Method for calculating road network traffic operation carbon emission by using vehicle trajectory data

Abstract

The invention discloses a method for calculating road network traffic operation carbon emission by using vehicle trajectory data. The method comprises the following steps: firstly, segmenting vehicle GPS trajectory data according to timestamps and serial numbers to obtain a trajectory subset of each vehicle; determining a vehicle stop interval based on the minimum instantaneous speed threshold value, and dividing the trajectory data into driving sections to obtain a driving trajectory; constructing an urban road network model, calculating a spatial distance between a track and a road section, screening candidate road sections according to a shortest distance threshold value, obtaining a matched road section in combination with a minimum difference value of an azimuth angle, mapping track points to the road section, and obtaining re-expanded driving section track data; according to the timestamps and the road section serial numbers, segmenting according to the road section serial numbers to obtain each road section track subset; idle speed and normal driving intervals of a road section are identified through spatial-temporal clustering, a MOVES model database is introduced to perform classified calculation on carbon emissions of the corresponding intervals, the carbon emissions are summarized to obtain road section emissions, and road network traffic carbon emissions are summarized to obtain road network traffic carbon emissions. According to the method, macrotopography and microscopic real-time working condition data are fused, and a road section carbon emission measuring and calculating method is perfected.

Description

Technical Field

[0001] This invention pertains to methods for calculating carbon emissions from roads, specifically a method for calculating carbon emissions from road network traffic operations. Background Technology

[0002] Currently, in order to take effective measures to reduce emissions and protect the environment and human health, there is a great deal of research in the industry on the prediction and monitoring of carbon emissions from motor vehicle driving on road networks. However, current research has some limitations.

[0003] A patent document published on September 1, 2023, with publication number CN116681323A, discloses a method, electronic device, and medium for calculating carbon emissions based on personal travel chains. The method includes: dividing an individual's single trip into travel modes and energy types; calculating a basic carbon emission factor using the energy lifecycle method based on the divided travel modes, energy types, and corresponding energy consumption; dividing an urban area into several grids; calculating the spatiotemporal distribution of vehicle speed-gradient for each grid based on historical traffic data; calculating a carbon emission correction factor based on road construction information corresponding to each grid; calculating the carbon emissions of each segment in the personal travel chain based on the grid and time of each trip, the spatiotemporal distribution of vehicle speed-gradient, and the carbon emission correction factor; and summing the carbon emissions of each segment to calculate the overall carbon emissions of the personal travel chain. This method is hampered by the fragmentation of personal data, the need to match travel modes and energy types one by one, making it cumbersome, unsuitable for various scenarios, and inefficient.

[0004] A patent document published on August 22, 2023, with publication number CN116628395A, discloses a method and system for calculating road vehicle carbon emissions based on urban terrain features. The method includes: a historical carbon emission dataset for the city to be measured, which includes at least road-related information, vehicle-related information, and terrain-related information for the city during a historical statistical period; training a carbon dioxide emission factor determination model using the historical carbon emission dataset; acquiring carbon emission data for the city during a target statistical period; determining the carbon dioxide emission factor for the city during the target statistical period based on the carbon emission data using the carbon dioxide emission factor determination model; and determining the total carbon dioxide emissions for the city during the target statistical period based on the carbon emission data and the carbon dioxide emission factor. This method emphasizes macro-level terrain over micro-level dynamics and relies on historical data, failing to completely overcome the limitations of the time dimension, leading to measurement lag.

[0005] A patent document published on July 23, 2024, with publication number CN118378060A, discloses a method, device, and medium for identifying key features and spatiotemporal patterns of private car carbon emissions. The method includes: identifying private car travel information based on GPS trajectory data to estimate private car carbon emissions; extracting road network features and / or urban function features from geographic information systems and point-of-interest (POI) data; inputting the road network features and / or urban function features into first and second carbon emission prediction models, respectively, using private car carbon emissions as training labels to predict private car traffic carbon emissions during peak and off-peak hours; calculating the SHAP value corresponding to each input feature by sequentially removing input features from the first and second carbon emission prediction models, sorting them, and using the top K1 and K2 input features as key features affecting private car carbon emissions during off-peak hours; clustering the SHAP values ​​corresponding to all input features, and using the resulting discrete clusters as the first spatiotemporal pattern. This method is limited to a single vehicle type, requires dedicated POI data support, cannot cover all types of motor vehicles, and has a narrow scope of application.

[0006] A patent document published on December 6, 2024, with publication number CN119091634A, discloses a method for predicting carbon emissions from motor vehicle driving on road networks. This method divides road networks into segments using intersections as boundaries, collects road network data for each segment, and constructs a road network data model. It then collects real-time dynamic traffic information data on traffic flow, vehicle driving conditions, and vehicle types for each segment, constructs a calculation model for carbon emissions from motor vehicle driving on urban road networks, and calculates the carbon emissions data. Next, it constructs a grouping model for carbon emissions from motor vehicle driving on road networks, obtaining grouped carbon emissions data for urban road networks. Based on the grouped carbon emissions data obtained in step S3, it uses a long short-term memory neural network to construct a prediction model for carbon emissions from motor vehicle driving on road networks, predicting carbon emissions from motor vehicle driving on road networks. However, this method only focuses on time-series prediction and lacks specific consideration of the micro-level operating conditions of vehicles.

[0007] In summary, there is an urgent need to develop a simple method for calculating road carbon emissions that integrates macroscopic terrain parameters with microscopic data such as real-time vehicle operating conditions and covers all types of motor vehicles. Summary of the Invention

[0008] To address the aforementioned existing technologies, this invention provides a method for calculating carbon emissions from road network traffic operations using vehicle trajectory data. This method integrates macroscopic terrain parameters with microscopic data such as real-time vehicle operating conditions (e.g., road network segments, segment slopes, and GPS point data), completely eliminating the limitations of the time dimension. It can cover all types of motor vehicles without requiring dedicated data support, thus broadening the application boundaries of the technology. While retaining high-precision prediction capabilities, it strengthens the closed-loop adaptation of "calculation-analysis-application," using a simple modeling logic to balance real-time performance with the needs of multiple application scenarios, thereby improving the practicality and adaptability of the technology.

