Urban distribution vehicle parking point identification method and device based on trajectory data

Abstract

The invention relates to an urban distribution vehicle parking point identification method and device based on trajectory data, and the method comprises the steps: S1, obtaining all vehicle trajectory point data, and carrying out the vehicle merging, time sorting and abnormal data correction of the vehicle trajectory point data, and obtaining a trajectory point data sequence of each vehicle; s2, on the basis of the track point data sequence of each vehicle, screening out track points with instantaneous speeds smaller than a first speed threshold value as instantaneous stop points; s3, screening out instantaneous stay points of which the stay time is greater than a first time threshold value as effective stay points; s4, main road staying points and non-freight staying points are removed from the effective staying points; and S5, clustering all the remaining effective stop points to obtain a plurality of potential freight temporary stop points. Compared with the prior art, the method has the advantages of being capable of relieving traffic jam caused by freight and the like.

Description

Technical Field

[0001] The present invention relates to the field of transportation, and in particular to a method and device for identifying a stop point of an urban distribution vehicle based on trajectory data. Background Art

[0002] Urban freight delivery vehicles (hereinafter referred to as "urban delivery vehicles") have become a significant factor in aggravating urban traffic congestion and environmental pollution while improving delivery efficiency. With the continuous growth of urban delivery demand, a large number of urban delivery vehicles frequently shuttle between busy urban center areas, which not only significantly increases the burden on the transportation system, but also poses a potential threat to the sustainability of the urban environment. In addition, urban delivery vehicles face severe parking problems. The lack of legal parking spots forces vehicles to adopt illegal or inappropriate parking behaviors. This not only violates traffic regulations, but also aggravates traffic congestion, thus having a negative impact on the city's traffic flow and the quality of life of residents.

[0003] Especially in the practice of green freight distribution demonstration projects, the innovative concept of "temporary stopping points for urban distribution vehicles" was proposed. These recommended stopping points are mainly for urban distribution vehicles to stop briefly and load and unload goods. Most of them are located on the road. The purpose is to alleviate the parking problem in the central area of ​​the city, while improving the efficiency of urban freight distribution and its environmental friendliness.

[0004] There are many limitations in the existing technology for urban freight distribution research. For example, some studies have mostly planted freight travel research on tourist travel research and borrowed from tourist travel behavior patterns. This method ignores the fundamental differences between freight and passenger travel in terms of purpose and frequency. For example, passenger travel patterns usually focus on commuting and non-commuting travel needs, while freight travel focuses more on transportation efficiency and cost. In addition, freight behavior is directly affected by business operation rules and market demand, which is very different from the contingency and personal selectivity of passenger travel. Therefore, directly applying the passenger travel model to the freight scene often cannot accurately depict the actual driving characteristics and transportation needs of freight vehicles, and more attention should be paid to the analysis of freight travel characteristics. Some existing technologies also provide some stop point identification involving urban freight, but although the stop point identification of such existing technologies can better reflect long-term parking needs and fixed parking needs, it cannot reflect the temporary parking needs in freight, and cannot alleviate the road congestion problem caused by temporary parking of freight vehicles in cities. Summary of the invention

[0005] The purpose of the present invention is to provide a method and device for identifying stop points of urban distribution vehicles based on trajectory data.

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

[0007] A method for identifying a stop point of a city distribution vehicle based on trajectory data, comprising:

[0008] Step S1: Acquire all vehicle trajectory point data, and obtain the trajectory point data sequence of each vehicle by merging the vehicle, sorting the time and correcting the abnormal data;

[0009] Step S2: based on the trajectory point data sequence of each vehicle, selecting the trajectory points whose instantaneous speed is less than the first speed threshold as instantaneous stop points;

[0010] Step S3: Screening out instantaneous stay points whose stay time is greater than the first time threshold as valid stay points;

[0011] Step S4: Eliminate main road stop points and non-freight stop points from the valid stop points;

[0012] Step S5: Cluster all remaining valid stop points to obtain multiple potential temporary freight stop points.

