the actual stop position of the pending stop point set; Step S5-2-2: Based on the position of the pending stop point set, extracting from the target vehicle positioning data sequence a front moving point cluster located before and adjacent to the pending stop point set, and a rear moving point cluster located after and adjacent to the pending stop point set; Step S5-2-3: Based on the front moving point cluster and the rear moving point cluster, combined with the actual stop position of the to-be-determined stop point set, determine the turning angle of the to-be-determined stop point set: K=min{I1-I,I0} Where: α(s c ) is the turning angle of the set of pending stop points, K is the number of single moving points selected for calculation, and w k is the weight of the moving point, Forward moving point m i-k , actual stop position, rear moving point m I+k The turning angle, dis(m i-k ,m I+k ) is the forward moving point m i-k and then move point m I+k , I1 is the serial number of the last rear moving point in the rear moving point cluster, I is the serial number of the last stopping point in the pending stopping point set, and I0 is the serial number of the first rear moving point in the front moving point cluster.

9. A data-driven urban green freight new energy vehicle parking purpose identification method according to claim 8, characterized in that: The step S5-2-1 specifically includes: Step S5-2-1-1: Determine whether the number of stop points in the pending stop point set is 1 or 2. If yes, calculate the average longitude and latitude based on all the stop points based on time weighting as the actual stop position of the pending stop point set. Otherwise, execute step S5-2-1-2; Step S5-2-1-2: determine whether the number of the stop points in the pending stop point set is 3. If yes, take the longitude and latitude of the second stop point as the actual stop position of the pending stop point set. Otherwise, execute step S5-2-1-3. Step S5-2-1-3: Based on the second to the second to last stop points in the set of pending stop points, calculate the average longitude and latitude based on time weighting as the actual stop position of the set of pending stop points.

10. A data-driven urban green freight new energy vehicle parking purpose identification method according to claim 9, characterized in that: In the steps S5-2-1-1 and S5-2-1-3, the weight of the kth stop point in the set of pending stop points involved in the calculation is: Where: t k+1 is the timestamp of the k+1th stop point, t k is the timestamp of the kth stop point, The timestamp of the last stop point in the stop point set. The timestamp of the first stop point in the stop point set.

Images