[0009] To address the aforementioned technical problems, this invention proposes a method for calculating carbon emissions from road network traffic using vehicle trajectory data. The method mainly includes: sorting vehicle GPS point trajectory data in ascending order by timestamp and vehicle number, and segmenting the data according to vehicle number to obtain a subset of trajectory points for each vehicle; calculating the instantaneous acceleration of the trajectory points for the same vehicle to obtain expanded trajectory point structure data; constructing a low-speed trajectory point dataset for a specified vehicle number using a minimum instantaneous speed threshold, and then using a spatiotemporal clustering method to obtain the dwell intervals of vehicles with that vehicle number; dividing the expanded trajectory point structure data into multiple driving segments based on the time boundaries of the dwell intervals, and extracting driving trajectory data; constructing a road network data model using the urban road network as the research object; and processing the driving trajectory data... The trajectory data is used to calculate the spatial distance between the driving trajectory points and candidate road segments using the shortest distance threshold. The road segment corresponding to the smallest difference between the azimuth angle of each point in the driving trajectory data and the azimuth angle of the candidate road segment is selected as the matching road segment, thereby mapping the driving trajectory points to the candidate road segments and obtaining the expanded driving segment trajectory point structure data. After sorting according to the matching timestamp and road segment number, the road segments are segmented according to their numbers to obtain a subset of trajectory points for each road segment. Subsequently, a spatiotemporal clustering method is used to identify the idling range and normal driving range of the road segments, and the MOVES model database is introduced to achieve trajectory behavior classification. The carbon emissions of vehicles in the corresponding range are calculated, and the overall carbon emissions of the road segment are summarized. The overall carbon emissions of all road segments are then summarized to obtain the carbon emissions of road network traffic operation. The specific steps are as follows:

[0010] Step 1: Sort the vehicle GPS point trajectory data in ascending order by timestamp and vehicle number, and segment it according to vehicle number to obtain a subset of trajectory points for each vehicle. This constitutes the original trajectory point structure data, which includes fields such as vehicle number, vehicle type, timestamp, longitude, latitude, instantaneous speed, and direction of travel. Calculate the instantaneous acceleration for the trajectory point data of the same vehicle, and add this instantaneous acceleration as a new field to the original trajectory point structure data to obtain the expanded trajectory point structure data.

[0011] Step 2: Using the expanded trajectory point structure data of the same vehicle number obtained in Step 1, filter the trajectory points with instantaneous speeds less than 1 km / h to obtain a low-speed trajectory point dataset. Then, perform spatiotemporal clustering on the low-speed trajectory point dataset using the ST-DBASCAN clustering method to obtain a cluster set for the specified vehicle number. This cluster set represents the dwell interval of the vehicle with that vehicle number.

[0012] Step 3: Using the cluster set of the specified vehicle numbers obtained in Step 2, extract the time boundary of each stop interval. The time from the end time of the previous stop interval to the start time of the next stop interval between two adjacent stop intervals is a driving segment. Divide the expanded trajectory point structure data into multiple driving segments through the time boundaries of the stop intervals to obtain driving trajectory data.

[0013] Step 4: Taking the urban road network as the research object, the road segments are divided with intersections as boundaries. Road network data for each road segment is collected, and a road network data model is constructed. The road network data model includes the azimuth and slope of all road segments.

[0014] First, candidate road segments of the driving trajectory data are obtained. The shortest distance to all road segments is calculated for the driving trajectory data. Road segments with a shortest distance within a 100-meter threshold are defined as candidate road segments of the driving trajectory data. Then, the azimuth angle is calculated and the nearest road segment is determined. The azimuth angle of each point of the driving trajectory data is calculated and the difference is made with the azimuth angle of the candidate road segments. The road segment with the smallest difference is defined as the matching road segment of the driving trajectory data point. The number of the matching road segment is added as a new field to the expanded trajectory point structure data to obtain the further expanded driving segment trajectory point structure data, which is denoted as the expanded driving segment trajectory point data.

[0015] Step 5: Sort the extended driving segment trajectory point data of all vehicles obtained in Step 4 by timestamp, then by segment number, and then divide it into segments according to the segment number to obtain a subset of trajectory points for each segment.

[0016] Step 6: Using the subset of trajectory points for each road segment obtained in Step 5, combined with the idle speed threshold, filter the idle trajectory points. Use the ST-DBASCAN spatiotemporal clustering method to perform spatiotemporal clustering on the idle trajectory points to obtain the road segment idle speed interval. Calculate the temporal and spatial complement of the road segment idle speed interval to obtain the normal driving interval of the road segment.

[0017] Step 7: Import the MOVES model database, calculate the idling carbon emissions using the idling range of the road segment combined with the carbon emission factors under different vehicle types' idling conditions, calculate the carbon emissions under different vehicle normal driving conditions using the mapping relationship between the carbon emission factors of the normal driving range of the road segment and the operating mode, and add the idling carbon emissions and the carbon emissions under the normal driving conditions to obtain the overall carbon emissions of the road segment.

[0018] Furthermore, the specific implementation method of step 3 of the method of the present invention includes the following steps:

[0019] Step 3-1) The vehicle number obtained in Step 2 is... The cluster set is represented as:

[0020] ;

[0021] in, It represents the number of dwell intervals and also the number of clusters;

[0022] Step 3-2) Extract the time boundaries of the dwell intervals: from each dwell interval Get start time and end time , in It is the cluster number. ,in yes The earliest timestamp in the middle, yes The latest timestamp;

[0023] Step 3-3) Segment driving points based on the extracted time boundaries of the dwell interval: Using the time boundaries, segment the vehicle... The entire set of trajectory points The non-stop points in the process are divided into multiple driving segments, as expressed by the formula:

[0024] travel section , ; Vehicle number The set of all trajectory points, For any point in the set of trajectory points, The timestamp of any trajectory point. and It is a stay area Time boundary, It is a parking area Time boundaries;

[0025] Steps 3-4) Represent the obtained driving trajectory data as a set of driving segments: Define the set of driving segments. , It represents the number of time intervals for staying.