[0013] The process of abnormal data correction in step S1 includes:

[0014] Step S1-1: Delete duplicate trajectory points;

[0015] Step S1-2: if the time difference between adjacent trajectory points is less than the first set time length, a new trajectory point is inserted between the two trajectory points using a linear interpolation method;

[0016] Step S1-3: Delete all trajectory points whose speed is within the first speed threshold range;

[0017] Step S1-4: Delete all trajectory points whose speed is not 0 and whose longitude and latitude have not changed;

[0018] Step S1-5: Delete all trajectory points with zero speed and abnormal longitude and latitude changes.

[0019] The step S1-5 comprises:

[0020] Step S1-5-1: if the speed of the track point is 0, calculate the distance between the track point and the track point before and after the track point, if the distance is greater than the first threshold, delete the track point, otherwise, execute step S1-5-2;

[0021] Step S1-5-2: Calculate the average speed between the front and rear trajectory points as the speed of the trajectory point.

[0022] The process of determining the first speed threshold comprises:

[0023] According to the trajectory point data sequence of the vehicle, the first mixed distribution model based on speed is used for fitting to obtain the speed of the saddle point;

[0024] The velocity of the saddle point is taken as the first velocity threshold.

[0025] The mathematical expression of the first mixed distribution model is:

[0026]

[0027] Where: f(v; ω1, μ1, σ1, ω2, μ2, σ2) is the first mixed distribution model, v is the velocity, ω1 is the weight of the first log-normal distribution, ω2 is the weight of the first normal distribution, μ1 is the mean of the first log-normal distribution, σ1 is the standard deviation of the first log-normal distribution, μ2 is the mean of the first normal distribution, σ2 is the standard deviation of the first normal distribution, Lognorm(v; μ1, σ1) is, Norm(v; μ2, σ2) is the first normal distribution.

[0028] The first time threshold is determined by the cumulative probability density of the vehicle stay time.

[0029] The first time threshold is the stay time at which the cumulative probability density of the vehicle stay time is 80%.

[0030] In the step S3, after the instantaneous stay points are merged, the instantaneous stay points whose stay time is greater than the first time threshold are screened out as valid stay points;

[0031] The merging process includes:

[0032] Step S3-1: Determine whether the distance between the stop point and the previous stop point is less than a second set threshold distance. If yes, execute step S3-2:

[0033] Step S3-2: Determine whether the stay time of the stay point is less than a second set threshold time. If yes, merge the stay point with its previous stay point.

[0034] The step S4 comprises:

[0035] Step S4-1: Eliminate all valid stop points that are on the main road;

[0036] Step S4-2: for all remaining valid stay points, three points of interest closest to each valid stay point are obtained as associated points of interest for the valid stay point;

[0037] Step S4-3: extract keywords of the associated points of interest of each valid stop point. If the three associated points of interest of any valid stop point all contain non-freight keywords and none of them contain freight keywords, then the valid stop point is removed as a non-freight stop point.

[0038] A device for identifying stop points of urban distribution vehicles based on trajectory data, comprising a memory, a processor, and a program stored in the memory, wherein when the processor executes the program, the method described in any one of claims 1 to 9 is implemented.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] 1. The five-step method of vehicle trajectory data preprocessing, vehicle instantaneous stop point identification, vehicle effective stop point identification, vehicle freight business point identification, and vehicle potential freight temporary stop point screening can accurately obtain the data information of urban distribution vehicle stop points, providing key data support for the site selection and planning of temporary stop points for urban freight distribution.

[0041] 2. Vehicle trajectory preprocessing: By integrating and sorting vehicles in time, we can obtain the trajectory data information of each vehicle in chronological order. We can further identify and correct the four abnormal data in the sorted trajectory data: data duplication, data missing, data out of bounds, and vehicle speed data abnormality.

[0042] 3. A mixed distribution model consisting of lognormal distribution and normal distribution is used to fit the vehicle speed distribution, so as to identify the vehicle instantaneous stop point from the preprocessed vehicle trajectory data.

[0043] 4. In order to exclude non-freight stop points caused by traffic lights or short-term traffic jams, this application sets a time threshold through the cumulative probability distribution of vehicle stop time, and excludes trajectory data below the time threshold; further, in order to exclude repeated stop trajectory data, data points with similar vehicle stop time and stop space are merged to obtain the final valid stop point data.

[0044] 5. Urban distribution vehicles may stay on the road for a long time due to traffic congestion, and these stops are not necessarily freight business points. Therefore, it is necessary to further identify freight business points from valid stops. This application is divided into two steps. First, the stops on the main roads are eliminated. Secondly, POI is used to eliminate the stops located in service areas, gas stations and other non-freight activity locations. This application uses a simple method that does not rely on matching vehicle trajectories with urban road networks to eliminate non-business stops.

[0045] 6. This application takes into account that temporary stops mainly serve cargo loading and unloading activities in a short period of time, so the data of freight business points with a longer duration (for example, this application prefers 1 hour) are eliminated by setting a time threshold. Considering that this application is mainly for the identification of urban distribution vehicle stop points in specific urban areas, data points in industrial areas outside the city and within the city (where it is not easy to generate trajectory data of urban distribution vehicles) are excluded. Finally, the urban distribution vehicle stop point data that conforms to the actual situation of the city is obtained, providing key data support for the site selection and planning of temporary stop points for urban freight distribution. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 Classification of all data of a city for this application;

[0047] Figure 2 A curve diagram between the average speed and the probability density function of the pre-processed data of a city fitted by the mixed distribution model of this application;

[0048] Figure 3 This is the cumulative probability distribution curve of the stay time for this application;

[0049] Figure 4 This is a statistical table of the merged data points with similar vehicle stay time and stay space in this application;

[0050] Figure 5 A statistical table of freight business point data selected from valid stop point data for this application;

[0051] Figure 6 For this application, step S5 deletes the statistical table of data with dwell time longer than the threshold from the freight business hour;

[0052] Figure 7 For Figure 6 The remaining data is then removed from the statistical table of data outside the city;

[0053] Figure 8 For Figure 7 The remaining data is then eliminated from the statistical table of industrial areas;

[0054] Fig. 9 It is a schematic diagram of the main steps of the present invention. DETAILED DESCRIPTION

[0055] 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.

[0056] A method for identifying stop points of urban distribution vehicles based on trajectory data, such as Fig. 9 As shown, including:

[0057] Step S1: Acquire all vehicle trajectory point data, and obtain the trajectory point data sequence of each vehicle by merging the vehicle, sorting the time and correcting the abnormal data;

[0058] The format of vehicle trajectory point data includes a tuple consisting of license plate, instantaneous speed, total mileage, longitude, latitude, and reporting time information. The longitude and latitude provide the spatial location information of the vehicle, the reporting time provides the time information of the vehicle, and the instantaneous speed and total mileage reflect the driving status of the vehicle. The details are shown in Table 1:

[0059] Table 1

[0060]

[0061] The trajectory data of each vehicle is sorted in chronological order, as shown in Table 2:

[0062] Table 2

[0063] PlateNum Speed TotalDis Ln Lat PDATime License Plate Instantaneous speed (km / h) Total mileage (km) longitude latitude Reporting time ED004XX 16.7 37252 120.6306 31.13101 2022 / 5 / 1 6:53 ED004XX 13.5 37252 120.6306 31.13101 2022 / 5 / 1 6:54 ED004XX 0 37252 120.6306 31.13101 2022 / 5 / 1 6:55

[0064] According to the formats shown in Tables 1 and 2, the vehicle trajectory data are integrated and sorted in chronological order. Then, the abnormal vehicle data is identified and corrected. The correction methods mainly include processing four situations: data duplication, data missing, data out of bounds, and abnormal vehicle speed data.