[0026] Furthermore, in step 4 of the method described in this invention, the process of obtaining candidate road segments for driving trajectory data is as follows: defining... For the number Any trajectory point in the set of vehicle travel segments; , Indicates the number is The set of vehicle travel segments, calculate To all sections shortest distance ; retain satisfaction The road segments are selected as candidate road segments; the set of candidate road segments for each point is denoted as . The process of calculating the azimuth and determining the nearest road segment is as follows: Calculate and Azimuth between The difference between the azimuth angle of the data point and the azimuth angle of all road segments in the road network data model is calculated, and the road segment with the smallest difference is defined as the matching road segment of the driving trajectory data point.

[0027] Furthermore, the specific implementation method of step 6 of the method of the present invention includes the following steps:

[0028] Step 6-1) Filter idling trajectory points: For each road segment number, filter the idling points corresponding to the subset of trajectory points, and set the idling threshold to [value missing]. km / h, the set of idling points for each road segment is denoted as . ;

[0029] Step 6-2) The ST-DBASCAN spatiotemporal clustering method is used to obtain the idling speed range of road segments. This method can accurately capture the characteristics of multiple vehicles operating in the same road segment and at the same time, which is beneficial for accurately estimating carbon emissions. The ST-DBASCAN spatiotemporal clustering method for obtaining the idling speed range of road segments is explained in detail below:

[0030] (1) Spatial distance calculation: Defined as the spherical distance between two trajectory points, the latitude and longitude coordinates of the trajectory points in the dataset are extracted using the Haversine formula. and Calculation, where ,

[0031] ;

[0032] ;

[0033] ;

[0034] ;

[0035] in, For temporary variables to store intermediate values, The radius of the Earth is expressed in meters.

[0036] (2) Calculation of time difference: Defined as the time difference between two trajectory points.

[0037] Unit: seconds; where, , Representing trajectory points respectively timestamp;

[0038] (3) Definition of core parameters:

[0039] Spatial distance threshold, A value of 5 meters represents the spatial error range at the same point of contact.

[0040] The time difference threshold, The value is 40 seconds, used to distinguish between continuous and discrete stops;

[0041] Minimum number of points threshold A value of 5 indicates that at least 5 low-velocity points are required to form a dwell cluster.

[0042] (4) Set of idling points for each road segment The points in the cluster are clustered according to the following rules:

[0043] ① Determining the core points

[0044] If point satisfy: but As the core point;

[0045] ②Density accessibility,

[0046] point From point Density is achievable if a point array exists. Each point is a core point, and adjacent points satisfy a spatiotemporal threshold.

[0047] ③ Cluster formation,

[0048] The road segments formed by mutually accessible points are numbered as follows: A cluster, It is a section of road A cluster set; that is ,in It is the cluster number. ,b;

[0049] Road section number is The set of clusters for the idling range can be represented as:

[0050] ;

[0051] in: It is the road section number. The number of clusters, i.e., the number of clusters.

[0052] Effective idle speed clustering must meet the following requirements: ;

[0053] Step 6-3) Obtaining the normal driving range of the road segment: Take the complement of the idle speed range in time and space to obtain the normal driving range of the road segment.

[0054] Furthermore, in the method described in this invention, step 7 specifically includes the following steps:

[0055] Step 7-1) Calculation of carbon emissions at idle speed:

[0056] Vehicle type is The carbon emission factor of the vehicle under idling conditions is expressed as The unit is grams per second; the specific value is obtained by querying the idle emission rate data for the corresponding vehicle type in the MOVES model database, i.e.: Idle carbon emission rate corresponding to VType in the MOVES model database;

[0057] The set of vehicles involved in the idling range is defined as follows: For any vehicle numbered The vehicle type is ,in The carbon emission factor under idling conditions is Idle time is ,in for The time when the vehicle enters the idle range is when the vehicle is in The difference between the earliest and latest timestamps in the data; then the idling carbon emissions of all vehicles on this road segment are:

[0058] ;

[0059] Step 7-2) Carbon emission calculation under normal driving conditions:

[0060] 7-2-1) Calculate the trajectory points of :

[0061] VSP-driven emission modeling introduces the MOVES model and uses engine specific power (VSP) to characterize the relationship between vehicle emissions and transient driving conditions. VSP integrates influencing factors such as speed, acceleration, and road gradient.

[0062] Vehicle power-to-weight ratio Vehicle specific power is used to characterize the relationship between vehicle emissions and transient driving conditions. The calculation formula is: ;in: The instantaneous speed of the vehicle is expressed in meters per second. Acceleration is expressed in meters per second (m / s·s). The slope of the road section is the ratio of the elevation gain to the horizontal distance. The specific power of the engine is expressed in kilowatts per ton; 1.1 in the formula is the acceleration correction coefficient, which is dimensionless; 9.81 is the gravitational acceleration, which is expressed in meters per second; 0.132 is the equivalent acceleration coefficient for rolling resistance, which is expressed in meters per second; 0.000302 is the comprehensive empirical coefficient for air resistance, which is expressed in l / m.

[0063] 7-2-2) Obtaining the running mode of trajectory points :

[0064] The operating mode (OpMode) is determined by both the speed range and the VSP range. Let... For a certain operating mode carbon emissions (unit: ).

[0065] Speed ​​range determination: For road segment numbered as trajectory points Perform speed unit conversion, speed The unit is kilometers, convert The unit is miles, that is ;

[0066] like The mapping relationship between the operating mode OpMode and the engine specific power VSP is as follows:

[0067] VSP < 0, OpMode is 11; 0 ≤ VSP < 3, OpMode is 12; 3 ≤ VSP < 6, OpMode is 13; 6 ≤ VSP < 9, OpMode is 14; 9 ≤ VSP < 12, OpMode is 15;

[0068] like The mapping relationship between the operating mode OpMode and the engine specific power VSP is as follows:

[0069] When VSP < 0, OpMode is 21; when 0 ≤ VSP < 3, OpMode is 22; when 3 ≤ VSP < 6, OpMode is 23; when 6 ≤ VSP < 9, OpMode is 24; when 9 ≤ VSP < 12, OpMode is 25; when 12 ≤ VSP < 18, OpMode is 27; when 18 ≤ VSP < 24, OpMode is 28; when 24 ≤ VSP < 30, OpMode is 29; when 30 ≤ VSP, OpMode is 30.

[0070] like The mapping relationship between the operating mode OpMode and the engine specific power VSP is as follows:

[0071] VSP < 6, OpMode is 33; 6 ≤ VSP < 12, OpMode is 35; 12 ≤ VSP < 18, OpMode is 37; 18 ≤ VSP < 24, OpMode is 38; 24 ≤ VSP < 30, OpMode is 39; 30 ≤ VSP, OpMode is 40.

[0072] Combining the velocity of each trajectory point and Determine the operating mode by referring to the table. That is, each trajectory point corresponds to a certain operating mode;

[0073] 7-2-3) Obtain the vehicle's operating mode :

[0074] Obtain the running mode obtained in step 7-2-2 by querying the MOVES model database. The specific figures for the emission rates are as follows; For operating mode The emission factor of carbon emissions q, expressed in grams per second;

[0075] The number is The carbon emission formula for vehicles normally traveling on the road section is:

[0076] ;

[0077] in, It is the time difference between two adjacent trajectory points in the original trajectory. ;

[0078] Step 7-3) The carbon emissions at idle speed are added to the carbon emissions under normal driving conditions to obtain the overall carbon emissions of the road section.

[0079] Compared with existing methods for calculating carbon emissions from road network traffic operations, the method of this invention is simpler, integrates macroscopic terrain parameters and microscopic data such as real-time vehicle operating conditions, and achieves road carbon emission calculation covering all types of motor vehicles.

[0080] (1) In this invention, vehicle trajectory data is used to calculate carbon emissions, which overcomes the limitations of data fragmentation and the need to match travel modes and energy types one by one, and breaks through the limitation of a single vehicle type. The modeling process is relatively simple and the calculation efficiency is improved.

[0081] (2) This invention takes into account the identification of vehicles operating in normal driving and idling states on different road sections. By aggregating all trajectory points on the road section to the corresponding road section, the speed of these trajectory points is used to filter idling trajectory points, and then the ST-DBASCAN spatiotemporal clustering method is used to perform spatiotemporal clustering to obtain the idling range of the road section. Therefore, the method of this invention does not have the shortcomings of focusing on macro-terrain and micro-dynamics, or relying on historical data to cause calculation lag. It integrates macro-terrain parameters and micro-data such as real-time vehicle operating conditions to achieve full coverage of all types of motor vehicles.

[0082] (3) The present invention considers the path selection characteristics of vehicles traveling between stopping points when matching vehicle trajectory data with road network. The vehicle stopping interval identification adopts the ST-DBASCAN clustering method. By clustering individual vehicle trajectories, the vehicle stopping interval is identified. This not only supports the calculation of the total carbon emissions of the road network, but also supports the rapid measurement and monitoring of road network carbon emissions under the conditions of changes in road traffic network structure and management methods. Attached Figure Description

[0083] Figure 1 This is a flowchart illustrating the calculation of carbon emissions from road network traffic operations using vehicle trajectory data, as described in this invention. Detailed Implementation

[0084] The design concept for calculating carbon emissions from road network traffic using vehicle trajectory data proposed in this invention is as follows: First, vehicle GPS trajectory data is sorted in ascending order by timestamp and vehicle number, and then segmented according to vehicle number to obtain a subset of trajectory points for each vehicle. Instantaneous acceleration is calculated for the trajectory points of the same vehicle to obtain expanded trajectory point structure data. Then, a low-speed trajectory point dataset is constructed for a specified vehicle number using a minimum instantaneous speed threshold. A spatiotemporal clustering method is then used to obtain the dwell intervals of vehicles with that vehicle number. Based on the time boundaries of the dwell intervals, the expanded trajectory point structure data is divided into multiple driving segments, and driving trajectory data is extracted. Furthermore, a road network data model is constructed using the urban road network as the research object. The driving trajectory data is then compared with... The spatial distance between road segments in the road network is used to filter candidate road segments using the shortest distance threshold. The road segment corresponding to the smallest difference between the azimuth angle of each point in the driving trajectory data and the azimuth angle of the candidate road segments is selected as the matching road segment, thereby mapping the driving trajectory points to the matching road segments and obtaining the expanded driving segment trajectory point structure data. Finally, the data is sorted according to the matching timestamp and road segment number, and segmented according to the road segment number to obtain a subset of trajectory points for each road segment. Subsequently, a spatiotemporal clustering method is used to identify the idling range and normal driving range of the road segment to achieve trajectory behavior classification. The carbon emissions of vehicles in the corresponding range are calculated, and the overall carbon emissions of the road segment are summed to obtain the carbon emissions of the entire road network traffic operation.

[0085] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but the following embodiments are by no means intended to limit the present invention.

[0086] The specific steps of the method for calculating carbon emissions from road network traffic using vehicle trajectory data proposed in this invention are as follows: Step 1: Sort the vehicle GPS point trajectory data in ascending order by timestamp and vehicle number, and then segment it according to vehicle number (first sort by timestamp in ascending order, then by vehicle number in ascending order, and then segment it according to vehicle number) to obtain a subset of trajectory points for each vehicle, which constitutes the original trajectory point structure data including fields such as vehicle number, vehicle type, timestamp, longitude, latitude, instantaneous speed, and driving direction; calculate the instantaneous acceleration for the trajectory point data of the same vehicle. The instantaneous acceleration can be approximated as the first-order difference of velocity with respect to time by the difference between adjacent points. Add this instantaneous acceleration as a new field to the original trajectory point structure data to obtain the expanded trajectory point structure data.

[0087] Step 2: Using the expanded point structure data of the same vehicle number obtained in Step 1, filter the trajectory points with instantaneous speeds less than 1 km / h to obtain a low-speed trajectory point dataset. Then, perform spatiotemporal clustering on the low-speed trajectory point dataset using the ST-DBASCAN clustering method to obtain a cluster set for the specified vehicle number. This cluster set represents the dwell interval of the vehicle with that vehicle number. The result obtained by using the ST-DBASCAN clustering method is common knowledge in this industry.