[0065] like Figure 1 As shown in FIG. 1 , the classification of all data in a city includes 91,984,192 normal data, 12,467,985 corrected data, and 389,553 deleted data. The specific process of abnormal data correction in step S1 includes:

[0066] Step S1-1: Delete duplicate trajectory points;

[0067] Step S1-2: if the time difference between adjacent track points is less than the first set time length, a new track point is inserted between the two track points using a linear interpolation method. In this embodiment, the first set time length is 3 minutes;

[0068] Step S1-3: deleting all trajectory points whose speeds are within a first speed threshold range. In the embodiment, the first speed threshold range is 0-150 km / h;

[0069] Step S1-4: Delete all trajectory points whose speed is not 0 and whose longitude and latitude have not changed;

[0070] Step S1-5: Delete all trajectory points with zero speed and abnormal longitude and latitude changes, including:

[0071] Step S1-5-1: if the speed of the track point is 0, calculate the distance between the track point and the track point before and after the track point, if the distance is greater than the first threshold, delete the track point, otherwise, execute step S1-5-2;

[0072] Step S1-5-2: Calculate the average speed between the front and rear trajectory points as the speed of the trajectory point.

[0073] In this embodiment, the distance between two estimated points is calculated by longitude and latitude, and the mathematical formula is:

[0074]

[0075] d=R*c

[0076] Where: ΔLng is the latitude difference, ΔLat is the longitude difference, R is the radius of the earth, and d is the distance.

[0077] Step S2: based on the trajectory point data sequence of each vehicle, selecting the trajectory points whose instantaneous speed is less than the first speed threshold as instantaneous stop points;

[0078] The process of determining the first speed threshold includes:

[0079] According to the trajectory point data sequence of the vehicle, the first mixed distribution model based on speed is used for fitting to obtain the speed of the saddle point;

[0080] The velocity of the saddle point is taken as the first velocity threshold.

[0081] The mathematical expression of the first mixed distribution model is:

[0082]

[0083] Where: f(v; ω1, μ1, σ1, ω2, μ2, σ2) is the first mixed distribution model, v is the velocity, ω1 is the weight of the first log-normal distribution, ω2 is the weight of the first normal distribution, μ1 is the mean of the first log-normal distribution, σ1 is the standard deviation of the first log-normal distribution, μ2 is the mean of the first normal distribution, σ2 is the standard deviation of the first normal distribution, Lognorm(v; μ1, σ1) is, Norm(v; μ2, σ2) is the first normal distribution.

[0084] The first logarithmic normal distribution represents the data deviation caused by the drift of trajectory data when the vehicle is stationary, and the first normal distribution represents the trajectory data when the vehicle is in a normal driving state on the road;

[0085] According to the first mixed distribution model Figure 1 The vehicle trajectory data of a certain city is fitted, and the results are shown in Table 5:

[0086] Table 5

[0087] parameter Estimated value <![CDATA[ω1]]> 0.00267 <![CDATA[μ1]]> 8.79287 <![CDATA[σ1]]> 13.37820 <![CDATA[μ2]]> 13.94185 <![CDATA[σ2]]> 6.63715

[0088] like Figure 2 As shown in the figure, it is a curve diagram between the average speed and the probability density function of the pre-processed data of a city according to the mixed distribution model. Figure 2 It can be seen that the fitting parameters show that the average speed follows a mixed distribution including a main normal distribution component (large proportion) and an almost negligible lognormal distribution component (small proportion), and the speed at the saddle point (i.e., the local minimum point of the probability density function) is 0.16 km / h. This speed is used as the speed threshold of the instantaneous stop point, and the trajectory points with an average speed less than the speed threshold are regarded as instantaneous stop points.

[0089] Step S3: Screening out instantaneous stay points whose stay time is greater than the first time threshold as valid stay points;

[0090] The first time threshold is determined by the cumulative probability density of the vehicle's stay time. In this embodiment, the first time threshold is the stay time at which the cumulative probability density of the vehicle's stay time is 80%.

[0091] In step S3, after the instantaneous stay points are merged, the instantaneous stay points whose stay time is greater than the first time threshold are screened out as valid stay points;

[0092] The merging process includes:

[0093] Step S3-1: Determine whether the distance between the stop point and the previous stop point is less than a second set threshold distance. If yes, execute step S3-2:

[0094] Step S3-2: Determine whether the stay time of the stay point is less than a second set threshold time. If yes, merge the stay point with its previous stay point.

[0095] like Figure 3 As shown in the figure, it is the cumulative probability distribution of the residence time, which is given by Figure 3 It can be seen that 80% of the stop times are less than 9.7 minutes, the cumulative distribution slope corresponding to the stop time of less than 10 minutes is large, and the slope of the stop time of the stop time of more than 10 minutes is significantly reduced. In view of the fact that there are many short stops in the actual driving of urban distribution vehicles, and the setting of the time threshold of non-freight activities in previous studies, this application regards the stop points within 10 minutes as non-freight stop points and excludes them, and the remaining ones are recorded as the initial effective stop points of the vehicle;

[0096] like Figure 4As shown, in order to further exclude data information of repeated stay phenomena, data points with similar vehicle stay time and stay space are merged to obtain the final valid stay point data; the merging rule for data points with similar vehicle stay time and stay space is that if the distance between adjacent instantaneous stay points is less than the distance threshold of 500 meters and the stay time interval does not exceed the time threshold of 10 minutes, the latter instantaneous stay point is merged into the previous one, and the moment of the first instantaneous stay point is set as the start stay time, and the moment of the last instantaneous stay point is set as the end stay time, and the stay time is calculated in minutes. If the instantaneous stay point still meets the merging conditions after merging, it will continue to be merged until the time and distance thresholds are no longer met. After merging the adjacent stay points in time and space distance, this application initially obtained 1,126,968 valid stay point data.

[0097] Step S4: Eliminate main road stop points and non-freight stop points from the valid stop points, including:

[0098] Step S4-1: Eliminate all valid stop points that are on the main road;

[0099] Step S4-2: for all remaining valid stay points, three points of interest closest to each valid stay point are obtained as associated points of interest for the valid stay point;

[0100] Step S4-3: extract keywords of the associated points of interest of each valid stop point. If the three associated points of interest of any valid stop point all contain non-freight keywords and none of them contain freight keywords, then the valid stop point is removed as a non-freight stop point.

[0101] The freight keyword is equivalent to the reserved keyword, and the non-freight keyword is equivalent to the excluded keyword. In this embodiment, the specific details are shown in Table 3:

[0102]

[0103] The final results are shown in Table 4

[0104] Table 4

[0105]

[0106]

[0107] like Figure 5 As shown, it is a data schematic diagram after excluding the stay points on the main roads and the stay points located at gas stations, service areas, auto repair and other non-freight activity places from the 1,126,968 valid stay point data initially obtained in step S3, and using POI to eliminate non-freight business data, about 991,705 freight business points are obtained, which will serve as an important basis for the subsequent extraction of the urban distribution vehicle travel chain.

[0108] Step S5: Cluster all remaining valid stop points to obtain multiple potential temporary freight stop points.

[0109] like Figure 6-8 As shown, the data with a stay time of more than 1 hour is removed from the vehicle freight business point data obtained in step S4; 496,524 data are left, and 51,562 data outside the study city are removed from the 496,524 data, leaving 444,962 data, and 252,915 industrial land stop points are deleted from the 444,962 data, and finally a total of 192,047 potential temporary stop points located in non-industrial land are obtained. These 192,047 data are the potential temporary stop point data for vehicles.

[0110] From the above, it can be seen that this application has a total of more than 110 million trajectory data from 6152 urban freight distribution vehicles. After the five-step method of vehicle trajectory data preprocessing, vehicle instantaneous stop point identification, vehicle effective stop point identification, vehicle freight business point identification, and vehicle potential temporary stop point screening, it finally obtained 192047 data information that can reflect the temporary stop points of urban distribution vehicles in the urban area, providing key data support for the site selection and planning of temporary stop points for urban freight distribution.