[0088] Step 3: Using the cluster set of vehicle numbers obtained in Step 2, extract the time boundaries of each stopping interval. This involves obtaining the start and end times from each stopping interval, where the start time is the earliest timestamp within the interval, and the end time is the latest timestamp. The time from the end time of one stopping interval to the start time of the next adjacent stopping interval constitutes a travel segment. The expanded trajectory point structure data is divided into multiple travel segments based on the time boundaries of the stopping intervals. Each travel segment is a continuous sequence of points, containing only non-stopping points, arranged in chronological order, thus obtaining the travel trajectory data, including:

[0089] Step 3-1) The vehicle number obtained in Step 2 is... The cluster set is represented as:

[0090] ;

[0091] in, It represents the number of dwell intervals and also the number of clusters;

[0092] Step 3-2) Extract the time boundaries of the dwell intervals: from each dwell interval Get start time and end time , in It is the cluster number. ,in yes The earliest timestamp in China yes The latest timestamp;

[0093] Step 3-3) Segment driving points based on the extracted time boundaries of the dwell interval: Using the time boundaries, segment the vehicle... The non-stop points in the trajectory point set are divided into multiple travel segments, expressed by the formula:

[0094] travel section ,in ;in Vehicle number The set of all trajectory points, For any point in the set of trajectory points, The timestamp of any trajectory point. and It is a stay area Time boundary, It is a parking area The time boundary. Only includes non-stop points; the time for stop points is within [time range]. Within the range;

[0095] Steps 3-4) Represent the obtained driving trajectory data as a set of driving segments: Define the set of driving segments. , This refers to the number of stopping intervals. Each travel segment is a continuous sequence of points, containing only non-stop points, and arranged in chronological order.

[0096] Step 4: Taking the urban road network as the research object, road segments are divided with intersections as boundaries. Road network data for each segment is collected to construct a road network data model. The road network data model includes the number, azimuth angle, and slope of all road segments. For the driving trajectory data in the driving segment set, the spatial distance between the driving trajectory point and the candidate road segment is calculated using the shortest distance threshold. The road segment corresponding to the smallest difference between the azimuth angle of each point of the driving trajectory data and the azimuth angle of the candidate road segment is taken as the matching road segment, thereby mapping the driving trajectory point to the matching road segment. This process is commonly referred to as map matching in the industry. Adding the matching road segment number to the driving segment trajectory point structure data yields a further expanded driving segment trajectory point structure data; specifically:

[0097] Step 4-4) Obtain candidate road segments for driving trajectory point data: Calculate the shortest distance to all road segments from the driving trajectory data in the driving segment set, and define road segments with a shortest distance within a 100-meter threshold as candidate road segments for driving trajectory data; define... For the number Any trajectory point in the set of vehicle travel segments; , Indicates the number is The set of vehicle travel segments, calculate To all sections shortest distance ; retain satisfaction The road sections were selected as candidates. For a distance of 100 meters, the candidate road segment set for each point is recorded as follows: .

[0098] Steps 4-5) Calculate the azimuth angle of adjacent driving data points: Calculate the azimuth angle point by point for the driving trajectory data, i.e., calculate... and Azimuth between The calculation formula is as follows:

[0099]

[0100] in: (Latitude to radians); (Longitude difference in radians).

[0101] Steps 4-6) Determine matching road segments: For the candidate road segment set Each section of the road , and the azimuth angle obtained in steps 4-5) Find the minimum cyclic difference:

[0102]

[0103] The road segment with the smallest difference is defined as the matching road segment for that driving trajectory data point. This process maps the driving trajectory point to the matching road segment, and is commonly referred to in the industry as map matching. Adding the matching road segment number to the driving segment trajectory point structure data results in the expanded driving segment trajectory point structure data, which will be referred to as expanded driving segment trajectory point data in the following description.

[0104] Step 5: Sort the extended driving segment trajectory point data of all vehicles obtained in Step 4 first by road segment number, and then by timestamp. That is, summarize all trajectories that pass through the same road segment to obtain a subset of trajectory points for each road segment.

[0105] Step 6: Using the subset of trajectory points obtained in Step 5 and the idling speed threshold, filter the idling trajectory points. Use the ST-DBASCAN spatiotemporal clustering method to perform spatiotemporal clustering on the idling trajectory points to obtain the idling speed intervals for each road segment. Calculate the temporal and spatial complements of the idling speed intervals to obtain the normal driving intervals for each road segment; including:

[0106] Step 6-1) Filter idling trajectory points: For each road segment number, filter the idling points corresponding to the subset of trajectory points, and set the idling threshold to [value missing]. km / h, the set of idling points for each road segment is denoted as . ;

[0107] Step 6-2) The ST-DBASCAN spatiotemporal clustering method is used to obtain the idling speed range of road segments. This method can accurately capture the characteristics of multiple vehicles operating in the same road segment and at the same time, which is beneficial for accurately estimating carbon emissions. The ST-DBASCAN spatiotemporal clustering method for obtaining the idling speed range of road segments is explained in detail below:

[0108] (1) Spatial distance calculation: Defined as the spherical distance between two trajectory points, the latitude and longitude coordinates of the trajectory points in the dataset are extracted using the Haversine formula. and Calculation, where ,

[0109] ;

[0110] ;

[0111] ;

[0112] ;

[0113] in, For temporary variables to store intermediate values, The radius of the Earth is expressed in meters.

[0114] (2) Calculation of time difference: Defined as the time difference between two trajectory points.

[0115] Unit: seconds; where, , Representing trajectory points respectively timestamp;

[0116] (3) Definition of core parameters:

[0117] Spatial distance threshold, A value of 5 meters represents the spatial error range at the same point of contact.

[0118] The time difference threshold, The value is 40 seconds, used to distinguish between continuous and discrete stops;

[0119] Minimum number of points threshold A value of 5 indicates that at least 5 low-velocity points are required to form a dwell cluster.

[0120] (4) Set of idling points for each road segment The points in the cluster are clustered according to the following rules:

[0121] ① Determining the core points

[0122] If point satisfy: but As the core point;

[0123] ②Density accessibility,

[0124] point From point Density is achievable if a point array exists. Each point is a core point, and adjacent points satisfy a spatiotemporal threshold.

[0125] ③ Cluster formation,

[0126] The road segments formed by mutually accessible points are numbered as follows: A cluster, It is a section of road A cluster set; that is ,in It is the cluster number. ,b;

[0127] Road section number is The set of clusters for the idling range can be represented as:

[0128] ;

[0129] in: It is the road section number. The number of clusters, i.e., the number of clusters.

[0130] Effective idle speed clustering must meet the following requirements: ;

[0131] Step 6-3) The normal driving range of a road segment is the complement of the idling range in time and space. That is, the area that is not clustered as idling is the normal driving range of the road segment. Taking the complement of the idling range in time and space is the normal driving range of the road segment.

[0132] Step 7: Introduce the MOVES model database, use the road segment idling range combined with the carbon emission factors under different vehicle types idling conditions to calculate idling carbon emissions, use the mapping relationship between the carbon emission factors of the road segment normal driving range and the operating mode to calculate the carbon emissions of different vehicles under normal driving conditions, and add the idling carbon emissions and the carbon emissions under normal driving conditions to obtain the overall carbon emissions of the road segment.

[0133] The carbon emission factor represents the amount of carbon emissions per unit time or unit of mileage. This study uses the carbon emission factor per unit time, which varies depending on the vehicle type. In practical applications, specific values ​​can be obtained by querying the idle emission rate data for the corresponding vehicle type in the MOVES model database.

[0134] Idle carbon emission rate corresponding to VType in the MOVES database.

[0135] Step 7 includes the following specific steps:

[0136] Step 7-1) Calculation of carbon emissions at idle speed:

[0137] Vehicle type is The carbon emission factor of the vehicle under idling conditions is expressed as The unit is grams per second; the specific value is obtained by querying the idle emission rate data for the corresponding vehicle type in the MOVES model database, i.e.: Idle carbon emission rate corresponding to VType in the MOVES model database;

[0138] The set of vehicles involved in the idling range is defined as follows: (Deduplicated vehicle IDs), for any vehicle ID as The vehicle type is ,in The carbon emission factor under idling conditions is Idle time is ,in for The time when the vehicle enters the idle range is when the vehicle is in The difference between the earliest and latest timestamps in the data; then the idling carbon emissions of all vehicles on this road segment are:

[0139] The carbon emissions obtained The unit is grams (g).

[0140] Step 7-2) Carbon emission calculation under normal driving conditions:

[0141] 7-2-1) Calculate the trajectory points of :

[0142] VSP-driven emission modeling introduces the MOVES model and uses engine specific power (VSP) to characterize the relationship between vehicle emissions and transient driving conditions. VSP integrates influencing factors such as speed, acceleration, and road gradient.

[0143] Vehicle power-to-weight ratio Vehicle specific power is used to characterize the relationship between vehicle emissions and transient driving conditions. The calculation formula is: ;in: The instantaneous speed of the vehicle is expressed in meters per second. Acceleration is expressed in meters per second (m / s·s). The slope of the road section is the ratio of the elevation gain to the horizontal distance. The specific power of the engine is expressed in kilowatts per ton; 1.1 in the formula is the acceleration correction coefficient, which is dimensionless; 9.81 is the gravitational acceleration, which is expressed in meters per second; 0.132 is the equivalent acceleration coefficient for rolling resistance, which is expressed in meters per second; and 0.000302 is the comprehensive empirical coefficient for air resistance, which is expressed in l / m.

[0144] 7-2-2) Obtaining the running mode of trajectory points :

[0145] The operating mode (OpMode) is determined by both the speed range and the VSP range. Let... For a certain operating mode carbon emissions (unit: ).

[0146] Speed ​​range determination: For road segment numbered as trajectory points Perform speed unit conversion, speed The unit is kilometers, convert The unit is miles, that is ;

[0147] The mapping relationship between operating mode OpMode and engine specific power VSP is shown in Table 1. The judgment number is... trajectory points of the road segment The speed range to which it belongs, The mapping relationship between the operating mode OpMode and the engine specific power VSP is shown in the left column of Table 1. The mapping relationship between the operating mode OpMode and the engine specific power VSP is shown in the middle column of Table 1. The mapping relationship between the operating mode OpMode and the engine specific power VSP is shown in the right column of Table 1.

[0148] Table 1. Mapping relationship between operating mode (OpMode) and engine specific power (VSP)

[0149]

[0150] Combining the velocity of each trajectory point and By combining the corresponding speed columns, find Determine the operating mode by referring to the table for the corresponding interval. That is, each trajectory point corresponds to a certain operating mode. And speed mph corresponding =12.

[0151] 7-2-3) Obtain carbon emissions from vehicles traveling normally on a certain road segment:

[0152] Obtain the running mode obtained in step 7-2-2 by querying the MOVES model database. The specific figures for the emission rates are as follows; For operating mode carbon emissions The emission factor is expressed in grams per second.

[0153] The number is The carbon emission formula for vehicles normally traveling on the road section is:

[0154] ;

[0155] in, It is the time difference between two adjacent trajectory points in the original trajectory. ;

[0156] Step 7-3) The carbon emissions at idle speed are added to the carbon emissions under normal driving conditions to obtain the overall carbon emissions of the road section.

[0157] Step 8: Sum the total carbon emissions of the road segments obtained in Step 7 to obtain the final carbon emissions of the road network traffic operation.

[0158] In summary, based on objective natural laws, this invention utilizes vehicle trajectory data to calculate carbon emissions generated by traffic operations on a road network. Vehicle trajectory data is collected from onboard positioning data and includes information such as position and speed, thus possessing objectivity. Actual measurement data shows that carbon emissions vary significantly across different road gradients and operating conditions. This invention establishes the relationship between vehicle driving status and road segments / operating states through vehicle trajectory data. It possesses measurement objectivity, does not require subjective factors set by humans, and can accurately calculate the total emissions of vehicles across the entire road network.