[0111] 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 method for identifying stop points of urban distribution vehicles based on trajectory data, characterized in that: include: Step S1: Acquire all vehicle trajectory point data, and obtain the trajectory point data sequence of each vehicle by merging the vehicle, sorting the time and correcting the abnormal data; Step S2: based on the trajectory point data sequence of each vehicle, selecting the trajectory points whose instantaneous speed is less than the first speed threshold as instantaneous stop points; Step S3: Screening out instantaneous stay points whose stay time is greater than the first time threshold as valid stay points; Step S4: Eliminate main road stop points and non-freight stop points from the valid stop points; Step S5: Cluster all remaining valid stop points to obtain multiple potential temporary freight stop points.

2. A method for identifying stop points of urban distribution vehicles based on trajectory data according to claim 1, characterized in that: The process of abnormal data correction in step S1 includes: Step S1-1: Delete duplicate trajectory points; Step S1-2: if the time difference between adjacent trajectory points is less than the first set time length, a new trajectory point is inserted between the two trajectory points using a linear interpolation method; Step S1-3: Delete all trajectory points whose speed is within the first speed threshold range; Step S1-4: Delete all trajectory points whose speed is not 0 and whose longitude and latitude have not changed; Step S1-5: Delete all trajectory points with zero speed and abnormal longitude and latitude changes.

3. A method for identifying stop points of urban distribution vehicles based on trajectory data according to claim 2, characterized in that: The step S1-5 comprises: Step S1-5-1: if the speed of the track point is 0, calculate the distance between the track point and the track point before and after the track point, if the distance is greater than the first threshold, delete the track point, otherwise, execute step S1-5-2; Step S1-5-2: Calculate the average speed between the front and rear trajectory points as the speed of the trajectory point.

4. The method for identifying stop points of urban distribution vehicles based on trajectory data according to claim 1, characterized in that: The process of determining the first speed threshold comprises: According to the trajectory point data sequence of the vehicle, the first mixed distribution model based on speed is used for fitting to obtain the speed of the saddle point; The velocity of the saddle point is taken as the first velocity threshold.

5. A method for identifying stop points of urban distribution vehicles based on trajectory data according to claim 4, characterized in that: The mathematical expression of the first mixed distribution model is: Where: f(v; ω1, μ1, σ1, ω2, μ2, σ2) is the first mixed distribution model, v is the velocity, ω1 is the weight of the first log-normal distribution, ω2 is the weight of the first normal distribution, μ1 is the mean of the first log-normal distribution, σ1 is the standard deviation of the first log-normal distribution, μ2 is the mean of the first normal distribution, σ2 is the standard deviation of the first normal distribution, Lognorm(v; μ1, σ1) is, Norm(v; μ2, σ2) is the first normal distribution.

6. A method for identifying stop points of urban distribution vehicles based on trajectory data according to claim 1, characterized in that: The first time threshold is determined by the cumulative probability density of the vehicle stay time.

7. A method for identifying stop points of urban distribution vehicles based on trajectory data according to claim 6, characterized in that: The first time threshold is the stay time at which the cumulative probability density of the vehicle stay time is 80%.

8. The method for identifying stop points of urban distribution vehicles based on trajectory data according to claim 1, characterized in that: In the step S3, after the instantaneous stay points are merged, the instantaneous stay points whose stay time is greater than the first time threshold are screened out as valid stay points; The merging process includes: Step S3-1: Determine whether the distance between the stop point and the previous stop point is less than a second set threshold distance. If yes, execute step S3-2: Step S3-2: Determine whether the stay time of the stay point is less than a second set threshold time. If yes, merge the stay point with its previous stay point.

9. The method for identifying stop points of urban distribution vehicles based on trajectory data according to claim 1, characterized in that: The step S4 comprises: Step S4-1: Eliminate all valid stop points that are on the main road; Step S4-2: for all remaining valid stay points, three points of interest closest to each valid stay point are obtained as associated points of interest for the valid stay point; Step S4-3: extract keywords of the associated points of interest of each valid stop point. If the three associated points of interest of any valid stop point all contain non-freight keywords and none of them contain freight keywords, then the valid stop point is removed as a non-freight stop point.

10. A device for identifying stop points of urban distribution vehicles based on trajectory data, 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 9 is implemented.

Images