[0159] Although the present invention has been described above in conjunction with the accompanying drawings, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many improvements and changes under the guidance of the present invention without departing from the spirit of the present invention, and these improvements and changes are all within the protection scope of the present invention.

Claims

1. A method for calculating road network traffic operation carbon emissions using vehicle trajectory data, characterized in that, the vehicle GPS point trajectory data is sorted in ascending order according to the timestamp and vehicle number in turn and segmented according to the vehicle number, obtaining the trajectory point subset of each vehicle; the instantaneous acceleration of the trajectory points of the same vehicle is calculated, obtaining the extended trajectory point structure data; a low-speed trajectory point data set is constructed for the specified vehicle number with the minimum instantaneous speed threshold, and then a space-time clustering method is used to obtain the stay interval of the vehicle with the specified vehicle number; the extended trajectory point structure data is divided into multiple driving segments according to the time boundaries of the stay interval, and the driving trajectory data is extracted; a road network data model is constructed based on the urban road network as the research object; the spatial distance between the driving trajectory data and the road segments of the road network is calculated, and the candidate road segments are selected with the shortest distance threshold, and the road segment corresponding to the minimum difference between the azimuth angle of the driving trajectory point and the azimuth angle of the candidate road segment is taken as the matching road segment, so as to map the driving trajectory point to the matching road segment and obtain the re-extended driving segment trajectory point structure data; the matching road segments are sorted according to the timestamp and vehicle number in turn, and then segmented according to the road segment number, obtaining the trajectory point subset of each road segment; then, the space-time clustering method is used to identify the idling interval and the normal driving interval of the road segment respectively, the MOVES model database is introduced, and the trajectory behavior classification is realized; the carbon emissions of the vehicle corresponding to the interval are calculated, and the overall carbon emissions of the road segment are obtained; the overall carbon emissions of all road segments are summarized to obtain the road network traffic operation carbon emissions.

2. The method for calculating road network traffic operational carbon emissions using vehicle trajectory data according to claim 1, wherein, The specific steps of the method are as follows: Step 1, the vehicle GPS point trajectory data is sorted in ascending order according to the timestamp and vehicle number in turn and segmented according to the vehicle number, obtaining the trajectory point subset of each vehicle, which constitutes the original trajectory point structure data including vehicle number, vehicle type, timestamp, longitude, latitude, instantaneous speed, and driving direction fields; the instantaneous acceleration of the trajectory point data of the same vehicle is calculated, and the instantaneous acceleration is added as a new field to the original trajectory point structure data to obtain the extended trajectory point structure data; Step 2, using the extended trajectory point structure data of the same vehicle number obtained in step 1, the trajectory points with instantaneous speed less than 1 km / h are selected to obtain the low-speed trajectory point data set, and the ST-DBASCAN clustering method is used for space-time clustering to obtain the clustering set of the specified vehicle number, which is the stay interval of the vehicle with the specified vehicle number; Step 3, using the clustering set of the specified vehicle number obtained in step 2, the time boundaries of each stay interval are extracted, and the end time of the previous stay interval to the start time of the next stay interval of the adjacent two stay intervals is a driving segment; the extended trajectory point structure data is divided into multiple driving segments by the time boundaries of the stay interval, obtaining the driving trajectory data; Step 4, taking the urban road network as the research object, the road segments are divided at the intersection boundaries, the road network data of each road segment is collected, and a road network data model is constructed, which includes the azimuth angle and road slope of all road segments; First, obtain the candidate road section of the driving trajectory data, calculate the shortest distance from the driving trajectory data to all road sections, and define the road section within the 100-meter threshold as the candidate road section of the driving trajectory data; Then, calculate the azimuth and determine the nearest road section, calculate the azimuth of the driving trajectory data point by point and subtract the azimuth of the candidate road section, and define the road section with the smallest difference value as the matching road section of the driving trajectory data point, and add the number of the matching road section as a new field to the extended trajectory point structure data to obtain the re-extended driving section trajectory point structure data, denoted as extended driving section trajectory point data; Step 5, sort all the extended driving section trajectory point data of the vehicle obtained in step 4 according to the timestamp, then sort according to the road section number, and segment according to the road section number to obtain the trajectory point subset of each road section; Step 6, screen the idling trajectory points using the trajectory point subset of each road section obtained in step 5 in combination with the idling threshold, and perform spatial-temporal clustering on the idling trajectory points using the ST-DBASCAN spatial-temporal clustering method to obtain the road section idling interval; obtain the road section normal driving interval by taking the complement of the road section idling interval in time and space; Step 7, introduce the MOVES model database, calculate the idling carbon emission using the road section idling interval in combination with the carbon emission factor under the idling working condition of different vehicle types, calculate the carbon emission under the normal driving state of different vehicles using the mapping relationship between the carbon emission factor and the operation mode of the road section normal driving interval, and add the idling carbon emission and the carbon emission under the normal driving state to obtain the overall carbon emission of the road section; Step 8, aggregate and sum the road section overall carbon emission to finally obtain the entire road network traffic operation carbon emission.

3. The method for calculating road network traffic operational carbon emissions using vehicle trajectory data according to claim 1, wherein, The specific implementation method of step 3 includes the following steps: Step 3 - 1) Represent the cluster set of vehicle numbers obtained in step 2 as: ​ ； wherein, is the number of stay intervals, and is the number of clusters. Step 3-2) Extracting time boundaries of stay intervals: obtaining start time and end time from each stay interval , where is the cluster number, where is the earliest timestamp of , and is the latest timestamp of ; Step 3-3) Segmentation of driving points according to the time boundary of the extracted dwell interval: using the time boundary, the non-dwell points in the whole trajectory point set of the vehicle are segmented into multiple driving segments, which is expressed by the formula: ​ driving segment , ; vehicle number all trajectory point sets, any one point in the trajectory point set, the timestamp of any one trajectory point. and the time boundary of the stay interval , the time boundary of the parking interval ; Step 3-4) represent the obtained driving trajectory data as a driving section set: Definition of a set of travel segments , is the number of stay intervals.

4. The method for calculating road network traffic operational carbon emissions using vehicle trajectory data according to claim 1, wherein, In step 4, The process of obtaining the candidate road segments of the driving trajectory data is as follows: defining any one trajectory point in the driving segment set of the vehicle numbered ; , representing the driving segment set of the vehicle numbered , calculating the shortest distance to all road segments ; retaining the road segments satisfying as candidates, with a value of 100 meters, and recording the candidate road segment set of each point as ; The process of calculating the azimuth and determining the nearest road segment is as follows: Calculate and Azimuth between The difference between the azimuth angle of the data point and the azimuth angle of all road segments in the road network data model is calculated, and the road segment with the smallest difference is defined as the matching road segment of the driving trajectory data point.

5. The method for calculating road network traffic operational carbon emissions using vehicle trajectory data according to claim 1, wherein, The specific implementation method of step 6 includes the following steps: Step 6-1) Screening idle speed trajectory points, screening idle speed points for each road section numbered corresponding trajectory point subset, idle speed threshold value taken as km / h, obtaining idle speed point set of each road section recorded as ; Step 6-2) obtain the road section idling interval using the ST-DBASCAN spatial-temporal clustering method, (1) Spatial distance calculation: defined as the spherical distance between two trajectory points, the latitude and longitude coordinates of the trajectory points in the dataset are extracted using the Haversine formula and calculated, where , ； ； ； ； wherein is a temporary variable for storing an intermediate value, is the radius of the earth in meters; (2) Time difference calculation: defined as the time difference between two trajectory points, , in seconds; wherein, , respectively represent the timestamps of the trajectory points ; (3) Core parameter definition: a spatial distance threshold, is 5 meters; is a time difference threshold, is 40 seconds; minimum point threshold value, is 5; (4) Set of idling points for each road segment The points in the cluster are clustered according to the following rules: ① Core point determination, If the point satisfies: then is a core point; ② Density accessibility, point from point density is reachable if there exists a point sequence , where each point is a core point and adjacent points satisfy the spatio-temporal threshold ③ Cluster formation, The road segments formed by mutually accessible points are numbered as follows: A cluster, It is a section of road A cluster set; that is ,in It is the cluster number. ,b; The set of idle intervals clustered by road segment number may be represented as: ； wherein: is a road segment number is a number of clusters, The effective idle speed clustering needs to meet: ; Step 6-3) road section normal driving interval acquisition: the complement of the idling interval in time and space is the road section normal driving interval.

6. The method for calculating road network traffic operational carbon emissions using vehicle trajectory data according to claim 1, wherein, The specific implementation method of step 7 includes the following steps: Step 7-1) idling carbon emission calculation: The vehicle type is The carbon emission factor of the vehicle under idle operating condition is expressed as , unit: g / s; the specific value is obtained by querying the idle emission rate data of the corresponding vehicle type in the MOVES model database, that is: The idle carbon emission rate corresponding to VType in the MOVES model database; The vehicle set involved in the idling section is defined as: For any vehicle numbered , the vehicle type is , wherein , the carbon emission factor under the idling condition is , and the idling time is , wherein is The time of entering the idling section is the difference between the earliest timestamp and the latest timestamp of the vehicle in ; and the idling carbon emission of all vehicles on the section is: ； Step 7-2) normal driving state carbon emission calculation: Step 7-2-1) Calculating the Trajectory Point of : Vehicle specific power The vehicle specific power is a formula used to characterize the relationship between the emissions of a motor vehicle and the transient driving conditions The formula is: ; wherein: V is the instantaneous speed of the vehicle in meters per second; a is the acceleration in meters per second squared; s is the slope of the road segment, being the ratio of the vertical distance to the horizontal distance; P is the engine specific power in kilowatts per ton; 1.1 in the formula is an acceleration correction factor, dimensionless; 9.81 is the acceleration due to gravity in meters per second squared; 0.132 is the rolling resistance equivalent acceleration coefficient in meters per second squared; and 0.000302 is the air resistance comprehensive empirical coefficient in 1 / meter. Step 7 - 2-2) Acquiring a running mode of the trajectory point : Speed interval determination: the speed unit conversion is performed on the track point with the road section number , and the speed in kilometers is converted into the speed in miles, i.e. ;​ If , the mapping relationship between the operating mode OpMode and the specific engine power VSP is as follows: VSP<0, OpMode is 11; 0≤VSP<3, OpMode is 12; 3≤VSP<6, OpMode is 13; 6≤VSP<9, OpMode14; 9≤VSP<12, OpMode is 15; If , the mapping relationship between the operating mode OpMode and the specific engine power VSP is as follows: VSP < 0, OpMode is 21; 0 ≤ VSP < 3, OpMode is 22; 3 ≤ VSP < 6, OpMode is 23; 6 ≤ VSP < 9, OpMode is 24; 9 ≤ VSP < 12, OpMode is 25; 12 ≤ VSP < 18, OpMode is 27; 18 ≤ VSP < 24, OpMode is 28; 24 ≤ VSP < 30, OpMode is 29; 30 ≤ VSP, OpMode is 30; If , the mapping relationship between the operating mode OpMode and the specific engine power VSP is as follows: VSP < 6, OpMode is 33; 6 ≤ VSP < 12, OpMode is 35; 12 ≤ VSP < 18, OpMode is 37; 18 ≤ VSP < 24, OpMode is 38; 24 ≤ VSP < 30, OpMode is 39; 30 ≤ VSP, OpMode is 40; combining the speed and determining the operating mode from a table i.e. each trajectory point corresponds to an operating mode; Step 7 - 2 - 3) Acquiring the operation mode of the vehicle : the emission rates for the pollutants q for the operating mode obtained in step 7-2-2) by querying the MOVES model database; for the operating mode emission factors for the pollutants q in grams per second; The vehicle carbon emission formula for normal driving of the road segment numbered is ； wherein is the time difference between two adjacent trajectory points in the original trajectory points; ​ Step 7-3) the idle carbon emission and the carbon emission in the normal driving state are added to obtain the overall carbon emission of the road section.

Images